UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Chinampa agriculture : the operation of an intensive pre industrial resource system in the Valley of… Robertson, Alastair J. 1983

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


831-UBC_1983_A8 R62.pdf [ 13.85MB ]
JSON: 831-1.0095765.json
JSON-LD: 831-1.0095765-ld.json
RDF/XML (Pretty): 831-1.0095765-rdf.xml
RDF/JSON: 831-1.0095765-rdf.json
Turtle: 831-1.0095765-turtle.txt
N-Triples: 831-1.0095765-rdf-ntriples.txt
Original Record: 831-1.0095765-source.json
Full Text

Full Text

CHINAMPA AGRICULTURE: THE OPERATION OF AN INTENSIVE PRE INDUSTRIAL RESOURCE SYSTEM IN THE VALLEY OF MEXICO by ALASTAIR J . ROBERTSON B.A., The University of British Columbia, 1976 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF ARTS in THE FACULTY OF GRADUATE STUDIES (Department of Geography)  We accept this thesis as conforming to the required standard  THE UNIVERSITY OF BRITISH COLUMBIA October 1983 ©  Alastair J. Robertson  In p r e s e n t i n g  this thesis  r e q u i r e m e n t s f o r an of  British  it  freely available  agree that for  Library  s h a l l make  for reference  and  study.  I  f o r extensive copying of  h i s or  be  her  copying or  f i n a n c i a l gain  g r a n t e d by  publication  s h a l l not  be  Date  (3/81)  of  further this  Columbia  thesis  head o f  this  my  It is thesis  a l l o w e d w i t h o u t my  G&OCrgflPHV  The U n i v e r s i t y o f B r i t i s h 1956 Main Mall V a n c o u v e r , Canada V6T 1Y3  the  representatives.  permission.  Department o f  University  the  s c h o l a r l y p u r p o s e s may  understood that  the  the  I agree that  permission by  f u l f i l m e n t of  advanced degree a t  Columbia,  department or for  in partial  written  ABSTRACT  During the 14th and 15th Centuries the complex of marshes and shallow lakes that occupied the low elevations of the Valley of Mexico were thoroughly modified by the construction of thousands of small agricultural platforms known as chinampas, and of an elaborate network of dykes, canals and aqueducts.  A quantified model of the hydrological  systems of the Valley indicates that under natural conditions water levels fluctuated to such an extent that the productivity of chinampa platforms would have been limited.  The larger hydraulic installations  smoothed these fluctuations in the upstream portions of the lake complex and so permitted the agricultural exploitation of these areas.  On the  platforms seedbedding and other intensive horticultural techniques were employed which allowed the maximum utilization of the controlled hydraulic environment.  Canals were exploited not only for irrigation water  but also as reservoirs for plant nutrients, which promoted a conservative use of materials in the system as a whole.  Communities of non-  crop plants were maintained on the platform border and these protected crops from climatic extremes, and may also have been important in regulating populations of insect pests.  The hydraulic installations were  managed by high-ranking officials of the Aztec state, while the chinampa platforms were to a large extent managed by their cultivators. Although managerial decisions were made by individuals of widely different rank, there was a tendency for decision-making to occur at the lowest hierarchic level compatible with the ability of the manager to command the  ii  labour and materials necessary to implement the decision.  This allowed  managers to respond rapidly and efficiently to small variations in the conditions of the physical systems that they managed.  The available  data indicate that the chinampas produced high and sustained yields per unit area and did so without requiring large subsidies of energy and materials.  The physical, ecological and managerial principles of  chinampa agriculture are therefore relevant to the design of modern agricultural systems that seek to exhibit these general properties.  iii  TABLE OF CONTENTS Page ABSTRACT  ii  LIST OF TABLES  vii  LIST OF FIGURES  ix  ACKNOWLEDGEMENT  xi i i  CHAPTER I.  INTRODUCTION  CHAPTER II. CHAPTER III.  1  THE HYDROLOGY OF THE BASIN OF MEXICO  9  THE COLONIZATION OF THE LAKE COMPLEX  44  Chinampa Agriculture without Water Level Control  44  Chinampa Agriculture and Water Level Control in Limited Areas  59  Chinampa Agriculture and Water Level Control in Large Areas  65  Late Aztec Hydraulic Installations CHAPTER IV.  THE ECOLOGY OF CHINAMPA AGRICULTURE  The Chinampa Platforms  73 92 92  Flora and Fauna  101  Seedbeds and Transplanting  120  Weeding and Pest Control  126  Material Cycles  130  Agricultural Calendar  136  iv  Paje Yields and Energy Ratios  145  The Principles of Chinampa Agriculture  150  CHAPTER V. THE MANAGEMENT OF CHINAMPA AGRICULTURE The City State  158  The Confederation and the Tribute State  162  The Lake Aztec State  166  The Construction of Hydraulic Installations  172  The State and its Hydraulic Installations  178  The Management of Agriculture  183  The Evaluation of Aztec Management  188  CHAPTER VI. SUMMARY AND CONCLUSIONS APPENDIX.  RECONSTRUCTION OF THE 16TH CENTURY HYDROLOGY OF THE BASIN OF MEXICO  Part I.  211  The Use of Modern Data in Reconstructing 16th Century Conditions  Part II.  198  212  Derivation of Mean Annual Values for the 16th Century Hydrological System  222  1.  Drainage Basins  222  2.  Physiographic Regions  222  3.  Precipitation  222  4.  Evapotranspiration  223  5.  Agricultural Evaporation Supplement  224  v  Page 6.  Surface Discharge  225  7.  Undifferentiated Runoff  226  8.  Lake Discharge  226  9.  Mean Annual Totals  226  Estimation of Mean Monthly Distributions  229  1.  Monthly Distribution of Precipitation  229  2.  Monthly Distribution of Surface Discharge  230  3.  Primary and Secondary Evaporation  233  4.  Monthly Distribution of Evaporation  238  5.  Monthly Distribution of Surplus  238  6.  Monthly Distribution of Undifferentiated Runoff  244  7.  Lake Tlahuac and Lake Chalco  251  Estimates of Variation in Lake Stage  262  1.  Assumptions  262  2.  Estimated Variations of Lake Stage with Normal  Part III.  . Part IV.  Precipitation 3.  262  Estimates of Variations of Lake Stage with Precipitation at Exceptionally High or Low Values  BIBLIOGRAPHY  264 266  vi  LIST OF TABLES Page TABLE I.  Absolute Chronology and Period Designations for the Valley of Mexico  TABLE II.  45  Widths of Canals in the Chinampas of San Luis and San Gregorio  99  TABLE III.  Common Aquatic Plants of the Chinampas  103  TABLE IV.  Edible and Medicinal Plants of the Chinampas  108  TABLE V.  Common Highland Mesoamerican Crops  112  TABLE VI.  Consumptive Use of Water in Centimeters for Localities at Low Elevations in the South of the Valley of Mexico  132  TABLE VII.  Nutrient Content of Organic Fertilizers  134  TABLE VIII.  Aztec Agricultural and Ritual Calendars  139  TABLE IX.  Yields of Shelled Corn, Mixquic  146  TABLE X.  Yields of Saleable Portions of Selected Vegetables, Mixquic  TABLE XI.  147  Summary of Estimates of 16th Century Mean Annual Water Balance, Basin of Mexico  TABLE XII.  Stations Used in Calculating Monthly Values of Climatic Variables  TABLE XIII.  227  231  Primary Evaporation as a Proportion of Total Evaporation  237  vii  Stations Used in Estimation of Distribution of Secondary Evaporation Estimated 16th Century Mean Monthly Water Balance, Basin of Mexico Assumed Areas and Elevations of the Antique Lakes  viii  LIST OF FIGURES No. 1.  Page The watershed of the Basin of Mexico in the 16th Century  2.  10  Estimated annual water transfers in the Basin of Mexico during the 16th Century  3.  13  Subsidiary drainage basins and physiographic regions in the Basin of Mexico in the 16th Century  15  4.  Mean annual isohyets in the Basin of Mexico, 1920-1959  17  5.  Mean monthly precipitation averaged for seven stations throughout the Basin of Mexico  19  6.  Average mean isotherms in the Basin of Mexico, 1920-1958  20  7.  Average monthly temperatures, Moyoguarda, 1951-1975  21  8.  Average monthly potential evaporation, Moyoguarda,  9. 10.  1951-1975  21  Estimated monthly variation in stage of Antique lakes  21  Average minimum isotherms in the Basin of Mexico, 1920-1958  11.  22  Maximum, mean and minimum temperatures, Moyoguarda, 1951-1975  24  12.  Frequency of frosts, Moyoguarda, 1951-1975  25  13.  Mean adjusted potential evaporation in the Basin of  14.  Mexico  27  Daily discharge of Rio Amecameca, 1947  30  ix  No.  Page  15.  Schematic section of spring in Sierra Ajusco  32  16.  Annual precipitation at Tacubaya, 1878-1975  36  17.  Variability and precipitation for selected stations in the Basin of Mexico  18.  37  Annual values for adjusted potential evaporation and precipitation at Moyoguarda, 1951-1975  19.  Estimated variations in stage of Lake Chalco for varying annual precipitation values  20.  50  Estimated variations in the extent and location of the inundated margins of Lake Chalco  21.  40  52  Length to area ratios of hypothetical dykes in Lakes Xochimilco and Chalco  64  22.  Chinampa platforms in the vicinity of Xochimilco  67  23.  Chinampas of San Gregorio Atlapulco, on the southern shore of Lake Xochimilco  69  24.  Lake Aztec hydraulic installations  74  25.  Reconstructions of 16th Century aqueducts  76  26.  The effects of dyking on normal variations in the stage of the Central and Southern Lakes  27.  Estimated normal variations in the stage of the Southern and Central Lakes with water levels controlled  28.  81  84  Estimated interannual variations of stage in the controlled lake complex, without centralized management  x  85  No. 29.  Page Estimated interannual variations of stage of the controlled lake complex, with centralized management  30.  Schematic representation of nutrient transport in chinampa agriculture  31.  137  Principal settlements of the Valley of Mexico during the 16th Century  32.  89  159  Changing lake levels and precipitation in the Basin of Mexico, 1600 B.C. to present  213  33.  Geological sections of the Basin of Mexico  218  34.  Location of meteorological stations in the Basin of Mexico  35.  232  Precipitation and discharge for Xochimilco and Chalco basins  36.  234  Water balance diagrams for selected stations in the Basin of Mexico  37.  240  Surplus and discharge for the Basin of Rios Magdalena and Esclava, Basin of Mexico  38.  242  Surplus and Discharge for Xochimilco and Chalco basins, Basin of Mexico  39.  243  Estimated mean monthly discharge of Rios de la Compana and San Francisco, Valley of Mexico  40.  245  Estimated mean monthly discharge of Rio Amecameca, Chalco basin  246  xi  No. 41.  Page Observed yields of springs in the Xochimilco and Chalco basins, 1903-04  42.  248  Graphs used for estimates of monthly distributions of infiltrated flows (I) and indirect flows (J)  43.  Estimated monthly distribution of surplus and runoff, Chalco basin  44.  250  250  Observed stage of Lake Tlahuac and estimated inputs of Lake Chalco  261  xii  ACKNOWLEDGEMENT  I would like to thank the many individuals who have helped me in the course of this work.  In particular I would like to thank Susan  Penner and Susan Leslie for their unfailing encouragement, and Irene Hull for preparing the manuscript under difficult conditions.  I would  like to thank Dr. Alfred H. Siemens and Dr. Michael Church who have been enormously supportive of my work: much of what may be of value in this thesis stems from their good counsel, while I alone am responsible for its shortcomings.  My special thanks are also due to Alfred  Siemens and to Rick Clark for their personal concern and great efforts on my behalf:  without them I could not have completed the work and I  am extremely grateful to them.  Finally, I would like to thank those  inhabitants of the chinampa villages who made me welcome and who instructed me in the intricacies of their ancient and beautiful landscape.  xi i i  1  CHAPTER I  INTRODUCTION  The industrialized countries have been remarkably successful in developing agricultural systems with high returns on investments of scarce inputs.  In the United States, where labour has historically  been in short supply, the productivity of labour in agriculture has increased enormously in the last century:  in 1880 about five people  could be supplied with food and fibre from the work of a single agricultural labourer, while today the figure is close to 80.  At the same time  there has been a large increase in the output for each unit input of land:  since 1950 the productivity of land planted to wheat in the  United States has increased twofold, and that of corn by a factor of 2.4.  Similar increases have been accomplished in Western Europe and  other developed regions of the world (Rasmussen 1982:80). These gains in productivity have been achieved by the massive substitution of machinery for labour, the extensive use of artificial fertilizers and biocides, and by the development of varieties of crop species that thrive in the controlled conditions of modern agriculture. In order to produce these materials and transport them to the fields, to maintain and operate machinery and to apply agricultural chemicals large investments of energy are required.  It is true that compared  to the total amount of energy consumed in the developed nations, energy use in agriculture is quite small, amounting in the United States to only 2.8% of the total consumption, and in Western Europe to 4.9%. Where energy is in adequate supply these investments in agriculture are  2  well justified (F.A.O. 1977:95).  However, there are two major problems  associated with the practice of subsidising agriculture with large amounts of commercial energy.  First, the cost of energy is increasing  and, to the extent that it is doing so, the high levels of use in agriculture are becoming increasingly unacceptable.  In the capital-  short regions of the world this factor is already critical, and the peasant farmers who constitute the majority of the world's agriculturalists are simply unable to pay for the energy that is required to develop the productivity of their agricultural systems.  Second, although the  yields of agriculture in the developed countries have risen with increasing investments of energy, the rate of increase in production has not kept pace with the rate of increase of energy use.  In the opinion of  some specialists the current line of development of Western agriculture has reached the point of diminishing returns, and some alternative line of development is required if agricultural production is to continue to increase (Kinne & McClure 1977:24). For both these reasons, there is a growing interest in certain aspects of pre-industrial agriculture, in particular, in the fact that such systems tend to be more efficient in their use of energy than are those of the developed nations.  As an example, maize production in the  United States returns about 3? units of energy for each unit of energy invested, while maize production by traditional methods in Mexico and Guatemala returns about 30 units of energy for each one invested (F.A.O. 1977:93; Fluck & Baird 1980:47).  However, any advantage that these  traditional systems might offer in terms of their efficient use of energy are offset by the fact that their returns per unit investment of both land and labour are generally quite low.  The traditional systems of  3  Middle America produce about ? ton/hectare, when fallow land is included in the calculation, and something of the order of 1 to H tons per agricultural worker per year (Redfield & Villa 1934:52, 55; Stadelman 1940:103, 147, 149; Steggerda 1941:118; Villa 1945:60, 61). Commercial corn production in the United States, in contrast, produces about 5 tons/hectare, and about 10 times that much per agricultural labourer (F.A.O. 1977:95). Low yields are characteristic of many traditional systems, but not of all of them.  Some intensive pre-industrial systems have not only  the relatively high energy efficiency that is generally characteristic of traditional systems, but also yields per unit area that compare favourably with those of modern agriculture.  It is systems of this  type that are of particular interest because they provide indications of how modern agriculture might be developed in an environment of scarce or expensive energy resources. It is one such system, chinampa cultivation in the Valley of Mexico, that is the subject of the present thesis.  "Chinampa" is an  indigenous Nahautl term that denotes a small rectangular platform, constructed of soil and decayed vegetable matter, typically about 10 meters by 100 meters and surrounded on three or four sides by canals or ditches.  The sides of the platforms are held in place by living willow  trees, and the surfaces are intensively cultivated.  Thousands of these  chinampas were constructed by the Aztecs and their predecessors in the Valley of Mexico during the 14th and 15th Centuries, together with the complex hydraulic installations that acted to maintain the water in the canals and ditches at a more or less constant level.  The chinampas  yielded between three and four metric tons of maize per hectare, as well  4  as a significant quantity of horticultural products, and these yields compare favourably with those of modern agriculture. Furthermore, chinampa cultivation had an estimated energy efficiency of between 10:1 and 15:1 which, although lower than the figures for other traditional systems, is still significantly larger than the 3.5:1 efficiency of modern practices (derived from Parsons 1976:245). This thesis is intended to explain how the Aztecs contrived to produce high yields and, at the same time, to make an efficient use of energy in chinampa agriculture.  A part of the answer is sought in  purely technical aspects of the system, and the hydrology and ecology of the chinampas are discussed at length.  However, the question cannot be  resolved solely in technical terms, and certain aspects of Aztec social and political organization are also considered, with a view to elucidating the role of management in the accomplishments of chinampa agriculture. The frame of reference is broad, including elements from both sides of the traditional division between "man" and "land", but it is not unbounded. The scope of the thesis is limited by considering the chinampas as a physical-cultural resource system with the essential function of transforming the "natural stuff" in the environment of a culture into materials that are required or desired by the participants in that culture (Zimmerman 1964:17).  Thus, although the discussion  impinges upon certain aspects of the physical landscape of the Valley of Mexico, it is restricted to those aspects that have a direct bearing on the production of crops in the chinampas.  Similarly, discussion of  the social, political and religious life of the Aztecs is limited to those aspects that were directly related to agriculture.  5  The chinampa system falls quite readily into three major subsystems, namely, the lake complex of the Valley of Mexico and the hydraulic installations by means of which its function was modified; the ecological system by means of which nutrients and water available in the lake complex were transformed into useful agricultural products, and a management system by means of which the structure of the two physical subsystems was modified and their functions controlled. Each -  of these three subsystems is treated in turn, and in each case the structure of the subsystems is- described and the efficiency with which they performed their functions is evaluated. There are numerous problems with the systemic approach to the study of interactions between a culture and its physical surroundings. There are few well established research procedures, and the models and constructs that are employed have not been as thoroughly tested by experience as would be desirable.  The purview includes subject matter  that has traditionally been the province of both the sciences and the arts, and there are difficulties in combining the languages,, methodologies and theories of these disparate approaches within a single study. However, there is one major advantage to the systemic approach that more than outweighs these disadvantages.  The traditional formulation  for the discussion of any instance of man's use of the land has been to conceive of the two entities as being distinct and externally related, and to study the relations between the two.  As Geertz has pointed out,  this formulation must finally result in the gross question of the extent to which each entity conditions the function of the other, and elicits the equally gross response: (Geertz 1971:1-11).  "To a degree, but not completely"  The systemic approach overcomes this problem by  6  conceiving of any instance of man's use of the land not as a relationship between two separate entities, but as a single entity which contains both physical and cultural elements, which has a coherent organization, and which performs a discernible function.  The advantage of this formu-  lation over the traditional one is that the question of how a complex system such as an agricultural system works can be rendered quite incisive.  By conceiving of the combined physical and cultural system as a  single entity one can ask such relatively precise questions as  "How is  the system organized?"; "To what extent and in what manner is it stable?"; "To what degree is it efficient in its use of energy and other inputs?" or "What mechanisms regulate the functioning of the system?".  The present  paper asks these questions of the chinampa system of cultivation. A great deal of the material that follows concerns the centralized and bureaucratic apparatus of the Aztec state and its role in the function of chinampa agriculture, and this obviously has a relevance to the discussion of Wittfogel's theories of "oriental despotisms" (Wittfogel:1957). However, in the present work, although note is taken of the work of Wittfogel and other researchers in this domain, the question of whether or not the state owes its origins to the development of hydraulic agriculture is not pursued. Nor in fact could this issue be resolved with reference to chinampa agriculture alone.  The chinampas were not con-  structed until quite late in the history of the Valley of Mexico, when many of the institutions of the state were already in place.  If  the question was to be pursued with reference to the Valley of Mexico, then the discussion should include the development of Teotihuacan and flood-water irrigation and, quite probably, the development of drained field agriculture in the Lowlands of Mexico as well.  Some aspects of  7  this large issue are relevant to an understanding of the function of chinampa agriculture and are discussed where appropriate, but the larger issue itself is not treated explicitly in the present work. The chinampas are discussed as they existed in the late 15th and early 16th Centuries, in the period immediately prior to the Spanish Conquest of Tenochtitlan in 1521.  Evidence for their condition at this  time is drawn from the few small areas of debilitated chinampas that s t i l l persist in the Valley of Mexico; from those modern data on the physical condition of the Valley that can be extrapolated back to the 16th Century with some assurance; from the practices of contemporary Indian agriculturalists in Mexico and Guatemala; from early Spanish colonial documents and from ethnohistorical and archaeological sources. There are many problems with these sources.  The extant chinampas are  particularly frustrating as a source of information on the 16th Century chinampas, because although one knows that some of the forms and practices that can be observed today are the same as ones that existed in the 16th Century, it is often difficult to distinguish truly indigenous features from those which have developed since the Spanish Conquest. Modern data have been used extensively in the reconstruction of the 16th Century hydrology of the Valley, but because the Valley has undergone considerable physical modification in the past half millennium modern data are frequently unrepresentative of past conditions.  The problem  of using these modern data is discussed at length in the appendix to this thesis.  The other sources of information on pre-Hispanic condi-  tions are frequently unreliable or fragmentary, and on some aspects of chinampa agriculture we have virtually no information at a l l .  However,  the combination of modern climatic dates of historical records of  8  drainage and land use patterns, and of assumptions based on these records has allowed a quantified and consistent reconstruction of 16th Century hydrological conditions.  There is sufficient information on  agricultural practices to allow a broad description of the ecology of the chinampas, and, although the sources on Aztec management practices are poor, they are at least sufficient to sketch in the main characteristics of the organization and administration of chinampa agriculture. The available sources provide a sufficient basis for a description of the structure and function of chinampa agriculture, and an explanation of how the chinampas worked.  9  CHAPTER II  THE HYDROLOGY OF THE BASIN OF MEXICO  The chinampas were constructed in the complex of lakes and marshes that occupied the low elevations of the Basin of Mexico. In order to understand the problems of constructing and operating the chinampas and their associated hydraulic installations, it is necessary to understand something of the behaviour of the lake complex and of the hydrological system of which they were a part. The Valley of Mexico is located on the Mexican Mesa Central, which is an immense uplifted sedimentary block, split with fractures and faults.  Since the Lower Tertiary Era, igneous materials have been  extruded through these lines of weakness, covering the fault block, and dividing it into a number of more or less distinct compartments, of which the Basin of Mexico is one (Figure T).  It is separated from the  Lerma Valley to the east by the Sierra de Las Cruces, and from the Puebla Valley to the west by Sierras Rio Frio and Nevada. The south of the Valley is bounded by Sierra Ajusco, and the north by low ranges of discontinuous hills.  These Sierras are all the products of vulcanism  that has occurred sporadically since the Mid Tertiary, and that continues to the present.  Their peaks are generally composite volcanoes, while  their foothills and lower slopes are formed of overlapping and interpenetrating layers of andesites, basalts and volcanic ash.  The floor of  the Valley has filled with alluvial materials, consisting largely of bentonitic clays with intercalated layers of ash and lava.  10  Figure 1.  The Watershed of the Basin of Mexico in the 16th Century (C.H.C.V.M.1964; Sanders, Parsons, Santley 1979:Map 20).  11  The Valley is a closed drainage basin, with a total area close to 9,600 square kilometers.  The peaks that define its boundary are  generally between 3,000 and 4,000 meters above sea level, but are slightly lower in the north, and in the southeast they reach a maximum of 5,452 meters in the peak of Popocatepetl.  Above the 2,700 meter  contour approximately, the Basin is truly mountainous, with steep slopes and thin and stony soils.  Below the 2,700 meter contour the land is  generally hilly, but with several quite large peripheral valleys, including the Avenidas de Pachuca in the north, the Teotihuacan Valley, and numerous smaller valleys and pockets of level ground. At the 2,300 meter contour there is an abrupt change in relief as the foothills give way to the alluvial plains of the main Valley floor.  These  decline to a minimum elevation of 2,237 meters above sea level, and comprise about 20 percent of the area of the Basin. The hydrology of the Basin has been enormously modified since the 16th Century.  It now drains through a breach in its northern wall into  the Rio Lerma system, but at the time of the Conquest it was an entirely closed drainage basin.  Water that was precipitated on the surrounding  mountains and hills drained towards the lowlying central areas where it was stored in a complex of lakes and marshes until it was lost from the system by evaporation.  The main floor of the Basin was occupied  by six more or less distinct bodies of water, known collectively to the Aztecs.as the Lakes of the Moon.  These lakes were at slightly differ-  ent elevations, with Lakes Zumpango and Xaltocan in the North, and Lakes Chalco and Xochimilco in the South, all draining into Lake Texcoco in the centre of the Basin.  The Lake of Mexico in the western  12  part of Lake Texcoco was formed artificially by the construction of the 16 kilometer dyke known as the Albarradon de Nezahualcoyotl.  All of  the.lakes lost water by evaporation, but Texcoco did so at a rather greater rate than the others and because it was never flushed through with fresh water it became quite saline. Figure 2 is a schematic diagram that summarizes the main characteristics of the hydrology of the Basin of Mexico.  It shows estimates  of the volumes of water that in the course of a normal year during the late 15th and early 16th Centuries passed through the Basin. The seven major divisions of the chart correspond to the seven major drainage basins that together constitute the Basin of Mexico, and these are mapped in Figure 3.  These basins are natural watersheds except that  the Mexico-Texcoco basin has been divided into a western and an eastern section, corresponding with the artificial separation of the Lake of Mexico from Lake Texcoco.  Inputs of precipitation (P) are indicated on  the left side of individual basins, and outputs by evaporation (E) on the right.  Each watershed is divided into three physiographic regions,  each of which had quite distinct drainage and evaporation characteristics: mountains (m), valleys (v) or lakeshore plains (p), and lakes, both extant and reconstructed (1) are distinguished.  The flow of water  between these various watersheds and physiographic regions is represented by lines whose thicknesses are directly proportional to the volumes of water transmitted in the course of a normal year.  The separate basins  were linked in the manner shown except that some seepage may have occurred between Lakes Chalco and Xochimilco, on one hand, and Lake Texcoco, on the other, and that the Aztec hydraulic installations allowed the discharge of water from Lake Xochimilco either directly into Lake  C h a l c o  X o c h i m i l c o  B a s i n  70U  Qm 36 29  658  Rv 353  Qm 12 8  336  628  Qv U l 17 J0|N\  Rv 323  Ev H i.86 , 769 Sgw,l  ©H  30 655  N o r t h e a s t e r n  B a s i n s  El 10 h ©  U 513  P a c h u c a  1282  B a s i n  M e x i c o  T e x c o c o  Figure 2.  Q Pach. 252  PI  o f  B a s i n  Estimated Annual Water Transfers in the Basin of Mexico during the 16th Century.  For Legend and Notes, see page following.  Q N.E. 156 El 53 7757  L a k e  Rm 250 775  511  Qv 3 5  Sgw,l  t h e  363  Rm M5  Pv 326 576  o f  Em 516  Pm 778 547  Pm +>! 868  Pv 370  B a s i n  B a s i n  C u a u t i t l a n  B a s i n  14  Figure 2:  Legend and Notes  Symbols (+)  Inputs of water  Basin Boundaries  0  Outputs of water  Direction of Flow  123  Volume in m x 10  123  Depth per unit area in mm  3  6  Large Volume Storage  Line thickness proportional to volume of water. Variables P  Precipitation  E  Evapotranspiration  Q  Surface Runoff  A  Consumptive use of water in agriculture  R  Subsurface and Undifferentiated Runoff  Physiographic Regions m  mountains  P  lakeshore plains  v  valleys  1  lakes  Notes 1)  For Precipitation (P), Evapotranspiration (E) and, in the Xochimilco and Chalco basins only, Surface Runoff (Q). The values shown are directly confirmed by C.H.C.V.M. publications for basins and physiographic regions.  2)  For Surface Runoff (Q) in the remaining five basins, the distribution is directly continued by C.H.C.V.M. publications for basins but not for physiographic regions.  3)  The value for the agricultural evaporation supplement is available only for the Basin of Mexico as a whole, and is distributed between subsidiary basins and physiographic regions in a manner not directly continued by C.H.C.V.M. publications.  4)  The values shown for Runoff (R) are inferred, and are not directly continued by C.H.C.V.M. publications.  5)  The sources used, and the means by which values for the diagram were calculated, are described in Appendix Part II.  15 '99"W  Figure 3.  Subsidiary Drainage Basins and Physiographic Regions of the Basin of Mexico in the 16th Century (modified from C.H.C.V.M. 1964 V I I I : 333).  16  Texcoco, or via the Lake of Mexico. The chart is based on modern hydrological and climatic data for the Valley of Mexico, and the problems associated with the use of these data to describe past conditions are discussed in Part I of the Appendix. Part II of the Appendix gives an account of the calculations employed in producing the chart. The Basin is located between 19° and 20° north, and its climate is essentially tropical, but modified by the effects of elevation. During the summer months it is supplied with water from the Gulf of Mexico, transported by the North East Trades, and precipitated by orographic and convectional lifting mechanisms.  During the winter,  the Gulf is cooler, and the air passing over it is more stable, with the result that the Trades are deflected to the south by the mountain barrier, and the Valley receives relatively little moisture. Such precipitation as does occur during the winter months is the product of frontal lifting mechanisms, caused by occasional inclusions of cold polar air.  The distribution of precipitation within the Valley is shown  in the map of mean annual isohyets (Figure 4).  Values range from less  than 500 millimeters to more than 1300 millimeters, with a general tendency for the lower figures to occur in low elevations towards the north of the Valley, and the higher figures at high elevations in the south.  This distribution is clearly reflected in the chart, Figure 2,  where it is evident that the bulk of the water entering the system occurs as precipitation on the mountains, and that, on a unit area basis, the mountains in the south receive rather more precipitation than those in the North. Precipitation shows a strong seasonal distribution, which is typical of tropical climates, and this is evident in the plot of mean  Figure 4.  Mean Annual Isohyets in the Basin of Mexico, 1920-1959 (C.H.C.V.M. 1964 VIII:243).  18  monthly values, averaged for a number of stations throughout the Basin (Figure 5).  The frontal lifting mechanisms of the winter months produce  very little of the year's total.  Values increase steadily through the  Spring, until May, when large volumes of warm, moist air arrive from the Gulf, and the ground is warm enough for convectional lifting mechanisms to operate.  Values are high from June through to September, during  which four months a little over 70 percent of the year's total precipitation occurs.  The rainy season is interrupted by a brief, relatively  dry spell during August, which is known locally as the "veranillo" or "little summer", and which is common throughout eastern and central parts of Mexico. This phenomenon is related to an interruption of the Trade Winds by the expansion of low pressure cells lying off the eastern seaboard of the United States.  During September the sun moves southwards,  the Gulf cools, the Trades cease to ascend the mountains, and precipitation begins its decline to low winter values. As would be expected at these elevations, temperatures are rather lower than is normal in the Tropics.  The average mean isotherms show  values ranging from more than 16° to less than 7° (Figure 6).  Again,  the distribution is related to relief and latitude, and temperatures are generally lower at high elevations, and towards the south of the Valley.  Temperatures fluctuate in the course of the year with a cycle  that corresponds approximately with the passage of the sun overhead, but with the summer peaks smoothed somewhat by the effects of the rainy season (Figure 7).  Mean minimum temperatures in the Valley quite  regularly fall below freezing (Figure 10).  The low values at high  elevations are accounted for by altitude, but they also occur at low elevations when pockets of cold, stable air are trapped on the Valley floor.  19  M  M  J  J  0  Figure 5. Mean Monthly Precipitation, averaged for 7 stations throughout the Basin of Mexico (C.H.C.V.M. Boletfn Hidrologico).  N  D  Figure 6.  Average Mean Isotherms in the Basin of Mexico, 1920-1958 (C.H.C.V.M. 1964 V I I I : 2 4 9 ) .  21  Figure 8. Average Potential Evaporation, Moyoguarda, 1951-1975 (C.H.C.V.M. Boletin Hidrologico n.d.). E  2241  M  Figure 9.  A  M  J  J  A  S  O  N  Estimated Variations in Stage of Antique Lakes (Appendix).  D  Figure 10.  Average Minimum Isotherms in the Basin of Mexico, 1920-1958 (C.H.C.V.M. 1964 VIII:253).  23  Temperatures in the Valley also vary from year to year.  Figure 11  shows data from Moyoguarda, and it can be seen that maximum and minimum temperatures have a range of 7° and 7.5°, respectively, and that mean temperatures do not vary by more than 3° for the period shown. The seasonal distribution of temperatures also varies from year to year, and because maize is a crop that is particularly susceptible to frost damage, minimum temperatures are of considerable importance.  Figure 12 shows  the frequency of freezing temperatures during the winter months, again at Moyoguarda, and it can be seen that frosts are virtually certain during January and February, that there is a better than even chance that they occur in November and March, and that occasionally they occur as early as October or as late as April. Water is transferred from the ground to the atmosphere either by evaporation from exposed surfaces or by transpiration from growing plants.  The two processes together are known as evapotranspiration.  Because the transfer is diffuse, and because it is determined by a large number of factors relating to the capacity of the atmosphere to absorb water, the properties of the evaporating surfaces and vegetation from which water is lost and the availability of water it is extremely difficult to make estimates of the rates at which water is actually lost by evapotranspiration.  The procedure that has proved most satisfactory  in the Basin of Mexico involves the calculation of actual evapotranspiration rates from daily precipitation and potential evaporation data. Precipitation is employed as an indicator of the amount of water available for evaporation.  Potential evaporation is conceived of as the  volume of water that would be evaporated i f water was freely available, and it is approximated as the volumes lost from standard evaporating  Figure 11.  Maximum, Mean and Minimum Temperatures, Moyoguarda, 1951-1975 (C.H.C.V.M. Boletfn Hidrologico n.d.).  100 90 ~  80  CD  § 70 60 50 o c  Q) Z3  cr CO 1_  LL  40 30 20 10 0  Figure 12.  0  N  D  J  M  Frequency of Frosts, Moyoguarda, Basin of Mexico 1951-1975 (C.H.C.V.M. Boletin Hidrologico n.d.). S/  26  pans.  The evaporation climate of the mountains and the valleys of the  Basin are quite different, and the C.H.C.V.M. employs different methods for estimating actual evapotranspiration losses in the two regions.  In  the present reconstruction of 16th Century conditions, C.H.C.V.M. estimates of actual evapotranspiration losses have been retained in the mountains throughout the Basin, in the valleys of the Pachuca and Northeastern basins and in the Teotihuacan Valley.  In those areas of the  Basin that were occupied by the complex of lakes and marshes modern estimates of actual evapotranspiration losses have been replaced with values for an adjusted potential evaporation rate, which is employed by the C.H.C.V.M. as an approximation of actual evapotranspiration losses from areas of open water and saturated soils.  In addition to these  various categories of evaporation, the C.H.C.V.M. makes use of an agricultural evaporation supplement, intended as an expression of the consumptive use of water in agriculture.  On the grounds that Aztec  agriculture, although less profligate in its use of water than modern irrigated agriculture, was more widespread, this supplement has been retained.  The procedures employed both by the C.H.C.V.M. and the  present writer are described more fully in the Appendix. Figure 13 shows mean annual values for the adjusted potential evaporation rate used as an approximation of losses from open water and saturated soils. far,  Like the other climatic parameters discussed so  it is evident that evaporation rates are strongly influenced by  both elevation and latitude:  values are lowest at high elevations in  the south of the Basin, and highest at low elevations in the centre and north of the Basin, reaching a maximum of more than 1500 millimeters in the areas of Lake Texcoco and Lakes Zumpango and Xaltocan. Actual  27  Figure 13. Mean Annual Adjusted Potential Evaporation in the Basin of Mexico.  Note: Adjusted Potential Evaporation is equal to 70% of Potential Evaporation (C.H.C.V.M. 1964 V I I I : 2 6 3 ) .  28  evapotranspiration shows a similar pattern of distribution, but the range of values is rather greater.  In the mountains where temperatures  are low and drainage is effective in removing water values are between 270 and 425 millimeters, with the higher values in the north.  Values  in the lowlying areas of the Basin are generally between 800 and 1000 millimeters, but reach a maximum of more than 1500 millimeters in the small areas of open water that remain in the Northeastern basins (C.H.C.V.M. 1964 VIII:333). Evaporation rates vary seasonally and Figure 8 (page 21 above) shows mean monthly potential evaporation rates at Moyoguarda, a station at low elevations in the Xochimilco basin.  Values are at a minimum  during the winter months when both temperature and precipitation values are low, and they increase through the spring as temperatures increase. Peak values occur in April and May, when temperatures are quite high, but before the rainy season has begun.  During the rainy season evapora-  tion rates remain high but, because the air is moist and has a low absorption capacity, they do not increase in proportion to the increase in temperature.  In September and October both precipitation and tempera-  ture values decline, and evaporation rates fall to their low winter values. The effects of these evaporation losses on the distribution of water within the Basin are evident in Figure 2.  The Pachuca and North-  Q  eastern basins receive about 2.3 x 10 cubic meters of precipitation in the course of a normal year but, as about 90% of this volume is lost by local evaporation, only modest volumes drain southwards into Lakes Zumpango and Xaltocan.  These volumes are augmented by drainage off the  slopes of the Sierra de las Cruces, in the Cuautitlan basin but, because  29  evaporation rates are high in Lakes Zumpango and Xaltocan, virtually all of the incoming volumes are lost, and a meagre 13 x 10 cubic meters drain southwards into Lake Texcoco.  The Chalco, Xochimilco  q  and Mexico basins receive about 2.4 x 10 cubic meters of precipitation per year, which is only slightly larger than the figure for the two northern basins, but evaporation losses are lower, and about 25% of this volume drains west, and then north, into Lake Texcoco. The mountains and hills of the Basin are drained by numerous rivers and streams that converge on the Valley floor.  None of these  watercourses are large, and most descend through narrow, elongated watersheds as directly to the Valley floor as the tortured topography of the region allows.  Figure 14 shows daily discharge data for Rio  Amecameca, which drains the slopes of Popocatepetl in the extreme southeast of the Valley.  There is very little discharge during the  winter months, when precipitation values are low, and in the spring the flow is intermittent.  Precipitation during the summer months is  convectional and the discharge data show a series of distinct hydrographs, each one of which marks the passage of runoff from an individual storm.  The catchment area of Rio Amecameca is large compared to other  watercourses in the Basin, and the frequency of storms within the catchment area is relatively high. individual storms overlap, out the rainy season.  Because of this the hydrographs of  and Rio Amecameca discharges water through-  In addition, there may be a small baseflow  resulting from snowmelt on Popocatepetl, but this is not evident in this graph.  In smaller watersheds where the frequency of storms is lower,  and where there is no discernible baseflow, discharge tends to be intermittent during the rainy season, and to cease altogether during the dry season.  31  It is evident from Figure 2 that measured surface discharge accounts for only a small part of the volumes of water that drained from the mountains.  Large areas of the Basin are composed of fractured  basalts and andesites which are extremely porous, and in these areas most of the drainage is subterranean.  Many streams lose water by  seepage, others disappear entirely into fissures and cracks, and large volumes of water percolate into the underlying rocks without entering channeled flow at a l l .  In the Valley today these infiltrated volumes  are extracted by mechanical pumping, but in the past they seeped into aquifers in the Valley floor, or supplied the numerous springs that occurred along the margins of the lake complex.  Springs were particu-  larly common along the western shores of the Lake of Mexico, and along the southern shores of Lakes Chalco and Xochimilco. One of the latter is shown schematically in Figure 15.  Measurements made in 1903 and  1904, before the aquifers were entirely depleted by pumping, show that some of the larger springs had minimum yields of more than 800 l.p.s. (Marroquin y Rivera 1914:151-54). The subsurface hydrology of the Basin has been enormously modified by pumping, and modern data are of little value in describing preHispanic conditions.  The present reconstruction makes the assumption  that volumes of groundwater stored in the Valley remained approximately constant over long periods, and from this it follows that all of the water that percolated into aquifers at higher elevations sooner or later made its way by seepage or springflow to the Valley floor, and became available for loss by evaporation.  This is discussed more fully in  Part I of the Appendix, but the assumption is made that infiltration represented a shift of water within the Basin, and not a loss from it.  Figure 15.  Schematic Section of Spring in Sierra Ajusco (C.H.C.V.M. 1964 1:27).  33  This assumption is carried through in Figure 2, where all of the water available in the mountains (defined as Pm - Em - Am) is shown draining into the lakes (1), lakeshore plains (p) or valleys (v). The transfer occurs either by surface discharge (Qm) or by undifferentiated runoff (Rm).  For the whole of the Basin, surface discharge  accounts for only 19% of the water that drains off the mountains, and values for subsidiary basins range from 46% in the Mexico basin, where the underlying rocks are relatively impermeable, to less than 7% in the combined Xochimilco and Chalco basins, where infiltration capacities are extremely high.  The remaining volumes are transferred  by undifferentiated runoff (Rm), which is defined as the difference between the volumes of water available in the mountains and the volumes transferred by surface discharge (Rm = Pm - Em - Am - Qm). Because undifferentiated runoff is defined in this way, it includes volumes of water that drain on the surface in overland flow or in channels that are too small to merit metering, as well as volumes that drain beneath the surface.  Although the proportion of subsurface flow  to the total flow of undifferentiated runoff cannot be determined with any precision, the indications are that it is substantial and that a large part of the drainage of the Valley occurs beneath the surface. This characteristic of the hydrology of the Basin is important because it affects lake recharge patterns.  Subsurface flows are very much  slower than those that occur on the surface and, to the extent that the lakes were recharged by subsurface flows, inputs would have responded in a very dampened and delayed fashion to variations in precipitation. Water from the hillsides drained towards the centre of the Basin and accumulated in the Lake complex.  The supply of water varied from  34  season to season and so, under natural conditions, the levels of the lakes also fluctuated.  By means of the calculations described in the  Appendix, Part III, it has been possible to produce a simple model of the normal fluctuations of inputs and outputs of the lake complex and, from this, to make a rough estimate of normal seasonal variations in stage.  These estimates are shown for Lakes Chalco, Xochimilco and  Mexico-Texcoco in Figure 9 (page 21 above).  Because so great a propor-  tion of the inputs of the lake complex were in the form of infiltrated flows, and because these flows are assumed to have lagged behind precipitation, fluctuations in stage were quite out of phase with fluctuations in precipitation.  During the spring, when evaporation rates and inputs  from the preceding year's precipitation were declining, lake levels declined, and during the summer at the height of the current year's rainy season, lake levels were at a minimum.  At the end of the rainy  season evaporation rates declined and water from the current year's precipitation began to accumulate in the lakes, with the result that stage increased.  During the winter, when evaporation rates were at  their lowest, inputs peaked, and lake levels reached their maximum. Once within the lake complex water that was not lost by evaporation from the upstream lakes drained towards Lake Texcoco.  The Chart  clearly indicates that most of the water precipitated on the Northern basins was lost to the system either by local evaporation or by evaporation from Lakes Zumpango and Xaltocan, and that in normal years, only small volumes were discharged into Texcoco.  In contrast, relatively  little of the precipitation in the Southern basins was lost by evaporation and large volumes of water drained off the slopes of Sierras Nevada, Ajusco and Las Cruces into the Southern lakes, and were subsequently  35  discharged into Lake Texcoco.  With its relatively large area and with  evaporation rates estimated in excess of 1600 millimeters per annum, an estimated 777 million cubic meters of water were lost annually from the latter lake, and the loss was sufficient to balance the budget of this closed hydrological system. Thus far the climate and hydrology of the Basin have been discussed with reference only to mean values but in fact the climate of the Basin, and the hydrological variables that are derived from i t , differ considerably from year to year.  Figure 16 shows annual precipi-  tation figures for Tacubaya. The mean value for the period shown is 699 millimeters, but values in individual years range from a low of 397 millimeters, or 57% of the mean, to a high of 1098 millimeters, or 157% of the mean.  Tacubaya is in no way exceptional in this respect.  Figure 17 illustrates a moderate positive correlation between mean annual precipitation and precipitation variability.  Not only does the  annual rainfall of the Basin vary considerably from year to year, but the variation tends to be greater where precipitation values are high. The variation is not entirely random, as is evident in the 10year overlapping mean for the Tacubaya figures, which show a trend towards a recurring cycle of about 50 years.  These relatively wet and  dry spells correspond with irregularities in the behaviour of the Thermal Equator, which are not entirely understood but which modify the behaviour of the North East Trades, and so modify precipitation in the Valley. Even although these longer term fluctuations have some effect in reducing the differences between extreme values in short periods, consecutive years can have quite large differences in annual values.  The year  1957, for example, was a particularly dry one with precipitation only  37 13001  1100 E  JE c o  9001 Q. O  x  x X  700  500  100  200  Standard  Figure 17.  300  Deviation  Variability and Precipitation, for selected stations in the Basin of Mexico (calculated from C.H.C.V.M. Boletfn Hidrologico n.d.).  400  38  72% of the mean, while the following year was unusually wet, with a value of 149% of the mean. The variability of precipitation can be expressed as the probability of a given percentage of the mean actually occurring in any individual year.  These values have been calculated for the central  highlands of Mexico for different precipitation classes (Wallen 1955:81). The mean annual precipitation for the entire area of the Basin of Mexico is close to 700 millimeters, and WalleVs calculations indicate that in any individual year there is a 46% probability that precipitation will actually equal the mean, a 9% probability that precipitation will equal 71% of the mean, and a 5% probability that it will equal 125% of the mean.  In other words, precipitation over long periods can be expected  to equal mean values about once every other year, but to be about 25% above or below mean values about once every six or seven years. Precipitation varies not only with respect to the total volumes precipitated within the year, but also with respect to the time of year at which it occurs.  If the rainy season is defined as comprising those  months during which precipitation is equal to or greater than 10% of the year's total, then at Tacubaya it normally lasts for four months. But in 6% of years, it lasts for three months or less, in 30% of years it lasts for five months, and in 9% of years it lasts for six months. There is no clear correlation of a long rainy season with high annual precipitation:  both 1957, which was a dry year, and 1975, which was a  wet one, had rainy seasons of six months.  The starting date of the  rainy season is critical to some aspects of agriculture, and this too is quite variable.  Data from 15 stations in the south of the Basin  show that in 48% of years the rainy season, as defined, starts in June,  39  but it starts in May in 30% of years, in July in 15% of years, and occasionally it starts as late as August, or as early as April (C.H.C.V.M. Boletin Hidrologico n.d.). Actual evapotranspiration rates appear to vary significantly from year to year, but the characteristics of the variation are quite different in different physiographic regions.  The lowlying areas of  the Basin are poorly drained, and because water is freely available it is the absorption capacity of the atmosphere that is likely to limit evaporation losses.  In these areas, actual rates can approach potential  rates, and these are likely to be higher in warm dry years than in cool moist ones. confirm this.  Some actual evapotranspiration data are available that The data are for 1957 and 1958, which were exceptionally  dry and wet years, respectively (C.H.C.V.M. 1964 V:257, 259).  In the  valleys, lakeshore plains and lakes there was a difference of 13% of the mean annual value between the two years, and losses were higher in the dry year than in the wet one.  Some data for the adjusted potential  evaporation rate that is used to estimate losses from marshes and lakes are also available, and are shown in Figure 18, together with annual precipitation values.  The evaporation values, with a standard deviation  of 165.5 vary more widely about the mean than do precipitation values, which have a standard deviation of 133.3.  The two sets of data are  positively correlated, but with a coefficient of 0.21, the correlation is extremely weak.  In the area of the lake complex actual evapotrans-  piration losses are approximated by the adjusted potential rate shown in the figure, and in other lowlying regions of the Basin the actual rates may be quite close to this adjusted potential rate.  If  Moyoguarda is at all representative of conditions on the valley floor,  Figure 18.  Annual values for adjusted potential evaporation and precipitation at Moyoguarda, 1951-1975 (C.H.C.V.M. Boleti'n Hidrologico n.d.).  41  then the figures indicate that actual evapotranspiration losses may vary quite widely from year to year, and that the variation is unrelated to variations in precipitation. The situation in the mountains is very different.  Drainage is  efficient in these areas, and it is likely to be the availability of water rather than the absorption capacity of the atmosphere that limits evapotranspiration losses.  Under these circumstances, losses would be  higher in wet years than in dry ones, and this is confirmed by the data for 1957 and 1958 when there was a difference of 21% of the mean between the two years and losses were consistently highest during the wet year. Data are not available for prolonged periods, but the indications are that actual evapotranspiration losses in the mountains may vary considerably from year to year, and that the variability is positively related to variability in precipitation. The relationship between annual values of precipitation and of actual evapotranspiration is important because it is the difference between these two that determines the volumes of water available for soil moisture recharge and runoff.  There is, however, very little data  available and it is not possible to describe the relationship in any detail or with an assurance.  The data available for 1957 and 1958 show  that over the entire area of the Basin there was very little difference between total actual evapotranspiration losses in the two years, and amounted to 2.5% of the mean.  In the calculations of annual variability  in lake stage that are described in the appendix, evaporation rates have been held constant and precipitation has been taken as the source of variability of inputs to the lake complex.  42  The inputs to the lakes may have been quite highly variable, but this would not have resulted in correspondingly large variations in lake stage.  The lakes occupied shallow, saucer shaped depressions  that in the present reconstruction are modeled as shallow spherical segments (Appendix, Part IV). As a consequence of this geometry, there would have been a tendency for any increase in lake inputs to be translated into increases in the area of the lake, as well as increases in stage.  Both stage and area would have been more responsive to variations  in inputs when water levels were low than they would have been when water levels were high.  During dry periods, relatively modest inputs would  have caused the area of the lake to expand, and stage to increase, but during wet periods, when the area of the lakes was large, any input volumes would have been distributed over the whole of the lake, and increases in stage would have been quite small.  At the same time, the  outputs of the lake were proportional to stage.  Any increases in input  that resulted in an increase of stage would also have resulted in an increase in output, and in all the lakes except Texcoco, this too would have acted to reduce the variability of lake stage relative to the variability of inputs.  These mechanisms would have dampened fluctuations  in inputs, but they would not have eliminated them.  Calculations carried  out in the Appendix indicate that there was a significant seasonal variation in the stage of the lakes, and a rather greater interannual variation. The chinampas were constructed in the complex of lakes and marshes that occupied the lowlying areas of the Valley of Mexico, and the problems of constructing and operating the chinampas were intimately related to the behaviour of the lake complex and of the hydrological system of which it was a part.  Virtually all aspects of the hydrological system  43  had some consequence for the development and operation of the chinampas, but two features were of fundamental importance.  First, inputs of water  varied considerably both from season to season and from year to year. Much of this variability was translated into fluctuations in stage in the lake complex.  Agriculture requires as constant an environment as  possible, and the early phases of the development of chinampa agriculture were centred on the problems of smoothing fluctuations in lake stage. Second, the Valley of Mexico was a closed drainage basin within which excess volumes of water could be relocated, but from which they could not be discharged.  Any reduction in fluctuations in one part of the  lake complex necessarily resulted in increased fluctuations in some other part of the complex.  As the areas of the chinampas were extended,  they began to impinge on this attribute of the hydrological system, and the later phases of the development of the chinampas were centred on the problem of the optimum distribution of fluctuations within the lake complex.  A great deal of what follows is essentially concerned  with the problem of controlling variability in the closed drainage basin of the Valley of Mexico.  44  CHAPTER III  THE COLONIZATION OF THE LAKE COMPLEX  It is useful to distinguish three phases in the development of the chinampas, namely, chinampa agriculture without water level control, with water level control in limited areas, and with water level control in whole lakes.  Chronologically these three stages are not well defined,  and they undoubtedly overlapped and merged into each other.  However, the  categories are useful because each stage involved the solution to quite distinct sets of technical and administrative problems, and each impinged on quite different aspects of the function of the lake complex. It should be noted that there is really very little information on the development of chinampa agriculture.  Much of what follows is  based not on direct evidence, but on inferences from quite general characteristics of the hydrology and history of the Valley of Mexico. The new period designations developed by Parsons and others are used in this thesis, and these are shown together with an absolute chronology in Table I. Chinampa agriculture without water level control The practice of transforming marshes and shallow lakes into agricultural platforms is unlikely to have originated in the Valley of Mexico.  There is mounting evidence from the Yucatan that chinampas,  or something very similar, were an important component of Classic Mayan subsistence and, if this is the case, then the form probably originated in the lowlands of Tabasco and Veracruz in the remote past,  45  TABLE I ABSOLUTE CHRONOLOGY AND PERIOD DESIGNATIONS FOR THE VALLEY OF MEXICO Approx. Absolute  New Period  Old Period  Chronology  Designation  Designation  Late Horizon Second Intermediate Phase 3  Late Aztec Early Aztec  Second Intermediate Phase 2  Late Toltec  Second Intermediate Phase 1  Early Toltec  Middle Horizon  Late Classic  First Intermediate Phase 4  Early Classic  First Intermediate Phase 3  Terminal Formative  First Intermediate Phase 2  Late Formative  First Intermediate Phase 1  Middle Formative  Early Horizon  Early Formative  1520 1400 1200 950 700 400 100  A. C. B. C.  200 500 800 1100  Source:  Parsons 1974:90.  46  and was introduced to the Valley of Mexico, along with so many other cultural traits, sometime during the Early Horizon (Mathney 1978). There is some direct evidence that raised platforms may have been used for agriculture in marshy areas of the Valley during the First Intermediate, Phase 1.  At a site dating from this period, near Tlaltenco  on the western shores of Lake Texcoco there are the remains of a platform which was about 2 meters higher than the contemporary lake bed, founded on a layer of rocks, and held in place by stakes driven into the lake floor (Palerm 1973:83).  The evidence is not conclusive, but the platform  may have been used for agriculture, and been a precursor of chinampa cultivation.  There is rather more substantial evidence that chinampa  cultivation was practiced during the Middle Horizon and in Lake Chalco, near Xico, in Lake Xochimilco, near Culhuacan, and on the eastern shores of Lake Texcoco there are the remains of a number of platforms which were very probably used for agriculture, all dating from this period (Apenes 1940:30; West and Armillas 1950:169; Nunley 1967:578; Armillas 1971:658). These scattered pieces of evidence suggest that the practice of cultivating poorly drained land by use of drainage ditches or raised platforms may have been known at quite an early date, but this does not imply that the form was also important at an early date.  Parsons'  thorough analysis of settlement data in the Basin indicates that the Valley floor was not colonized to any significant extent prior to the Second Intermediate Phase 1 (Parsons 1974:81-108).  This is the period  immediately following the decline of Teotihuacan, and it is characterized by the dispersal of population throughout the Valley, into areas that  47  had only been sparsely settled during the Middle Horizon.  Two quite  large centres were established on Lake Chalco during this period, one one the northwest shore of Cona Xico in the centre of the lake, and the other about 2 kilometers to the southeast of modern Chalco.  In addition  to these, some smaller settlements developed at the foot of Vulcan Tecutli and around the periphery of Lake Xochimilco. All of these settlements were at elevations well below the 2250 meter contour, and well within the area of the lake complex. Of the centres that were established in the Valley during the SI-1, Tula became the most important, and during the SI-2 it came to dominate the Valley.  Populations in the southern and central parts of  the Valley declined during this period, and the sites on Lakes Chalco and Xochimilco were abandoned.  The reasons for this decline are obscure,  but they may have been associated with the relations between Tula, in the northwest corner of the Valley of Mexico, and the equally commanding centre of Cholula in the Puebla Valley, which reduced the southern and central portions of the Valley of Mexico to a buffer zone between two powerful and contentious states (Parsons 1974:107). Towards the end of the SI-2 Tula declined in importance, and the site was sacked and abandoned sometime during the early part of the 13th Century. This apparently removed a constraint on population growth in the southern and central parts of the Basin of Mexico and a number of large settlements were established on and around the shores of the lake complex, including most of the towns with which we are familiar through the accounts of the early Spanish colonists. It is clear that the low elevations of Lake Chalco were first colonized during the SI-1, but this does not necessarily indicate that  48  the early development of chinampas also occurred at that time. Lorenzo has presented evidence that mean water levels in the lake complex were declining during the SI-1 and that by the end of the period they were considerably lower than they had been three-hundred years earlier (Lorenzo 1956, cited in Sanders 1970:88).  Because the floor of the  Chalco basin has such gentle gradients, even slight reductions of the mean water level would have rendered large areas of the lakeshore plain secure from inundation.  Under these circumstances it could be  that the colonization of the lakeshore plains during the SI-1 consisted of nothing more than the extension of existing techniques of agriculture onto the emerging areas of the lakeshore plain.  If this is not the case,  then it is hard to understand why refugees from Teotihuacan, dispersing to areas that were only sparsely populated, should have developed novel and extremely labourious techniques for cultivating swamps, rather than simply settling at slightly higher elevations and cultivating adequately drained land by techniques with which they were already quite familiar. The situation during the SI-3 was quite different.  Lorenzo's  researches indicate that water levels were rising, and it would have been necessary to undertake some form of drainage work even to maintain existing areas of cultivation.  At the same time the population of the  Valley was increasing quite rapidly, and this would have impeded a retreat from the Valley floor and stimulated the development of techniques for cultivating high water table land (Parsons 1974:103). For these reasons it seems probable that chinampa agriculture was developed during the SI-3, rather than during the SI-1, even although the lowlying areas of the basin had been colonized at the earlier date.  49  In its natural state the lake complex consisted of a series of ecological zones arranged more or less concentrically around the centres of the various lakes, and the character of these zones was broadly defined by the intersection of hydrological variables with variations in relief. On the basis of the calculations described in the Appendix it has been possible to estimate the extent of fluctuations in Lake Chalco, under uncontrolled conditions.  Figure 19 shows variation in stage over a 10  year period in which years 2 and 8 had precipitation 25% above the mean, years 4 and 6, 25% below the mean and the remaining years had normal values.  As the data cited for Tacubaya indicate, these extremes of  precipitation are quite realistic, but their frequency is overrepresented in the figure.  On average the extremes depicted would have been  more likely to have occurred over a 20 or 25 year period, rather than over the 10 years shown here.  The regime of the lake was dominated by  the Tag time implicit in the infiltrated flows by means of which the lake acquired the bulk of its input water.  Seasonal maxima and minima in  stage lagged some 5 or 6 months behind precipitation maxima and minima, with the result that fluctuations in stage and precipitation were almost exactly out of phase.  The lag times involved also meant that the effects  of a particular year's precipitation became evident only towards the end of that year, and persisted over the subsequent two or three years. In this example the high precipitation values in year 2 results in maximum values for stage during the winter of years 3 and 4, and they are s t i l l evident in the above average values of year 5.  Similarly, the  low precipitation values in year 6 hold stage at below average values until year 9.  The effects of the geometry of the lake depressions in  smoothing variations when water levels are high is also evident in the  E l e v a t i o n a b o v e  -5  m  T3  -5 L/> FD R+ O -"•  _i. 3  -a  CU  _ l . c+ c+ fD CD CL c+ -J.  <  O  CO  <  CU'  -S  3  a> C+ 1 I. c o ro 3 to in > 3  -a  T3 d  ro c+ 3 CU Q-tQ  _ i . ro X  "  o  "O - H CU  -s r— C+ CU  TCl—l ro < —' o 3" CU  •  o o -b  o < CU  -s << 3  ia  CU  3 3  CU  og  m e a n  s e a  l e v e l  ( m )  51  figure.  During year 7 when the mean annual water level was estimated as  2239.2 meters, the range of monthly values was 1.36 meters, while in year 3 when the mean annual value was 2241.4 meters, the range was only 0.31 meters.  Although the values shown in the figure are only estimates,  it is clear that the regime of the lake was characterized by pronounced seasonal variations in stage, and by rather larger inter-annual variations, and that the variations tended to be dampened during wet periods when stage was high, and emphasized during dry periods, when stage was low. The operational environment of the early colonists was defined by the relationship between these fluctuating water levels and the gentle gradients of the Valley floor.  These interacting variables are shown  schematically in Figure 20, where values for stage are the mean and extreme values show in Figure 19, and the profile shown for the lake follows from the assumptions stated in the Appendix, Part IV.1. Under normal conditions the central portion of the lake would have contained open water, and would have done so throughout the year.  During dry  periods the area of this standing water would have diminished and during exceptionally dry periods it would have disappeared altogether. During wet periods the lake would have expanded.  Surrounding the lakes  there would have been a series of marshy margins.  Those closest to the  lake would have been exposed only during the summers of exceptionally dry periods.  Those at a slightly greater remove would normally have  been flooded during the winter and exposed during the summer,.but during exceptionally dry or exceptionally wet periods they would have been either exposed or inundated throughout the year.  The outer margins would  normally have been exposed throughout the year, but they would have been  Elevation above mean sea level.(m) ho  UJ  r o •To  NO  —i-  — I —  Wet Year Margins CQ -5 fD ro  o  3  Mean Year Margins  1. m  3  OCu rtfD O-  3  cu  to c+ -j.  3  CU rtfD Q.  <  -s -s  CQ -J. i. Co c+ CO — i .  o 3  O - h to  |— _1. Cu 3 fD c+ O  3"  3" fD  Cu fD —J X o rto fD  a  Q n  D f  o 3  Q Q O Q D  3  rt-  X3  -a fD 3  a > 3 Q_  CL — i -J. o  X o CU  3. x  o o o  rt-  -a —1. ai o 3 -s r+ o i — i -n < •  c+  3" fD  2  2  fQ D D 00 a  CfD Q  o  -< fD Q  2  2 2  2 2 2  fQ D Q D f Q D •D X cn5'? (/)O X cn cn cn tn cn a a a a" a" O Q CfD O C Q C Q C Q C fD fD D f D fQ C D fQ C D fQ fD Q D -fD < Q Q D X  53  flooded during the winters of exceptionally wet periods.  Given the  extremely gentle gradients of the region, these margins would have been quite wide.  The calculations suggest that, on average, a strip of land  about 700 meters wide would regularly have been liable to inundation, and beyond this there would have been a further 1000 meters that would have been flooded occasionally.  On the eastern shores of Lake Chalco,  where gradients are more gentle than the average, the periodically inundated margins would have been rather wider.  The lake margins would  have been surrounded by a zone of high water tables that was not liable to inundation and this would have merged into the well-drained outer lakeshore plain. The margins and the lakeshore plains were poorly drained and liable to inundation, but they did offer two great advantages to agriculturalists.  First, they had nutrient-rich soils covered with a  light growth of vegetation that could be removed with neolithic tools. The most common soils in the Basin are the Chestnuts that are distributed over the lower slopes of the hillsides.  These are loams or sandy loams,  with a moisture retention capacity in the vicinity of 250 millimeters per meter depth.  In areas where they are not heavily eroded they are gener-  ally about 1 - 3 meters deep, and rather deeper where they have accumulated in depressions in the hillsides.  They are quite good agricultural soils  and are amenable to improvement by cultivation practices.  The soils that  occur in the larger subsidiary valleys and in the area of the old lakeshore plains are Chernozems. These are clay loams with a crumbed structure, a moisture retention capacity of about 300 millimeters per meter depth, and they are high in organic matter.  They are generally quite deep,  between 7 and 8 meters at the mouth of the Teotihuacan Valley, and they  54  are good agricultural soils (Sanders 1970:73).  It is difficult to dis-  tinguish the effects of variation in soil quality on yields from the effects of differing cultivation practices, but in the Teotihuacan Valley today the Chernozems of the Valley floor can yield about twice as much as the Chestnuts of the lower slopes under similar dry fallowing systems, and this gives some indication of the advantage of the soils of the Valley floor (Charlton 1970:88). The second advantage of the lake margins for agriculturalists was that water was present in large volumes at times when precipitation values were low.  Maize is a crop of the humid lowlands and, where its  cultivation is entirely dependent on rainfall, production is marginal in areas where precipitation is less than 600 millimeters per year (Sanders et al. 1979:222-24).  The southern part of the Basin is usually  quite adequately watered and the mean annual precipitation in the areas of Lakes Xochimilco and Chalco is upwards of 700 millimeters, but in the variable conditions of the region, values occasionally fall below 600 millimeters.  The Moyoguarda data cited above show values of less  than 600 millimeters in four of the 25 years of the record (Figure 18). If Lorenzo's evidence is correct and the climate of the Basin was passing through a dry phase during the Second Intermediate, then the frequency of dry years may have been higher than this.  In addition,  the cyclical fluctuation that is evident in the precipitation data for Tacubaya suggests that there would have been periods of twenty or thirty years when the frequency of dry years would have been higher than usual, alternating with similar periods of lower than normal frequencies.  Under  these circumstances, some form of irrigation would have been valuable in reducing the frequency of crop failure due to droughts.  55  Irrigation can be valuable not only in augmenting the volumes of water available for agriculture but also in adjusting the seasonal distribution of water availability.  Maize is a crop that is particularly  susceptible to damage by frost and it was noted that frosts are common even at low elevations in the south of the Valley.  Under normal cir-  cumstances maize is planted in April or May at a time of year when precipitation values are increasing and it is harvested in September before there is any great danger of frost.  However, the rains are often  late and if planting is delayed until the start of the rains, then the probability of frost damage at the end of the six-month growing season increases.  If it so happens that the rains are late in the same year that  frosts are early, the conjunction can be ruinous.  Irrigation is valuable  in these circumstances because it permits planting to take place in April or May regardless of whether or not the rains have started, and the harvest can take place six months later before there is any great probability of damage by frost (Sanders et al. 1979:222-24). The initial phase of the colonization of the Valley floor undoubtedly consisted of the extension of existing techniques of agriculture onto the relatively well-drained areas of the outer lakeshore plains, but as the area under cultivation extended into the marshy margins of the lake complex, novel techniques would have to have been developed for dealing with the problems of high water tables and occasional inundations. It might have been possible to cultivate maize on the seasonally inundated margins of the lakes without any modification at all to local drainage patterns.  In modern Tabasco there is a seasonally inundated  area of river flood-plain that supports an agricultural system known  56  locally as the "marceno" or "March planting" (Orozco-Segovia and Gliessman 1979: 11 ).  Maize is planted on areas that have been exposed by declining  water levels and it is harvested six months later just before water levels rise and the ground is again inundated. On occasions the water levels rise a little early in the season and flood the land before the corn is harvested, but this appears to create no major difficulties and the harvest is carried out from boats.  The system is extraordinarily  productive and yields from a single planting are normally between 4 and 5 tons per hectare, and are occasionally double that amount (Orozco-Segovia and Gliessman 1979:11).  In Tabasco the high yields are a consequence of  the high fertility of the s i l t in which the maize is cultivated. The practice is not in any way confirmed for the Valley of Mexico but something similar might have been possible.  Maize could have been  planted quite close to the April shoreline of the lake and, because water levels declined during the summer and tended not to return to April levels until the following September, the crop could have been clear of the water for the period of the growing season and could have been harvested before the rising water flooded the land.  Apart from access to nutrient-rich  soils the practice would also have allowed an April planting in moist soils, even if the rains were late.  If one assumes on the evidence of  the modern chinampas that the planting took place on land that was between 25 centimeters and 50 centimeters above April water levels, then the schematic diagram indicates that a strip of land about 300 meters.wide would have had the appropriate conditions.  The location and  dimensions of this zone would have been determined by the location of the April shoreline in any particular year, but a zone with these conditions would have existed in every year.  57  Something like the marceno system may have been practiced in the Valley of Mexico, but its usefulness would, have been limited by poor drainage and soil aeration.  The most simple means of reducing these  problems would have been the excavation of trenches.  The configuration  of the chinampas and the practice in marshy areas in other parts of Middle America today suggest that the ditches would have been excavated at right angles to the shoreline and about 10 meters or so apart (West and Armillas 1950:172-73; Vanegas 1978:5).  Ditches would have  permitted the accelerated drainage of water from the lake margins and by this function alone would have reduced the problems of waterlogging and increased the utility of seasonally or intermittently inundated land.  They would have had the additional function of increasing the  area of land in which pre-season irrigation was available: by excavating them to a depth a few centimeters below the mean April water-level, ditches could have provided standing water for"bucket irrigation in the drier areas of the outer margins and inner lakeshore plains.  How far  it would have been worth extending these trenches away from the lake would have depended in part on an evaluation of the cost of excavating the ditches, in terms of the effort required, against the benefits of pre-season irrigation.  If one assumes that 20 centimeters of water in  the bottom of a ditch was sufficient for bucket irrigation, that this depth was required in April, and that it was not worth the effort of excavating ditches more than one meter deep, then in the conditions shown in the schematic diagram, it would have been worth extending the ditches to a distance of about 1 kilometer from the mean shoreline. However, a further consideration is introduced as soon as any permanent installation is associated with water levels that vary from  58  year to year.  The ditches could only have been fully utilized when  April water levels were at or close to mean values.  In dry periods,  April water levels would have fallen below the level of the bottom of the ditches and water would not have been available.  The problem could  have been overcome by increasing the depth of the ditches, but this would have added to their cost.  In wet periods those areas close to the  mean shoreline would have been flooded and the labour expended in excavating ditches there would have been wasted.  Because water levels varied  from year to year the return on effort expended in excavating ditches would have been reduced. The area of cultivable land could also have been increased by raising soil surfaces in the more frequently inundated margins of the lake.  The benefits of such an undertaking would have been much the same  as the benefits of excavating ditches, but the costs would have been higher and would have been exacted in the form of the work required to transport the necessary volumes of material.  If one assumes that plat-  forms 10 meters wide were separated by ditches 1 meter wide, that the surfaces of the platforms were 30 centimeters above mean April water levels and that ditch bottoms were 20 centimeters at least below mean April water levels, then in the conditions shown in the diagram a strip of land 10 meters wide, extending 200 meters on either side of the mean shoreline, would have required the transport of about 670 cubic meters of material to raise the surface to the necessary elevation.  This work  may have been carried out in some areas but, because the platforms would have been liable to flooding during wet periods and would have been without water during dry ones, it seems unlikely that large areas of raised platforms would have been constructed, and that the early  59  colonization of high water table land would have been largely restricted to the excavation of ditches. Chinampa agriculture and water level control in limited areas The value of permanent installations on the margins of the lakes was limited by the large year-to-year variations in water levels.  If  water levels could have been held constant, the value of ditches and agricultural platforms could have been enormously increased.  Lake  stage is a function of the difference between input and output rates and it can be held constant only when the two rates are matched. In the categories used in the present discussion, lake inputs were by direct precipitation, by surface and subsurface discharge from the hillsides and, in all the lakes except Chalco and Zumpango, by discharge from lakes upstream.  Lake outputs were by evaporation and, in  all the lakes except Texcoco, by discharge to lakes downstream. Most of these variables were not amenable to regulation. cipitation is quite beyond control.  Pre-  Evaporation and subsurface drainage  could be modified to some extent by controlling vegetation cover, slope gradients and soil quality, but the flows are too diffuse to allow any precise control.  Streamflow could be regulated by check dams but this  would have required the construction and integrated operation of a large number of small installations and, in all the basins except Mexico and Cuautitlan, streamflow is too small a proportion of the total input to the lake complex to have allowed anything but slight modifications to water levels. The discharge of upstream lakes was technically amenable to control but, because it involved the exploitation of upstream storage  60  capacity in a manner that was not of benefit to the inhabitants of upstream watersheds, it also required a degree of coercive political control.  The small polities of the early colonists were unable to  exercise this type of control over their upstream neighbours and, because of this, they would not have been able to regulate the discharge of upstream lakes.  This aspect of the development of chinampa  agriculture is discussed more fully below with reference to the growth of the Mexica state and the management of chinampa agriculture. The only remaining variable by means of which the level of water in the lakes could have been controlled was discharge as an output, and it was this variable that virtually all of the purely hydraulic installations of the lake complex were intended to regulate.  Essentially, lake  levels could be held constant by insuring that all inputs of water were translated into outputs as rapidly as possible.  To accomplish this  two types of installation were necessary, namely, dykes and canals. We have very little direct information on early hydraulic installations in the Basin, but it is possible to make some quite plausible assumptions from more general characteristics of the region. Drainage canals were probably constructed by straightening and deepening natural watercourses, and it is possible that the major canals of the 16th Century were in much the same locations as the original watercourses. The Canal Amecameca ran along the southern shores of Lakes Xochimilco and Chalco draining the discharge of Rio Amecameca and other smaller streams in the south and southeast of the Basin, as well as the discharge of the many springs along the southern shoreline of the two lakes.  At a point near Xochimilco the canal turned north and flowed  into Lakes Mexico and Texcoco, collecting as it passed the discharge  61  of rivers in the southwestern corner of the Valley (Figure 24).  A second  large canal ran along the northern shores of Lakes Chalco and Xochimilco, but this may have been a Colonial rather than a pre-Hispanic installation. Many of the streams that drained the slopes of the Sierra de Las Cruces were canalized, and.their discharge directed through the Lake of Mexico into Lake Texcoco (Palerm 1973:113 note 48). In theory, water levels could have been held constant by the use of canals above.  If the canals were excavated down to the same level as  the minimum stage of a lake, excess volumes would simply drain out of the lake and the level would not fluctuate.  In practice, there were  three reasons why water levels would not have been maintained at minimum values and why fluctuations in stage would not have been controlled by canals alone.  First, water was required for irrigation, and i f the lakes  were maintained at low levels, water would have been available only in small areas in the centre of the lakes.  Second, it was noted above,  with reference to the regime of Lake Chalco, that the geometry of the lake basins acted to dampen, fluctuations in both the stage and the extent of the lake when water levels were high, and to emphasize them when water levels were low.  As the aim of the hydraulic installations  was to minimize fluctuations in stage, this property of the lakes of the complex would have provided an incentive for maintaining stage at higher rather than lower values.  Finally, in the south of the Valley  at least, it would not have been possible to hold water levels at minimum values.  The difference in elevation between Lakes Chalco and  Xochimilco, on the one hand, and Lake Texcoco, on the other, was so slight that high water levels in Lake Texcoco were above the low water levels in the other two lakes.  Thus, when the levels of Texcoco rose  62  to the level of the water in Lakes Chalco and Xochimilco, the gradient between them would have been eliminated and the levels of all three lakes would thereinafter have risen together.  For these reasons, dykes  as well as canals were required for the control of fluctuations in stage. With dykes, water could be backed up and maintained sufficiently high above the levels of downstream lakes that the gradient persisted, and drainage could occur even when stage downstream increased.  With high  water levels, the natural tendency for fluctuations to decrease with increase in stage would reinforce the effects of the dyke in holding levels constant and, in the large areas behind the dyke, water could be stored for irrigation. The locations of the earliest dykes in the lake complex are unknown.  There are mentions of early hydraulic works in the vicinity  of Culhuacan and Xochimilco and at^ Xico and Tlapacoya on Lake Chalco (Palerm 1973:113 note 48).  In addition, the early importance of sites  along the western shores of Lake Mexico-Texcoco suggest that dykes may have been constructed across the mouths of the many small streams and rivers that drained the slopes of the Sierra de la Cruces.  Although  there is no more, detailed evidence than this, some sites in the lake complex were more suitable for dyking than others, and it is probable that these were the first to be adapted.  The earliest dykes were prob-  ably constructed across the mouths of small indentations in the shoreline of the lake complex and, given the minute scale of traditional agricultural practices, they were probably only a few meters long. There are a number of locations in the southern lakes that would have been suitable for larger dykes.  If one assumes dykes were approximately  the same height, the costs of the dyke in terms of the amount of effort  63  required to construct it would have been roughly proportional to its length, and its benefits in terms -of the control of fluctuations would have been roughly proportional to the area contained by the dyke. Thus the ratio  of the length of the dyke to the area contained is an indica-  tion of the efficiency of any particular installation.  Figure 21 shows  a number of sites in the southern half of the Basin, evaluated in these terms.  Smooth or protruberant sections of shoreline are clearly unsuit-  able (Dyke No. 1).  Islands near the shoreline and small bays would have  been appropriate for dyking and the hydraulic works mentioned in the vicinity of Xochimilco and Tlapacoya may have exploited these topographic features (Dykes Nos. 3 and 5).  The dykes across the mouths of  lakes, Nos. 7 and 8, are technically more complex than those with lower efficiency ratios, and are to be discussed in the following section. Within the areas contained by the dykes, water could held at a constant level.  Ditches could be excavated with the assurance that they  would contain water for bucket irrigation in April in all but extraordinarily dry years and,with the assurance that flooding would occur only in exceptionally wet years, it would have been worth the effort required to raise soil surfaces above the levels of the water.  To the  extent that dykes and canals were successful in holding water levels constant the problems of the earlier colonists would have been resolved and the potential benefits of chinampa agriculture could be realized. It is not possible to put any precise dates on the construction of these earlier installations and the development of chinampa agriculture in restricted areas of the lake complex.  Some small bays and inlets may  have been colonized as early as the Second Intermediate Phase One and  64  No.  Dyke Length  Enclosed Area  Ratio  km 1 .  2. 34. 56. 78.  Figure 21.  7-3 3-9 2.3 4.0 3-2 0.8 3-3 2.5  4.6 3.9 3-9 10.0 6.9 3-6 116.8 98.8  0.6 1 . 0  1.7 2.5 2.2 4.5 35-4 39-5  Length to Area ratios of hypothetical dykes in Lakes Xochimilco and Chalco.  65  others may not have been dyked until after the emergence of the Aztec state in the Late Horizon.  However, it seems likely that the develop-  ment occurred largely during the Second Intermediate Phase Three, when population was pressing on carrying capacity at existing technical levels and when there was an incentive to.solve subsistence problems with novel agricultural practices. Chinampa agriculture and water level control in large areas Dykes across the mouths of bays and inlets would have allowed control of water levels in limited areas and within these areas chinampas could have been constructed and operated.  However, these were not the  most suitable sites for dyking and it is evident from Figure 21 that dykes across the discharge points of lakes would have been very much more efficient.  In the final phase of the development of chinampa  agriculture, dykes were constructed that controlled water levels in large areas and that substantially modified the function of the whole of the lake complex. Lake Chalco was the most suitable of the lakes for colonization and it was almost undoubtedly the first of the lakes to be fully adapted for chinampa agriculture. The lake was shallow and it had a well defined mouth with a small island at Cuitlahuac that would have facilitated the construction of a dyke.  There was only a slight gradient  between Chalco and Xochimilco so that there was no great natural head of water to contend with, but it was elevated sufficiently high above the level of Lake Texcoco so that a drainage gradient persisted in all but unusually wet periods and water could be discharged whenever it was expedient to do so.  In general, Lake Chalco bore a basic functional  66  resemblance to the small, independent and open drainage systems that had been colonized at an earlier date and, apart from issues arising from the fact that a rather greater area was involved, the colonization of Lake Chalco did not require the solution to any novel technical problems. Lake Xochimilco had good water supply and drainage characteristics and it too was well suited to conversion to chinampa agriculture, but it may not have been quite as well suited as Lake Chalco.  It was  about 1.5 meters above the level of Lake Texcoco and, although the difference in elevation is not great, it would have necessitated the construction of installations that were considerably larger and more robust than those constructed at Cuitlahuac. Certainly, at the time of the Conquest the Aztecs had found it necessary to construct two, and possibly as many as four, separate dykes between the Ixtapalapa peninsula and the shoreline at Coyoacan. For this technical reason and for reasons associated with the failure of the inhabitants of the Xochimilco basin to form a unified administrative apparatus prior to the ascendancy of the Mexica it is likely that the mouth of Lake Xochimilco was not dyked until the 15th Century.  This is not to suggest that Lake Xochimilco was without  chinampas at an earlier date.  Ethnohistorical sources indicate that  there were chinampas in the vicinity of Culhuacan during the early 14th Century, and the large areas of chinampa cultivation in the vicinity of Xochimilco were undoubtedly constructed before the Mexica came to power (Peterson 1962:93; Vaillant 1966:110; Davi.es 1977:37-38). Figure 22 shows some of the extant chinampas in the vicinity of Xochimilco and, although they may not have been installed prior to the rise of the Mexica, they are probably among the oldest of the extant chinampas.  Figure 22.  Chinampa platforms in the vicinity of Xochimilco (CIA Mexicana Aerofoto SA 1977).  68  However, these areas of chinampas are probably best understood as the product of incremental extensions to controlled areas along the lakeshore, rather than as the infilling of areas contained behind a dyke at the mouth of the lake.  Even i f the Aztecs were not responsible for  the dyking of the mouth of Lake Xochimilco, they did complete the colonization of the lake, and all the indications are that during the 15th Century the area was thoroughly reworked.  Existing chinampas were  realigned and tidied up and any areas that had not previously been colonized were converted to the orderly and uniform chinampas, some of which can be seen in the vicinity of San Gregorio today (Figure 23). Because the growth of modern Mexico City has obscured so many physical traces we know very little of the history of settlement and chinampa agriculture in that region, but it is clear that the combined Mexico-Texcoco was of all the lakes in the Basin the least suitable for conversion to chinampa agriculture.  The model indicates that because  of its greater area natural seasonal fluctuations in stage may have been less than in the other lakes.  But because above-normal volumes of water  would have been disproportionately accommodated in Lake Mexico-Texcoco it is likely that the range in values for stage over a period of years was greater than was the case in the other lakes.  At the same time the  lake lacked a drainage outlet so that all losses were by evaporation, with the result that the waters of the lake became quite brackish. The salinity may have been unequally distributed and it is possible that there was a sufficient flow of fresh water from the slopes of Sierra de las Cruces and from Lake Xochimilco that the salts did not accumulate in the western parts of Mexico-Texcoco to the extent that they did elsewhere.  Figure 23.  Chinampas of San Gregorio Atlapulco, on the southern shore of Lake Xochimilco (CIA Mexicana Aerofoto SA 1977).  70  The most feasible method of claiming brackish marshes and lakes for agriculture would have been to construct dykes across the inlets along the foot of Sierra de las Cruces.  These would have been filled  with the discharge of freshwater from the streams flowing off the h i l l sides and, during periods of seasons of low water levels, the water could have been discharged into the main body of the lake, flushing the areas contained by the dykes.  If the operation was repeated sufficiently  often, most of the salts could be washed from the soils, and the area could be used for agriculture.  The dykes could have been extended in  increments to form quite large compartments that could be flushed in this manner.  Given the historical importance of the settlements on the west-  ern shores of Lake Mexico-Texcoco it is likely that at least a few sections of the shoreline had been colonized in this manner prior to the foundation of Tenochtitlan in the mid 14th Century.  There are also  some indications that parts of the lake itself had been settled at quite an early date.  Archaeological evidence from Tlatelolco indicates that  the site may have been settled in the 13th Century and there are some ethnohistorical indications that the island upon which Tenochtitlan was established was inhabited before the Mexica arrived (Davies 1977:37). The traditional date for the founding of Tenochtitlan is 1325, but there has been some confusion concerning the transposition of the Nahua calendar and it is now thought that a more probable date is 1345 (Davies 1977:37).  Soon after it was founded, dykes were constructed  connecting the new city to the shoreline at Tlacopan, Tepeyac, Ixtapalapa and Coyoacan, with construction probably proceeding in that order (Madrid-Mendizabel 1946:3).  These radiating dykes divided the  western part of Mexico-Texcoco. into a number of large sectors, each of  71  which could have been subdivided by smaller dykes, filled with fresh water and claimed for agriculture.  The final stage in this line of  development was the construction of the Albarradon de Nezahualcoyotl in 1449 which extended for 16 kilometers from a point on the shoreline to the northeast of Tenochtitlan to the Ixtapalapa peninsula (MadridMendizabel 1946:21). This strategy for claiming saline portions of the lake complex could have succeeded as long as adequate supplies of fresh water were available.  The smaller compartments on the western shores of the lake  could have been supplied by streamflow from Sierra de las Cruces, but the larger compartments and sectors may have required additional supplies.  The Aztecs constructed a number of aqueducts which may have  been intended to fulfill this function.  These included aqueducts from  the springs of Atzcapotzalco which may have been completed at quite an early date, the Chapultepec aqueduct, the remains of which can still be seen in modern Mexico City, and an aqueduct from the springs of Coyoacan, which was completed in the last years of the 15th Century (Madrid-Mendizabel 1946:21).  The largest supply of fresh water to the  Lake of Mexico was from Lake Xochimilco. The canals that directed this water north to Tenochtitlan were important as transport routes, but the configuration of the system suggests that an equally important function was the provision of fresh water to Tenochtitlan, both for drinking and for the operation of chinampa agriculture in the constituent compartments of the Lake of Mexico. With these supplements of water, levels in the compartments of the Lake of Mexico could be maintained above the levels of Lake Texcoco, but only so long as the level of water in Lake Texcoco did not rise too  72  high.  However, the levels of water in Lake Texcoco fluctuated consider-  ably.  Calculations with the model of the hydrological system suggest  that in normal years the levels may have varied by as much as 70 centimeters and it is recorded that on at least one occasion the level of Texcoco approached that of Chalco, which was about 2 meters above the level of the lower lake (see Figure 28 below).  Even under less extreme  circumstances, dykes in the Lake of Mexico would have to have been constructed to very large tolerances i f they were to remain above the levels of Lake Texcoco in the majority of years. The only method of reducing variations in the stage of Lake Texcoco was to exploit the upstream storage capacity of the lake complex.  It would have been possible to avert dangerously high water  levels in Lake Texcoco by holding back water in the southern lakes and in Lakes Zumpango and Xaltocan in the north.  The latter lakes were  more important in this respect than the mean annual figures charted in the diagram of the hydrological system of the Basin suggest (see Figure 2 above).  In normal years very little water was discharged from the  northern lakes but calculations with the model indicate that in wet years significantly large volumes did drain southward into Lake Texcoco. This finding is confirmed by early Spanish colonial documents that single out Rio Cuautitlan as one of the major causes of flooding in Mexico City (Madrid-Mendizabel 1946:21).  For these reasons it would have been impor-  tant for the inhabitants of Tenochtitalan to exploit the upstream storage capacity of the system.  The existence of a dyke at Ecatepec, across the  discharge point of the northern lakes, is not entirely substantiated for the period prior to the Conquest, but such an installation would have been important in preventing floods in Tenochtitlan during wet years.  73  There were at least two dykes across the mouth of Lake Xochimilco and these too would have been valuable in averting damaging rises in the level of Lake Texcoco.  The fact that these dykes would have been  critical to the development of the Lake of Mexico, while they would merely have been advantageous to the development of Lake Xochimilco, suggests that they were constructed in the 14th Century by the Mexica rather than at an earlier date by the Xochimilca or the Culhua of the Ixtapalapa peninsula. Because the Lake of Mexico was the least suitable of the lakes for conversion to chinampa agriculture and because its colonization required the solution of quite novel technical problems, it was the last of the lakes to be reclaimed.  It was settled at a later date than the  other lakes of the system and the work on hydraulic installations continued right up until the time of the Conquest.  The problems of creat-  ing and then reclaiming the Lake of Mexico were enormous, and it is probable that the Lake was never fully claimed for chinampa agriculture. Late Aztec Hydraulic Installations Figure 24 shows the major dykes, aqueducts and canals of the lake complex as they were in the early part of the 16th Century. The Nahua word for "dyke" or "causeway" is "cuepotli" which translates literally as "roadway of turf," and suggests how the earlier and smaller dykes may have been constructed (de Lameiras 1974:29). were more, elaborate.  Later installations  The Albarrad6n de Nezahualcoyotl was formed by  driving two parallel rows of stakes into the lakebed and by filling the space between them with earth, rubble and.aquatic vegetation (Palerm 1973:83).  The dykes radiating from Tenochtitlan were made of  A. B. C. D. E. F. G. H. I. J, K.  Tenochtitlan Tlatelolco Tepeyac Culhuacan Ixtapalapa Churubusco Albarradon de Ahuizotl Albarradon de Nezahualcoyotl Atzcapotzalco aqueduct Chapultepec aqueduct Coyoacan aqueduct dyke canal aqueduct  16thc. lake shore 2250 m contour  Figure 24.  Late Aztec Hydraulic Installations (Palerm 1973).  75  similar materials, but they were faced with stone and mortar and may have been plastered as well (Palerm 1973:83).  Many of the dykes served  not only to control water levels but also as transport routes.  The  dyke at Cuitlahuac was wide enough for four horsemen to ride abreast, and the dyke running from Ixtapalapa to Tenochtitlan was twice as wide as that (Palerm 1973:46, 49). We have little direct information on the 16th Century canals, but the network probably bore a general resemblance to the waterways in the extant chinampas and consisted of a few larger canals formed by straightening and deepening natural watercourses and numerous smaller canals for local drainage, irrigation and transport. There were a number of aqueducts in the system which consisted of ditches mounted on causeways.  Figure 25A shows a reconstruction of the  aqueduct from Atzcapotzalco to Tenochtitlan which was built between 1417 and 1427, and which was made of earth and stone held in place by wooden stakes.  Figure 25B shows, a section of the Chapultepec aqueduct  which took thirteen years to construct, and.which was completed in 1466. The structure was of wood and stone, held in place with mortar, and provided with a double channel so that each in turn could be cleaned without interrupting the flow of water (C.H.C.V.M. 1964 1:149-50; Palerm 1973:52).  The aqueducts were fed from springs and, from the descrip-  tion of the disasters associated with the Coyoacan aqueduct which was completed in 1497 or 1498, it is evident that the discharge was fed first from the spring into a pool or reservoir formed by the construction of a retaining wall around the spring and thence to the head of the aqueduct.  The impression given in the sources is that the retaining  wall was quite high and the reservoir quite deep (Palerm 1973:85).  76  B. Figure 25.  Chapultepec Aqueduct. Reconstructions of 16th Century Aqueducts (C.H.C.V.M. 1964 1:149, 150).  77  At various points in the Lake of Mexico and probably in the other lakes as well, dykes and aqueducts were breached to allow the passage of water and, in some of the larger canals, of canoes as well.  In some  instances, the openings were no more than unadorned interruptions in the line of the dyke, as was the case in the Albarradon de Nezahualcoyotl (Palerm 1973:34).  Dykes that also served transport needs were equipped  with wooden bridges over the openings as was the case with the dykes that converged on Tenochtitlan (Palerm 1973:32).  We also have a des-  cription of the Chapultepec aqueduct in which a conduit "as wide as an ox" conducted the flow of the aqueduct over the water below (Palerm 1973:52). The dykes maintained water at different levels.  This feature  of the system is most dramatically illustrated by an incident during the battle for Tenochtitlan when the Aztecs deliberately breached dykes in the hope that the Spanish invaders would be drowned in the sudden rush of water that followed (Diaz 1976:317).  Even without such inci-  dents as examples, it is clear from the configuration of the system that the dykes maintained differences in stage, and it seems probable that the Aztecs could control these levels quite accurately (Palerm 1973:49). The sources provide no clear descriptions of the devices used to regulate water levels.  Weirs are technically quite straightforward and  these may have been employed by the Aztecs.  There are some modern  examples of Indian hydraulic systems in which water levels are controlled by moving and replacing heavy stones in apertures at different elevations in a dyke.  Early Spanish-Nahua dictionaries have quite a large  vocabulary relating to the opening or the blocking-up of apertures, to the materials employed for this purpose, and to the individuals whose  78  job it was to perform the task, and this suggests that similar devices may have been used in the lake complex in the 16th Century (de Lameiras 1974:31-34). Some of the dykes were breached by canals and we know that there was a great deal of canoe traffic between the various lakes.  This  implies the existence of sluices or locks that would permit the passage of canoes and, at the same time, maintain differences in water levels, but the sources offer no adequate descriptions of these mechanisms. The only direct reference is in a passage of one of Cortez letters, 1  which reads: Y ya junto la ciudad esta una puente de madera de diez pasos de anchura y por allf esta abierta la calzada porque tenga lugar el agua de entrar y salir, porque crece y mengua, y tambien por fortaleza de la ciudad.porque quitan y ponen algunas vigas muy leungas y anchas de que la dicha puente esta hecha, todas las veces que quieren; y de estas hay muchas por la ciudad. (Quoted in Palerm 1973:52) The text is not clear.  The last lines would.seem to suggest some kind  of a drawbridge by. means of which access to the city could be restricted, but the first lines associate the opening of the bridge with allowing the water to "come and go," because it "rises and falls," and this would seem to indicate the existence of some kind of sluice gate. Whatever the installations were, there were many of them throughout the city. The configuration of the hydraulic system and the direct evidence that we have indicate that the Aztecs had some means of regulating water levels and, at the same time, permitting the passage of canoes. But without any concrete information we are left to speculate on the nature of these devices.  79  The chinampa platforms themselves were constructed in the areas where water levels were controlled and we know that virtually the whole area of both Lakes Xochimilco and Chalco were converted to chinampa agriculture.  Chinampas were also constructed in the Lake of Mexico, but  because the area is now concealed by Mexico City it is not clear how thoroughly the lake was colonized (Calnek 1972).  Sections along the.  western shoreline and in the vicinity of Tenochtitlan and other towns on the lake were converted to chinampas for both agricultural and residential purposes.  The description of the bridge in the Chapultepec  aqueduct includes the note that the water in the canal below the aqueduct was saline, which suggests that something less than the whole of the lake was colonized (Palerm 1973:52).  In addition to the areas in  the southern lakes, there are indications that chinampa agriculture was practiced on the eastern shores on Lake Texcoco, at the mouth of the Teotihuacan Valley, and in restricted areas in Lakes Zumpango and Xaltocan. (Apenes 1943; Gibson 1964:268; Palerm 1973). The: map of the hydraulic installations of the Basin at the time of the Conquest is.in effect a diagram of the structure of the system. Taken in conjunction with the model of the hydrological system of the Basin that is described in the Appendix it is possible to infer a great deal about the manner in which these various installations functioned. The dykes across the mouths of Lakes Chalco and Xochimilco and the Albarradon de Nezahualcoyotl did not simply maintain upstream water levels constant as did the earlier dykes across the mouths of small bays and inlets, but they quite radically altered the function of the lake complex as a whole.  Water levels were a function of the volumes of  water stored within the complex.  These volumes were determined by  80  differences in the rates of precipitation and evaporation for in the whole of the Basin of Mexico they were entirely beyond control.  Levels  in upstream portions of the lake complex could be controlled by passing surplus volumes downstream, but the surplus volumes were not eliminated from the lake complex, they were only relocated within it and caused the water levels to rise in the downstream lakes.  Any decrease in the  fluctuations in one part of the system could only be accomplished by increasing fluctuations in some other part. Figure 26A shows an estimate of normal seasonal fluctuations in the four southern lakes of the complex under natural and uncontrolled conditions.  Figure 26B shows exactly the same data, but with the addition  of a dyke at Cuitlahuac which held levels in Lake Chalco as close to constant as possible.  Because evaporation rates in Lake Chalco during  the early part: of the summer were slightly higher than the sum of all inputs, levels-declined in spite of the dyke, but the decline was only slight.  The water that in uncontrolled conditions had been stored in  Lake Chalco was now distributed between the other three lakes, with the result that their levels fluctuated more widely than they had before the dyke was in place.  Figure 26C shows the same data again, but with  the addition of a further dyke at the mouth of Lake Xochimilco.  Because  evaporation rates in Xochimilco are at no season of the year greater than input rates, the levels in that lake could be held quite constant. Lakes Mexico and Texcoco, on the other hand, had to accommodate the excess from both of the upstream lakes and their levels fluctuated yet more widely.  Figure 26D shows the effect of the construction of the  Albarradon de Nezahualcoyotl on the function of the system.  It can be  seen that the dyke would only have been marginally effective in controlling  £ u <y cn o  c o  CJ  E O  o c o  J'F'M'A'M'J'J'A'S'O'N'D A . •30  w i t h o u t  J 'F'M'A'M'J'J 'A'S'O'N'D J'F'M'A'M'J'J 'A'S'O'N'D  d y k e s  Chalco r = 10  ^20 i/i £ o  ai cn o  1  •10  o  o  QJ  £  ^ -10 D O  .o  ° -20 c o  o -30 -40 -50 J FMAMJ JAS0ND B .  Figure 26.  w i t h  l e v e l  o f  J FMAMJ JAS0ND C h a l c o  J FMAMJ  JAS0N0  c o n t r o l l e d  The Effects of Dyking on Normal Variations in the stage of the central and southern lakes (continued on following page). (Appendix Parts III and IV).  82  C .  w i t h  l e v e l s  o f  J FMAMJ JASOND D .  w i t h  E .  a s  D ,  l e v e l s  b u t  o f  w i t h  C h a l c o  a n d  X o c h i m i l c o  J FMAMJ JASOND C h a l c o ,  u p s t r e a m  X o c h i m i l c o  s t o r a g e  c o n t r o l l e d  J FMAM J J A S O N D a n d  u s e d  M e x i c o  t o  h o l d  J FMAMJ J ASOND  c o n t r o l l e d  M e x i c o  c o n s t a n t  83  fluctuations in the Lake of Mexico.  Evaporation rates in the Lake of  Mexico are high; during the spring and summer they greatly exceed inputs with the result that the levels fluctuated considerably.  Lake Texcoco,  the final destination of all the excess water in the system, fluctuated very widely.  The lakes were linked in series and, with the exception  of Lake Chalco, it was technically possible to control levels in any of the lakes not only by regulating outputs downstream, but also by regulating inputs from lakes upstream.  Figure 26E, for example, indi-  cates that it would have been possible to hold the levels of Mexico quite constant by exploiting the storage capacity of Lake Xochimilco. Figure 27 shows the same data as Figure 26D but with an indication of the relative elevations of the four lakes.  The differences in  elevation are taken from.18th Century figures, but they do not conflict with what we do know of the lake complex in the 16th Century (Sanders 1976).  The dyke at Cuitlahuac held median levels of Lake Chalco some  50 centimeters above the level of Lake Xochimilco, and Lake Xochimilco was held 1.5 meters above the level of the Lake of Mexico. The Albarradon de Nezahualcoyotl held the Lake of Mexico 50 centimeters above the level of Lake Texcoco. In normal years quite low dykes would have been sufficient to maintain differences in water levels, but in wet years the complex was liable to flooding.  Figure 28 shows the same four lakes and their  relative elevations, but it shows the variations in lake levels over a period of three consecutive years, in which year 1 had precipitation 25% in excess of normal values, year 2 had precipitation 25% below normal values, and year 3 had normal precipitation.  The period is assumed  to have been preceded by a series of normal years.  The variations  84 2241  C h a l c o  -j=  2240  X o c h i m i l c o  2239 H  M e x i c o  UJ 2238 A  T e x c o c o  2237  1  J  1  F  Figure 27.  M  1  1  A  1  M  1  J  1  J  1  A  S  1  0  1  Estimated Normal Variations in the Stage of the Southern and Central Lakes with water levels controlled (Appendix Parts III and IV).  1  N  D  f-  2241 +  Chalco dyke  Chalco mean stage Xochimilco dyke  > O Q)  in  c a  Xochimilco mean stage  2240  E CD > O  a c o a 2239 >  Mexico dyke  LU  k  Mexico mean stage  Texcoco mean stage  2238  J FMAM  Figure 28.  J J A S O NDJ F M A M J  J ASONDJ  FMAMJJASOND  Estimated Inter-annual variations of stage in the controlled Lake Complex without centralized Management (Appendix Parts III and IV).  CO CJ1  86  in precipitation are large, but they are not entirely unrealistic, and the extremes shown would probably have occurred in the course of a 27 year period.  It is also assumed that the dykes projected 50 centimeters  above median upstream water levels and that they could accommodate a rise of this much before their storage capacity was exhausted.  It is  assumed that by means of the canal network linking Lakes Xochimilco, Mexico and Texcoco the discharge of the former could have been delivered to either of the latter at the discretion of the managers of the system. Lakes Zumpango and Xaltocan in the north of the lake complex are not shown, but the discharge of the Cuautitlan Basin is included in the inputs of Lake Texcoco (Appendix Parts III and IV). At outset the level of Texocois at its normal highwater mark and the levels of the other three lakes are at their median values. During the spring and early summer of year 1 , when evaporation rates are high and before the current year's precipitation has taken effect, levels in all the lakes except Xochimilco decline, with Chalco reaching minimum levels in May, and Mexico and Texcoco in June.  In June or July  the effects of the current year's precipitation begin to take effect and levels begin to rise.  In July the levels in Chalco and Mexico reach  their median values, and excess volumes are passed downstream to Texcoco, whose levels rise accordingly.  In September the level of Texcoco rises  to the median value of Mexico; thereinafter the gradient between the two is eliminated, and their levels rise in tandem.  In the same month  the storage capacity of the two northern lakes is exceeded and the discharge of the Cuautitlan basin drains into Texcoco.  The excessive  precipitation of year 1 continues to accumulate in the lake complex and water levels continue to rise.  In October of year 1 the levels of  87  Mexico and Texcoco rise above the level of the Albarradon de Nezahualcoyotl and the saline waters of Texcoco can mingle with the fresh waters of the Lake of Mexico. Levels continue to rise during the fall and winter of year 1 . In February of year 2, five months after the peak in precipitation, lake storage reaches its maximum and the levels of Mexico and Texcoco come very close to the level of Xochimilco. In April of year 2 evaporation rates exceed input rates and levels begin to decline. In May the levels of Chalco fall slightly and reach a minimum in June, after which water is retained until the median level is reestablished. Levels in Mexico and Texcoco decline steadily during the spring of year 2, but the rate of decline decreases as the effects of the precipitation begin to accumulate in the lake complex and, during the last part of year 2 and the early part of year 3, the rate of decline is very low.  As a result of the low precipitation of year 2, inputs to the lake  complex in the early part of year 3 are not sufficiently large to offset evaporation losses.  In Chalco stage declines earlier than in  the preceding year and to a lower value, and in Xochimilco there is a slight fluctuation in stage.  The stage of Mexico and Texcoco declines  quite steeply, and in June they fall below the elevation of the dyke separating the two.  Stage is at a minimum in July, at which time the  current year's precipitation begins to take effect, and levels rise towards their winter peak. During the three-year period shown in Figure 28, the level of each lake was held as close to median values as possible, and the managers of the individual basins were assumed to have acted in an uncoordinated and self-interested fashion.  Managers in Chalco and  Xochimilco simply discharged unwanted volumes downstream, and managers  88  in Mexico could do little other than tolerate the resulting floods. Figure 29 assumes the same precipitation values as in the preceding figure but, in this case, the managers of the Mexico basin are assumed to have control of the upstream lakes, and the storage capacity of Xochimilco and Chalco are exploited to reduce flooding, in Mexico. The results are similar to those shown in the preceding figure, but there are some significant differences. In the first year, the stage of Mexico is held constant for 9 months rather than 5 as in the preceding diagram.  In the second year  the stage of Mexico and Texcoco is above the level of the dyke that separates them for 16 rather than 22 months, and the maximum height of the flood is some 30 centimeters less than in the preceding figure. The differences are not very great, but in less extreme conditions than those depicted here, the capacity to increase the time during which the stage of Mexico could have been held constant and to reduce the height and duration of flooding may have been critical for the inhabitants of the Lake of Mexico. The diagrams, and the model from which they are derived are only approximations of 16th Century conditions, but they do provide an insight into the functioning of Aztec hydraulic installations. evidently quite complex:  The system was  in the model there are several alternatives  available to the manager with regard to the routing of water, the location 1  of storage and the timing of discharge; these provided a number of options for dealing with any particular contingency.  The real system was  very much more complex and the range of options open to the Aztec managers correspondingly larger.  The system required a considerable level of  Chalco Chalco dyke  2241  CD >  Chalco mean stage Xochimilco dyke  Xochimilco  o  CD CO  c o  ]- Xochimilco mean stage  2240  CD  E  CD > O  JD O  c o  I- Mexico dyke  "a 2239 > LU  [• Mexico mean stage  U  2238  J FMAMJ  Figure 29.  J ASONDJFMAMJ  JASONDJ  FMAMJ  Texcoco mean stage  JASOND  Estimated Interannual Variations in Stage of the Controlled Lake Complex, with Centralized Management (Appendix, Parts III and IV).  CO  90  experience and skill on the part of the managers.  This involved not  only experience of the consequences of particular decisions on the function of the system, but also experience in relating present observations to the future behaviour of the system.  Because there was a lag  between precipitation and the resulting variations in lake storage the possibility of anticipating future stage from present precipitation existed.  Because stage was not simply determined by precipitation, but  related as well to volumes of water already within the hydrological system, skill was needed in making inferences of the moisture content of the Valley from such indicators as spring flow or the condition of plants, and in relating these to the future condition of the lake complex.  The system also required decision-making not on the basis of  achieving simple or unique goals, but on the evaluation of the relative costs.and benefits of different strategies.  In the simplified and  extreme situation shown in the diagram, the major choice was between flooding Mexico to a great depth for a long period, but maintaining the production of Xochimilco largely uninterrupted or reducing the extent of flooding in Mexico, but increasing the extent of the disruption of production in Xochimilco. The Aztec managers would have had to reconcile a very much more complicated set of costs and benefits. The diagram makes it clear that floods occurred when the storage capacity of the lake complex was exhausted, and it was the availability of storage that finally set limits on the expansion of chinampa agriculture.  In this closed drainage basin the area within which water levels  were controlled could only be increased by reducing the areas in which water levels could fluctuate freely, by reducing the storage capacity of the system and by increasing the probability of flooding.  Chinampa  91  agriculture could only be extended to the point where the benefits of increased production were offset by the increased probability of flooding.  Whether or not the Aztecs reached this point is uncertain. At  the time of the Conquest the Valley had a population of about 300,000, which was very much higher than it had ever been before.  The historical  record shows a high incidence of famines and it has been estimated that a serious disruption of production occurred about once every four years (Harris 1977:147ff).  The indications are that the population was press-  ing very hard on the carrying capacity of the region and, under these circumstances, it is at least possible that the Aztecs were tempted to increase the area of chinampa agriculture up to or beyond the limits imposed by available storage capacity and that they, like so many of their unfortunate descendants, had no alternative but to overexploit their resource base.  92  CHAPTER IV  THE ECOLOGY OF CHINAMPA AGRICULTURE  The modified lake system was one of the two major physical subsystems of chinampa agriculture. individual chinampa platforms.  The otherwas the multitude of  The basic characteristics of these  major subsystems are quite different.  The lake complex was a large and  unique entity in which each of the lakes had different characteristics and each of the major dykes or canals performed distinctly different functions.  The chinampa platforms, in contrast, were small and numer-  ous and, in terms of their micro-relief and biota, each one was very similar to all the others.  Where the focus of the preceding sections  was on the large areas of the lake complex, the focus of the following pages is on the ecology of a typical, small chinampa platform. The Chinampa Platforms Prescott, in his Conquest of Mexico, published in 1843 has this to say about the chinampas: The chinampas, that archipelago of wandering islands . . . have nearly disappeared. These had their origin in the detached masses of earth, which, loosened from the shores, were still held together by the fibrous roots with which they were penetrated. The primitive Aztecs,.in their poverty of land, availed themselves of the hint thus afforded by nature. They constructed rafts of reeds, rushes and other fibrous materials, which, tightly knit together, formed a sufficient basis for the sediment that they drew from the bottom of the lake. Gradually islands were formed, two or three hundred feet in length, and three or four feet in depth, with a rich stimulated soil, on which the economical Indian raised his vegetables and flowers for the market of Tenochtitlan. Some of these chinampas were firm enough to allow  93  the growth of small trees, and to sustain a hut for the residence of the person that had charge of i t , who with a long pole, resting on the sides or bottom of the shallow basin, could change the position of his little territory at pleasure which with its rich freight of vegetable stores was seen moving like some enchanted island over the water. (Cited in Willey 1939:86). The idea that the chinampas once floated, but have since become rooted, is widespread in both popular and academic literature.  In parts of  the lake complex today there are dense mats of floating vegetation and, i f the load is distributed with boards or bundles of twigs, it is possible to walk on them, and so it may be true that nature offered a hint to the primitive Aztecs.  However, the idea that anything resemb-  ling the extant chinampa platforms once floated is certainly untrue. Some of the early Spanish visitors to the Valley of Mexico described rafts with flowers and vegetables growing on them. They may have been describing canoes loaded with seedlings that were being transported from one place to another, or they may have been describing floating seedbeds as opposed to floating chinampa platforms.  Floating  seedbeds are quite feasible, and in Burma today plants are propagated on a thin layer of soil distributed over the surface of mats of floating aquatic vegetation.  Although the practice is not known in the modern  chinampas, it could have occurred in the past (National Geographic 1974:835 -838). Clavijero, writing his Historia de Mexico in 1780 made use of the earlier sources, but he confused descriptions of boats loaded with seedlings or of floating seedbeds with descriptions of the chinampas themselves and generated a description of the chinampas that is the basis of Prescott's account.  The misunderstanding was endorsed by no  less an authority than Humboldt in his  Political Essay,  and therein-  after it became firmly entrenched in the literature on the Valley of  94  Mexico.  Leicht and Wi1 ley have both shown how the confusion arose, and  several modern authors have demonstrated that the chinampas did not float and could never have done so, but the idea is appealing and tourists continue to be lured to the "Jardines Flotantes" in Xochimilco and continue to be disappointed when they find that the gardens do not float (Leicht 1937; Wilken 1979). Chinampas are no longer constructed in the Valley of Mexico, but it is probable that the earliest chinampas were formed by the excavation of ditches into high water table land, with the spoil from the ditches being distributed over the intervening soil surfaces.  Palerm distin-  guishes chinampas formed in this fashion as "chinampas tierra adentro" or inland chinampas (Palerm 1973:238). As the cultivable area was extended into increasingly high water table land, spoil from the ditches became insufficient to raise the soil surface to the necessary elevation, and volumes were supplemented by additions of the water weed and soil.  The weeds grew in dense and  interwoven mats, known as "cespedes", and sections of these mats were cut and towed to the site of the chinampas where they were sunk with loads, of earth and turf ferried out from the shore.  The operation was  repeated until the material accumulated to a sufficient height about the level of the water (Armillas 1971:653; West and Armillas 1950:174-76). The further the chinampas were extended from the shoreline, the greater the volume of the material required, and where water was a meter or so deep the chinampas were in effect artificial islands resting on the bed of the lake.  This type of chinampa Palerm distinguishes as "laguna  adentro" (Palerm 1973:238).  This description is from historical records,  but it is plausible and confirmed by the practice of modern chinamperos  95  who raise the surface of chinampa platforms with alternating layers of mud and aquatic weeds. Once the necessary volume of material had been accumulated the sides of the chinampas were held in place with "ahuejote" trees.  These  may have been inserted as stakes or piles driven into the lake bed, which happened to take root and.grow into trees, or they may have been propagated deliberately by cuttings, as they are in the chinampas today. Whether the majority of the chinampas were originally of the "laguna adentro" or the "tierra adentro" form is uncertain and depends very much on the depth of the lake at the time that the chinampas were constructed.  It has been suggested that the lakes were extremely shal-  low and best described as marshes rather than as open bodies of water, and that the chinampas were predominantly of the "tierra adentro" type (Sanders et al. 1979:280).  However, if water levels were maintained  artificially high, as was suggested in the preceding sections of this work, the islands may have been formed very largely of imported materials, and been of the "laguna adentro" type.  It is equally possible  that the chinampas were initially formed by excavation, but with the development of dyking they were subsequently raised.  Whether "laguna"  or "tierra adentro," the resulting landscape was very much the same. Chinampas are generally rectilinear, but they occur in a variety of shapes.  Air photographs of Xochimilco show square chinampas, long  narrow chinampas, and chinampas of irregular compound shapes (Figure 22, page 67 above).  Their dimensions vary considerably as well and  range from minute chinampas with only a few square meters of planting area up to large chinampas 30 or 40 meters wide and several hundred meters long.  The literature provides numerous estimates of the normal  96  dimensions of rectangular chinampas.  An excavation of fossil chinampas  shows them to have been about 3 or 4 meters wide by 40 meters long (Parsons 1976:24).  The chinampas that Clavijero saw seemed to him to  be about 6 meters by 18 or 20 meters (Quoted in Leicht 1937:381). Vanegas states that chinampas in Mixquic are traditionally about 8 meters wide and 20 or 30 meters long, while Wilkin gives dimensions of the chinampas in the San Gregorio area as between 3 and 10 meters wide and up to 200 meters long (Vanegas 1978:14; Wilken 1979a:2). Air photographs show that the chinampas in the vicinity of San Gregorio vary enormously in length and range from 25 meters to nearly 400 meters, with a median value of about 75 meters (Figure 23, page 69 above).  In other respects they are quite uniform. They are all long  narrow rectangles, and they are very rarely more than 15 meters wide, with the vast majority being between 10 and 12 meters wide including the strip of riparian vegetation growing, along the sides of the canals. Chinampas in the vicinity of San Gregorio may not be typical of the Aztec chinampas as a whole, but they were probably among the last of the chinampas to be constructed.  If this is the case, they embody a  great deal of experience in chinampa agriculture, and their dimensions may be close to the ideal. It has been suggested in the literature that the height of the surface of the chinampa platform above the level of the water in the surrounding canals is one of the critical dimensions of chinampa agriculture, and. that if the difference between the two levels is appropriately adjusted., water is supplied to the plants by root suction and capillary action, and the need for hand irrigation is eliminated, or at least considerably reduced (Wilken 1979b:4).  This may be the case, but  97  we have few clues, either historical or contemporary, as to what this distance might.be.  One colonial source gives the figure as one "vara",  or 0.84 meters, and another gives it as one "pied", or 0.28 meters (Gibson 1964:256).  The modern chinampas are not a reliable guide. The  pumping of fossil water has caused subsidence, the profile of the landscape has been distorted and the heights of chinampa platforms vary widely.  Many modern platforms are flooded to a depth of a meter or  more, the lowest cultivated platforms are about 20 centimeters above the water level and, at the other extreme, some of the platforms near the shoreline of the antique lake are more than 2 meters above the level of the surrounding canals.  Nor do the modern "chinamperos" offer  information that clarifies the issue.  They do not want their chinampas  to be waterlogged and they do not enjoy lifting water from the canals to a great height, but if their chinampas are between 30 centimeters and-100 centimeters above the water level, they are quite satisfied. If the technique of passive irrigation is to work, then the difference between the levels of platform surfaces and the water table must be related to the depths to which crops extend their roots and the difference should be such that the roots extend to the moist but well aerated capillary fringe, but do not extend into the layer of saturated soils below the water table. Maize was the predominant crop and it probably rooted to a greater depth than the vegetables and flowers that were cultivated in the winter months.  Ideally, water levels would have fluctuated slightly in the  course of the year to accommodate variations in the depths to which the crop plants rooted. This may actually have occurred in the chinampas. Under uncontrolled conditions, water levels were higher in the winter  98  than is the summer.  If the hydraulic installations had the effect of  damping the fluctuations rather than eliminating them completely, then the dampened seasonal, variation in stage would have corresponded with the seasonal variations crop rooting depths.  If variations in stage  were entirely eliminated, then it is probable that the levels of chinampa surfaces would have been adjusted to the requirements of maize, which was the most important crop, and that bucket irrigation would have been necessary in the winter months. These considerations do not resolve the question of the ideal elevation of chinampa platforms, but merely shift it to the problem of how close to the water table crops could be cultivated.  We do not know  how far downwards maize extended its roots; modern hybrid varieties root to a depth of about 80 centimeters but this may not have been the case with the varieties grown in the chinampas (Epstein 1973:50). There are a number of uncertainties about the behaviour of plant roots in different conditions of soil moisture, and the maize grown in the chinampas may have had some resistance to overly moist conditions. The behaviour of moisture in soils is determined by minute characteristics of soil chemistry and structure, and the chinampa soils are not sufficiently well described to allow estimates of how high above the water table the capillary fringe might have extended.  The problem of  how high the chinampas were above the water table is a complex one, and the most that can be said at present is that if there was a difference in elevation of between 25 centimeters and 75 centimeters they appear to function adequately.  99  Many of the canals in the modern chinampas are blocked with weeds or entirely silted up, but they are not entirely obliterated, and on the ground it is generally possible to make a rough estimate of how wide the canals would have been when they were in better condition. A count of 52 canals in the area of San Gregorio and San Luis revealed the distribution shown in Table II.  The figures indicate a  tendency toward two quite different types of canal.  On the one hand,  there are narrow access canals that are just wide enough to permit the passage of the small local "canoas". 80 and 100 centimeters wide.  The "canoas" are usually between  On the other hand, there are the larger  canals that allowed two boars, travelling in opposite directions to pass each other, and that also served as the main drainage canals.  In the  San Luis area, the largest canal that was measured is about 8 meters across, but in other areas of the extant chinampas, there are canals of up to thirty meters. TABLE II WIDTHS OF CANALS IN THE CHINAMPAS OF SAN LUIS AND SAN GREGORIO Width of canal in meters  % of canals  less than 0.75  15.4  0.76 - 1.75  53.8  1.76 - 2.75  17.3  2.76 - 3.75  5.8  more than 3.75  7.7  TOO  A few of the canals are so narrow that they would not have permitted the passage of boats; these invariably run along one side of a chinampa only, so that access is available on the other side.  There  are a number of other canals that are uneconomically wide for use by a single boat, but not quite wide enough for boats to pass each other. However, the overall tendency is towards either narrow access canals or large arterial canals.  The canals are generally quite deep, with values  ranging from about 1 meter in the smaller canals to 3 or 4 meters in the longer ones.  Air photographs show that in the San Luis area there  are about 750 meters of canals per hectare.  This figure, combined with  the data on canal widths, indicates that when the canals were in good condition, they occupied about 13% of the total area of the chinampas. Ahuejote trees are planted along the perimeters of the chinampa platforms and hold the banks of the canals in place.  In well maintained  chinampas, the banks of the canals are vertical and sometimes slightly undercut, and the edge of the cultivable area can extend to within 50 centimeters in plan, of the edge of the water.  If one allows a figure  of 75 centimeters for this distance, then a further 11% of the total area of the San Luis chinampas is. occupied by trees and riparian vegetation. These measurements were made in an area where the chinampas were particularly neatly aligned, where there is only one major canal, and where there are no deep pockets of water that would have impeded colonization, and so they may overrepresent the proportion of land that was cultivated in the entire Xochimilco and Chalco basins.  The calcu-  lations correspond almost exactly with Armillas estimates and it seems 1  clear that about 75% of the area of the chinampas was cultivable, and that of the remaining area about half was occupied by canals and half by  101  vegetation on the borders of the platforms (Armillas 1971:660). Flora and Fauna We are accustomed to making a clear distinction between domesticated species of plants and wild ones, but in traditional agricultural systems the distinction is often unclear and there are many plants that fall between these two extremes.  Some of these are useful wild  plants which are tolerated by farmers even if they are not deliberately propagated.  Others are plants that seed spontaneously but which are  encouraged by selective weeding and fertilization once they are established.  Others are plants that grow wild, but that are on occasions  deliberately planted by farmers to supplement the wild supplies.  In  the chinampas the distinction between wild and cultivated plants is particularly difficult to make because the whole landscape is manmade and because most.of the plants in the landscape are useful to the "chinampers" in one way or another.  In the summary of the flora that  follows, groups are not distinguished so much by their status as wild or cultivated, but by the niches in which they grow.  The plants of  the canals, the platform borders and the platform surfaces are treated separately. In the.modern chinampas the most evident water plant is the water hyacinth known locally as "lirio" or "huachinango" and scientifically as Eichornia crassipes Kunth.  It is a widespread aquatic week  that is native to Brazil and that was introduced to Mexico during the 19th'Century.  It is a prolific plant, and it has contributed to the  decline of the chinampas by clogging canals and by displacing native  102  species that have a higher nutrient content and a greater value as green fertilizers (C.H.C.V.M. 1971). Table III gives a list of some of the more common native aquatics that can s t i l l be found in the less polluted areas of the chinampas. Many of the reed, reddmaces and rushes (Cyperaceae, Typhus spp„ Juncus spp.)  were valuable for their structural properties.  Both the leaves  and stems of these species could be woven into mats, baskets, ropes and other utensils, and the leaves of reedmace can be used to make cloth, although not a particularly satisfactory one (_Bursche 1971:33; Dr. A. Lot pers. com.; Sanders 1979:293).  Huts and protective coverings for  seedbeds were also made from these materials. larly valuable.  Typhus spp.were particu-  The stems are quite buoyant and were probably the main  components, of the fleeting seedbeds discussed above.  Their dried leaves  expand when wetted and are widely used by North American Indians as a caulking material (Erichsen-Brown 1979:213-14).  In addition, the rhizomes  are edible, raw or cooked, and they can be processed to provide a flour that is as nutritious as that of wheat or rice.  They also give high  yields and modern experimental plots have produced 7,000 kilograms per hectare or more (Brown et al. 1976:101).  This use is not confirmed  for the Aztecs, but the reedmaces might have been used as food. The Umbellifers, and Potamogeion spp. and Polygonum spp. are a highly variable group of plants that can adapt to a range of microhabitats.  Polygonum arnphium, the water smartweed, has two quite distinct  forms, and prospers as well in moist but exposed soils as it does in a meter's depth of water, and other plants of this group were nearly as adaptable.  These plants were the main constituents of the "cespedes" or  interwoven mats of floating vegetation which grew on the lakes.  They are  103  TABLE III COMMON AQUATIC PLANTS OF THE CHINAMPAS Mexican Common Name  Scientific name TYPHACEAE  Typhus T.  CYPERACEAE  latifolia angustifolia  Cyperus C.  Bourgaei  hermaphroditus  Eeleodhavis H.  S. JUNCACEAE  GRAMINAE  laoushis  americanus  Junaus J.  palustvis  acioulavis  Scivpus  Tule cuicho  Great reedmace  Tule  Lesser reedmace  Tule grande Tule, Tulillo  esoulentis  C.  English Common Name  bathcus  effusus  Tulillo Cerbatcina Tule, Tulillo  Common spike rush  Estapil  Bulrush  Zacatule Tulillo  Rush  Tulillo  Rush  Hoja flecha  Arrowhead  Hoja flecha  Arrowhead  Berro  Narrow leaved water parsnip  Pasp alum humgoIdtianuum Evagrostis Glycera  ALISMACEAE  fluitans  Sagittaria sagitti S.  mexicana  folia  maovophylla  NAJADACEAE .  Najas  flexilis  UMBELLIFERAE  Berule  eveota  Hydvoootyle vanunculoides H. verticeI  lata  104  TABLE t i l (continued) COMMON AQUATIC PLANTS OF THE CHINAMPAS Mexican Common name  Scientific name UMBELLIFEME  Berro largo  L-Claeopsis ocoidentalis L.  sahaffneriana  POTAMOCETONACEAE Potamogeton P. P.  POLYGONACEAE  '  luaens  Achicorillo cambrai  foliosa  Achicorillo cambrai  Polygonum  Chi Lillo  Water smartweed  Chi Lillo  Pale smartweed  amphibium  lapathifolium  P. acre  Chi Lillo  P.  Chi Lillo  Ladysthumb  Attacuetzon  Water l i l y  persicarioides  Nymphea mexiaana • N.  alba  Nympha  N. ampla N. LEMNACEAE  elegans  Chilecastle  Duckweed  Lemna minor  Chilecastle  Common duckweed  L.  Chilecastle  Gibbons duckweed  Chilecastle  Ivy duckweed  Chilecastle  Great duckweed  Wolff-ia  L.  spp.  gibba tvisulea  Spyrodela  Source:  Pondweed  fluitans  P. pechumtus  P.  NIMPHAEACEAEA  English Common name  polyrrhysa  Sanchez y Sanchez 1978.  105  generally good plants for use as green fertilizers and are s t i l l used today by the "chinamperos" for this purpose.  Many of these plants are  floating leaved and contain pockets of air which are valuable for aerating soils, and they are generally quite high in nutrients. Potamogeton spp. have the property of accumulating crusts of calcium which could allow selective fertilizing of crop plants. and Berula erecta both have medicinal properties.  P. amphibium  The former is used  in treating stomach upsets and the latter is edible and has a marked value (Bursche 1971:99; Dr. A. Lot pers. com.). Several species of Lemnaceae grow on the canals and are known locally as "chilecastle". in the world.  The duckweeds are the most prolific plants  A single square inch of Lemna minor can expand to 1.2  acres in a mere 55 days and Lemna spp. can yield up to 60 kilograms dry weight per hectare per day.  The Lemnaceae are excellent green  fertilizers, both because they have airpockets in the leaves and because they are exceptionally high in nutrients.  Typically, nitrogen consti-  tutes 6% or 7% of the dry weight, phosphorus and potassium 1.4% to 3.0% each, and calcium about 1%. Many species of duckweed are edible. Wolffia spp. have 20% protein, 44% carbohydrates and 5% fats, and some species.are cultivated as a.vegetable in Southeast Asia. Spyrodela polyrrhyza has 35% to 45% protein, compared to 37% for soybeans.  In  general, the edible Lemnaceae are more nutritious than the majority of terrestrial crops (Brown et al. 1976:184ff).  There is no direct evi-  dence that the Aztecs exploited this resource, but they did gather a substance called "tecuitlatl" from the surface of the lakes for consumption ,. and they may have harvested the Lemnaceae as well.  106  "Tecuitlatl" is a substance about which there has been a great deal of confusion.  Sahagun listed it among the lower animals, and  described it as a green or purple scum which grew on the surfaces of the lakes, and which was collected, dried and formed into edible cakes. Gomara adds that the natives "hold the opinion that this skume or fatness of the water is ye cause that such great number of fowle cometh to the lake" (Quoted in Deevey 1957:227).  The substance may possibly have  been Spirode!a polyrrhyza which has leaves that are reddish coloured on the underside, but it is more likely that it was a combination of two types of algae, Spirolina spp. and Oscillatoria spp.  Spiro!ina is a  filamentous blue-green algae that grows in large masses on the lakes and is harvested as a food in the Valley today (Deevey 1957:227-28; Sanders 1979:290; McDowell 1980:727):  It is extremely nutritious, with  70% protein by dry weight, and it contains all eight essential amino acids.  Oscillatoria is also a filamentous algae and it commonly grows  intertwined with Spirolina.  At certain seasons of the year it turns  the water in which it grows to the colour of blood, and it is highly motile.  A clump left on a saucer will climb the sides, and this would  account for Sahagun's classification (Prescott 1978:208).  Both species  are salt loving, and so they were probably restricted to the chinampas of the Lake of Mexico, if they grew in the chinampas at a l l . The narrow strips between the cultivated areas of the chinampa platforms and the edges of the canals support the growth of a large number of different species.  The most evident of these are the Ahuejote  trees, which are a species of willow, Salix bonplandiana Kunth. They are narrow upright trees, with short erect branches, looking rather like poplars.  They grow to a height of 12 meters, although they rarely reach  107  t h i s height  i n the chinampas.  They are g e n e r a l l y spaced about 2 meters  a p a r t a l o n g the margins o f the p l a t f o r m s  which amounts to some 700 o r  800 t r e e s per h e c t a r e , and they are propagated by hardwood The r o o t s are e f f e c t i v e platforms,  cuttings.  i n c o n s o l i d a t i n g the edges o f the chinampa  but they p r o b a b l y do not extend v e r y f a r below the  o f the w a t e r .  surface  In the modern chinampas the t r e e s are i n f e s t e d w i t h a  species of tent c a t e r p i l l a r  moth, Malacosoma a z t e c a , a n d ,  remedies are a v a i l a b l e , the Mexican a u t h o r i t i e s of organizing t h e i r  a p p l i c a t i o n , and so i t  few y e a r s the chinampas w i l l  appear t o be i n c a p a b l e  i s probable t h a t w i t h i n a  be s t r i p p e d o f t h e i r  A h u e j o t e t r e e s ( S r . J . Cornejo p e r s .  although  n e c e s s a r y and g r a c e f u l  com.).  The v a s t m a j o r i t y o f the t r e e s i n the chinampas are S a l i x bonp l a n d i a n a , but o t h e r s p e c i e s are o c c a s i o n a l l y p r e s e n t .  S. l a s i o l e p i s  i s q u i t e common, and S . a c i c u l a r i s and S . acumulata may o c c u r as w e l l . Sambucus m e x i c a n a , a n o t h e r s p e c i e s o f w i l l o w , as h o l d i n g c a n a l banks i n p l a c e i t  can be f o u n d , and as w e l l  has the a d d i t i o n a l  value that  its  l e a v e s a r e the b a s i s f o r an i n f u s i o n used i n t h e ' t r e a t m e n t o f coughs. B u d d l e j a . c o r d a t a and A l n u s f i r m i f o l i a ,  a s p e c i e s o f a l d e r , are a l s o  found (West and A r m i l l a s 1950, 1 9 7 5 ; . S a n c h e z y Sanchez 1978; W i l k e n 1979b:13). The borders o f the chinampa p l a t f o r m s  s u p p o r t the growth o f many  s p e c i e s o f g r a s s e s and h e r b a c i o u s p l a n t s , and some o f the more i m p o r t a n t o f these are l i s t e d i n Table IV. they a r e e d i b l e .  Some o f t h e s e p l a n t s  Chenopodium n u t t e l l i a  are u s e f u l  because  o r " q u e l i t e s " and P o r t u l a c a o l e r a -  c e a , " v e r d a l o g a " , are s e m i - d o m e s t i c a t e s , sometimes growing w i l d and sometimes c u l t i v a t e d  in p l o t s .  These n a t i v e s are on t h e i r way t o  becoming d o m e s t i c a t e s , i n c o n t r a s t w i t h the European c e l e r y  which  108  TABLE IV EDIBLE AND MEDICINAL PLANTS OF THE CHINAMPAS Scientific Names  Notes Lovegrass.  GRAMINEAE Eragrostis  spp.  E.  diffusa  E.  mexicana  Setaria  Humid places Foxtail grasses, related to millet.  spp.  Edible. S.  Edible.  genioulata  Glycera  fluitans  1  "Zacate cerdoso". Grows on waterlogged ground. Grains and florets harvested from 7,11  water surface and used as food Zea  "Tesoutle".  mexicana  "Teociatle".  AMERANTHACENE Ameranthus  spp.  Alteranthera  achyrantha  Common weeds, with edible leaves. "Vendaloga de Puerca".  Edible.  1  COMPOSITAE Ambrosia  spp.  Bidens spp. B.  laevis  Ragweeds; widespread weeds, some of 1 which are edible. Spanish needles, some edible. 1 Burmarigold.  "Te de Milpa", grows 67  B. aurea B.  pilosa  in humid places. ' "Acahual".^ Growing on ditchbanks, "Acahual bianco".^  109  TABLE IV (continued) EDIBLE AND MEDICINAL PLANTS OF THE CHINAMPAS Scientific Names Montanoa  tomentosa  Notes "Cihuaptli" - woman's medicine. Used by Aztecs and today as diuretic; aid to childbirth, lactation, menstru5  ation and possibly effective. Parthenium Tagetes  hysterophorus luaida  Used as an analgesic.  7  Infusion of roots used against colic.'  SOLONACEAE Datura  stramonium  Jimsonweed.  "Toloache": narcotic with  a wide range of medicinal uses. .. exotic. Solanum  nigram  Possibly  8  Used in treating sores and wounds.  7  CHENOPODACEAE Chenopodium spp.  Lambsquarters "Quelite".  Widely used  as leafy greens, growing both wild and 12 C  ambroslodes  3 7  cultivated in chinampas. ' ' ' Wormseed "Apazole"; medicinal uses,  C.  graveolens  cultivated at times. "Epezatl"; used by Aztecs and today in treatment of dysentery, intestinal 5 7  Atriplex  spp.  worms, asthma, and so effective. Similar use as leafy greens.  '  no  TABLE IV (continued) EDIBLE AND MEDICINAL PLANTS OF THE CHINAMPAS Scientific Names  Notes  PAPAVERACEAE Argemone  mexicana  "Chicalote"; "Ampola blanca" herb tea.  A.  ochroleuca  7  "Ampola audi 11a".  Leaves and stems  contain substance similar to morphine, used for eye and intestinal problems.  7  Other  Families:  Plantago  mexicana  Begonia  gracilis  "Acaxicote"; used by Aztecs as emetic, 57 and effective. ' Used as emetic.  Cuphea  augustifolia  Used in treating wounds and tumours.  Anacacia  atroplupura  Rumex spp.  Oxalis  spp.  7  7  Used in treating fever. Pocks and Sorrels.  Used in treating 3 79 sores and diarrhea. ' ' Wood Sorrels. Edible. Diuretics and Antiseptics.  1  Sanders 1974:287.  7  10  7  Sanchez y Sanchez 1978.  2 Erichsen-Brown, 1979:413.  8  USDA 1971.  3 Vanegas 1978:19.  9  Erichsen-Brown 1979:222.  4 Dressier 1953:128.  1 0  Erichsen-Brown 1979:334.  5 de Montellano 1975:219.  1 1  Pohl 1968:103.  6 Wilkinson & Jaques 1972.  Ill  now grows as a weed in the chinampas (Vanegas 1978:21).  Some of the  plants are used as herbs or in the preparation of herbal teas, and others are valued for their medicinal properties.  Some of the latter  have been identified from Sahagun's account of Aztec medicine and others are the basis of folk remedies which may well have been known to the Aztecs (Montellano 1975).  "Teocintle", Zea mexicana, is a  weed that sometimes occurs in the chinampas, and where it grows near maize it cross-pollinates quite freely to produce hybrids that are similar to the early, pre-historic races of maize (Vanegas 1978:21). The modern chinampas are not a good guide to the selection of crops that were grown there by the Aztecs.  Soon after the Spanish  arrived they established gardens near Mexico City and started to raise European vegetables.  These were amenable to cultivation by Indian  practices and, because the market for these exotic products increased along with the growth of Mexico City while the demand for the.native vegetables that satisfied Indian tastes declined along with the Indian population, the new crops spread rapidly through the chinampas. The extent to which the introductions displaced the indigenous crops rather than supplementing them is unclear, but it is certain that the selection of crops now grown in the chinampas is not representative of the selection cultivated by the Aztecs. Table V is a list of common Mesoamerican crops, omitting only those that are cultivated exclusively in the humid lowlands.  Some of  these crops were undoubtedly cultivated by the Aztecs, and others may have been.  The list includes some 48 different species and, while it  includes most of the major crops, it may omit some minor ones. A casual examination of a single garden plot in highland Guatemala showed  112  TABLE V COMMON HIGHLAND MESOAMERICAN CROPS  CEREALS Maize  vars. "Chinampero" (probably several vars.) 3  Zea mays  "Chalqueno Conico". Most varieties are 6 month 1 2 28  hybrids with maize result. ' '  Amaranth  S. Watts, "HUATLI"' "alegria" 3  Amercinthus  leucoparpus  A.  L.; A. sanguineus;  cruentis A.  leucocarpus  A.  tricolor.  is the most widespread; of consid-  able importance to Aztec economy.  Because of  religious associations, cultivation actively sup2 3 4  pressed by Spanish. Panicum  sonorum  BEANS  Leaves also eaten. ' '  Beal panic grass, occasionally grown as a  cereal. "AYECOTE" "frijol" Phaseolus  coccineus  L . , common bean.  P. vulgaris  L . , Lima bean.  P. lunatus  A. Grey, Tepary bean.  P. acutifolius  Macfad.  P.. limensis  SQUASHES  Canivala  L . , scarlet runner, roots edible.  spp.  , Jack bean.^'^  CUCURBITACEAE Cucurbita  ficifolia  Bouche, "chilacayote" oldest of normally  cultivated species. C. mixta,  Pang., green striped, cushaw squash.  C. moschata  Duch., cushaw squash.  113 TABLE V (continued) COMMON HIGHLAND MESOAMERICAN CROPS  C. pepo  L . , summer squash.  Sechium  edule  Langenavia  SW.  "chayote" "CHAyOTLI"  3  (Mol) "tecomate" bottle gourd valuable  sicevavia  for seeds, flowers, fruit and roots.  L.  sicevavia  is African in origin, but it was definitely cultivated prior to the Conquest, and probable that fruits floated across the Atlantic. A further 4 7 89 21 spp. in Mexico, and most are edible. ' ' ' CHILE  CAPSICUM annuum L.  Capsicum  L.  C. fvutescens  TOMATO  C. pubescens  R & p. ' '  Lycopevsicum  esculentum  Physatis  4  8  1 0  Mi 11, "xitomatl" "jitomate"  Brot, "tomatl".  ixocarpa  In Mixquic 2 varieties  recognized: ORNAMENTALS  Polyanthus  tubevosa  Cav, dahlia, edible tuberous roots.  Dahlia  pinnata  Dahlia  coccinea  Tagetes  "Miltomatl", "Coztictomatl". 13 L . , tuberose. 14  evecta  Cav, "ACOCOXOCHITL" dahlia.  "CEMPOALXOCHITL", 'twenty flower' marigold. 4 12 0  \  • Day of the dead. T. patula Tigvidea  J  1  5  '  1  4  Taluana  '  (LE) Kerr, "OCELOXOCHITL" 'jaguar flower'  pavonia  acutifolia,  Chivodendvom  Medicinal properties. '  6  edible c o r n s . ' ' Plumevia  14  3  17  "CACALOXOCHITL" 'crowflower'. ' 4  pentadactylon,  mexicana  15  15  "MACPALOXOCHITL" 'handflower'. '  "YOLOXOCHITL" 'heart f1ower'.  4  15  15  114 TABLE V (continued) COMMON HIGHLAND MESOAMERICAN CROPS  Callandvia  "TLACOXILOXOCHITL" ghost f1ower'. 1  anomala,  15  "COATZONTECOXOCHITL" 'snake flower'. Helianthus  "ITZQUIXOCHITL" sunflower; o i l ;  cinnus,  edible seeds. ' 15  18  "COYOLXOCHITL" edible tuberous roots.  Bomarea edulis,  15  19  "COYOLXOCHITL".  19  B. acutifolia,  OTHER PLANTS Chonopodium spp. \  widely used as leafy greens throughout 21  \ North & South America.  Growing both 12 Atriplex spp. J wild and cultivated in chinampas. C. Nuttalliae Safford, "Quelite" seeds, leaves, unripe 20 28 fruit clusters eaten. ' C. ambrosiodes "Apazote" wormseed with medicinal uses; 20 28 3  cultivated occasionally. C. album,  '  "Quelite" maple leaved goosefoot.  Once thought  to be introduced, but now known as native. Most 21 widespread of "Quelites", and cultivated. 12 Vovtulaea spp. growing both wild and cultivated. 22 Salvia hispanioa L . , "Chia" seeds used to prepare drink. 3  I  Hyptis suaveolens Port., Crotalaria longirostrata  o  "Chia grande". Hook & Arn., "Chi pi 1 in" leguminous  23 pot herb. 24 Euphorbia  pulcherrima  Mentha spp.  M. spicata  Wilid., "Noche Buena", Spurge. spearmint, is commonly cultivated  now, is not native, and probably replaced • 26  M.  avvensvs.  115 TABLE V (continued) COMMON HIGHLAND MESOAMERICAN CROPS Allium  oepa  L . , "XONOCATL", onions.  27  14 Gossypium  hirsutum  3  "IZCATL", cotton.  Notes: Mexican common names, upper and lower case in double inverted commas.  a  k Nahuatl name, upper case in double inverted commas. Literal translation of Nahuatl name, upper and lower case in single  c  inverted commas. Sources: 1  Cornejo, J . , pers. com.  16  Dressler 1953:147.  2  Vanegas 1978:21.  17  Ibid.:149.  3  Dressier 1953:121.  18  Ibid.:136.  4  West & Armillas 1950:181.  19 lbid.:125. Ibid.:128 Iy  5  Dressier 1953:141.  20  6  Ibid.:142, 143.  21  7  Ibid.:130, 131, 136, 146.  22  Dressler 1953:146.  8  Vanegas 1978:18.  23  Ibid.:130.  9  Cutler & Whitaker 1961:470.  24  Ibid.:236.  10  Dressler 1953:126.  1 1  Ibid.:132.  Vanegas 1978:19.  12  Erichsen-Brown 1979:413.  25 Sanchez y Sanchez 1978:330. Erichsen-Brown 1979:285. 26  27  13  Dressler 1953:144.  28  14  Ibid.:132.  29  15  Schilling 1938:27.  Peterson 1962:167.  Sanders 1979:234, 236.  Vanegas 1978:24  116  that some 26 different species were being cultivated, and a survey of garden plots in Tabasco showed that 143 different species were present (Anderson 1952:136-37; Gliessman 1979b:7).  While the chinampas may  not have been quite as diverse as the gardens of the humid lowlands, it is probable that a very large number of different species were cultivated. level.  The list gives no indication of the diversity at the varietal A single visit to the market in Xochimilco produced 14 different  varieties of Capsicum, al1 of which were reportedly grown locally. The six commonly cultivate species of beans are liable to great variation, as are the six commonly cultivated species of squash.  Cucurbita ficifolia  in particular is liable to hybridization with other species which results in many novel forms, and the variety is increased yet further by the presence of 21 other native.species of Cucurbita that are not normally cultivated, but that are nevertheless edible (Cutler and Whitaker 1961: 470). Maize was and is the single most important crop grown in the chinampas.  The variety that is most commonly cultivated is a six month  maize, known as "Chalqueno conico", and it is possible that a number of other local "criollo" varieties were grown as well. crop that is not well known in temperate latitudes.  Amaranth is a  It is usually  cultivated for its seeds, but it is valuable as a leafy vegetable as well.  The brilliantly coloured seeds were employed by the Aztecs in  ritual activities and for this reason the Spanish actively suppressed the cultivation of Amaranth, but it was an important component of the Aztec economy.  One of the lists of tribute due to Tenochtitlan includes  200,000 bushels of Amaranth, in addition to 280,000 bushels of maize and 230,000 bushels of beans, which is an indication of the relative importance  117  of these crops (Dressier 1953:121).  Chiles, squashes and tomatoes were  all cultivated'in the chinampas, but it is probable that the species of tomato that is familiar to us, Lycopersicum esculentum, was less widely cultivated than the Husk Tomato, Physalis ixocarpa:  Various species  of Portulaca, Chenopodium and Atrip!ex were probably cultivated as green vegetables.  Xochimilco was a centre for floriculture, as it is today.  Many species were cultivated including dahlias which the Aztecs called "Acocoxochitl" and marigolds or "cempoaloxochitl"; the latter had important ceremonial functions and still adorn graves on the Mexican "Day of the Dead". In addition to its diverse flora the chinampa landscape was home to a large number of animal species.  The following is a compilation of  those that are of particular interest or economic importance. Several groups of insect were important. valued because they were edible.  Some of these were  "Axayacatl" were the Coryxid water  bugs of which five species were of some economic importance: Ahuetlea mexicana.Llave., Krisousacorixa azteca.Jacewski., K. femorata, Corisella texcocana and C. mercanaria were all valuable for their eggs.  These  were called "ahuatli" or "water amaranth" and were cultivated by the Aztecs.  Stems of reeds were placed in the water, and on these the bugs  laid their eggs.  The stems were then removed, and the eggs were stripped  off and'formed into small cakes.  "Aneneztli" were the eggs of dragon-  flies which were also collected and eaten.  "Cuculin" were the pupae  of salt flies, Ephydra hians, and in the late summer these came to the surface of the lakes in great numbers and were collected.  "Izcuaatli"  were the larvae of the same species, which were "little grubs like earthworms [and were] slender and pullilant in their multitude and  118  density."  These, too, were collected and eaten (Deevey 1957:224-28;  Gibson 1964:339-41).  Also included among the insects are a number of  species that we regard as crop pests.  The illustrations to Sahagun's  Historia show insects that are clearly identifiable as grasshoppers, the larvae of June Beetles, Phyllophaga spp., leaf-eating catepiliars of various types, corn ear worms, which are the larvae of the Noctid moth Heliothis zea, and the larvae of the Geometer moths.  The Nahua term for  the latter-is "tetatamachluhqui" which translates literally to our common name for the creature, which is "measuring worm" (Curran 1937). Although not directly confirmed, it is undoubtedly the case that aphids, leafhoppers, bean beetles, cucumber beetles, leaf beetles and other common pest species were present in the chinampas. The Crustaceans were represented by the "acocile" or freshwater crayfish, Cambarellas moctezumae, and by a species of freshwater shrimp, either Macrobrachium sp. or Palaemos sp.  Among the amphibians, frogs  were caught and eaten, together with the "axolotl" which is the larval form of the. salamander Ambystoma sp.  These grew to about 8 inches in  length and were rolled in a maize dough and roasted.  Water snakes and  turtles were also hunted (Gibson 1964:339-41; Flannery 1968:83-84; Niederberger 1979:133-34, 139). Fish were an important resource.  A large proportion of these  were Athernidae, including Chirostoma humboldtianum, C. Jordani and C. pegani, some of which were the freshwater whitefish and others the saltwater "charales".  Fish of the minnow family, Cyprinidae, were  important' including Notropis azteca, Evarra spp. and Algansea tincella. The Goodeid'Giradinichthys viviparus were also caught.  The Aztecs  distinguished species as "xohulin", "ajolote", "amilotl" and "iztacmichin",  119  but the identification of these is not clear.  An estimated one million  fish were caught annually during the 17th Century, and at least this number may have been caught by the Aztec fishermen.  The older living  chinamperos report that in their youth the occasional fish, caught in the chinampas, was one of the small advantages of their agricultural systems (Deevey 1957:224-28; Gibson 1964:339-41; Vanegas 1978:41; Niederberger 1979:133-34). Migratory birds came to the Valley of Mexico in large numbers during the winter, having passed the summer in prairie provinces of Canada, and these waterfowl were a further resource for the Aztecs (Deevey 1957:224-28; Gibson 1964:339-41; Vanegas 1978:41; Niederberger 1979:133-34).  Sahagun enumerated some 40 varieties of duck and geese  which were commonly hunted by the inhabitants of the Valley. A number of species of birds other than these also inhabited the chinampas. Although they are not mentioned in the literature and their economic importance may have been slight, they were probably of some value in controlling insect pests (Sr. J . Cornejo pers. com.). The foregoing description of the chinampa platforms and of the biota that they supported gives some indication of the structural properties of the chinampas as an ecological system.  The chinampas  had some 700 to 800 Ahuejote trees per'hectare, which would have amounted to a considerable standing biomass.  Waterweeds on tropical  lakes can accumulate to about 70 tons per hectare, dry weight and, although the figure would not have been as high in the well-maintained canals and in the cool conditions of the Valley of Mexico, aquatic vegetation may have contributed significantly to the total biomass (Brown et al. 1976).  In traditional systems in Tabasco, where maize  120  is produced by united cropping techniques, the standing biomass of cultivated crops can amount to 19.5 tons per hectare, dry weight, and the figures in the chinampas may have approached this value (Gliessman & Amador 1979b:Tables 1 & 2).  In Mixquic, the annual yields of the sale-  able portions of vegetables average 3.9 kilograms per hectare, and as vegetables are cultivated during the winter months, this suggests that the biomass of the crop system may have been quite high throughout the year. The chinampas would appear to have rated quite high on most indices of diversity.  The microtopography was varied, and they supported  plants of different life forms and different dimensions, which would confer a high pattern diversity on the system.  The preceding tables  list 118 different species of plants that were likely to have grown in the chinampas, and so the species diversity of the system was probably quite high.  Many of the plants had medicinal properties, or were  used as herbs or spices, which suggests that the biochemical diversity of the system may also have been quite high.  The chinampas supported  some perennial plants and annuals were.present throughout the year, with the result that the diversity of the system persisted through every season.  In general, the information available on the chinampas  suggests that, for an agricultural system whose primary purpose was the production of maize, they supported a substantial and varied biomass, and did so throughout the year. Seedbeds and transplanting Agriculture can be considered as two linked endeavours, one being to promote the growth of plants that are particularly inept at propagating  121  themselves, and the other being to suppress the growth of plants that are much stronger competitors than the crops, and of populations of animals against which the crops have few defences of their own. In the chinampas seedbeds, known by the Spanish terms "camellone" or "almaciga", were at the heart of the set of techniques for propagating crop plants.  Seeds were started in the beds while other  crops were maturing in the fields.  When the mature crops were harvested  the seedlings were planted in their place, and new seeds were started in the beds.  In this fashion each plant was accorded only as much  space as it required at any particular stage of its development, and the physical infrastructure of the chinampas was used to its maximum advantage. The Aztecs brought the use of seedbeds to a high art, and the techniques that they developed are s t i l l used in the chinampas. A small plot is selected for the construction of a seedbed; in this plot the soil is carefully leveled, tilled and weeded.  A board is held  on its edge along the edge of the seedbed, and soil is gathered from outside the bed and pressed into the angle formed by the board and the ground.  The operation is repeated until the whole of the plot is  surrounded by a low earth wall 15 centimeters"to 20 centimeters high, which is vertical on its inside face and sloping on its outside one. Work then shifts to the canals where "agualodo" or bottom mud is scooped up and ladled into boats.  A cloth scoop attached to the end of  a long pole, called a "zoquimaitl" is used for this job.  The bottom  mud is. then transported to the site of the seedbed and all the roots and other debris'are removed.  Water is mixed in with the mud until it is  quite liquid, and then the mud is ladled into the seedbed.  In the past  122  a "zoquimaitl" or a bucket made out of the central part of the maguey cactus was used, but today plastic buckets are the rule.  Because the  mud is quite liquid it flows over the area contained by the earth walls and forms a perfectly level surface.  The seedbed is then left overnight,  during which time much of the water in the liquid mud drains into the prepared soils below, and the level of the mud falls by about one third. The mud that remains is quite firm and cohesive, but with a spongy and resilient quality to it.  The layer of mud is then sliced at small and  regular intervals along both dimensions of the bed to form a number of uniform blocks.  The cuts are made with a knife attached to the end of  a long handle or with a device called an "almagigero", which consists of between 5 and 10 knives fixed to the end of a single handle. The resulting blocks are called "chapines" or, in Mixquic, "tlacopehuales" (Vanegas 1978:3). Measurements are made throughout the operation using fractions of the hand with fingers extended downwards to measure the depth of the "agualodo" when it is poured into the seedbed, and multiples of fingerwidths to determine the distances between the cuts.  The resulting  "chapines" are closely matched to the sizes of the plants that are to be grown in them and are usually about five fingers square by six or seven fingers deep.  But occasionally they are minute, about one and a  half fingers square by two deep, or quite large, eight fingers or so square and ten or more deep. Once the "chapines" have been cut, a board is placed across the seedbed to distribute the weight of the "chinampero", and a hole is made in each "chapine" with a finger or a twig.  A single seed or  several seeds are placed in each hole, a pinch of fertilizer may be  123  added, and the hole is closed and smoothed over.  Most plants are seeded  individually, but for small herbs half a dozen seeds may be placed in a single hole. When the seedlings have grown to f i l l the capacity of the "chapine", the "chapines" are broken apart.  They separate easily along  the lines of the cuts and from the friable soil beneath, and the blocks keep their form so that the seedlings can be moved without disturbing their roots.  Some crops are exported from the chinampas at this stage.  Others, particularly flowers, may be planted into larger "chapines" and vegetables are planted out into larger plots. sowing and transplanting varies.  The interval between  "Cilantro" or coriander is transplanted  after about 20 days, maize after 30 or 40 days, onions after 60 days, and chiles after 90 or 100 days (West and Armillas 1950:77). The fields are carefully prepared to receive the transplants. A green fertilizer of aquatic vegetation is dug into the soil, either over the whole area.of the plot, or only in the spots that are to receive the transplants.  Sometimes furrows are formed so that irrigation water  will flow to the roots of the plants, and maize is planted in small mounds of soil as is generally the practice in Middle America. When the hole is excavated a handful of "agualodo" and compost may be added before the seedlings and their "chapines" are deposited. The success of the technique depends on the quality of the "agualodo", and the modern "chinamperos" will travel a considerable distance to secure supplies of mud that they consider to be of good quality.  A few samples of chinampa muck have been analyzed, and one that  was considered to be of good quality was low in nitrogen, which is characteristic of submerged soils,, but was otherwise rich in nutrients (Wilken 1979a:Table 1).  The "agualodo" has a high cation exchange capacity  124  which facilitates the uptake of nutrients by plants (Vanegas 1978:11). Very few samples of chinampa mucks have been analyzed, and the modern soils may not be representative of pre-Hispanic conditions because they are to some extent enriched by artificial fertilizers in the runoff from the platform surfaces; but it seems probable that the "agualodo" was generally well supplied with plant nutrients. ture of the bottom mud is important.  The texture and struc-  Samples from the San Luis area  were smooth textured and very sticky and cohesive, which is the property that allows the muck to.be formed into "chapines".  They are also very  dark and have a high organic matter content, over 60% by dry weight in one sample.  The organic matter is important not only because i t is a  source of nutrients, but because it imparts to the mucks their light, spongy character and allows the roots of the seedlings to develop freely. The "chinamperos" of San Luis complain that the local "agualodo" has an increasing tendency to stunt the growth of seedlings by constricting their roots.  This may be related to the death.of a large number of the  Ahuejote trees, whose falling leaves are an important constituent of the organic matter in the chinampa mucks (Gliessman pers. com.). In addition to the seedbeds that can be seen today on the platform surfaces, the Aztecs probably made use of the floating seedbeds mentioned.above.  These were the features that Clavijero confused with  the chinampas. themselves to create the myth of the "floating gardens". They are mentioned by Padre Jose Acosta in his Historia Natural y Moral de las Indias and they are described by Fray Hernando de Ojea, who wrote in 1608 as follows:  125  In this lake (Lake Chalco) the Indians make use of a very interesting thing which are movable gardens some 20 or 30 "pieds" (6.6 - 10.2 m) long, and as wide as they please, lying on the surface of the water, on waterweeds, rushes and reeds, on which they sow seedbeds with their vegetables, such as pimiento, lettuce, cabbage etc., for transplanting elsewhere, and these can be moved with cords, from one place to another. (Quoted in Leicht 1957:377. Trans. A.J.R.) The description is perfectly clear, and both of these authors are quite definite that the seeds were grown on these improvised rafts and not simply transported on them.  The same authors provide descriptions of the  chinampa platforms themselves and, as these are accurate, there is no good reason for supposing that their descriptions of floating seedbeds are not similarly accurate.  Technically the device is quite feasible  and is in use in other parts of the world today, and so it seems probable that the Aztecs did employ floating seedbeds but that the practice has since been abandoned. Seedbeds are s t i l l used for the cultivation of flowers in the modern chinampas, but most other crops are sown directly into the fields. The chinampas are an intensive agricultural system in decline and it is characteristic.of such systems that they become less intensive as the decline proceeds.  On these general grounds it is likely that seedbeds  were much more thoroughly exploited in the past than they are today. This seems to be confirmed in the case of maize, which is now sown directly into the fields from which it is to be harvested.  Many "chinam-  peros" report that they used to grow maize in seedbeds but that they now no longer do so, and West and Armillas note that during the 1940 s 1  maize was grown in seedbeds in Mixquic which is rather more remote from Mexico City than the other extant areas of chinampas and rather more  126  conservative in its practices, but that it was not so cultivated in Xochimilco (West and Armillas 1950:1).  Fray Alonso Ponce who travelled  in New Spain between 1585 and 1587 notes that maize was cultivated in seedbeds in the chinampas, and that the practice was peculiar to the area (Leicht 1937:377).  Because it is so great a departure from normal  Middle American practice, maize is the least likely of all crops to be started in seedbeds, and yet it seems clear that it was so planted.  If  this is the case it is likely that virtually all crops in the past were started in seedbeds. There is some linguistic evidence that the Nahua word "chinampa" or "chinamitl" should be translated as "seedbed" and that it was not used to refer to the agricultural platforms.  The latter were probably  called "tlateles", which is a derivation of the term for "a pile of earth".  The fact that we now use the word "chinampa" to refer to the  platforms may arise from a misunderstanding on the part of the early Spanish colonists (de Lameiras 1974:22-23).  Whether physically or  etymologically, seedbeds are very closely bound up with the practice of chinampa agriculture. Weeding and Pest Control Seedbeds were elaborate and successful devices for propagating populations of crop species that were not well equipped to propagate themselves.  The devices used by the "chinamperos" for suppressing the  populations of unwanted species were very much less elaborate. Weeds do not appear to have been a great problem in the seedbeds. The ground upon which the beds were laid out was carefully weeded, and once the "agualodo" was in place it suppressed the growth of weed  127  seedlings.  The mud which was dredged up from the bottom of the canals  may have contained few fertile weed seeds, and the layers of grass and other materials that were laid over the surface of the seedbeds may have helped prevent invasions by weeds during the few weeks of the life of the seedbed.  In general the seedbeds in the modern chinampas seem to  be quite free of weeds. In the modern chinampas, the crops are mulched when they are planted out, which suppresses the growth of weeds to some extent. nevertheless, weeding is necessary.  But,  The modern "chinamperos" weed their  maize about a month after it is sown when the plants are about 50 centimeters high, but most "chinamperos" do not undertake any subsequent cultivation.  It is probable that in the past the maize was more frequently  weeded than it is now.  The Aztecs may have weeded their crops at inter-  vals of about six weeks, which is the case with other intensive agricultural systems in Middle America today.  During the winter months  weed growth is to some extent held back by declining temperatures but, even so, weeding is a chore that has to be carried out occasionally. In the modern chinampas something like half of each year's crop of maize is lost to insect pests and plant diseases (J. Cornejo pers. com.).  We know from Sahagun's works that many of the pests that cause  this damage were present in the Aztec chinampas and yet it appears that, although the potential for damage existed, it was not realized. There are no historical references to damage by pests, and the older "chinamperos" today recall that in their youth pests were simply not a problem. Nor was this because the "chinamperos" had some curious and efficient means of eliminatingunwanted animals; as is generally the case with  128  traditional agricultural systems in Middle America, the repertoire of agricultural techniques includes no methods or devices for combating insect pests other than pinching the occasional caterpillar off a plant if it happens to be noticed. The means by which populations of insect and other pests are regulated in traditional agriculture is a topic that is not well understood.  Even if such a practice is not imposed by the climate, most  traditional systems interrupt production during one season of the year. One of the benefits of this practice is that populations of pests die back somewhat during the period when the crops are not present in the fields.  In the chinampas the seasonal alternation of maize with  horticultural products would have had this effect on populations that were entirely dependent on maize.  The fact that the winters in the  Valley of Mexico are quite cold may have further regulated pest populations, but it is evident from the condition of the modern chinampas that the climate alone is not capable of preventing unacceptable depradations. There is mounting evidence that a diversified flora in the vicinity of crops is an important element in the regulation of populations of pests.  If there is a diversity of microhabitats, popu-  lations of predatory and parasitic species can survive during those periods when crops are not in the fields, and when populations of the pest species upon which they depend are at low levels.  When the crops  are reestablished, the pest species multiply but, because the predatory and parasitic species are available in the vicinity, populations do not increase to epidemic proportions.  129  There are difficulties with this theory, one of them being that the diversified flora that allows the predatory and parasitic species to survive may also contribute to the survival of the pest species. However, workers in the field are becoming increasingly convinced that it is mechanisms of this sort that account for the low levels of crop damage by insect pests in traditional agroecosystems (Janzen 1973:1214; Gliessman et al. 1978).  As indicated in the preceding sections, the  chinampas have a highly diversified flora and it seems probable that these biological control mechanisms were at work on behalf of the Aztec agriculturalists. A similar set of mechanisms at a different scale may also act to regulate populations of soil pathogens. important role.  Birds, too, may play an  One of the reasons cited for the current problems of  the chinampas is that bird populations in the district have been hunted almost to extinction (J. Cornejo pers. com.). Also related to the control of pests is the attitudes of the farmers themselves.  Farmers in the Yucatan accept that a certain pro-  portion of their crops will be lost to pests.  They anticipate the loss  by planting something additional "for the birds," and the attitude appears to be quite general among Indian farmers.  Attitudes about the  pests themselves are also rather different from our own.  According to  Sahagun, the Aztecs considered harmless many of those species that we regard as pests.  The corn ear worm, that we look upon with disgust, was  seen by the Aztecs as a tasty morsel that occasionally appeared in the ears of corn (Curran 1937).  The Aztecs, along with other traditional  farmers, accepted a level of damage that we would not.  130  We hardly understand these mechanisms at a l l , but in the absence of any plausible alternative it seems that the Aztecs relied chiefly on the "balance of nature" to regulate populations of unwanted organisms. The balance is a fragile one and in the modern chinampas it has been disrupted, but in the past it appears to have maintained populations of pests and pathogens at acceptable levels. Material Cycles Ecosystems are open with regard to materials.  They receive  materials from their environment, cycle them within the system for a greater or lesser period, and then discharge them from the system. Water entered the chinampas by the various routes discussed at considerable length in the preceding section of this work. Once within the chinampas it was stored in the canals or in the underlying aquifers, and it was cycled by capillary action within the chinampa soils and by irrigation. The most obvious need for irrigation is to satisfy the water requirements of crop plants and to allow their cultivation in those areas and at those times of year when rainfall alone is inadequate. In climates when evaporation rates are high, irrigation is also required to wash away the salts that accumulate on the soil surface.  In the Lake  of Mexico, where the water was brackish and evaporation rates were high, this may have been a major problem.  Irrigation in the chinampas is now  carried out with plastic buckets, watering cans and mechanical pumps, but in the past maguey buckets may have been used and the "chinamperos" also employed a large wooden spoon attached to the end of a long pole, which is called a "texpetlatl".  131  The Comision Hidraulica de la Cuenca Del Valle de Mexico (C.H.C.V.M.) has made an estimate of the consumptive use of water for different crops grown by modern methods in the vicinity of the chinampas, and the results are shown in Table VI. It is clear that rainfall obviates the need for irrigation during the summer months, except during August where there is a slight decline in precipitation values. During the winter months when precipitation values are low, irrigation is required. Maize is irrigated during April and May to insure that the seeds' germinate whether or not the rains are on schedule.  The fact  that it is not irrigated during the "Veranillo", when other crops are, is attributable to the fact that it transpires water at a relatively low rate.  The figures estimated by the C.H.C.V.M. represent the needs of  crop plants grown at low elevations in the southern parts of the Basin of Mexico, and they probably overrepresent the consumptive use of water in the chinampas. Mulching is an important element of chinampas agriculture. In Mixquic, seed beds are sometimes covered with a layer of finely sifted soil as soon as they are planted.  This protective layer is left in  place during the few.days that it takes the seeds to germinate, after which it is carefully swept away and replaced with a layer of grass (Vanegas 1978:16).  In other areas of the chinampas, the initial appli-  cation of soil is omitted, but seedbeds are covered with any materials that will protect the seedlings, including plastic sheeting or old newspapers, as well as grass or leaves.  When plants in the "chapines"  have grown a few inches, a framework of branches or weeds is erected over the seedbeds, and this is covered with grass, leaves or other materials.  When the crops are planted out, a mulch of chopped up water  132  TABLE VI CONSUMPTIVE USE OF WATER IN CENTIMETERS FOR LOCALITIES AT LOW ELEVATIONS IN SOUTH OF VALLEY OF MEXICO Flowers  January  18  15  -  15  February  18  15  -  15  March  18  15  12  15  April  18  15  6  15  May  -  -  -  -  June  -  -  -  -  July  -  -  -  -  August  18  15  -  15  September  -  -  -  -  October  -  -  -  -  November  18  15  -  15  December  18  15  -  15  Source:  Maize  Clover  Vegetables  C.H.C.V.M. VI:135.  weeds is applied, either over the whole surface of the plot or in the immediate vicinity of each plant.  The more delicate plants are pro-  tected with cones of staw, grass or newspaper, which are known as "abrigos" or "overcoats". These mulches and coverings are valuable because they provide nutrients for plants, suppress weed growth to some extent, and provide protection against low temperatures during the winter months, but one of  133  their chief effects is to reduce evaporation losses and to prevent the soils and seedbeds from drying out.  The soils of the chinampas  have good tilth and a high organic matter content, and the Ahuejote trees reduce windspeeds and provide shade, both of which features further reduce the rate at which water is lost by evaporation.  With regard to  the latter point, it may be significant that the chinampas of the San Luis area are laid out on an approximate north-south axis, which maximizes the extent of the shaded area. The initial fertility of the chinampas was high, and the Aztecs maintained and enhanced this fertility with a variety of other nutrientrich materials.  Water weeds were used as a green fertilizer, and the  duckweeds in particular had a high nutrient value.  Both waterweeds and  crop residue are composted in heaps on the modern chinampas, and there is some linguistic evidence that the practice is pre-Hispanic. The nutrient value of leaf mould and rotten wood was understood by the Aztecs. They distinguished soil enriched with these materials as "Quauhtalli"; the materials are imported for use in the modern chinampas, and again it is probable that this is a continuation of an old established practice.  The Aztecs had no large animals, but turkey manure, "totolcuit-  latl", was used as a fertilizer and bat dung was imported from caves at Ixtapalapa, and from Guerrero and Morelos, for use in cultivating chiles (West and Armillas 1950:167-68; Gibson 1964:306; de Lameiras 1974:25, 28; Sanders 1979:290).  The Nahua words for "fertilize" and  "defecate" are derived from the same root, and the Aztecs were undoubtedly aware of the value of human waste as a fertilizer.  Settlements  were more widely dispersed in the 16th Century chinampas than they are today, and undoubtedly night soil was informally distributed around the  134  as is the case with the home maize fields in many parts of Middle America today.  There is also some evidence that sewage was system-  atically collected from dwellings and public latrines in the larger towns and delivered in boats to the chinampas for use as fertilizer (Armillas 1971:654).  There are two difficulties with this practice,  one being that there is a danger of spreading pathogenic bacteria, and the other being that fresh sewage has a high initial toxicity for plants and can kill them completely if concentrations are too high. Both these problems can be overcome quite simply by anaerobic digestion; if the material is composted for one or two months in pits, it is quite safe to use as a fertilizer, but we have no information on Aztec practices in this respect (Chaney 1973:69).  Quite a wide range of materials  were used as fertilizers by the Aztecs.  These are listed in Table VII,  together with an indication of their nutrient content. TABLE VII NUTRIENT CONTENT OF ORGANIC FERTILIZERS Material  %  Bat Guano  1  P  1.5 - 2 3  1  0.85  0.45  1 - 3  0..5 - 4.,0  0.0 - 0.5  6 - 7  Trierweiler and Utzinger 1975:18. Brown et al. 1976:150.  2  4  % P 2 Duckweed  % K0  6 1  1  ° 2°5  4..5 - 8..5  1  Sewage Sludge (digested)  2  5 - 10  Poultry Manure  Poultry Manure (fresh)  N  1..4 - 3.,0  % K 1.5 - 3.0  135  Of all the macronutrients, nitrogen is the one that is most frequently a limiting factor on crop production. Atmospheric nitrogen is ubiquitous but it cannot be used in this form by plants, and the materials such as bat dung, bloodmeal, fish meal and bone meal in which there is a high concentration of nitrogen in a form suitable for applying to fields are quite scarce.  A common solution to this problem is  to incorporate legumes or one of the other species of plant that are host to nitrogen fixing bacteria into the crop cycle, either by rotating them with the main crop or by planting them as companions of the main crop.  The chinampas were used intensively and it is very unlikely  that maize was rotated with a leguminous crop on an annual basis.  Cer-  tainly, air photographs taken in 1947 and the photographs accompanying Schilling's 1938 article on the chinampas show that virtually all of the cultivable land was planted to maize during the summer months, which precludes the possibility of a regular annual rotation.  It is possible  that a crop such as beans was rotated with maize on a seasonal basis, and that at least part of the land was planted to beans during the winter months when maize was not in the fields.  Alternatively, maize could  have been grown with squashes and beans as companions, which is a widespread Middle American practice.  This practice is not common in the  modern chinampas, but it is hard to say whether this is because the chinampas are now being cultivated less thoroughly than in the past or because the practice never occurred in the chinampas.  Measurements made  in Tabasco show that the yield of maize cultivated with beans and squash is about 10% higher than when it is cultivated alone.  If the Aztecs  exploited every available means to increase the yields of their chinampas, then it is probable that polyculture was part of their repertoire of agricultural techniques (Gliessman and Amador 1979:Tables 1 and 2).  136  The structure of the chinampas and the techniques of chinampa agriculture tended to promote a cyclical and conservative use of nutrients (Figure 30).  The canals functioned as a repository for  nutrients, which entered by surface runoff from the surrounding h i l l sides or by leaching and runoff from the surface of the platforms. Dissolved nutrients were incorporated in the tissue of aquatic plants, and the water in the canals moved so slowly that even very fine suspended solids accumulated in the bottom mud. Nutrients were removed from the canals in irrigation water, and by the collection of water weeds for use as mulches and green fertilizers, and of bottom mud for the construction of seedbeds.  Crop residues and water weeds were com-  posted and applied as fertilizer.  Any materials that decomposed to  such an extent that they could not be gathered up were finally returned to the canals in runoff and accumulated there for future application to the crops on the platform surface.  There was undoubtedly some loss of  nutrients to the system in slow drainage of water, but this loss was made good by inputs of runoff from surrounding hillsides.  Nutrients  were also lost in the exportable components of crops, but if it really was the case that sewage sludge was collected, then a proportion of these would have been returned to the chinampas, and any shortfall made up by the import of other fertilizer materials.  No precise data  are available, but it seems clear that the chinampas made a very conservative use of nutrients. Agricultural Calendar The modern chinampas are not cultivated as intensively as they were in the past, but the general sequence of agricultural activities  137  Figure 30.  Schematic Representation of Nutrient Transport in Chinampa Agriculture.  138  appears to be much the same now as it was during the 16th Century. This is broadly confirmed by the sequence of Aztec ritual activities, about which we do have some information.  Most of the major events of the  agricultural year were maked with ceremonies.  These are shown in  Table VIII, together with notes on the major activities in the modern chinampas, supplemented by the few historical references that are available. It is clear that the agricultural calendar was dominated by the production of maize.  In the modern chinampas the preparations begin in  February, when the last of the winter crops are removed from the fields and the ground is prepared for sowing.  In the 16th Century, when maize  was started in seedbeds, it is probable that the winter growing season was extended into March.  The ritual activities in February were largely  concerned with assuring an adequate supply of rain for the spring planting.  In March and April, the maize was sown either directly into the  fields as it is now or, more probably, into seedbeds.  The sowing was  anticipated slightly by the festival of Xipe Totec, the god of spring and regrowth, and various ceremonies involving the adoration of corn deities and of the seed corn itself were performed.  The maize was  transplanted in May, 30 or 40 days after it had been sown, at about the same time of year as the modern "chinamperos" weed and fertilize their fields.  At the end of May children were sacrificed to the rain god  Tlaloc, presumably in an effort to insure the continuation of the rains. June is a slack period for modern "chinamperos", but the Aztec farmers may have weeded their fields at this time.  Young ears of corn were con-  sidered a delicacy, as they are today, and during July they became available.  The Aztecs marked the season with a sacrifice, after which  TABLE VIII AZTEC AGRICULTURAL AND RITUAL CALENDARS  Agricultural Month  J A N U A R Y  Activities  7  No. 17  of winter crops  goddess. 18  1  squash planted  p  6.  "Old Princess".  Sacrifice of woman impersonating  8. Women and children beaten to make them cry, stimulating rain.  1. IZCALLI.  2. "Resuscitation".  5. Firegod.  8.  3. XIUHTECUHTLI.  4. "Lord of the Year".  At the end of this month were five "useless days",  prepared for planting  1. ATCOUALCO. 2. "Lack of water".  3. CHALCHIUHTLICUE.  5. Goddess of rivers, lakes, and oceans.  ground  c  3. LLAMATECUHTLI. 4.  NEMONTEMI, required to correct the solar calendar.  Y I  H  2. "Severe weather".  5. Creation and fertility goddess.  r  M A R  1. TITITL.  harvest  p E B R U A R  Aztec Solar Months and Ceremonial Activities  2  4. "Jewelled Robe.  7. TLALOC, "He who makes things  grow" in charge of rain, hail, thunder, floods.  8. Sacrifice of a child.  1. TLACAXIPEUALIZTLI.  3. XIPETOTEC.  2. "The flaying of men".  the Flayed One". 5. God of spring and regrowth.  4. "Our Lord,  6. Priests dressed in the  flayed skin of sacrificial victims, representing the germinating seed in the dead husk.  TABLE VIII (continued) AZTEC AGRICULTURAL AND RITUAL CALENDARS  Agricultural Month  Activities  No. 3  maize sown  4  3. COATLICUE. 4. "Serpent Skirt". 6. Offerings of flowers.  7. TLALOC.  1.  HUEL TOZOZTLI. 2. "long fast".  3. CENTEOTLI and CHICOMECOATL.  5. Maize deities.  6. Seed corn blessed  and worshipped. maize transplanted or weeded and fertilized  5  1. TOXCATL. 2. "Dug or slippery".  3. TEZCATLIPOCA and HUITZILIPOCHTLI.  4. & 5. See months 9 & 12.  2  6  1. EIZALQUALIZTLI.  2. "Bean porridge".  3. TLALOC.  4. & 5. See month 13.  6. Two children drowned by tying them to a boat loaded with sacrificial  J  hearts, and sinking the boat.  U N E  2. "Short fast".  4. "Corn God" and "Seven Snake".  L  Y  1. TOZOZIONTLI.  8. Child sacrifice.  P R I  A  2  5. Earth and fertility goddess.  A  M  Aztec Solar Months and Ceremonial Activities"  7  1. TECUHILHUITONTLI.  2. "Little feast of Lords".  4. "Salt Women". 5. Goddess of saltwaters.  3. HUIXTOCIHUATL.  TABLE VIII (continued) AZTEC AGRICULTURAL AND RITUAL CALENDARS  Agricultural Month  J U L Y  Activities  selective harvest of young ears  No. 8  Aztec Solar Months and Ceremonial Activities 1. HUEITECUHILHUITL.  2. "Great Feast of Lands".  loose, simulating corn tassels; after sacrifice, "elotes" eaten.  of corn 1. TLAX0CHIMAC0.  2. "Birth of flowers".  4. "Hummingbird of the Left Hand". A  maize  U G U  ripening in  S E P  4. "Young  Corn Ear". 6. Sacrificed slave girl impersonating deity; women's hair worn  9  S T  3. XILONEN.  3. HUITZILIPOCHTLI.  5. Tribal god of Mexica; god of  hunting and war. 10  fields  1. X0C0TLHUETZI. God".  amaranth harvested  2. "Fall of fruits".  5. Fire god.  3. HUEHUETEOTL. 4. "Venerable Old  6. Prisoners drugged, burned but not killed on fire,  dragged out with hooks and sacrificed.  2  11  1. OCHPANIZTLI.  2. "Month of brooms".  filth or sins".  5. Earth and corn goddess.  T  yellow as maize decapitated.  t M  warriors.  3. TLAZ0LTE0TL. 4. "Eater of 6. Girl with face painted as  8. Efforts to avoid rain; ritual review of  TABLE VIII (continued) AZTEC AGRICULTURAL AND RITUAL CALENDARS  Agricultural Month B E R 0 C T 0 B E R N 0 V E M B E R D E C  Activities mai ze harvested  No. 12  Aztec Solar Months and Ceremonial Activities 1. TEOTLECO. 2. "Return of Gods". 3. TEZCATLIPUCA. 5. Texcocan god of war and hunting.  chiles planted  4. "Smoking Mirror".  8. Return of gods celebrated with  furnace sacrifice; ritual drunkenness.  tomatoes planted  13  1. TEPEILHUITL.  2. "feast of the mountains".  3. TLALOC.  4. "He who  3  makes things grow".  5. God of rain, hail, thunder, floods, conceived  to live on top of mountains. vegetables  Amaranth used.  6. Four women and one man sacrificed.  Ritual cannabalism.  and flowers planted  14  2. "bird month".  5. God of war and hunting. 15  start of harvest of  1. QUECHOLLI.  1. PANQUETZALIZTLI. above.  3. MIXCOATL. 4. "Cloud Serpent".  6. Ceremonial hunt; making of weapons.  2. "feast of flags".  3. HUITZILIPOCHTLI.  4. & 5. as  6. "Flowery Wars", ritual combats with neighbouring states in  which prisoners were taken for sacrifice.  TABLE VIII (continued) AZTEC AGRICULTURAL AND RITUAL CALENDARS  Agricultural Month E M B E R Note:  Activities vegetables and flowers chiles transplanted  No. 16  Aztec Solar Months and 1. ATEMOZTLI.  2. "Fall of waters"  Name of the solar month Meaning of the name of the month. Name of principal deity associated with month. Meaning of name of deity. Function of deity. Character of ceremonies. Other deities associated with month. Character of ceremonies associated with other deities.  Sources: 1 2 3  3. TLALOC. 4. & 5. as above.  The Aztec solar calendar comprised 18 months of 20 days each, together with five "useless days" which were required to correct the calendar to the solar year.  Explanation of numbers: 1. 2. 3. 4. 5. 6. 7. 8.  7  Ceremonial Activities  Vaillant 1965:187-190, 200-202. Peterson 1959:125-136, 167. Gibson 1964:121.  144  the "elotes" were harvested.  During August the rains decline and at  the end of August amaranth, which matures slightly faster than maize, was harvested, providing some insurance against a late failure of the corn crop.  During the latter part of August and the early part of  September, ritual activity was directed at achieving hot, dry weather so that the corn would mature and dry out; towards the end of September it was harvested, as it is now.  After the harvest the feast of "the  return of the gods" was celebrated, with ritual drunkenness that was perhaps an expression of relief that the harvest was safely secured. As soon as the maize is harvested, the winter cycle begins with the planting of chilies and other vegetables and flowers. continue through the f a l l .  These activities  The harvest of these various winter crops  begins in January and continues through the early spring. A great deal of work is involved in these horticultural activities and the crops are important, but not as important as maize.  During the winter, Aztec  ritual attention shifted from agriculture to hunting and war. The modern chinampas are cultivated quite extensively and there is a tendency to avoid the use of seedbeds for crops other than flowers, and to sow directly into the fields once the preceding crop has been harvested.  In the past this was probably not the case, and the succes-  sion of crops overlapped with a second crop sown in the seedbeds before the first was harvested from the fields.  By exploiting seedbeds in this  way the agricultural calendar may have been far more intricate than is indicated by this generalized summary, and with irrigation, fertilization and careful timing it would have been quite possible to maintain the entire area of chinampas in continuous production.  145  Yields and Energy Ratios Table IX gives the yields of shelled maize from seven plots in modern Mixquic and shows values ranging from 1.5 tons per hectare to 5.1 tons per hectare, with a mean value of 3.9 tons per hectare. The Xochimilco chinampas are extremely polluted and large areas have been abandoned, but the C.H.C.V.M. estimates that the yields from the areas that are cultivated are about 3.5 tons per hectare.  How close these  modern figures may be to pre-Hispanic yields is uncertain.  The higher  modern values reflect some use of chemical fertiliziers, which would suggest' that Aztec production may not have reached these levels. However, the chinampas are now in very poor physical condition and they may have been more productive in the past than they are now.  The modern  chinampas are not well looked after, and the low yields from some of the Mixquic plots are entirely attributable to poor management.  During  the 16th Century the work of agricultural labourers was carefully supervised by state appointed officials, and it is probable that the whole area of the chinampas was very carefully managed.  Both of these factors  could have resulted in 16th Century yields that were at least as high as modern average yields, and quite possibly very much higher than that.  If  one assumes that the lack of chemical fertilizer was offset by the better conditions and management, then a yield between three and four tons per hectare of maize would seem to be a reasonable estimate of the productivity of the Aztec chinampas, and most modern writers support estimates in this range (Sanders and Price 1968:148; Cainek 1972:111; Parsons 1976:242-45; Vanegas 1978:6).  The estimate is of the yields of  the cultivable areas of the chinampas only.  Given that about one quarter  146  TABLE IX YIELDS OF SHELLED CORN, MIXQUIC  Source:  Plot No. 1  4612 Kg/Ha  Plot No. 2  1519 Kg/Ha  Plot No. 3  3923 Kg/Ha  Plot No. 4  4585 Kg/Ha  Plot No. 5  3037 Kg/Ha  Plot No. 6  5098 Kg/Ha  Plot No. 7  4543 Kg/Ha  Vanegas 1978:22-23.  of the total area of the chinampas.is occupied by uncultivable ditches and canals, this gives a yield of between 2250 and 3000 kilograms per hectare for the whole area of the chinampas. The chinampas yielded a number of crops"in addition to maize. Table X gives yields of the saleable portions of various vegetables cultivated in modern Mixquic.  The figures are for year-round production  of vegetables and are not representative of the 16th Century chinampas, but they suggest that the yield of horticultural products could have been quite high.  If, as an approximation of 16th Century yields, one  takes the average of the yield shown in Table X and assumes that vegetables were cultivated only during the six winter months, that the yield of indigenous varieties grown during the off season was only half as great as the yields of modern varieties grown throughout the year, and that in any one year only half the cultivable land was sown to winter vegetables, then the chinampas yielded about 5 tons per hectare  147  TABLE X YIELDS OF SALEABLE PORTIONS OF SELECTED VEGETABLES, MIXQUIC Celery Acelga Spinach  5.4 kg/m p 2  6.6 kg/m 2.2 kg/m  2  Coriander  1.7 kg/m  2  Source:  Vanegas 1978:20.  of vegetables from cultivable land, or about 3.75 tons per hectare overall, in addition to the yield of 2.25 to 3 tons per hectare of maize.  To put  these figures in perspective, the yield of maize from hand-tilled extensive systems in other parts of Mexico and Guatemala is generally in the vicinity of one ton per hectare, or less than 0.25 tons per hectare i f fallow land is included in the calculations.  Yields of commercial corn  production in the United States are rather more than 5 tons per hectare. An alternative to expressing yields as output per unit area is to express them as output per worker.  Sanders has estimated that in the  Aztec economy, a single "chinampero" could have worked between 0.5 and 0.75 hectares of chinampa land (Cited in Parsons 1976:242).  If the lower  figure is combined with the estimate of yields per hectare, then a single "chinamperos" could in the course of a year have produced between 1.125 and 1.5 tons of maize, together with about 1.875 tons of vegetables. The estimate for maize seems reasonable, and it conforms with data from North West Guatemala and the Yucatan that indicates that farmers in extensive traditional agricultural systems produce about 1.2 or 1.3 tons  148  of maize in the course of a growing season (Redfield and Villa 1934:52, 55; Stadelman 1940:103, 147, 149; Steggerda 1941:118; Villa 1945:60, 61). Sanders' higher figure leads to estimates of 1.7 to 2.2 tons of maize plus 2.8 tons of vegetables as the annual output of a single worker, but although this may have been possible in the productive chinampa land, the figure seems too high. If energy values are assigned both to the output and the input of an agricultural system, then the ratio of the output to input gives a useful measure of the energy efficiency of the system.  In the present  calculations, inputs and outputs are expressed as the energy equivalent of a kilogram of maize.  The outputs of the chinampas included 2250 to  3000 kilograms per hectare of maize, together with some 3750 kilograms per hectare of assorted vegetables.  Expressed as a percentage of the  calorific content of maize, vegetables range in value from about 16% in the case of leafy greens to about 120% in the case of beans.  Bearing  in mind that beans were an important secondary crop, it seems reasonable to employ a figure of 40% as an expression of the amount of vegetables that could be substituted or exchanged for a given volume of maize.  Using this figure, the total output of each hectare of chinampa  land amounted to between 3750 and 4500 kilograms of maize or maize equivalents. Agroecosystems receive inputs of energy in the form of solar radiation and of cultural energy.  The latter term includes all inputs  of energy in the construction of capital installations and equipment, in the production and transport of'materials that are required in the operation of the agricultural system, and in the modification of the  149 physical structure of the system by tilling, cultivating, harvesting or any other agricultural activity.  In comparative studies of the efficiency  of agricultural systems solar energy is generally ignored and the calculations are restricted to the inputs of cultural energies.  It is clear  that a great deal of energy was invested in constructing the chinampas and their associated hydraulic installations but, because this investment was amortized over many generations, the proportion of the initial investment that should be debited to the energy budget of any particular year of operation is so small as to be negligible.  The tools employed  in chinampa agriculture were generally quite simple and would not, for the most part, have required more than a few hours' labour to construct. Even if such items as "zoquimaitls" or "texpetlatls" had to be repaired or replaced several times in the course of a year, the investment of energy they represent is slight.  The only major piece of capital equip-  ment that the "chinamperos" required was a boat or a share in a boat.  If  modern practices are representative of Aztec times, the boats were constructed by specialists and were expensive to acquire, but with careful maintenance their purchase price could be amortized over several years. A good boat would last for 15 or 20 years before it had to be replaced. Bat-guano, leaf mould and sewage sludge were imported to the chinampas, but we have no indication of the quantities involved or of the amount of labour employed in assembling and transporting these materials.  Apart  from these costs, the only major input of energy to the chinampas was in the form of ongoing agricultural labour.  Using the diet of modern  Mexican "campesinos" as a guide, a single agriculturalist required something of the order ot 160 kilograms of maize for a year's subsistence, together with other foodstuffs, bringing the total dietary input of each  150  worker to about 200 kilograms of maize or maize equivalent each year (Parsons 1976:243).  If one assumes two workers per hectare and allows  a further 100 kilograms of maize or maize equivalent to cover the costs of boat construction and material transport, then the energy equivalent of 500 kilograms per hectare can be taken as an estimate of the input of cultural energy to chinampa agriculture which, with an output of 3750 to 4000 kilograms, gives an energy ratio of between 7.5 and 8.0.  If one  assumes that a single worker could have managed 0.75 hectares of chinampa land and allows 100 kilograms per hectare for additional costs, then the input falls to 367 kilograms per hectare, and the efficiency increases to between 10.2 and 12.3. It is clear that even slight modifications to the assumptions would make quite large differences to these estimates of yields and energy efficiency.  However, the estimates do appear to be quite consis-  tent with data on other systems of traditional agriculture and they are generally confirmed by other published accounts of the yields of chinampa agriculture. The Principles of Chinampa Agriculture The chinampas can be thought of as comprising three separate but complementary ecosystems, each with its distinctive characteristics, and each contributing in a different fashion to the function of the system as a whole.  One of these ecosystems consisted of the canals, which  occupied about 13% of the area of the chinampas, and which supported the growth of many of the species of aquatic plants that had flourished in the lake complex prior to its colonization for agriculture. Aquatic plants are generally high yielding, and the fact that the canals of the  151  extant chinampas must be frequently cleaved of vegetation if they are to remain navigable suggests that the productivity of this ecosystem may have been quite high. The platform borders and the community of plants that they supported comprised a second ecosystem, which occupied about 11% of the area of the chinampas.  This ecosystem included Ahuejote and other  trees, which, with a density of about one tree per one and a half square meters, amounted to a considerable standing biomass.  It included species  of different dimensions, life-spans and growth habits, and it appears to have had a high and persistent pattern diversity, at least in comparison to other ecological systems within the chinampas.  Taken together,  these characteristics suggest that the platform borders were quite generalized and had some of the characteristics of stable and mature ecosystems . The remaining area of the chinampas was occupied by crop plant communities of the platform surfaces.  A large number of different  species were grown and polyculture was practiced.  Different species  may have been grown together from seeds planted in the same holes, as is the case in some extant Middle American agroecosystems.  More  probably, different species were interspersed in the same rows, in adjacent rows, or in adjacent plots, as is the practice in the chinampas today.  The crop communities may have been least diverse during the  summer when maize was the predominant crop but, even then, uninterrupted pure stands were limited to the area of a single chinampa platform, and it is possible that within the area of a single platform, stands of maize were interspersed with stands of amaranth, flowers or other crops.  During the winter, when a greater number of species were  152  cultivated, the plots would have been smaller, and the diversity of the community correspondingly greater.  The practice of polyculture contrib-  uted a degree of complexity and structural diversity to the crop ecosystems of the chinampas, which would have enhanced their stability and resilience.  However, the crop species were short lived, vulnerable to  the depredations of pests and the competition of uncultivated plants, and, like crop ecosystems generally, those of the chinampas tended to be immature and unstable. The three ecological systems were closely interrelated. The canals and the platform borders provided a number of biotic resources that were of direct economic importance, but their major role in the chinampas appears to have been to provide support and protection for the crop ecosystems.  The canals acted as reservoirs for the materials that  were cycled within the chinampas.  Water was stored in them and nutrients  were stored in the water, in the bottom mud, and in the tissue of aquatic plants.  These materials were available throughout the year.  When they  were required they were withdrawn from the canals and applied to the platform surfaces.  The canal ecosystems not only stored water and nutri-  ents, but they also transformed them into materials that were valuable for their structural properties.  Reeds and similar plants were used for  the construction of floating seedbeds and shelters, and they were woven into a variety of useful artifacts.  Many of the plants that were used  as green fertilizers were valuable as well for the pockets of air trapped in their leaves, which maintained the tilth of the chinampa soils, and it was the peculiar spongy character of the bottom mud that made it indispensable to the construction of seedbeds.  The canals and platform  borders acted to condition the microclimate of the chinampas.  The supply  153  of water in the canals may have been effective in maintaining a high level of moisture in the vicinity of crops. The Ahnejote trees protected the platform surfaces from wind and sun, and in so doing, smoothed fluctuations in temperatures, evaporation rates and other climatic variables.  Finally, the platform borders may have played an important  part in regulating populations of pests and pathogens.  The assertion  is not entirely substantiated, but it is probable that generalized ecosystems in the vicinity of crop ecosystems provide a refuge for birds and for predatory and parasitic insects and microorganisms, with the result that these are available to suppress populations of pests and pathogens when the crops are in the field.  It was suggested  that the canal border ecosystems were quite generalized, at least in comparison to the crop ecosystems, and they may have performed this function within the chinampas. The relationship between the crop ecosystem, on the one hand, and the canal and platform border ecosystems, on the other, was intimate and intricate.  The crop ecosystems exploited the non-crop  ecosystems for their stock of water, nutrients and structure, and for protection from climatic variables, but they did so in a manner that allowed the latter to continue functioning unimpaired. The demand for microclimate control and pest regulation was mitigated to the extent that the heterogeneity of the crop ecosystems themselves permitted the performance of these functions.  The exploitation of the canals for  materials was frugal, and most of the water and nutrients that were removed were returned by the passive processes of decay and runoff, where they again became available for future use.  Any exploitation  that represented a degradation of the structure of the canal ecosystems  154  was offset by the high productivity of the aquatic plant communities. The relationship between the crop and non-crop ecosystems could be described as a commensurate one in which the crop ecosystems benefitted from the presence of the other two ecosystems, but the latter neither benefitted nor suffered from the association. The chinampas are here viewed as comprising three separate ecosystems, and the hypothesis is put forward that the primary role of the two non-crop ecosystems was to provide support and protection for the fragile crop communities of the platform surfaces.  Much work  remains to be done before the hypothesis is substantiated and, in particular, the role of the narrow strip of trees and herbs along the borders of the canals in suppressing the growth of pest populations requires investigation.  However, the hypothesis does account for the  facts that are available, and it is a useful one because it provides a number of insights into the origins and functions of chinampa agriculture. First, it sets chinampa agriculture securely in its developmental context.  One of the earliest and most widespread forms of  agriculture in Middle America is Rosa cultivation, which essentially consists of inserting a small crop ecosystem into a generalized natural ecosystem.  The crop ecosystem is relatively unstable, but it is  supported by nutrients from the natural ecosystem, and to some extent it is protected from, climatic flux and the explosive growth of pest populations by the natural ecosystem.  The association is managed in  such a way that the crop ecosystem benefits and the function of the natural ecosystem is not, in the long term at least, impaired. The chinampas differed from Rosa cultivation in all particulars, not least  155  in the fact that the generalized, non-crop ecosystems were as artificial as the crop ecosystems themselves, but the principle that underlies the two forms is the same.  In both, a fragile crop ecosystem is supported  and protected by the existence of a more generalized ecosystem in its close vicinity.  The chinampas were an elaboration of a pre-existing  form, and the multiple ecosystem construct emphasizes this. The hypothesis also allows a fairly succinct expression of the means by which the chinampas produced high yields without incurring a commensurate cost in cultural energy inputs.  The high yields of the  chinampas, or of any other intensive agricultural system, were a function of the immaturity of the crop ecosystem.  The costs of maintaining  an ecosystem in an immature state are exacted in the measures that are required to provide a suitable growing environment for the fragile crop plants, and to protect them from diseases, pests and competing plant species.  To some extent, both these functions were performed  in the chinampas by increasing the intensity of human labour, and this was reflected in an energy ratio  that was lower than that of more  extensive traditional agroecosystems.  However, to a large extent,  these functions were performed in the chinampas by the canal ecosystems that provided the crop ecosystems with readily assimilable nutrients and by the platform border ecosystems that protected the crops from climatic fluctuations and the growth of pest populations.  The cost of  these functions was not eliminated but, to the extent that it was performed by the non crop ecosystems, it was exacted in the form of the solar energy required to maintain these two ecosystems, rather than in the form of human labour applied directly to the crop ecosystems.  156  Finally, the multiple ecosystem construct emphasizes the fundamental distinction between the principles of chinampa agriculture and of modern western agriculture.  Both support crop ecosystems that are  fragile, specialized and unstable, and both include means for provisioning and protecting the crop ecosystems.  However, the chemical and  mechanical means that are employed by Western agriculturalists exact a very high cultural energy cost, with the result that the energy efficiency of Western agricultural systems tends to be very low.  In  the chinampas the same functions were performed by secondary ecological systems, which although.they may not have been quite as effective as the techniques of Western agriculture, exacted a very low cultural energy cost with a result that the overall efficiency of chinampa agriculture was considerably higher than the efficiency of commercial Western agriculture, while the yields were at least as high, and possibly very much higher.  157  CHAPTER V  THE MANAGEMENT OF CHINAMPA AGRICULTURE  Management is here understood as the process by which decisions are made that have the effect of modifying the structure, and thence the function, of managed physical systems.  The term is defined broadly  and in the present discussion the individual "chinampero" who makes decisions with regard to a particular plant and the high order administrative institution that makes decisions with regard to the whole of the lake complex are both understood as being managers.  It was sug-  gested that the development of the hydraulic works necessary for the control of water levels was an essential element of the development of chinampa agriculture and, this being the case, the most important function of management in the developmental phases of chinampa agriculture was the organization of the labour and materials necessary for the construction of hydraulic installations.  The management of these  installations in the Valley of Mexico, as elsewhere, was intimately bound up with the government of the state.  To a very large extent the  development of institutions for the management of hydraulic resources was identified with the development of the state itself.  Because  of this a brief account of the social organization and political history of the Aztecs is a necessary preliminary to the discussion of the management of the chinampas. It is convenient to distinguish three phases in the development of the Aztec state.  In the first phase, during the unsettled period  following the collapse of Toltec civilization in the late 12th Century,  158  the Valley was colonized by refugees from Tula and by immigrants from beyond the borders of the Basin, and the political entity that is best described as the city-state emerged.  In the second phase, city-states  were aggregated into progressively larger tribute states, culminating in the formation of the Aztec state in 1433.  In the final phase, s t i l l  underway at the time of the Spanish Conquest, the basic institutions of the large tribute state were modified and reworked, and the institutions of the centralized and bureaucratic state emerged.  Each of  these three phases exhibited distinctive characteristics in the relationship between the state and its hydraulic resources, and the character of the management institutions that finally emerged was determined by the processes that occurred during this period of evolution. The City-State Following the collapse of Teotihuacan in the late 9th Century, and of Tula in the late 12th Century, the surviving inhabitants of these great centres established themselves at settlements in other parts of the Valley.  Among the more important settlements established  at this time were Atzcapotzalco, which was founded by refugees from Teotihuacan, and Xico, Coatepec and Culhuacan, which were founded by Toltec groups (Figure 31).  During the 13th and 14th Centuries this  population was augmented by the arrival of Otomi and Nahua tribes from regions to the north of the Valley.  The new arrivals were known  collectively to the more civilized inhabitants of the Valley as "chichimeca" or "dog-people" which is a term that carries all the significance of "barbarian". The Otomi tribes arrived first and settled towards the north of the Valley in such centres as Tenayuca, Zumpango  159  Figure 31.  Principal Settlements of the Valley of Mexico in the 16th Century (Sanders, Parsons and Santley 1979 Map 20; Palerm 1973).  160  and Xaltocan, and one group allied itself to the settlement of Coatepec which became the nucleus of an important state whose centre subsequently shifted to Texcoco.  The Nahua tribes arrived at a slightly later date  and established themselves at such centres as Xochimilco, Cuitlahuac, Mixquic and Chalco, towards the south of the Valley.  The Mexico were  one of the last of the Nahua tribes to arrive, and they did not establish their settlement at Tenochtitlan until 1345 (Wolf 1959; Peterson 1962; Gibson 1964; Vaillant 1966; Davies 1977; Leon-Portilla (ed.) 1977). The Chichimec tribes were attracted by the civilizations of the Valley and soon acquired many of the traits of civilization, but they also brought with them cultural traits and institutions of their own. One of these was the "calpulli", and an understanding of this term is essential to an understanding of the subsequent political and social developments.  "Calpulli" or in the plural, "calpultin", translates  literally as "big house" or "big family", and it was originally an endogamous lineage group, distinguished from other such groups by a patron deity, a name and battle insignia, and by a common if putative ancestry.  The institution may have originated in some migratory  prehistory of the Chichimec tribes, but in the settled context of the Valley it underwent considerable modification and acquired territorial connotations.  By .the time of the Conquest the term referred not  only to a lineage group, but also to the tract of land claimed by that group and to the village or neighbourhood of a town occupied by the group.  These various associations of the term caused con-  siderable confusion in early Colonial accounts, much of which has persisted in the modern literature (Wolf 1959:135; Carrasco 1971; Munzon 1977).  The "calpulli" was administered by a council called the "huehuetque",  161  which was headed by a senior member called the "caipullec". Members of the council may originally have been elected from the population at large, but in the late period of Aztec history they were drawn only from the upper classes.  The "calpullec" was elected by other members of the  council and held his office for life.  The "huehuetque" was responsible  for the organization of education, civil order, public works projects and religious and military activities, and it was provided with officials for the execution of these various undertakings.  The "calpulli" was  also the institution in which was vested the ownership of agricultural land, and the "huehuetque" administered the use of this land. The larger "calpultin" were divided into segments, called "tlaxilacalli", which consisted of groups of closely related families, and which in conformity with the territorial associations of the "calpulli" also carried the significance of a street or a block within a neighbourhood of a town.  The smaller "calpultin" and the "tlaxilacalli"  of the larger "calpultin" were made up of individual families. Membership of a "calpulli" gave an individual certain rights, including the right to the use of a plot of "calpulli" land.  It also carried with  it certain obligations including the obligation of cultivating the plot of land and of providing goods or services for the support of public works projects, religious or military activities and any other undertakings that were endorsed by the "calpulli". The "calpullecs" served on a higher order council called the "tlatocan", which translates literally as "place of discourse", and this council was headed by a "tlatoani" or "speaker".  The "tlatocan"  may have originated as an institution roughly corresponding to a tribe, but as the Nahua tribes settled in the Valley of Mexico it came to be  162  better described as the administrative body of a small city-state. A city-state comprised agroup of "calpultin", either genealogically related or in a close political alliance, together with the villages and towns that the "calpultin" occupied and the surrounding territories to which they laid claim.  The smaller city-states consisted of a town  and a few outlying settlements, and they were administered by a single "tlatoani".  The larger city-states were made up of several towns, each  with its territories and outlying settlements and each with its own "tlatoani".  In this case the "tlatoani" of the largest and most influ-  ential town was in effect the ruler of the state. The Confederation and the Tribute State The city-states were quite stable institutions, and they became the constituents of larger political aggregations.  Two rather different  types of supra-city-state polities developed in the Valley of Mexico, the confederation and the tribute state.  Elements of the political  confederation were evident in a number of the larger states that developed in the Valley, including the "Triple Alliance" that was the core of the Aztec empire, but the form appears to have achieved its fullest expression in the Chalco-Amecameca Confederation.  The Chalco  basin was.occupied by many different tribal groups, including two Toltec groups settled at Xico and Tlapacoya, and several Nahua groups, including the Mixquica, the Chalca and the Amecameca. Sometime during the 13th Century these groups formed a confederation which at one time included twenty-six separate city-states, each with its own "tlatoani" (Davies 1977:46).  As far as one can tell from the limited historical  sources, this was a true confederation in which all the participants  163  acted together in affairs of common concern, but in which each citystate retained authority over its internal affairs.  The Confederation  that developed in the Chalco basin was powerful and able to withstand attacks by other large states in the Valley; it was stable and persisted as an independent entity until 1465, at which time it s t i l l had thirteen separate "tlatoani" (Davies 1977:46). A more common device for the formation of'polities at higher levels than the city-state was the tribute state.  The Nahua and Otomi  tribes of the Valley of Mexico were nothing if not bellicose, and there was constant warfare in the Valley.  Wars were quite formalized affairs,  and they were considered to be completed when one side succeeded in sacking the chief temple of the city of the other.  The glyph that  means "victory" is a picture of a burning temple (Peterson 1962:160). When the battle was completed the two sides sat down to negotiate the value of the tribute due to the victors.  The payments were made both  in the form of goods, including the specialized products of the conquered city.as well as common subsistence staples, and in the form of services, including both labour on public works projects and military service as the involuntary ally of the conquering state.  Wars were endemic and, as  the outcomes of successive campaigns, alliances and rebellions accumulated, all of the cities of the Valley became enmeshed in an extremely complex network of tribute payments.  As the result of past victories and  defeats, a single town might be the recipient of tribute from several other towns, and be at the same time the donor to as many more (Gibson 1964). The tribute states tended to be unstable and their borders shifted as the result of rebellions by imperfectly subjugates tributaries  164  or the attacks of powerful neighbours.  For example, the city-state of  Cuitlahuac, which was unfortunate enough to be located midway between the territories of three powerful neighbours, was at different times a part of the territory of Culhuacan, Xochimilco and the Chalco-Amecameca Confederation.  However, little by l i t t l e , the more powerful states  consolidated their hold on their territories and by the mid-14th Century the earlier profusion of small city-states had been reduced to the relative simplicity of.five large tribute states and the Chalco-Amecameca Confederation, with only a few small city-states maintaining a semblance of independence in the shadows of their more powerful neighbours. The Chalco-Amecameca Confederation controlled most of the Chalco drainage basin.  The Xochimilea state had at one time included a large part of the  Chalco basin and considerable areas in Morelos, outside the Valley, but by the mid-14th Century it had contracted to include only areas within the Valley, to the south and southwest of Lake Xochimilco. The Culhuaque controlled the Ixtapalapa peninsula and, although its territory was quite small, the state was pretigious because of its associations with the ancient city of Tula.  Atzcapotzalco was the centre of the Tepanec  state that extended over most of the western parts of the Valley and that included the recently established Mexica city of Tenochtitlan as one of its tributaries.  Xaltocan appears to have been the dominant  centre in the north of the Valley, but not a great deal is known about this region.  The Acolhua domain extended over the eastern parts of  the Valley.  It was administered from the city of Texcoco, which was  unrivalled as a centre of the arts and the sciences of highland Mesoamerican civilization.  165  It was inevitable that one of these states should attempt to overpower the other five and gain control of the entire Valley. The Tepaneca came very close to doing this.  In 1353 Tezozomoc became the  "tlatoani" of Atzcapotzalco and embarked on an unprecedented career of empire building.  In 1357 Culhuacan was defeated by the Tepaneca  and their allies, who included the Mexica. A few years later Cuitlahuac and Mixquic were conquered by the Mexica, acting on behalf of the Tepaneca as well as in their own self interest.  Campaigns were initi-  ated against the Xochimilca in the 1360's, largely by the Mexica, but with Tepanec support, but the Xochimilca were not finally defeated until 1391.  In 1365 the Mexica embarked on their protracted war with  the Chalco-Amecameca Confederation, which lasted for about ninety years. In 1395 the Tepaneca successfully attacked Xaltocan, and by the end of the 14th Century all of the Valley except for Texcoco and the ChalcoAmecameca Confederation was a part of the Tepanec empire.  Tezozomoc  initiated campaigns against Texcoco in the early part of the 15th Century which resulted in the defeat of Texcoco in 1416 or 1418.  Itzcoatl,  the "tlatoani" of Texcoco was killed and his son, Nezahualcoyotl, was forced into exile.  The Tepaneca installed a puppet ruler in Texcoco,  and only the Chalco-Amecameca Confederation resisted assimilation into the Tepanec empire.  Tezozomoc came very close to uniting the whole of  the Valley, but he did not quite succeed.  In 1426, at the reported  age of 106, he died and the Tepanec empire began to fragment. 1428 Nezahualcoyotl regained control of Texcoco.  In  Shortly after, the  Tepanecs of Tlacopan and the Mexico of Tenochtitlan rebelled against Atzcapotzalco and joined the Texcocans.  In 1433 this "Triple Alliance"  166  defeated Atzcapotzalco, and the Aztec empire was established. The "Triple Alliance" acquired all the tributaries of Tepanec empire and effectively controlled the entire Valley, excepting only the ChalcoAmecameca Confederation, which resisted incorporation into the Aztec state until 1468. The Lake Aztec State The concentration of political power continued after the victory of the "Triple Alliance" in 1433.  Tlacopan was at the outset the least  powerful of the allies and it was quickly relegated to a subordinate status,.leaving Texcoco and Tenochtitlan in control of the empire. The Mexica had transformed themselves from a minor tribe to a major power in the Valley in less than a hundred years, and they were an efficient and assertive people.  Nezahualcoyotl was an able and res-  pected ruler and, while he remained alive, the Texcocans were able to resist the assertive tendencies of the Mexica and to maintain a balance in the power of the two states.  However, the power of Texcoco began to  decline after the death of Nezahualcoyotl.  By 1500 the decline had pro-  ceeded to such an extent that the Mexica were able to impose a puppet ruler on Texcoco and Tenochtitlan became the single most powerful city, not only in the Valley but in the whole of Middle America. The process by which political power was concentrated, and progressively larger polities were formed, was accompanied by a modification of social and political institutions.  The "calpulli" had been  an efficient device for marshalling the resources of the city-state and the resources of several city states could be organized by means of the tribute system, but finally both institutions were inadequate for  167  the administration of a large empire.  They were cumbersome and ineffi-  cient when operating over a large population or through many hierarchic levels and, by maintaining local rulers in place, they tended to foster the identity of the constituent parts of the empire rather than the identity of the empire itself.  These institutions were s t i l l in opera-  tion at the time.of the Conquest which is why we have a fairly detailed knowledge of how they functioned, but they were in the process of being replaced by a very much more centralized administrative apparatus and by a rigidly stratified social system. Many of the institutions of the late Aztec state had been transmitted to the Nahua and Otomi tribes by the survivors of Teotihuacom and Tula, and subsequently modified and integrated with more purely Nahua institutions.  Like the earlier societies of the Valley, the Aztec  state was divided into a number of ranked social classes. The major distinction was between the ruling class and those they ruled.  The former were "tecuhtzin" (sing, "tecuhtli").  The class  included numerous ranked groups or individuals ranging from the "hueytlatoaui" or "great tlatoani" of Tenochtitlan, down to the lesser nobility, or "pipi 1tin".  The status of "tecuhtli" carried with it  certain rights, including exception from tribute payments, and certain duties, including submission to a set of laws that were somewhat more harsh than those that governed the bulk of the population (Peterson 1962:123). At the time of the Conquest the lower class was divided into a number of ranked groups.  The "macehualtin" (sing, "macehual1i") are  probably,best understood as a class of free commoners, and included skilled craftsmen and agriculturalists who had rights under the "calpulli"  168  system.  "Tlalmaitl" were also agriculturalists, but they had fewer  rights and were obliged to live and work on the large tracts of land that were set aside for the support of high ranking individuals or institutions. class.  "Tlacotin" (sing. "Tlacotli") were of the lowest social  The term is often translated as "slave", but although individ-  uals in this condition were obliged to work for their masters and could be bought and sold, they had a number of civil rights and were not considered as property or chattels. The social distinctions were marked by sumptuary regulations that, for instance, prohibited the wearing of lip-plugs by any but "tecuhtzin", and of jade or golden lip-plugs by any but high ranking "tecuhtzin".  The distinctions were perpetuated by an educational system  in which the children of lords and those of commoners were educated in different types of schools.  At the time of the Conquest, a certain  amount of social mobility persisted and, by exceptional military accomplishments, it was possible for an individual of low status to become a member of the ruling class, but the indications are that status was to an increasing extent becoming hereditary.  At the same  time the distinctions between "macehualtin" and "Halmaitl" were becoming blurred, and a single undifferentiated lower class was beginning to develop. These horizontal social strata were crosscut by three major vertical divisions, each of which was represented at all social levels and all of which were united at the apex by the overlapping offices of the "tlatoani" of Tenochtitlan. There was a well developed military hierarchy.  All males owed military service and the members of each  "calpulli" fought as a unit, with military specialists on the "calpulli"  169  council functioning as officers.  With a social status bordering on  that of the nobility were the specialized military orders of "eagle" and "jaguar" warriors.  High level officers were drawn from the ranks  of the nobility, and two senior commanders, the "tlacatecatl" and "tlacochcalatl", the commander of warriors" and the "man of the javelin house", were two of the four senior advisors to the "tlatoani" of Tenochtitlan.. The latter carried the title of "tlacatecuhtli" or "chief of warriors", and was in one of his functions the commander in chief of the army. There was an equally well developed religious hierarchy made up"of various types of "tlamacazqui" or full-time religious functionaries.  At the lower levels were junior priests and priestesses,  including the "quacu.illi" who were.associated with the "calpultin", and the "tlamacaztequiuaque" who were combination priests and warriors. High priests of the many deities had the status of nobles and were 1  known collectively as "tlenamacac", and the high priests of Huitzilipochtli and Tlaloc, two of the most important Aztec deities, functioned with thetwo high ranking military officials as the chief advisors to the "tlatoani" of Tenochtitlan.  The latter included among his many  functions that of the high priest of Quetzalcoatl. The third major hierarchy was administrative.  Ranking immedi-  ately below the "tlatoani" of Tenochtitlan were the two military and two religious functionaries, one of whom carried the additional title of "ciuacoatl" or "woman serpent", and who functioned as the vice-ruler of the state.  Ranking below these was a council of thirteen high  making dignitaries called the "tecuhtlatoque".  This council included  high ranking military and religious officials, nobles with special responsibilities for the administration of justice, and the "tlatoani"  170  of the more prominent city-states.  The administration was served by a  large number of officials of lower rank, including scribes or "tlacuilos" and "calpixque" or "house attendants".  These latter functioned in many  capacities, but most particularly as tax collectors and tribute gatherers.  "Centecpanpixque" had the responsibility of collecting dues from  twenty families, and each group of five of these officials were under the authority of a "macuiltecpanpixque". served by an elaborate court system. of relatively slight importance. messengers'or policemen  The administration was also  Low level judges dealt with matters  They were assisted by scribes, by  called "topolli",;and by cryers or "tecpoyotl".  These judges were subject to strict rules of conduct and were responsible to higher order courts which, in turn, were responsible to a supreme court, the "tlacxitlan", which was presided over by the "hueytlatoani" himself.  The authority of the central government was extended to the  constituent city-states of the empire through the agency of local "tlatoani".  These offices were hereditary and generally filled by  local nobility, but it came to-be the case that the local "tlatoani" had to be confirmed in their office by the central government, and they came to function rather as state appointed provincial governors rather than as independent local rulers. The transition to a centralized and bureaucratic government and to a stratified society was accompanied by a change in the institutions of land tenure.  By military means, land was withdrawn from the "cal-  pulli" system and passed into various categories of state ownership with designated tracts allocated to.the support of specified individuals or institutions.  Numerous classes of land were distinguished, of which  the most important were lands for the support of religious activities  171  with particular, tracts set aside for the support of the priests of particular gods; I'milchimalli" or "lands of the shield" and "yaotlalli" or "war fields" which were for the support of military ventures; "tecpantlalli" were government lands set aside for the support of judges, scribes, and other officials as well as for singers and entertainers; and "tecuhtlalli" and "pillalli" were for the support of individual nobles.  Particularly large areas were for the satisfaction  of the needs of the "hueytlatoani".  These lands were ascribed to  individuals, but they are probably better understood as being for the support of the offices held by those individuals (Moreno 1977:418ff). At the same time that land passed from "calpulli" to state ownership, the status of the agriculturalists passed from that of "macehualtin" to "tlalmaitl". The former paid a fixed share of whatever tribute their "calpulli" owed, while the latter were obliged to give up all their surplus production to the "calpixque", and were allowed to retain only that portion that was considered necessary for their subsistence. The city-states had their origins in tribal groups and they were quite stable and enduring polities.  They were aggregated into  larger polities by means of the tribute system, but the tribute empires were unstable and tended to fragment into their constituent city-states. The institutions of the late Aztec empire acted to reduce these divisive tendencies.  Local rulers came to function as regional governors; land  was withdrawn from the control of the city-states and incorporated in large estates that were under the direct control of the central government, and a rigid class system was imposed that undermined local, ethnic allegiances and defined the role of the individual in terms of his  172  place in the state as a whole.  The transformation was not complete at  the time of the Conquest and many of the institutions of the "calpulli" and the city-states remained in place, but it is clear that the older institutions were being replaced by the apparatus of the centralized bureaucratic state and that all the resources of the Valley were being brought under the direct control of the government of Tenochtitlan. The Construction of Hydraulic Installations It was suggested that the first phase in the development of chinampa agriculture was the excavation of ditches and, possibly, the construction of platforms around the periphery of the lake complex. The work could have been carried out in increments over long periods by small family groups, or it could have been carried out quite rapidly by working parties organized under the "calpulli" system.  Technically,  the devices are amenable to construction in either fashion, but in the context of the social organization of the Nahua settlers it seems likely that the "calpulli" was involved in the work. Small dykes across the mouths of streams could have been undertaken by individual families, but with the dyking of larger areas, technical considerations arise that suggest that the work was organized at a higher hierarchic level.  The early dykes were simple constructions  of earth and stone and they would have been liable to erosion.  If they  were constructed in small increments, then effort would have been wasted when the partially completed structures were eroded away or when labour was expended on the maintenance of structures that in their partially completed state served no useful purpose.  This would seem to  suggest that'the larger dykes would have been completed quite rapidly, and this would have required the organization of large labour forces.  173  A second characteristic of dykes is that although they do not themselves occupy a large area, they may control water levels in large areas.  It  is unlikely that the inhabitants of sites suitable for dyking would have behaved in an altruistic fashion, and would have invested a considerable effort in constructing a dyke that was of benefit to others beside themselves.  Under these circumstances it seems more likely that  the work was organized by some institution capable of marshalling labour from the entire area to be serviced by the dyke, and of applying it at the site of the dyke. The "calpulli" and "tlatocan" institutions were clearly well suited for the performance of these functions.  That the "calpulli"  could command the labour and material resources of its members was a well established tradition, and the "tlatocan" could similarly call on the resources of its constituent "calpultin".  Furthermore, both  institutions were quite stable, endorsed by history and bound by kinship, and so they were appropriate for the organization of maintenance work on the installations once they were completed. Small installations could be constructed and managed with the resources available to the "calpulli" or "tlatocan", but with larger installations the participation of institutions capable of commanding greater sources of labour and materials was required. When hydraulic installations became so large that they required a significant part of the resources of the city-state as a whole for their construction and operation, the state itself became the institution that was necessary for the management of the undertaking. At this point the government of the state became identified with the management of hydraulic  174  agriculture, and the subsequent history of the development of institutions for the management of hydraulic agriculture becomes identified with the development of the state itself. The period during which the tribute states were the dominant political institutions of the Valley of Mexico is historically remote, and the hydraulic installations that were constructed during this period were obscured by later constructions, with the result that we know virtually nothing of the works that were organized by these polities.  The importance of the Tepanec empire suggests that it might  have enlarged its. resource base by constructing large areas of chinampas along the western shores of what was to become the Lake of Mexico. Because of its high salinity, Lake Texcoco was unsuitable for colonization for agriculture, but the Texcocans are associated with an expertise in hydraulic engineering and they may have been able to exploit the discharge of fresh water from the slopes of Sierra Nevada to construct some areas of chinampas along the eastern shores of Lake Texcoco. There is some evidence of early chinampas in the vicinity of Xochimilco and Culhuacan, and both of these states may have been successful in colonizing quite large areas of Lake Xochimilco. Lake Chalco was shallow and amply provided with fresh water which made it the most suitable of the lakes for colonization.  There are ethno-historical  indications'that hydraulic agriculture was practiced at an early date and it is probable that this lake was the first to be fully controlled and adapted to chinampa agriculture.  The tribute states had the capacity  to organize labour and materials for large hydraulic works and, although the assertion cannot be substantiated in any detail, it is probable that the governments of the tribute states organized the colonization of quite large areas of the lake complex for chinampa agriculture.  175  Virtually all of the large hydraulic installations that the Spanish encountered when they arrived in the Valley of Mexico were constructed after the victory of the "Triple Alliance" in 1433, and the ethnohistorical sources provide descriptions of the construction of two of these large installations. was constructed in 1466.  The Albarradon de Nezahualcoyotl  The work was initiated by Moctezuma the  Elder, the "tlatoani" of Tenochtitlan, who . . . sent messengers to the kind of Texcoco . . . asking that he prepare a plan to prevent the flooding of the city (of Tenochtitlan) because many buildings were ruined and had already collapsed. Nezahualcoytl came quickly to Mexico (Tenochtitlan) and told Moctezuma that the best and most efficient remedy would be to construct a barrier of wood and stone which would hold back the water so that it would not reach the city (Palerm 1973:83). Torquemada's account then proceeds to details of the materials employed in the construction, but he adds that both of the rulers initiated the work with their own hands (Davies 1977:92). We also have some detailed information on the construction of the aqueduct from Coyoacan to Tenochtitlan in the last years of the 15th Century and of the disastrous flooding that was caused by the aqueduct. The Mexica required the water for the operation of their chinampas because, without it, their plants were "drying out and withering." With the approval of all his advisors, Ahuitzotl, the "huetlatoani" of Tenochtitlan, sent two high ranking courtiers to the ruler of Coyoacan with his request for the use of the water, together with a. military detachment to back up the request.  Tzutzumatzin, the ruler of  Coyoacan, acceded to the request because "he and his republic were the vassals of (Ahuitzotl) and were obliged to comply," but he added a  176  warning to the effect that a great deal of water was involved, and that i f the project was carried through it might result in flooding in Tenochtitlan (Davies 1977:92).  In spite of this warning and against the  advice of his advisors who were persuaded of the danger, Ahuitzotl proceeded with the construction, having first found it necessary to arrange the assassination of Tzutzumatzin. Labour was provided by Tenochtitlan's tributaries, as were the materials required for the construction.  In addition to labour,  Texcoco provided light and heavy stone; Xochimilco provided tools and canoes loaded with earth; Calco provided timber; and the Tepaneca provided heavy stone.  "Grand masters" and "architects" were employed,  as were the best craftsmen that could be found (Davies 1977:92). The aqueduct was completed and, after the appropriate and elaborate rituals and ceremonies had been performed, the flow of water from the springs at Coyoacan was directed through.the aqueduct.  Nothing untoward  occurred for a few days but after a while the waters of the Lake of Mexico began to rise dangerously.  Ahuitzotl attempted a number of  entirely supernatural solutions to the problem, but these were ineffectual and he was finally obliged to ask the advice of Nezahualpil1i, the son of Nezahualcoyotl and the ruler of Texoco.  Nezahualpilli "assembled  all the architects who were in his realm" and visited the scene of the disaster (Ixtlixochitl, quoted in Palerm 1973:111).  He then criticized  Ahuitzotl for having killed the respected ruler of Coyoacan, for having ignored the warnings that he had been given, and for having offended the gods.  He then gave the advice that the aqueduct should be dis-  mantled, that the flow of water should be allowed to return to its original course, and that a girl should be sacrificed to propitiate the  177  relevant deities.  These measures were duly carried out, and the flood  subsided. These accounts of the construction of the Albarradon de Nezahualcoyotl and of the Coyoacan aqueduct include virtually all of the information that we have on Aztec management practices.  The most  notable feature of the accounts is that the rulers of the two most powerful cities in the Valley are described as being personally involved in the projects, and the rulers of Texcoco in particular are described as having a special expertise in hydraulic matters.  While it seems  clear that these individuals did participate personally, it may be the case that their participation was entirely in a ceremonial capacity and that the decisions that are attributed to them may in fact only have been made in their names.  However, it may be that the texts are liter-  ally true and that Nezahualcoyotl and Nezahualpilli were, as individuals, highly qualified hydraulic engineers.  There are in addition the refer-  ences to "councillors", "architects" and "grand masters" which would seem to imply the existence of a group of individuals who had particular competence in hydraulic matters.  That the rulers should even aspire to  be considered as a part of this group indicates that the hydraulic managers were of high social standing.  The texts also contain refer-  ences to skilled craftsmen and to divers "who had learned to go under water."  The early Nahua-Spanish dictionaries contain several terms  that describe jobs associated with hydraulic matters:  "atzaqui", for  instance, means "one who stops the flow of water" and "tlapechoani" means "one who blocks up a hole with something heavy" and carries with it the implication that the hole is one through which water flows (de Lameiras 1974:28).  This suggests that the specialists commanded  178  the services of skilled workers who specialized in jobs relating to the manipulation of water.  The only other concrete information that we  have with reference to the management of water resources are statements to the effect that the canals were cleaned out on specified.days, and that the lower courts.were charged with the responsibility for seeing that the work was carried out (Palerm 1973:239). The regularity of the layout of some of the existing areas of chinampas, and the experience of Colonial administrators who found that the various lakes of the complex could not be managed individually, give the impression that the hydraulic installations of the Valley of Mexico were' the product of centralized planning, and that the planners had administrative control over the whole of the lake complex (Armillas 1971: 660; Calnek 1972:111; Palerm 1973:22).  The few historical details  that are available on the planning, construction and management of the hydraulic engineering works confirm this impression and indicate that it was the governing institution of the Aztec state itself that was responsible for the management of hydraulic installations in the lake complex. The State and its Hydraulic Installation There are a number of instances where general political or social characteristics of the states that developed in the Valley of Mexico appear to reflect or to correspond with specific characteristics of the lake basins in which the states developed.  These correspondences  betweeen the physicaland social systems are most clearly evident in the contrasts between the Mexica state and the Chalco-Amecameca Confederation.  179  Of all the lakes in the Valley, Lake Chalco was the most suitable for conversion to chinampa agriculture.  It was shallow, well  provided with fresh water from springs along its southern shoreline, favoured with a well defined mouth, and elevated sufficiently high above Lake Texcoco that discharge could occur in all but exceptionally wet periods.  It was also the first in the linked series of southern  lakes, with the consequence that it suffered none of the complications of fluctuations passed down from hydraulic installations upstream. The Lake of Mexico, in contrast, was the least suitable of the lakes for colonization.  It was brackish and it had no well defined mouth  that could be easily dyked.  It was initially at the same level as  Lake Texcoco, with the consequence that drainage could only occur by the artificial elevation of water levels, and even then it could be prevented by quite slight increases in the level of Texcoco.  Because  it was the penultimate lake in the series, its regime was liable to disruption by hydraulic activities in the lakes upstream. There are several ways in which the contrasts between the two states that developed in these basins can be related to contrasts in the physical characteristics of the basins.  The first of these con-  cerns the times at which the two states emerged.  The origins of the  Chalco-Amecameca confederation are obscure, but it probably emerged towards the end of the 12th Century and it was a significant power in the Valley by the mid 13th Century.  The Mexica state, in contrast,  was not even established until 1345, when Tenochtitlan was founded, and it was not a significant power in the Valley until the last quarter of the 14th Century (Davies 1977:47).  The differences in the times at  which the two states developed may be related to the fact that Lake  180  Chalco was suitable for conversion to chinampa agriculture, while the Lake of Mexico was not. A second contrast is between the internal structures of the two states.  The Chalco-Amecameca confederation appears to have been a true  confederation in which some twenty-six independent city-states formed a voluntary, stable and enduring alliance in order to further projects that were in their common interest.  The fact that it was a coopera-  tive institution that developed in the Chalco basin may at least in part be related to the physical characteristics of the basin itself. It functioned as an open system from which any excess volumes of water could be freely discharged.  A dyke constructed at Cuitlahuac would  have been of benefit to all the agricultural communities on or around the lake, and it was therefore in the interests of each of these communities to cooperate with the others in the construction of the dyke and'its associated canals and secondary dykes.  The Mexica state was  unambiguously a tribute state, assembled by force and maintained against the will of its constituent cities and city-states.  This characteristic  and its association with the hydrology of the basin is illustrated by the events surrounding the construction of the Coyoacan aqueduct, when the Mexica found it necessary to assassinate the ruler of Coyoacan, and to threaten the use of their army in order to obtain the supplies of water that they required.  Coercion appears to have been an integral  part of Mexica water management practices, and this appears to reflect the fact that the Lake of Mexico was not well suited for colonization and that fresh water was in short supply. The contrasting characteristics of the basins of Lakes Chalco and Mexico seem also to have had some effect on the manner in which  181  the states that developed in those basins conducted their relationships with polities outside the areas of the drainage basins in which they were located. date.  The Chalco-Amecameca Confederation emerged at an early  It was a.powerful polity that was capable of fending off the  Tepanec and Mexica armies in the late 14th Century and of resisting the Aztec armies until long after all the other states in the Valley had succumbed.  However, although it was powerful, it never extended  its power beyond the boundaries of the Chalco basin.  Because of its  upstream location, water level could be brought under control as effectively as possible with the prevailing techniques by the construction of. the dyke at Cuitlahuac, and the fact that the polity that developed in the basin never extended its power beyond the basin may be related to the fact that there was no technical advantage in doing so. Again, the situation in the Lake of Mexico was quite different. With its downstream location it was the recipient of excess volumes of water passed on from Lakes Chalco and Xochimilco, and even modifications to the regimes of Lakes Zumpango and Xaltocan in the north of the Valley could have had quite immediate consequences.on the levels and behaviour of the Lake of Mexico. At the same time, there was no natural difference in the level of Lakes Mexico and Texcoco, with the result that levels in the Lake of Mexico had to be artificially maintained if discharge was to occur at all and, even then, high water levels in Texcoco frequently prevented discharge.  The only technically feasible mechanism  available to the Mexica for alleviating their problems was to exploit the upstream capacity of the lake complex.  Because such an expedient  was not in the interests of the polities located upstream, the only means by which the Mexica could make use of this storage capacity was  182  to gain political control of the upstream polities and coerce them into accepting the floodwaters that would otherwise cause damage in Tenochtitlan.  This appears to be precisely what the Mexica did. Within a  few years of founding Tenochtitlan, they embarked on aggressive campaigns against Culhuacan, Cuitlahuac, Mixquic, Xochimilco and the ChalcoAmecameca Confederation and, although the campaigns were related to a number of purely political factors, they were in precisely the areas that the Mexica needed to control in order to avert flooding in the Lake of Mexico.  Subsequent campaigns as a member of the "Triple  Alliance" gave the Mexica control of the entire lake complex. The Chalco-Amecameca Confederation developed at an early date, its constituent city states were linked by cooperative institutions and, although powerful, it did not extend its power beyond the boundaries of the Chalco basin.  These general characteristics of the state appear  to bear some relationship to the fact that Lake Chalco was suitable for colonization, that fresh water was in sufficient supply to satisfy the needs of all the constituent city states, and that there was no technical advantage to extending power beyond the basin.  The Mexica state devel-  oped late in the history of the Valley, its constituent city states were bound together by force, and it finally expanded its power over the whole of the watershed, and beyond.  These general characteristics of  the state appear to be associated with the difficulty of colonizing the Lake of Mexico, with the necessity of acquiring supplies of fresh water at the expense of those city states already using i t , and with the need to control the upstream storage capacity of the lake complex.  183  The Management of Agriculture The management of hydraulic installations was only one aspect of the management of the chinampas': equally important was the management of the more purely agricultural aspects of the system.  Direct  information on the management of the chinampas is again quite limited but it is not as tightly limited as is the case with respect to hydraulic management.  Agriculture is closely associated with social  and political patterns, which were matters in which the Spanish Chroniclers had some interest, and so there is a larger basis for inference about the management of agriculture than there is about the management of hydraulic installations.  At the same time, management  of agriculture at low hierarchic levels is an aspect of a peasant society, of a "little tradition" of practices which have persisted through the intervening five hundred years in a way that the "great traditions" of the Aztec state have not, with the result that many of the techniques and activities of contemporary Indian farmers bear at least a general resemblance to the practices of their pre-Hispanic forbears.  For both these reasons our understanding of agricultural  management in Aztec society is better than our understanding of hydraulic management. Decisions that had a consequence in the physical structure of the chinampa agroecosystems were made at various hierarchic levels; the decisions that involved the least expenditure of energy and that had consequences in the physical system for the shortest periods of time were made at the lowest hierarchic levels.  The basic organizational  unit of Middle American society is the farming family.  This was as true  in Aztec society as it is in the more remote parts of Middle America  184  today.  Early records from Cortez estate in Morelos show that the 1  family typically consisted of a single married couple and their children, but that this unit was occasionally extended to include the spouses of the children and to include the children of the next generation.  In Tepoztlan, in the 16th Century, the average household included  5.4 individuals and 1.4 married couples (Carrasco 1971:363ff).  There  are some indications that "chinampero" families were slightly larger, but it is unlikely that the average family comprised more than about half a dozen individuals. Full time agricultural activity was restricted to the male members of the family, and a father and his sons made up the typical unit of agricultural labour; women assisted in the performance of light work and at exceptionally busy seasons.  Under the "calpulli" system,  it was the head of the farming family who held usufruct rights to land. Generally the head of the family was the father, but widows could hold rights to the use of the land so long as they had male relations to perform the work.  The head of the family was in effect the manager  of the land, and it was he who made most of the short-term decisions about which of the available selection of crops should be planted, where within his small domain they should be planted, and when to carry out such commonplace tasks as sowing, weeding, fertilizing and irrigating. The situation of the "tlalmaitl" families who were tied to the large estates of the late Aztec period is a little less clear.  We know  that the agricultural activities in these estates were closely supervised by "calpixque" who lived in the vicinity, and who directed the activities of tenant farmers (Parsons 1976:236).  It is possible that  these officials functioned as labour bosses, organizing every detail  185  of the work of groups of agricultural labourers.  Certainly, officials  known as "tequitl" performed this function in the execution of public work projects, and they may have performed a similar function in agricultural labour.  However, such an arrangement runs counter to all  contemporary Middle American practice, and it seems more likely that the heads of the "tlalmaitl" families acted as the managers of their small plots in the same way that the heads of "macehualtin" families did, albeit under the supervision of resident "calpixque". Resident "calpixque" supervised the work of the "tlalmaitl" families, and the "calpulli" councils performed the same function with respect to the "macehual" families.  Although individuals in this  latter group had rights to. the use of land, the rights were conditional on the land being cultivated.  If it was not cultivated in two consecu-  tive years the family could be dispossessed and replaced by a more industrious one (Carrasco 1971:363ff).  The "calpulli" council was pro-  vided with officers who could enforce the decision if necessary. The use of the land appears to have been very closely supervised in the late Aztec state; maps were kept by the supervisors, which showed individual plots, the name of the owner, the number of people in the family, and which included a description of the soil in the plot (Gibson 1964:300).  Whether these records were kept by the "calpixque"  or the "calpulli" is unclear, but their existence suggests that the supervision was very careful indeed.  The effect of this supervision  was to insure that the agricultural land was kept in production, and that it produced as much as possible.  Even if the powers of the  supervisors were limited to the removal of the farmers who performed poorly, their decisions resulted in modifications to the physical  186  structure of the land, and in the sense in which the term is used here, both "calpulli" councils and "calpixque" functioned as agricultural managers. The "calpulli" councils and the "calpixque" were responsible to higher order institutions, but it appears unlikely that the governors of cities or states were actively involved in the day-to-day management of agriculture.  Once the goal of maximizing yields had been established  and the responsibility of realizing the goal had been delegated to lower order authorities, the function of the government in the day-to-day management of agriculture was discharged.  The overt concern of the higher  order institutions would have been to enlarge the resource base, and in the chinampas the decisions that were made with a view to accomplishing this goal would have related to the construction of hydraulic works and impinged on hydraulic management. However, the higher order institutions did manage agricultural resources, but they did so in a manner very much more diffuse but fundamentally more important than the manner in which the lower order institutions did. There is an interesting text that relates to the war between the Mexica and the Tepaneca, following the death of Tezozomoc. At issue was the question of whether or not the Mexica should rebel against their former masters.  On one side of the issue were the  Mexica nobility, who favoured the war, and on the other side, "the people", who did not.  After some discussion, Tlacaelel, the "woman  serpent" or vice-ruler of the Mexica, spoke on behalf of the nobles, saying "If we are unsuccessful in our undertaking, we will place ourselves in your hands, that our bodies may sustain you, and you may thus take your vengeance and devour us in dirty and broken pots." The  187  spokesman of the people responded, "and thus we pledge ourselves, i f you should succeed in your undertaking, to serve you and pay tribute, and be your labourers and build your houses, and to serve you as our true lords." The outcome of this strange bargain was, of course, that the nobles won the war and the service of the people (Davies 1977:65).  The story may be apocryphal or it may have been deliberately  fabricated at a later date to justify the privileged position of the nobility.  But whether true, apocryphal or a fabrication, it indicates  that the development of the centralized and stratified state was an issue that was overtly discussed, and the consequence of deliberate decisions rather than the unconscious expression of underlying developmental trends.  If this is the case, then it is also the case that the  creation of estates and the apparatus necessary for their administration was the product of deliberate decisions, and to the extent that these novel institutions modified the physical properties of the agroecosystems by insuring that they were well managed, the governing councils of the emerging Aztec state were acting deliberately and overtly as agricultural managers. In summary, management decisions were made at a variety of different hierarchic levels, by a variety of individuals and institutions within the Aztec state.  The government itself was directly involved  in the construction of hydraulic installations and in the creation of the administrative apparatus for the maintenance and agricultural exploitation of those installations.  Decisions made at these high  hierarchic levels involved the expenditures of large amounts of cultural energy and made an impact on the physical structure of the managed systems that endured for long periods; in the case of the decisions made  188  with regard to the layout of the chinampas, the effects are s t i l l evident half a millennium later.  Decisions were also made at lower  hierarchic levels, by the "calpixque" and the "calpulli" councils who oversaw the production of individual "chinampero" families. At the lowest hierarchic level were the decisions made by the head of the farming family with regard to the use of his small agricultural plot and to the plants that were cultivated there.  These decisions  regulated the use of only a small amount of energy and had consequences in the structure of the physical systems for a correspondingly short period of time. The Evaluation of Aztec Management Management involves the making of decisions that have the effect of modifying the physical structure and thence the function of the managed systems.  Essentially this process is to do with the transmis-  sion and.modification of information.  Management may be conceived  of as a circuit in which information about the state of a physical environment is collected by the managers, processed and supplemented in various ways, and returned to the physical system in the form of directives for the modification of the system. stages are involved.  Four quite distinct  First, the physical system is perceived, either  directly by the managers or indirectly by agents who report to the managers.  Second, the perceived state of the system is compared to  some model held by the managers of a desired state of the system.  The  model of the desired state of the system is derived from the goals of the society in which the managers are operating, modified by the manager's knowledge of what is or is not feasible with regard to the  189  nature of the system and the nature of the technology available.  Third,  a strategy is selected by the managers that will, in the event of a discrepancy between the perceived and desired states of the system, bring the perceived state into conformity with the desired state or that, in the event of a conformity between the two states, maintain that conformity.  The set of strategies available to the managers are  derived again from general characteristics of the society in which the managers operate, in particular, from the technical and administrative capabilities of the society.  Fourth, and finally, the selected strategy  is implemented on the physical system, either by the managers themselves or by their agents.  When the modification has been implemented on the  system, the whole process is then repeated. This is a highly generalized conception of management and it can be applied to the management of any physical system in any society, at any scale.  The model is useful because it provides some means of  evaluating the performance of the management of a physical resource system. As the model indicates, management involves the transmission of information, and a potential source of inefficiency in resource management is poor communication. There are a number of reasons for suggesting that the Aztecs were quite efficient at communicating information about the state of their resource systems.  The management systems com-  prised only a few individuals, and these were people who through shared culture and experience were to a large degree mutually intelligible. This is evident in the case of management at the level of the farming family, where communication was between members of the same family. A few more individuals may have been involved when management was carried  190  out at higher hierarchic levels, by "calpulli" councils or by "calpixque". However, the fact that agricultural overseers were resident in the areas that they administered suggests that tours of inspection were among their duties, and that they were appraised of the states of the systems that they managed by personal observation, and any instructions that they issued were extremely unlikely to have been misunderstood by the heads of the farming families who implemented the decisions of the "calpixque" on the systems.  Even at high hierarchic levels, there was little scope  for confusion in communications.  The history of the Valley of Mexico  is so.complicated that one tends to forget how small the Valley is. Managers who were resident in Tenochtitlan or Texcoco could easily have visited all the localities over which their domain extended. Certainly, the records state explicitly that Nezahualpilli visited the site of the ill-fated aqueduct at Coyoacan before he suggested methods of dealing with the crisis.  Even if the hydraulic managers did not  base their decisions on personal observations, it is unlikely that more than one individual was interposed between the managers and the managed system.  As such individuals would have been known personally  to the manager, and would have been specialists in hydraulic matters, the possibility of confusion would have been slight.  Rather larger  numbers of individuals were involved in the execution of management decisions, but the labour force was cooperative and worked willingly. It was supervised efficiently by "tequilatos" and the work itself, although onerous, was quite straightforward.  Managers undoubtedly  visited the sites of construction to check on progress, and all of this would have reduced to a minimum the possibilities for confusion and misunderstanding in the execution of management decisions.  191  It is a property of hierarchic systems that as messages are passed up through the system information is lost by aggregation; data that at low levels in the system may be discrete are aggregated as they ascend to the apex of the system (Simon 1962).  Thus a management  system with many hierarchic levels responds to very much less specific information.about the state of the system than does one with only a few hierarchic levels, or, one might say that management with few hierarchic levels is capable of a very much more detailed and precise response than management with many levels. There appears to have been a tendencywithin the Aztec state to manage physical systems at the lowest hierarchic level that was feasible.  Each farmer was the manager of his own small plot and he  could respond to quite particular conditions of the system, whether by rectifying problems or exploiting variations.  The scale of management  was so small that the Aztec chinamperos could have treated each plant individually, as their descendants do today.  "Tender loving care"  is a property of agricultural management with few hierarchic levels. It is true that the work of the farmers was supervised, but the supervisors did not substitute for the function of the plot managers, they merely insured that the managers fulfilled their functions.  At the  other extreme of scale, the construction of "hydraulic installations was managed at a high level in the administrative hierarchy simply because high level administrative institutions were necessary for the organization of labour on the appropriate scale, but this is not to say that the management system itself had many hierarchic levels.  If  the managers were supplied with information by specialists who functioned as surveyors and if they issued instructions to labour supervisors,  192  the number of hierarchic levels involved would have been quite small even if a large number of individuals participated in the actual execution of the work.  Considering the fairly simple technologies involved,  there was no need for many levels of management and, if this was the case, then the hydraulic managers as well as the individual farmers could have responded to quite specific characteristics of the systems that they managed. Beyond fulfilling basic subsistence needs resource systems are intended to satisfy social goals, and when a society attempts to satisfy more than one set of goals from the same resource system there is a potential for inefficient management.  Either compromise decisions  must be made that satisfy neither set of goals entirely, or one set of goals must be forcibly imposed on the others. conflicts of interest within Aztec society.  There were undoubtedly The Mexica's requirements  for a plentiful supply of fresh water in the Lake of Mexico conflicted with the interests of those cities such as Coyoacan where the water was already in use, and the fact that the Mexica were forced to exploit upstream storage capacity for the control of floods in Tenochtitlan conflicted with the interests of the inhabitants of the Chalco and Xochimilco basins.  At the same time the interests of the ruling classes  conflicted with those of the ruled.  The conflicts were resolved by  force and, to the extent that the maintenance of an army and a police force reduced the agricultural labour force, consumed agricultural surplus: and disrupted agricultural production, the need for enforcement impaired the efficiency of Aztec management.  However, although the  use of force was itself an inefficiency, the coercive apparatus of the Aztecs was well organized and was probably as efficient as it was  193  possible for such a set of institutions to be.  At the same time, any  tendency that the state may have had.to dissever along social or political boundaries was held to a minimum by an ethic that emphasized duty to the state above all other values, and by an awesome and sanguinary state-sponsored religion that communicated the idea that the existence of the universe was dependent on submission to the authority of the state (Wolf. 1959:144; Peterson 1962:110-11 ).  There were conflicting  goals in Aztec society, but the goals of the ruling class were efficiently propagandized and enforced and the potential for inefficient management was. held to a minimum. Management is based on the knowledge that the managers have of the systems that they manage.  All the indications are that the Aztec  managers had an extremely thorough knowledge not only of their resource systems, but of all aspects of their physical environment.  One chronicler  notes that "almost a l l , including the boys, know the names of all the birds., animals, trees and herbs., knowing as many as a thousand varieties of the latter, and what they are good for (Zorita 1963:163).  This is  what Polyani calls "personal knowledge", learned by individual observation and long apprenticeship (Polanyi 1944).  It is knowledge that is  restricted to the small localities in which most individuals lived out their lives in pre-industrial societies, but within these small areas the knowledge is extraordinarily detailed—the accumulation not just of a single lifetime's observation, but of the observations of many generations transmitted by precept and example.  It is clear that the Aztec  agricultural managers based their decisions on this type of knowledge, learned by apprenticeship to their fathers.  The purview of the  hydraulic managers was somewhat larger and may not have been quite as  194  intimate as the knowledge of the farmers but, even so, the areas involved were small, the specialists probably served long apprenticeships, and the knowledge upon which they based their decisions would have been extremely detailed. A final quality of Aztec resource management was that all aspects of the process had been quite thoroughly tested by experience.  Pre-  industrial societies generally tend to be quite conservative, and when development and change do occur, they occur over long periods of time by incremental modifications to existing practices rather than by the rapid adoption of new ones.  This was evidently the case in such basic  aspects of agricultural management as the selection of crops and the repertoire of techniques for weeding, fertilizing and harvesting which the chinampas shared with other, older Middle American agricultural systems.  Even the practices that were peculiar to the chinampas such  as seedbedding and transplanting with employed bottom mud are unlikely to have been the consequence of specific inventions, but were developed slowly over long periods.  It was suggested above that even the basic  strategy underlying chinampa agriculture was a development of the basic strategy of Rosa cultivation.  Similar considerations apply to the  development of hydraulic installations which, it was argued, although they developed quite rapidly, developed by the elaboration of existing facilities and practices—practices that had already been quite thoroughly tested by experience.  What is true of the physical installa-  tions is also true of the institutions by which the installations were managed and operated.  The elaborate apparatus of the 16th Century was  not suddenly imposed on the society, but developed from the earlier "calpulli" institutions and from the residual institutions of the  195  earlier cultures of Tula and Teotihuacan.  Virtually all of the physical  installations of the Valley and of the administrative institutions had been validated by generations of experience.  It appears to have been  the case that only very late in the 15th Century did the Aztecs begin to employ procedures that had not been validated by experience, and that caused serious technical problems. There is a yet more fundamental aspect to this characteristic of experiential validation of techniques and procedures.  Any culture  orders its experience of the objective world in terms of a set of religious beliefs and an epistemology that provide models of the manner in which the physical environment functions; some of those models provide a more adequate basis for interaction with the objective world than others.  We do not know in what terms the Aztec farmers understood  the processes that they manipulated.  The goddess Tlazolteotl, or  "filth eater", thrived on the confession of sins, but there also indications that she was a representation of the earth and that she thrived on filth in the quite literal sense of the term, and this may have been an expression of the role of organic fertilizers in regenerating the soil and promoting plant growth (de Lameiras 1974:39). Modern Mayan farmers believe that their maize plants are endowed with spirits so that if the plot in which the plant is growing is not weeded the spirit of the maize will be offended and move to a cleaner field, and the Aztec farmers may have held similar beliefs (Vogt 1969: 46).  If the construct of the earth as a deity who thrives on filth  results in fertilization, and if the construct of sentient maize taking umbrage results in the weeding of plots, the plants will do well, and the constructs are validated.  We do not know how the Aztec farmers  196  understood the processes that they manipulated, but whatever models or epistemologies they did employ, they were models that had been tested by generations of farmers and that had been validated by the accumulated experience of those generations. Management is the control mechanism of resource systems.  Its  essential function is to maintain or induce conformity between the physical state of the system and some state of the system that is desired by the managers.  The argument is made here that Aztec manage-  ment was efficient in communicating information, whether information about the current state of the system or information that resulted in modifications to the system; that the information which was communicated to the managers and upon which they based their decisions amounted to a thorough and detailed description of the systems that they managed; that the managers were not confronted with the need to make compromise decisions that satisfied conflicting goals, and that the techniques and procedures employed by the managers were for the most part quite thoroughly tested by experience. The desired state of the chinampas was the one that resulted in high yields, and Aztec management was effective in bringing this about.  Management decisions were based on an intimate and tested  knowledge of plant cultivation, on the conditions of the immediate environment of agriculture, and of the hydrological system in which the chinampas were embedded.  The management subsystems operated at  a scale very close to the scale of the managed systems with the result that it was capable of a rapid response to changes in the condition of the systems and of a specific response to minute spatial  197  variations within the systems.  Aztec management was, in short, effi-  cient at fulfilling the goal of maximizing food production.  198  CHAPTER VI  SUMMARY AND CONCLUSIONS  At the outset the question of how the chinampas worked was posed and it was suggested that a useful method of answering the question was to provide a description of the structure of the system and an account of its functions.  The chinampas were discussed in  terms of their three major subsystems, namely, the hydrological system in which the platforms were constructed, the ecological systems in which crop plants were raised, and the management system that regulated the function of the physical subsystems.  These three subsystems have been  treated at some length, and a number of conclusions have been presented as to their structure, function and interrelationship. It was argued that the lake complex of the Basin of Mexico had two properties in particular that were of potential advantage to agriculturalists:  it had water available throughout the year and it was  particularly rich in nutrients.  In addition to these potential advan-  tages, however, it had the major disadvantage that, as the consequence of large climatic variations, water levels in the lakes fluctuated so greatly both from season to season and from year to year that installations constructed in the complex were either rendered unusable in many years, or else had to be built to impractically large tolerances.  The  only means of exploiting the lake complex for its water and nutrients was to reduce the values of these fluctuations.  The only method of  doing this was to construct a complex of dykes and canals that removed  199  excess volumes of water during times of surplus and that held water back during times of shortage.  The costs of such constructions were  high and they were not constructed in any significant numbers until a point had been reached when the population of the Valley pressed so hard on its carrying capacity at existing technological levels that there was no alternative but to invest the necessary effort.  From  the mid 14th Century onwards, hydraulic installations were constructed and water levels in increasingly large areas of the lake complex were brought under control.  With the levels controlled the complex could  be colonized and its stocks of water and nutrients exploited for agriculture.  The immediate function of the hydraulic works was to control  water levels, but their function in the chinampa agriculture system was to make available material inputs to the crop systems at values which could be exploited. The second major system of chinampa agriculture comprised the multitude of chinampa platforms together with the small and essentially similar ecological systems that each"platform supported.  It was suggested  that the ecosystems of the chinampas are best understood as comprising three quite separate systems, each one of which had different structural properties, and each of which performed different but complementary functions in the chinampa plant community as a whole.  One of the sub-  systems was an agroecological system which performed the function common to all such systems of combining nutrients with solar energy to produce biological materials that were of use to man. Like any other agroecological systems, the crop plant communities of the chinampas were valued for their high growth rates, but for reasons relating to the nature of ecological systems high growth rates could only be had at  200  the expense of stability.  The chinampas were high-yielding and, with  yields of three or four tons per hectare, they compared favourably with modern agricultural systems.  To the extent that they were high yielding,  however, they were also vulnerable to pests and the competition of plants more.robust than crop plants are.  High yields could only be had  at the cost of large investments of energy in producing appropriate growing conditions for the crop plants and in protecting them from depredatory animals and competing plants.  It was in this capacity that  the two other sub-ecological systems of the chinampas functioned. The canals and the platform borders, together with the plant communities they supported, performed the function of supplying the crop systems with the materials they required and with protection from insect pests, soil pathogens and variations in microclimatic values.  To the extent  that the non-crop ecosystems performed they reduced the amount of labour that had to.be expended directly in maintaining the fragile crop plants and increased the yields of chinampa agriculture.  In the majority of  agricultural systems these functions are performed at the cost of cultural energy, either in the form of human or animal work or in the form of commercial energy products.  The genius of chinampa agriculture  lay in the fact that ecological systems were employed to perform much of this work and by this means to provide the necessary inputs of energy, not in the form of scarce cultural energy, but in the form of solar energy. The third major subsystem in the chinampas was the management system by means of which the physical structures of the other subsystems were modified and their functions regulated.  The limited  evidence on the management of chinampa agriculture suggests that Aztec  201  managers operated at levels in their society that were roughly commensurate with the size of the labour force that was required for the execution of their decisions.  Most of the day-to-day managerial decisions  made with regard to the modifications of the agroecosystems and their supporting non-crop ecosystems were made at low levels in the social hierarchy by the heads of the farming families and the necessary labour was provided by the small familial work force.  The construction of  hydrological installations required the expenditure of rather greater amounts of energy and these undertakings were managed by institutions or individuals who were situated sufficiently high in the social hierarchy that they could command the necessary resources.  For the larger  undertakings, at least, such large resources of labour and materials were required that the management of the undertakings was carried out at the level of the government of the state itself and, quite possibly, by the rulers of the state. It was further argued that in spite of differences in dimensions and other attributes, the management of chinampa. platforms and of hydrological installations shared some basic characteristics.  In both  cases only a few individuals were involved in the management of the system.  There was' a tendency for the management subsystems to comprise  only a few hierarchic levels, the consequences being that information necessary for management was communicated with a minimum of confusion and aggregation, and that the managers could respond rapidly and accurately to small variations in both the temporal and physical characteristics of the systems which they managed.  In agricultural management  the scale was so fine that plants could be treated individually. Because the managers lived out their lives in small localities and  202  acquired their knowledge by personal observation and long apprenticeship, the knowledge upon which they based their decisions was extremely detailed and quite particular to the systems that they managed.  Because  the development of agricultural techniques and practices and, to a lesser extent, the techniques of water control had been developed over long periods and in small increments, most of the procedures and strategies adopted by the managers had the quality of being experientially validated and thoroughly tested.  Finally, Aztec management was charac-  terized by a certain clarity of purpose and uniformity of goals. Management was intended to produce agricultural surplus for the support of the state, or, more specifically, for the support of the ruling classes of the state.  Managers of high social standing worked willingly  towards this goal because it was in their interests to do so.  Managers  of low social standing also worked towards this goal, not because it was in their interests to do so, but because the entire coercive apparatus of the state obliged them to do so.  The Aztec managers had efficient  communications, a thorough knowledge of the systems that they managed and a well tested set of strategies.  For these reasons they were efficient  in the construction of'installations for the regulation of water levels in the lake complex, in operating these once they had been constructed, and in exploiting the agricultural potential of the lake complex that the hydraulic installations made available. The chinampas, like any other material resource system, were a technical device for transforming the materials of the physical environment into goods for thei satisfaction of human needs and cultural desires.  The chinampas converted nutrients and water into food-  stuffs and by some criteria they appear to have performed this function  203  quite efficiently.  If efficiency is understood in terms of the ratio of  output to input, then the efficiency of a resource system must be evaluated in terms of a specific input; the inputs that are most generally of interest with respect to agricultural systems are land, labour and energy.  The chinampas appear to have been quite efficient with  regard to outputs per unit input of land; exact figures are not available, but the impression is that the chinampas yielded between three and four tons of shelled maize per hectare and possibly a quantity of horticultural products in addition to this.  These yields are respect-  ably high, which is to say, they were very much higher than the yields of virtually all other Middle American agricultural systems and not very far short of the yields of modern corn production. Like other preindustrial systems, the chinampas were labour intensive and their output per agricultural worker was probably in the vicinity of one and a half tons of maize per year, which is a small fraction of the output of agricultural workers in the developed economies.  The returns per unit  investment of energy were again respectably high and with an Energy Ratio of between 10 and 15 they were considerably more efficient than modern commercial agriculture, but only about half as efficient as more extensive traditional systems in Middle America. A second aspect of how efficiently a resource.system performs its function relates to the question of how long it could have continued to perform its function without in some way impairing its ability to do so in the future.  By this criterion, too, the chinampas appear to have  been quite efficient.' So long as the canals and platforms were maintained and new trees planted to replace those that died, the internal  204  structure of the system could have persisted indefinitely.  With regard  to their external relations, the chinampas exploited an exceptionally fertile locality and they did so extremely frugally.  A large part of  the nutrients that were withdrawn from the canals and applied to the platform surfaces were returned to the canals by the processes of decay and surface runoff, where they again became available for future use. The export of nutrients from the platform surfaces was held to a minimum by exporting only the usable portions of crops.  Much of the crop  that was exported from the platforms was consumed locally and the nutrients that they contained were returned to the canals in the form of human waste and household dejecta.  Sewage may have been systematically  collected and returned to the chinampas for use as fertilizer.  Nutrients  were used so conservatively in the chinampas that it is unlikely that they were consumed or exported from the system at a greater rate than the one at which they became available to the chinampas in the course of their larger natural cycles.  If this was the case, then the chinam-  pas did not reduce the capacity of their environment to continue to support them and the system could have continued to function indefinitely.  Nor is it likely that i f the 16th Century hydrological conditions  had been maintained water would ever have become a limiting factor on chinampa production. The chinampas were also quite self-sufficient in terms of energy use.  The only major input of cultural energy was in  the form of the human labour expended in agricultural activities and, to some extent, in the activities associated with procuring lime, bat-dung and.leaf mould. But because these expenditures could have been met from the production of chinampa agriculture they did not represent a subsidy  205  of any kind.  Otherwise, the only input of energy was in the form of  solar radiation and, so long as the sun shone, the chinampas could have continued to function.  Chinampa agriculture functioned quite  independently of any significant source of non-renewable materials or of cultural energy; its inputs were entirely limited to those nutrients and supplies of water that became available in the course of their natural cycles and to solar radiation.  Under these circumstances they  could have persisted indefinitely, and the fact that large areas of chinampas were s t i l l in production at the beginning of this century is direct proof of the stability of the form. There may have been no limits on the amount of time that the chinampas could have persisted, but there were limits to the extent to which the area of chinampa agriculture could be increased.  In a closed  drainage basin such as the Valley of Mexico, the volume of water present at any particular time was determined by the difference between evaporation and precipitation.  The running discrepancy between the values of  these two variables caused changes in the volumes of water stored in the lake complex; these, in turn, caused fluctuations in water levels. Because the Aztecs had no control over the volumes of water present, the most that.their dykes and canals could accomplish was the redistribution of the volumes of water that were already present in the system.  They could prevent water levels from rising or falling in one  part of the lake complex, but only at the cost of increasing fluctuations in some other part of the complex.  Calculations with the hydrological  model suggest that Lake Texcoco had sufficient capacity to absorb the fluctuations of Lakes Xochimilco and Chalco and some small areas in other parts of the system, and that if chinampa agriculture had remained  206  restricted to these areas it would have been quite stable.  The hydraulic  managers of the Aztec state, however, extended water level control into progressively larger areas of the Lake of Mexico and, in so doing, reduced the volume of storage capacity available in Lake Texcoco. The fact that the city of Tenochtitlan stood in the waters of the Lake of Mexico at the time of the Spanish Conquest indicates that the Aztecs had not quite reached the limits to which they could extend chinampa agriculture.  The fact that floods were common in the Lake of Mexico  in the years prior to the Conquest and that it was occasionally necessary to impose floods on the two southern lakes in order to reduce flooding in Tenochtitlan suggests that the Aztecs were very close to the limits to which the chinampas could have been extended.  There  were limits on the development of the chinampas, but up to those limits the system appears to have been remarkably stable and productive. At the beginning of this essay it was suggested that the chinampa system.of cultivation might have a relevance in the larger context of the development of agriculture in the 20th Century and that it might provide an indication of how the energy efficiency of modern agriculture might be improved without sacrificing high yields. Some aspects of chinampa cultivation do seem to provide such indications. The chinampas made a very conservative use of materials:  exports  from the system were minimized and materials were as far as possible recycled within the system.  The chinampas were particularly well endowed  with nutrients and water but, because exploitation was frugal, the rate at which materials in their environment were consumed was very lowprobably lower than the rates at which the materials became available in the course of their large natural cycles.  Because of this the  207  availability of material inputs was unlikely ever to have become a limiting factor on chinampa production. Had the chinampas been less frugal in their use of materials they would have required the input of greater volumes which would, in turn, have increased the energy investment in the system and reduced its efficiency.  This suggests  that the techniques and devices by means of which materials were conserved in the chinampas could serve as models for similar devices in other agroecosystems.  Such practices as the use of canals as  repositories for nutrients, the exploitation of productive aquatic ecosystems for fixing atmospheric nitrogen and for converting nutrients into forms that are readily accessible to plants, the careful cycling of crop residues and field litter, the selection of crops that are closely adapted to local climatic conditions and the use of large perennial plants and mixed cropping techniques for the regulation of microclimatic conditions could be usefully replicated.  To the extent  that such practices reduce the use of water and nutrients, they also reduce the volumes of materials that must be imported to the agroecosystem and they reduce the cultural energy costs implicit in the provision.of these materials. A second feature of the chinampas was a tendency for certain necessary agricultural functions to be performed, not actively by the agriculturalists, but passively as a consequence of the structure of the system itself.  The canals collected and stored nutrients and so  reduced the amount of work involved in procuring these materials; the relative elevations of the canals and platform surfaces reduced the amount of work required in irrigation; and, to the extent that the noncrop ecosystems supported and protected the crops, they reduced the  208  amount of effort that the agriculturalists had to expend in these activities.  In each of these instances, the function was performed,  but the onus for the performance was shifted from the agriculturalist to some aspect of the structure of the system.  Work was involved in  the construction and maintenance of the inorganic and.non-crop biotic components of the system.  But in the case of the inorganic installa-  tions the great stability of the chinampas allowed the investment to be amortized over extremely long periods, and in the case of the biotic infrastructure the work performed by the "chinamperos" was slight compared to the amount of solar energy that was harnessed to the benefit of the crop plant communities.  The tendency of chinampa agriculture  to exploit passive processes that were vested in the structure of the system and that did not require high cultural energy investments, rather than to expend considerable effort in the direct performance of these functions contributed to the energy efficiency of the system and is another feature of the chinampas that deserves emulation. A third quite basic property of chinampa agriculture was that the scale of the various constituent artificial systems was very close to the scale of the natural•systems upon which they impinged. With regard to the ecological systems, the areas that were treated and managed as a unit were very small and, because they were small, certain types of functions occurred that could not have occurred had the units been larger.  Passive irrigation depends on the availability  of water in close proximity to crops; trees can regulate microclimate only within an area of a few square meters; the range over which a generalized ecosystem can extend its beneficial influences is limited by the range over which it influences microclimate, or by such factors  209  as the range of insect predators.  The range over which ecological  interactions of this type can operate is not well known, but the example of the chinampas suggests that it is quite small and that, if the interactions are to occur, any agricultural system that seeks to exploit them must be quite small. itself a desirable quality.  This is not to suggest that small size is of The hydrological elements of chinampa  agriculture extended over the whole area of the lake complex and, because the different parts of the lake complex were interrelated, it was necessary to manage the lake complex as a single unit in order to modify its function successfully.  What seems to be suggested by the example of  the chinampas is not any absolute criteria for the dimensions of energy-efficient agricultural systems, but the need for a concern with matching the scale of the managed units with the scale of the natural processes that are manipulated in these units, and that at a point where a discontinuity between the scale of the managed system and the scale of the managed processes occurs the system should be extended not by increasing the area of the units but by replicating them. The chinampas, then, do seem to offer some indications of strategies for the increase of the energy efficiency of agricultural systems without incurring high energy costs.  This is not to suggest  that it would be desirable to recreate the 16th Century chinampas in every detail.  Much of the work in the chinampas was drudgery and the  productivity of each worker was below the level at which even basic needs could be satisfied in a modern economy.  What is suggested is  that underlying and informing the practice of chinampa agriculture are certain principles and strategies that have the effect of reducing energy inputs to the system, and that by the study of the chinampas  210  and similar intensive pre-industrial agricultural systems, these principles can be understood and used as the basis for the development of biotic resource systems that are efficient in transforming soil and water and sunlight into useful goods.  211  APPENDIX  RECONSTRUCTION OF THE 16th CENTURY HYDROLOGY OF THE  BASIN OF MEXICO  212  Part I  The Use of Modern Data in Reconstructing 16th Century Conditions  In the calculations that follow modern data are used as the basis of a description of the hydrology of the Basin of Mexico in the late 15th and early 16th Centuries. The use of data that are 500 years out of date requires some discussion.  The impression given in  the literature is that the climate of the Valley of Mexico has not changed very greatly during the past 500 years (C.H.C.V.M. 1964 V: 132).  Temperatures are conditioned by latitude and elevation; so  modern figures are probably quite representative of pre-Hispanic conditions except in the immediate locality of larger built-up areas where urban heat islands have developed (Jaurequi 1973-74).  Precipi-  tation values have changed slightly during the past half millennium. Lorenzo has analyzed the elevations of lakeshore settlements at different archaeological periods and from them has produced estimates of long term changes in mean lake levels and in precipitation (Lorenzo 1956).  The results of these analyses are shown in Figure 32 and it  can be seen that over the entire period of occupance there have been some fairly large changes in precipitation values.  However, in the  relatively short period that has elapsed since the Conquest the changes in values have been quite small, with mean values declining from 600 millimeters to just above 500 millimeters.  As the lake complex was  largely supplied with water from areas where mean annual precipitation is in excess of 1000 millimeters the decline is proportionately less.  213  -\  1600  1  1 1 1200  Figure 32.  1 1 800  1 1—~i 1 400 BC/AD  1 1 400  1 1 1 1 1 1 r 800 1200 1600 2000  Changing Lake Levels and Precipitation in the Basin of Mexico, 1600 B.C. to Present (Lorenzo 1956, cited in Sanders 1970:88).  214  Insofar as evaporation rates are tied to temperature and precipitation, they too may have changed somewhat during the past 500 years, and modern values may differ from pre-Hispanic values. Whether or not these differences in precipitation and atmospherically conditioned evaporation are significant depends very largely on the degree of accuracy required.  The present work aspires only to  a very rough approximation of pre-Hispanic conditions, and the differences are assumed to be insignificant.. Modern precipitation data are used without modification, except that some of the isohyets in the vicinity of Mexico City have been smoothed to obscure the effects of the urban heat island in the area.  Modern Potential Evaporation data,  adjusted in the manner recommended by the Comision Hidraulica de la Cuenca del Valle de Mexico (C.H.C.V.M.) as being most appropriate to the estimation of evaporation losses from open water and saturated soils have been applied to the areas of the antique lake complex to provide estimates of evaporation losses from the complex.  This may  result in some inaccuracies in the reconstruction of 16th Century conditions, but within the present frame of reference they are considered acceptable. A much greater source of error is the use of modern data on evaporation rates from groundsurfaces on hillsides, and on streamflow and infiltrated flows in the same areas.  Although not as extensively  modified as the Valley floor, the hillsides have been greatly altered since the 16th Century. Prior to the Conquest the lower slopes of the southern and central parts of the Valley were thoroughly colonized for agriculture.  Indian agricultural techniques involved considerable  use of green manures which, amongst their other benefits, improved the  215  water retention capacities of soils.  Polyculture and the practice of  growing perennials along the boundaries of agricultural plots tended to insure plant cover through much of the year.  Virtually all of the  lower and steeper slopes of the Basin were terraced and irrigation techniques involved the use of thousands of small dams and distribution ditches.  In addition, the upper slopes of the Valley were more  thoroughly forested than they are today.  The combined effect of these  factors was to retain a considerable volume of precipitation on the hillsides.  Soon after the Conquest the Indian population suffered a  savage decline as the result of the introduction of European diseases to which it had no immunity, and land that had been under careful Indian management came under Spanish control.  The Spanish substituted  the monoculture of wheat for the polyculture of maize, beans and squash, which would have reduced the effectiveness of crops as persistent ground cover.  They substituted the plough for the digging  stick which may have reduced the soil moisture retention capacities, and which resulted in the removal of much of the perennial growth that had bounded the small Indian plots.  Most of the terraces and many of the  small irrigation systems were abandoned.  Cattle and sheep were intro-  duced soon after the Conquest, and their numbers grew rapidly.  It is  estimated that by 1600 some 5,000 head of cattle and 76,000 sheep were roaming more or less without restraint over the Valley (Simpson 1955:55). Cattle's hooves compact the soil and the close cropping teeth of sheep discourage the regrowth of vegetation.  The cumulative effect of these  changes on surface runoff would have been to increase the ratio of runoff to storage on the hillsides.  Streamflow would have increased,  and the interval between precipitation and stream discharge would have  216  decreased.  With the removal of vegetation cover and the accelerated  movement of water, the hillsides would have been more liable to erosion and suspended loads would have contributed to a more rapid silting of the lake complex.  The effect of these changes on evaporation and trans-  piration rates are not as clear.  On one hand, actual evaporation rates  would have been reduced by the more rapid removal of water from the vicinity of precipitation but, on the other hand, the reduction of ground cover and the exposure of soil surfaces would have allowed evaporation mechanisms to operate more efficiently, and potential evaporation rates may have increased.  The effects of these changes  on infiltrated flows are entirely obscure, but it is very likely that they were affected the same way.  These changes are particularly  critical in the present work where the focus is on the function of the lake complex.  The hillsides were the catchment areas for the lakes,  and changes in values per unit area that may have been quite slight in the large area of the hillsides would be magnified when concentrated in the relatively small area of the lake complex. It would, in theory at least, be possible to make some systematic adjustment to modern data so that they more accurately represent preHispanic conditions.  However, this work has not been carried out by  other writers and it is beyond the capacities of the present one. The figures used for evaporation and discharge from the hillsides are modern and are used without adjustment.  Because the ratio of runoff to evapora-  tion may have changed since the Conquest, there may be inaccuracies in the calculation of mean annual volumes; because the lag time between precipitation and discharge may have altered, there may be inaccuracies in the calculation of the monthly distributions of these volumes.  217  Unfortunately, there are no means presently available for correcting these inaccuracies, nor for estimating how great they may be. Modern data are also used in the calculation of evaporation losses from the lakeshore plains.  This too may lead to erroneous  results, not only because land use patterns have changed, but also because the areas were very much moister during the 16th Century than they are now and this would have modified evaporation rates.  The  overall effect of these errors on the reconstruction of 16th Century conditions is probably not great, simply because the areas involved are not great. The hydrology of the Basin floor has been so thoroughly modified by the drainage of the lake complex, by the pumping of subsurface water and by imports of water from outside the Basin, that modern data on drainage and storage characteristics clearly bear little relationship to past conditions and thus these data are not used.  In their place  some assumptions about the behaviour of water in the Valley floor have been made on the basis of general characteristics of the geology of the region. Basin.  Figure 33 shows two sections of the southern parts of the  It can be seen that the Valley floor overlies a more or less  level layer of Lower and Mid Tertiary volcanic rocks and that it is bounded by the products of Upper Tertiary and Quaternary vulcanism. At various, points in the region the underlying rocks project towards the surface and the surrounding rocks approach each other laterally to form a series of constrictions that divide the whole area of the Valley floor into a number of distinct compartments, each corresponding with the basin of one of the antique lakes.  These compartments are filled  with recent alluvial materials and with water.  The alluvial materials  218  Quaternary Volcanic Rocks Quaternary Alluvial Fill Tarango Formations  m  Upper Tertiary Volcanic Rocks Middle and Lower Tertiary Volcanic Rocks Possible contact with Sedimentary Marine Rocks  Elevations in m x 1000 above M.S.L.  Figure 33.  Locations of Sections  Geological Sections of the Basin of Mexico (C.H.C.V.M. 1961:69).  219  are largely Bentonitic clays and they contain as much as 83% water by volume (Fox 1965:533).  The clays do not so much contain the waters  of the lakes as sink to the bottom of them, and the enveloping volcanic rocks contain both water and clay.  When water levels rose above the  level of the alluvial f i l l , lakes occurred, and when they rose above the levels of the igneous constrictions, lake discharge occurred. The effect is of a single body of water in each of the compartments, some 801 meters deep near Xochimilco, but with 20% of the lower 800 meters filled with alluvial materials (C.H.C.V.M. 1961:22).  This description  is of course much simplified and the stratigraphy of the alluvial deposits is extremely complex, but it is evident that in the upper few meters of the Valley floor there is no very clear distinction between groundwater storage and lake storage. Water storage in the floor of the Basin is largely subterranean and so too is the discharge of water into these reservoirs.  The geology  of the surrounding mountains and hills is tortuous in the extreme with many kinds of igneous rocks overlaying and interpenetrating each other. In the Xochimilco and Chalco watersheds, some 57.2% of the surface of the lake catchment area is composed of Quaternary fractured basalts, which are extremely porous, and a further 14.6% of the area is composed of older basalts, which are only slightly less porous.  The remaining  areas of the mountains and hills are composed of andesites and volcanic ashes, into which water also percolates quite freely.  With these large  areas of permeable rocks it is estimated that for the entire area of the Basin about 29% of precipitation is infiltrated, compared with only 3.6% which is discharged in surface streamflow (C.H.C.V.M. 1964 V:133, 141; C.H.C.V.M. 1967:86-87).  The Chalco and Xochimilco regions are  220  exceptionally permeable, but even in the other watersheds of the Basin a very high proportion of precipitation is infiltrated. In the light of this information the calculations that follow make the assumption that there was a relatively free exchange of water between surface and groundwater reservoirs.  It is also assumed that  the lakes were charged directly with surface runoff from their catchment areas and that groundwater reservoirs were supplied by flows of indirect and infiltrated runoff.  Finally, it is assumed that over long  periods stocks of groundwater remained approximately constant, and that inputs to these stocks were balanced by outputs either in the form of evaporation from the lakeshore plains or by seepage and indirect flow to the lakes.  The lakes in turn lost water either by evaporation from  their surfaces or by discharge to lakes downstream.  These assumptions  are perhaps over-generalized, but with respect to mean annual volumes they may not result in great inaccuracies.  However, there is a major  problem associated with the lag times, and this probably results in substantial errors in calculating the monthly distributions of inputs to the lakes.  This problem is discussed more fully in Part III of  this Appendix. In sum, modern climatic and hydrological data are used extensively in this reconstruction of the conditions of the 16th Century. The use of modern data on precipitation and evaporation insofar as it is conditioned by atmospheric variables may distort this reconstruction to some extent, but in general they are probably quite representative of pre-Hispanic conditions.  Modern data on drainage and drainage  patterns within the area of the Valley floor are not used because it is clear that they bear little or no resemblance to pre-Hispanic conditions.  221  The simplified model of these patterns that is substituted in their place may result in some inaccuracies, particularly with respect to monthly distributions.  Modern data on evaporation from ground surfaces and on  streamflow, infiltration and indirect flows in the catchment areas of the.lakes are used--and this is liable to be a source of considerable error in the reconstruction.  Unfortunately, there is no convenient  means of making systematic adjustments to these data so that they provide a better representation of pre-Hispanic conditions.  However,  the reconstruction does appear to be worthwhile, even if inaccurate, partly because many of the arguments in the main body of this work are based on relative rather than on absolute values and are not invalidated even i f many of the absolute values are in error, and partly because it does increase our understanding of the problems that Aztec engineers confronted and surmounted and of the means by which they contrived a successful system of hydraulic agriculture.  222  Part II  Derivation of Mean Annual Values for the 16th Century Hydrological System  1.  Drainage Basins The seven separate drainage basins are slightly modified from the  Hydrological Zones used in C.H.C.V.M. Evaporation studies.  The boundar-  ies of the watersheds used in these calculations are shown in Figure 3. 2.  Physiographic Divisions The boundaries of the mountains (m) are those used in the  C.H.C.V.M. Evaporation studies. The lakeshore plains (p) are here taken as existing only in the five central and southern basins, and are taken as the areas between the 2,250 meters and boundaries of the mountain. The valleys (v) comprise the Teotihuacan Valley and the lowlying regions of Pachuca and the Northeast, and their boundaries are those used in the C.H.C.V.M. evaporation studies. The lakes (1) comprise the area of the lake and marsh complex that originally occupied the central portions of the Valley, together with the few small lakes that persist in the Northeast, and they are taken as the area enclosed by the modern 2,250 meter contour. 3.  Precipitation (P) Precipitation values are derived without modification from the  mean annual isohyets (1920 - 1959), published by the C.H.C.V.M. and reproduced in Figure 4.  Maps of the Isohyets and of the Drainage  223  Basins were prepared at the same scale and overlaid, and volumes were calculated by measuring the area in each precipitation class, multiplying the resulting value by the precipitation class value, and summing the volumes so calculated for each physiographic region of each basin. The same procedure was used.for calculating evaporation losses. The total volume for precipitation calculated in this manner for the Basin of Mexico was 6,836 x 10 m , compared with 6,933 x 10 m calculated by the C.H.C.V.M. (C.H.C.V.M. 1964 V:333). 4.  Evapotranspiration (E) a)  Mountains (m), Valleys (v) and the Teotihuacan Valley (t). Values for total evaporation losses have been published by  the C.H.C.V.M. and are available by Study Regions and fractions of those regions (C.H.C.V.M. 1964 V:333).  Because the regions employed  in the C.H.C.V.M. evaporation studies do not exactly correspond with those used in the present study, it was in some instances necessary to proportion a single given value between two regions, and this was done on the basis of measures of areas.  Apart from this, the values given  by the C.H.C.V.M. are used without modification. b)  Lakeshore Plains (p) and Lakes (1) The model assumes a free exchange of water between the antique  lakes and the groundwater reservoirs beneath both lakes and lakeshore plains.  Because of this both these regions were treated in the same  manner.  Modern values for evaporation losses in these areas were not  employed.  Instead, an adjusted potential evaporation rate (Epo. adj)  equal to 70% of the potential evaporation rate was substituted.  This  measure is used by the C.H.C.V.M. as an estimate of losses from open  224  water and saturated soils, and is plotted in Figure 13 (C.H.C.V.M. 1964 V:44).  Values for the present reconstruction were derived by the same  method used for estimates of precipitation, for the area of the antique lakes and the lakeshore plains. 5.  Agricultural Evaporation Supplement (A) Evaporation rates, as calculated by or derived from C.H.C.V.M.  publications, include values for water transpired by wild plants, but do not account for the consumptive use of water in agriculture. To account for these losses, the C.H.C.V.M. employs an agricultural supfi 3  plement (A), which is equal to 680 x 10 m for the entire area of the Basin.  This value has been applied without modification in the present  reconstruction, on the assumption that Aztec agriculture, although less profligate in its use of water than modern agriculture, was more widespread; i.e., spray irrigation was not used in the past, but many areas that were cultivated in the 16th Century are now abandoned (C.H.C.V.M. 1964:VII).  Unfortunately, this single datum is the only information  presently available on the agricultural supplement, and so it has been necessary to distribute the value by the following formula: B. x F. x 680 A.  =  where A. is the volume of supplementary evaporation in a particular region.  225  b\ is either 34%, as it is in the mountains of Xochimilco, Chalco, Mexico or Texcoco, or 100% of the lakeshore plains of those basins, or 25% of the lakeshore plains of Cuautitlan. These fractions are intended as a representation of the areas of cultivable land in the Basin, and the northern parts of the Basin are assumed to have supported only a little rainfed agriculture. F^ is a factor to include variations in the evaporation climate, is equal to the primary evaporation divided by the area of the section for which values are being calculated.  For an explana-  tion of primary evaporation, see Appendix Part III, 3, below. 6.  Surface Discharge (Q) The figures used in the diagram are the measured volumes of  runoff for each of the Zones of the Valley.  In the Xochimilco and  Chalco basins, data are presently available for all of the major streams, and as the drainage areas of these streams correspond with the areas of the mountains of these basins, all streamflow has been allocated to the mountains (C.H.C.V.M. 1964 111:91-94; 141-148).  In the remaining  areas of the Basin of Mexico, only summary data are available at present, which give the total discharge for each of the hydrological zones employed by the C.H.C.V.M., and the total catchment area for that discharge.  In the basins for which complete data are not available,  all discharge has been allocated to the mountains, unless the given catchment area exceeds the measured area of the mountains, in which case the surplus area was allotted to the valleys of lakeshore plains and the total volume of discharge allocated proportionately. ;  226  7.  Undifferentiated Runoff (R) Measured surface discharge accounts for only a small portion of  the drainage in the Basin. Water transfer also occurs by flow in surface channels that are too small to merit metering, by seepage from streams between metering stations and by subsurface flows through fractured Basalts and Andesites. In the absence of any useful estimates of these various flows, they are aggregated as Undifferentiated Runoff (R) which is defined as: R = P -(E - A - Q) These values were calculated for each physiographic region and each basin. 8.  Lake Discharge (QI) Discharge from lakes is simply calculated as the sum of all the  inputs, less the outputs of evaporation, for each of the basins. The assumption is made that over the long term lake storage remained constant. 9.  Mean Annual Totals The values calculated by these various means are shown in  Table XI , and they are incorporated in the Flow Diagram, Figure 2.  227  TABLE XI SUMMARY OF ESTIMATES OF 16TH CENTURY MEAN ANNUAL WATER BALANCE, BASIN OF MEXICO  0)  (3) E m xl0  Area km  (2) P m xl0  Chalco Mountains Plains Lakes Total  880.2 107.7 140.9 1128.8  904.3 85.6 112.1 1102.0  336.4 134.0 183.2 653.6  Xochimilco Mountains Plains Lakes Total  407.4 28.1 99.7 535.2  370.6 14.5 70.7 455.8  Mexico Mountains Plains Lakes Total  741.9 142.5 178.8 1063.2  Texcoco Mountains Teotihuacan Plains Lakes Total Cuautitlan Mountains Plains Lakes Total  2  3  6  3  6  (4) (5) A Q m xlO m xl0 3  6  (6) R 3 6 m xlO 1n  88.0 49.0  40.9  439.0  137.0  40.9  439.0  132.5 31.5 94.1 258.1  41.0 13.0  5.0  683.5 88.7 111.3 883.5  276.4 123.8 194.6 594.8  780.4 172.2 153.1 483.7 1589.4  575.0 108.1 96.6 305.4 1085.1  301.5 205.3 179.8 777.7 1464.3  853.2 232.8 317.2 1403.2  558.1 158.9 216.4 933.4  323.0 264.4 469.6 1057.0  -  -  54.0  -  -  -  6  +311.4  -  •-  192.0  76.0 64.0  151.6 22.9  179.5 65.8  140.0  174.5  245.3  79.0 155.0  39.8  154.7  234.0  39.8  154.7,  115  124.0 5.9  111.0 153.0  115  129.9  264.0  -  3  192.0  5.0  -  (8) QI m xl0  -  +143.7  +148.7  -613.2  -238.6  228  TABLE XI (continued) SUMMARY OF ESTIMATES OF 16TH CENTURY MEAN ANNUAL WATER BALANCE, BASIN OF MEXICO  (1) Area  (2) P  km  m xl0  Pachuca Mountains Valleys Lakes Total  1421..9 631..3 7,.8 2061,.0  778.,4 326.,2 4..0 1108..6  516..0 486..0 9.,9 1011..9  12.,3 3.,2  250.,1 323.,0  15..5  399..1  Northeast Mountains Valleys Lakes Total  1232,.6 561,.8 45,.8 1840,.2  868,,0 369.,8 30,.2 1268..0  417..0 641..7 52,.3 1111,.9  36.,0 16.,6  415..0 353.,0  52..6  768..0  +157.0  Total Watershed Mountains Plains Valleys Lakes Total  6317.6 664.2 1365.3 1273.9 9621.0  4737.9 444.3 804.1 850.1 6836.4  2302.8 733.5 1333.0 1781.4 6150.7  458.2  2462.1  +5.7  2  3  (3) E 6  (4) A  m x10 3  6  m xl0 3  (5) Q 6  m xl0 3  -  -  680.0  (6) R 6  m xl0 3  (8) QI 6  -  m xl0 3  6  +96.7  229  Part III  Estimation of Mean Monthly Distributions  The specific intent of these investigations into the hydrology of the Basin of Mexico in the 16th Century is to come to an understanding of the behaviour of the lake complex in which the chinampas were located.  Figure 2 shows estimates of the volumes that in the  course of a normal year passed through the lake complex, but it gives no indication of the monthly distribution of these volumes.  The cal-  culations described in the following pages are intended to produce estimates, however approximate, of all inputs to the various lakes and of all evaporation losses. 1.  Monthly Distribution of Precipitation The mean annual, figures calculated above for precipitation in  the various watersheds or fractions of watersheds have been distributed in proportion to the measured monthly values of precipitation at a station or a group of stations in, or in the vicinity of, each of the watersheds, by use of the following formula:  PI.  PI. x P2. t -\  where PI. is the value of precipitation in month i for the area. Pl  t  is the calculated mean annual volume of precipitation for the area for which values are being calculated.  230  P2.. is the monthly value of precipitation at the station or stations in the vicinity. P2 is the total annual precipitation for the station in the t  vicinity. Values for individual watersheds were estimated with reference to particular stations, as shown in Table XII. Figure 34 shows locations of the stations.  The results of these calculations are shown in  Table XV, Rows 1, 6 and 9. 2.  Monthly Distribution of Surface Discharge a)  Xochimilco and Chalco The C.H.C.V.M. has kept records or made estimates of the larger  river basins in the Valley of Mexico.  Of these data, only those for  the Xochimilco and Chalco basins are presently available, and these are used without modification (C.H.C.V.M. 1964 111:91-94; 141-146). b)  Mexico, Texcoco, Cuautitlan, Pachuca and the Northeast In the remaining basins only figures for mean annual discharge  are presently available (C.H.C.V.M. 1964 111:6-9).  These have been  distributed between individual months on the assumption that the proportion of discharge to precipitation in any individual month is the same in the whole of the Basin as it is in the Xochimilco and Chalco watersheds.  The following formula has been used to generate figures  for the remaining basins:  0 m g m  = Pm Qt _ fm Qt Pt 100  231  TABLE XII STATIONS USED IN CALCULATING MONTHLY VALUES OF CLIMATIC VARIABLES (Figure 34 shows location of stations) Basin  Subdivision  Chalco  Mountains  San Rafael Mil pa Alta Juchitepee Amecameca Repetadora TV  Valley  Chalco Tlahuac Los Reys S. Luis Ameca  Mountains  S. Francisco Ajusco El Guarda  Valley  Moyoguarda San Gregoria  Mountains  Pena Pobre Huixcuilacan Tlalnepantla  Xochimilco  Mexico  Station  Data 1961-1975  mean mean  Source *  # # * *  Valley  Tacubaya Xoco  1961-1975 1961-1975  Mountains  Texcoco S. Rafael Chapingo Texcoco Teotihuacan  mean 1961-1975 mean mean mean  Cuautitlan  Tizayuca Tlalnepantla  mean mean  #  Pachuca  Pachuca Tizayuca  mean mean  # #  Northeast  Pachuca Teotihuacan  mean mean  # #  Texcoco  Valley Teotihuacan  *  # * # # #  C.H.C.V.,M. Boletfn Hidrologico Nos. 81-90; 128; 136; 138; 161; 162.  # Thornthwaite 1964.  232  Amecameca San R a f a e l 3- R e p e t a d o r a T.V. Juchltepec 4. San L u i s Ameca 56. Chalco San F r a n c i s c o 78. Milpa A l t a Tlahuac 910. San G r e g o r i o 11. A j u s c o 12. Moyoguarda 1. 2.  F i g u r e 34.  13- E l Guarda  14. Los Reyes 15- Xoco 16. Pena Pobre 17- Tacubaya 18. H u i x c u i l a c a n 19- T l a l n e p a n t l a 20. Texcoco 21. Chapingo 22. T e o t i h u a c a n 23. T i z a y u e a 24. Pachuca  L o c a t i o n o f m e t e o r o l o g i c a l s t a t i o n s i n the B a s i n o f Mexico ( C . H . C . V . M . 1964 V : 2 4 9 ) .  233  where 3 fi the proportion of total annual discharge in m x 10 in any particular basin and month.  Qm is  Pm is the precipitation for the same basin and month. Qt is the mean annual discharge for the same basin. Pt is the mean annual precipitation for the same basin. fm is a factor equal to the difference between precipitation in any particular month in the Xochi-Chalco basin, expressed as a % of mean annual precipitation, and the discharge in any particular month in the Xochi-Chalco basin expressed as a % of total discharge. Monthly discharge as % of total discharge and monthly precipitation as a % of total precipitation in the Xochi-Chalco basin are graphed in Figure 35.  Fm is the vertical distance between the two plots,  with positive values when precipitation is greater than discharge, and negative values otherwise. The results of these calculations are given in Table XV, Rows 13 and 15. 3.  Primary and Secondary evapotranspiration In its calculations of evapotranspiration values the C.H.C.V.M.  distinguishes three different types of evapotranspiration, namely, total evaporation (Eto), primary evaporation (Epr) and secondary evaporation (Esc), which are simply related as: Eto = Epr + Esc  Figure 35.  Precipitation and Discharge for Xochimilco and Chalco basins (C.H.C.V.M. 1964 111:91-93; 141-146; Thornthwaite 1964:383, 386).  235  Primary evaporation is defined as: The inevitable evaporation from the soil surface directly from local precipitation. It excludes all evaporation supplied by drainage, either surface or subsurface, from neighbouring zones, as well as evaporation produced by other imports of water, i.e., by pumping or diversion of rivers by dams (C.H.C.V.M. 1964 V:243) and the context makes it clear that the term "evaporation" should be understood as "evapotranspiration".  Values for primary evaporation  in the Valley of Mexico are calculated by the formula: Epr  = Z Epo  1  + J] P - 2jEpo  2  +K  where Epo.j is potential evaporation measured in evaporating pans, in mm, on days when P > Epo. P  is precipitation in mm during months when P < 40 mm.  Epo is Epo-| during months when P < 40 mm. 2  K  is an empirically derived adjustment which is equal to 50 mm where P is greater than 800 mm, and equal to zero where P is less than 800 mm (C.H.C.V.M. 1964 V:333).  Secondary evaporation is understood as the evaporation of volumes of water that are imported to a region by drainage or any other means, and it is calculated by a variety of different procedures, depending on the nature of the evaporating surface.  The agricultural supplement  is an addition to the total evaporation values, and is not included in secondary evaporation. There is no physical basis for the distinction between primary and secondary evaporation.  Water is lost up to limits imposed by the  236  condition of the evaporating surface and of the atmosphere, and the process is the same whether water became available by precipitation or some form of drainage.  The distinction is employed by the C.H.C.V.M.  as a computational convenience which yields useful results in the Valley of Mexico, and because much of the data is expressed in these terms, it is retained in the following calculations.  The annual  values for primary and secondary evaporation were calculated using the following terms: Epr  = Primary evaporation.  Esc  = Secondary evaporation.  Etot  = Total evaporation losses calculated by the C.H.C.V.M. for the Valley of Mexico today (C.H.C.V.M. 1964 V:333).  Epo.adj  = Adjusted potential evaporation, equal to 0.7 Epo used by the C.H.C.V.M. as an estimate of losses from from large areas of open water and saturated soils. A plot of these values is shown in Figure 13.  E16  = Estimated total evaporation losses for the Valley of Mexico in the 16th Century,  f  = Factor published by the C.H.C.V.M., reproduced here as Table XIII, expressing primary evaporation as a proportion of total evaporation per hydrological zone.  The values for evaporation losses were calculated as follows: a)  In the mountains (m) E16  = Epr = Etot (C.H.C.V.M. 1964 V:333)  237  b)  In Northeast, lakes and valleys; Pachuca, valleys El 6 = Etot  c)  Epr  =  f(Etot)  Esc  = Etot - Epr  In all remaining areas E16 = Epo. adj Epr =  f(Etot)  or, in Xochimilco 1, p; Texcoco 1, p, t, where (f)«Etot > P (C.H.C.V.M. 1964 V:333) Epr  = 0.95 P  Esc  = El6 - Epr.  TABLE XIII PRIMARY EVAPORATION AS A PROPORTION OF TOTAL EVAPORATION Hydrological Zone  Source:  Primary Evaporation (Epr) as T of Total Evaporation (Etot)  Xochi-Chalco  61.0  S.O. Cd. de Mexico  60.5  N.O. Cuautitlan  60.0  N. Pachuca  62.0  N.E. Apcun  55.0  Texcoco  73.0  C.H.C.V.M. 1964:145.  238  4.  Monthly Distribution of Evaporation The monthly distributions of primary evaporation are subsumed  in the calculations of monthly distributions of surplus (see 5-. below). The mean annual values estimated for secondary evaporation have been distributed in proportion to the distribution of potential evapotranspiration derived from the Thornthwaite equations, for a station or group of stations in or in the vicinity of individual basins. The Thornthwaite equations are used because they are available for more stations than are figures on the distribution of evaporation published by the C.H.C.V.M.  For stations where both Thornthwaite and C.H.C.V.M.  data are available, the distributions correspond quite closely, so the use of the Thornthwaite figures would not appear to result in too great an inaccuracy in spite of the fact that the Thornthwaite equations do not provide a very adequate description of the Basin of Mexico in other respects. Values for secondary evaporation losses from lakes were estimated with reference to potential evapotranspiration data.  Values for second-  ary evaporation losses from lakeshore plains and from valleys, and for losses by the agricultural supplementary evaporation were estimated by the same procedure except that figures for actual rather than potential evapotranspiration were used.  Values for individual basins were esti-  mated with reference to stations shown in Table XIV. The results of these calculations are shown in Table XV, rows 4, 17, 18 and 19. 5.  Monthly Distribution of Surplus In order to make our estimate of the monthly distribution of  undifferentiated runoff (R), it was necessary to estimate the monthly  239  TABLE XIV STATIONS USED IN ESTIMATION OF DISTRIBUTION OF SECONDARY EVAPORATION Basin  Station  Reference  Xochimilco and Chalco  Moyoguarda  *  Mexico  Mexico City  #  Texcoco  #  Teotihuacan Tizayuca Tlalnepantla  #  .Texcoco Teotihuacan Cuautitlan Pachuca  Pachuca  Northeast  Ti zayuca Pachuca Teotihuacan  *  # # #  C.H.C.V.M. 1964.  # Thornthwaite 1964  occurrence of surplus.  Surplus is intended as an expression of the  volumes of water that become available for surface or subsurface runoff after evaporation has occurred and after any depletions of soil moisture are replenished.  There are a variety of methods for estimat-  ing surplus and one of the most widely used is the set of Thornthwaite equations, whereby surplus is estimated from daily precipitation and evapotranspiration data.  Thornthwaite water budgets have been calcu-  lated for several stations in the Basin of Mexico, and the results are shown in Figure 36. The Thornthwaite calculations have been carried out for stations that are for the most part at low elevations in the Valley, and they do  240 mm 250  2001  150  100  50  J  F  M  A  M  J  J  A  S  O  N  J  D  F  M  A  M  J  Huixquilacan  Mexico City  1 9 ' 2 2 ' N . 99'12'W. elev. 2700m  1 9 ' 2 6 ' N . 99'08'W.  J  A  S  O  N  D  J  J  F  M  A  M  J  J  A  S  0  N  D  Texcoco  elev. 2259m.  19"31*N. 98"52'W.  mm 150  elev. 2216m  mm  100H  100'  /  :f \  50H / ^  J  — i — i — i — i — i — i — i — i — i — i F  M  A  M  J  Teotihuacan 19'42'N. 9 8 ° 5 2 ' W .  J  A  S  O  elev. 2294m  N  J D  J  F  ILi  M  A  M  \> \\\ \\  50  •7'/  N  J  J  A  S  O  N  D  J  J  20"06'N 98'44'W  F  M  A  M  elev.  1 9 ° 5 0 ' N . 98"59'W.  2447m  Precipitation  —  Potential Evapotranspiration Actual Evapotranspiration  F i g u r e 36.  J  J  A  S  O N  Tizayuca  Pachuca  Water Balance Diagrams f o r S e l e c t e d S t a t i o n s i n the B a s i n o f Mexico ( T h o r n t h w a i t e 1964).  elev. 2270m  D  J  241  not provide a very adequate description of the evaporation climate at high elevations, where most of the surplus is generated.  This is  evident in Figures 37 and 38 which show surplus and surface discharge values for watersheds in the south of the Valley.  In both examples  discharge increases in April, well before the time of year when surplus becomes available.  Because surplus is defined as including dis-  charge, this should not be possible.  The Thornthwaite equations make  the assumption that all soil moisture deficits are recharged before surplus becomes available, but even i f this assumption is relaxed and surplus is assumed to occur as soon as precipitation exceeds actual evapotranspiration, it s t i l l does not occur sufficiently early in the year to account for the increase of discharge in April and May. For the purposes of this discussion an estimate of the distribution of surplus has been made on the basis of the following assumptions: a)  that the mean annual volume of surplus in any basin or part  of a basin is equal to the mean annual volume of precipitation minus the mean annual volume of primary evaporation for that basin. b)  that the relationship between precipitation, discharge  and surplus in the whole of the Basin of Mexico is the same as the relationship between these variables in the Xochi-Chalco basin, and . in the drainage basin of Rios Magdalena and Esclava. c)  that surplus begins to occur at a time t-j when discharge  begins to occur. d)  that in the conditions of rapid runoff that occur in the  Basin, surplus ceases to occur at a time to a minimum at the end of the rainy season.  when precipitation falls  242 500  4 00  Figure 37.  Surplus and Discharge for Basin of Rios Magdalena and Esclava, Basin of Mexico (C.H.C.V.M. Boletin Hidrologico n.d.; Thornthwaite 1964:383).  Figure 38.  Surplus and Discharge for Xochimilco and Chalco basins,- Basin of Mexico (C.H.C.V.M. Boletin Hidrologico n.d.).  244  e)  that the volume of surplus and soil moisture recharge at  any time t .when t^ < t n  <t  2  is directly proportional to the volume  of precipitation at time t . n  f)  that at any time t  the proportion of surplus to soil  moisture recharge is the same as the proportion of t  - t-j to t,, - t-j.  On the basis of these assumptions some rather inelegant calculations were carried out, yielding the estimates for the distribution of surplus that are shown in Table XV, Rows 3, 8, 11 and 16. 6.  Monthly Distribution of Undifferentiated Runoff (R) There are no good, direct indicators of when these flows  became available in the lake complex.  One possibility was that hydro-  graphic separations performed on mean annual streamflow records might provide some indication of the properties of base and indirect flows. However, most of the streams are intermittent and cease discharge during the winter months.  Rios de la Compana, San Francisco, Amecameca,  Magdalena and Esclava are normally perennial, but their annual discharge profiles do not appear to be suitable for this type of analysis (Figures 39 and 40).  All of these streams show a slight decrease in the rate  of decline during the winter months, but as this corresponds with, and is almost certainly caused by a slight peak in precipitation during the winter, it provides no information on baseflows.  When this peak is  smoothed out, the recession limbs of these rivers show a smooth and steep decline, and the results of any attempts to separate the flows depend entirely upon the selection of points for analysis and not at all on the objective characteristics of the streamflow profiles.  Figure 39.  Estimated Mean Monthly Discharge of Rios de la Compana and San Francisco, Valley of Mexico (C.H.C.V.M. 1964 111:141, 142)  Figure 40.  Estimated Mean Monthly Discharge of Rio Amecameca, Chalco Basin (C.H.C.V.M. 1964 111:141)  247  A second possibility was that data on the yield of springs in the Xochimilco and Chalco regions would provide some indication of subsurface flows.  Some data are available for the years of 1903 and  1904, before mechanical pumps were installed (Marroquin y Rivera 1914:151-154).  However, these are fragmentary, and the fact that  different government agencies making measurements of the same springs at the same time recorded values that differed by as much as 100% suggests that they may also be unreliable.  The measurements made by  one of these agencies are shown in Figure 41, but because they are so fragmentary it is difficult to make any evaluation of the data. A priori, it would seem that the yield of springs should in some way be tied to variations in precipitation, but the available data show very little evidence of this.  It is true that in the summer of 1903  most of the larger springs show peak yields between one and three months after peaks in precipitation, but any hypothesis of a consistent lag is frustrated by data on the winter of 1903-04 which shows an increase in the output of Quetzalapa and some of the other springs at a time of year when precipitation values are low.  This group of  springs is not even a representative sample because there were many smaller and submerged springs about whose yields we have no information, and so it is not even possible to make an evaluation of the extent to which'the springs had the effect of smoothing inputs to the lake complex by counter-variation. In the absence of any useful indicators of the seasonal distribution of the flow of undifferentiated runoff into the lakeshore plains and the lakes, the following procedure was adopted.  First, undifferen-  tiated runoff (R) was divided into two parts, infiltrated flow (I) and  248  Figure 41.  Observed Yield of Springs in the Xochimilco and and Chalco basins (Marroquin y Rivera 1914:151-54) 1903-04.  249  and indirect flow (J), such that PI  = I + J.  The infiltrated flow (I) is taken directly from estimates made by the C.H.C.V.M. of the annual rates at which deep aquifers, depleted by pumping, are recharged by the infiltration of precipitation.  It is  assumed that the Valley of Mexico in the 16th Century was an entirely closed drainage basin, from which it follows that volumes infiltrated at high elevations sooner or later made their way to the aquifers underlying the Valley floor, and thence to the lake complex where they became available for lake discharge, storage or evaporation. The deep aquifer recharge volumes are here used as an approximation of the volumes that became a part of this deep subsurface flow.  The indirect  flow (J) is intended to represent volumes draining in channels too small to merit metering, or that were lost from streams by seepage before they were metered.  This was calculated as the difference between  undifferentiated runoff (R) and infiltrated flows (I). The monthly distributions of these flows were then estimated on the basis of the following assumptions: a)  that the flows in any one year begin at the times estimated  for the start of surplus. b)  that indirect flows peak one month after the peak of surplus  and that the flows of infiltrated water peak two months after the peak of surplus. c)  that the rise side of the curve of indirect and infiltrated  flows is approximated by an S-shaped curve, drawn by eye, and shown in Figure 42.  250  I  i  i  A Figure 42.  i  i  i _  B Graph used for estimates of monthly distributions of infiltrated flows (I) and indirect flows (J). Note: A and B indicate the start and peak of surplus, respectively.  150  Figure 43.  Estimated Monthly Distribution of surplus and runoff, Chalco Basin. S - surplus; Q - surface discharge; I - infiltrated flow; J - indirect flow.  251  d)  that the decay rate of the regression limb is approximated  by a constant K = 0.65 for indirect flow, and K = 0.75 for infiltrated flow, which figures are based on the characteristically steep recession limbs of the various perennial rivers in the southern parts of the Basin. Values for particular basins were estimated by adjusting the time axis of the graph, Figure 42, so that it corresponded with the period during which surplus was estimated to occur in that basin, and the total volumes of infiltrated and indirect flows were proportioned against values taken from the vertical axis of the graph. Figure 43 shows the values that resulted for the Chalco basin, together with estimated data for discharge and surplus.  The values calculated  for the monthly distributions of infiltrated and indirect flows were then added, to provide an estimate of the monthly inputs of water by subsurface flows into the lake complex.  These values are shown in  Table XV, Rows 12 and 14. 7.  Lake Tlahuac and Lake Chalco Some modern data are available for the stage of Lake Tlahuac,  which is a small residue of the antique Lake Chalco.  These data  exist only for the short period from 1950 to 1955, before which time there are no records of stage and after which time the residual lake was supplied with seasonal additions of treated water from Mexico City, with the result that fluctuations in stage were considerably reduced.  The observations were made well after pumping of groundwater  on a massive scale had commenced, and so the extent to which.the regime of the modern Lake Tlahuac resembles that of antique Lake Chalco is  TABLE XV ESTIMATED 16TH CENTURY MEAN MONTHLY WATER BALANCE, BASIN OF MEXICO  Abbreviation  Variable  Row Numbers  Derivation or Source in Appendix Part III  Physiographic Regions  P  Precipitation  1,6,9  Section 1  m  mountains  Epr  Primary Evaporation  2,7,10  Epr = P - S  v  valleys  S  Surplus  3,8,11,16  Section 5  p  plains  A  Agricultural Supplement  4,17  Section 4  1  lakes  Om  Output from mountains  5  Om = Sm - Am  R  Undifferentiated Runoff  12,14  Section 6  Q  Surface Discharge  13,15  Section 2  Esc  Secondary Evaporation  18,19  Section 4  I  Total Input to Lake Complex  20  Sum of rows 12-19  TABLE XV (continued) Row No. Variable 1 2 3 4 5 6 7 8 9 10 11  Jan.  Feb.  Mar.  Apr.  May  June  July  Aug.  Sept.  Oct.  Nov.  Dec.  Total  +P.m -Ept.m =S.m -A.m  14.5 14.5  7.2 7.2  13.6 13.6  38.0 33.0  164.5 89.5  168.1 66.1  160.0 39.2  153.7 14.8  67.8 48.8  19.0 4.1  7.2 7.2  904 336  0.0 4.1  0.0  0.0 3.3  5.0 6.8  90.4 63.9 26.5  102.0 10.4  120.8  138.9  84.9  14.9  0.0  568  10.3  75.0 11.2  10.1  3.3  -1.8  16.2  63.8  91.6  1.6 1.6  -4.6 0.8 0.8  1.4  8.0 6.5  14.4 10.2  17.8 10.9  13.7 5.9  6.3  0.0 2.1  0.0 1.0  1.4 0.0 1.8  3.8 3.5  110.7 16.8  5.5 9.4  88  -4.1  7.7 77.2  5.0-  =0.m +P.p  9.0 129.9  0.3 4.9  1.5 10.4  4.2  6.9 23.2  7.8 17.9  4.0 8.2  0.9 0.7 0.2  2.1  1.8 0.0  4.6 0.3  8.6 1.8  8.6 9.3  -0.3 8.5  0.8 0.3  21.3 0.5  15.2  13.3 2.6  55.6 7.7 5.6  59.0 3.9  58.3 1.8  5.2  3.8  0.7 2.5 0.0 0.0 2.8 4.6 6.4 41.5  -Epr.p =S.p +P.1 -Epr.l =S.l  4.6  13  +R.m +Q.m  39.9 0.5  1.0 0.0 29.7 0.3  14 15  +R.p +Q.p  1.7  1.1  0.7  16 17 18  +S.1 -A.p  0.0 0.0 2.3 3.8  0.0 0.0 2.5 4.2  0.0 0.0 1.9  19  -Esc.l  6.8  20  = 1.1  6.0 30.0  3.1 9.5  17.6  8.0  12  -Esc.p  0.0  0.9 0.5  0.7  0.0 0.3 3.8 6.2  0.0 1.8  11.2 -4.3  13.0 -9.8  5.7 9.5  18.8 13.8 5.0 18.1 6.3 1.8 0.0 5.0 6.2 10.3 12.3 2.4  8.7 8.1  2.3  15.0 8.2  21.9 12.3 9.6  30.1 8.1  46.4 7.7  4.1  5.3  0.0 8.2  0.0 9.3  0.0 8.5  5.8  0.0 9.6 5.6  5.0  4.3  0.0 0.3 3.1  9.6 11.7 23.4  9.3 11.3 42.8  8.3 10.0 54.9  7.0  5.1  8.8 56.5  7.0 49.0  1) Estimated Water Balance of the Chalco Basin  1.1  -5.0 0.5 0.5  480 86 53  0.0 0.7  33 112  0.7 0.0 52.1  69 43 439 41 33 0 43 49 81 114 312  TABLE XV (continued) Row No. Variable 1 2 3 4  +P.m -Ept.m =S.m -A.m  5 6  =0.m +P.p  7 8 9  -Epr.p =S.p +P.1  10  -Epr.l  11 12  =S.l +R.m +Q.m +R.p  13 14 15  +Q.p  16 17  +S.1  18  -Esc.p  19 20  -Esc.l  -A.p  = 1.1  Jan.  Feb.  Mar.  Apr.  May  June  July  Aug.  Sept.  Oct.  Nov.  Dec.  Total  7.1 7.1  2.6 2.6  5.2 5.2  0.0 2.1  0.0 2.3  0.0 2.9  12.2 10.7 1.5 3.4  28.2 19.9 8.3  63.4 33.7 29.7  3.0 2.8 0.2  371 133 238  3.2  41  -2.3  54.9 2.7  0.6  1.0 0.9  0.1 0.1  0.1 0.1  0.0  0.0 2.6  1.3 0.1 6.5  2.5 2.2  197 15  0.2  43.9 3.4 3.2 0.2  -2.3  0.1 0.1 0.0  -1.9 0.6  24.0  0.3  -2.9 0.2  2.6 -0.4  2.5  -2.1  66.0 4.4 61.6 3.8 57.8  4.5 2.3 2.2  4.7  75.3 16.3 59.0 4.1  26.7 -0.5 27.2  5.0 3.3  76.8 28.5 48.3 4.4  0.3  0.1  16.0 15.1 0.9  13.0 12.1 0.9  12.0  0.0 0.6  71  11.0 1.0  4.7 4.3 0.4  0.0 0.6 0.6 0.0  0.6 0.0  67 4  19.4 1.1  23.3  24.9  24.8  22.8  0.4 0.2  0.0 0.2  0.0  0.1  0.9 0.1  192 5  0.1  1  0.0  0.0  0.0  0.0  0.0  0  0.9 1.5  1.0 1.4 1.7  0.4 1.1 1.5  0.0 0.8  0.0 0.7 1.0  2.4 19.8  2.1 21.2  1.1 1.6 21.5  4 13 17  1.5 19.7  27 145  0.3 0.0 1.4 1.4 0.0 18.2 0.0 0.1 0.0 0.0 0.6 018 1.4 15.5  1.4  25.0 2.6 2.5 0.1 12.1  0.6  1.0  0.6 0.0  1.0 0.0 10.2  2.6 0.0  6.3 0.2  11.5 0.6  7.5  6.3  12.5  0.0 0.0  0.0 0.0 0.0  0.5  8.1 0.9  0.0  0.0  0.0 0.2  0.0 0.6  1.5  1.7 2.1 2.9  0.1 0.0 0.9 1.5  14.0 0.0 0.1 0.0 0.0 0.7  0.0 0.0 0.5  0.9 1.6  0.6 2.2  10.9  6.9  2)  0.0 1.0 1.3 2.7 2.5  2.0 3.1 0.4  2.9  1.2  2.0 2.8 8.4  2.5 0.2  2.0 2.7 15.3  Estimated Water Balance of the Xochimilco Basin  14 1  TABLE XV (continued) Row No. Variable 1 2 3  +P.m -Ept.m =S.m  4 5 6  -A.m =0.m +P.p  7 8  -Epr.p  9 10 11 12 13 14 15 16 17 18 19 20  =S.p +P.1 -Epr.l =S.l  Jan.  Feb.  Mar.  Apr.  May  June  July  Aug.  Sept.  Oct.  Nov.  Dec.  Total  8.2 8.2  6.2 6.2  47.2 34.5 12.5  121.7 69.4 52.3  146.4 62.0 84.4  129.3 18.5  5.5 5.5  110.8  8.7 43.6  8.1 76.3  11.9 4.8  0.0  9.3 3.2 5.9  89.3 7.7 81.6  48.6 3.8 44.8  15.7 3.8  0.0 4.3 -4.3 0.5  19.1 17.1 2.0 6.2 -4.2  125.2 35.9  0.0 3.9 -3.9  10.9 10.9 0.0 5.4  684 276 408 76 332  17.9 10.5  18.6 8.4 10.2  19.1 6.0 13.1  12.4  22.3 13.1 9.2  23.2 10.5  23.7 7.5 16.2  19.1 3.7 15.4  7.3  12.7 17.9 2.7  19.5 26.5  23.3  2.7 12.7  4.0 16.2  8.1 10.7  7.9  1.3 1.3 0 1.5 1.5  -5.4 1.2  2.9 2.5  0.5  1.2  0 0.6 0.6  0 1.5  0  0  3.7 3.3 0.4  1.5  0.4  4.4 1.5 7.3 5.4 1.9 5.2  0 15.9 3.3 1.9  11.7 1.8 1.4  8,3  5.9  1.2 0.9  0.4  0.3  0.2  1.6 0.7 0.2  0.7 0.4  -A.p  0.0 2.5  -Esc.p -Esc.l = 1.1  3.3 7.7 8.0  0.0 2.7 3.6 8.8 0.1  0.0 3.5 4.7  0.4 4.2 5.5  1.9 6.6 8.8  12.1 -9.7  14.3 -15.2  16.6 -21.2  +R.m +Q.m +R.p +Q.p +S.1  3)  2.6  7.4  7.0 1.3 1.1 9.2 8.5 11.3 15.8 -9.7  12.7  15.0 14.9  3.6  10.5 14.5 36.9  7.0 103.8 15.2 2.8  47.0 4.0 7,1 15.4 7.2 9.5 12.9 67.2  Estimated Water Balance of the Mexico Basin  6.0 38.8 5.3 0.5 4.8 6.7  7.1 0.6 0.4 0.2 0.8  0.7 6.0  0.6 0.2  24.6 29.5  24.2  3.9 4.5 6.0 5.7 7.6 11.2 44.0  9.0 3.4 1.4 0.2 3.9 5.2 8.9 20.2  4.6 -4.6 0.5 0.5 0 0.6 0.6 0 21.4 4.6 2.5 0.7 0.0 3.2 4.3 8.2 13.5  89 39 50 111 49 62 180 152 27 23 62 64 85 146 149  TABLE XV (continued) Row No. Variable 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20  +P.m -Ept.m =S.m -A.m =0.m +P.p -Epr.p =S.p +P.1 -Epr.l =S.l +R.m +Q.m +R.p +Q.p +S.1 -A.p -Esc.p -Esc.l = 1.1  Jan.  Feb.  Mar.  Apr.  May  June  July  Aug.  Sept.  Oct.  Nov.  Dec.  Total  7.2  5.2 5.2  10.4 10.4 0 5.1  19.8 18.0 1.8  47.7 37.1 10.6  103.5 66.8  119.1 62.8  115.0 47  87.7 26.4  42.0 10.2  10.9 5.5  5.7 5.4  575 302  36.7  57.1  61.3  31.8  9.4 1.2  9.6  9.1  7.7  0.3 4.1  273 79  18.9 18.4 0.5  48.0 40.8 38.8  6.6 25.2  5.4 4.8  27.1 34.2  68.0 8.8 59.2  14.7  0.6 4.5  31.9 2.5  13.6 1.1  4.3 0.2  -3.8 2.2 2.2  194 205 195  52.5  20.7 19.1 1.6 21.2  5.8 5.4 0.4 20.0  0.0 2.7 2.7  10 305  3.7 1.1  1.6 1.2  0.0 1.6  0.0 0.4  14.2  9.2  7.2.  15 155  17.6 37.5  11.4  817 27.3  191 487  -24.6  -614  7.2 0 3.7 -3.7  0 4.1  -5.1  2.3 2.3  -4.1 2.1 2.1  6.0 -4.2  4.8  8.2  4.8  8.1  0.0 3.4 3.4 0 12.4 0.3  0.0 1.8 1.8 0 9.3 0.3  0.1 11.3 11.2  1.0  0.9  0.0 4.9 4.9 0 6.9 0.5 0.8  0.0 0.0  0.0 0.0  5.7 6.9 25.8  6.0 7.4 29.2  -24.5  -32.1  0.0 0.0 8.4  0.1 5.2 0.7 0.6 0.0 0.1  26.8 26.2 0.2  1.3 54.9 52.9  2.0 61.9 58.9 3.0 12.3 8.4 0.4  4.7 2.8 0.5  2.0 6.9 6.8 0.4  0.0 0.6  0.0 2.0  0.0  16.5  20.4 25.1  19.9 24.7 50.2 -70.7  10.4 40.4  10.0 12.4 47.7  20.4 55.5  -51.0  -63.5  -83.8  4)  32.9  52.6 -82.0  3.0  37.9 35.6 2.3 58.3 54.8 3.5 17.7 7.7 0.8 0.0 3.5 19.5 24.0 48.2  53.6 34.4  48.7 3.8 20.6 6.5 1.0 0.0 3.8 18.0 22.0  42.9 -62.0 -51.0  Estimated Water Balance of the Texcoco Basin  -41.7  29.7 -21.1  0 16.7 0.7 1.2 0.0 0.0.  290 15 154 40 10 0  TABLE XV (continued) Row No. Variable  Jan.  Feb.  Mar.  Apr.  May  June  July  12.0 11.6 0.4  15.6 14.2  16.5 14.2  1.4  2.3  40.2 29.4 10.8  104.1 63.3 40.8  101.0 53.6 47.4  1 2  +P.m -Ept.m  7.5 7.5  3  =S.m  0.0  4 5  -A.m =0.m +P.p  -  -  -  -  -  0.0 2.1  0.4 3.4  1.4 4.5  2.3 4.7  10.8  40.8  11.4  2.1  3.4 0  4.4 0.1  4.5 0.2  10.4 1.0  29.7 26.4 3.5  4.6 4.5  6.0 5.8 0.2  6.4  15.6  6.1  14.4 1.2  6 7 8 9  -Epr.p =S.p  0  +P.T  2.9  10 11 12  -Epr.l =S.l +R.m  2.9  13 14 15  +Q.m +R.p +Q.p  16 17 18  +S.1 -A.p  19 20  0 10.4 1.1  0.1  6.7 0.1  5.6 0.1 0.1 5.8  -Esc.p  0.0 4.7 5.2  -Esc.l = 1.1  15.0 -6.6  17.0 -11.8  7.1  9.1 2.5  6.4  2.6 4.2 0.1 0.2 7.8 8.6 23.5 -25.7 5)  Including imports from Pachuca  0.3 6.1 1.5  6.0 7.4  -  -  Aug. 85.9 39.5 46.4  -  Sept.  Oct.  Nov.  Dec.  Total  108.8 40.4 68.4  39.9 15.4 24.5  19.5 8.9 10.6  7.0 7.0  558 323  0.0  235  -  -  -  24.5 11.4 9.2 2.2  10.6 5.6  0.0 2.0  235 159  4.7  2.0 0  137 22  15.4 12.6  7.6 6.4 1.2  2.7  216  2.7 0  11.7  11.3 2.6  187 29 111  -  -  47.4 28.7  46.4 24.5  68.4 31.0  24.6 4.1  20.5  25.0  4.0  40.2  39.2  35.6  33.8  33.3 27.9  6.0 42.1  4.6 6.7  5.4 9.2  5.4 10.4  34.3 7.8 11.3  21.9 3.5 1.1  22.6  18.1  25.4  2.8 11.7 11.2  5.0 1.2  7.3 1.2  7.7 0.6  5.4  6.5 0.8 5.4  7.8  2.8  7.1 7.7 0.3 1.2  13.7 15.1  11.0 12.2  9.8 10.8  7.1 7.9  5.9 6.5  115 127  28.0 -15.6  24.9 4.9  21.8 -8.4  17.3 -4.3  15.8 -6.9  283 -187  3.3 0.1 0.3  3.1 0.4 1.2  8.0 8.9  11.5  4.6 15.1  12.7  16.6  14.6 16.1  27.7 -33.3  32.3 -38.4  30.6 -24.5  29.1 -16.4  Estimated Water Balance of the Cuautitlan Basin  0.9  7.4 0.0 0.0  124 68* 6 29  TABLE XV (continued) Row No. Variable  Jan.  Feb.  Mar.  Apr.  May  June  July  Aug.  Sept.  Oct.  Nov.  Dec.  Total  10.1 10.1 0.0  21.9 21.9 0.0  28.8 27.5 1.3  31.9 28.8 3.1  67.7 55.0 12.7  142.0 103.0 39.0  118.3 77.2 41.1 _  107.4 63.0 44.4 _  136.9 65.7 33.5 69.4 31.9 17.5 67.5 33.8 16.0 _ _ _ _  14.1 11.1 3.0 _  778 516 262 _  1 2 3 4  +P.m -Epr.m. =S.m -A.m  5  =Q.m  0.0  0.0  1.3  3.1  12.7  39.0  41.1  44.4  67.5  33.8  16.0  3.0  262  6 7 8 9 10 11  +P.p -Epr.p =S.p +P.1 -Epr.l =S.l  4.2 4.2 0.0 0.1 0.1 0.0  9.1 9.1 0.0 0.1 0.1 0.0  12.1 12.1 0.0 0.1 0.1 0.0  13.4 13.1 0.3 0.2 0.2 0.0  28.4 27.9 0.5 0.3 0.3 0.0  59.4 44.4 3.9 0.7 0.7 0.0  49.6 45.5 4.1 0.6 0.6 0.0  45.0 40.6 4.4 0.6 0.6 0.0  57.4 50.7 6.7 0.7 0.7 0.0  27.5 24.1 3.4 0.3 0.3 0.0  14.0 12.4 1.6 0.2 0.2 0.0  5.9 5.6 0.3 0.1 0.1 0.0  326 301 25 4 . 4 0  12 13 14 15  +R.m +Q.m +R.p +Q.p  23.3 0.1 2.2 0.0  18.0 0.3 1.8 0.1  12.9 0.4 1.4 0.1  10.1 0.3 1.0 0.0  10.4 0.9 0.8 0.2  13.0 2.1 1.0 0.5  20.1 1.8 1.3 0.5  25.9 1.6 2.0 0.4  28.8 2.2 2.5 0.6  30.1 1.2 2.7 0.3  29.6 0.8 2.7 0.2  27.8 0.3 2.6 0.1  250 12 22 3  16 17 18  +S.1 -A.p -Esc.p  0.0 0.0 5.0  0.0 0.0 7.6  0.0 0.0 10.4  0.0 0.0 11.5  0.0 0.0 18.7  0.0 0.0 27.7  0.0 0.0 24.8  0.0 0.0 23.1  0.0 0.0 22.7  0.0 0.0 17.0  0.0 0.0 10.0  0.0 0.0 6.5  0 0 185  19 20  -Esc.l =1.1  0.3 20.3  0.4 12.2  0.5 3.9  0.6 -0.7  0.7 -7.1  0.6 -11.7  0.6 -1.7  0.6 6.2  0.5 10.9  0.5 16.8  0.4 22.9  0.3 24.0  6 96  6)  Estimated Water Balance of the Pachuca Basin  TABLE XV (continued) Row No. Variable  Jan.  Feb.  Mar.  Apr.  May  June  July  Aug.  Sept.  Oct.  Nov.  Dec.  Total  10.4 10.4  18.2  38.2 32.2  80.7 57.3  143.2  143.7  135.4  148.0  75.5  31.7  13.5  18.2  29.5 27.6  37.0  15.5  9.3  9.9  0  1.9  6.0  23.4  66.9 76.8  49.0  0  83.7 59.5  868 417  86.4  111.0  60.0  22.4  3.6  451  =0.m +P.p  4.4  12.6  61.1  61.2  57.7  13.5  5.7  370  57.8  58.5  54.7  0 0.6 0.6  3.3 4.9 4.8  2.7  0.8 1.1 1.0  0.5  n  =S.l  0  0  0.1  0.1  3.9 5.1 4.9 0.2  353 17  5.0 4.8 0.2  3.0 4.7 4.5 0.2  5.6 0.1 0.5  10  0 0.4 0.4  30.1 2.1  12.7  =S.p +P.1 -Epr.1  33.6 0.8 2.8 2.7  63.1 59.2  32.2  4.4  16.3 16.1 0.2  34.4  -Epr.p  7.8 7.8  12  +R.m +Q.m +R.p +Q.p  38.6. 0.6  29.4 0.8  17.1  21.7  33.5  1.5  16.7 2.2  3.8  2.5  0.0 0.3  0.0 0.5  0.0 0.5  0.0 0.0 9.8 1.4  0.0 0.0 14.2  0.0 2.5 0.2  0.0 3.1 0.2  0.0 1.8 0.1  0.0 1.0 0.1  0.0 0.4 0.0  0 17 1  2.0  0.0 0.0 17.3 2.4  0.0 1.4 0.1  5.9 0.0 2.8 0.2  38.1 5.8  0 46.1 0.2  0.0 0.1 0.0 0.0 7.8  6.3 0.0 2.6  43.1 5.9  0.1 49.4 0.5  19.3  6.9  -0.3  1 2  +P.m -Epr.m  3 4 5  =S.m  6 7 8 9  13 14 15 16 17 18 19 20  -A.m  +S.1 -A.p -Esc.p -Esc.l = 1.1  1.3 30.2  7)  12.6 0 1.0 1.0 0 21.1  1.3 1.3 0  0.0 29.5 2.7 -9.8  2.6 2.5 0.1 50.2  30 29 1 415 36  0.1 0.0 43.1 2.6  0.0 39.3 2.5  0.0 37.3 2.4  0.0 37.6 2.1  0.0 28.3 1.8  0.0 15.3 1.5  0.0 9.5 1.3  0 289 24  -15.0  0.6  12.0  17.5  24.5  34.2  35.9  156  Estimated Water Balance of the Northeastern Basins  260  unclear.  Figure 44 shows the mean monthly levels of Lake Tlahuac as  observed and the fluctuations of inputs estimated for the antique Lake Chalco, calculated by summing the monthly values of streamflow (Q), precipitation (P), and undifferentiated runoff (R), minus the volumes estimated for primary evaporation.  The two curves correspond  closely but, because of the uncertain resemblance of modern to preHispanic conditions and because the values shown for Lake Chalco do not include the effects of secondary evaporation, the correspondence may not entirely validate the procedures employed in this reconstruction.  261  Figure 44.. Observed stage of Lake Tlahuac and estimated inputs of Lake Chalco (Boletin Hidrologico, n.d.).  262  Part IV  Estimates of Variation in Lake Stage  1. Assumptions In order to produce estimates of the variations in stage of the antique lakes, the following procedures were employed: a)  The areas of the lakes were derived from maps and are assumed  to have been as shown in Table XVI, column 2. b)  The relative elevations of mean lake surfaces are assumed to  have been as shown in Table XVI, column 3. c)  The average elevation of all five lakes are assumed to have  been 2240 m above mean sea level, which coupled with the two preceding assumptions yields the values for individual lakes shown in Table XVI, column 4. d)  The shape of each lake depression was.approximated as a spheri-  cal segment with a mean volume equal to the mean area of the lake multiplied by one meter.  A graph was then constructed for each sphere  lake that related lake stage to volumes in storage, and stage values for calculated volumes in storage were then read from the graph. 2.  Estimated.Variations of Lake Stage with Normal Precipitation The volumes of water entering a lake complex in any month i  were estimated using the figures from the Table XIV of this Appendix, and were taken as:  263  TABLE XVI ASSUMED AREAS AND ELEVATIONS OF THE ANTIQUE LAKES (3)  (2)  (1)  (4)  mean area  relative mean  mean elevation  Lake  (km )  elevations (m)  above m.s.1 (m)  Texcoco  483.7  0.0  2238.1  Mexico  178.8  +0.5  2238.6  99.7  +2.0  2240.1  Chalco  140.9  +2.5  2240.6  Cuautitlan  317.2  +5.0  2243.1  2  XochimiIco  I.  = R. + Q. + S.l. - A.p - Esc.p - Esc.l. +Q.1. i  i  where, in each month i I.  = total inputs  R.j  = undifferentiated runoff  Q.  = streamflow  S.1.  = lake surplus  A.p.  = agricultural evaporation supplement from plains, valleys and the Teotihuacan Valley  Esc.p. 1  = secondary evaporation from plains, valleys and the Teotihuacan Valley  Esc.l^  = secondary evaporation from lakes  Q.l.  = discharge from lakes upstream  264  For models of uncontrolled conditions, the discharge from a lake in any month i was estimated as:  Q.1.  = 0.5 Q.1. + 0.5 I.  where Q1 . y is equal to the annual discharge from the lake, per text Figure 2, divided by 12.  The volume remaining after discharge was  then located on the graphs described in 1. d) above to give the corresponding value of stage. 3.  Estimates of Variation of Lake Stage.with Precipitation at Exceptionally High or Low Values Only one variable was modified in order to produce these  estimates, namely precipitation itself. to be as they were in normal years.  All other values were assumed  There is some justification for  this with regard to evaporation rates.  Lines of equal potential evapor-  ation in 1957 and 1958, which were exceptionally dry and wet years, respectively, do not vary consistently; in some of the basins losses were higher in the wet year than in the dry and in others, lower (C.H.C.V.M. 1964 V:257, 259) In order to produce the estimates shown in text Figure 19, 28 and 29, the annual value for precipitation for the whole of the Basin was increased or decreased by 25% and distributed between basins and physiographic regions in the same proportions as in normal years. For each region, the annual precipitation value was reduced by the normal value for primary evaporation, to yield an annual value for surplus.  In the lakes, the value for surplus was pro-rated against  265  the normal monthly distributions.  In the remaining regions values for  surplus were distributed against normal values for discharge (Q) and runoff (R).  Monthly values for discharge were obtained by proportioning  the volumes against normal values, and monthly values for runoff were calculated from Figure 43.  The procedure was carried out individually  for each year of the sequence and overlapping values in consecutive years were summed. The resulting figures were then reduced by normal monthly values for secondary evaporation, and the agricultural supplement, to provide estimates of the total inputs to the lake complex in each successive year. For text Figure 19 (page 50), fluctuations of Lake Chalco over a 10-year period, monthly lake discharge and stage were estimated in the manner described for uncontrolled normal variations in Part IV, 2, above.  For text Figures 28 and 29 (pages 85 and 89, respectively),  stage was estimated in the same way but inputs of water were either added to lake storage or discharged, depending on the state.of the system and the decisions imputed to the managers.  266  BIBLIOGRAPHY  Anderson, Edgar 1971 Plants, Man and Life. Press. Apenes, Ola 1943 "The  Tlateles  Berkeley:  of Lake Texcoco."  University of California  American Antiquity 9:29-32.  Armillas, Pedro 1971 "Gardens on Swamps." Science 174:653-61. Boulding, Kenneth 1956 "General Systems Theory - The Skeleton of Science." Management Science 2:197-208. Brady, N.C. 1974 The Nature and Properties of Soils.  New York: Macmillan.  Bribiesca Castrejon, Jose Luis 1960 "Hidrologia Historica del Valle de Mexico." Hidraulica en el Mexico 14(3):43-61. Brown, L.N. et al. 1976 Making Aquatic Weeds Useful. Academy of Sciences.  Ingenieria  Washington, D . C : National  Bursche, Eva M. 1971  A Handbook of Water Plants.  London: Fredrich Warne & Co.  Cainek, Edward E. 1972 "Settlement Pattern and Chinampa Agriculture at Tenochtitlan." American Antiquity 370 ): 104-115. Carrasco, Pedro 1971 "Social Organization of Ancient Mexico." In Wauchaupe (.Ed.) Handbook of Middle American Indians 10:349 ff. Chaney, R.L. 1975 "Sewage Sludge: benefits and problems." Garden Record 31:1.  Brooklyn Botanic  Charlton, Thomas 1970 "Contemporary Agriculture in the Teotihuacan Valley." Occasional Papers in Anthropology No. 3, Department of Anthropology, Penn. State University:253-385.  267  Coe, Michael 1964 "The Chinampas of Mexico."  Scientific American 211:90-98.  Comision Hidrologica de la Cuenca del Valle de Mexico 1961 Informe sobre la geologia de la Cuenca del Valle de Mexico y zonas colindantes. Mexico: Secretaria de Recursos Hidraulicos. Comision Hidrologica de la Cuenca del Valle de Mexico 1964 Hidrologia de la Cuenca del Valle de Mexico. Secretaria de Recursos Hidraulicos.  Mexico:  Comision Hidrologica de la Cuenca del Valle de Mexico 1971 Estudio de Lirio Aquatico. Mexico: Secretaria de Recursos Hidraulicos. Comision Hidrologica de la Cuenca del Valle de Mexico 1965 Datos Diaros del Valle de Mexico. Boletfn Hidrol6gico:17. Curran, Charles H. 1937 "Insect Lore of the Aztecs Revealing Early Acquaintance with Many of our Pests." Natural History 39(3):196-203. Cutler, H.C. & T.W. Whitaker 1961 "History and Distribution of the Cultivated Cucurbits in the Americas." American Antiquity 26:469-485. Davies, Nigel 1973 The Aztecs, A History.  London: Macmillan.  Deevey, Edward S., Jr. 1957 "Limnological Studies.in Middle America with a Chapter on Aztec Limnology." Transactions of the Connecticut Academy of Arts and Sciences 39:213-328. Deutsch, Karl W. 1966 The Nerves of Government: Models of Political Communication and Control. New York: The Free Press. Diaz, Bernal 1963  The Conquest of New Spain.  Harmondsworth:  Penguin.  Dressier, R.I. 1953 "The Precolumbian Cultivated Plants of Mexico." Harvard University Botanical Museum Leaflets 16(6). Epstein, Emmanuel 1973 "Roots." Scientific American May:48-58. Erichsen-Brown, Charlotte 1979 Use of Plants for the Past 500 Years. Creeks Press.  Ontario:  Breezy  268  Flannery, Kent 1968 "Archaeological Systems Theory in Early Mesoamerica." In B.J. Meggers (Ed.) Anthropological Archaeology in the Americas. Washington: Anthropological Society of Washington. Fluck, Richard C. and C. Direlle Baird 1980 Agricultural Energetics. Connecticut:  Avi Publishing.  Food and Agriculture Organization of the United Nations 1976 The State of Food and Agriculture. Rome. Fox, D.J. 1965 "Man Water Relations in Metropolitan Mexico." Geographical Review 55(4):523-545. Garrett, W.E. and D. Jeffery 1974 "Burma's Leg Rowers and Floating Farms." National Geographic 145(6):826-845. Geertz, Clifford 1971 Agricultural Involution: in Indonesia. Berkeley:  the process of ecological change University of California.  Gibson, Charles 1964 The Aztecs Under Spanish Rule. Press.  Stanford:  Stanford University  Gliessman, S. et al. 1978 "The Ecological Basis for the Application of Traditional Agricultural Technology in the Management of Tropical Agroecosystems." Presented at 2nd International Congress of Ecology: Israel. Gliessman, S. and A. Moises Amador 1979 "Ecological Aspects of Production in Traditional Agroecosystems in the Humid Lowland Tropics of Mexico." Prepared for the V International Symposium on Tropical Ecology: Kuala Lumpur. Harris, Marvin 1977 Cannibals and Kings: The Origins of Culture. Random House.  New York:  Harris, Marvin and Marshall Sahlins 1979 "Cannibals and Kings: An Exchange." (Letters by Harris and Sahlin). New York Review of Books, June 28:25. Harrison, P.D. and B.L. Turner II 1978 Pre-Hispanic Maya Agriculture. New Mexico Press.  Albuquerque: University of  Janzen, Daniel H. 1973 "Tropical Agroecosystems." Science 182:1212-19.  269  Jauregui, Ernesto 1973 "The Urban Climate of Mexico City."  Erdkunde 27-28:298-307.  Kinne, Ivan and Thomas A. McClure 1977 "Energy in the Food System." In Daniel Lapedes (Ed.) The McGraw-Hill Encyclopedia of Food, Agriculture and Nutrition. New York: McGraw-Hill. Kormandy, E.J. 1969 Concepts in Ecology.  Englewood Cliffs, N.J.: Prentice-Hall.  de Lameiras, Brigette y Armando Pereyra 1974 "Terminologia Agrohidraulica - Prehispanica Nahua." Secretaria de Educacfon Publica - Instituto Nacional de Antropologfa y Historfa, Col. Cientifica No. 13, Mexico. Leicht, Hugo 1937 "Chinampas y Almacigas Flotantes." Biologia 12. Mexico: U.N.A.M. McDowell, Bart 1980 "The Aztecs."  Anales de Instituto de  National Geographic 158(6):704-751.  Madrid Mendizabal, Fernando 1946 Breve resena historica de los principales problemas hidraulicos y sus derivados, que han tenido que resolverse para hacer habitable la Ciudad de Mexico. Mexico. Marroquin y Rivera, M. 1914 Memoria Descriptiva de las obras de prevision de aguas potables para la Ciudad de Mexico. Mexico. Matheney, Ray T. 1978 "Northern Maya Lowland Water Control Systems." In P.D. Harrison and B.L. Turner II (Eds.) Prehispanic Maya Agriculture. Albuquerque: University of New Mexico Press. Miller, James G. 1965a "Living Systems: 193-237.  Basic Concepts."  Behavioural Science 10:  1965b "Living Systems: Structure and Process." Behavioural Science 10:337-379. Moreno, Manuel M. 1931 La organizacion politca y social de los Aztecas. Munzon, Arturo 1977 "El Calpulli." a Los Aztecas.  Mexico.  In Miguel Leon-Portilla (Ed.) De Teotihuacan Mexico: U.N.A.M.  270  Niederberger, Christine 1979 "Early Sedentary Economy in the Basin of Mexico." Science 203: 131-142. Nunley, P. 1967 "A hypothesis concerning the relationship between Texcoco fabricmarked pottery, tlateles, and chinampa agriculture." American Antiquity 32:515-522. Odum, Howard 1971 Environment Power, and Society. Odum, H.T. and E.C. Odum 1976 Energy Basis for Man and Nature.  New York: Wiley. New York:  McGraw-Hill.  Orozco-Segovia, A. and S.R. Gliessman 1979 "The Marceno in Flood-Prone Regions of Tabasco." Prepared for the Symposium on Mexican Agroecosystems, XLIII International Congress of Americanists: Vancouver. Ortiz de Montellan, Bernal 1975 "Empirical Aztec Medicine."  Science 188:215-220.  Palerm, Angel. 1968 "The Agricultural Basis of Urban Civilization in Mesoamerica." Reprinted in Y.A. Cohen (Ed.) Man in Adaptation: The Cultural Present. Chicago: Aldine. Palerm, Angel 1973 Obras hidraulicas prehispanicas en el" si sterna lacustre del Valle de Mexico. Mexico: I.N..A.H. Parsons, J.R. 1974 "The Development of a Prehispanic Complex Society: A Regional Perspective from the Valley of Mexico." Journal of Field Archaeology 1:81-108. Parsons, J.R. 1976 "The Role of Chinampa Agriculture in the Food Supply of Aztec Tenochtitlan." In C E . Cleland (Ed.) Cultural Change and Continuity. New York: Academic Press. Peterson, F. 1962 Ancient Mexico: an Introduction to Prehispanic Cultures. New York: Capricorn. Pohl, R.W. 1978 How to Know the Grasses: The Pictured Key Nature Series. Dubuque, Iowa: Wm. C. Brown Company.  271  Polanyi, Karl 1944 The Great Transformation. Winston.  New York:  Holt, Rinehart and  Prescott, G.W. 1970 How to Know the Freshwater Algae: The Pictured Key Nature Series. Dubuque, Iowa: Wm. C. Brown Company. Price, B.J. 1971 "Prehispanic Irrigation Agricultural in Nuclear America." Latin American Research Review 6(3):3-60. Rasmussen, Wayne D. 1982 "The Mechanization of Agriculture." 76-89. Redfield, R. and R. Alfonso Villa 1934 Chan Kom: . A Mayan Village. Institute. Sanchez y Sanchez, Oscar 1978 La Flora del Valle de Mexico.  Scientific American 247(3):  Washington, D . C : Carnegie  Mexico:  Editorial Herrero S.A.  Sanders, William T. 1962 "Cultural Ecology in Nuclear Mesoamerica." pologist 64(l):34-44.  American Anthro-  Sanders, William T. and Barbara J. Price 1968 Mesoamerica: the Evolution of a Civilization. Random House.  New York:  Sanders, William T. et al. 1970 "Teotihuaccin Valley Project, Final Report." Occasional Papers, Department of Anthropology 1(3). Pennsylvania State University. Sanders, William T. 1976 "Natural Environment." In E.R. Wolf (Ed.) The Valley of Mexico: Studies in Pre-Hispanic Ecology and Society. Albuquerque: University of New Mexico Press. Sanders, William T., J. Parsons, and R.S. Santley 1979 The Basin of Mexico: Ecological Process in the Evolution of a Civilization." New York, San Francisco, London: Academic Press. Schilling, Elizabeth 1938 Die "Schwimmenden Garten' von JXochimilco.  Kiel.  272  Simmons, I.G. 1974 The Ecology of Natural Resources.  London: E. Arnold.  Simon, Herbert A. 1962 "The Architecture of Complexity." Proceedings of the American Philosophical Society 106:467-482. Simpson, L.B. 1952 "The Exploitation of Land in Central Mexico in the 16th Century." Ibero America 36. Soustelle, J. 1961 The Daily Life of the Aztecs. Nicholson.  London: Weidenfeld &  Stadelman, Raymond 1940 "Maize Cultivation in Northwestern Guatemala." Contribution to American Anthropology and History. Publication #523. Washington, D . C : Carnegie Institution. Steggerda, M. 1941 "The Maya Indians of Yucatan." Contribution to American Anthropology and History. Publication 531. Washington, D.C: Carnegie Institution. Thornthwaite, CW. 1964 Average Climatic Water Balance Data of the Continents, Part VI. North America excluding the United States. Publications in Climatology 17(2). Centerton, N.J.: CW. Thornthwaite Associates. TrierweiTer, J.F. and J.D. Utzinger 1975 "Nutrients Needed by Plants." In Wearne (Ed.) Natural Gardening Handbook:15-20. New York: Brooklyn Botanical Garden. United States Department of Agriculture 1971 Common Weeds of the United States. Publications, Inc. Vaillant, G.C 1951 The Aztecs of Mexico.  Middlesex:  New York: Dover  Penguin.  Vanegas, R.C. 1978 Las Chinampas de Mixquic. Thesis, Fac. de Ciencias. U.N.A.M.  Mexico:  Vickers, Geoffrey 1956 Control, Stability and Choice. 9th Wallberg Memorial Lecture, University of Toronto, 30 October. Villa Alfonso, R. 1945 The Maya of East Central Quintana Roo. Carnegie Institution.  Washington, D . C :  273  Vogt, E.Z. 1969 Zinacantan: A Maya Community in the Highlands of Chiapas. Cambridge: Harvard University Press. Wall en, C.C. 1955 "Some Characteristics of Precipitation in Mexico." Geografiska Annaler XXXVI:l-2. Watt, K.E. 1968 Ecology and Resource Management: New York: McGraw-Hill.  A Quantitative Approach.  West, R. and P. Armillas 1950 "Las Chinampas de Mexico." Cuadernos Americanos 50:165-182. Wilken, Gene C. 1979a "Mucks, Mucking and Soils of the Chinampas of Mexico." XLIII International Congress of Americanists: Vancouver. 1979b: "A Note on the Buoyancy of the 'Floating' Chinampas of Mexico." XLIII International Congress of Americanists: Vancouver. Wilkinson, R.E. and H.E. Jazues 1979 How to Know the Weeds: The Pictured Key Nature Series. Dubuque, Iowa: Wm. C. Brown Company. Willey, N.C. 1939 "El Embrujo de las Chinampas." Hispanic American Historical Review 19. Wittfogel, Karl A. 1957 Oriental Despotism.  New Haven: Yale University Press.  Wolf, Eric 1959 Sons of the Shaking Earth. Press.  Chicago:  University of Chicago  Wolf, Eric (Ed.) 1976 The Valley of Mexico: Studies in Prehispanic Ecology and Society. Albuquerque: University of New Mexico Press. Zimmerman, E.W. 1964 Introduction to World Resources.  New York:  Harper and Row.  Zorita, Alonso de 1963 Life and Labor in Ancient Mexico: The Brief and Summary Relation of the Lords of New Spain. Translated by Benjamin Keen. New Jersey: Rutgers University Press.  


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items