UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

A functional classification of Canadian cities Maxwell, James William 1964

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

Item Metadata

Download

Media
831-UBC_1964_A8 M39.pdf [ 10.34MB ]
Metadata
JSON: 831-1.0104958.json
JSON-LD: 831-1.0104958-ld.json
RDF/XML (Pretty): 831-1.0104958-rdf.xml
RDF/JSON: 831-1.0104958-rdf.json
Turtle: 831-1.0104958-turtle.txt
N-Triples: 831-1.0104958-rdf-ntriples.txt
Original Record: 831-1.0104958-source.json
Full Text
831-1.0104958-fulltext.txt
Citation
831-1.0104958.ris

Full Text

A FUNCTIONAL CLASSIFICATION OF CANADIAN CITIES by JAMES WILLIAM MAXWELL B.A., The University of British Columbia, 1959 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF . . A/ THE REQUIREMENTS FOR THE DEGREE OF MASTER OF ARTS in the Department of GEOGRAPHY We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA July, 1964 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree .that the Library shall make it freely available for reference and study. I further agree that per-mission for extensive copying of this thesis for scholarly purposes may be granted by the Head of my Department or by his representatives. It is understood that, copying or publi-cation of this thesis for financial gain shall not be allowed without my written permissions, Department of Geography The University of British Columbia, Vancouver 8, Canada Date July, 1964. ABSTRACT The major cit i e s i n Canada have been classified i n terms of their functional structure i n order to develop an overview of the Canadian urban milieu, A quantitative method of classification based on census labour force s t a t i s t i c s has been used to identify the functional charac-ter of c i t i e s . Examination of the traditional techniques of c i t y functional classification reveals that a good quantitative classification scheme must recognize that a l l c i t i e s are multifunctional, that changing city size affects city functional structure, and that urban functions are essentially dichotomous by nature, having distinct " c i t y -serving" and "city-forming" characteristics. Generally the city-serving a c t i v i t i e s are ubiquitous, being found in almost a l l centers and usually having relatively constant levels of importance i n the functional profiles of c i t i e s . In contrast, the city-forming functions appear sporadically in c i t i e s and have a great range in the importance they exhibit i n city functional structures. This importance ranges from complete domination to no representation at a l l for some functions. Because city-forming activity reveals the essential functional role of a city, only this a c t i v i t y should be uti l i z e d when classifying c i t i e s i n terms of • • * XXX function. The "minimum requirement" technique as developed by Ullman and Dacey has been used to classify the c i t i e s because i t conforms most closely to theoretical considerations, using only city-forming activity as the basis of c l a s s i f i c a -tion and allowing for the effect of changing city size on city functional profiles. In addition i t provides for a measure of c i t y functional specialization. The position of an activity i n a city's functional profile should be examined on two distinct planes: (1) i t s importance relative to that of other functions in the city's functional structure, and (2) i t s importance in the city's functional profile relative to i t s importance in the functional profiles of a l l other c i t i e s . The activity that occupies the highest position i n a city's functional structure—determined by the proportion of city-forming employment in the different functions—is termed the city's "dominant" function. A function that engages an atypically high proportion of a city's c i t y -forming employment in relation to the proportion usually found in the function in most c i t i e s i s called a "distinctive" function. By determining the dominant and distinctive functions of c i t i e s and analyzing the distribution patterns of functional relative importance and city functional speciali-zation, several observations can be made regarding the character of the functional performance of c i t i e s . i v The findings of the classification exercise generally coincide with observations based on qualitative data and with the results of other similar quantitative studies. Trade and manufacturing are the key urban functions both i n the cit y -serving and city-forming profiles of c i t i e s . The propensity for functional specialization decreases with increasing city size. Gity size, however, i s not the only factor governing city specialization. Elements such as the importance and kind of manufacturing i n the functional profile, and the degree of "isolation" a ci t y experiences are also important factors affecting city specialization. The distribution patterns of relative importance of the key urban a c t i v i t i e s are extremely uneven and indicate that a fundamental difference in functional performance exists between the c i t i e s of the densely populated St. Lawrence Lowlands and southern Ontario—the "heartland"--and the cit i e s of the remaining parts of Canada—the "periphery". Heartland c i t i e s are generally more specialized and emphasize manufacturing to a greater degree than do the peripheral c i t i e s . The latter, except for a very few resource-oriented manufacturing centers, are quite diversified and are inclined to have an important involvement with functions associated with distance such as wholesale trade and transportation. V The Canadian heartland and periphery are geographic r e a l i t i e s . They dif f e r i n hi s t o r i c a l , economic, and to some degree, i n cultural development. That their c i t i e s reflect these differences seems like a reasonable and to-be-expected conclusion. ACKNOWLEDGMENTS The writer wishes to express his sincere thanks and appreciation to the authorities of the Geographical Branch, Department of Mines and Technical Surveys for the use of Branch f a c i l i t i e s during the course of this project. To Dr. W.G. Hardwick i s tendered special acknowledgment for his continued interest and encouragement maintained over many months and across nearly three thousand miles. TABLE OF CONTENTS CHAPTER PAGE. I. CLASSIFICATION OF CITIES: A CRITICAL REVIEW . . 1 The Problem. 1 Statement of the problem 1 Importance of the study. 2 Definitions of Terms Used 3 City functional structure 3 Canadian regions: heartland and periphery . 4 Organization of the Remainder of the Thesis. . 5 Chapter I. 5 Chapter II 6 Chapter III 7 Chapter IV 7 Review of the Literature on Functional Classi-fication of Cities 7 Qualitative studies 7 Quantitative studies 19 Quantitative source material . • 22 Methods of identifying functional impor-tance 25 City-type classes 3 8 A critique of quantitative studies i n city functional classification 40 Method of Study 50 v i i CHAPTER PAGE Cities studied • • • 50 Source material. 51 Methods of analysis 59 II. THE RELATIVE IMPORTANCE OF URBAN FUNCTIONS IN CANADIAN CITIES. 63 A Measure of Functional Importance . . . . . . 64 The standards • 64 The measure 77 Functions and Cities 82 The Distribution of the Relative Importance of Functions in Canadian Cities 97 The distribution of the relative importance of the sporadic functions 101 Extraction 101 Manufacturing, wholesale trade, and transportation 103 Government service . . . 108 The distribution of the relative importance of the ubiquitous functions 110 Summary: "heartland", "periphery", and Canada's urban geography 117 v i i i CHAPTER PAGE III. THE DOMINANT FUNCTIONS AND SPECIALIZATION OF CANADIAN CITIES 122 Measures of Functional Dominance and Specialization 122 Functional dominance . . . . . . . 122 Functional specialization 123 Dominant Functions of Canadian Cities. . . . . 135 Functional Specialization of Canadian Cities • 148 The relationship of city specialization to dominant functions and city location . . . 148 A quantitative measure of the association of city specialization with city size, "isolation", and manufacturing activity. • 154 Comments on "city types" 160 IV. SUMMARY AND CONCLUSIONS 174 The Canadian Urban Milieu: An Overview. . . . 174 City Functional Classification: Retrospect and Prospect 173 BIBLIOGRAPHY 189 APPENDIX . . . . . . 195 LIST OF TABLES TABLE PAGE I. Examples of Source Material Utilized i n Quantitative Studies of City Functional Structure, 26 II. Example of Conversion of Census Industrial Categories into City Functional Categories • • 29 III. Examples of City-Type Classes Used i n City Functional Classification Studies. . . . . . . 41 IV. Cities Examined: Population, Location, and Legal Constituents • 52 V. Census Industrial Categories and City Functional Categories Used i n the Study • 58 VI. Minimum Percentages Employed i n Canadian Cities of Varying Size Classes, 1951. 66 VII. Values of Excess Employment for Determining Classes of Functional Importance • . 80 VIII. The Mean Percentage Distribution of Urban Labour Force Among Functions i n Canadian Cities . . . 83 IX. The Position of Functions as "City-Servers": Functions Ranked by Decreasing "Expected Minimum Requirement Values" for Cities with Populations of 10,000, 100,000, and 1,000,000 Inhabitants 87 X TABLE PAG: X. Specialization Indexes and Dominant Functions of Canadian Cities* 139 LIST GF FIGURES FIGURE PAGE 1, Canadian Cities of 10,000 Population and Over, 1951 56 2„ The Relationship of Functional Importance to City Size, and Regression Lines showing "Expected Minimum Requirement Values" for Functions 69 3. Selected "Expected Minimum Requirements" for Twelve Functions, Based on Regression Lines. • 7$ 4. The Distribution of Wholesale Trade and Manufacturing as "Distinctive" Functions i n Canadian Cities 105 5« The Distribution of Retail Trade and Recreation as "Distinctive" Functions in Canadian Cities. I l l 6» Functional Specialization and Dominant Functions of Canadian Cities . . . . 142 7. "City Types" i n Canada . . . . . 163 CHAPTER I CLASSIFICATION OF CITIES: A CRITICAL REVIEW Cities serve manifold functions in the economy and culture of a people. A l l c i t i e s have some functions i n common; a l l c i t i e s have some functions that are peculiar to their site and situation, to the people whom they serve; and a l l c i t i e s have some functions peculiar to their development and history; hence c i t i e s may be classified more effectively on the basis of their functions as c r i t e r i a than perhaps according to any other attribute. 1 I. THE PROBLEM Statement of the problem. It was the purpose of this study to examine certain aspects of city functional character as manifested i n Canadian urban centers. The objectives of the study were (1) to investigate the tech-niques available for city functional classification and to develop a taxonomy suitable for classifying Canadian c i t i e s ; (2) to classify c i t i e s by their distinctive and dominant functions and degree of specialization; and (3) to describe and analyze: (a) the distribution patterns of the relative importance of functions among c i t i e s , and (b) the distribution patterns of city types and c i t y specialization. l nUrban Functions," Economic Geography. XXI (April, 1945), 78. 2 Importance of the study. Study of the functional character of urban centers i n Canada has been for the most part limited to evaluation of the functions of individual cit i e s or, at most, to study of special groups of c i t i e s . To the writer's knowledge no published studies exist which attempt to examine the functional character of Canadian ci t i e s by taking into consideration a l l principal c i t i e s and objectively relating the functional structure of each to that of the others. By dealing with a l l major Canadian centers in one study i t was anticipated that an overview of the functional activity of the cities would emerge. Such an overview, i t was hoped, would be a contribution to the knowledge of Canadian urban geography. The study was macro i n scale, consequently certain limitations were inherent i n the work. The scope of the study prevented a detailed analysis of each city and individual urban a c t i v i t i e s . Consequently, the explanatory comments on the several distribution patterns of functional importance contain no "new" information; they are restate-ments of well known locational principles viewed i n a ; 2Three examples of such studies are: Louis Trotier, "Some Functional Characteristics of the Main Service Centers of the Province of Quebec," Cahiers de Geographie de Quebec. No. 6 (April-September, 1959). 243-259; J. Spelt. The Urban  Development i n South Central Ontario (Assen, The Netherlands: Van Gorcum & Company, 195$); Ira M. Robinson, New Industrial  Towns on Canada's Resource Frontier (Chicago: University of Chicago, Department of Geography, 1962). 3 Canadian context. The contribution of the study l i e s , i t i s thought, i n the way the data have been developed to obtain the overview of urban centers and their functions. II. DEFINITIONS OF TERMS USED City functional structure. The functional structure of a cit y was interpreted i n this study as being the summation of a l l the socio-economic acti v i t y of a c i t y . The units i n the summation are the proportions of the city's labour force employed i n a l l i t s different a c t i v i t i e s . In essence, "functional structure" has been equated with "employment structure" as i t was thought that the a c t i v i t i e s i n which a city's people earn their l i v i n g are the best indicators of city functions. The terms "city a c t i v i t y " and "city function" were taken to be synonomous i n this study. Two kinds of functions were recognized when characterizing the functional structure of c i t i e s : distinctive functions and dominant functions. The distinctive functions i n a city are defined as those ac t i v i t i e s whose proportions of the city labour force are extremely high i n relation to their average propor-tions i n a l l c i t i e s studied. They are very important i n the functional structure or profile of the cit y relative 4 to their importance i n other c i t i e s . The dominant function in a cit y i s defined as the activity engaging the largest proportion of the cit y labour force. It i s the most impor-tant function i n the city's functional structure or profile, relative to the other functions i n the city . Canadian Regions: heartland and periphery. For the purpose of this study Canada was broken down into two regions; the "heartland" and the "periphery". The densely populated areas of the St. Lawrence Lowlands, the Eastern Townships and southern Ontario constitute the heartland and the remaining settled parts of Canada make up the periphery. The periphery has been divided into three components: the western periphery containing the four western provinces and northwestern Ontario; the "northern" periphery consisting of northern Ontario and Quebec; and the eastern periphery composed of the lower St. Lawrence area of Quebec and the Atlantic provinces. Areas not containing c i t i e s of ten thousand population and over were not considered; thus, the Northwest Territories and the most northerly areas of Quebec, Ontario, and the western provinces are ignored i n this "regional system". Of the eighty Canadian c i t i e s studied, forty-five are i n the heartland, fifteen i n the western periphery, eight in the northern periphery, and twelve i n the eastern 5 periphery. Figure 1 on page 5.6 indicates the regional location of the c i t i e s . III. ORGANIZATION OF THE REMAINDER OF THE THESIS The study was essentially an exercise i n the classification of ci t i e s by functional specialization. This involved a problem i n comparative analysis since c i t i e s of greatly different size and situation had to be dealt with on the same plane i n the analysis. An important part of the study involved the development of classification techniques which make allowance for city size differences and the associated bias introduced i n city functional structure. The techniques applied i n the study were developed i n similar studies carried out elsewhere, mainly in the United States. As i n most works of this kind, census labour force st a t i s t i c s were used as the basic source material i n the classification scheme. Chapter I. The remainder of Chapter I i s divided into two parts. The f i r s t part i s a review of the l i t e r a -ture on the functional classification of ci t i e s ; the second i s an outline of the methods and materials that were used in the study. In the literature review the various concepts, techniques and materials u t i l i z e d i n previous studies are examined. The problems encountered, both conceptual and 6 technical, are noted and a critique of the different approaches i s offered. In the outline of methods and materials the approaches that were taken in this study to the problems of functional classification are briefly described. The detailed description of techniques i s l e f t to those chapters where they apply. Chapter II. Because cit i e s are multifunctional i t was considered necessary to examine city functions i n a systematic fashion as well as i n terms of dominant ac t i v i t i e s so that no aspect of city functional structure remained hidden. Chapter II i s devoted to a systematic discussion of the place of urban functions i n Canadian c i t i e s . A measure of relative importance of urban a c t i v i t i e s i n the functional structure of c i t i e s was developed i n order that the objective of the chapter be met. In the f i r s t part of the chapter the measure of relative importance i s outlined, and in the second part, the characteristics of urban functions and city functional structure are discussed. In the remaining section of the chapter the position of the functions in c i t i e s , as indicated by their relative importance, i s discussed. 7 Chapter III, In Chapter III the distribution of city types i n Canada i s described and analyzed. To determine the patterns of city types and specialization, c i t i e s were classified by dominant function and degree of specialization. This discussion i s the complement of that presented i n Chapter II, Both views are required to provide a balanced view of urban functions and the functional structure of c i t i e s . The problem encountered here was the development of techniques suitable for identifying dominant functions and measuring the degree of city specialization. The techniques u t i l i z e d i n measuring the relative importance of functions were further developed to meet this problem. Chapter IV, Chapter IV contains a summary of the findings of the preceding chapters and a concluding statement giving an assessment of the approaches and techniques of functional classification of c i t i e s . IV. REVIEW OF THE LITERATURE ON FUNCTIONAL CLASSIFICATION OF CITIES Qualitative Studies The f i r s t studies to examine urban functions and to classify c i t i e s by function were made late i n the nineteenth and early i n the twentieth 8 centuries,3 These f i r s t attempts at functional c l a s s i f i -cation were based on qualitative analysis. S t a t i s t i c a l data necessary for quantitative study were not available, their appearance having to await the development of modern enumeration devices. These f i r s t studies, while unable to classify c i t i e s with any real objectivity, were important however, because they provided a great deal of preliminary information on urban functions and city functional structure. Classic among these works i s the study of M. Aurousseau on urban and rural characteristics.^- In this report, Aurousseau presented theory on city existence and location that i s s t i l l valid. He identified function as being the essential element of urban character and was able to relate the performance of urban functions to location. . . . i t i s at once evident that function i s the driving force i n the l i f e of towns.... A town comes into being either at a point having those characteristics of nodality which enable i t to discharge that partic-ular function to the best advantage or at a point a r t i f i c a l l y endowed with nodality.5 3Many of these early studies are cited i n : CD. Harris, "A Functional Classification of Cities i n the United States," Geographical Review. XXXIII (January, 1943), 86-99; P.E. James and C.F. Jones (eds.), American Geography: Inventory & Prospect (Syracuse: Syracuse University Press, 1954), pp. 143-44; G. Alexandersson, The Industrial Structure of American Cities (Lincoln, Neb.: University of Nebraska Press, 1956), pp. 20-21. •^M. Aurousseau, "The Distribution of Population: A Constructive Problem," Geographical Review. XI (October, 1921), 563-92. 5lbid.. 569. He went further to identify six functions that urban centers perform and presented comments on the structure and location of c i t i e s of specific functional type. Recognition was also given to the fact that urban centers, although essentially multifunctional, usually have one dominant function. Within the national boundary are numerous urban groups which exist for the exercise of the following six functions: administration, defense, culture, production, communication, recreation.•.• as a l l towns are placed i n nodal situations, many are conveniently situated for the discharge of more than one function. There i s generally one phase of activity, however, which overshadows the r e s t . 0 Aurousseau's comments on the character of urban ac t i v i t i e s are especially significant. Later theories on urban character were anticipated when he spoke of "primary" and "secondary" occupations existing in c i t i e s ; the former concerned with the basic function of the city, the later concerned with the maintenance of the labour force engaged i n the primary occupations. These early investigators, by identifying the major urban functions and noting, in preliminary fashion, the relationship between functional performance and location l a i d the groundwork i n the development of theory necessary for the rational classification of c i t i e s . Their contri-6 I b i d . 10 butions were further refined i n a later study by Harris and Ullman.7 These authors recognized, as had earlier writers, that the existence of c i t i e s depends on the service they perform for their hinterlands. They noted the close association existing between functions and locational factors and were able to present a classification of c i t i e s based on the locational characteristics of functions. This classification, containing three categories, represents a keen insight into the role of ci t i e s and implicated a l l the primary or basic functions c i t i e s perform. The categories were: 1. Cities as central places performing comprehensive services for a surrounding area.... 2. Transport c i t i e s performing break-in-bulk and a l l i e d services along transport routes, supported by areas which may be remote i n distance, but close in connection because of the city's strategic location on transport channels.... 3. Specialized-function c i t i e s performing one service such as mining, manufacturing, or recreation for large areas, including the general tributary areas of hosts of other c i t i e s . . . , 8 The multifunctional character of c i t i e s was also acknowledged by the authors in the statement that, "most c i t i e s represent a combination of the three factors 7c.D. Harris and E.L. Ullman, "The Nature of C i t i e s , " Annals of the American Academy of P o l i t i c a l and Social  Science, CCXLII (November. 1 9 4 5 ) , 7-17. ~ ~ SIbid., 7-9. 11 [central place services, transport services, special-function services], the importance of each varying from city to city."9 Further c l a r i f i c a t i o n of the origin, character, and locational aspects of urban functions important to cit y classification has resulted from work i n fi e l d s a l l i e d to geography. The economist, R.U. Rat c l i f f , i n a succinct explanation of the economics of urbanization, identified the fundamental forces of urbanization, described the primary and secondary functions, and related them to locational f a c t o r s . 1 0 He states that, "the explanation of the urban organization of society w i l l be found i n the socioeconomic a c t i v i t i e s that require the concen-tration of people, buildings, and machines within relatively small areas". 1 1 Although recognizing social forces as being the f i r s t urbanizing factors—survival, as well as religious and administrative a c t i v i t i e s demanded places of defense and assembly—he states that i n the modern industrial state, economic forces are the principal urbanizing factors. "The 9 I b i d . 10R.U. Rat c l i f f , Urban Land Economics (New York: McGraw-Hill Book Company, 1949), pp. 19-59. n I b i d . , p. 20 . 12 a c t i v i t i e s of man that are fundamental as agglomerative forces i n the formation of c i t i e s are extraction, manufacturing, and trade"* 1 2 Ratcliff observes, as did Aurousseau, that other functions occur in c i t i e s ; those secondary ac t i v i t i e s that service the primary functions. These are identified as: financial, business, personal and professional services. The social functions: administrative, religious, cultural, and recreational a c t i v i t i e s , that led to the development of the f i r s t urban centers are, for the most part, also secondary functions having been relegated to this position by the emergence of the industrial state. Only i n isolated cases do they continue to play their former roles of primary city functions. The distinction made in the early studies between primary and secondary functions was given formal recog-nition i n the economic base theory, formulated in the late 1920's and refined i n the 1930's and 1940's. The main points of this theory have been outlined concisely by Pfouts: This theory may be characterized br i e f l y by saying that i t divides urban economic ac t i v i t y into two categories: exporting industry that brings money into the community from the outside world and 1 2 I b i d . , p. 29. 1 3 non-exporting industries whose goods and services are sold within the community. The exporting industries are referred to as basic industries and the non-exporting industries are called service industries• 3 i t i s also contended i n discussions of the theory that the exporting oi* basic industries provide the source of urban growth; they are "city building" i n d u s t r i e s . 1 4 J.-W. Alexander has outlined three geographic qualities of the theory which have relevance to the classification of c i t i e s . x 5 fje states that the theory: (1) reaffirms the relationship between city and hinter-land by providing a view of economic ties which bind a city to other areas; (2) permits the most satisfactory classification of c i t i e s i n terms of regional function: Cities are more accurately distinguished by their basic economy than by their total economy because the basics express a city's service to i t s region. For such purpose, the nonbasics 'cloud the picture' and therefore should be subtracted from the total economy as one endeavors to distinguish industrial c i t i e s from commercial c i t i e s from government c i t i e s , etc.l° Sometimes called "nonbasic industries". 1ZfR.W. Pfouts (ed.), The Techniques of Urban Eco- nomic Analysis (West Trenton, New Jersey: Chandler-Davis Publishing Company, I 9 6 0 ) , p. 1. For additional comment on the economic base theory see: W. Isard et a l . , Methods  of Regional Analysis: An Introduction to Regional Science (Cambridge, Mass.: The M.I.T. Press, I960), PP. 189-205. •^ J.W. Alexander, "The Basic-Monbasic Concept of Urban Economic Functions," Economic Geography, XXX (July, 1954), 2 4 6 - 6 1 . l 6 I b i d . , 2 5 1 - 2 5 2 . 14 Thirdly, the "basic-nonbasic r a t i o " — t h e ratio between total labour force in basic activity and total labour force i n nonbasic activity—may be of significance i n distinguishing types of citi e s since different types of citi e s appear to have different ratios. Several intensive studies of individual ci t i e s were conducted i n the 1940's using the concepts of the economic base theory. 1? Although these studies revealed much about the economic profiles of individual c i t i e s , certain tech-nical and conceptual problems emerged with the application of the theory. It was found d i f f i c u l t to determine what was basic and what was nonbasic activity, and then more d i f f i c u l t to develop measures for separating basic from nonbasic activity. These problems precluded the use of the "basic-nonbasic r a t i o " as a criterion i n differentiating city types i n studies involving the comparative analysis of many c i t i e s . More significant i n placing limitations on the usefulness of the theory has been the criticism levelled at certain conceptual aspects of the theory, especially that directed at the proposition that the basic ac t i v i t y of a city may be used as a criterion i n predicting city growth.x^ Regardless of the shortcomings of the 1 7 I b i d . , 255-259. l % o r critiques of the economic base theory see: Pfouts, op.cit., pp. 213-341. 15 theory however, some of i t s concepts have been useful in the identification of distinctive and dominant functions i n c i t i e s . Many analysts interested i n classifying c i t i e s have based at least some parts of their classification scheme on concepts of the economic base theory. The influence of the theory i s evident in such statements as "...•the city serving structure' was subtracted from the total structure before the classification was made",19 and "the problem of functional classification i s best under-taken by considering functional specialization in terms of the kinds of export activity of a community which creates an inflow of money to the community,w2<^ The division of city functions into basic and service categories has resulted i n some interesting observations regarding the place of the two components i n the functional profile of c i t i e s , Alexandersson has noted that the distribution patterns of urban a c t i v i t i e s among c i t i e s can be grouped into two broad types: sporadic and ubiquitous distributions,21 He termed those a c t i v i t i e s ^Alexandersson, op, c i t . , p, 22, 20 O.D. Duncan and A.J, Reiss, Jr., Social Charac- t e r i s t i c s of Urban and Rural Communities: 1950 (New York: John Wiley & Sons, 1956), p. 21b, cited by E.D. Duncan et a l , , Metropolis and Region (Baltimore: The Johns Hopkins Wess, I960), p. 34. pi Alexandersson, op. c i t , , pp, 13-14. 16 which are not represented i n many c i t i e s , hut assume very important or dominant roles i n some ci t i e s as being sporadic a c t i v i t i e s . Activities that are found in a l l c i t i e s were considered ubiquitous a c t i v i t i e s ; only infrequently do these functions occupy distinctive or dominant positions i n city functional structure. Generally the sporadic a c t i v i t i e s can be considered as being basic functions. The very character of their distributions suggest they are exporting a c t i v i t i e s since when they do occur i t i s usually i n great magnitude and they tend to dominate the functional profile of a city. Mining i s an example of a sporadic activity; i t s importance i s non-existent or insignificant i n most c i t i e s , but i n a few i t occupies the position of extreme importance, being the sole basic activity of the city. The characterizing of the ubiquitous functions as being either basic or service i s somewhat more complex. Barring special circumstances however, most ubiquitous a c t i v i t i e s can be considered service functions. Only when certain characteristics of the sporadic-type distributions "invade" the normal ubiquitous pattern i s this not valid. For example, a l l ci t i e s have an education function. In most cit i e s this i s a service function (city-serving), hence, the relative importance of the 17 function among ci t i e s remains relatively constant. Highly specialized college towns, however, do not f i t the pattern. They have atypically high values for the education function. In this case an activity, normally found to be service or city-serving i n character, has assumed the role of a basic or city-forming function. This characteristic of the sporadic ac t i v i t i e s found occasionally i n the distribution patterns of primarily ubiquitous functions allows the isolation of those instances where ac t i v i t i e s normally considered secondary functions have assumed the role of primary a c t i v i t i e s . These observations on the relationship between basic and service functions, and sporadic and ubiquitous d i s t r i -butions have provided an important aid in the identification of distinctive and dominant functions i n c i t i e s as well as in supplying insight into the role of individual ac t i v i t i e s i n many c i t i e s . To summarize, the contributions of the qualitative studies, significant to city functional classification, are noted in the form of observations on c i t i e s and city functions which the studies have made possible. These observations have provided the foundation on which city classification, u t i l i z i n g s t a t i s t i c a l data, has been b u i l t . (1) function i s the essential element of urban character. (2) urban functions are those activities that require the concentration of people, goods, and services i n small areas for their performance. (3) c i t i e s are essentially multifunctional, but usually they have one and sometimes several dominant functions. (4) city function and location are closely related because the existence of a city depends on the functions i t performs for: (a) i t s immediate hinterland (central place functions). (b) the nation (transport and special-function services). (5) trade, manufacturing, and extraction are the most important urban a c t i v i t i e s in an industrialized state. (6) urban a c t i v i t i e s can be divided into two components: basic and service. Basic a c t i v i t i e s are the "exporting", "city-forming" or "primary" a c t i v i t i e s of the c i t y . Service a c t i v i t i e s are the "non-exporting", "city-serving" or "secondary" ac t i v i t i e s which maintain the basic functions. Cities are best classified by their basic a c t i v i t i e s , since they are the raison d'etre of / the city and i l l u s t r a t e most clearly the functional relationship between the city and i t s hinterland, and the state« (7) The distribution patterns of urban a c t i v i t i e s among cit i e s usually approximate one of two patterns: sporadic or ubiquitous. Basic a c t i v i t i e s are usually characterized by sporadic patterns, whereas service a c t i v i t i e s usually have ubiquitous distributions. The fact that service a c t i v i t i e s are ubiquitous confirms the concept that c i t i e s are multifunctional, a l l having a standard repertoire of service functions. The sporadic distribution patterns of the basic a c t i v i t i e s i l l u s t r a t e the variab i l i t y among ci t i e s in their functional roles within the state. The specific functional role a city f u l f i l l s i s determined by i t s site and situation, and the economy of the state. It now remains to be seen how these observations have been ut i l i z e d in the classification of c i t i e s by quantitative means. Quantitative Studies The findings of the early qualitative studies established the fact that function i s the key to an understanding of city existence and location. Once 2 0 detailed s t a t i s t i c a l data on ci t i e s were available i t was possible to attempt the identification of distinctive functions i n a large number of ci t i e s through quantitative study. The object was to identify distinctive and dominant functions, distinguish city types, and to relate the locational characteristics associated with specific functions to the actual locations of c i t i e s . The f i r s t functional classifications of c i t i e s based on quantitative data appeared i n 1943. C.D. H a r r i s 2 2 classified a l l American c i t i e s of 10,000 population and over into eight city-type classes, and W. William-01sson23 published a classification of urban centers in northern Sweden i n which towns were grouped into three categories. These two studies established the general format which most of the later works of similar type were to follow. The majority of the studies to classify c i t i e s by function i n the Harris-William-Olsson tradition appeared in the late 1940's and 1950's. They covered urban centers 22C.D. Harris, "A Functional Classification of Cities in the United States," Geographical Review, XXXIII (January, 1943), 86-99. 23utredning angaende NorrLands naringsliv," Statens offentliga utredningar ( 1943:39) , cited by G. Alexandersson, The Industrial Structure of American Cities (Lincoln, Neb.: University of Nebraska Press, 1956), p. 21. See also: W. William-Olsson, Ekonomisk-geografisk karta over Syerige (Economic geographical map of Sweden), Stockholm, 1946. 221. i n the European countries, New Zealand and Japan as well as those i n the United States and in a part of Canada.2^ These studies, with some notable exceptions, follow con-ceptual approaches to the problem similar to those used by Harris and William-Olsson, but they dif f e r i n technical detail from the f i r s t studies and from each other. A multitude of city types have been recognized and numerous ways of identifying distinctive functions have been established. The one factor which identifies these studies as members of a distinct group of studies i s their common ^Alexanders son cites the functional classification studies carried out up to 1955. G. Alexandersson, The  Industrial Structure of American Cities (Lincoln, Neb.: University of Nebraska Press, 1956), p. 21. Similar studies published on city functional classification since the appearance of Alexandersson's work include: I. Morrissett, "The Economic Structure of American Ci t i e s , " Papers and Proceedings of the Regional Science Association. IV (1958), 239-56; L. Kosinski, "Problem of the Functional Structure of Polish Towns," Przeglad Geograficzny (Polish Geographical Review), XXXI (Supplement, 1959), 35-67, (in English); J.W. Webb, "Basic Concepts in the Analysis of Small Urban Centers of Minnesota," Annals of the  Association of American Geographers. XLIX (March, 1959), 55-72; L. Trotier, "Some Functional Characteristics of the Main Service Centers of the Province of Quebec," Cahiers de Geographie de Quebec. No. 6 (April-September 1959), 243-259; E.L. Ullman and M.F. Dacey, "The Minimum. Requirements Approach to the Urban Economic Base," Papers  and Proceedings of the Regional Science Association, VI (I960), 175-194; Y. Watanabe, "An Analysis of the Function of Urban Settlements Based on S t a t i s t i c a l Data - A Func-tional Differentiation Vertical and Lateral." The Science  Reports of the Tokoku University (Geography), No. 10 (September, 1961), 63-94. 22 use of census data as the source material for classification. They are quite distinct from studies which u t i l i z e other data in addition to census material for classification purposes.25 Usually the methods of analysis are more complex i n these latter studies, hence, fewer c i t i e s are normally dealt with i n them. This study i s modelled after the former type where only one source of data i s used, but where many c i t i e s are involved i n the study. The basic problems encountered when formulating city classifications based on s t a t i s t i c a l data include: (1) the choice of quantitative source material, (2) the selection of methods to identify distinctive and dominant functions in c i t i e s , and (3) the establishment of city-type classes. The classification schemes that have been developed using single sources for data, are discussed below i n terms of these three problems. Quantitative source material. The s t a t i s t i c a l data selected as source material for a quantitative study i n city functional classification must meet two requirements. ^Examples of such studies are: O.D. Duncan et a l . , Metropolis and Region (Baltimore; Johns Hopkins Press, I960): M. Palomaki. The Functional Centers and Areas of  South Bothnia, Finland* (University of Helsinki, Publi-cations of the Department of Geography. No. 35. . Vammala, Finland: Vammalan Kirjapaino Oy, 1963). 23 They must be available i n a standard form for a large number of c i t i e s , and they must be presented i n a form that allows them to be grouped into classes which can be equated with specific urban functions. Comprehensive sta t i s t i c s covering a l l major c i t i e s i n a state are found only i n the national censuses, consequently most functional classifications of c i t i e s have been based on census data. A question arises as to what type of census material should be used i n measuring the importance of a function in a city relative to the other functions. Several possible bases are available, among them, figures on employment and value added. Discussion usually surrounds the choice of the most appropriate data for measures and often no agreement i s reached among individuals on which are the best. This problem of selecting bases for measurement has been somewhat ameliorated by results of a study carried out by J.W. Alexander and J.B. Lindberg. ° These results suggest that there i s no significant difference among most measures available i n the census, including data on employment and value added, for rating functional ac t i v i t y . Employment figures have been used almost exclusively in city functional studies. This has occurred pz: J.W. Alexander and J.B. Lindberg, "Measurements of Manufacturing: Coefficients of Correlation," Journal  of Regional Science, III (Summer, 1961), 71-81.. i 2 4 primarily because data on value added are not as compre-hensive and detailed as those on labour force, and labour force stat i s t i c s are more easily related to population figures. Two sets of labour force stat i s t i c s are available in the censuses of most industrialized countries. One set, collected at place of residence and giving the number of workers classified by industry and occupation, usually appears i n volumes of the Census of Population. The other set i s collected from manufacturing and commercial establishments and gives the number of workers by place of work classified by industrial groupings. The latter s t a t i s t i c s normally appear in censuses of manufacturing and commerce which are often carried out annually. Most quantitative studies of city functions have ut i l i z e d labour force stat i s t i c s for place of residence classified by industry. Although an element of error i s introduced by using figures collected by place of residence rather than place of employment, these figures are compre-hensive, covering a l l a city's labour f o r c e . 2 7 Statistics 2 7By using stat i s t i c s collected by place of residence two assumptions are made: (1) most workers are employed i n the city where they l i v e , and (2) the character of the labour force commuting to jobs within the city from surrounding areas i s similar to that of the labour force residing in the c i t y . The errors introduced when using these s t a t i s t i c s have been shown to be insignificant for the most part. See: J.F. Hart, Functions and Occupational Structures of Cities of the American South," Annals of the Association of American  Geographers. XLV (September, 1955), 271. 25 given in censuses of manufacturing and commerce normally cover only the labour force in industry and trade; usually omitted are data on government and some professional groups. Statistics classified by industrial groups, rather than by occupation, are preferred in city functional study because census industrial categories are more readily equated with specific functions. For example, i t i s possible to obtain directly from the census tables on labour force classified by industry, the number of workers in manufacturing, wholesale trade, transportation and other functions. Census data grouped by occupation do not allow this ready conversion of census categories into city functional categories. Included for summary purposes are Tables I and II il l u s t r a t i n g , respectively, the kinds of source materials u t i l i z e d i n cit y functional studies, and a conversion of census industrial categories into city functional categories. Methods of identifying functional importance. The methods developed to identify distinctive and dominant functions in city classification studies can be grouped into.three broad categories: (1) those based on empirical observations, (2) those based on s t a t i s t i c a l measures such as means, and (3) those based on "minimum requirements." 26 TABLE I EXAMPLES OF SOURCE MATERIALS UTILIZED IN QUANTITATIVE STUDIES OF CITY FUNCTIONAL STRUCTURE Author of Publication Study Year Source Material Harris 1943 United States Bureau of the Census, Fifteenth Census of the United  States; 1930* Population"; Vol. IV, "Occupations" (Wash-ington: Government Printing Office, 1 9 3 3 ) . United States Bureau of the Census, Census of Business: 1935. Retail Distribution, Vol. II, "County and City Summaries'* (Washington: Government Prin-ting Office, 1 9 3 6 ) ; Wholesale  Distribution, Vol. I l l , "Cities and Counties',' (Washing-ton: Government Printing Offiee, 1 9 3 7 ) . United States Bureau of the Census, Biennial Census of Manufac- turers : 1935 (mimeographed press release for each state). Note: Because occupation figures were not classified by industry in the 1930 U.S. Census of Population, (for example, employment in trade was not broken down into wholesale and r e t a i l trade), Harris used employment figures, classified by industrial groups, from the Census of Business and Manufacturers as his main basis of c l a s s i f i c a -tion. aC.D. Harris, "A Functional Classification of Cities i n the United States," Geographical Review. XXXIII (January, 1 9 4 3 ) , 86-99. TABLE I (continued) 27 Author of Publication Study Year Source Material Pownall 1953 New Zealand Department of Labour and Employment, Location and Decen- tralization of Industry—  District Office Returns t A p r i l t  1950* (Wellington, New Zealand: Department of Labour and Employment), Note: Twice yearly (April and October) the New Zealand Department of Labour and Employment collects sta t i s t i c s stating the number of workers gainfully employed i n 69 different industrial codes, Pownall has used the April 1950 s t a t i s t i c s as the basic material for his c l a s s i f i -cation scheme, Trotier e 1959 Dominion Bureau of Stati s t i c s , Ninth Census of Canada: 1951. Labour Force, Vol. IV (Queen's Printer, 1953). Note: The Census of Canada gives labour force s t a t i s t i c s classified by industry and occupation for urban centres of 10,000 and over. Similar data on centers 1,000-10,000 population are available on request, Trotier based his classification on these labour force stat i s t i c s classified by industry. bL.L. Pownall, "The Functions of New Zealand Towns," Annals of the Association of American Geographers. XLIII -December, 1953), 33W$0. ~ — -CL, Trotier, "Some Functional Characteristics of the Main Service Centers of the Province of Quebec," Cahiers de  Geographie de Quebec. No. 6 (April-September, 1959), 243. 28 TABLE I (continued) Author of Study Publication Year Source Material Hart<* 1955 Nelson e f 1955 Alexandersson1 1956 MorrissetS 1958 Webbn . 1959 Ullman & Dacey1 I960 United States Bureau of the Census, Seventeenth Census of the United  States: 1950. Population, Vol. II "Characteristics of the Population" (Washington: Govern-ment Printing Office, 1952). Note: The 1950 U.S. Census of Popu- lation gives labour force figures classified by industry as well as occupation. This has allowed classification schemes to be constructed on this one source. No supple-mental data are required as was the case when Harris developed his functional classi f i c a t i o n . dJ.F. Hart, "Functions and Occupational Structures of Cities of the American South," Annals of the Association of  American Geographers, XLV (September, 1955), 268-286. eH.J. Nelson, "A Service Classification of American Cities.".Economic Geography. XXXI (July, 1955), 189-210. ^G. Alexandersson, The Industrial Structure of American Cities (Lincoln, Neb.: University of Nebraska Press, 1956). S i . Morrissett, "The Economic Structure of American Cities , " Papers and Proceedings of the Regional Science  Association, IV (1958). 239-256. ' kj.W. Webb, "Basic Concepts in the Analysis of Small Urban Centers of Minnesota," Annals of the Association of American Geographers. XLIX (March, 1959), 55-72. AE.L. Ullman and M.F. Dacey, "The Minimum Require-ments Approach to the Urban Economic Base," Papers and  Proceedings of the Regional Science Association, VI (I960), 175-194. 29 TABLE II EXAMPLEa OF CONVERSION OF CENSUS INDUSTRIAL CATEGORIES INTO CITY FUNCTIONAL CATEGORIES Census Categories by Industry Groups Functional Categories Agriculture, Forestry, and fisheries Mining Construction Manufacturing Railroads & railway express ) service ) Trucking service & warehousing) Other transportation ) Telecommunications ) U t i l i t i e s & sanitary service Wholesale trade Food & dairy produce stores, ) & milk r e t a i l ) Eating & drinking places ) Other r e t a i l trade ) Finance, insurance & real estate Business services) Repair services ) Private households Hotels & lodging places ) Other personal service ) Entertainment & recreation ) Medical & other health services Educational services, government Educational services, private Other professional & related services Public administration Industry not reported Omitted Mining Omitted Manufacturing Transportation and communication Omitted Wholesale trade Retail trade Finance, insurance & real estate Omitted Omitted Personal service Professional services Public administration Omitted aExample i s taken from H.J. Nelson, "A Service Classi-f i c a t i o n of American Cities," Economic Geography. XXXI (July, 1955), 190. Census categories are from the 19^0 U.S..Census of Population. 30 Representative of methods in category one i s the one devised by CD. Harris. 2** Harris' method calls for the examination of the labour force st a t i s t i c s of ci t i e s "of well recognized types." 29 On the basis of the percentage breakdown of the labour force employed i n various indus-t r i a l categories in these c i t i e s , he established specific percentage values for different functions which must be equalled or exceeded for the function to be considered dominant. For example, a city's manufacturing employment must be at least 60 per cent of the total employment i n manufacturing, retailing, and wholesaling before manufac-turing i s considered the dominant function of the cit y . For wholesaling to be a dominant function, however, only 20 per cent of the labour force i n manufacturing, retai l i n g , and wholesaling must be employed in wholesaling. 3 0 As Harris was interested primarily i n classifying c i t i e s by dominant function and only incidentally interested in the •^Harris, op. c i t . . 86-39. 2^Ibid., 87. Unfortunately Harris did not identify the c i t i e s "of well recognized types'* that he selected and the c r i t e r i a he uses to classify c i t i e s must be accepted on f a i t h . The reputation that this classification study enjoys suggests that his choice was sound; however, this method of determining c r i t e r i a for identifying functional importance i s essentially subjective because of the judgment involved i n selecting type c i t i e s . 3 0 I b i d . . 88. 31 relative importance of a l l functions in the city functional pr o f i l e , he provided no means for evaluating functional relative importance within or among c i t i e s . Other analysts using methods similar to Harris 1 include Kneedler, Jones, and Hart. 3 1 The methods used by Pownall, Nelson, Steigenga and Webb are representative of those in category two.^2 These methods are based wholly on s t a t i s t i c a l measures. Usually average percentage values of employment i n various census industrial categories are calculated from figures for a l l c i t i e s under study and these values are used as a "normal 31G.M. Kneedler, "Economic Classification of Citie s , " The Municipal;Year Book, 1945 (Chicago: The International City Manager's1 Association, 1945), pp. 30-33, 48-68; V. Jones, "Economic Classification of Cities and Metro-politan Areas." The Municipal Year Book, 1954 (Chicago: The International City Manager's Association, 1954), pp. 31-36, 62-70, 81-108; J.F. Hart, "Functions and Occupational Structures of Cities of the American South," Annals of the Association of American Geographers, XLV (September, 1955), 269-286. 32 L.L. Pownall, "The Functions of New Zealand Towns." Ann. Assoc. Am. Geog.. XLIII (Dec, 1953), 332-350; H.J. Nelson, "A Service Classification of American Cities, "Econ. Geog.. XXI (July, 1955), 189-210; W. Steigenga, ?'A Comparative Analysis and a Classification of Netherlands' Towns," Tiidsckrift  voor Economische en Sociale Geografie. XLVI (19$5), 105-19, cited by J.W. Webb, "Basic Concepts i n the Analysis of Small Urban Centers of Minnesota," Ann. Assoc.  Am. Geog.. XLIX (March, 1959), 55; J.W. Webb, "Basic Concepts i n the Analysis of Small Urban Centers of Minnesota." Ann. Assoc. Am. Geog.. XLIX (March, 1959), 55-72. 32 structure" against which the employment structure of individual c i t i e s i s compared. In some studies national percentage figures for labour force employed i n various a c t i v i t i e s are selected as the "normal structure". The positive deviations of each city's percentage values from the "normal" or "expected" values for the respective industrial categories are measured to determine the relative importance of the various a c t i v i t i e s . Distinctive a c t i v i t i e s are usually determined by ranking city percentage values for a c t i v i t i e s according to their degree of deviation from the "norms"; the activity with the largest degree of deviation i s considered the most distinctive function in the city. Several refinements i n methodology have been introduced in these studies. Pownall and Webb, in developing their "normal structures", have taken city size into consideration. They grouped the ci t i e s under study by size classes and calculated average percentage values for employment i n census industrial categories for each of the city-size classes. Percentage values for individual ci t i e s were then compared with the average values of the appropriate city-size group. The division of c i t i e s into size groups was made on the assumption that city functional structure i s modified with changes in city 33 s i z e — a city of 10,000 population cannot be expected to have a functional role similar to that of a city of 100,000 population. This assumption has been proven to be correct, 3 Another refinement was provided by Steigenga and Nelson with their introduction of the standard deviation in the measurement of the deviation of city values from the "normal" or "average" values. The use of the standard deviation permits the comparison of the degree of functional variation, thus f a c i l i t a t i n g greatly the comparison of the relative importance of different functions in and among c i t i e s . The methods in category three have been used in the most recent functional studies of c i t i e s , and are out-growths of the economic base theory. One of the major points of the base theory i s the breaking down of city functional activity into two components: basic and service sectors. In most of the early economic base studies attempts were made to identify what ac t i v i t i e s belonged i n each sector. Often a tedious job was involved and results were not always satisfactory. Later students, the f i r s t to publish in English being Alexandersson, developed a new tact in u t i l i z i n g the theory. While recognizing the dichotomous nature of city functional activity as stated i n the economic 33Morrissett, op. c i t . . 2A9-250 34 base theory, these students, rather than trying to allocate individual a c t i v i t i e s to either basic or service sectors, focused their attention on the question, what minimum proportions of employment in urban acti v i t i e s are required to keep the city viable? Alexandersson stated the question as, ....what ratios in different industries are a necessary minimum to supply a city's own population with goods and services of the type which are produced in every normal city?34 This approach to city functional study has been called the "minimum requirement approach". The minimum requirement of a city may be equated with i t s service or nonbasic sector. It i s the "city-serving" sector. When using this approach to develop methods for identifying distinctive and dominant functions, the "minimum requirement" for each urban activity being studied must be calculated. These minimum values provide a "base structure" against which the values of individual c i t i e s can be compared. By measuring each city's degree of "excess" activity over and above the minimum values, a measure of "city-forming" activity i s obtained, as well as a measure of functional importance. Alexandersson determined the minimum requirements ^^"Alexandersson, op. c i t . , p. 17. 35 —he termed them "k-values"—for census industrial categories by calculating the percentage of city labour force in the respective industrial categories and arraying each category's city values in increasing order. The city value at the 5th percentile in each of the arrays was then selected as the minimum requirement or "k-value". This value was selected rather than the lowest value in the array "to avoid extreme ratios representing such agglomerations • ••• [which] are not towns in the ordinary sense".3'' Alexandersson wished to avoid dormitory c i t i e s and other abnormal centers so that his "k-values" would, as nearly as possible, be representative of "normal" c i t i e s . Once the minimum requirements ("k-values") were calculated, the "excess" was determined i n each activity for a l l ci t i e s under study. Activities were rated i n terms of "excess" value, the acti v i t y experiencing the greatest excess being considered the dominant function. A refinement to Alexandersson's work was contributed by Morrissett who investigated the effect of cit y size on minimum requirements. Using Alexandersson's method, Morrissett calculated a series of "k-values" for cit i e s grouped by city-size classes. His work demonstrated that 3 5 I b i d . eity size has a real effect on city functional structure and should be taken into consideration when the minimum requirement of ci t i e s i s being c a l c u l a t e d , H e observed that: (1) small c i t i e s are much less diversified than large c i t i e s . (2) as population increases, diversification increases and specialization decreases. (3) the proportion of persons employed i n most industries i s higher i n large c i t i e s than i n small. That i s , the minimum requirement increases with city size, (4) differences i n functional structure i s greater between small- and medium-sized cities than between medium-sized and 37 large c i t i e s . Further refinements of Alexandersson1s technique 38 have been made by Ullman and Dacey, Their method not only takes city size into consideration when calculating minimum requirements, but also provides a means of using 3 6 x r o t i e r , i n his study of Quebec centers, used Morrissett's "k-values" i n determining the relative importance of urban functions. Trotier, loc. c i t . ^^Morrissett, loc. c i t . 3%llman and Dacey, loc, c i t . 37 empirical data to determine the minimum requirement values. They selected at random a standard number of c i t i e s within specified city-size classes and calculated the percentage values for each city's labour force employed in fourteen census industrial categories. Within each size class the lowest city percentage value for each industrial category was noted. Once the minimum city values for an industrial category were known for each size class, they were plotted on semi-logarithmic paper against the logarithm of city population and a regression line was calculated for the points. The regression lines are used to read off the minimum requirement values i n each acti v i t y for c i t i e s of a l l sizes. This method would appear to be superior to Alexandersson's because i t avoids the necessity of arbitrary choice i n the selection of minimum requirement values—Alexandersson a r b i t r a r i l y chose the 5 t h percentile values i n his arrays of city values. The use of regression lines, rather than Alexandersson*s "k-values", for determining minimum requirement values i s especially desirable when only a small number of c i t i e s i s being studied. If Alexandersson's method i s used, the 5 t h percentile i n the arrays of city values w i l l often be the lowest value in the arrays and could well belong to 3* a city with an abnormal structure. In addition, Ullman and Dacey's method allows minimum requirement values to be calculated for each city on the basis of the city's exact population size rather than on the basis of city-size classes where broad ranges i n city population are encountered, as in Morrissett's work. City-type classes. In a l l classification problems the question arises as to what c r i t e r i a should be used to classify individuals and what kinds of categories should be established. In city functional classification a concensus appears to have been reached regarding the kinds of c r i t e r i a that should be used. Most analysts have used distinctive or dominant functions and/or degree of city functional specialization to classify c i t i e s . City-type classes, however, have varied greatly in kind and number among studies. This multiplicity has resulted from the necessity of basing classes on the city functions recognized, which in turn are based on census industrial categories. The specific city a c t i v i t i e s recognized, and hence the city-type classes formulated, depend on how the census industrial categories are subdivided and combined. As many as thirty-six census industrial categories have been recognized as separate city functions and used as 3 9 city-type classes (Alexandersson); i n contrast, W, William-Olsson, using six census industrial categories, formulated only three city-type classes. Despite the apparent diversity, systems of city-type classes can be characterized as belonging to either one of two types; systems using mutually exclusive c i t y -type classes where a city appears i n only one of the established categories, or systems using non-exclusive city-type classes where a city appears in as many categories as i t has distinctive functions. The classification studies u t i l i z i n g methods similar to those devised by Harris have been the major users of the mutually exclusive city-type classes, while investigators such as Alexandersson, Nelson and Webb have been advocates of non-exclusive city-type classes. The classification systems using mutually exclusive classes, although offering the most compact and convenient form of classification, have one major disadvantage. Because c i t i e s appear i n one city-type class only in these classification systems, the implication i s made that c i t i e s are unifunctional when i n fact they are essentially multi-functional in character, J«W, Webb has stated the criticism of this approach concisely. 40 The f i n a l i t y of such terms as manufacturing town, r e t a i l trade center, and mining town seems to exclude the possibility that a mining town may also have a significant function as Qa r e t a i l trade center or as a manufacturing town.^ To avoid this problem most of the more recent studies have used non-exclusive city-type classes, thus recognizing that a city may be classified in terms of several speciali-t i e s . Cities are characterized by l i s t i n g their distinctive functions i n order of relative importance and, in some studies, grouping c i t i e s by degree of functional special-ization. Table III gives examples of the two basic classification systems that have been used in city functional classification studies. A critique of quantitative studies i n city functional  classi f i c a t i o n. The general methodology in city c l a s s i -f i c a t i o n i s clear. It i s the objective of the classification process to identify what i s "distinctive" i n the functional profile of a city i n comparison with other c i t i e s , to determine the dominant functions of c i t i e s , and to relate city functions to locations. As indicated in the preceding section, this has been done i n most studies by (1) establishing "standards" of city functional structure 39webb, op. c i t . , 55 41 TABLE III EXAMPLES OF CITY-TYPE CLASSES USED IN CITY FUNCTIONAL . CLASSIFICATION STUDIES Mutually exclusive city-type Non-exclusive city-type classes classes (a city appears i n one (a c i t y may appear i n as many category only) categories as i t has distinctive functions) Harris a Nelson** Manufacturing ci t i e s (2 sub Manufacturing classes divided on basis of manufacturing specialization) Retail centers Retail trade Diversified c i t i e s 0 Diversified 0 Wholesale centers Wholesale trade Transportation centers Transportation & communication Mining towns Mining University towns Resort and retirement towns Professional service Personal service Public administration Financing, insurance and real estate aC.D. Harris, "A Functional Classification of Cities i n the United States," Geographical Review, XXXIII (January, 1943), 88. DH.J. Nelson, "A Service Classification of American Ci t i e s , " Economic Geography. XXXI (July, 1955), 196. c a residual category i s usually present i n both systems of city-type classes. Cities which have no distinctive functions according to the definitions formulated in a study to determine functional importance are placed i n the residual category. They are often referred to as diversified c i t i e s since no one function appears to be of major importance i n them. 42 against which the functional structure of individual c i t i e s i s scrutinized, and (2) using measures^ of functional importance and specialization which are based on the established "standards", to classify c i t i e s into city-type classes. When carrying out these two steps certain qualities of city character should be considered i f the classification scheme i s to be effective. Studies of urban character have revealed that function i s the essential element of the city; that most ci t i e s are multifunctional; that c i t i e s have a c t i v i t i e s present in their functional profiles i n different proportions; that city a c t i v i t y can be broken down into two components: basic and service. It has been noted that the service a c t i v i t i e s , with few exceptions, exist i n a l l ci t i e s i n relatively constant proportions. When exceptions do occur and an abnormally high proportion of a city's employment is i n one of the service a c t i v i t i e s , i t i s usually found that the activ i t y i s operating in the capacity of a basic activity and not in i t s customary role of "city-server". ^The units i n these measures have been proportions of c i t y employment in specified a c t i v i t i e s . It has been the practice in a l l studies to equate functional importance with proportions of city labour force employed in various census industrial categories. The conceptual aspects of this practice have been questioned, but i n light of the findings of Alexander and Lindberg (loc. c i t . ) , and in the absence of better procedures, this practice i s accepted as the best means available for developing measures of functional importance. 43 The proportions of the different basic a c t i v i t i e s in c i t i e s have been shown to vary greatly among centers being reflections of the functional relationship existing between individual c i t i e s and their hinterlands—local, national, and international. It has also been observed that city size has a direct bearing on city functional structure. These observations suggest that the "best" c l a s s i -fication scheme w i l l : (1) use only basic act i v i t y in characterizing city functional profiles since only the basic activity reveals the relationship existing between city function and location. (2) u t i l i z e "standards" of city functional structure that: (a) have some basis in theory, and (b) are designed to compensate for changes i n city size. (3) use measures of functional importance that are based on relative rather than absolute values in order that an unbiased comparison of functional importance in and among citi e s i s made possible. (4) use a system of city-type classes that provides for the recognition of the multifunctional character of ci t i e s by allowing c i t i e s to be classified in terms of several specialities. 44 The "standards" of city functional structure that have been established i n classification studies include empirically derived structures (Harris), s t a t i s t i c a l measures (Pownall, Nelson, Webb), and minimum requirement profiles (Alexandersson, Ullman and Dacey). The empirical structures were the f i r s t to be used. They have the advantage of being taken directly from reality and in this respect enjoy an advantage over s t a t i s t i c a l l y derived profiles. This kind of standard, however, i s only as good as the analyst's judgment i n selecting c i t i e s representative of different functional types. Furthermore, using i t and the associated system of city-type classes means that a city i s characterized by i t s dominant function only; no account i s taken of other functions that may be distinctive i n the city functional profile. This procedure tends to negate the concept that c i t i e s are multifunctional. In addition, no allowance appears to have been made for differing city sizes in the studies using this kind of "standard". The most frequent s t a t i s t i c a l measure to be used as a "standard" of functional structure has been the arithmetic mean. The mean enjoys the advantages of being easily and objectively determined, and the standard deviation, the best measure of deviation, i s based on the 45 mean. Using the mean, analysts have been able to take city size into consideration by calculating means for a series of city-size classes and comparing a city's values for functional activity with the appropriate class means. By using deviation from the mean values of city a c t i v i t y as a measure of functional importance, c i t i e s have been c l a s s i -f i e d i n terms of a l l their distinctive functions rather than by dominant function alone, thus recognizing the multi-functional concept. When using deviation from the mean as a measure of functional importance, relative rather than absolute values should be used. Pownall, i n his study of New Zealand towns,^x took city size into consideration by using mean values calculated for several city-size classes as his "standards", but he used absolute values of deviation to measure functional importance, thereby precluding effective comparison of functional activity i n and among c i t i e s . This was improved on i n such studies as Nelson*s42 where the standard deviation was used to measure functional importance, allowing; the relative values of importance to be determined. Nelson, however, did not allow for changes i n city size; he used one set of mean values for a l l c i t i e s studied regardless 41pownall, loc. c i t . i o Nelson, loc. c i t . 46 of population. J.W. WebtA3 has also used mean values as standards. He developed weighted location quotients^ to get functional and specialization indices. These indices incorporated two distinct values: (1) the importance of the function i n a city relative to i t s importance i n a l l c i t i e s , and (2) the importance of the function in a city relative to the importance of other functions in the city . Although representing new uses of location quotients, these indices are of limited use as interpretative devices because of their complexity. The fact that two values are combined in the indices makes i t impossible to ascertain the position of individual functions i n cit i e s without referring to the data from which the indices are derived. The most serious technical d i f f i c u l t y i n using the mean as a "standard" of city functional activity i s associ-ated with activities whose distributions of relative ^Webb, loc. c i t . ^Location quotients, taken i n the context of Webb's study, are devices for comparing the importance of a function i n an individual city with i t s importance in some basic aggregate (in the case of Webb's study, the aggregate was made up of mean values of functional activity i n the c i t i e s studied). Relative measures are used. In essence a location quotient i s a ratio of ratios. For example: Number of persons employed in given industry i n the city Total number of persons employed i n the city Number of persons employed i n given industry in a l l c i t i e s studied  Total number of persons employed i n a l l c i t i e s studied 47 importance are highly sporadic. For these functions the means are misleading as "standards". For example, an activi t y that i s not represented in 70 out of 80 ci t i e s under study, but has extremely high percentage values i n the remaining 10 c i t i e s w i l l have an unrealistically high value for a "standard". The 70 cit i e s without the activity w i l l be shown to be deficient i n the function when i n fact they may represent the "normal" situation. The major conceptual disadvantage of the mean when used as a '•standard" of functional activity i s that i t does not have a foundation i n re a l i t y as does the empirically derived structure, nor a grounding in theory as has the minimum requirement profi l e . There i s no reason for believing that the mean structure i s the most typical or normal functional p r o f i l e . In essence, i t i s a s t a t i s t i c a l abstraction which i s d i f f i c u l t to interpret when used in this context. Of the three kinds of "standards" established only the minimum requirement profile i s based directly on theory. In the studies u t i l i z i n g this "standard", a conscious attempt has been made to use only basic activity to charac-terize city functional structure; a l l other studies have used total city activity. 48 The f i r s t attempts to u t i l i z e minimum requirement profiles involved the use of relatively crude techniques. Alexandersson had to resort to arbitrary methods to determine his "k-values", and when measuring functional importance he used absolute rather than relative values.^5 in addition, although noting that city size affected the minimum requirement profiles of c i t i e s , his analytical method does not allow for changing city size. Despite the limitations of his techniques, Alexandersson was s t i l l able to classify ^Alexandersson used fixed cut-off points of excess employment to rank a c t i v i t i e s by classes of importance, and ignored the variable quality of his base unit—the "k-values". He developed three classes of functional importance based on the amount of excess employment over the minimum require-ment: "k" + 20.0 per cent, "k" + 10.0-19.9 per cent, and "k" + 5.0-9.9 per cent. Since "k-values" varied from 0.0 per cent to 8.0 per cent for different a c t i v i t i e s , his classes of functional importance discriminate against ac t i v i t i e s having only small percentages of city labour force even when highly concentrated i n a ci t y . ,For example, the "k-value" of activity A i s 1.0 per cent of city labour force and a city has 5.0 per cent of i t s labour force i n activity A. In this instance the excess employment i s 4 per cent of ci t y labour force or 4.00 times greater than the "k-value". In activity B the city has 18 per cent of i t s labour force, and the "k-value" for this activity i s 8.0 per cent; therefore excess employment i s equal to 10.0 per cent of city labour force or 125 times the "k-value". Despite the fact that the degree of concentration of activity A in the city i s greater than that of activity B, "A" i s not registered as a distinctive function because the "excess" employment does not exceed the "k-value" by 5.0 per cent of the city labour force. Activity B, although not as concentrated in the ci t y as "A" i n relative terms, has excess employment of 10 per cent thus placing the city i n the second of the three classes of functional importance in terms of this function. 49 c i t i e s in terms of several specialities rather than by the dominant function alone. As noted i n an earlier section, Morrissett improved on Alexandersson's method by calculating "k-values" for cit i e s grouped by size classes. By studying trends in "k-values" he was able to make several generalizations regarding the relationship between city size and functional structure. Although making a significant contribution i n this study, Morrissett used Alexandersson's crude technique of arbi-t r a r i l y selecting certain points in the arrays of city values to act as "k-values" or minimum requirement values. It was not u n t i l Ullman and Dacey's paper appeared that a f a i r l y objective and rational means of determining minimum requirements was made available. The method presented in the paper conforms with most of the requirements associated with the "best" classification scheme. Only a procedure for evaluating the importance of a function i n a city relative to i t s importance in a l l c i t i e s i s omitted. Such a procedure can easily be developed by extending the use of some of the authors' data. Like Webb, Ullman and Dacey developed an index of specialization which allows c i t i e s to be ranked by degree of functional specialization, thus allowing a correlation to be made between kinds of distinctive and dominant functions and specialization. 50 To summarize i t may be said that none of the studies cited has included a l l the best methods of analysis that have been presented; however, when they are taken together, the methodology i s f a i r l y complete. By selecting the best elements in each method i t i s possible to approximate most of the conditions l a i d down for a "best" classification scheme, , In the next section i s presented an outline of the method used in this study. It represents a partial synthesis of the methods described in this section, V. METHOD OF STUDY Cities studied. In 1951 there were 106 incorporated urban centers i n Canada with populations of ten thousand and over. Many of these municipalities are in Census Metropolitan Areas and others, although not of metropolitan rank, are contiguous. By recognizing Census Metropolitan Areas as individual cities and grouping other contiguous municipalities to form single units, the total number of cit i e s for this study was reduced to eighty. The cluster-ing of adjoining municipalities allowed "geographical" rather than "legal" c i t i e s to be used, thereby obtaining a better approximation of "functional" centers. Unincorporated urban centers of ten thousand population and over, as well as a l l c i t i e s under ten 51 thousand population were omitted from this study because the labour force st a t i s t i c s necessary for functional study are not published for them. Kirkland Lake (unincorporated; 1951 population: 18,392) i s the only large center affected by this lack of data. Table IV gives the population and constituent municipalities or Census Metropolitan Area for a l l c i t i e s studied, and Figure 1 shows city location. Source materials. Canadian census labour force st a t i s t i c s for place of residence classified by industry were u t i l i z e d exclusively as source material i n this study.^ They represent the most comprehensive data available for city functional study. In addition, the form i n which the statist i c s are presented allowed some of the key values derived in American studies of city functional structure to be u t i l i z e d i n analyzing Canadian c i t i e s . Labour force s t a t i s t i c s from the 1951 Census of  Canada were used because the required labour force data for the 196.1 Census of Canada were not available at the time the study was made. Since the emphasis of the project was on methodology as well as on fact finding, the use of somewhat dated source material was not thought to be a 46nominion Bureau of Statistics, Ninth Census of Canada: 1951* Labour Force, Vol. IV, Tables 17 and 21 (Ottawa: Queen's Printer, 1953). 52 TABLE IV CITIES EXAMINED: POPULATION, LOCATION, AND LEGAL CONSTITUENTS C i t y a Population (1951) Location Component Municipali-ties or Census Metro-politan Areas (CM.A,-) Arvida 11,073 northern Arvida periphery Barrie 12,514 heartland Barrie B e l l e v i l l e 19,519 20,593 t? Belleville Brandon western Brandon periphery Brantford 36,727 heartland Brantford Brockville 12,301 n Brockville Calgary 139,105 western Calgary CM.A. 15,387 periphery Charlottetown eastern Charlottetown periphery Chatham 21,218 heartland Chatham Chicoutimi 23,216 northern Chicoutimi 16,899 periphery Cornwall heartland Cornwall Drummondville 14,341 tt Drummondville Edmonton 173,075 western periphery Edmonton CM.A. Edmundston 10,753 eastern Edmundston 16,018 periphery Fredericton eastern periphery Fredericton Gait 19,207 25,536 heartland Gait Glaee Bay eastern Glace Bay periphery Granby 21,939 heartland Granby Grand'Mere 11,089 tt Grand'Mere Guelph 27,336 tt Guelph Halifax CM.A.V Halifax 133,931 eastern periphery a"Geographical" rather than "legal" c i t i e s have been used where possible. Census Metropolitan.Areas are considered as single c i t i e s , as are clusters of contiguous municipalities. *CM.A. for which unpublished data were supplied by the Dominion Bureau of Statistics. 53 TABLE IV (continued) City Population Location Component Municipali-(195D ties or Census Metro-politan Areas (C.M.A.) Hamilton 259,685 heartland Hamilton C.M.A. Joliette 16,064 n Joliette Jonquiere 21,618 northern Jonquiere • periphery Kingston 33,459 56j853 heartland Kingston Kitchener- n Kitchener, Waterloo Waterloo Lethbridge 22,947 western Lethbridge 121,516 periphery London heartland London C.M.A. Magog 12,423 t* Magog Medicine Hat 16,364 western Medicine Hat periphery Moneton 27,334 eastern Moncton periphery Montreal 1,395,400 heartland Montreal C.M.A. Moose Jaw 24,355 western Moose Jaw periphery New Waterford • 10,423 eastern New Waterford 22,874 periphery Niagara Falls • heartland Niagara Falls North Bay 17,944 northern North Bay periphery O r i l l i a 12,110 heartland O r i l l i a Oshawa 41,545 281,908 n Oshawa Ottawa tt Ottawa C.M.A. Owen Sound 16,423 tt Owen Sound Pembroke 12,704 n Pembroke Pentictpn 10,543 western Penticton periphery Peterborough 38,272 heartland Peterborough Fort William- 66,108 Western Port Arthur, Port Arthur periphery Fort William Prince Albert 17,149 western Prince Albert periphery Quebec 274,327 heartland Quebec C.M.A. Regina 71,319 western Regina 11,565 periphery Rimouski eastern Rimouski periphery Rouyn 14,633 northern Rouyn periphery 54 TABLE IV (continued) City Population Location Component Municipali-(195D ties or Census Metro-politan Areas (C.M.A.) St. Catharines 37,934 heartland St. Catharines St. Hyaeinthe 20,236 n St. Hyaeinthe St. Jean 19,305 t» St. Jean St. Jerome 17,685 n St. Jerome (Terrebonne) St. John's 67,749 eastern St. John's C.M.A. periphery St. Thomas 18,173 heartland St. Thomas Saint John 73,337 eastern Saint John C.M.A,* periphery Sarnia 34,697 heartland Sarnia Saskatoon 53,268 western Saskatoon periphery Sault Ste. Marie 32,452 northern Sault Ste. Marie 26,903 periphery Shawinigan heartland Shawinigan Falls Sherbrooke 50,543 tt Sherbrooke Sorel 14,961 tt Sorel Stratford 18,735 n Stratford Sudbury 42,410 northern Sudbury periphery Sydney 31,317 eastern Sydney periphery Thetford Mines 15,095 heartland Thetford Mines Timmins 27,743 northern Timmins periphery Toronto 1,117,470 heartland Toronto C.M.A. Tra i l 11,430 western T r a i l 10,085 periphery Trenton heartland Trenton Trois-Rivieres 64,741 tt Trois Rivieres, .. Cap-de-la-Madeleine Truro 10,756 eastern Truro periphery Valleyfield 22,414 heartland Valleyfield (Salaberry-de) Vancouver 530,728 western Vancouver C.M.A. periphery Victoria 104,303 western Victoria C.M.A.* periphery 55 TABLE IV (continued) City Population Location Component Municipali-(1951) ties or Census Metro-politan Areas (C.M.A.) Vic t o r i a v i l l e We Hand Windsor Winnipeg Woodstock 13,124 15,332 157,672 354,069 15,544 heartland western periphery heartland Vi c t o r i a v i l l e Welland Windsor C.M.A. Winnipeg C.M.A. Woodstock Figure 1. Canadian cities of 10,000 population and over* 1951. 5 7 serious limitation. The facts revealed in the study can be considered as "cross-sectional" views; they reveal the functions of Canadian c i t i e s at a particular point in time. The 1 9 5 1 Census of Canada breaks down the services performed by the labour force into twelve industry divisions which are further broken down into major groups. The f i r s t column of Table V l i s t s a l l the census industry divisions and major groups which are relevant for this study. The second column of the table presents the city functions derived from the census categories for use in the study. Those census categories representing "rural functions" (functions which require broad surfaces for their performance) were omitted as urban a c t i v i t i e s since their inclusion in a city's employment structure depends on the arbitrary positioning of city boundaries. The "Not Stated" category was also omitted. The remaining census categories, and the respective functional classes derived from them, represent broad groups in the employment structure of the city. Although highly aggregated these categories are sufficiently detailed to distinguish in reasonably specific terms the functional character of c i t i e s . Also, by using these categories i t was possible to draw on the techniques and findings of American studies in determining functional profiles because the classes can be equated with 53 TABLE V CENSUS INDUSTRIAL CATEGORIES AND CITY FUNCTIONAL CATEGORIES. USED IN THE STUDY . Census Industry Divisions & Major Groups City Functions .significant to the study a recognized i n the study Agriculture. • • . . . Omitted Forestry and Logging Omitted Fishing and Trapping Omitted Mining, Quarrying and O i l Wells. . . . . Extraction Manufacturings . . . . Manufacturing E l e c t r i c i t y , Gas, and Water Public U t i l i t i e s Construction . . . . . • Construction Transportation, Storage, Communication . Transportation Trade Wholesale Trade Wholesale Trade Retail Trade . . Retail Trade Finance, Insurance, and Real Estate. . . Finance, Insurance and Real Estate Service ' K Community Service 0 Community Service Government Service 0 • Government Service Recreation . . . . . . . Recreation Business Service" Business Service Personal Service e Personal Service Not Stated ..' . . . . . . . . . Omitted aDominion Bureau of Statistics, Ninth Census of Canada: 1951. Labour Force, Vol. IV, Tables 1? and 21 (Ottawa: " Queen's Printer, 1953). ^"Community Service" includes labour force i n education, health, religion and welfare work. Includes government personnel and persons employed by private agencies doing work in health, welfare and educational f i e l d s . c"Government Service" includes labour force employed in work peculiar to public service only. ^"Business Service" includes labour force employed in accounting, advertising, engineering and sci e n t i f i c services, labour and trade organizations, law, and other business services. e"Personal Service" includes labour force employed i n barbering and hairdressing, dry cleaning and laundry, restau-rants, motels and hotels, private households, undertaking, photography, and other personal service. 59 those used in similar studies carried out in the United States. Methods of analysis. To meet the objectives of the study the analytical techniques selected had to make i t possible to: (1) determine the importance of a function i n a city relative to i t s importance in a l l c i t i e s . (2) determine the importance of a function relative to other functions i n a city . (3) determine the degree of functional specialization in a city. None of the studies discussed f u l f i l l s a l l these require-ments; however, by combining some of the techniques that have been developed, i t was possible to meet the three requirements enumerated. The Ullman-Dacey method of city functional study conforms most closely to theoretical concepts; i t was adopted as the basis of the method utili z e d in this study. Despite their advances these authors did not, as had earlier analysts, devise a measure of the importance of a function i n a city relative to i t s importance in a l l c i t i e s ; consequently, Nelson's use of the standard deviation was applied to achieve this end. The "standards" of city functional structure established for this study are based on Ullman and Dacey's 60 minimum requirement profiles. When possible, the minimum requirement values calculated by these authors were used i n this study as they are thought to be more representative of the "normal" situation than similar values for Canadian c i t i e s . Ullman and Dacey!s values are based on random samples of American c i t i e s and take into consideration a great many more c i t i e s than can be considered when Canadian c i t i e s form the basis of the values. In this study, as in a l l studies based on the minimum requirement approach, employment over and above the minimum employment necessary to keep the city viable i s equated with basic ac t i v i t y . Functional importance was determined by the amount of "excess" employment in a function. To determine the importance of functions in a city relative to their importance i n a l l c i t i e s , mean values of excess employment for a l l functions were calculated and the standard deviations obtained. The amount of excess employment for each function i n a city was then stated i n terms of standard deviations—a function might have an excess employment value less than one standard deviation, between one and two, or over two standard deviations above the mean. By ranking functions in terms of standard 61 deviations of excess employment, a technique for measuring functional importance using relative values was obtained. This allowed a comparison to be made of the relative importance of different functions among c i t i e s , and provided a way of determining a city's distinctive functions. This system i s similar to that used by Nelson in determining distinctive functions, except that only basic activity, rather than total city activity, was used in the measurement. Ullman and Dacey's method of identifying the dominant function in c i t i e s was used. The importance of functions relative to other functions in a city was determined by ranking a c t i v i t i e s according to their percentage share of total city excess employment. The function having the largest percentage share was presumed to be the dominant activity. City functional specialization was determined in relative terms by using a specialization index devised by Ullman and Dacey. This index i s based on the distribution of city excess employment among the various functions; the index value depends on the degree of "evenness" or "unevenness" i n the distribution. Summarizing, the methods outlined here in general form make possible the determination of a city's distinctive functions, dominant function, and degree of functional specialization, three characteristics which had to be determined to meet the objectives of the study. CHAPTER II THE RELATIVE IMPORTANCE OF URBAN FUNCTIONS IN CANADIAN CITIES Because ci t i e s are multifunctional i t i s desirable to study them not only in terms of their dominant functions, but also in terms of their relative strength i n a l l the functions ci t i e s perform. This i s best accomplished by examining systematically the role of urban functions i n a l l c i t i e s . It was the purpose in this chapter to develop a measure giving the importance of a function in a city relative to i t s importance i n a l l c i t i e s and, using this measure, to determine, describe, and analyze the distribution of relative importance for functions among c i t i e s . The methods of Ullman and Dacey, and Nelson, outlined in the preceding chapter, were u t i l i z e d to develop the measure. In the f i r s t part of the chapter, the measure of functional importance i s described. In the second section certain observations regarding the proportions of city labour force in different functions are made, and the reactions of functions to changes i n city size are noted. 64 The balance of the chapter deals with the distribution patterns of functional relative importance. I. A MEASURE OF FUNCTIONAL IMPORTANCE The "standards". Minimum requirement profiles were used as "standards" in this study. For each of the eighty ci t i e s under study the percentage of the urban labour force 1 i n each of thirteen census categories was calculated. The c i t i e s were then divided on the basis of population into four a r b i t r a r i l y determined size-groups: group one: 10,000 - 19,999 population 35 c i t i e s " two: 20,000 - 29,999 " 15 c i t i e s " three: 30,000 - 99,999 " 17 cit i e s " four: 100,000 & over " 13 cit i e s For every function (census category), the percentage figure for the city with the minimum per cent employed i n each city group was noted. For example, the lowest proportion of city labour force employed i n manufacturing for the cit i e s of group one i s 2.7 per cent (New Waterford); of group two, 5.2 per cent (Glace Bay); of group three, 11.8 per cent (Regina); and of group four, 12.4 per cent lUrban labour force i s defined as total labour force minus the labour force in the "Agriculture", "Forestry and Logging", "Fishing and Trapping", and "Not-Stated" categories of the 1951 Census of Canada industry classification. 65 (Halifax). Table VI shows the minimum percentages employed in c i t i e s of the varying size-classes for each of the functions (census categories) recognized. Ullman and Dacey consider such values to represent the empirical minimum requirements for a viable city, that i s , the service or nonbasic sector of city activity . As noted in the review of the literature minimum requirement values are modified by changes i n city size. In order to relate these values to city size, Ullman and Dacey plotted the minimum values i n their city groups for each activity against the logarithm of city population and calculated linear regression equations for the plotted points.3 By reading minimum requirement values off the regression lines extreme values representative of "abnormal" ci t i e s are avoided. Also, by using the regression lines, the values are determined according to a city's exact population rather than by a broad population class. This procedure has been followed i n this study. The minimum requirement values given in Table VI were plotted against 2Ullman and Dacey, "The Minimum Requirements Approach to the Urban Economic Base," Papers and Proceedings  of the Regional Science Association, VI (I960), 177. 3¥hen minimum requirement values were plotted against the logarithm of population, Ullman and Dacey found that the plotted points closely f i t a straight l i n e . 66 TABLE VI MINIMUM PERCENTAGES EMPLOYED IN CANADIAN CITIES . . OF VARYING SIZE CLASSES, 1951 Cities (population i n thousands) 10.0-19.9. 20.0-29.9 30.0-99.9 .100 & over (35 cities) (15 cities) (17 cities) (13 cities) - per cent of city urban labour force* -Extraction 0.0 0.0 0.0 0.0 Manufacturing 2.7 5.2 11.8 12.4 Public U t i l i t i e s 0.4 0.5 0.6 0.8 Construction 1.8 3.3 4.0 5.a Transportation, etc. 2.4 3.2 3.6 5.1 Wholesale Trade 0.4 1.0 1.5 3.0 Retail Trade 4.2 8.3 9.1 10.3 Finance, Insurance, & Real Estate 1.1 1.1 1.7 2.5 Community Service 4.7 5.3 5.1 5.7 Government Service 1.9 1.6 2.5 3.0 Recreation 0.2 0.4 0.4 0.4 Business Service 0.2 0.3 0.6 1.2 Personal Service 4.4 4.0 4.2 5.4 Total 24.4 34.2 45.1 55.6 Source: Calculated from the Ninth Census of Canada. 1951 *Urban labour force i s defined as total labour force minus the labour force in the "Agriculture", "Forestry and Logging", "Fishing and Trapping" and "Not Stated" categories of the 1951 Census of Canada industry classification. Function (Census Category) the logarithm of city population and linear regression equations were calculated for the plotted points. The values as given by the regression equations are considered to represent the "expected minimum requirements" for c i t i e s . The term "expected" i s used because, by definition, seme of the city minimum values w i l l f a l l below those given by the regression lines.^ Because the minimum city-labour-force percentage values identified, and the regression lines based on them, depend on hew the cit i e s are grouped into population size-classes, a second grouping of c i t i e s , using different class intervals, was made and regression equations calculated for the minimum city percentage values subsequently selected. This was done to ascertain the effect of ar b i t r a r i l y grouping c i t i e s in determining minimum percentage values. Six classes were used in the second grouping scheme in contrast to the four used in the original scheme. Emphasis in the second scheme was placed on differentiating at the higher population levels. Where in the f i r s t grouping scheme, ci t i e s above 30,000 population were divided into only two classes ("30,000—99,999" and "100,000 & over) ^Regression lines represent the "average" f i t , therefore the extremely high and low point values on which lines are based, w i l l not f a l l on the lines. 68 in the second grouping, cities over 25,000 were divided into four categories. The classes used in the second scheme were: group one: 10,000 - 14,999 population 18 cities n two: 15,000 - 24,999 " 27 c i t i e s " three: 25,000 - 49,999 " 14 ci t i e s " four: 50,000 - 99,999 " S cit i e s " f i v e : 100,000 -299,999 " 9 ci t i e s n six: 300,000 & over " 4 ci t i e s Once the equations for the regression lines had been calculated, they were graphed. The Ullman—Dacey regression equations for those a c t i v i t i e s having equivalence in the Canadian and American censuses were also plotted; six of the thirteen functions recognized are in this category. The diagrams in Figure 2 show the relationship of the regression lines to each other for twelve of the thirteen functions. Because "extraction" has minimum requirement values of zero in a l l city-size classes, regression lines were not needed to determine minimum values for this function. Comparison of the two sets of regression lines derived from Canadian data suggests that the effects of a r b i t r a r i l y determining the city-size classes are, for the most part, not particularly significant. For most ac t i v i t i e s the lines are remarkable similar. There i s one t r a i t , however, in which the two sets of lines appear to d i f f e r consistently—*>the degree MANUFACTURING FINANCE, INSURANCE. REAL ESTATE C i t i e s * R e g r e s s i o n l i n e s g i v i n g " E x p e c t e d M i n i m u m R e q u i r e m e n t V a l u e s " • H e a r t l a n d 0 W e s t e r n P e r i p h e r y U l l m a n - D a c e y v a l u e s : — E a s t e r n P e r i p h e r y C a n a d i a n v a l u e s Q N o r t h e r n P e r i p h e r y f o u r c i t y c l a s s e s : • s i x c i t y c l a s s e s : 1— * S e e F i g u r e 1 f o r t h e l o c a t i o n s o f c i t i e s a n d r e g i o n a l b o u n d a r i e s Figure 2. lines The relationship of functional importance to city size, and degression showing "expected minimum requirement values" for functions. PUBLIC UTILITIES RECREATION -!1 1 . 1 | " ~ i i i r i 1 1 : 1 0 1 : • 1 0 1 - \ 1 0 -& o - I 0 _ ~-- -*o - D _ * P ' I s/5 0 a o Valleyfield p o 'Niagara Falls " l j • o * o ,°: a a 1 •*. • 1 1" no II i i i 1 1 j 1 1 1 10 12 Employment as a Per Cent of City Labour Force 0.8 1.2 1.6 2.0 2.4 Employment as a Per Cent of City Labour Force TRANSPORTATION, ETC. /' • i i i i i i i i i i : - /' -/ 1 -II -\ / .' • . a Fort William-Port Arthur 8 0 I ' ° -i • b ' • ° *• ° /••• "° °' £ Monclon 0 Moose Jaw O O North Bay o St.' Thomas ' I CO 0 8 12 16 20 24 28 Employment as a Per Cent of City Labour Force CONSTRUCTION - I — I — 1 1 — I -8 12 16 20 24 28 32 (employment as a Per Cent of City Labour Force C i t i e s * • H e a r t l a n d O W e s t e r n P e r i p h e r y A E a s t e r n P e r i p h e r y D N o r t h e r n P e r i p h e r y * S e e F i g u r e 1 f o r t h e l o c a t i o n s o f c i t i e s a n d r e g i o n a l b o u n d a r i e s R e g r e s s i o n l i n e s g i v i n g 1 ' E x p e c t e d  M i n i m u m R e q u i r e m e n t V a l u e s " U l l m a n - D a c e y v a l u e s : C a n a d i a n v a l u e s f o u r c i t y c l a s s e s : s i x c i t y c l a s s e s : - :  Figure 2. (continued) COMMUNITY SERVICE 3 12 16 20 , 24 28 32 Employment as a Per Cent of City Labour Force PERSONAL SERVICE Employment ELS a Per Cent of City Labour Force GOVERNMENT SERVICE BUSINESS SERVICE ^ O Regina St. John's • St. Jean & Fredericton 8 12 16 20 24 . ! Employment as a Per Cent of City Labour Force Employment as a Per Cent of City Labour Force C i t i e s * . H e a r t l a n d O W e s t e r n P e r i p h e r y ^ • - - E a s t e r n P e r i p h e r y D N o r t h e r n P e r i p h e r y * S e e F i g u r e 1 f o r t h e l o c a t i o n s o f c i t i e s a n d r e g i o n a l b o u n d a r i e s R e g r e s s i o n l i n e s g i v i n g " E x p e c t e d M i n i m u m R e q u i r e m e n t V a l u e s " U l l m a n - D a c e y v a l u e s : C a n a d i a n v a l u e s f o u r c i t y c l a s s e s : :  s i x c i t y c l a s s e s : = Figure 2. (continued) EXTRACTION ° Edmonton Q 0 Calgary • Rouyn - Thotford Minus New Waterford d _ 1 1 1 1 1 I I I I I 15 25 35 45 55 Employment as a Per Cent of City Labour Force C i t i e s * • H e a r t l a n d O W e s t e r n P e r i p h e r y — E a s t e r n P e r i p h e r y _ D N o r t h e r n P e r i p h e r y * S e e F i g u r e 1 f o r t h e l o c a t i o n s o f r c i t i e s a n d r e g i o n a l b o u n d a r i e s R e g r e s s i o n l i n e s g i v i n g " E x p e c t e d  M i n i m u m R e q u i r e m e n t V a l u e s " U l l m a n - D a c e y v a l u e s : — C a n a d i a n v a l u e s f o u r c i t y c l a s s e s ; s i x c i t y c l a s s e s : F O R M U L A E O F T H E L I N E A R R E G R E S S I O N E Q U A T I O N S G I V I N G T H E " M I N I M U M R E Q U I R E M E N T " V A L U E S F O R E M P L O Y M E N T I N F U N C T I O N S F O R C I T I E S O F A L L P O P U L A T I O N S I Z E S . N o t e : A l l e q u a t i o n s a r e o f t h e f o r m : l o g y = a + b x. W h e r e y i s c i t y p o p u l a t i o n a n d x i s t h e m i n i m u m r e q u i r e m e n t ; v a l u e o f e m p l o y m e n t i n p e r c e n t o f c i t y l a b o u r f o r c e . j S o l v i n g f o r x: j x = l o g y - a 1 b ' | T h e y - i n t e r c e p t a n d s l o p e v a l u e s o f t h e r e g r e s s i o n e q u a t i o n s f o r e a c h o f t h e f u n c t i o n s a r e : I i y - i n t e r c e p t (a) s l o p e (b) Function " f o u r c l a s s " city- "six c l a s s " city- U l l m a n and " f o u r c l a s s " city- "six c l a s s " c i t y - U l l m a n and groupin ; scheme grouping scheme Dacey grouping scheme grouping scheme Dacey M a n u f a c t u r i n g 3 8 0 7 1 3 9 6 0 6 3 . 2 2 0 0 + 0 0 9 9 0 0 . 0 6 2 7 0 . 3 5 2 6 + F i n a n c e , I n s u r a n c e a n d R e a l E s t a t e 3 2 2 8 6 3 7 1 9 4 0 . 8 3 3 8 0 8 6 1 5 0 . 5 4 8 0 2. 31'62 W h o l e s a l e T r a d e 3 9 0 5 7 3 9 1 5 3 2 . 2 7 3 1 ' 0 4 3 8 7 0. 3 8 8 7 1 . 5 6 7 1 R e t a i l T r a d e 2 6 0 3 3 2 9 1 5 4 - 8 . 4 2 3 8 * 0 2 7 4 1 0 . 2 2 4 5 1 . 0 2 3 8 * -P u b l i c U t i l i t i e s 1 6 1 4 5 3 5 5 6 9 n o d a t a * 5 5 0 6 7 1 . 8 3 7 2 no data* R e c r e a t i o n 3 5 1 9 9 2 9 9 9 3 1 . 9 8 9 2 3 0 0 8 0 4 . 0 4 2 0 5 . 0 0 2 0 T r a n s p o r t a t i o n e t c . 2 7 5 5 8 3 0 9 7 6 0 . 4 9 9 0 * 0 5 2 5 5 0 . 4 1 4 1 1 . 1 4 9 0 * C o n s t r u c t i o n 3 4 6 7 3 3 4 0 8 1 1 . 7 3 0 7 0 2 9 3 5 0 . 2 7 9 1 0 . 8 8 6 1 C o m m u n i t y S e r v i c e 1 5 0 9 2 0 6 5 0 7 no data* 0 5 9 6 4 0 . 7 5 8 8 n o data* P e r s o n a l S e r v i c e 0 2 9 2 9 0 7 4 3 9 n o data* 0 9 5 7 2 0 . 8 4 8 8 n o data* G o v e r n m e n t S e r v i c e 2 8 2 8 9 3 0 5 2 1 2 . 4 6 1 6 - 0 8 0 1 8 0 . 6 9 3 2 1 . 2 6 5 1 B u s i n e s s S e r v i c e 3 8 7 7 2 3 9 5 6 6 n o data* 1 2 5 7 6 1 . 0 9 3 6 n o data* E x t r a c t i o n n o c a l c u l a t i o n n o c a l c u l a t i o n n o c a l c u l a t i o n n o c a l c u l a t i o n n o c a l c u l a t i o n n o c a l c u l a t i o *: U l l m a n a n d D a c e y v a l u e s a r e n o t a p p l i c a b l e t o C a n a d i a n c e n s u s d a t a b e c a u s e c l a s s i f i c a t i o n p r o c e d u r e s a r e d i f f e r e n t i n t h e -C a n a d i a n a n d A m e r i c a n c e n s u s e s f o r t h e s e f u n c t i o n s . + : T h e A m e r i c a n c e n s u s c a t e g o r i e s " d u r a b l e " a n d " n o n d u r a b l e " m a n u f a c t u r i n g u s e d b y U l l m a n a n d D a c e y a r e t o g e t h e r e q u i v a l e n t t o t h e C a n a d i a n c a t e g o r y " m a n u f a c t u r i n g " . Figure 2. (continued) 73 of slope. The regression lines based on the "four-class" scheme are steeper for ten of the twelve functions. This, i s due, i t i s suggested, to the limited number of large Canadian c i t i e s , their wide dispersal, and the subsequent lack of "city-systems" involving many large c i t i e s . Because there are so few big ci t i e s in Canada and because they are scattered across the country, they function i n relative isolation and hence are more diversified than cit i e s of similar size i n well-integrated "city-systems" such as the Manufacturing Belt of the United States. In such city-systems exchange among ci t i e s reaches large proportions and the propensity for interdependence i s high. Opportunities for functional specialization are much greater for a city i n such a system than for a city operating i n relative isolation where functional activity i s largely between the city and i t s hinterland and where exchange between the city and other ci t i e s i s relatively insignificant. In the highly special-ized city of the system, concentration on one or few functions means that only "token a c t i v i t y " or the minimum activity necessary to keep the city viable needs to be maintained in other functions. In contrast, the non-system city must be more self-sufficient because i t cannot depend to the same degree on other cities for assistance, and because i t must be much more concerned with the needs of i t s hinterland. 74 This situation i s not conducive to the maintenance of only-minimum requirements i n functions. Such a situation, i t i s suggested, exists for most large Canadian c i t i e s , hence the possibility of finding only the minimum activity necessary to keep a city viable i s remote for these large centers. This w i l l cause the "minimum" city percentage values for employment in the higher population classes of the city-grouping schemes to be greater than they possibly should be. The minimum percentage values given by both city-grouping schemes for the smaller cities are thought to be reasonably valid. For c i t i e s up to 50,000 population, i t i s considered that "city-systems" do exist i n Canada—for example, the city-system of southern Ontario. The possi-b i l i t i e s for specialization and interdependence are greater for these c i t i e s than for the large centers; therefore, the probability of some ci t i e s having minimum employment values approximating the "actual" minimum requirement values i s considered reasonably high for these smaller centers. This thesis would appear to be substantiated by the fact that the regression lines for the four and six-class city-grouping schemes correspond f a i r l y closely for most functions at the lower population levels. At the higher population levels the validity of the minimum percentage values comes into question. In light of 75 the previous discussion, the minimum values identified can be expected to be higher than the "actual" values at these population levels. The effects of the "high" values w i l l be greater i n the six-class city-grouping scheme than i n the four-class scheme because the former has more classes at the higher levels of population than the latter. Because of this feature, the "expected minimum requirement" values given by the regression equations based on the four-class scheme have been given priority. Not only i s less emphasis placed on the higher population levels i n the four-class scheme, but by using fewer classes, each class contains more ci t i e s , thus the problem of basing a minimum value on an exceedingly small group of cit i e s i s avoided. For reasons similar to those stated i n the previous discussion, the Ullman-Dacey equations are considered to have a greater validity than either set of equations based on Canadian data. Because there are many American cities at a l l population levels, Ullman and Dacey were able to select at random a standard number of ci t i e s at several population levels giving them a more "balanced" sample than can be obtained using Canadian c i t i e s . Also, c i t i e s equivalent in size to the largest Canadian c i t i e s are found i n American city-systems. The problem of determining minimum requirements based on "non-system" c i t i e s i s , 76 therefore, avoided when using American data. It i s hypothesized that the probability of obtaining minimum employment values very close to the "actual" minimum require-ment values for a city i s much greater using the American data than the Canadian data, especially at the higher population levels. This hypothesis would appear to be borne out by comparison of the regression lines i n Figure 2. For the six functions where the Ullman—Dacey values are applicable, the regression lines based on the American data are steeper than Similar lines derived from Canadian data. The Ullman—Dacey equations give lower "expected minimum requirement" values at the higher population levels than do the equations based on Canadian data. In addition, the divergence between the American and Canadian values increases with population increase, indicating the increasing effect of the high "minimum" employment percentage values i n the large Canadian c i t i e s . The Ullman—Dacey regression lines, therefore, are considered more representative of the "actual" minimum requirements than similar equations based on Canadian city values. The United States with i t s much larger number of cit i e s for which minimum values of employment may be selected, and i t s "city-systems" which include c i t i e s the size of most of the larger Canadian centers, provides a 77 more favourable area from which to draw representative values. Differences in the urban socio-economic environment of Canada and the United States are not considered to be of sufficient significance to invalidate the application of American values to the Canadian situation^. The same economic functions are performed, using similar technology, i n the ci t i e s of both countries and the social mores of urbanites are not greatly different in the two nations. Where the Ullman—Dacey values were not applicable, the regression equations based on the "four-class" city-grouping scheme were utiliz e d ; this occurred for six of the twelve functions for which regression lines are required. Figure 3 shows the regression lines selected in determining the "expected minimum requirements" for functions. The values given by the equations were used as "standards", in rating the importance of functions in and among c i t i e s . The measure. Only basic activity, as given by employment above the "expected minimum requirement n, was used to determine the importance of a function, and the measure -*A precedent i s not being established by applying Ullman and Dacey's regression equations to Canadian data. Trotier used Morrissett's "k-values" i n analyzing Canadian census data for Quebec c i t i e s . Louis Trotier, 'Some Functional Characteristics of the Main Service Centers of the Province of Quebec, "Cahiers de Geographie de Quebec, No. 6 (April-September, 1959), 243-59. A l l equations are of the f o r m : loiz y = a + bx ; where y is c i t y population and x is the "expected minimum requirement value" . The y-intercept a n d the slope of the equations are : Function y-intercept (a) slope (b) Manufactu ring' 3 2200 0 3526 P u b l i c U t i l i t i e s 2 1. 6145 5 5067 C o n s t r u e t i o n ' 1. 7307 0 8861 T r a n s porta t i o n 2 2 7558 0 5255 Wholesale Trade ' 2 2731 1. 5671 Retail T r a d e 2 2 6033 0 2731 Finance , etc.^ 0 8338 2 3162 Community S e r v i c e 2 1. 5092 0 5964 Government Service^ 2 4616 1. 2651 Recreationl 1. 9892 5 0020 Business S e r v i c e 2 3 8772 1. 2576 Personal S e r v i c e 2 0 2929 0 . 9572 ^Equation after U l l m a n and Dacey. 2 E q u a t i o n based on the " f o u r - c l a s s " city-grouping scheme of Canadian ci t ies . Figure 3. Selected "expected minimum requirements" for twelve functions, based on regression lines. 79 was taken i n relative rather than absolute terms. Four classes of functional importance were recognized in the study. They are defined i n terms of standard deviation values of excess employment in a function. For each function, Class I includes a l l ci t i e s whose value of excess employment in the function under consideration i s equal to the mean value of excess employment for the function, plus a value of over two standard deviations above the mean. Class II includes a l l c i t i e s whose values of excess employ-ment exceed the mean by between one and two standard deviations, and Class III includes a l l c i t i e s whose excess employment value in the function i s between the mean and one standard deviation above the mean. Class IV includes a l l c i t i e s where excess employment values are below the mean excess employment value for the function. Table VII shows the values determining the classes of importance for each function. Cities were rated i n each function in terms of the classes of functional importance. An activity was considered "distinctive" in the functional profile of a city when the city qualified for one of the f i r s t three classes of functional importance. Cities have as many "distinctive" a c t i v i t i e s as they have excess employment values qualifying them as being of Class I, II, or III in functional importance. A function SO TABLE VII. VALUES OF EXCESS EMPLOYMENT FOR DETERMINING CLASSES OF FUNCTIONAL IMPORTANCE,.. Mean value Standard Function of excess deviation employment* of excess employment* per cent of city labour force Extraction 3.76 12.49 Manufacturing 30.91 18.25 Public U t i l i t i e s 1.04 1.01 Construction 3.85 2.23 Transportation, etc. 5.56 5.11 Wholesale Trade 2.63 2.34 Retail Trade 5.52 2.91 Finance, etc. Community Service 1.22 0.92 4.23 3.63 Government Service 5.06 6.31 Recreation 0.15 0.14 Business Service 0.65 0.31 Personal Service 2.63 1.74 *Both the mean values and the standard deviations were calculated on the basis of a l l eighty c i t i e s . In the calculations, an excess employment value of zero was assigned to c i t i e s i n those few instances where city excess employment values are negative. a i TABLE VII (continued) Glass IV Class III Class II Class I Function ( mean value (between. (between (over two of excess mean value one and two standard employment) and one standard deviations standard dev- deviations above the iation above above the mean) the mean) mean) per cent of city labour force Extraction <3.a 3.a - 16.2 16.3 - 2a.7 2a. a+ Manufacturing <30.9 30.9 - 49.2 49.3 - 67.4 67.5 + Public U t i l i t i e s <1.0 1.0 - 2.0 2.1 - 3.1 3.2 + Construction <3.a 3 .a - 6.1 6.2 - a.4 a.5+ Transportation, etc. Wholesale Trade <5.6 5.6 - 10.7 10.a - 15.a 15.9+ <2.6 2.6 - 5.0 5.1 - 7.3 7.4+ Retail Trade <5.5 5.5 - a.4 a.5 - 11.3 11.4+ Finance, etc. Community Service <1.2 1.2 - 2.1 2.2 - 3.1 3.2 + <4.2 4.2 - 7.9 a.o - 11.5 11.6 + Government Service <5.1 5.1 - 11.4 11.5 - 17.7 17.a+ Recreation <0.15 0.15- 0.29 0.30- 0.43 0.44 + Business Service <0.65 0.65- 1.46 1.47- 2.27 2.28 + Personal Service <2.6 2.6 - 4.4 4.5 - 6.1 6.2+ 82 was not considered "distinctive" i n the structure of a city when the city qualified only for the Class IV rating of importance. By mapping the ratings of cities for each function a picture of the distribution of the relative importance of functions was gained. Because the measure i s i n relative terms, i t was possible to compare the distribution patterns of different functions as well as examine the relative importance of individual functions among c i t i e s . In section three the distribution patterns are described and analyzed. I I . FUNCTIONS AND CITIES R.U. Ratcli f f was quoted in Chapter I as stating that manufacturing, trade and extraction are the fundamental ac t i v i t i e s of man leading to urbanization in an industrial state. The case for manufacturing and trade (including i t s inseparable associate, transportation) i s well stated by the average percentage distribution of employment among functions for the citi e s under study (Table VIII). Nearly 60 per cent of the urban labour force in the "average" city i s occupied in these functions. Manufacturing alone accounts for over one-third of the employment. The case for extraction, however, i s not apparent from the mean percentage values; i t accounts for only 3.8 per cent of the labour force i n the 83 TABLE VIII THE MEAN PERCENTAGE DISTRIBUTION OF URBAN LABOUR FORCE AMONG FUNCTIONS IN CANADIAN CITIES Mean Function percentage Standard value deviation Manufacturing 34.6 18.0 Retail Trade 12.4 2.7 Community Service 9.2 3.8 Transportation, etc. 8.9 5.3 Personal Service 7.0 1.7 Construction 7.0 2.4 Government Service 6.7 6.4 Wholesale Trade 4.0 2,6 Extraction 3.8 12.5 Finance, etc. 2.8 1.1 Public U t i l i t i e s 1.6 1.0 Business Service 1.1 0.9 Recreation 0.6 0.2 84 average city. This value, in any event, i s not too meaningful since the standard deviation i s over three times the mean value. The original urbanizing functions—social and adminis-tration activities—although shown to be subordinate to the economic functions in terms of employment, s t i l l maintain important positions in the average city p r o f i l e . Together community and government service account for 15.9 per cent of employment in the average c i t y . The remaining 20 per cent of average city employment i s taken up by what may loosely be called ancillary urban a c t i v i t i e s . These functions —finance, business and personal service, entertainment, construction, and public u t i l i t i e s — a r e by their very nature essentially secondary a c t i v i t i e s . In some instances, however, they do emerge as primary functions in c i t i e s . While the mean percentage distribution of employment among functions provides a useful way to examine the positions of functions in a general fashion, i t i s not entirely satisfactory for detailed study. Three reasons for this are apparent: (1) the standard deviations for most functions (Table VIII) are quite large indicating a large degree of va r i a b i l i t y around the mean values; hence relatively non-representative mean values, (2) i t i s known that city $5 functional profiles change with city size; by using the mean employment distribution for a l l c i t i e s to gauge the position of functions, these changes are obscured, and (3) by using total city employment as a basis of study, "abnormal" functional values, usually representative of functions with very high "export" values attributable in most cases to the uniqueness of city location, are included i n the base. Such inclusions tend to give skewed values for functions. This latter point i s closely related to that raised in the f i r s t statement regarding v a r i a b i l i t y around the mean functional values. To avoid the use of relatively non-representative values and to reflect changes in functional structure due to city-size differences, the positions of functions i n city profiles were studied i n terms of expected minimum requirement values for three city-size levels. In addition to avoiding the influence of unique city locations on the base values, the use of expected minimum requirement values as references gives additional precision to the problem. The question to be answered—what are the positions of activities i n the city functional p r o f i l e — i s refined and restated as: what are the positions of functions in maintaining the city, that i s , what i s their role as "city servers". This s i g n i f i -cant refinement i s introduced because minimum requirements 36 represent the nonbasic or city-serving sector of city a ctivity. J.W. Alexander has stated that ".... the nonbasics 'cloud the picture' ...." when attempts are made to identify a city's service to i t s region. 0 Conversely, i t may be argued that the basics "cloud the picture" when attempting to determine the role of functions as city servers. The discussion to follow deals with the position of functions i n city functional profiles as city servers. Later sections deal with the analysis of distinctive and dominant functions, that i s , the role of functions as city builders. In Table IX i s presented the relative positions of functions as city servers i n cit y functional profiles at three population levels. Preliminary study of this table suggests that the rankings of functions are very similar to those presented in Table VIII (page 33) where functions are ranked by mean values. For example, the same ac t i v i t i e s are found i n the six top positions i n each of the ranking l i s t s . In fact, there i s a significant correlation between the average ranking of functions based on minimum require-°J.W. Alexander, "The Basic-Nonbasic Concept of Urban Economic Functions," Economic Geography, XXX (July, 1954), 251. I 37 TABLE IX THE POSITION OF FUNCTIONS AS "CITY-SERVERS"*: FUNCTIONS RANKED BY DECREASING "EXPECTED MINIMUM REQUIREMENT VALUES" FOR CITIES WITH POPULATIONS OF 10,000, . 100,000 AND 1,000,000 INHABITANTS City of 10,000 population City of 100,000 population City of 1,000,000 population Retail Trade Community Service Personal Service Construction Transportation Manufacturing Finance Government Service Wholesale Trade •Retail Trade -R e t a i l Trade -Community Service v. -.Manufacturing Manufacturing ' ^ Community Service Personal Service \ --Transportation X Transportation Construction vx'GC Government Service Finance Wholesale Trade Public U t i l i t i e s eBusiness Service Entertainment - s / ^ P u b l i c U t i l i t i e s Business Service ' ^-Entertainment Extraction ^Extraction Personal Service Construction Government Service XWholesale Trade Finance -Business Service -Public U t i l i t i e s »- Entertainment -Extraction *City-serving activity i s defined as activity directed towards the servicing of the city's own needs. The ranking of functions i s based on expected minimum requirement values given by the regression equations presented in Figure 3. 88 ments and the ranking of functions by mean values.7 This would appear to indicate that a general relationship exists between the positions of functions in city functional profiles as city servers and their positions i n profiles i n undifferentiated roles as both city servers and city builders. The implication i s that the same general structure of relative importance exists for functions regardless of 7m order to compare rankings of functions based on mean employment values with those based on minimum require-ment values, the "average ranking" of functions for the city functional profiles in Table IX was calculated. This was done by weighting functions by their rank values.in the three ranking l i s t s , adding the three values for each function, and l i s t i n g the functions by the sum of the three rank values. The function with the highest sum, in this case extraction with a value of thirty-nine ( i t ranked thirteenth in each l i s t ) , f a l l s at the bottom of the average ranking. The activity with the lowest sum i s placed at the top of the ranking. Retail trade ranked f i r s t in each of the three l i s t s i n Table IX giving i t a sum of three, the lowest of any function. The average ranking of functions i s : 1. Retail Trade 7. Government Service 2. Community Service a. Finance, etc. 3. Manufacturing 9. Wholesale Trade 4. Personal Service 10. Public U t i l i t i e s 5. Transportation, etc. 11. Business Service 6. Construction 12. Recreation 13. Extraction Kendall's coefficient of rank correlation was calculated for the two ranking schemes. A coefficient value of 0.795 was obtained with a normal deviate of 3.72 indicating that a significant correlation exists between the two schemes. whether the city-service or the city-forming roles of functions are considered. Comparison of the tables also reaffirms the leading positions of trade, manufacturing, and community service since they occupy the top ranks in both ranking schemes. The position of extraction i s exceedingly weak in Table IX, indicating that i t i s completely unimportant as a city-serving function. Since i t ranks higher by mean values, i t probably can be considered primarily a city-forming activity. By examining separately the ranking of functions for the three population levels i n Table IX, additional characteristics of functions and ci t i e s are revealed. There are a total of twelve changes in the relative positions of act i v i t i e s i n the table associated with shifts i n city size. Eight occur with the shift from a population level of ten thousand to one of one hundred thousand, and six with the population change from one hundred thousand to one million. Only two act i v i t i e s maintain the same relative position regardless of city size; r e t a i l trade and extraction. Retail trade dominates the city-serving profile at a l l population levels reflecting perhaps the essential role of trade as a prime urbanizing force. The urban center as an economic entity i s composed of numerous specialized unit 90 Trade among these units i s basic to city survival. Intra-city as well as intercity exchange i s necessary for the city to become and remain viable. Extraction ranks last at a l l population levels in the ranking of functions. Such a position would appear to contradict R a t c l i f f T s observation regarding the forces responsible for urbanization. Indeed, the position of extraction as a city-serving function i s non-existent, but the overwhelming dominance i t displays in the functional structure of some ci t i e s when total city a c t i v i t y i s considered (see Figure 2) demonstrates that i t i s an important urban function, albeit a wholly basic (city-forming) one. Of the functions altering their relative positions, manufacturing experiences the greatest s h i f t . It rises from the sixth rank for the lowest level (10,000) to the third for the one hundred thousand population level and f i n a l l y to the second rank for the highest population level (1 ,000,000). Four other functions (government and business service, wholesale trade and transport) experience rises in relative position with increasing population although not of the same magnitude as manufacturing. Six functions undergo decreases i n relative position with increasing population. Two of these, personal service and 91 financing drop one position with each population increase, while community service drops only with the last increase (100,000 to 1,000,000). Construction, public u t i l i t i e s and entertainment decline with the f i r s t population s h i f t , but retain their new positions with the last increase i n population. One suspects that at least some of these shifts in relative position are due in part to the differing qualities of functions i n their capacity to "expand" their service to a larger population using the same employment unit. One employee i n r e t a i l trade, for example, may be able to service, say from one to f i f t y persons, whereas an employee in finance may be able to handle from one to five hundred persons. In this example, i t might be said that r e t a i l trade i s relatively "inelastic" while finance i s very "elastic" i n respect to the amount of servicing one employment unit can achieve. At the lower population levels the minimum requirement values for functions w i l l f a l l within a relatively narrow range because basic services must be maintained even i f underemployment i s experienced. As population increases the "expansion potential" of functions i s more f u l l y exploited and, because the expansion capacity differs among functions, an increasing spread among 9 2 minimum requirement values w i l l occur. One other reason for the shifts in position may be due to the highly aggregated nature of the functions used i n the study. It i s possible that the "content" of the functional groups—retail trade, manufacturing, etc.—varies with city size and thus influences the ranking of functions. The service provided the city of one million by the manufacturing function may well di f f e r from that performed for the city of ten thousand by the same function. Despite these two hypotheses, both of which require further investigation to be substantiated, the total effect of the shifts in the relative positions of functions seems to imply that there are some distinct differences in the make-up of the "city-service" structure i n c i t i e s of different size. Such differences are especially noticeable between city sizes of ten thousand and one hundred thousand population. The role of the city as an "economic entity" appears to be played down somewhat at the lower population levels, and functions associated with urbanization prior to industrialization—community and personal service—occupy dominant positions. As city size increases the functions representing the more purely economic role of the c i t y — manufacturing, transportation and business service—come to 93 the fore and the social functions lose some of the importance they enjoyed at lower population levels. It seems that the role of the ci t y as an "economic mechanism" i s approximated more closely at the higher than at the lower population levels i n terms of city-service structure. While changes i n the ranking of a c t i v i t i e s in city functional profiles have been discussed, no reference has yet been made to the actual changes the minimum requirement values experience with changing city size. Study of Figure 3 (page 78) shows that there are different rates of change associated with each function. A l l functions except extraction, however, show an increase i n minimum require-ment value with population increase; r e t a i l trade and manufacturing experience the greatest increases judging by the slope values i n Figure 3. These increases have a fundamental effect on the functional structures of centers because the total minimum requirement values of c i t i e s increase with city-size increases. Whereas only one-quarter of the city labour force (24 .99%) i s required to maintain a city of ten thousand, 40 per cent i s required for the city of one hundred thousand and over half (55.51%) of the city labour force i s required for a city of one million. Two consequences, one the corollary of the other, 94 result from the reactions of minimum requirement values to changes in city size. F i r s t , functional specialization can be expected to occur more frequently and be of greater magnitude in small ci t i e s than in big ones. Because the total minimum requirement increases as population increases, a much smaller proportion of city labour force i s available for basic activity i n a big cit y than in a small one--45 per cent in a city of one million population in contrast to 75 per cent in a city of ten thousand population. The small city, therefore, with i t s large block of '^unattached" labour force that can be directed into one or few export functions, has a far greater potential for specialization than the big center. Secondly and conversely, the big ci t y , by being able to accommodate such a large proportion of i t s labour force in city-serving functions, relies to a lesser degree on "outside" activity than the small center; thus i t i s more self-sufficient. Ullman and Dacey believe that the trend in minimum requirement values where large c i t i e s have higher total minima than small c i t i e s , i s consistent with theory, ".... since the larger the city the larger the number of specialties that can be supported and the more self-contained the city can be."** Morrissett, taking %llman and Dacey, op. c i t . , 180. 95 a somewhat different tact, also concludes that big ci t i e s approximate self-sufficiency more closely than small c i t i e s . 9 He states that for the manufacturing functions which he studied only one out of fourteen i s ubiquitous in cities of ten thousand population, whereas nine out of the fourteen are ubiquitous for cities of one m i l l i o n . ^ He concludes that "the fact that a l l but a few industries are ubiquitous i n large c i t i e s i s another way of saying that big ci t i e s appear to be more nearly self-sufficient than small c i t i e s . " H Defining the terms "sporadic" and "ubiquitous" in a way similar to Morrissett»sl2 and applying these c r i t e r i a to data on Canadian cities reaffirms Morrissett's conclusions. At the ten thousand population level seven of Morrissett, "The Economic Structure of American Cities, " Papers and Proceedings of the Regional Science  Association, IV (1958). 245. ~~ " l^Ibid. Morrissett defined the terms sporadic and ubiquitous quantitatively as follows: " i f the "k-value" of an industry i s less than one-fourth of the national percentage, the industry w i l l be called sporadic; and i f the ratio i s one-fourth or greater, i t w i l l be termed ubiquitous." n i b i d . -^Quantitative definitions of the terms sporadic and ubiquitous as applied to data on Canadian c i t i e s : If the expected minimum requirement value of a function i s . less than one-third of the average percentage for the ci t i e s studied, the function i s called sporadic; and i f the ratio i s one-third or greater, i t i s termed ubiquitous. 96 the thirteen functions studied are classed as sporadic, but at the one million level only two functions remain sporadic. In brief, analysis of the trends for minimum require-ment values of functions reveals that: (1) the probability of functional specialization decreases with increasing city size, (2) the probability of functional diversification increases with increasing city size, and (3) self-sufficiency i n c i t i e s increases as city size increases* To summarize, the relative positions of ac t i v i t i e s i n the functional profile of cit i e s have been discussed in terms of mean values of total city employment, expected minimum requirement values, and i n light of city-size differences. The reaction of minimum requirement values to city-size changes has been noted and i t s effect on city functional structure has been described. The findings, based on quantitative data, generally coincide with qualitative observations made by earlier analysts, as well as with quantitative results of other similar studies. Trade and manufacturing were found to be the dominant functions in both the total functional structure of the city and in the city-serving profile, although manufacturing, and some other economic functions, were found to be weaker as city servers at the lower population 97 levels. The social functions, although secondary to the economic a c t i v i t i e s , are s t i l l significant i n the functional profile especially i n the small c i t i e s . There are some significant changes i n and influences on c i t y functional structure associated with changing city size. These are: (1) the large city i s more purely an "economic mechanism" than the small city judging from the shifts i n the ranking positions of functions with city-size changes. (2) there i s a greater capacity and potential for functional specialization i n small c i t i e s than i n large ones. (3) large c i t i e s are more diversified and self sufficient than small ones. III. THE DISTRIBUTION OF THE RELATIVE IMPORTANCE OF FUNCTIONS IN CANADIAN CITIES Alexandersson has noted that the distribution pattern of urban a c t i v i t i e s can be grouped into two broad classes: sporadic and ubiquitous. For convenience, the distribution patterns of functional relative importance examined were grouped into these two classes for discussion purposes. As stated i n Chapter I, sporadic a c t i v i t i e s can normally be considered basic or city-forming functions. The ubiquitous 98 ac t i v i t i e s , except when possessing abnormally high values i n city employment profiles, are usually service or city-serving functions. In those isolated cases when ubiquitous ac t i v i t i e s do have atypically high values i n c i t i e s , they should be considered basic functions. Because of the highly aggregated nature of the functional groups used i n this study—thirteen categories i n contrast to Alexandersson»s t h i r t y - s i x — i t was not possible to apply AlexanderssonTs c r i t e r i a i n classifying a c t i v i t i e s as being either sporadic or ubiquitous. His reasoning, however, was followed when allocating functions to the classes. It was assumed that the ubiquitous functions w i l l appear in relatively constant proportions i n c i t i e s except when they take on city-forming status. In contrast, the sporadic a c t i v i t i e s can be expected to range greatly i n their relative importance from city to cit y . These assumptions led to the combined use of (1) the coefficient of variation of city employment i n functions^, and (2) the scatter diagrams given i n Figure 2, (pages 69-72), i n classifying functions as being either sporadic or "^The coefficient of variation for city employment in a function i s the standard deviation for employment i n the function as a percentage of the arithmetic mean for employment i n the function. 99 ubiquitous* Functions Goeff, of Variation Extraction 332,2 Government Service 95*5 Business Service 75.6 Public U t i l i t i e s 65,6 Wholesale Trade 64*8 Transportation, etc, 59.1 Manufacturing 52,2 Community Service 41*5 Finance, etc, 41.2 Construction 34»2 Recreation 33.3 Personal Service 24*2 Retail Service 21.6 The a c t i v i t i e s considered by Ratcliff to be the principal urban functions—manufacturing, trade (and transportation), and extraction—rank high as sporadic, hence basic, functions when "sporadic" i s defined in terms of a high degree of variation from a constant level. However, three a c t i v i t i e s , usually considered as service functions and as having ubiquitous distributions, have higher rankings than wholesale trade, transportation and manufacturing. Examination of the scatter diagrams i n Figure 2, (pages 69-72), shows that two of these functions—business service and public u t i l i t i e s -have high coefficient values because on the few occasions when atypically high percentage values occur for them the values are so very high, they have increased the coefficient of variation markedly. Except for these isolated instances when extremely high values appear (once for business service, 100 six times for public u t i l i t i e s ) , the distribution patterns i n the scatter diagrams for these two functions resemble the patterns of activ i t i e s with much smaller coefficient of variation values. For this reason these a c t i v i t i e s were classified as ubiquitous functions. Government service, although having a tendency to appear i n similar proportions in many c i t i e s (Figure 2, page 71) was considered to have too many deviations (abnormally high values) not to be classed as a sporadic a c t i v i t y . The dictionary meanings of "sporadic" and "ubiquitous" cannot be applied s t r i c t l y here. Perhaps different terms should have been selected. However, because the reasoning behind the classification of functions as being sporadic or ubiquitous has been outlined, the exact terms applied were not considered of great importance. Because the functional categories used are highly aggregated, the classification of functions as sporadic or ubiquitous should be considered i n relative rather than i n absolute terms. The classification of a c t i v i t i e s used i s outlined below: Sporadic" Functions "Ubiquitous" Functions Extraction Wholesale Trade Transportation, etc. Manufacturing Government Service Business Service Public U t i l i t i e s Community Service Finance, etc. Construction Recreation Personal Service Retail Trade 101 The distribution patterns of functional relative importance i n ci t i e s are discussed i n terms of the Canadian "heartland" and "periphery". This regional framework, already introduced on page 4 and i n Figure 2 (pages 69-72), i s a useful conceptualization when analyzing city functional character at the national level, because the distributions of c i t i e s within the two regions differ so greatly. The Distribution of the Relative Importance of the Sporadic  Functions Extraction. Of a l l the distribution patterns, that of extraction best f i t s Alexandersson's definition of a sporadic pattern. Extraction appears as a distinctive function i n only seven c i t i e s ; i t i s completely absent i n twenty-six centers and accounts for less than 1 per cent of city labour force i n forty-two others. In six of the seven cities where i t i s distinctive, however, i t rates as "Class I" meaning that i t accounts for at least 28 per cent of c i t y employment; only manufacturing requires a higher excess employment value for a Class I rating of "distinctive-ness" and no city achieves i t . Five of these c i t i e s are located i n the northern and eastern periphery zones, and one i s i n the heartland. Because extraction, i n each case mining, i s also the dominant function for the c i t i e s , further 102 discussion of them i s postponed to Chapter III. The seventh city where extraction i s a distinctive function (Glass III) i s i n the western periphery—Lethbridge. Only 4 per cent of this city's labour force i s i n extraction (coal mining) giving l i t t l e evidence of the fact that mining was the i n i t i a l city-forming function i n the city. Lethbridge, like the modern "pioneer" towns of Kitimat and Thompson, was developed as a company town in 1885 i n conjunction with the exploitation of coal deposits used to supply the Canadian Pacific Railway. The fact that extraction i s absent i n one-third of the c i t i e s , and accounts for less than 5 per cent of the labour force i n a city where i t provided the basis for city formation brings into question the position of the function as a principal urban act i v i t y . The fact remains though, that extraction requires the concentration of people and machines i n a small area, thus creating an urban situation. The type of urban center associated with this function, however, i s usually quite different from that associated with the other principal urban a c t i v i t i e s . It i s wholly site oriented because extraction a c t i v i t i e s are inevitably oriented to the immobile natural site of the resource. The necessity of placing a high priority on site factors often means that the situational aspects for other city-forming a c t i v i t i e s are unfavourable, thus tending to limit the size of the center while increasing the propensity for specialization* Cities over ten thousand population i n 1951 accounted for only one quarter (27,5$) of the Canadian labour force i n extraction, i n contrast to 80 per cent of the nation's labour force i n wholesale trade, and 68 per cent of the national employment in manufacturing* Extraction, then, can be considered most important as a city-forming act i v i t y i n urban centers under ten thousand population. As the size of the center increases, the relative importance of extraction tends to drop at a very fast rate, other c i t y -forming a c t i v i t i e s becoming the mainsprings of urbanization* Manufacturing, wholesale trade and transportation. The distributions of relative importance for these important city-forming functions il l u s t r a t e the essential elements, not only of ci t y functional character, but also of Canadian economic geography; they demonstrate clearly the effect of distance on the economy of Canada, Two of the distributions are almost identical, those of wholesale trade and trans-portation. The important elements i n both patterns are: (1) the almost complete lack of "distinctiveness" i n the 104 heartland, and (2) the almost unbroken high rating of relative importance i n the periphery. Less than ten of the forty-five heartland c i t i e s have either function showing as a distinctive activity. In contrast, i n the western periphery, trans-portation and wholesale trade do not occur as distinctive functions i n only three of the fifteen c i t i e s , A complete transposition takes place with the manufacturing distribution (see Figure 4)* Manufacturing appears as a distinctive function in only five of the thirty-five peripheral c i t i e s while i t i s distinctive i n thirty-two of the heartland c i t i e s . That i s , there i s an exceedingly high, almost exclusive, concentration of manufacturing relative importance i n the heartland. In the periphery i t simply does not rate as a distinctive function except in purely resource-oriented manufacturing centers. These almost diametrically opposed patterns (manufacturing versus wholesale trade and transportation) provide some indications of the differing characteristics of these functions. Trade and transportation, while requiring the concentration of f a c i l i t i e s and people for their performance, are functions of space (distance). They are concerned with the collection, transfer, and distribution of goods. They can be expected, therefore, to reach their highest levels of relative importance i n those areas where Figure 4. The distribution of wholesale trade and manufacturing as " d i s t i n c t i v e " functions i n Canadian c i t i e s . 106 extensive a c t i v i t i e s (rural functions as defined i n Chapter I) are of prime importance. Such i s the situation i n the eastern, and especially, the western periphery. A very large proportion of the total "economic energy" generated i n these zones i s "consumed" i n overcoming the great distances associated with extensive a c t i v i t i e s such as the wheat-production and ranching functions of the brown and dark brown s o i l regions of the Great Plains. The c i t i e s i n these zones, as the l o c i of the "economic energy", reflect in their functional profiles this involvement with distance. Manufacturing, on the other hand, i s focused on adding value to goods through their conversion, and obtaining a net return by so doing. To maximize the net return, the costs of assembling inputs and distributing outputs must be kept minimal, but also, so must be the cost of the conversion process i t s e l f . This latt e r objective introduces factors not directly involved i n wholesale trade and transportation, such as economies of agglomeration, and internal and external scale economies. Economies that accrue through agglomeration and economies of scale tend to promote the concentration of manufacturing activity i n areas where large populations (easy access to markets) and other manufacturing ac t i v i t i e s (allows specialization and gives more opportunities for substitution) are located. Only when 107 the costs of assembling inputs i s high in relation to total costs, when no substitutes are available, or when the weight-loss i s high i n the manufacturing process w i l l resource or input location factors override those factors favouring market-oriented locations. Such influences tend to propagate population concen-trations already established and give such concentrations of people and economic activity a self-generating capacity. This feature perhaps more than any other factor, i s respon-sible for the continued growth of economic activity i n the heartland. The i n i t i a l causes of concentration i n the heartland were: (1) relatively easy access from the Old World and the United States 1 "core", (2) the ava i l a b i l i t y of resources demanded in world markets—fur and timber, and (3) the existence i n the heartland of a favourable resource base for agriculture. The growing agricultural population demanded manufactured goods; this demand was met by imports i n i t i a l l y , but because the inputs for many manufacturing processes were found i n the heartland, local manufacturing soon developed i n response to the local market, A " c y c l i c a l " relationship developed between rural and urban act i v i t i e s i n the heartland with growth i n one sector allowing expansion i n the other. Because of the excellent I 108 locational relationships existing in the heartland among the factors involved in urbanization—a good agricultural base, the existence of energy resources and raw materials, a large labour supply (and market), and port f a c i l i t i e s , a l l within a relatively small area—the heartland has been able to extend i t s market area to the other regions of Ganada and become the focal point of the nation. The distributions of relative importance of wholesale trade and transportation, and manufacturing i n the nation's c i t i e s t e l l this story more clearly than any of the other distribution patterns. The functional profiles of c i t i e s are very reliable indicators of regional character. Government Service, Only i n the capital c i t i e s and "garrison towns" does this function assume city-forming status, and even then i t i s often secondary to the economic ac t i v i t i e s as a city builder. For example, government service does not have a distinctive role i n either Winnipeg or Toronto—both capital c i t i e s . It attains Class I ratings (excess employment value over 17.8 per cent of city labour force) i n the larger c i t i e s only i n the national capital, and i n the provincial capitals where both legislative and military functions are 109 located (Victoria and Halifax). Glass I ratings of relative importance also exist for the garrison towns of St. Jean (military college) and Trenton (air force base). Class II ratings are found i n the provincial capitals of Fredericton, Regina, and St. Johns, and i n the military and administrative centers of Kingston (military college and penal institution), Pembroke (army camp), and Barrie (army camp). Except for three centers, the twelve remaining c i t i e s where government service i s distinctive (Class III) are located i n the periphery. Of the three i n the heartland one i s a provincial capital—Quebec City, and the other two are regional service centers on the edge of the main manufac-turing zone of southern Ontario--Belleville and London, This function, like transportation and wholesale trade, tends to be relatively more important i n the peripheral c i t i e s , i l l u s t r a t i n g again the concern of these c i t i e s with the problem of distance and large, relatively sparsely populated hinterlands. The administration functions "consume" proportionately more "energy" i n the periphery where f r i c t i o n of distance i s high, than i n the highly developed heartland where settlement i s relatively dense and easy access exists to a l l areas. 110 The Distribution of the Relative Importance of the  Ubiquitous Functions As expected there are no distinct gaps or concen-trations in the distribution patterns of the ubiquitous a c t i v i t i e s . The great contrasts existing between the periphery and the heartland in the patterns of relative importance for the sporadic functions are not found. Some slight differences i n functional relative importance, above that created by ubiquitous functions assuming cit y -forming status, are, however, exhibited by the distributions and serve to il l u s t r a t e some significant differences i n city functional structure within regions, especially the heartland. The distribution patterns of functional relative importance in the heartland for the ubiquitous functions— for example, r e t a i l trade and recreation, (see Figure 5 ) — when compared with the pattern for manufacturing, (Figure 4, page 105), serve to ill u s t r a t e the "binodal" nature of the region's urban structure. The manufacturing pattern indicates two areas of concentration i n the heartland. One consists of the cities near Montreal, the other of the c i t i e s at the western end of Lake Ontario, The patterns for the ubiquitous functions show that, 112 generally, the importance of these acti v i t i e s i s highest i n the c i t i e s outside the manufacturing "cores". In the c i t i e s making up the manufacturing concentrations, the ubiquitous ac t i v i t i e s tend to have low ratings of relative importance except for the few instances when they attain strong city-forming status* These distributions reflect some important charac-t e r i s t i c s of the heartland's economy and the nature of the c i t i e s i n this region. Manufacturing within the heartland i s shown to be highly concentrated i n two areas, and the ci t i e s making up these "cores" have very few functions other than manufacturing showing up as distinctive i n their profiles. The heartland c i t i e s f a l l i n g between and beyond the two manufacturing "cores", however, demonstrate a significant degree of "distinctiveness" i n the ubiquitous functions. Moreover, included i n this latter city group are most of the few heartland ci t i e s where wholesale trade and transportation;show as distinctive functions. These ci t i e s are i n fact very similar i n functional structure to the majority of the peripheral c i t i e s where servicing a hinterland i s the principal raison d'etre of the city. The c i t i e s i n the manufacturing "cores" on the other hand, are the basic units i n "city-systems" where intercity and 113 intracity interaction i s often relatively more significant than city-hinterland activity. The relatively few times ubiquitous functions take on city-forming status are of interest simply because they are exceptions to the general situation. Although i t i s not possible to comment on a l l the occasions that the ubiquitous, city-serving functions assume a city-forming capacity, some of the more outstanding ones are mentioned. Two of the best examples of this phenomenon are given by the business service and public u t i l i t i e s functions. Seventy-nine of the eighty c i t i e s have less than 3 per cent of city labour force in business service. One city , however, Pembroke, has over 7 per cent of i t s labour force i n business service, making this function very significant, i n relative terms, i n the city's functional p r o f i l e . Pembroke has a relatively high concentration of employment in engineering and s c i e n t i f i c service. The c i t y i s located i n the Ottawa Valley where the exploitation of white pine provided the impetus for settlement and made Pembroke a major administrative center (county seat) as well as an important wood-processing c i t y . The administrative function of the city i s further reflected through the significant business service employment rate. The city i s a site of a d i s t r i c t forester's office, and nearby i s a large forest research station, as well as an important military establishment. There i s also a major provincial f i s h hatchery and bird sanctuary proximate to the c i t y . In addition the city i s the head-quarters of engineering firms associated with forestry operations, and legal surveying. The distribution of relative importance for the public u t i l i t i e s function follows closely that of business service. The function appears in an almost standard and unspectacular proportion i n a l l c i t i e s but a few located i n the heartland. The two most outstanding exceptions are Niagara Falls and Valleyfield, both associated with well-known hydro power generating sites: Niagara Falls and Beauharnois. In these instances the public u t i l i t i e s function i s definitely a basic activity. The other heart-land c i t i e s with high ratings for public u t i l i t i e s are a l l associated with hydro power, either i t s generation, transmission, or administration. Barrie, for example, i s one of the nine regional offices for Ontario Hydro. The few times that community service attains relatively high ratings i n city functional profiles serve to i l l u s t r a t e , to a very limited degree, the duality that exists i n Canadian culture. The three c i t i e s having the 115 highest ratings for this function are located in Quebec, and seven of the top fifteen c i t i e s i n community service are found in predominantly French-speaking areas. Educational and especially religious activity would appear to be very important (in relative terms) i n the functional structure of c i t i e s within the realm of French Canadian culture. In St. Hyaeinthe, for example, over 38 per cent of the employment i n community service i s i n religious a c t i v i t y . Although the point cannot be pursued very far on the limited evidence presented here, i t i s thought that a closer approximation of the "cathedral" town w i l l be found i n Quebec than i n any other province. With the exception of the "finance, insurance and real estate" function, the remaining ubiquitous activities attain atypically high values only i n a few peripheral c i t i e s . As expected, the c i t i e s where finance, etc. attains the greatest importance are i n the heartland. Toronto, as the financial heart of the nation, has the highest rating i n Canada for this function. The two insurance centers of London and Kitchener-Waterloo closely follow. Retail trade and recreation maintain their ubiquitous characteristics throughout, showing no extremely high atypical values for any city. Generally though they 116 are more significant i n the peripheral c i t i e s than i n those of the heartland. Personal service i s very distinctive i n a few peripheral c i t i e s , but again i t i s insignificant i n the heartland. The one remaining function, construction, i s unlike other city functions in that the proportion of the labour force employed i n the construction industry i s i n large part determined by the current growth or expansion of a community. Its importance indicates the dynamics of growth more than the actual functional relationships of a cit y . An outstanding exception to this thesis i s found in one city, although i t i s primarily one of scale, Edmonton has a high rating of "distinctiveness" for 1 construction. In large part this i s due to the high growth rate i n the city brought on by petroleum and natural gas discoveries i n i t s hinterland, but some of i t i s due, i t i s suggested, to the situation of Edmonton i n relation to northern development. Edmonton i s the "gateway to the Arctic" and as such i s the home office of many construction firms and residence of workmen concerned with northern construction. These two factors have led to construction's high rating i n Edmonton, The distribution patterns of relative importance for the ubiquitous functions, by themselves, do not reveal any 117 significant variations at the national level i n city-functional character. However, when used to supplement similar patterns for the sporadic a c t i v i t i e s , they act to refine the overview of cit y functional performance, especially i n the heartland. Summary; "Heartland". "Periphery" and Canada's Urban  Geography The foregoing discussion has shown that the distribution patterns of relative importance for urban functions i n Canada are characterized by extreme unevenness. The heartland, enjoying excellent locational relationships for many manufacturing processes—the key urban fu n c t i o n -i s the site of urban concentration. The region contained 52 per cent of Canada's population i n 1951, yet i t occupies only 2,5 per cent of the nation's land area1**-; forty-five of the eighty c i t i e s studied are in the heartland. Indicative of the different densities existing i n the two regions i s the average road distance between c i t i e s . Cities i n the heartland are, on the average, 280 miles apart by road, while those i n the western periphery are separated by 14-The percentage values were derived by determining the population and area of a l l census divisions that make up the heartland, and comparing the total valuesibrthe heartland with the national figure. Source: the 1951 Census of Canada, 1X8 15 an average road distance of 740 miles • Whereas great concentration of economic activity, financial power, and population exist i n the heartland, the periphery i s characterized by huge sparsely settled areas, giving i t almost vassal status i n i t s relationship with the heartland* The distribution of relative importance for the manufacturing, wholesale trade and transportation functions i l l u s t r a t e , i n a very definite manner, the fundamental difference existing between the majority of the peripheral c i t i e s and most of the heartland c i t i e s . The periphery, because of i t s involvement with the f r i c t i o n of distance, must devote much more of i t s "economic energy" to overcoming this f r i c t i o n than must the heartland where distal f r i c t i o n i s relatively inconsequential. These fundamental differences in situational characteristics, reflected by the high rating of "distinctiveness" for wholesale trade and transportation i n the peripheral c i t i e s , and manufacturing i n the heartland c i t i e s , provide the basic empirical evidence supporting the concept of "heartland" ci t i e s and "peripheral" c i t i e s . ^Average road distances were derived by determining the road distance between ci t i e s using mileage charts on o f f i c i a l provincial and o i l company road maps* A more detailed account of the concentration of activity i n the heartland can be found in, D.W. Slater, "Trends in the Industrial Location i n Canada," Resources  for tomorrow Conference Background Papers. Vol* I (Ottawa: The Queen's Printer, July, l 9 & l ) , pp. 410-11. 119 Concentrations i n national distribution patterns of socio-economic phenomena, although evident i n spectacular fashion i n Canada, are not peculiar to this country alone, E.L. Ullman has delineated and described the "core" and "fringe" areas of the United States, He states that the American core, located in the northeastern part of the nation, has 43 per cent of the nation's population and $2 per cent of the total national income although i t occupies only 7 per cent of the land.* 0 A.E. Smailes i n his studies of urban phenomena i n Britain refers to the "hour-glass" shaped area of maximum urban concentration: "The most striking feature [of the distribution of urban centers i n England and Wales] i s that a great majority of towns are concentrated within a belt extending north-west from the south-east coast and London region to the industrial areas of Lancashire and West Yorkshire, A large proportion of the urban groups i s found clustered i n areas within forty miles radius of central London and central Manchester . . . In fact, the form of the main area of concentration may be likened to an hour-glass, with i t s main axis running through London and Manchester and i t s waist about Northampton,!' Similar concentrations of population are also found in continental Europe; for example, the Po Valley of Italy, l^Edward L, Ullman, "Regional Development and the Geography of Concentration," Papers and Proceedings of the  Regional Science Association, IV (1958). 181. x'Arthur E. Smailes, "The Urban Mesh of England and Wales," Transactions and Papers of the Institute of British  Geographers. (1946), 89. . ' ' ~" """ the basins of the Seine and Meuse i n France and the Rhine and Elbe i n Germany. To describe the distribution of urban functions i n terms of a "heartland" and a "periphery" or a "core" and a "fringe" seems reasonable because concentrations of popu-lation such as those outlined above have been shown to be intimately associated with urban functions, especially manufacturing. x o Ullman states that 70 per cent of the United States industrial employment i s located i n the American core»^9 Smailes claims that ". . .the industrial population i n Britain i s the major determinant of the arrangement of urban centres."20 He further states that: The peculiarly uneven distribution of towns and c i t i e s that i s characteristic of Britain i s directly attributable to the high degree of localization of the natural foundations of the industrial and commercial l i f e of this country. The distribution of the industrial population i s thus seen as cause rather than effect of the urban pattern. 2^ l^This statement must be qualified to the extent that i t applies only to industrialized countries. Population concentration and urbanization are not necessarily closely associated i n non-industrialized nations. •^Ullman, loc. c i t . ^Smailes, oc. c i t . t 90• 2 1 I b i d . 121 Allan Rodgers i n portraying graphically the variation in the distribution of the manufactural employment i n Italy has shown the coincidence of the important manufacturing areas with the densely populated core area of I t a l y — the Po River b a s i n , 2 2 The close association of population concentration and urbanization, i n the industrialized countries at least, i s really self-evident, especially when recalling R a t c l i f f f s definition of urban activ i t i e s ; , ". • .activities that require the concentration of people, buildings, and machines within relatively small areas." 23 Urbanization thus leads automatically to concentrations of human activity, and hence to unevenness i n the distributions of socio-economic phenomena. The examples of concentration cited above, together with the evidence presented for the Canadian scene where city functional structure was shown to vary funda-mentally between regions, provide strong support for the "heartland-periphery" concept and i t s application to the analysis of urban character at the macro or national scale. 2 2 A l l a n Rodgers, "Regional Industrial.Development with Reference to Southern Italy," Essays On Geography and Eco-nomic Development, Norton Ginsburg, ed., Department of Geography, University of Chicago, Research Paper no, 62 (Chicago: University of Chicago Press, I 9 6 0 ) , Ghap. IX, pp. 143-73. 2 3 R . U . Ratcliff, Urban Land Economics (New York: McGraw-Hill Book Company, 1 9 4 9 ) , p. 20. ~" CHAPTER III THE DOMINANT FUNCTIONS AND SPECIALIZATION OF CANADIAN CITIES In the preceding chapter urban functions were studied systematically i n order that their relative importance i n a l l c i t i e s could be assessed* In this chapter dominant act i v i t i e s i n cit i e s are discussed, and city specialization as shown by an index developed by Ullman and Dacey i s examined* Finally, a discussion of city types and their distribution i s presented. Section one of the chapter deals with the techniques u t i l i z e d i n identifying dominant functions and determining cit y specialization. In the second and third sections the distribution of ci t i e s by dominant activ i t i e s and degree of specialization, respectively, are examined. The fourth and f i n a l section includes the discussion of city types and their relationship to factors of location. I. MEASURES OF FUNCTIONAL DOMINANCE AND SPECIALIZATION Functional dominance. As described i n Chapter I the dominant function of a city i s here defined as the activity having the largest share of to t a l city excess 123 employment. The identification of dominant functions involved the u t i l i z a t i o n of techniques described i n Chapter II, As outlined i n Chapter II excess employment values were calculated when identifying distinctive a c t i v i t i e s i n c i t i e s . These values, calculated by subtracting a city's expected minimum requirement values from the city's actual employment values for functions, were further applied to determine dominant a c t i v i t i e s . The excess employment values for a l l ac t i v i t i e s within a city were added to give t o t a l city excess employment. The percentage share of this total associated with each function was then calculated and the function having the largest share was considered to be the city's dominant activity. The activity identified need not necessarily be a distinctive function in addition to being the dominant one, although the probability of this being so is high. Again, because only basic activity (excess employment) i s used, this method of determining dominant functions would appear to be superior to methods based on tota l city employment where the included ". • .nonbasics •cloud the picture'. . .".*• Functional specialization. The problem of determining the degree of functional specialization i n ci t i e s i s closely Ij.W. Alexander, "The Basic-Nonbasic Concept of Urban Economic Functions," Economic Geography, XXX (July, 1954). 251. ~~ 124 related to that of identifying distinctive and dominant ac t i v i t i e s * Norms or standards must be established against which the specialization of individual c i t i e s can be measured, and units of measurement must be developed* Several methods of determining functional speciali-zation have been devised, most of which have used census employment data as source material. The most common standards or reference levels against which individual c i t i e s or other areal units are compared i n evaluating specialization include the distribution of employment among selected economic a c t i v i t i e s for a nation, the average employment distribution i n selected a c t i v i t i e s for a l l c i t i e s or areas under study, and equal distribution of employment i n a l l a c t i v i t i e s selected. P.S. Florence and associates used the national employment profile as a base when calculating specialization indices on a state level i n the United States. 2 The employment structure of each state (percentage distribution of employment i n a l l economic activities) was compared to that of the United States—the nation—using the latter as 2P.S. Florence, W.G. F r i t z , and R.G. Grilles, "Measures of Industrial Distribution," Industrial Location  and National Resources. U.S. National-Planning Board (Washington, D.C.,: U.S. Government Printing Office, 1943). 125 a frame of reference for a balanced structure. Deviations between the state and national values for individual economic ac t i v i t i e s were summed, disregarding positive and negative signs, and the resulting t o t a l was taken as a specialization index. Both J,W. Webb^ and A, Rodgers^ have used average employment values in selected a c t i v i t i e s for the c i t i e s under study as base levels i n constructing specialization indices. Webb's index i s similar to Florence's i n that i t i s the sum of a l l a city's functional indices. As described i n Chapter I, Webb's functional indices are weighted location quotients based on mean values of employment i n a l l functions for the c i t i e s he studied. If a city has a specialization index of unity i n Webb's system, the employment structure of the city conforms absolutely to the mean employment profile for the citi e s studied. As the index increases in value, specialization increases, Rodgers has developed an index whose value ranges from zero to one thousand; specialization increasing with 3 j . W. Webb, "Basic Concepts i n the Analysis of Small Urban Centers of Minnesota," Annals of the Association  of American Geographers. XLIX (March, 1 9 5 9 ) , 5 5-72, * Rodgers, "Some Aspects of Industrial Diversi-fication i n the United States," Economic Geography, XXXIII (January, 1 9 5 7 ) , 1 6 - 3 0 . -126 with increase in the index value. An index of zero indicates that the employment distribution in the city i s identical to the mean employment distribution for a l l c i t i e s under consideration. Absolute specialization i s indicated by an index value of 1000. This value indicates that the entire employment i n a city i s in one function. In using the mean employment structure as the base of their specialization indices, these analysts are assuming that the mean structure i s the "balanced" city structure. Equal distribution of employment among functions was used as a base in the study by R.C. Tress.^ He used equal employment i n twelve major industrial groups as a basis for least specialization (or absolute diversification) and computed the deviation of a city*s employment profile from this base as an indicator of the degree of specialization. Rodgers states that this i s ". . .relatively meaningless because an index constructed on such a base [equal employment in each function] would vary tremendously based on the kinds and numbers of industrial groups measured."^ Tress, "Unemployment and Diversification of Industry," The Manchester School. IX (1938), cited by A. Rodgers, rtSome Aspects of industrial Diversification i n the United States," Economic Geography, XXXIII (January, 1957), 18. ^A. Rodgers, "Some Aspects of Industrial Diversi-fication i n the United States," Papers and Proceedings of the  Regional Science Association, I (1955), B-2. -•••• • 127 Each of these standards has some serious conceptual faults. F i r s t , i t i s questionable whether any of the bases represent "balanced" employment structures. Secondly, as stated by Rodgers, ". . .no area can be a microcosm of the nation as a whole";'' therefore, the comparison of individual c i t i e s with the national employment profile or with the average employment structures for a l l c i t i e s under study i s probably not a completely valid means of determining specialization. In re a l i t y specialization, and i t s antonym diversification, are d i f f i c u l t to define i n this context, as i s a "balanced" structure. The particular "employment mix" representative of a balanced profile probably differs with city location and size, and the national economy. Because of the d i f f i c u l t i e s encountered when using the described standards, a different approach was taken in determining specialization i n this study. A specialization index developed by Ullman and Dacey based on the expected minimum requirement profiles of individual eities was adopted. This index assigns values to c i t i e s on the basis of the distribution of total city excess employment among the city 7 l b i d . , B-3. ^E.L. Ullman & M.F. Dacey, "The Minimum Requirement Approach to the Urban Economic Base," Papers and Proceedings  of the Regional Science Association. VI (I960), 189. 128 functions. Ullman and Dacey have designed their index so that large deviations (concentrations i n one or very few functions) are accentuated. The formula 9 for determining the index value i s : / where: nS t t i s the index of specialization, " i " refers to each of the thirteen functions, n\?^ n to the percentage of total city urban labour force employed i n each of the " i * functions, "M^ the expected minimum requirement for each function, and nT±n the sum of a l l the functions. An index value of unity indicates that the d i s t r i -bution of excess employment among functions i s of the same magnitude as the distribution of total minimum requirement employment among functions. That i s , i f a city's expected minimum requirement values for each function are multiplied by a common factor, the product in each case would be the city's actual employment percentage value for the respective functions. For example, i f the census industrial categories are aggregated into five functional classes, a city would have a specialization index of one, i f the conditions i n Example A. (page 129) held. In this case the actual city employment values exceed the expected minimum values by a factor of 2.5 i n each function. Excess employment i s 9 I b i d . Function Mi Pi-Mi (excess employment) EXAMPLE A (Pi-Mi) 2 (Pi-Mi) 2 (riPi-XiMi) 2 Mi ZiMi Specializa-tion Index Total expected minimum require-ment value _ 40.00 100.0 Per cent of Gity Urban Labour Force 90.00 a 10.0 25.0 15.0 225.00 22.5 b 5.0 12.5 7.5 56.25 11.25 c 2.0 5.0 3.0 9.00 4.50 d 15.0 37.5 22.5 506.25 33.75 e 8.0 20.0 12.0 144.00 18.00 60.0 90.00 90.00 s i no 130 distributed among the functions i n the same proportions as the t o t a l minimum requirement employment i s distributed among them* The distribution of excess employment would appear to provide a reasonable base reference when measuring speciali-zation* The use of an unrealistic standard such as the national employment structure or the average employment profile for c i t i e s under study i s avoided, and because excess employment i s determined as a function of city size, the effect of changing city size i s allowed for. Also, by assigning an index value of unity to c i t i e s whose d i s t r i -bution of excess employment i s of the same magnitude as that for the total minimum requirement employment, and increasing the value of the index as deviation from this pattern increases, a rational plan of action i s followed. It seems logical to expect the distribution of excess employment among functions i n "diversified 1* or "well balanced" cit i e s to follow closely the distribution of the minimum requirement employment. Similarly, deviation from this format would indicate a tendency towards an "unbalanced" city structure and hence specialization. City functional specialization as given by an index based on the Ullman-Dacey type i s discussed i n Section III. 131 At this point, i t i s necessary to digress and present some comments on Ullman and Dacey*s application of their specialization formula. These analysts appear to have misinterpreted their own index. They state that . .a city with i t s excess distributed evenly above each of the 14 minima represents the most diversified or least specialized city.1*^*® They also state that, "The higher the number [the index value] the more specialized [the c i t y ] ; the lower the more d i v e r s i f i e d . t t ^ This would mean that a city whose excess employment was distributed evenly among i t s functions would have the lowest index value. Exami-nation of the properties of the index reveals that this i s not so. This i s made evident by comparing the following example of city functional values where excess employment i s distributed evenly among functions, to the earlier example (page 129) where the excess employment i s distributed i n constant proportions relative to the minimum requirement values. The minimum requirement values are the same for both examples; this assumes that the ci t i e s i n examples "A" and "B" are equivalent i n population. In the earlier example ("A'*) the index value i s 1 0 I b i d . , 190. n i b i d . i EXAMPLE B Function 1^ P i P ^ (excess employment) (Pi-Mi) 2 (Pi-Mi) 2 (ZiPi-ZiMiV M± ZiMi Specializa-tion Index Per cent of City Urban Labour Force a 10 22 12 144 14.4 b 5 17 12 144 2S.8 c 2 14 12 144 72.0 d 15 27 12 144 9.6 e 8 20 12 144 18.0 Total expected minimum |_ 40 100 require-ment value (Pi-Mi) 2 * 142.8 90.0 = 1^51 133 unity, but i n the later example ("B"), i t i s 1.58, suggesting that "B" i s more specialized than "A" even though excess employment i s distributed evenly among functions for nB n* These findings are at variance with Ullman and Dacey's statement regarding excess employment distribution. Observation of city characteristics suggests that to expect the city of example "A'* to be less specialized would appear rational since i t i s unlikely that equal amounts of excess employment would be found in a l l a c t i v i t i e s i n the least specialized c i t i e s . Because the minimum requirement values vary among activ i t i e s i t i s to be expected that excess employment values w i l l also vary from function to function i n the least specialized c i t y . Equal amounts of excess employment i n a l l a c t i v i t i e s , of necessity, indicates that some functions w i l l be relatively more important than others because the base from which excess employment i s measured differs among functions. For example, i f a l l ac t i v i t i e s i n a ci t y have excess employment values of 5 per cent, the function whose minimum requirement value i s 1 per cent w i l l be relatively more important i n the city than the function whose minimum i s 4 per cent; hence, there i s a trend towards specialization when excess employment values are identical for a l l functions. 134 Ullman and Dacey appear to have overlooked one other significant item when applying their index. At one point the specialization index formula calls for the division of the formula operates satisfactorily; however, i n the case of one function—extraction—Mi » 0 regardless of city size. This i s true of both Ullman and Dacey's study and the present study. Since the relationship "X * 0* i s undefined (mathematically impossible) i t i s not possible to apply the formula i n the case of extraction i n a mathematically correct manner. Ullman and Dacey appear to have equated the relation-ship [(Pi*Mi) 2 f Mi; when Mi s 0] to zero. Apart from being mathematically incorrect this practice, although not altering index values significantly for c i t i e s where extraction i s unimportant, does give very misleading index values for c i t i e s where extraction i s significant. The index values for extraction centers are much lower than they should be. To avoid this problem in this study, an arbitrary value of 0.01 per cent of city labour force was assigned as the minimum requirement value for extraction, regardless of city size, when calculating specialization 135 indices. By using 0.01 rather than zero, the high specialization of typical mining centers i s reflected by the specialization index; at the same time the a r b i t r a r i l y selected value (0.01) i s small enough to avoid serious distortion of index values for non-extraction centers. The following example il l u s t r a t e s this point. In case "A", the minimum requirement value for extraction i s zero; in ease "B" i t i s 0.01. A l l other values are the same i n both cases. Comparison of the two cases shows the need for assigning a value other than zero as the minimum requirement value of extraction. In case A, where the minimum value i s zero, extraction i s ignored by the formula, giving an unrealistically low index value. Extraction i s taken into account i n case B and the index subsequently reflects the sample city's great specialization i n this function. II DOMINANT FUNCTIONS OF CANADIAN CITIES Manufacturing i s the dominant function i n sixty-one of the eighty c i t i e s . This high frequency of occurrence demonstrates i n a dramatic way the overwhelming importance of this activity as a city builder. Because the function appears dominant so often, two classes of dominance are recognized: "Manufacturing I", and "Manufacturing II". CASE A Function (Pi-Mi) ( P i - % ) ^ (Pj-Mj) 2 (ZTiPj-ZiMi)1 Specializa-tion Index extrac-tion Total expected minimum require-ment value Per cent of City Urban Labour Force 0.00 25.00 25.00 625.00 ? (0) b 10.00 12.00 2.00 4.00 0.40 c 15.00 18.00 3.00 9.00 0.60 d 20.00 25.00 5.00 25.00 1.25 e 3.00 20.00 17.00 289.00 96.33 .48.00 100.00 (Pi-Mi) 2 M. 98.58 56.33 98.58*56.33 = 1.75 ON CASE B Function % Pi (Pi-Mi) (Pi-Mir (Pi-MiT (ZiPi-ZiMir Specializa-" tion Index M-i ZiMi extrac-tion Per cent of City Urban Labour Force 0.01 25.00 24.99 624.50 62450.00 b 10.00 12.00 2.00 4.00 0.40 c 15.00 18.00 3.00 9.00 0.60 d 20.00 25.O0 5.00 25.00 1.25 e 3.00 20.00 17.00 289.00 96.33 Total expected minimum require-ment value 48.01 100.00 (Pi*Mj)' Mi " s 62548.58 56.30 62548.58 * 56.30 3 1.110.99 138 Included i n the f i r s t class are a l l cit i e s where 50 per cent or more of total city excess employment i s accounted for by manufacturing. In the second class are a l l c i t i e s where manufacturing i s dominant, but accounts for less than 50 per cent of total city excess employment. Only five other functions register as dominant a c t i v i t i e s , "sharing" the remaining nineteen c i t i e s . Extraction i s dominant i n six c i t i e s , transportation i n five , government service i n f i v e , r e t a i l trade in two, and community service in one. Table X gives the dominant function of each cit y . Of the sporadic functions, only wholesale trade f a i l s to appear as a dominant activity. As expected, a dominant rating comes to very few ubiquitous functions (two), and only three of the eighty ci t i e s have such functions as their dominant activity. The pattern of functional dominance in c i t i e s (Figure 6) substantiates the observations made i n Chapter II on the role of functions in c i t i e s ; simultaneously, i t emphasizes even more the unique position of manufacturing among the urban functions. It reflects the heavy concen-tration of manufacturing in the heartland. Thirty-one of the thirty-six c i t i e s with ratings of "Manufacturing I" are located i n the heartland and a l l of these, save two, are i n either one or the other of the two manufacturing nodes 139 TABLE X SPECIALIZATION INDEXES AND DOMINANT FUNCTIONS OF CANADIAN CITIES...... City Special-ization Dominant Function Periphery: Heartland Index Charlottetown 1.16 Retail Trade (Lethbridge*) 1.25 Rimouski 1.36 Community Servi ce Prince Albert 1.40 Manufacturing II Brandon• 1.41 Transportation,.etc. Penticton 1.44 Saskatoon 1.51 Manufacturing II Moncton 1.55 Transportation, etc. Moose Jaw 1.56 n Saint John 1.67 Manufacturing II Chicoutimi 1.72 t» St. John1s 1.75 n Truro 1.77 n (Calgary*) 1.77 tr (Edmonton*) 1.80 tt Fredericton 1.86 Government Service Barrie 1.87 Manufacturing II Medicine Hat 1.97 tt Winnipeg 2.00 tt/ t» B e l l e v i l l e 2.04 Fort William-Port Arthur 2.08 n Regina 2.09 Government Service Quebec 2.11 Manufacturing II St. Thomas 2.14 n Kingston 2.17 London 2.17 tt Chatham 2.26 tt North Bay 2.35 Transportation, etc. Edmundston 2.35 Manufacturing II •The regular index value i s very high because of the extraction function; this function, however, accounts for less than 5 per cent of city labour force. Because of this, a special index was constructed. It i s similar to the regular index except that extraction has been omitted i n i t s compilation. TABLE X (continued) 140 City Special- Dominant Function ization Periphery: Heartland Index Vancouver Brockville Owen Sound O r i l l i a Joliette Pembroke Toronto Trois Rivieres Sherbrooke Niagara Falls Trenton Stratford Sarnia Montreal St, Hyaeinthe Victoria Woodstock Cornwall Guelph r Jonquiere Shawinigan St, Catharines Sault Ste, Marie Sydney < Kitchener-Waterloo Windsor Peterborough Halifax St, Jean St, Jerome Vic t o r i a v i l l e Brantford Hamilton 2.40 2,71 2.73 2.74 2.88 3.41 3.51 3.52 3.61 3.62 3.63 3.37 3.90 3.92 3.96 4.01 4.14 4.16 4.22 4.40 4.53 4.56 4.7© 5.21 5.36 5.41 5 . 4 2 5.46 5.55 5.56 5.59 5.67 5.38 Manufacturing II » tt: tt Manufacturing I Manufacturing II Manufacturing I tt tt tt Manufacturing II Manufacturing I. n tt tt Government Service Manufacturing I tt « » ** n 9 tt: Government Service Manufacturing I tt: n tr It TABLE X (continued) 141 City Special-ization Dominant Function Periphery: Heartland Index (Sorel*) 6,05 Manufacturing I Drummondville 6.24 n Granby 6.28 Grand »Mere 6.34 tt Valleyfield 6.39 tt We Hand 6.54 n T r a i l 6.36 nr. Gait 7.14 » Oshawa 7.50 tt Magog 7.56 tt Ottawa 7.84 Government Service Arvida • 8.96 Manufacturing I Sorel** 10.93 Lethbridge** . 11.69 Edmonton** 14.96 Calgary** 17.21 Rouyn 612.92 Extraction Sudbury 918.63 ». Thetford Mines 1139.76 «: Timmins 1400.66 It; Glace Bay 1697.89 tt New Waterford 1952.46 tt *The regular index value i s very high because of the extraction function; this function, however, accounts for less than 5 per cent of city labour force. Because of this, a special index was constructed. It i s similar to the regular index except that extraction has been omitted i n i t s compilation. **Normal specialization index values. .Figure 6. F u n c t i o n a l s p e c i a l i z a t i o n and dominant functions of C a nadian c i t i e s . 143 within this region. In addition, i t reveals a new element not shown by the distribution patterns given i n Chapter II: manufacturing i s the dominant function in many of the peripheral c i t i e s , although not distinctive in them. This means that even i n c i t i e s which are relatively isolated and concerned primarily with servicing a hinterland, manufacturing s t i l l accounts for the largest share of basic employment. It i s hypothesized that most of this manufacturing employ-ment w i l l be accounted for by industrial a c t i v i t y that i s purely market-oriented, and producing for the f i n a l consumer market. Industries that use local, ubiquitous materials such as water, whose products are highly perishable, where the manufacturing process involves weight gain, or where the size of market for one minimum-sized plant i s small (governed by the internal scale economies of the manu-facturing process) w i l l make up the bulk of the manufac-turing activity i n most of the peripheral c i t i e s . Such industries as baking, soft drink, and ice cream manufacturing are representative of this group. It i s not possible to say with certainty that a fundamental difference exists between most heartland and periphery c i t i e s i n the kinds of manufacturing activity associated with each because manufacturing was not broken down on the basis of relationships of inputs and outputs to resources and 144 markets. If such a grouping of the industries making up the manufacturing category was made, however, i t i s thought that the hypothesis would stand; market-oriented manufacturing associated with the f i n a l stages of production w i l l take up a proportionately larger share of city manufacturing act i v i t y i n the periphery than i n the heartland where the industries associated with the earlier and especially the intermediate stages of production for both f i n a l and non-f i n a l markets w i l l be relatively more important. As has been indicated, the heartland c i t i e s with "Manufacturing I" ratings are highly concentrated in two zones. These two city clusters represent the manufacturing heart of Canada. Because the cit i e s of these clusters are highly specialized, (illustrated in Chapter III), i t i s assumed that a high degree of industrial linkage has developed among the industries at both the intercity and intracity level; a l l types of industrial a c t i v i t y occurs from blast furnace activity, to electronics manufacturing and food processing. The c i t i e s within these "systems", although dominated by manufacturing, d i f f e r from "Manufacturing I" c i t i e s i n the periphery i n that the "manufacturing mix* within the heartland c i t i e s i s much more diversified. 145 Cities with large proportions of their total excess employment in manufacturing are the exception in the periphery. They d i f f e r markedly from most peripheral c i t i e s i n that they normally are not concerned with hinter-land servicing, at least not as a primary function. These manufacturing c i t i e s , with the mining centers, stand apart from the other peripheral c i t i e s , and, although based on different a c t i v i t i e s , have many t r a i t s i n common with the extraction centers. The most outstanding t r a i t , perhaps, i s the tendency for both types of centers to be "one-industry c i t i e s " where the dependence on one industrial a c t i v i t y i s extremely high. Without exception the peri-pheral c i t i e s with ratings of "Manufacturing I" are associ-ated with primary, resource-oriented industries. That i s , industries involved i n the i n i t i a l processing of crude materials where there i s a considerable loss of weight and often a high rate of fuel consumption i n the processing. The c i t i e s with "Manufacturing II" ratings are, for the most part, regional capitals or transportation centers where functions other than manufacturing provide the best indicators of the functional role. Because manufacturing, as an undifferentiated functional category, enjoys a position of such overwhelming importance, i t s dominance tends 146 to blanket the distinct and unique features of city functional profiles, thus giving an almost uniform and colourless impression of city functional structure. When studying the functional roles of c i t i e s , distinctive a c t i v i t i e s must be considered as well as dominant functions. This i s especially true when manufacturing i s the dominant, but not a distinctive a c t i v i t y i n a city as i s the case in most c i t i e s with ratings of "Manufacturing II". If the most distinctive functions of this group of c i t i e s are examined, the importance of the "central place" and "transport" functions soon becomes apparent. Wholesale trade, finance, recreation and community service rank high as distinctive functions followed closely by r e t a i l trade, and government service. Transportation ranks high as a distinctive function i n some "Manufacturing II" c i t i e s and serves to isolate those instances where cit i e s owe their existence primarily to transport functions although having manufacturing as the dominant activity. Fort William-Port Arthur provides the classic example of a city that has built up an important manufacturing function based on a transport break-in-bulk point. A l i s t i n g of ci t i e s with their dominant and distinctive functions i s given i n the appendix. It i s observed i n Figure 6 that only two of the 147 forty-five heartland c i t i e s do not have manufacturing as their dominant functions; one i s the national capital, the other an extraction center. The remaining seventeen c i t i e s with functions other than manufacturing being dominant are in the periphery. They are either regional capitals or transportation centers, or a combination of both, except for the extraction centers. Such a distribution i s to be expected i n light of the previous discussions on the fundamental economic differences existing between the heartland and periphery. To summarize, the distribution of dominant functions t e l l s much the same story as those of the distinctive functions. It portrays i n dramatic fashion the importance of manufacturing i n the heartland, delineating clearly the two clusters of manufacturing c i t i e s i n this region. It also reinforces the thesis that the peripheral c i t i e s are primarily concerned with functions associated with distance. This i s indicated by the number of c i t i e s i n the periphery with central place and transport functions as their dominant a c t i v i t i e s . The distribution pattern also introduces a significant additional element: manufacturing i s dominant in many c i t i e s , both in the heartland and the periphery, which are essentially 148 regional capitals. Study of this phenomenon has led to the hypothesis that the kind of manufacturing acti v i t y found i n regional capitals i s fundamentally different from that of the manufacturing c i t i e s making up the "city-systems". The hypothesis suggests that manufacturing acti v i t y associated with f i n a l stages of production, and producing for the consumer market i s relatively more important i n the regional capitals than in the "system" c i t i e s . I l l FUNCTIONAL SPECIALIZATION OF CANADIAN CITIES It was suggested i n Sections I, II and III of Chapter II that city functional specialization i s determined by several factors, the most important being: population size, geographic position, and the functions of the cit y . The application of the modified Ullman-Dacey specialization formula to Canadian c i t i e s has provided a quantified measure of city specialization which allows some of the theories regarding specialization to be evaluated. The relationship of city specialization to dominant  functions and city location. Table X (page 139), gives a l i s t i n g of the c i t i e s and their specialization indexes, ranked by increasing specialization. City specialization 149 i s also shown graphically i n Figure 6 (page 142), where the length of the bars and areas of the circles are proportional to the degree of city specialization. Preliminary observation of these materials reveals that a great range i n value i s experienced by the specialization index—a range of almost two thousand points. Despite this wide range, however, seventy of the eighty c i t i e s have index values f a l l i n g between one and ten, and another four have indexes between ten and twenty. Thus, only six ci t i e s have extremely high values. In each of these six ci t i e s extraction i s the dominant function. This observation would appear to substantiate the statement that the exclusive priority necessarily placed on resource location often precludes the possibility of other city-forming ac t i v i t i e s emerging i n extraction centers, thus promoting extreme specialization i n extraction. In the four ci t i e s with index values between ten and twenty, extraction i s again responsible for the relatively high values. In these c i t i e s , however, extraction accounts for less than 5 per cent of the city labour force, and i s a distinctive function in only one of them^-Lethbridge, 12 already discussed i n Chapter II. In Calgary and Edmonton— ^Because extraction occupies such a small place i n the functional profiles of these ci t i e s special indexes were derived for them. These indexes were constructed i n the same manner as the regular indexes except that the extraction function was omitted. This action gives specialization values more representative of the actual profiles of these c i t i e s . 150 two of the remaining three c i t i e s i n this category—the presence of extraction, albeit i n a low position in both city functional profiles, illustrates an important aspect of these centers* functional position. They are the focal points of the western Canadian o i l and natural gas industries. Almost a l l the employment reported for extraction in the 1951 Census of Canada for these c i t i e s i s associated with o i l prospecting, and crude o i l and natural gas production. The centers act as the head-office locations of numerous companies i n o i l and gas exploration and development. Many o i l d r i l l i n g firms, for example, operate i n the o i l fields from headquarters i n these eities where maintenance f a c i l i t i e s are established and business transactions carried out. Extraction also serves to point out an important functional characteristic of Sorel, the last city having an unrealistically high specialization index due to the extraction function. There i s a close relationship between Sorel and mineral exploitation on the north shore of the St. Lawrence estuary and gulf. Good harbour f a c i l i t i e s exist at Sorel and a smelter in the city processes iron-titanium ores from the north shore. Because of this close relationship i t i s thought that the extraction employment registered for Sorel probably includes many 151 workers whose homes are i n the city, but who work i n north-shore mining camps. Further examination of Table X (page 139), and Figure 6, (page 142) reveals that some clear trends exist i n the association of city functional specialization with city location and dominant functions. Apart from the extraction centers, most of the highly specialized c i t i e s are "Manufacturing I" centers. These c i t i e s are of two kinds: (1). those of the well-integrated "city-systems" i n the heartland, and (2) those purely resource-oriented, relatively isolated c i t i e s of the periphery. As outlined in Chapter II, the propensity for city specialization i s high i n "city-systems" where a high rate of city interaction allows a center to rely on other "system c i t i e s " for both needs and markets. The resource-oriented centers are highly specialized for much the same reasons as are the extraction c i t i e s . They are not located in areas with densely populated hinterlands where central place functions are required to a degree which would offset the manufacturing specialization. Further, because their industrial ac t i v i t y i s concerned with one or few major operations, rather than many as i s the case i n the "city-system" c i t i e s , the l i k e l i -hood of developing diversification even within the manufacturing function through industrial linkage i s small. 152 Specialization also reaches high levels i n a few government centers, notably the national capital. The reason for Ottawa's specialization i s obvious. Without the government function the city would be a relatively small regional capital serving the lower Ottawa Valley. It has attained i t s present size solely through i t s government function; industrial activity i s very insignificant. Victoria and especially Halifax, unlike Ottawa, are important regional capitals besides being government centers. The existence of both administrative and military a c t i v i t i e s i s sufficient, however, to give these c i t i e s a high degree of specialization i n government service. Functional diversification i s greatest among the c i t i e s of the periphery where serving a large hinterland i s of prime importance. These a c t i v i t i e s , plus a certain degree of isolation have worked to make the centers relatively self-sufficient. Charlottetown, for example, i s the least specialized city i n Canada—according to the modified Ullman-Dacey specialization index. It i s also one of the most isolated c i t i e s i n Canada whose primary role i s the servicing of a predominantly agricultural hinterland. The regional capitals of the periphery are closely followed in terms of specialization, by heartland c i t i e s with similar 153 functional structures. The heartland c i t i e s acting primarily as service centers tend to be slightly more specialized than their counterparts i n the periphery. This occurs, i t i s thought, because they are closer to the national "nucleus*, hence, not as isolated. The possibility of them receiving special a c t i v i t i e s such as national headquarters of large firms i s greater than i t i s for peripheral c i t i e s . London, for example, i s essentially a regional capital or service center—-the principal one i n southwestern Ontario. In addition, i t i s also an important financial and insurance center on the national l e v e l . Specialization i s further promoted i n the heartland service centers by their relative proximity to the specialized manufacturing centers of the "city-systems". Generally, manufacturing plays a more significant role i n heartland regional capitals than those of the periphery, as indicated by the dominant functions of c i t i e s . A l l regional capitals i n the heartland are "Manufacturing II** c i t i e s , whereas only about half of the peripheral service centers are so characterized; the balance having transportation, r e t a i l trade, or some other function as the dominant one. This greater emphasis on manufacturing i n the heartland regional capitals i s to be expected; f i r s t , because the heartland i s the "heart" of Canadian manufac-turing; secondly, because the distances between the regional 154 capitals and the centers of the "city-systems" are small and some "overflow" i n industrial linkage can be expected to occur* A quantitative measure of the association of city  specialization with city size, "isolation", and manufacturing  ac t i v i t yo To obtain additional quantitative evidence regarding city specialization, a rather crude technique was devised to examine the relationship of city specialization to: (1) city population, (2) city "isolation", and (3) c i t y manufacturing activity. From previous evidence i t was assumed that a causal relationship exists between city specialization and each of these factors. An easily employed technique, Spearman's coefficient of rank correlation, was used to assess the degree of association existing between city specialization and the three factors l i s t e d . Since c i t i e s had to be ranked by value for each factor, four measures had to be obtained. The modified Ullman-Dacey specialization index was used as a measure of city speciali-zation. To measure city "isolation" the total road distance from a city to i t s four nearest neighbours of ten thousand 1 5 5 1 3 population or over was used • This measure, i t i s readily admitted, i s very crude and open to much criticism; however, because of the great amount of time and effort required to construct a better index of "general accessibility" such as population potential, the crude index was adopted. The percentage of to t a l city excess employment i n manufacturing was taken as a measure of city manufacturing. Two sets of coefficients were calculated. One set was based on the ranking of a l l c i t i e s , the other on a l l c i t i e s except those where extraction or government service are the dominant functions—eleven cities were excluded on this count. Extraction and government centers were omitted because i t i s thought they represent special situations and do not reflect the general spatial-economic elements of the urban milieu to the same extent as other c i t i e s . Their location and degree of specialization are more directly related to individual factors such as; specific resource location, strategic position, and p o l i t i c a l consideration, than to the general economic fabric of the nation. The coefficient values are as follows: ^The total road distance from each city to i t s four nearest neighbours of ten thousand population or over was found by taking road distances from mileage charts on o f f i c i a l government and o i l company road maps. 156 Spearman's Coefficient of Rank Correlation Specialization Specialization Specialization and and and Population "Isolation" Manufacturing A l l c i t i e s (80) 0.088 0.395 0.469 A l l c i t i e s except those where govern-ment service and extraction are dominant (69) 0.105 0.604 0.958 aThe c i t i e s were ranked i n the following manner: (1) by decreasing degree of specialization (2) by increasing population size (3) by increasing t o t a l distance to the four nearest neighbours of ten thousand population and over (4) by decreasing percentage of excess - - employment in manufacturing Comparison of the values for the two sets of correlation coefficients serves to point out the special position of the extraction and government centers. The degree of correlation i s much less when a l l c i t i e s are included i n the rankings than when the government and extraction centers are excluded. This would appear to confirm the suggestion that specialization in such centers i s subject to population and accessibility factors to a much lower degree than i t i s in other types of functional centers. In these special c i t i e s the type of activity 157 occupying the dominant position i s the prime determinant of city specialization. For example, Ottawa, ranking f i f t h in population and forty-sixth i n "isolation", could be expected to be a reasonably well diversified c i t y because of i t s large population and moderately good accessibility* In specialization, however, i t ranks as the eighth most specialized city; only the extraction centers and Arvida— a one industry, resource-oriented "company town"—are more specialized. The coefficient values also reveal that the highest degree of association with city specialization i s given by manufacturing employment, closely followed by "isolation". The close relationship between specialization and manufac-turing i s not unexpected because of the very prominent position this activity plays in urban functional profiles. The relatively close association between degree of city isolation and specialization i s also to be expected in light of the discussion on "city-systems" and the need of regional capitals for "self-sufficiency". The unexpected development i s the low degree of association existing between population size and specialization. Some of the reasons for this probably result from the nature of the indexes used to characterize city functional structure and to measure specialization. Also the use of highly aggregated functional categories undoubtedly creates a bias because no allowance i s made for diversification within categories* Such diversification w i l l undoubtedly increase with city size* It i s thought, however, that the low degree of association between specialization and population also reflects, at least to some extent, the operation of other factors i n addition to city population size in the deter-mination of city specialization. Such factors as the two noted earlier—degree of isolation and the particular "dominant function" of the c i t y — p l a y a very significant role i n determining the degree of functional specialization of a city* For example, the regional capitals of Charlottetown, Rimouski, and Penticton—all under twenty thousand population—are much less specialized functionally than the manufacturing centers of Hamilton and Windsor or the government centers of Ottawa and V i c t o r i a — a l l over one hundred thousand population. To summarize, the following statements, based on the foregoing discussions, are presented: (1) the extraction centers are the most specialized c i t i e s , followed by the manufacturing centers. (2) generally, city specialization i s greatest i n the heartland and least in the periphery,, (3) within the heartland the manufacturing c i t i e s of the "city-systems" are the most specialized, with the exception of one extraction center, (4) some very distinct exceptions occur i n the general pattern of city diversification in the periphery. These exceptions are associated with the extrac-tion and purely resource-oriented manufacturing centers which experience extreme rates of specialization. Two government centers where both legislative-administrative and military functions are found also experience high rates of specialization, (5) specialization i n the heartland regional capitals tends to be sl i g h t l y greater than i n the peri-pheral counterparts due to the greater degree of isolation i n the periphery. Generalizing: (1) city functional specialization i s affected by the degree of isolation a city experiences. As city isolation increases, city specialization usually decreases. 160 (2) the particular "dominant function" of a city affects the rate of city specialization. Centers where extraction or government service are dominant often are much more specialized than would be expected from their degree of isolation and population size. (3) while population size i s a major factor governing the degree of city specialization (illustrated i n Chapter II), i t s influence on specialization rates i s often secondary to that of geographic position (isolation) and the type of dominant and distinctive functions of the city. IV COMMENTS ON "CITY TYPES" The data resulting from the application of the minimum requirement technique of city functional analysis to Canadian ci t i e s are presented in summary form in Table X (page 139), Figure 6 (page 142), and i n the appendix. These substantive materials represent "objectives attained" i n reference to the formal goals of the study. The descriptive notes presented in this section are of a supplementary and summary nature. 1 6 1 Reference was made in previous sections to "regional capitals", "transportation centers", and "specialized manu-facturing c i t i e s " , yet nowhere have these terms been defined. In this section a brief descriptive outline i s presented of Canadian cities viewed in terms of these "city types". The purpose i s to cl a r i f y what i s meant by "specialized manufac-turing city", "regional capital", et cetera, and to ill u s t r a t e further the basic differences that exist between the functional positions of ci t i e s i n the heartland and in the periphery. The terms used here to describe c i t i e s are of a general nature and were not applied with great precision. Only certain aspects of city functional structure were considered when grouping c i t i e s , thus the grouping scheme i s f a i r l y crude. The appendix should be consulted for f u l l details on distinctive and dominant functions, and degree of specialization of individual c i t i e s . Briefly, five groups of c i t i e s were identified on the basis of the importance of wholesale trade, manufacturing, and degree of specialization i n c i t i e s . Obviously, other city groups exist, but different c r i t e r i a are required to identify them. The c r i t e r i a selected and the groups recognized are considered to be the most significant ones at the macro level where the total Canadian urban milieu i s 162 i s being viewed. The influences of the two fundamental elements determining the nature of city functional profiles —distance and concentration—are recognized when using wholesale trade and manufacturing as c r i t e r i a ; each function i s intimately associated with one of the two elements. The percentages of a city's total excess employment (basic employment) i n manufacturing and wholesale trade, respectively were used as measures of importance for the two functions. These values for each city were plotted on an isometric graph (Figure 7), together with values of city population and functional specialization to f a c i l i t a t e the "characterizing" of centers. So that other functions were not completely ignored, special note was made when individual c i t i e s within groups were known to have important functional positions that are not identified by the grouping c r i t e r i a . In this way "transportation", "extraction", and "government" centers were identified. An examination of Figure 7 reveals two familiar patterns: specialization generally increases with increases in the importance of manufacturing, and heartland c i t i e s are generally more specialized than the peripheral c i t i e s . Further examination brings to light several f a i r l y distinct groups of centers. F i r s t , there are thirty-one c i t i e s with high to very high values of manufacturing importance, but Figure 7. "City-types" i n Canada. 164 with moderately low to low ratings for wholesale trade. Since these cities appear, for the most part, to have nearly minimal trading functions and, as a group, exhibit a high rate of specialization, they have been called "specialized manufacturing centers". A l l but four of these cities are located in the heartland. The four peripheral cities are resource-oriented manufacturing centers where the nature of the manufacturing processes, and either: (1) the lack of a large, densely populated hinterland that would promote central place functions, or (2) the assumption of hinterland servicing by a nearby center, have given rise to high specialization rates and a tendency for one industry to be dominant. In the heartland, i t i s the "specialized manufacturing centers" that make up the "city-systems". While a l l these c i t i e s exhibit a high specialization i n manufacturing, a f a i r degree of variation exists within the group. Gshawa, Gait and Magog, for example, present extreme cases of manufacturing specialization. These c i t i e s are almost juxtaposed to larger centers and are, i n a sense, "specialized suburbs" or components of "dispersed c i t i e s " . Victoriaville and Trois Rivieres, on the other hand, are important as regional capitals, although the central place functions are definitely secondary to manu-facturing i n their functional profiles. The city with the 165 lowest rating of manufacturing importance i n the group, Niagara F a l l s , i s a tourist center as well as a manufacturing city . This feature i s revealed by "distinctive" ratings for both the personal service (includes hotel, motel, and restaurant employment) and the recreation functions i n this ci t y . Secondly, there i s a group of eight c i t i e s which are very highly specialized, but have low rates of importance for both manufacturing and wholesale trade. Other functions, obviously, have the significant roles i n these c i t i e s . Because this feature (high specialization rates coinciding with low value in both manufacturing and wholesale trade) i s characteristic of so few c i t i e s , these centers have been called "special c i t i e s " . Included i n this group are highly specialized government and extraction centers. The two most specialized Canadian c i t i e s , the mining centers of New Waterford and Glace Bay, have the lowest ratings of manufacturing importance i n Canada. In some ways they resemble the highly specialized centers of Oshawa and Gait i n that they are components of a "dispersed ci t y " . The two c i t i e s , with Sydney and smaller suburbs, make up the iron and steel complex of Cape Breton Island. The functional structures of the c i t i e s are well integrated with one another. 166 The mining center of Rouyn presents an interesting case because i t s value of importance in wholesale trade i s higher than for any other "special" city. Here i s a case where a highly specialized extraction center also acts as a service center for surrounding areas. A third group of cit i e s i s characterized by relatively high values of wholesale trade importance, low values of manufacturing importance, and generally low levels of specialization. Since city functional diversi-f i c a t i o n ( l i t t l e specialization) and high ratings for trading functions are closely associated with hinterland servicing, the seventeen cit i e s of this group have been called "regional capitals: manufacturing relatively unimportant". Actually, there i s a moderately wide range in wholesale trade importance in this group. It i s the low value of manufacturing importance combined with low specialization rates that set this group apart. The range in wholesale trade importance appears to be related to population size of c i t y . As indicated in Figure 2 (page 69), the importance of wholesale trade in c i t i e s tends to increase with city size. This i s again shown i n Figure 7 where the importance of the trading function i s relatively greater in the large cities than in the small ones. Despite the rather 167 low ratings for wholesale trade i n centers such as Rimouski, Charlottetown and Prince Albert, however, these centers are quite definitely regional capitals. Although their rates for wholesale trade importance are low i n comparison to those of the larger c i t i e s within the group, they are higher than those of many larger ci t i e s outside the group. It i s i n these c i t i e s that the central place functions show their greatest strength i n city functional profiles. Significantly, a l l the c i t i e s of this group are located i n the periphery. These are the c i t i e s most concerned with collecting, transferring, and distributing goods. It i s not surprising to find that a l l five c i t i e s where transportation i s the dominant function are i n this group. Also i n the group are the two c i t i e s where r e t a i l trade i s the dominant function. Several c i t i e s act as p o l i t i c a l capitals as well as regional ones. Included are Edmonton, Regina, Fredericton, and Halifax. The f i r s t three maintain functional profiles typical of regional capitals despite the government function. Halifax, however, because of the presence of a large military establishment i n addition to other government a c t i v i t i e s , has a high rate of specialization. This i s the only atypical specialization index in the group* An additional four eities can be isolated on the basis of population size and relatively high values for both wholesale trade and manufacturing. They are the nation's "major metropolitan centers". The positions of these c i t i e s in Figure 7 symbolizes the two distinct groupings of Canadian c i t i e s : peripheral versus heartland "functional types". Winnipeg and Vancouver, located in the periphery are much more involved with wholesale trade (also transportation) than Toronto or Montreal, situated in the heartland. On the other hand, the latter two ci t i e s emphasize manufacturing to a greater extent than their peripheral counterparts. The implication i s clearj a now familiar thesis i s again stated: the problem of distance requires urban centers i n the periphery to "consume" greater amounts of "economic energy" i n achieving their functional objectives than i s required by the heart-land c i t i e s . In keeping with the position of the heartland, i t s metropolitan centers are the largest c i t i e s i n the nation and their influence i n f i l t r a t e s a l l segments of the national economy. The influence of the two peripheral metropolitan c i t i e s , in contrast, i s restricted to western Canada, The northern and eastern peripheries are essentially appendages of the heartland—as i s the western periphery, 169 but to a lower degree—and they do not have any citi e s which rank as national metropolises. They do have several c i t i e s that could be described correctly as "regional metropolises"; these have been included with the "regional capitals" i n this grouping scheme. The remaining twenty c i t i e s are "mid-way", in terms of specialization and importance of wholesale trade and manufacturing, among the specialized manufacturing centers, the regional capitals where manufacturing i s relatively unimportant, and the special c i t i e s . In this group of citie s are found some of the best examples of the multi-functional nature of city functional structure. It i s perhaps the most heterogeneous group of a l l because there i s no outstanding t r a i t common to a l l the c i t i e s within i t . In the groups previously described a l l c i t i e s share one distinctive characteristic, be i t : (1) a very high degree of specialization, (2) a high degree of manufacturing importance, or (3) a high rate of wholesale trade importance combined with low ratings for manufacturing and speciali-zation. The lack of any real "distinctiveness" i n terms of specialization, wholesale trade or manufacturing means that many of these cit i e s must be differentiated on the basis of other functional characteristics i f detailed 170 analysis i s required. In general terms, the larger centers are regional capitals with a moderate to heavy emphasis on manufacturing, and the smaller c i t i e s , while often serving as regional capitals, do not have very significant whole-sale trade functions. They are frequently involved with other types of activity such as government service and transportation functions. This group of ci t i e s has been called "regional capitals: manufacturing relatively important". The t i t l e i s not entirely satisfactory because the group i s heterogeneous and the name obviously does not apply to a l l i t s members; however, for lack of a better term the name was allowed to stand. Six of the twenty centers in the group are peripheral c i t i e s . They are a l l important regional capitals but also have significant manufacturing functions which are based on either local resources and/or advantages accruing from transportation factors. For example, the existence of local coking coal i s basic to Sydney's iron and steel complex, local clay deposits and natural gas reservoirs provide the principal materials for Medicine Hat's clay products industry, and regional forest resources feed the large pulp and paper operations i n Edmundston. Fort William-Port Arthur and Truro represent cases where trans-portation factors are of prime importance as underpinnings 171 of a city's manufacturing function. Local resources are relatively unimportant in the two c i t i e s as factors in manufacturing location; the advantages accruing to break-in-bulk points and points of "centrality" (Truro i s considered the "hub" of Nova Scotia) clearly provide the basis for manufacturing i n these centers. The existence of flour mills and o i l refineries i n Fort William-Port Arthur, for example, i s directly related to the break-in-bulk feature. These peripheral cities are only slightly more specialized than most of their counterparts which serve as regional capital but where manufacturing i s relatively unimportant. Sydney, however, i s much more specialized than the others because of i t s heavy emphasis on iron and steel. In this respect i t resembles closely the "specialized manufacturing centers". A f a i r amount of functional diversity i s exhibited by the heartland ci t i e s in this group. For example, Trenton, Barrie, and Pembroke are closely associated with military camps; St. Thomas i s an important railway center; Kingston i s perhaps the closest approximation in Canada to an "institutional" center, having important educational, military, religious, and penal f a c i l i t i e s ; Quebec City i s a p o l i t i c a l capital; and O r i l l i a i s the site of a large 172 mental hospital. Because of this diversity i t i s d i f f i c u l t to formulate general statements on these c i t i e s . Of note, however, i s the fact that only two of the fourteen heartland c i t i e s i n this group are members of the "city-systems"; the great majority are outside the "Systems" and are not typified by high specialization indexes as are "city-system" centers. Even the two c i t i e s within the "systems" are highly diversified in comparison to other "system" c i t i e s . Joliette i s the least specialized of the "city-system" centers, and only two other "system" cit i e s besides Joliette are more diversified than Sherbrooke. In essence, most of these heartland ci t i e s are basically regional capitals exhibiting functional profiles not too dissimilar from those of their counterparts in the periphery. Their rates of specialization, for example, are closer to those of the peripheral service centers than to the ones of "system" c i t i e s . As stated i n the preceding section, the propensity for obtaining special functions and moderately high rates of manufacturing activity i s greater for the heartland regional capitals than for those of the periphery. This greater opportunity of obtaining additional city-forming functions enjoyed by the heartland regional capitals has tended to produce more complex functional structures because the additional functions are superimposed on the basic "service-center type" of functional structure. Variety in the types of functional profiles i s hence greater among the heartland regional capitals than i n the periphery, even though manufacturing appears as the dominant function for a l l heartland regional centers. Summary: Accepting the limitations stated on page 161, this technique of city grouping—including the graphic portrayal in Figure 7—supplies a useful means of obtaining an overview of urban character in Canada. Although i t by no means allows an exhaustive analysis of city functional structure to be made, i t il l u s t r a t e s clearly the influence of the fundamental forces affecting city functional performance i n Canada: distance and concentration. CHAPTER IV SUMMARY AND CONCLUSIONS I. THE CANADIAN URBAN MILIEU: AN OVERVIEW The functional structure of cit i e s i n the Canadian "heartland" differs fundamentally from that of ci t i e s i n the "periphery" of Canada, The heartland c i t i e s are functionally more specialized and emphasize manufacturing to a greater degree than do the peripheral c i t i e s . The latter c i t i e s are generally quite diversified and are inclined to have a relatively more important involvement with functions associated with distance such as wholesale trade and transportation, although manufacturing i s s t i l l important i n their profiles. Several highly specialized ci t i e s do, however, exist i n the periphery; without exception they are associated with either the extraction function or resource-oriented manufacturing. These ci t i e s are few i n number and do not represent the norm, although they are v i t a l i n the national economy. In order to explain the dichotomous nature of c i t y functional profiles—heartland versus peripheral c i t i e s -several hypotheses have been stated including: 175 (1) heartland c i t i e s are generally more specialized  than peripheral cities because functional  specialization decreases with increasing geo- graphic isolation. Division of labour, the basis of city functional specialization, i s a function of distance. Increasing the f r i c t i o n of distance increases the d i f f i c u l t y of attaining high rates of division-of-labour because exchange of goods i s essential to a continuing system of specialized production. Specialization, there-fore, decreases as d i s t a l f r i c t i o n increases. Relatively isolated c i t i e s cannot rely on neighbouring centers to supplement their requirements to the extent that c i t i e s i n "city-systems" can. Moreover, most c i t i e s i n comparatively insular positions are primarily concerned with hinterland servicing where numerous urban functions come into play. This situation tends to produce self-sufficient and diversified c i t i e s . In contrast, a city located within a cluster of c i t i e s can rely on nearby centers for both needs and markets. The f r i c t i o n of distance i s relatively inconsequential, thus 176 allowing the f u l l advantages of division-of-labour to operate and thereby promoting functional specialization. In addition, because so many ci t i e s are located within a small area, the "diversifying" effect associated with hinterland servicing i s insignificant since the amount of such servicing performed by any one city of the cluster i s very small. Evidence resulting from a crude test of association between specialization and geographic isolation verified the applic-a b i l i t y of this hypothesis to a degree. The hypothesis i s not applicable, however, i n the case of the purely resource-oriented manufacturing peripheral c i t i e s or the extraction centers. This situation prompted the stating of a second hypothesis. ( 2 ) high specialization rates are experienced i n peripheral resource-oriented manufacturing  ci t i e s and extraction centers primarily because  of; (a) the lack of additional functional  opportunities, and (b) the nature of the manu-facturing and extraction processes. For these c i t i e s , resource location usually received an exclusive priority i n the selection of the city s i t e . Often this feature works against the situational requirements of other urban functions. The desired resource site may be located i n an area without a populated hinterland or where hinterland serving i s adequately managed by other centers. Secondly, most resource-oriented manu-facturing processes are at the primary stages of production. These stages usually require very large plants—even at the "minimum-size" l e v e l -i n order to be economic. This i s a consequence of high levels of wastage, high energy require-ments, and relatively low unit-values of output i n these manufacturing processes. For this reason many resource-oriented centers are completely dominated by one large plant, thus discouraging the development of diversification even within the manufacturing function through industrial linkage. Variation i n city functional structure has been found to be a function of distance, concentration, and type of manufacturing activity. In essence, the basic differences i n c i t y functional development stem from the simple fact that "the character of a city's functional profile reflects the character of the region the city serves". The Canadian 178 heartland and peripheral areas are geographic r e a l i t i e s . They d i f f e r greatly i n his t o r i c a l , economic, and to some degree, i n cultural development. The fact that their c i t i e s reflect these differences seems like a reasonable and torbe-expected conclusion. II. CITY FUNCTIONAL CLASSIFICATION: RETROSPECT AND PROSPECT In terms of technique, the most significant feature of the study i s that the application of a purely objective method of classification, based on census s t a t i s t i c a l data, has produced a general classification of Canadian cit i e s which i s rational i n light of background knowledge of the Canadian urban milieu. There are certain weaknesses in the study scheme, however. For example, i t was impossible to make any definite statements regarding the differing nature of the manufacturing activity i n heartland and peripheral c i t i e s , although certain hypotheses were entertained. This was a consequence of the failure to break down the manu-facturing activity into components based on the relationship of manufacturing inputs and outputs to resources and markets. Such a situation i s characteristic of most classification schemes which u t i l i z e census statistics as the sole source 179 of data i n the classification problem. In general terms, the use of census industry divisions as city functional categories in these studies, has meant that no direct relationship has been established between factors of city location and city functional characteristics. This limita-tion i s perhaps the most serious shortcoming of this study. It stems, not from any fault of the particular classification technique applied, but from weaknesses in the traditional conceptual approaches to city classification. Too often the motivation behind the production of a classification scheme has not been made clear, thus giving the impression that the scheme has been created simply for the sake of making a classification of c i t i e s . The quantitative classification schemes developed (including the one presented in this study) have, i t i s concluded, placed undue emphasis on the detailed manipulations of data, at the expense of conceptual considerations as to how these studies can best be applied in advancing the knowledge of the "location-function" relationships of c i t i e s . To be sure, most classification studies have attempted to analyze and explain these relationships through cartographic portrayal, but too often this has been an isolated and secondary exercise quite divorced from the classification 180 scheme i t s e l f . City functional classifications developed without some basic association with location factors, tend to be sterile inventories. What i s required are studies where the city functional categories imply definite locational characteristics, and which use city c l a s s i f i -cation as a tool i n the solution of specific problems related to the location and socio-economic characteristics of urban centers. If city functional classification i s accepted as an essential tool i n the analysis of problems associated with the type and spacing of urban phenomena, two areas of investigation that would lead to improved and more flexible classification schemes suggest themselves from this study. There are two limiting factors i n the quantitative classification schemes discussed. One has been mentioned— the use of census industry divisions as city functional categories. The census groupings at the intermediate and higher levels of aggregation tend to be quite heterogeneous and arbitrary. It i s obvious that the practice of using them as city functional categories, while convenient, does not give city functional groupings that have much significance i n terms of location factors. For example, to know that the dominant function of a city i s "manufacturing" t e l l s one 181 l i t t l e because this function i s the dominant activity i n most c i t i e s . The significant item to know i s what kind of manufacturing i s present i n the functional pr o f i l e , i n terms of markets and resources. Is i t resource-oriented, market-oriented or "footloose"? Is i t producing for the f i n a l or non-final market? What stages of production are involved, and what i s the propensity for the development of industrial linkage? The realignment of census material to give this kind of information would f a c i l i t a t e greatly the integration of data on city location and function. Investigation into ways and means by which this realignment can be achieved would, i t i s suggested, prove to be a profitable avenue of research. Progress along these lines has already been made i n the United States by researchers of Resources for the Future, Inc.* They have compiled a table, based on the 1947 Interindustry Relations  Study of the United States Bureau of Labour Statistics, which shows industries grouped into four classes on the basis of their relationship to raw materials. These classes are: Primary resource extractors F i r s t stage resource users ^Harvey S. Perloff et a l . , Regions. Resources, and  Economic Growth, (Baltimore! fhe Johns Hopkins Press, I960)• Appendix Table L, pp. 677-680, 182 Second stage resource users Industries for which resources have the most indirect significance, Otis Duncan and associates have used these industrial groupings in their study of metropolitan centers i n the 2 United States, By u t i l i z i n g additional information given i n the Resources for the Future table they were able to further classify industries on the basis of whether they produced primarily for a f i n a l or non-final market. While this scheme represents a significant step forward i n the study of "location-function" relationships of c i t i e s , i t i s not readily applicable to service and cultural categories as l i s t e d i n the census. Further research i s required to develop groupings of these a c t i v i t i e s which w i l l demon-strate the significant relationships that exist among functions within a city, between a city and i t s hinterland, and between the city and other c i t i e s . Once the realign-ment of the detailed census categories i s achieved, i t w i l l be possible to formulate meaningful functional groups with reasonable ease. The second limiting factor present i n the quantitative cit y classification schemes discussed, stems from the problem of selecting "functional standards" against which the i n d i -20tis Duncan et a l . . Metropolis and Region, (Baltimore: The Johns Hopkins Press, I960), pp. 200-209. 16*3 vidual city profiles are evaluated. Despite the "objective-ness" of these studies, subjectivity i s s t i l l very much involved when the "standards" are being established. For this reason the results of different studies vary. As an example of this, the results of a pilot study i n the functional classification of Canadian c i t i e s d i f f e r from the findings given i n this report. The same source material was used i n both studies, and the functional categories recog-nized were almost identical but different classification techniques were used. A classification scheme similar to J.W. Webb's was used i n the pilot study, while the minimum requirement technique was applied i n the f i n a l study. The greatest variance i n the results of the two studies occurs in the ranking of ci t i e s by the degree of functional specialization. Sydney, for example, ranks as the least specialized city according to the modified Webb method, but i t i s the twenty-eighth most specialized c i t y (out of a total of eighty cities) when the minimum requirement technique of classification i s applied. Perhaps i t w i l l never be possible to develop "standards" without the presence of some subjective elements i n the considerations, indeed, i t probably i s not even desirable. It i s desirable, however, to minimize the necessity of making arbitrary decisions, when establishing c r i t e r i a . The pos s i b i l i t y of 184 doing this has been greatly increased with the introduction of systems of multivariate analysis into the geographical literature. Multivariate techniques of analysis, not only allow the "testing" of variables as to their su i t a b i l i t y as measures of city functional character, but with electronic computer f a c i l i t i e s at hand, expand the number and types of characteristics that can be considered when typifying city functional profiles, • • ,the distinction between our process of classification and one that makes use of simpler methods, such as dividing towns according to their chief industry, i s that we have taken into account a very much wider set of characteristics and that the c r i t e r i a of classification emerged from the analysis i t s e l f . ^ This quote i s from a study that has pioneered the use of multivariate analysis in the classification of urban centers. Studies such as this appear to have a great potential in the development of classifications which u t i l i z e meaningful functional categories and are sufficiently flexible to be A. Moser and Wolf Scott, British Towns. A  s t a t i s t i c a l study of their social anct economic differences, Centre for Urban Studies. Report No, 2 (London: Oliver and Boyd, 1961), Component analysis was used i n this study. Another study in this same general area i s : Leslie J. King, "A Multivariate Analysis of the Spacing of Urban Settlements i n the United States", Annals of the Association of American Geographers, LI (March, 1961), 222-233. King has used multiple regression analysis in an attempt to "explain" the spacing of centers. He introduced certain urban functional characteristics into his analysis. 185 adapted to studies where specific problems in city "location-function" relationships are being solved. Additional studies along these lines would prove worthwhile. In so far as the Canadian macro urban scene i s concerned, a specific project i s suggested: A study which would relate the degree of city functional specialization to: (a) city manufacturing activity classified by the relationship of inputs and outputs to resources and markets, (b) "general accessibility" using "population - - potential" as the measure of accessibility, (c) city and hinterland populations. Such a study would require the development of city functional categories which are related to locational characteristics, and, simultaneously, would provide an excellent opportunity for applying techniques of multi-variate analysis i n evaluating the degree of city specialization "explained" by the several factors l i s t e d . Such a project can be considered a logical next step i n the analysis of urban "location-function" relationships at the macro scale. While such a program was beyond the scope of this study, several significant steps towards a better under-standing of both city functional classification and the Canadian urban milieu have been made in this study. The 186 review of the traditional classification techniques has revealed or re-emphasized the importance: (1) of dealing with c i t i e s as multifunctional entities, (2) of recognizing the dichotomous nature of urban functions (city-forming and city-serving) when classifying c i t i e s , (3) of considering the effect of changing city size on c i t y functional structure, and (4) of analyzing the performance of functions on two distinct planes—the relative and the absolute ("distinctive" and "dominant" functions). The quantitative classification techniques have been scrutinized and the minimum requirement method as developed by Ullman and Dacey has been found to be the technique conforming most closely to theoretical concepts on city structure. However, before the method could be applied to determine city specialization i t was found necessary to modify the Ullman-Dacey formula to satisfy mathematical requirements* Finally, the analysis of the distribution patterns of functional relative importance (distinctive functions), dominant functions, and city functional specialization led to the development of the "heartland-periphery" concept i n the Canadian context. Cities differ fundamentally i n functional structure depending on their location i n either the heart-land or the periphery. Part of the difference, especially is? the difference i n city functional specialization, i s considered to be associated with the degree of geographic isolation and the role of manufacturing i n the functional profile of c i t i e s . The study has presented a comprehensive view of city functional performance at the national level i n Canada using traditional techniques of city functional c l a s s i f i -cation. From this overview have evolved several hypotheses which received preliminary investigation. Further analysis awaits later study employing more sophisticated and powerful techniques of investigation. This study i n some respects marks a beginning i n the study of the urban milieu at the national level i n Canada. BIBLIOGRAPHY BIBLIOGRAPHY Alexander, John W. "The Basic-Nonbasic Concept of Urban Economic Functions," Economic Geography, XXX (July, 1954), 246-61. Alexander, John W. Economic Geography. Englewood C l i f f s , New Jersey: Prentice-Hall, Inc.,1963. Chap. XXIV, PP. 537-644* Alexander, John W., and James B. Lindburg. "Measurements of Manufacturing: Coefficients of Correlation," Journal of  Regional Science. I l l (Summer, 1961), 71-81. Alexandersson, Gunnar. The Industrial Structure of American Cities. Lincoln, Nebraska: University of Nebraska Press, 19567" Aurousseau, M. "The Distribution of Population: A Con-structive Problem," Geographical Review, XI (October, 1921), 563-92. Dickinson, Robert E. City Region and Regionalism. London; Kegan Paul, French, Trubner &Go., Ltd., 1947. Dickinson, Robert E. "The Scope and Status of Urban Geography," Land Economics. XXIV (August, 1948), 221-38. ' Dominion Bureau of Statistics. Ninth Census of Canada: 1951. Labour Force. Vol. IV, Tables 17 and 21. Ottawa: Queen's Printer, 1953. Duncan, Otis D., et a l . Metropolis and Region. Baltimore: The Johns Hopkins Press, I960. Duncan, Otis D,, and Albert J. Reiss, Jr. Social Charac- t e r i s t i c s of Urban and Rural Communities! 19^0. New York: John Wiley & Sons, Inc., 1956. \ Florence, P.S., W.G. F r i t z , and R.C. Gilles. "Measures of Industrial Distribution," Industrial Location and  National Resources t United States National Planning Board. Washington, D.C.: U.S. Government Printing Office, 1943. Chap. V. 190 Harris, Chauncy D. "A Functional Classification of Cities i n the United States," Geographical Review, XXXIII (January, 1943), 86-99. Harris, Chauncy D,, and Edward L. Ullman. "The Nature of Cit i e s , " Annals of the American Academy-of P o l i t i c a l  and Social Science, CCXLII (November, 1945), 7-17. Hart, John F. "Functions and Occupational Structures of Cities of the American South," Annals of the Association  of American Geographers. XLV (September, 1955), 269-286, Hoover, Edgar M, The Location of Economic Activity, New York: McGraw-Hill Book Company, Inc., 1948, Isard, Walter, et a l . Methods of Regional Analysis: an  Introduction to Regional Science? Cambridge, Massachu-setts : The Massachusetts Institute of Technology Press, I960. Jones, Victor. "Economic Classification of Cities and Metropolitan Areas," The Municipal Year Book. 1954. Chicago: The International City Manager's Association, 1954. PP. 31-36, 62-70, 81-108. King, Leslie, J. "A Multivariate Analysis of the Spacing of Urban Settlements i n the United States," Annals of  the Association of American Geographers, LI, (March, 1961), 222-33. Kneedler, G.M. "Economic Classification of Ci t i e s , " The  Municipal Year Book. 1945. Chicago: The International City Manager's Association, 1945. pp. 30-38, 48-68. Kosinski, L. "Problem of the Functional Structure of Polish Towns," Przeglad Geograficzny, XXI (Supplement, 1959), 35-67! Kostrowicki, J. "Basic Functions and Functional Types of Towns," Przeglad Geograficzny. XXIV (1952), 7-64. Mayer, Harold M., et a l . "Urban Geography," i n Preston E, James, and Clarence F. Jones (eds.). American Geography:  Inventory and Prospect. Syracuse, New York: Syracuse University Press, 1954. Chap. VI, pp. 143-66. . 191 Morrissett, Irving. "The Economic Structure of American Cit i e s , " Papers and Proceedings of the Regional Science  Association, IV (1959). 239-59. : Moser, C.A., and Wolf Scott. British Towns. A s t a t i s t i c a l  study of their social and economic differences. Centre for Urban Studies, Report No. 2. London: Oliver and Boyd, 1961. Mumford, Lewis. "The Natural History of Urbanization," i n William L. Thomas, Jr. (ed.). Man's Role i n Changing the Face of the Earth. Chicago! The University of Chicago Press, 1956. pp. 382*98. Nelson, Howard J. "A Service Classification of American C i t i e s j " Economic Geography, XXXI (July, 1955), 189-210. Palomaki, Mauri. The Functional Centers and Areas of South  Bothnia, Finland. Department of Geography, University of Helsinki, Publication No. 35. Vammala, Finland: Vammalan Kirjopaino Oy., 1963. Perloff, Harvey S., et a l . "Location Factors," Regions,  Resources, and Economic Growth. Baltimore: The Johns Hopkins Press, I960. Chap. VI, pp. 75-86. Pfouts, Ralph W. The Techniques of Urban Economic Analysis. West Trenton, New Jersey: Chandler-Davis Publishing Company, I960. Pownall, L.L. "The Functions of New Zealand Towns," Annals  of the Association of American Geographers. XLIII (December, 1953), 332-50. Rat c l i f f , Richard U. Urban Land Economics. New York: McGraw-Hill Book Company, Inc., 1^49 • Robinson, Ira M. New Industrial Towns on Canada's Resource  Frontier. Department of Geography, University of Chicago, Research Paper No. 73. Chicago: University of Chicago Press, 1962. Rodgers, Allan. "Some Aspects of Industrial Diversification i n the United - States," Papers and Proceedings of the Regional Science Association, I (1955). B-I—B-II. 192 Rodgers, Allan, "Some Aspects of Industrial Diversification i n the United States," Economic Geography, XXXIII (January, 1957), 16-30, Rodgers, Allan, "Regional Industrial Development with reference to Southern Italy," Essays on Geography and  Economic Development, Norton Ginsburg, Department of Geography, University of Chicago, Research Paper No, 62. Chicago: University of Chicago Press, I960, Chap. IX, pp. 143-73. Smailes, Arthur E. "The Urban Hierarchy in England and Wales," Geography. XXIX (June, 1944), 41-51. Smailes, Arthur E. "The Urban Mesh of England and Wales," Transactions and Papers of the Institute of British -Geographers. (1946). 87-101. ' Spelt, Jacob. The Urban Development i n South Central  Ontario. Assen, The Netherlands: Van Gorcum & Company, 1955. Steigenga, W. "A Comparative Analysis and a Classification of Netherlands Towns," Ti.jdsckrift voor Economische en  Sociale Geografic. XLVI (l$55), 10S-19. . . . . " Tiebout, Charles M. The Community Economic Base Study. Committee for Economic Development, Supplementary Paper No. 16, New York, 1962. Tress, R.G. "Unemployment and Diversification of Industry," The Manchester School. IX (1938). Trotier, Louis. "Some Functional Characteristics of the Main Service Centers of the Province of Quebec," Cahiers de  Geographie de Quebec. No, 6 (April-September, 1959), Ullman, Edward L, "Regional Development and the Geography of Concentration," Papers and Proceedings of the Regional Science Association. IV (1958). 179-98. . . . . Ullman, Edward L., and Michael F. Dacey, "The Minimum Requirements Approach to the Urban Economic Base," Papers and Proceedings of the Regional Science  Association, VI (I960). 175-94. ' * 193 "Urban Functions, n Economic Geography. XXI (April, 1945), 78. Watanabe, T. "An analysis of the Function of Urban Settle-ments Based on S t a t i s t i c a l Data - A Functional Differentiation Vertical and Lateral," The Science  Reports of the Tokoku University (Geography), No. 10 (September, 1961), 63-94* Webb, John W* "Basic Concepts in the Analysis of Small Urban Centers of Minnesota," Annals of the Association of American Geographers. XLIX (March, 1959), 55-72* APPENDIX APPENDIX DOMINANT AND DISTINCTIVE FUNCTIONS, AND SPECIALIZATION INDEXES FOR CANADIAN CITIES OF 10,000 POPULATION AND OVER, 1951 WESTERN PERIPHERY City Special- Dominant Distinctive Functions ization Function CLASS Index I :; II : III Lethbridge* 1.25 Retail Trade *1.25 i s the special index value for Lethbridge; ex-traction was omitted i n i t s construction. The city's regular specialization index value i s 11.69• Retail Trade Construction Wholesale Trade Finance, etc. Recreation Personal Service Extraction Public U t i l i t i e s Transportation, etc. Business Service Prince Albert 1.40 Manufacturing II Recreation Community Service Construction Transportation, etc. Wholesale Trade Retail Trade Finance, etc. Government Service Business Service Personal Service Brandon 1.41 Transportation, etc. Transportation, etc. Wholesale Trade Finance, etc. Community Service Recreation Public U t i l i t i e s Retail Trade Government Service Business Service Personal Service 196 Penticton 1.44 Transportation, etc. Saskatoon 1.51 Manufacturing II Moose Jaw 1.56 Transportation, etc. Calgary* 1.77 Manufacturing II *1.77 i s the special index value for Calgary; extrac-tion has been omitted i n i t s construction. The city's regular specialization index value i s 17.21. Edmonton* 1,80 Manufacturing II *1.80 i s the special index value for Edmonton; extraction has been omitted i n i t s construction. The city's regular specialization index value i s 14.96. Recreation Personal Service Construction Transportation, etc. Wholesale Trade Retail Trade Finance, etc. Business Service Wholesale Trade Finance, etc. Community Service Transportation, etc. Retail Trade Government Service Recreation Business Service Personal Service Transportation, etc. Recreation Public U t i l i t i e s Construction Wholesale Trade Retail Trade Finance, etc. Community Service Personal Service Wholesale Trade Finance, etc. Recreation Public U t i l i t i e s Construction Government Service Business Service Personal Service Construction Wholesale Trade Transportation, etc. Finance, etc. Government Service Business Service 197 Medicine Hat 1.97 Manufacturing II Recreation Transportation, etc* Construction Retail Trade Finance, etc. Government Service Personal Service Winnipeg 2.00 Manufacturing II Fort William-Port Arthur 2.08 Manufacturing II Wholesale Trade Finance, etc. Transportation, etc. Recreation Transportation, etc. Public U t i l i t i e s Construction Wholesale Trade Recreation Regina 2.09 Government Service Wholesale Trade Finance, etc. Government Service Public U t i l i t i e s Transportation, etc. Retail Trade Community Service Business Service Personal Service Vancouver 2.40 Manufacturing II Victoria T r a i l Lethbridge Edmonton Calgary 4.01 Government Service 6.86 Manufacturing I 11.69 ( 14.96 ( given above 17.21 ( Wholesale Trade Finance, etc. Transportation, etc. Business Service Government Service Finance, etc. Community Service Recreation Manufacturing Recreation Business Service 198 NORTHERN PERIPHERY City Special- Dominant Distinctive Functions ization Function CLASS Index I : II : III Chicoutimi 1.72 Manufacturing II North Bay 2.35 Transportation, etc. Jonquiere 4.40 Manufacturing I Sault Ste. Marie 4.70 Manufacturing I Arvida Rouyn 3.96 Manufacturing I 612.92 Extraction Sudbury 918.63 Extraction Community Service Construction Retail Trade Public U t i l i t i e s Wholesale Trade Business Service Personal Service Transportation, etc. Public U t i l i t i e s Construction Wholesale Trade Retail Trade Finance, etc. Community Service Personal Service Manufacturing Construction Retail Trade Personal Service Manufacturing Transportation Recreation Manufacturing Community Service Extraction Recreation Personal Service Construction Wholesale Trade Business Service Extraction Construction Recreation Personal Service 199 Timmins 1400.66 Extraction Extraction Public U t i l i t i e s Retail Trade Recreation EASTERN PERIPHERY City Special* Dominant Distinctive Functions ization Function CLASS Index I : II : III Charlottetown 1.16 Retail Trade Retail Trade Recreation Personal Service Wholesale Trade Finance, etc. Community Service Public U t i l i t i e s Construction Transportation, etc. Government Service Business Service Rimouski 1.36 Community Service Moncton 1.55 Transportation, etc. Community Service Construction Personal Service Public U t i l i t i e s Business Service Transportation, etc. Wholesale Trade Retail Trade Finance, etc. Transportation, etc. Retail Trade Wholesale Trade Finance, etc. Community Service Government Service Business Service Personal Service 200 Saint John 1.6? Manufacturing II St. John's 1.75 Manufacturing II Truro 1.77 Manufacturing II Fredericton 1.86 Government Service Edmundston 2.35 Manufacturing II Sydney 5»21 Manufacturing I Transportation, etc. Wholesale Trade Recreation Finance, etc. Community Service Government Service Personal Service Wholesale Trade Government Service Transportation, etc. Retail Trade Community Service Personal Service Personal Service Transportation, etc. Retail Trade Wholesale Trade Finance, etc. Community Service Recreation Business Service Public U t i l i t i e s Wholesale Trade Retail Trade Government Service Personal Service Construction Finance, etc. Community Service Recreation Business Service Personal Service Transportation Community Service Construction Retail Trade Manufacturing Transportation Wholesale Trade Retail Trade Recreation Personal Service I 201 Halifax 5•46 Government Service Glace Bay 1697•39 Extraction New Waterford 1952,46 Extraction Government Service Public U t i l i t i e s Construction Transportation Wholesale Trade Finance, etc. Extraction Extraction Retail Trade Recreation HEARTLAND City Special-ization Index Dominant Function Distinctive Functions CLASS I : II : III Barrie 1.87 Manufacturing II Public U t i l i t i e s Retail Trade Finance, etc. Government Service Construction Transportation, etc. Recreation Business Service Personal Service Bel l e v i l l e 2.04 Manufacturing II Recreation Transportation, etc. Retail Trade Public U t i l i t i e s Finance, etc. Government Service Business Service Personal Service Quebec 2.11 Manufacturing II Construction Wholesale Trade Finance, etc. Community Service Government Service 202 St, Thomas 2,14 Manufacturing II Transportation, etc. Retail Trade Public U t i l i t i e s Community Service Recreation Business Service Personal Service Kingston 2,17 Manufacturing II Community Service Government Service Retail Trade Finance, etc. Recreation Business Service London 2,17 Manufacturing II Finance, etc. Public U t i l i t i e s Construction Wholesale Trade Community Service Government Service Chatham 2,26 Manufacturing II Public U t i l i t i e s Retail Trade Recreation Construction Transportation Wholesale Trade Finance, etc. Community Service Business Service Personal Service Brockville 2,71 Manufacturing II Community Service Manufacturing Transportation, etc. Retail Trade Finance, etc. Recreation Business Service Personal Service 203 Owen Sound 2.73 Manufacturing II O r i l l i a 2.74 Manufacturing II Joliette 2.88 Manufacturing I Pembroke 3.41 Manufacturing II Toronto 3.51 Manufacturing I Trois Rivieres 3.52 Manufacturing I Wholesale Trade Retail Trade Manufacturing Public U t i l i t i e s Transportation, etc. Finance, etc. Recreation Business Service Retail Trade Community Service Recreation Manufacturing Public U t i l i t i e s Construction Finance, etc. Business Service Personal Service Community Service Manufacturing Public U t i l i t i e s Construction Retail Trade Business Service Personal Service Business Service Retail Trade Government Service Public U t i l i t i e s Construction Community Service Recreation Personal Service Finance, etc. Public U t i l i t i e s Wholesale Trade Business Service Manufacturing Public U t i l i t i e s Construction Community Service 204 Sherbrooke Niagara Falls Trenton Stratford Sarnia Montreal St, Hyacinthe Woodstock Cornwall 3.61 Manufacturing I 3.62 Manufacturing I 3.63 Manufacturing II 3.87 Manufacturing I 3.90 Manufacturing I 3.92 Manufacturing I 3.96 Manufacturing I 4.14 Manufacturing I 4.16 Manufacturing I Manufacturing Construction Wholesale Trade Community Service Public U t i l i t i e s Personal Service Manufacturing Recreation Business Service Grovernment Service Construction Transportation, etc. Retail Trade Recreation Personal Service Manufacturing Public U t i l i t i e s Transportation, etc. •Retail Trade Finance, etc. Construction Business Service Manufacturing Public U t i l i t i e s Transportation, etc. Finance, etc. Wholesale Trade Community Service Manufacturing Construction Manufacturing Public U t i l i t i e s Retail Trade Community Service Manufacturing Retail Trade Recreation Personal Service Guelph Shawinigan St, Catharines 4.22 Manufacturing I 4*53 Manufacturing I 4,56 Manufacturing I Kitchener-Waterloo 5.36 Manufacturing I Windsor Peterborough St• Jean St« Jerome Vic t o r i a v i l l e Brantford Hamilton Sorel* 5.41 Manufacturing I 5.42 Manufacturing I 5.55 Manufacturing I 5.56 Manufacturing I 5.59 Manufacturing I 5.6? Manufacturing I 5.88 Manufacturing I 6,05 Manufacturing I *6,05 i s the special index value for Sorel; extraction was omitted i n the construction. The city's regular specialization index value i s 10,93. 205 Manufacturing Community Service Public U t i l i t i e s Construction Manufacturing Manufacturing Construction Business Service Finance, etc. Manufacturing Manufacturing Manufacturing Finance, etc. 1 Government Service Manufacturing Manufacturing Construction Manufacturing Public U t i l i t i e s Construction Wholesale Trade Manufacturing Business Service Manufacturing Manufacturing Drummondville 6,24 Manufacturing I Manufacturing Personal Service Granby 6.28 Manufacturing I Manufacturing Construction Personal Service Grand'Mere Valleyfield We Hand Gait Oshawa Magog Ottawa 206 6,34 Manufacturing I Manufacturing Public U t i l i t i e s Construction 6,39 Manufacturing I 6.54 Manufacturing I 7.14 Manufacturing I 7.50 Manufacturing I 7.56 Manufacturing I 7.84 Government Service Public U t i l i t i e s Manufacturing Construction Manufacturing Business Service Manufacturing Manufacturing Manufacturing Government Service Finance, etc. Recreation Thetford Mines 1139.76 Extraction Extraction Public U t i l i t i e s Retail Trade 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

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"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0104958/manifest

Comment

Related Items