Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Variation in resistance and tolerance of black cottonwood to Melampsora occidentalis (Jacks) rust Wang, Jun 1991

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

Item Metadata

Download

Media
831-UBC_1991_A1 W36.pdf [ 7.08MB ]
Metadata
JSON: 831-1.0101045.json
JSON-LD: 831-1.0101045-ld.json
RDF/XML (Pretty): 831-1.0101045-rdf.xml
RDF/JSON: 831-1.0101045-rdf.json
Turtle: 831-1.0101045-turtle.txt
N-Triples: 831-1.0101045-rdf-ntriples.txt
Original Record: 831-1.0101045-source.json
Full Text
831-1.0101045-fulltext.txt
Citation
831-1.0101045.ris

Full Text

VARIATION IN RESISTANCE A N D T O L E R A N C E O F B L A C K COTTONWOOD T O  MELAMPSORA  (JACKS) RUST  OCCIDENTALS By  J u n Wang B.Sc. Nanjing Forestry University  A T H E S I S S U B M I T T E D IN P A R T I A L F U L F I L M E N T O F T H E R E Q U I R E M E N T S FOR T H E D E G R E E OF DOCTOR OF PHILOSOPHY  in T H E F A C U L T Y OF G R A D U A T E STUDIES FORESTRY  We accept this thesis as conforming to the required standard  T H E U N I V E R S I T Y O F BRITISH C O L U M B I A 1991  © J u n Wang,  1991  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 permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department  or  by  his  or  her  representatives.  It  is  understood  that  copying  or  publication of this thesis for financial gain shall not be allowed without my written permission.  Department The University of British Columbia Vancouver, Canada  DE-6 (2/88)  11  ABSTRACT  Ramets of 14 clones of western black cottonwood Populus  trichocarpa  Torr. & G r a y collected from the interior and coast of British Columbia were tested in the nursery for their growth performance after infection by the leaf rust Melampsora  occidentalis  Jacks. Linear relationships were demonstrated  between various growth parameters (total dry weight, stem dry weight, root dry weight, volume, diameter and height) and disease severity rated  as  diseased leaf-weeks divided by total leaf-weeks. Losses due to rust infection included reduced total dry weight and volume growth i n the year of heavy disease, the death of severely infected ramets during the following winter and reduced initial volume increment in the following growing season. The percentage  reduction in yield (total dry weight) was  greater than  the  cumulative percent leaf area infected, suggesting that the rust infected leaf parts act as sinks for photosynthate. The normal pattern of photosynthate allocation was altered in favour of the top growth of ramets. stem/root  dry  weight  increased  rapidly as  disease  Ratios of  level increased.  A  threshold infection level, below which no loss occurred, was not detected i n this pathosystem.  Significant variation in rust resistance of black cottonwood clones was detected both within and between the two geographic areas. Clones from the coast or warm, moist areas were, on average, more resistant than clones from  the  interior  or  cold,  dry  climates.  The phenomenon of induced  Ill  resistance was not detected at either the local or the systemic levels in black cottonwood challenged by the rust. Variation in rust tolerance among black cottonwood clones was demonstrated. Rust tolerance was defined as the slope of the relationship between yield of ramets (expressed as a proportion of controls) and disease severity (expressed as the proportion of the total number  of  leaf-weeks  infected).  A  negative  correlation  between  rust  tolerance and rust resistance was found. In the collection of clones tested, the positive effect of disease resistance on the growth of cottonwood clones was partially counterbalanced by the negative effect of reduced tolerance on the growth. The importance of this relationship in both natural and artificial selection of superior trees against disease is indicated.  IV  TABLE OF CONTENTS  Page ABSTRACT  ii  LIST O F T A B L E S  viii  LIST O F F I G U R E S  xi  ACKNOWLEDGEMENT  xiii  INTRODUCTION  1  LITERATURE REVIEW  6  2.1  Damage to Populus  Trees Caused by Melampsora  Leaf  Rust. 2.2  6  Variation in Resistance of Populus  Trees to  Melampsora  L e a f Rust Infection  10  2.3  Tolerance of Plants to Rust Disease  14  2.4  Induced Resistance in Plants  19  MATERIALS AND METHODS 3.1  24  T h e 1988 Test of Black Cottonwood Clones  24  3.11  Collection of cuttings of black cottonwood  24  3.12  Collection of inoculum of Melampsora  28  3.13  Experiment design  30  3.14  Inoculation and measurement  32  rust  V  3.15  Disease severity rating  33  3.2  The Test of Black Cottonwood Clones in 1989  38  3.3  Tests of induced Resistance in Black Cottonwood  39  RESULTS 4.1  4.2  42  Variation i n Rust Resistance of Black Cottonwood 4.11  T h e pattern of rust development  42  4.12  Variation of rust resistance  46  Effects of the Rust on the Growth of Ramets 4.21  T h e growth of ramets  4.22  Relationship between growth characteristics and disease severity  4.23  Variation i n Rust Tolerance of Black Cottonwood 4.31  T h e definition of tolerance  4.32  Relationship between rust resistance and rust  Induced Resistance in Black Cottonwood  The Effects of Melampsora  70 70  81  84 Rust on the Growth of  Black Cottonwood  5.2  66  74  DISCUSSION 5.1  50  The effect of rust incidence during 1988 on the  tolerance 4.4  50  55  1989 growth of ramets 4.3  42  84  5.11 T h e impact of leaf rust  84  5.12 The rationale for R D S T rating system  86  Disease Resistance and Induced Resistance  90  VI  5.3  5.21  Variation in disease resistance  90  5.22  Induced resistance  90  Disease Tolerance  92  5.31  The demonstration of variation i n disease tolerance  92  5.32  Disease tolerance versus disease resistance  93  CONCLUSIONS  98  BIBLIOGRAPHY  101  APPENDICES  109  Appendix 1 Composition of the fertilizer mix used in the experiment.  109  Appendix 2 The average (n = 56) weekly relative disease severity (RDSW) of each treatment (T).  110  Appendix 3 The average (n = 4) relative disease severity (RDST) values in log transformation of each treatment of each of the 14 clones.  Ill  Appendix 4 Plots of the total dry weight of black cottonwood ramets of each clone over their relative disease severity (RDST) ratings.  112  vii  Appendix 5 A method of testing parallelism of linear regression lines.  126  Appendix 6 Analysis of covariance of average initial volume increment (AlVI) of black cottonwood ramets in 1989, the final volume of a ramet in 1988 acts as a covariate.  128  Appendix 7 Parallelism test of relative reduction rates of stem dry weight of 13 clones (16 not included).  129  vm  LIST O F T A B L E S  Page  Table 1  Results of analysis of variance for relative disease severity (RDST) values with log transformation [ R D S T = M e a n + Clone (random) + Treatment (fixed) + C x T + Error].  Table 2  The average (n = 20) relative disease severity (RDST) values of each clone.  Table 3  48  49  Average of height (cm) and diameter (mm) of uninfected control ramets of each clone at the end of the first (1988) and second (1989) growing seasons.  Table 4  51  Linear regression equation coefficients and statistical significance of the relationship between total dry weight (TDW) and relative disease severity (RDST) of each of the 14 clones [ T D W = i + b R D S T ] .  Table 5  57  Linear regression equation coefficients and statistical significance of the relationship between stem dry weight (SDW) and relative disease severity (RDST) of each of the 14 clones [SDW = i + b R D S T ] .  59  IX  Table 6  Linear regression equation coefficients and statistical significance of the relationship between root dry weight (RDW) and relative disease severity (RDST) of each of the 14 clones [ R D W = i + b R D S T ] .  Table 7  60  Linear regression equation coefficients and statistical significance of the relationship between ratio of stem dry weight to root dry weight (SDW/RDW) and relative disease severity (RDST) of each of the 14 clones [SDW/RDW = i + bRDST].  Table 8  61  Linear regression equation coefficients and statistical significance of the relationship between volume (D^H) and relative disease severity (RDST) of each of the 14 clones [ D H = i + bRDST].  63  2  Table 9  Linear regression equation coefficients and statistical significance of the relationship between diameter (D) and relative disease severity (RDST) of each of the 14 clones [D = i + b R D S T ] .  Table 10  64  Linear regression equation coefficients and statistical significance of the relationship between height (H) and relative disease severity (RDST) of each of the 14 clones [H = i + b R D S T ] .  Table 11  65  Average of relative disease severity (RDST) values in 1988, volume (cm ) at the end of 1988, and 3  average initial volume increment (AIVI) (cm ) i n 3  1989 for 1988 diseased and control ramets.  69  X  Table 12  Relative reduction rate of total dry weight (RDR), relative reduction rate of stem dry weight and (RDR') products of R D R and R D S T (relative disease severity) of each clone.  Table 13  72  Parallelism test (Cunia 1973) of relative reduction rates of total dry weight (RDR) of 14 clones.  Table 14  M e a n number of uredial pustules of  73  Melampsora  oecidentalis per leaf resulting from a challenging inoculation 10 days after an inducing inoculation and probability that means do not differ. Table 15  M e a n number of uredial pustules of  82  Melampsora  oecidentalis per leaf or per half leaf resulting from a challenging inoculation after an inducing inoculation or wounding and probability that means do not differ.  83  LIST O F F I G U R E S  Page  Figure 1  Location of collection sites of black cottonwood clones. The coastal and interior clones are designated by the capital letter ' C and T respectively.  Figure 2  26  Young seedlings of Douglas-fir used to produce aecial inoculum of the leaf rust.  Figure 3  29  Ramets of black cottonwood on the experiment site in September 2, 1988 (left: diseased, right: control).  Figure 4  Symptoms of Melampsora  occidentalis  on black  cottonwood for a range of levels of infection. Figure 5  41  The average (n = 20) weekly relative disease severity (RDSW) of each clone.  Figure 7  34  The location of black cottonwood ramets in the greenhouse for induced resistance study.  Figure 6  31  43  Plot of the correlation coefficients (r) values between individual measurements of disease severity (RDSW) and the total disease severity (RDST).  Figure 8  Average height of uninfected clones from the interior and coast during the first growing season.  Figure 9  45  52  Average diameter of uninfected clones from the interior and coast during the first growing season.  53  Xll  Figure 10  Relative disease severity in total (RDST) values of dead and surviving ramets of each clone (+:  surviving,  o: dead). Figure 11  67  Plot of the relative reduction rate of total dry weight (RDR) of each clone against its average relative disease severity (RDST) value to show the linear relationship (solid line) between tolerance and resistance. The dotted line represents the relationship between R D R and R D S T in which R D R x R D S T is a constant, the value of that constant being selected so that the line is the best fit for the 14 observations.  Figure 12  75  Plot of yield (total dry weight as the proportion of uninfected control) of each ramet in clone 12 over its relative disease severity (RDST) value.  Figure 13  76  The predicted yield (total dry weight) of each clone under different disease hazard conditions.  80  Xlll  ACKNOWLEDGMENT  I would like to express my gratitude: to m y supervisor, D r . B. J . van der K a m p , for his wise guidance, inspiring  discussions,  technical  instructions  and  miscellaneous  help  throughout the course of project. to the members of my supervisory committee, D r . D . T . Lester, D r . P. L . Marshall, D r . D . P. Lavender and D r . R. J . Copeman for their valuable comments, suggestions on statistical problems and many editorial changes on the manuscript. to the university examiners, D r . K , K l i n k a and D r . J . H . Myers, and external examiner, D r . P. V . Blenis from University of Alberta for their constructive review of the dissertation and recommendation of references. to fellow Chinese students in the University of British Columbia, Xiaotian Shi, Qinghua L i , Qingli W a n g and G a n g L i for their generous help in graph drawing, word and data processing. finally to m y parents  for their  ardent  expectation  and  constant  encouragement.  The  financial support for the program from the  State Education  Commission, People's Republic of C h i n a , the University of British Columbia and D r . B. J . van der K a m p is acknowledged.  1  INTRODUCTION  Melampsora cottonwood Populus  oecidentalis trichocarpa  Jacks is a major leaf rust of western black Torr. & G r a y in British Columbia. T h e rust  infects cottonwood leaves via aeciospores i n spring and produces a massive inoculum of urediospores on the lower side of leaves in summer. In late summer, telia are formed on the lower side of leaves,  and the fungus  overwinters on the fallen leaves. Next spring, the telia produce basidia. T h e n the basidiospores produced by the basidia infect the newly flushed needles of Douglas-fir Pseudotsuga  menziesii (Mirb.) Franco, the alternate  host, and produce pyenia, and later aecia on the lower side of infected needles. The aeciospores so produced reinfect cottonwood. T h e amount of rust on cottonwood i n late summer varies a great deal from year to year, depending largely on weather conditions. In some years, virtually all leaves are infected by September, while in other years the rust may be hard to find at that time.  The biology of Melampsora The  leaf rust is well understood (Ziller 1974).  damage caused by leaf rust infection is also widely reported. This  damage includes premature defoliation (Toole 1967), increased susceptibility to frost and other parasitic fungi (Peace 1962; Spires 1976), and reduced increment or yield (van der Meiden and v a n Vloten 1958). Reductions in height and diameter growth (Spires 1976; Widin and Schipper 1981), a decline in accumulated volume, and mortality of the plant (Jokela and  2  Lovett 1976) have been reported. However, the relationship between rust infection and yield loss of Populus  trees has not been carefully quantified.  For instance, it is not known whether the biomass loss of cottonwood i n the year of infection is directly proportional to the rust severity, or i f there is a threshold infection level i n the Melampsora-Populus  pathosystem. The  threshold infection level here means that the rust will have no adverse effect on plant growth unless its buildup reaches a certain point, i.e. the threshold.  A n adequate measurement of disease severity is important in the quantification of loss from plant disease. Cobb (1892) first designed a scale to measure the severity of the rust by comparing the sketches of infected leaves (showing five degrees ranging from 1 - 50 % of the leaf area covered by pustules) with real leaves. V i s u a l and quantitative estimate of plant disease  through  assignment  of numeric values  to a range  of disease  conditions was initiated from Mckinney (1923). Horsfall and Baratt (1945) improved the numeric rating system by grading the disease logarithmically rather than arithmetically by recognizing the logarithmic aspect of visual acuity. Similar systems were adopted i n poplar rust rating. Schreiner (1959) measures the rust severity as the product of the degree of rust infection on the infected leaf and the percentage of similarly infected leaves on the tree. M a n y studies i n agriculture and forestry assessed the disease severity by assigning numeric codes or keys to defined disease conditions (Large and Doling 1963; Jokela 1966; James 1971; Sarri and Prescott 1975). Raymond et al. (1985) developed a weighted scale to take into account leaf position i n evaluating the severity of tan spot of winter wheat. The upper leaves which  3 contributed significantly to grain yield were given more weight.  Since disease development is a dynamic process, the time continuation of a disease epidemic should be an important factor i n an disease rating system.  The  issues  associated  with  models  using  one-time  disease  assessments (single or critical-point models) and those using several disease assessments made throughout the epidemic, such as multiple-point models and the area under the disease progress curve ( A U P D C ) , were recognized (James 1974;  T e n g and Johnson 1988). Those models were frequently used  in evaluating powdery mildew epidemics (Shaner 1973; al.  F r y 1978;  Fried et  1979). These systems and models of disease assessments, to different  extents,  achieved  reasonable  estimations  of disease  severity  and  high  correlations between grain yield and disease severity. However, if a study intends to examine the inherited relationship between cottonwood growth and leaf rust severity both qualitatively and quantitatively, the existing disease rating systems then may not satisfy the purpose completely. For example,  the percentage reduction in growth of the tree related to the  percentage of disease area could not be assessed on a numeric scale based rating system.  The invasion of plants by pathogens is always accompanied by some degree of resistance in the hosts. The Melampsora-Populus no exception.  Clones of black cottonwood  pathosystem is  differed significantly in their  ability to support urediospore production of Melampsora  oecidentalis in a  leaf disc test (Hsiang and van der K a m p 1985). However, the cases of variation  in  resistance  of  trees  under  field  conditions  are  not  fully  4 understood.  N o r is it known  if there  is induced resistance  in black  cottonwood trees. If induced resistance is present, and if the degree varies between trees or clones, then to measure disease resistance based on a single inoculation is misleading. Although there is an increasing number of reports concerning induced resistance i n plants, most of them deal with agricultural crops, and information on forest trees is scarce. M a t t a (1979) listed 149 references of induced resistance studies, but only seven of them were involved in woody plants. Furthermore, systemic induced resistance in plants caused by obligate parasitic fungi has rarely been reported, although a few instances of localized and systemic induced resistance stimulated by such fungi have been demonstrated  (Littlefield 1969; Hoes and Dorrell  1979).  Disease tolerance has been reported i n agriculture crops (Caldwell et al. 1934). However, the subject has not received much attention in forest pathology, nor has the relationship between disease tolerance and disease resistance ever been investigated. T h e term 'disease tolerance' designates the relative ability of a clone or a cultivar to sustain disease attack without much loss i n yield in comparison with others of the same species. It is different from the concept of disease resistance  in that it concerns the  growth performance of infected plants rather than just their superficial index of disease. The issue raised here is whether clones of black cottonwood will respond to the same amount of rust attack identically or differentially in terms of yield reduction.  This research project deals with the Melampsora  occidentalis-Populus  5  trichocarpa  pathosystem. It examines the relationship between rust severity  and a series of growth parameters for a sample of clones selected from the British Columbia cottonwood population. In order to achieve this goal, a new rust rating system was designed to record the proportion of infected leaf area and include the duration or accumulated effects of the disease. This new system was also used to assess the relative resistance of the set of cottonwood  clones  to  the  rust.  A  search  for  induced resistance  was  undertaken, in part to determine whether a valid assessment of resistance can be obtained by a single inoculation, and because of its inherent interest. Finally, a method for measuring tolerance was developed, the tolerance of individual clones compared, and the relationship between tolerance  and  resistance elucidated.  The goal of the research conducted was, therefore, to answer the following questions:  (1)  W h a t is the relationship between disease severity and growth loss of  black cottonwood? (2)  W h a t is the impact of leaf rust on the growth of black cottonwood? Is  there a threshold infection level in this pathosystem? (3)  Is  there  significant variation  i n both  rust  resistance  and  rust  tolerance among clones of black cottonwood? A r e they mutually related? (4)  Does black cottonwood exhibit induced resistance to Melampsora  at either the local or systemic level?  rust  6  LITERATURE REVIEW  2.1  Damage to Populus  Trees Caused by Melampsora  Studies on the pathosystems  Leaf Rust  of leaf rust and host plants in both  agriculture and forestry have shown the disturbed physiological functions i n the infected plants due to the growth of mycelium in the leaf tissue and the massive production of the urediospores on the surface of leaves.  These  include increased transpiration rate (Duniway and Durbin 1971; Siwecki and Przybyl  1981),  stimulated respiration (Daly and Sayre  1957) and  decreased photosynthetic capacity (Livine 1964), as well as the disturbed translocation of the photosynthate  (Griffiths 1984). T h e consequences of  such physiological abnormalities often lead the susceptible hosts to shed leaves prematurely and to fail to harden properly (Nagel 1949; Schipper and Dawson 1974). Therefore, it is not surprising that the rust infected poplars are more susceptible to frost and secondary pathogens Cytospora  and Dothichiza,  such as  and m a y eventually be killed by these (Peace  1962; Spires 1976).  A number of species of Melampsora  cause leaf rust on Populus  trees.  They have similar life cycles on different hosts and alternate hosts. A m o n g them, the American species Melampsora species Melampsora  larici-populina  attracted much research attention.  Kleb.  medusae T h u e m and Eurasian are most  common, and have  7  The reduced increment of host poplar trees due to the rust infection was first described by v a n der Meiden and v a n Vloten (1958). Chiba and Nagata  (1973) found a negative  relationship i n Populus  maximowiczii  Henry between rust scores made i n the nursery and height at 8 years of age.  Wilcox  and Farmer  incidence of Melampsora  (1967)  found  that  the correlations  between  medusae i n 1963 and the following season's growth  in diameter and height were -0.45 and -0.38 respectively i n a Mississippi plantation of eastern cottonwood Populus  deltoides Bartr. The rust infection  occurred i n late summer. Although not obviously reducing the current year's growth, it apparently interfered with accumulation of plant reserves and thus affected growth in the following season.  In a study of chemical control of the poplar leaf rust larici-populina  i n N e w Zealand,  Spires  (1974)  found  Melampsora  that  negative  correlations existed between height growth and rust rating (r = -0.171) and similarly between diameter growth and rust rating (r = -0.228) in the semievergreen Lombardy poplar Populus  nigra L . cv 'Sempervirens'. T h e impact  of rust infection was greater on diameter growth than on height growth. Protection of poplars from the rust by fungicides resulted i n better tree growth. In a continuation of the study, Spires (1976) further demonstrated that  the percentage  reduction i n height  and diameter  growth  of the  unsprayed controls were 24 percent and 47 percent for the semi-evergreen, and 30 percent and 43 percent for the '1-214' cultivar respectively compared with those protected by the best fungicide ('Bas 31703F'). Fresh coarse root weight of the unprotected controls was reduced by 93 percent i n the semi-  8 evergreen and 92 percent in '1-214' cultivar compared with the 'Bas 31703F' protected treatment. The structure of the root system of the poplars was also influenced by rust infection. In the effective fungicide treatments, the roots  were  larger  in diameter  and  penetrated  the  soil  further both  horizontally and vertically, so that more mechanical force was required to extract the trees from the soil.  Similar results were reported by Widin and Schipper (1976) from a study of the impact of Melampsora  medusae leaf rust on hybrid poplar at  Rhinelander, Wisconsin. They found that the percentage increase in basal diameter was greater in trees sprayed with fungicide than i n unsprayed trees. Basal diameter growth was reduced 23 percent by rust, whereas the height growth was only slightly reduced during the first year.  In a later study, Widin and Schipper (1981) planted cuttings of five hybrid poplar clones in paired plots at each of three locations. There was a total of four plots at each location; two were allowed to become naturally infected with rust and two were sprayed with fungicide to reduce or exclude rust infection. In each growing season, they measured the stem height and basal stem diameter three times. The trees were finally harvested by cutting at ground line after two growing seasons. D r y weight and volume of each tree were recorded. They showed that poplars which were protected from leaf rust produced more wood fiber than trees that were not protected. T h e leaf rust reduced wood volume and dry weight of all but the most resistant  hybrid clone. The results were  similar at  each location with  average losses for all clones ranging from 29 percent to 32 percent for dry  9  weight and 31 percent to 42 percent for volume.  Jokela and Lovett (1976) studied the cumulative effect of annual rust infection by Melampsora  medusae on growth and survival of seedlings of  eastern cottonwood in an Illinois plantation. The yield of cottonwood at age 15 was related to the average of leaf-rust scores observed in September of the second and third growing season. A linear relation between the volume reduction ( D H ) of trees and rust score was found. It illustrated that an 2  increase of 1 rust score class on a 5 - point scale (1 = light infection to 5 = severe infection) reduced yield by about 20 percent. Accumulated mortality at age 15, which averaged about 15 percent of the trees i n the lower three rust score classes, increased to 33 percent of the trees i n rust class 4 and to 50 percent in class 5.  10  2.2  Variation i n Resistance of Populus Trees to Melampsora  Leaf Rust  Infection  The differential reactions of poplar species and clones to the rust infection provides the basis for variation i n resistance.  Several studies  examined this aspect. T h e infection of eastern cottonwood by  Melampsora  medusae was initiated by stomatal entry (Shain and Jarlfors 1987). Some healthy-appearing haustoria were seen i n susceptible clones before uredial production. In the resistant clones, on the other hand, most haustoria and invaded host cells appeared necrotic within two days of inoculation or plasmolized rapidly i n infected portions of leaves.  Mlodzianowski et al.  (1978) showed that urediospore germination was more rapid on leaves of susceptible clones of black cottonwood than on those of Populus poitou'  'serotina de  (a resistant clone). Necrotic reactions of guard and mesophyll cells  in resistant poplars Populus  maximowicizii  and Populus  deltoides 'B-60'  were observed (Siwecki and Werner 1980). E M examination showed that starch grains i n some chloroplasts were broken down into small particles i n a very susceptible clone of black cottonwood. Other chloroplasts in this clone had an abnormal, strongly osmiophilic matrix  which was not found i n  resistant clones (Mlodzianowski and Siwecki 1976).  Reports on variation i n resistance Melampsora 1978;  of poplar species and clones to  leaf rust are numerous (Ciesla et al. 1975; Lemoine and Pinon  Heather et al. 1980; Ostry and M c N a b b 1985, 1986). A s early as  1937, Schreiner observed variation i n resistance i n different species, clones,  11 and progenies of Populus.  Nagel (1949) found that defoliation of 250  different poplar species and clones ranged from 6 percent to 98 percent upon attack by the leaf rust. Jokela (1966) identified a wide range of resistance to Melampsora  medusae  among clones  and seedling progenies of eastern  cottonwood i n Illinois.  Much  variation  in  encountered i n an eastern  resistance  to  Melampsora  medusae  was  cottonwood provenance and progeny test at  Wooster, Ohio (Thielges and A d a m s 1975). The test consisted of 228 clones originating from 76 parent trees. Disease resistance, rated on a visual scale from one to seven, showed significant differences among and within openpollinated families. Rust resistance was under strong genetic control. Broadsense heritabilities were estimated as high as 0.88 - 0.95. They claimed that major genetic gains i n Melampsora  rust resistance can be realized through  selection of cotton woods with observed field resistance. The selection for rust resistance resistant  should trees  result  i n higher annual volume increments,  are not defoliated i n late summer  because  and therefore  their  diameter growth is increased.  The genetic variation i n disease resistance has also been studied i n other  Populus  - Melampsora  pathosystems.  In Poland,  Krzan  (1981)  reported that of 265 clones tested, eastern cottonwood and its hybrids with Populus  nigra  maximowiczii populina.  L . , the Italian selections  1-214 and 1-154, and Populus  and its hybrids were more resistant to Melampsora  Populus  simonii  C a r r . and Populus  larici-  koreana Rehd. also had high  resistance. A high degree of resistance was inherited by progeny of Populus  12  maximowiczii. euramericana Melampsora  Gao  (1981)  and Populus larici-populina  identified  Populus  koreana,  Populus  x  deltoides var. 'missourinsis' as being resistant to i n northeastern China.  Gallo et al. (1985) investigated 49 families of controlled crosses for their  resistance  Melampsora  to Melampsora  magnusiana  leaf  rust,  Wagner. Populus  thought  to be caused by  tremuloides  demonstrated the highest resistance, Populus  Michx. families  tremula L . families, on the  other hand, were severely attacked. The interspecific hybrid families showed intermediate behavior after assessment of the rust infection at the end of both the first and second growing season. Additive genetic variation was manifested. Variation in rust virulence and host resistance of occidentalis  Melampsora  on black cottonwood was detected by Hsiang and van der K a m p  (1985) after inoculation of 10 isolates of the rust on each of 14 clones collected from the natural pathosystem i n British Columbia. The clones displayed significant differences i n their  ability to support urediospore  production. There was no differential reaction between clones and isolates. They believed that the absence of qualitative resistance and virulence was indicative of the horizontal resistance of cottonwood to the leaf rust.  Variation of rust resistance eastern  cottonwood. Young,  also occurred within single plants of  expanding leaves  resistant to the infection of Melampsora  were  found to be highly  medusae while the mature leaves  were quite susceptible (Shain and Miller 1982). Pinocembrin (PC), a major constituent of leaf resin of eastern cottonwood was identified to be inhibitory to the rust. A s leaves expanded, the amount of P C per unit of leaf area  13  decreased. The amount of P C on the surface of young leaves was thought to be sufficient to contribute substantially to their resistance.  14  2.3  Tolerance of Plants to Rust Disease  Although there is a general realization of disease tolerance as the ability of plants to endure the disease while still producing good crops, the concept of disease tolerance is not as unified as disease resistance. A number of definitions have been proposed by individual researchers, and there are at least three senses in which the term tolerance has been used.  The first sense is often used i n the case of viral diseases and concerns the appearances  and severity of the symptoms of infected plants. Thus  Matthews (1970) considered Anbalema tobacco to be tolerant because it remained almost symptomless, even though it was systemically infected by tobacco mosaic virus ( T M V ) . The second sense of disease tolerance relates to the existence of infection thresholds (Ellis 1954; G a u n t 1981). In this sense, if the yield of a plant is not affected or reduced when infection is below a certain level, the plant is said to be tolerant.  The third sense of disease tolerance is concerned with the relationship between the disease severity and growth or yield loss of plants. It is the sense used i n this study. According to Schafer (1971) tolerance  is the  capacity of a cultivar that results i n less yield or quality loss relative to disease practical  severity  or pathogen development than in other  applications,  Simons  (1966)  defined tolerance  cultivars. F o r as when two  varieties both exhibit appreciable numbers of large uredia at some time during the growing season, and one variety shows a significantly smaller  15  loss i n quantity than the other, then it can be regarded as the more tolerant of the two. This definition, in the author's words is less precise than his definition of 'true' tolerance which is when two varieties, exhibiting equal numbers of large uredia at any given time throughout the infection period, show significantly different quantitative responses to the infection.  The phenomenon that some plants can sustain considerable pathogen infection and still produce a good yield, was described i n agricultural crops as early as the turn of the century (Orton 1909). Salmon and Laude (1932) observed that the Fullhard variety of wheat produced a high yield though severely infected with the leaf rust Puccinia  triticina  Erikss. Fullhard  yielded 29 b u / A , with a leaf rust rating of 87 percent. Five other varieties yielded from 11 - 18 b u / A . with 80 - 88 percent of leaf rust Caldwell  et al.  (1934)  subsequently  verified this  result  ratings.  i n controlled  experiments. They compared the yield of rusted plants with those of sulfur protected plants and found that although Fullhard was the most severely infected variety i n the test, its yield was not reduced.  Ellis (1954) reported differences i n maize varietal susceptibility and of tolerance to maize rust Puccinia  polysora Underw., although all types were  fully susceptible i n a standard seedling test. T h e unsprayed hybrid was given an average score of 3.5 open urediosori per plant, and those of the local selected and unselected maize were 1.9 and 1.6 respectively. However, the yield of the hybrid was reduced by 17 percent, whereas that of the local varieties was completely unaffected. Ellis suggested that the tolerance of the local varieties m a y be an indication that they did not express their full  16  metabolic potentials under normal condition in the yield production. H e postulated the existence of a photosynthate reservoir, such that until the reservoir was exhausted no loss i n yield would be encountered.  B y comparing two pairs of 'isogenic' oat varieties (Clinton 59 and Clintland,  Benton and Bentland),  demonstrate coronata  Caldwell et al. (1958)  were  able to  a high level of tolerance to the crown-rust fungus  Puccinia  C d a . i n the Benton variety. There was little difference between the  members of these pairs of varieties i n appearance or yield i n the absence of crown rust. During the rust epidemics i n 1955 and 1956, the susceptible varieties Benton and Clinton 59 were completely infected as opposed to a trace of disease on the resistant varieties Bentland and Clintland. However, loss i n yield of the Benton variety in two years were nil or small as deduced from comparison with its resistant counterpart, Bentland. In contrast, the losses of the Clinton 59 were severe as deduced from comparison with its resistant counterpart, Clintland.  Direct comparison of the yield of the Clinton 59 and Benton varieties generated  the same result.  Both  varieties  appeared  to be completely  susceptible to crown rust. Clinton 59 was significantly the higher yielding of the two i n the absence of crown rust, yet under severe epidemics the Benton gave much higher yields i n the two years it was tested.  The result that the Benton variety is more tolerant was also reported by Clifford (1968). In his study, circular or square field plots, centrally inoculated with urediospore cultures of specific races of Puccinia  coronata  17  were used to describe the rust spread and intensification with time. The experiment  showed that  the  Andrew  and  Benton  varieties  developed  comparable rust levels throughout the season but that the Benton tolerated this level of rust better and yielded more than the Andrew did.  The relative tolerance of varieties of oats Avena sativa to races of the crown rust fungus Puccinia  coronata  var.  avenae Fraser  and L e d . was  studied by Simons (1966, 1968) and Simons and M u r p h y (1967). Split plots were used in the 1966  study, in which one half plot was rusted and the  other protected from rust by a fungicide. The relative tolerance  of oat  varieties was expressed by the ratios of yield or kernel weight of rusted to nonrusted plot pairs, with a ratio of 1.00  indicating complete  tolerance.  Three varieties, Clinton, Benton and Cherokee, were chosen as a measure of the others. The first two had been shown previously to differ significantly in tolerance to crown rust. The third variety was commonly regarded as being at least somewhat tolerant to infection by the rust fungus. Results showed that Clinton, relative to most of the other 21 varieties, was nontolerant. Two additional varieties, Richland and Markton, were also quite intolerant. Benton appeared to be less damaged by crown rust than Clinton. Kernel weight data from all tests showed Cherokee to be significantly superior to Clinton. In three of four tests with races 203 and 216, the kernel weight ratio of Cherokee was also significantly higher than the Benton. Several other  varieties  with susceptible  reactions  were  also  significantly more  tolerant than the Clinton and Benton.  Linear models were established i n assessment  of powdery mildew  18 severity in relation to grain yield of winter wheat cultivars (Lipps and Madden 1989). More susceptible cultivars had higher slope values year, indicating greater yield reduction per increase i n disease  each  severity.  They pointed out that the slope values can be considered a measure of the tolerance of the cultivar; smaller values indicate greater tolerance. The slope values for the different cultivar also varied with the year,  and those  cultivars considered most susceptible prior to the study (Hart, Becker, and Adena) had higher slopes than those considered less susceptible (Cardinal and Caldwell).  The mechanisms  of variation i n disease  tolerance  are not clear.  Clarke (1984) suggested that some of the damage caused by pathogens was avoidable because it may not be essential for parasite development. Orton (1909) proposed tolerance may result from exceptional vigor or a hardier structure of plants. Stakman and H a r r a r (1957) stated that tolerance may be due to compensating growth i n some cases.  19  2.4  Induced Resistance i n Plants  B y definition (Sequeira 1983), induced resistance is the resistance that is dependant on factors present only after the host is challenged by the pathogen. Two forms of induced resistance exist: localized and systemic. The former can be detected only i n the area immediately adjacent to the site of attempted penetration by the pathogen; the latter refers to resistance that occurs at sites i n the host distant from the point of initial interaction with a potential pathogen. Systemic induced resistance is generally the result of necrosis brought about by the inducing inoculation, and is usually detected by challenge inoculation at a different time and location on the plant.  There  have  pathogens.  Rahe  Phaseolus  vulgaris  carbonum  been  many  et al.  reports  (1969)  L . respond  r a c e l and Alternaria  varietal-nonpathogenic  of induced resistance  observed  that  etiolated  hypersensitively  to  to fungal  hypocotyls of  Helminthosporium  sp. 24 - 36 hrs after inoculation with a  race of Colletotrichum  Magn., the causal agent of bean anthracnose.  lindemuthianum  Sacc. &  T h e protection of excised  hypocotyls of the bean plants was localized within areas receiving spores of the incompatible race and was seen as an effect of the host on the growth of established intracellular hyphae of the incompatible race rather than on spore germination or penetration (Skipp and Deverall 1973).  Systemic Colletotrichum  protection lagenarium  was  elicited  in  cucumber  plants  against  Dass. by prior inoculation with the fungus (Kuc  20  et al 1975). T h e protection persisted to the fruiting period i f a second booster inoculation was applied after the first inducing inoculation. A single lesion produced significant protection. A direct relationship existed between the number of spores used for protection and the extent and duration of protection (Kuc and Richmond 1977). Four cultivars of watermelon and muskmelon were similarly induced by Colletotrichum  lagenarium  to resist  the pathogen both i n the greenhouse and i n the field (Caruso and K u c 1977a,  b). Resistance  of cucumber  to  anthracnose  systemically by a localized infection with tobacco  was also induced  necrosis virus ( T N V )  (Jenns and K u c 1977) and the bacteria Pseudomonas  lachrymans  (Caruso  1977).  Mclntyre and Dodds (1979) detected systemic protection of a variety of Nicotiana  tabacum  Phytophthora  parasitica  (WS 117)  with  var. nicotianae  tobacco  mosaic  virus  against  (Barda de Haan) Tucker (Ppn). In  later experiments, Mclntyre et al. (1981) found that tobacco mosaic virus ( T M V ) inoculation of a hypersensitive tobacco cultivar induced systemic and long lived resistance simultaneously against T M V , P p n and Pseudomonas tabacci Wolf and Foster. Separately, T M V also induced resistance Peronospora  tabacina  against  A d a m . , and reduced the reproduction of the aphid  Myzus persicae Sulz. Levels of protection and the time resistance developed in a given leaf were not the same for the different challengers.  Local antagonism between two cereal rust fungi was detected by Johnston and Huffman (1958). Inoculation with urediospores of  Puccinia  coronata on wheat seedlings susceptible to leaf rust prior to inoculation with  21 the wheat leaf rust fungus, resulted i n fewer rust pustules than similar plants inoculated only  with  the leaf rust  organism. Localized induced  resistance against the virulent race of Melampsora  lini (Ehrenb) Lev. also  occurs following prior inoculation of flax with an avirulent race of the fungus (Littlefield 1969). The induced resistance was effective for seven days after inoculation. T h e effect was nonspecific since the resistance was also induced by prior inoculation with the wheat stem and leaf rust pathogens. Hoes and Dorrell (1979) demonstrated a systemic induced resistance i n flax to flax rust infection. T h e protective effect lasted at least 30 days i n flax cultivar Redwood 65.  Alterations i n resistance  of plants following treatment  by abiotic  factors has been reported by a number of studies. Andebrahan and Wood (1980) observed that irradiation of etiolated bean hypocotyls with ultraviolet light for short periods immediately before inoculation greatly resistance to races of Colletotrichum and to Colletotrichum Here  induced  lagenarium  lindemuthianum  otherwise avirulent,  and Colletotrichum  susceptibility occurred.  Resistance  decreased  to  coffeanum Noack. these  pathogens,  however, was induced when inoculation was conducted 24 or 48 hours after ultraviolet irradiation. Doubrana et al. (1988) reported induction of systemic resistance to anthracnose caused by Colletotrichum  lagenarium  i n cucumber  by extracts from spinach and rhubarb leaves. Oxalate was identified as the active component of both extracts. Perucca transitory induction of resistance oxysporum,  and M a t t a (1988) found a  to vascular wilt of tomato  (Fusarium  f.sp. lycopersici) by root treatments with chloroform vapor and  immersion i n mannitol solution. Similar results were obtained by a hot  22 water treatment (Anchisi et al. 1985).  Not all plants tested can be induced to gain higher resistance. Paine and Stephen (1987) found no discernible effect of previous exposure to Ceratocystis  minor  (Hedge) Hunt, a fungus associated with the southern  pine beetle on the subsequent ability of lobololly pine Pinus  taeda L . to  respond to inoculation.  The mechanisms of induced resistance are poorly understood although altered metabolism has been reported (Hammerschmidt et al. 1982, Gessiler and K u c 1982). Sequeira (1983) suggested a cascading series of biochemical reactions which originate from enzymatic breakdown of the cell wall of plants upon injury or attack by pathogens  that spreads throughout the  plant. Dean and K u c (1986) identified a signal which was produced i n cucumber  leaves  inoculated  with  Collectotrichum  lagenarium.  Once  produced, the signal was swiftly translocated to other parts of the plants v i a the vascular system and resulted i n rapid sensitization of the plant's defense mechanisms.  Dean and K u c (1987) attributed  the induced protection of  cucumber to the rapid lignification of the epidermis and reduced penetration of it by the fungus.  Some characteristics of induced resistance can be summarized as follows:  1.  Induced resistance appears non-specific with regard to inducer and  challenger. Diverse biotic and abiotic agents are able to induce resistance of  23 plants against diverse challengers. However, a particular inducer does not necessarily induce resistance in all the plants.  2.  A positive relationship appears  to exist between  the  amount  of  inducer and the effectiveness of induction. Generally, more initial inoculum would result in more complete later protection of plants.  3.  Induction of resistance does not occur until some time interval after  the inducing inoculation. Induced resistance may persist for quite a long time or just be transitory depending on the types and intensities of the stimulation on plants.  4.  Induced resistance seems to be an energy demanding process, it is  limited and directed by the reproduction of plants. Resistance  was  not  induced in cucumber plants once they set fruits and was markedly reduced when induced at the time of flowering (Guedes et al. 1980).  24  MATERIALS AND METHODS  3.1  The 1988 Test of Black Cottonwood Clones  3.11  Collection of cuttings of black cottonwood  In middle and late A p r i l 1988,  cuttings of 15 black cottonwood trees  from the coast (Fraser Valley) and the interior of British Columbia were collected. The chosen trees were generally small in size. Shoots of similar position in the middle of crown were selected to maximize genetic  and  physiological uniformity of the material of each clone. Diversity of the experimental materials  was achieved by locating the  collection sites in  different areas. Trees were selected randomly with regard to their relative resistance to Melampsora  rust since all were uninfected at the time of  collection. Selections were made from five biogeoclimatic zones (Pojar et al. 1987)  of the province (Table 2, Figure 1). Some of these represents warm  and moist climates like the Coastal Western Hemlock zone. Some have cold winters, but w a r m and moist summers such as the Interior Cedar-Hemlock zone. The Interior  Douglas-Fir zone is hot and dry in summer, while  summers in Montane Spruce zone are moderately short, warm and moist. The climates in Sub-Boreal Spruce zone are severe in winters, but short, cool, and, in the subzones sampled, relatively dry in summers (Anonymous 1988) . A t the time of collection, most black cottonwood trees in the Fraser Valley had already flushed, while in the interior the plants were still in  25 quiescent state or just about to flush.  The cuttings were brought to the laboratory, clipped to a standard length of about 20 cm, and inserted into beakers containing cold water for initial rooting. After a week, small, white, delicate roots developed at the base of the cuttings. The cuttings were then planted into small (3 liter) plastic pots and grown outdoors i n the South Campus field at the University of British Columbia. The pots were filled with a soil mixture consisting of hemlock sawdust, peat and sand at a proportion of 4:2:1. A fertilizer mix, which contained necessary macro and micro nutrients for the growth of plants was added to the soil mixture (Appendix 1).  The ramets were kept in the small pots for one month. During this period, the cuttings flushed and initiated growth. Twenty eight ramets with uniform size of each clone except C 6 were then transplanted from the small pots into big plastic pots (11 liter) containing the same soil mixture. Finally, the ramets in the big pots were assigned to different treatments and allocated on the experiment growing shoot.  site. A l l ramets were  pruned to a single  Figure 1. Location of collection sites of black cottonwood clones. The coastal and interior clones are designated by the capital letter ' C and T respectively.  C l : Placid Lake  II: Fort St. James  C 2 : Placid Lake  12: Hixon  C 3 : Steelhead  13: 70 Mile House  C 4 : Matsqui  14: Prince George  C 5 : Vancouver  15: Vanderhoof  C6: Steelhead  16: Nakusp 17: L u m b y 18: Monte Lake 19: Merritt  28  3.12  Collection of inoculum of Melampsora  rust  About 100 young seedlings of Douglas-fir (Figure 2) were planted in a small square in the field of South Campus in late A p r i l . The planting site was roughly ten meters away from the m a i n experimental site. In early May,  fallen leaves  of black cottonwood bearing Melampsora  telia were  collected from various locations at the coast and in the interior. The leaves were mixed and laid on the ground between the Douglas-fir seedlings, with their lower sides facing up. Water was sprayed frequently to moisten the dry leaves of the black cottonwood and the needles of the Douglas-fir, so as to facilitate  the  germination of teliospores  and infection of needles  by  basidiospores. B y late M a y , the seedlings of Douglas-fir had flushed their new  shoots,  which  are  the  only  susceptible  part  of the  plant.  The  basidiospores, at the same time, were released from the telia and landed on these new needles to cause infection. In June, tiny yellow coloured aecia of Melampsora  occidentalis  were observed on the lower side of the needles.  When the aecia were mature and began to release aeciospores, the needles bearing the aecia were collected and used for initial inoculation of the black cottonwood ramets. E x t r a cottonwood ramets were also inoculated, and the urediospores produced on these were used for the second inoculation.  30  3.13  Experiment design  The main study area was located on the South Campus Field area of the University of British Columbia. The experiment occupied an area of approximate 12 x 10m. The site had a slight south facing slope. The ground surface of the site was first leveled and covered with black polythylene to control weeds. T h e n the big pots planted with ramets of black cottonwood were set out on the site at approximately a 0.5 x 0.5m spacing (Figure 3).  There was a total of five treatments in the experiment. They were arranged side by side from east to west with the first treatment at eastern end. Treatments  1-4  the  were designed for sequential rusting of the  ramets to generate different disease severities,  treatment 5 was treated  with a fungicide to exclude the rust and served as the control. Five ramets of each of 14 clones were randomly assigned to each treatment. For the purpose of a second year study, three additional ramets of each clone were assigned to the control section.  The ramets were arranged  randomly within each treatment. The  arrangement was completed in the middle of June. F r o m then on, the test ramets were carefully managed by regularly spraying with water, getting rid of weeds and insects. A liquid solution of Cyprex 65W (dodine 65%)  was  sprayed to run-off on the control ramets weekly from June to October to prevent rust infection. The application rate of the fungicide was 500g/10001 as suggested by the manufacturer.  Figure 3. Ramets of black cottonwood on the experiment site in September 2, 1988 (left: diseased, right: control).  32  3.14  Inoculation and measurement  The  first set  of measurements  cottonwood was conducted at  of height and diameter  the end of June,  before the  of black  initial  rust  inoculation. T h e diameter was measured at the base of the ramets. Both parameters  were  measured  at  monthly intervals until the  end of the  growing season. The product of square of diameter and height (D^H) of main stem was calculated as the index of volume.  The  black cottonwood  ramets  in treatment  1 received the  first  inoculation i n J u l y 4. The inoculation was done by rubbing the underside of the cottonwood leaves with needles of Douglas-fir bearing the aecia Melampsora  occidentalis.  Two leaves i n the middle of each ramet  of  were  inoculated. The inoculated leaves were covered with plastic bags for 40 hours to maintain humidity. Treatment 2 was inoculated on J u l y 28. A t this time, urediospores of the rust were used as inocula. The inoculation was performed by rubbing healthy cottonwood leaves with infected ones. A l l mature leaves were inoculated. A t the same time, all the leaves  i n treatment  1 were  reinoculated. Melampsora  symptomless  rust  progressed  rapidly i n August, and ramets in treatment 3 and 4 were naturally infected. Therefore, no artificial inoculation was applied on the ramets in these two treatments.  About one week after inoculation with aeciospores or urediospores, small blisters of light orange coloured uredinia began to appear on the lower  33  side of the inoculated leaves of the black cottonwood. Later, the uredia became larger and darker. Finally, black patches of telia were produced between the uredinia in the fall (Figure 4). The measurements of disease severity were begun in the middle of July, and conducted every week for 13 weeks. A measurement  frequency of one week was chosen because  corresponds roughly to the  it  length of a single urediospore cycle under  optimum conditions.  The ramets of black cottonwood were harvested at the beginning of November after leaf fall was complete. Four ramets of each clone from each treatment were randomly selected for harvesting. The plants were pulled out of the pots by hand and washed. T h e n the ramets were clipped into several segments and stored in paper bags. The paper bags were carried back to the laboratory and left open on benches to air dry for 50 days. Finally, the dry weight of roots and stems  were determined. The total dry  weight was the sum of the two. It did not include leaf weight.  3.15  Disease severity rating  The procedure for rating rust severity each week and the manner in which these weekly measurements were combined to give an overall disease rating was as follows:  1. Choose a leaf of medium size as the standard for the ramet being rated. 2. Estimate the total leaf area as the number of leaves of standard size (adjusted number of leaves).  34 Figure 4. Symptoms of Melampsora  oecidentalis  on black cottonwood for a  range of levels of infection.  A . V e r y light infection, rated as 4 percent. B . Light infection, rated as 20 percent. C . Moderate infection, rated as 45 percent. D . Moderate heavy infection, rated as 57 percent. E . Heavy infection, rated as 73 percent. F . V e r y heavy infection, rated as 91 percent.  36  37  3. Combine the area of uredial pustules leaf by leaf on the ramet and estimate this area as a number of standard leaves (adjusted number of diseased leaves). 4. Divide the adjusted disease leaf number of each weekly measurement by the adjusted leaf number to generate the relative disease severity for that week (RDSW) as the proportion of leaf area infected. 5. S u m the adjusted disease leaf numbers of the 13 weekly ratings and divide it by the summation of adjusted leaf numbers of 13  individual  measurements, this generates the relative disease severity in total (RDST) as the proportion of the total leaf area infected (ie. proportion of diseased leaf-weeks of all foliage).  Once a standard leaf was decided, each individual leaf was compared to it. The area of each leaf was estimated at a proportion of the area of the standard  leaf,  and then  proportions summed to  give total  leaf  area.  Similarly, the area occupied by uredial pustules were put together to meet the standard leaf size. Thus, both diseased and healthy leaf areas were converted into the same unit. The proportion of diseased leaf area on a ramet  then was determined. In the process  of each measurement,  the  adjusted number of leaves and adjusted number of diseased leaves were estimated visually with the aid of standard leaves. The area of an individual uredial pustule was considered to extend to the edge of the small chlorotic spot surrounding each pustule. The leaves that were infected, killed and shed before the final measurements  were considered to be one hundred  percent infected. Figure 4 shows some examples of different infection levels on individual leaves of black cottonwood.  38  3.2  The Test of Black Cottonwood Clones i n 1989  The  unharvested  ramets of the  black cottonwood were  another year's growth to test the extended effects of the  kept for  Melampsora rust  on their growth. The ramets from treatments 1-4 were combined to form a new group which consisted of four ramets for each clone. Four ramets from each clone i n the control section formed a second group. The plants in both groups were managed in the same way as in the first year, except that no artificial inoculation was applied and all plants were protected by fungicide. New soil mixture was added as required to fill the pots. The basal diameter and height of each ramet were measured each month starting in the late May,  1989.  A t the end of October, these ramets were also harvested and  the dry weight of the roots and stems determined.  39  3.3  Tests of Induced Resistance in Black Cottonwood  Young ramets of clone C 6 and clone 12 were used to determine whether induced resistance to Melampsora  rust occurs i n black cottonwood.  The test was conducted in the fall of 1989. susceptible to the Melampsora  Both clones were moderately  rust according to the 1988 experiment. The  ramets were maintained in a greenhouse, receiving 19 hours of fluorescent light to simulate long daylength (Figure 5).  Urediospores of Melampsora agents.  Spore  suspensions  were  occidentalis made  by  were first used as inducing dispersing freshly  collected  urediospores into distilled water and adding two drops of Tween-20 per 100ml. The concentrations of the spore suspensions were not determined, but a single suspension was used for each experiment. Spore suspensions were pipetted on the lower side of the cottonwood leaves with 5 drops/leaf or 3 drops/half leaf parted by the middle vein, and spread evenly over the surface. The inoculated leaves were enclosed i n plastic bags to keep them moist for 48 hours. Control ramets received the same treatment except that distilled water was substituted for the urediospore suspension.  Wounding of the cottonwood leaves was also tried to stimulate the reaction of infected ramets. Twelve holes (3mm in diameter) on a leaf or 6 holes on a half leaf were pierced with a surgical blade. The leaves or half leaves of control ramets were not wounded.  40  Four different trials were conducted to detect the phenomenon of induced  resistance  in  black  cottonwood.  Each  trial  consisted  of  two  treatments: one received both the inducing and challenging inoculations, while the second, the control, received only the challenge inoculation. E a c h treatment contained  8 ramets. The number of uredial pustules  resulting  from the challenging inoculation was counted in the end.  Systemic  induced resistance with urediospores  as  inducing agent  The  inducing inoculation was applied to a leaf in the middle portion of a ramet. The challenging inoculation was applied 10 days later to the leaf directly above and below the inducing inoculated leaf. The experiment was repeated once.  Systemic induced resistance with wounding as inducing agent A leaf in the lower portion of a ramet was wounded. Expecting a quick response from the host, the challenging inoculation was applied an hour later to a leaf above the wounded one.  Local induced resistance with urediospores as inducing agent  The inducing  inoculation was applied to one side of the middle vein of an attached leaf. The challenging inoculation was applied to the other side of middle vein of the leaf 10 days later.  Local induced resistance with wounding as inducing agent  A n hour after  one half of a leaf divided by the middle vein was wounded, the other half of leaf was given a challenging inoculation. Attached leaves were used.  Figure 5. The location of black cottonwood ramets in the greenhouse for induced resistance study.  42  RESULTS  4.1  Variation i n Rust Resistance of Black Cottonwood  4.11  The pattern of rust development  The general pattern of Melampsora  rust development during  the  infection period i n this experiment is presented in Figure 6. The spread of leaf rust was relatively slow during the first several weeks after initial inoculation with aeciospores. A s the inoculum increased, the rate of disease build up also increased. The rust progressed rapidly i n midsummer and reached its peak in the early fall. The pattern of the rust development was more or less the same for all the clones, but the entire process for resistant clones (C2, C 4 and 16) appeared to be delayed by two or three weeks in comparison with susceptible clones (II, 14 and 15). Although all clones were inoculated at the same time, the susceptible clones had a much shorter time of initial inoculum buildup and reached a high rate of disease spread earlier than the resistant ones. The pattern of rust spread for each treatment (including the control) was similar to that shown i n Figure 6 although there were different levels of rust severity among treatments (Appendix 2).  Date (month.day) Figure 6. The average (n = 20) weekly relative disease severity (RDSW) of each clone.  CO  44  The correlation between individual measurements of R D S W and the final  disease  ratings  (RDST) increased  with the  date  of the  RDSW  measurement (Figure 7). The best correlation occurred on September 16; a point in the epidemic at which infection on the most susceptible clones approached saturation. Hence, rating the rust at this date would be the best time for a single measurement. It also was the only date on which the order of clones from least to most resistant was identical to that of R D S T .  The spread of rust on individual ramets followed a pattern from bottom to top. The older leaves of ramets were generally infected earlier than  younger  leaves  because  of their  greater  susceptibility or  longer  exposure to the rust inoculum. The rust constantly spread upward as the young leaves of black cottonwood expanded. It was observed that the newly formed leaves were usually free of the rust. This may be because young leaves have not been fully exposed to the rust or because their physiological activities contribute to the resistance.  45  1-i  Date (month.day)  Figure 7. Plot of the correlation coefficients (r) between individual measurements of disease severity (RDSW) and the total disease severity  (RDST).  46  4.12  Variation of rust resistance  The  relative resistance of different clones was manifested in their  disease severity ratings. The total relative disease severity (RDST) was used as the index of disease resistance average R D S T  in this study. Since the variance  of clones on each treatment was  of  not homogeneous, log  transformation of the data was conducted. The transformed data achieved the homogeneity. Analysis of variance ( A N O V A ) was conducted through the system of statistics ( S Y S T A T ) (Wilkinson 1988). The same system was also used in the rest of the analyses. Results (Table 1) demonstrated significant difference of R D S T  values among clones, which indicated the  apparent  clonal variation of resistance. T h e significant difference among treatments indicated that the attempt to generate a range of disease severities on each clone by applying different inoculation treatments was successful. But the specific interpretation of the variation is difficult since the treatments can not be purely regarded either as different inoculation times or as different initial amounts of inocula. This also made the explanation of the interaction between clone and treatment difficult although it can be partly attributed to the fact that the high susceptible clones exposed to treatment 1 and 2 reached saturation before the end of the season, so that the differences between  treatments  were  smaller  for  them  than  for  resistant  clones  (Appendix 3). The significant interaction suggests varied responses of clones to different treatments. Nevertheless, the disease development of clones was generally parallel over time (Figure 6).  47 Linear contrasts of average R D S T following the A N O V A showed that clones from the coast were more resistant than those from interior, and within the Interior, northern clones (II - 15) were more susceptible than southern ones (16 - 19) (Table 2). A l l the contrasts are highly significant (p < 0.001).  48  Table 1. Results of analysis of variance for relative disease severity (RDST) values with log transformation [RDSTjj^ =  Mean +  Clonej (random)  Treatment^ (fixed) + C j x Tj + E r r o r ] . k  Source of Variation  DF  MS  Clone  13  5.5856  41.6833  <0.0001  4  70.5043  203.4924  < 0.0001  C x T  52  0.3465  2.5393  <0.0001  Error  210  0.1364  Treatment  F  P  +  49  Table 2. T h e average (n = 20) relative disease severity (RDST) values of each clone.  Clone  Ecological Z o n e l  Average R D S T ,  Cl C2 C3 C4 C5  CWH CWH CWH CWH CWH  0.199 0.121 0.180 0.087 0.171  11 12 13 14 15 16 17 18 19  SBS SBS SBS SBS IDF ICH ICH IDF MS  0.324 0.267 0.271 0.382 0.380 0.135 0.173 0.184 0.222  1: Based on biogeoclimatic classification of British Columbia: C W H = Coastal Western Hemlock, S B S = Sub-Boreal Spruce, I D F = Interior Douglas-Fir, I C H = Interior Cedar-Hemlock, M S = Montane Spruce.  50  4.2  Effects of the Rust on the Growth of Ramets  4.21  T h e growth of ramets  The characteristics of easy rooting and rapid juvenile growth of black cottonwood were  well demonstrated  by all of the clones used i n this  experiment. In the summer of 1988, the ramets grew quickly i n both height and diameter. B y the end of the first growing season, the most rapidly growing clone reached an average height of over 130 cm and a diameter of over 11 m m . However, there was considerable variation of growth i n height (p <  0.001) and diameter (p <  0.001) among clones ( A N O V A ) . Although  all clones had uniform initial size, the slowest growing clone had achieved only 45 percent of the height and 67 percent of the diameter of the most rapidly growing clone. The ramets began to cease growing i n height by late August. A t the same time, diameter growth also decreased. T h e interior clones ceased height growth earlier than the coastal ones. Table 3 shows the average height and diameter of control ramets at the end of the first growing season. These control ramets were never inoculated and received several fungicide (Cyprex) applications i n order to prevent infection. They remained free of infection until the end of August, and then developed a few uredinia. Clones originating from the Fraser Valley were taller and had a greater final diameter than those from the interior of the province. The growth performance of these clones is shown i n Figures 8 and 9.  51  Table 3. Average of height (cm) and diameter (mm) of uninfected control ramets of each clone at the end of the first (1988) and second (1989) growing seasons.  Clone  First Year  Second Year  Height  Diameter  Height  Diameter  Cl C2 C3 C4 C5  111.00 139.75 109.50 128.25 103.50  10.00 10.75 8.75 11.13 10.00  151.25 186.00 143.25 167.25 145.00  13.50 13.50 12.75 13.75 13.13  11 12 13 14 15 16 17 18 19  74.50 82.00 110.50 83.25 62.75 102.75 113.75 88.25 92.50  7.88 9.50 9.75 9.50 7.50 7.88 9.00 8.00 7.75  124.50 123.25 129.25 132.75 149.50 151.50 123.00 132.75 138.50  13.63 12.00 13.00 12.13 12.50 12.00 12.13 12.25 13.38  52  Date (month.day) Figure 8. Average height of uninfected clones from the interior and coast during the first growing season.  53  7  -  2  8  8.27  9.26  10.26  Date (month.day)  Figure 9. Average diameter of uninfected clones from the interior and coast during the first growing season.  54  A t the start of the second growing season i n 1989, all the clones flushed at approximately the same time i n early A p r i l . T h e pattern of growth of the ramets was similar to that i n the first year except that the time difference i n cessation of height growth between the interior and coastal clones was not so apparent. Variation i n final height (p <  0.001)  and diameter (p = 0.004) was still great at the end of the second growing season. However, it had decreased to some extent compared to that of first year. T h e slowest growing clone reached 66 percent of the height and 87 percent of the diameter of fastest growing clone. The relative positions of some clones i n the ranking of these parameters changed dramatically from the first year. Clone 15 and 17 are two good examples. The former ranked last i n both the diameter and height i n the first year, but jumped to ninth in diameter and fifth i n height i n the second year. Clone 17 dropped from third i n height and eighth i n diameter i n the first year to the last i n height and twelfth in diameter in the second year (Table 3).  55  4.22  Relationship between growth characteristics and disease severity  The total dry weight of a black cottonwood ramet at the end of the first year was the most representative parameter of its overall performance during  the growing season  among  the growth  parameters  measured.  Therefore, it was selected to describe the relationship between rust severity and tree growth. Similarly, R D S T was chosen as the parameter expressing disease severity over the growing period. Plots of these two parameters for each clone demonstrated a linear relationship between them (Appendix 4). Simple  linear  regression  was applied  to depict  the relationship. T h e  regressions were highly significant for all the clones (Table 4). Coefficients of determination ranged from 0.40 to 0.95. Slopes for the different clones varied tremendously from -188 i n clone C 4 to -33 i n clone 15, as did the intercepts which ranged from 25 to 71 grams.  In order to test the hypothesis of the existence of a threshold infection level  i n the black  cottonwood  and Melampsora  rust  pathosystem,  a  quadratic term was added to the linear model for each clone. If there is a damage threshold, a sigmoidal phenomenon at low disease levels between total dry weight and R D S T would be expected; this requires a significant negative quadratic factor.  However, only two clones (C2 and C4) had  negative quadratic factors  and these were not significant (p >  0.420).  Therefore, it is unlikely that there is a threshold infection level in this pathosystem. T h e remaining clones had a positive quadratic factor, however, these were not significant (p ^ 0.260) except for clones C l , 13 and 17 (p <  56  0.080). For these three clones, the quadratic models were only slightly better than the linear models judging from the values of the multiple coefficient of determination ( R ) and the standard error of estimate. The 2  linear models of these clones were adopted for the sake of consistency.  The statement of no infection threshold was confirmed by an another test. If a threshold existed, then total dry weight would be unaffected by disease severity at disease levels below that threshold. Fitting a straight line to such a relationship would lead to an overestimate of yield at low disease severity levels. The test consists of selecting ramets with either the lowest, the two lowest, or the three lowest R D S T values for each clone and determining  whether  their  dry  weights  lie  predominantly below  the  regression line for that clone. If so, a case for a threshold could be made. It turns out that the regression lines underestimate the yield of the lightly diseased ramets i n 9 of 14, 19 of 28, and 28 of 42 cases for the three sets of points respectively. Therefore, there is no evidence for a threshold infection level in this pathosystem.  57  Table 4. Linear regression equation coefficients and statistical significance of the  relationship between total  dry weight ( T D W ) and relative  severity (RDST) of each of the 14 clones [ T D W = i + b R D S T ] .  Clone  i  b  r  z  p  Cl C2 C3 C4 C5  59.11 71.06 39.13 69.34 48.56  -92.76 -134.46 -68.17 -188.28 -100.36  0.700 0.740 0.666 0.572 0.530  <0.001 <0.001 <0.001 <0.001 <0.001  11 12 13 14 15 16 17 18 19  33.58 45.64 49.78 41.26 25.47 31.89 47.94 31.05 30.03  -39.08 -60.42 -84.37 -43.16 -32.95 -42.79 -88.13 -55.55 -49.70  0.750 0.830 0.953 0.839 0.696 0.404 0.615 0.630 0.587  <0.001 <0.001 <0.001 <0.001 <0.001 0.003 <0.001 <0.001 <0.001  disease  58  F r o m the relationship between total dry weight and relative disease severity (Table 4, Appendix 3), it can be clearly seen that a great loss of total biomass production was caused by the rust infection i n all the clones. Total dry weight reduction, as a percentage of uninfected controls was greater than the associated R D S T values. F o r instance, an R D S T rating of 0.05 caused 10 percent total dry weight reduction i n clone C 5 and 9 percent total dry weight reduction i n clone 17.  The relationship between stem dry weight and disease severity was linear, as was the relationship between root dry weight and R D S T ratings. A s disease severity increased, both the stem and root weight decreased. These relationships were highly significant for all the clones (p <  0.003).  The coefficients of determination ranged from 0.391 to 0.920 for the stem dry weight  and from 0.601 to 0.910 for the root dry weight. T h e only  exception was clone 16, whose regression of stem weight on R D S T was not significant (Table 5 and 6).  The relationship between the ratio of stem weight to root weight of ramets and R D S T was linear within clones. T h e regression equations were significant for all the clones (p < 0.050) (Table 7). The results demonstrate that as disease severity increases, the root dry weight is more severely affected by the rust than the stem dry weight. It suggests a changed pattern of photosynthate allocation i n diseased ramets, such that photosynthate is preferentially allocated to the growth above ground.  59  Table 5. Linear regression equation coefficients and statistical significance of the  relationship between  stem dry weight (SDW) and relative  severity (RDST) of each of the 14 clones [ S D W = i + b R D S T ] .  Clone  i  b  r  z  p  Cl C2 C3 C4 C5  30.18 36.42 23.51 37.43 28.70  -39.45 -51.21 -33.49 -94.08 -50.06  0.658 0.453 0.616 0.457 0.413  <0.001 0.001 <0.001 0.001 0.002  11 12 13 14 15 16 17 18 19  15.84 21.14 30.86 18.31 12.96 15.42 22.29 15.37 16.69  -14.34 -17.31 -46.59 -14.62 -14.51 -5.44 -31.58 -19.67 -20.68  0.605 0.438 0.920 0.646 0.475 0.038 0.430 0.398 0.391  <0.001 0.002 <0.001 <0.001 0.001 0.407 0.002 0.003 0.003  disease  60  Table 6. Linear regression equation coefficients and statistical significance of the  relationship between  root  dry weight (RDW) and relative  severity (RDST) of each of the 14 clones [ R D W = i + b R D S T ] .  Clone  i  b  r^  p  Cl C2 C3 C4 C5  28.93 34.65 15.62 31.91 19.86  -53.31 -83.26 -34.68 -94.20 -50.30  0.689 0.863 0.634 0.658 0.601  <0.001 <0.001 <0.001 <0.001 <0.001  11 12 13 14 15 16 17 18 19  17.74 24.51 18.92 22.95 12.51 16.47 25.65 15.68 13.34  -24.73 -43.11 -37.78 -20.54 -18.44 -37.35 -56.55 -35.88 -29.02  0.780 0.870 0.910 0.891 0.773 0.616 0.621 0.691 0.630  <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001  disease  61  Table 7. Linear regression equation coefficients and statistical significance of the  relationship between ratio of stem dry weight to root dry weight  ( S D W / R D W ) and relative disease severity (RDST) of each of the 14 clones [SDW/RDW = i + bRDST].  Clone  i  b  r  z  p  Cl C2 C3 C4 C5  1.00 1.01 1.22 1.13 1.47  1.71 2.41 6.85 1.55 3.91  0.557 0.676 0.785 0.220 0.361  <0.001 <0.001 <0.001 0.037 0.005  11 12 13 14 15 16 17 18 19  0.72 0.75 1.41 0.57 0.97 0.85 0.82 0.91 1.10  1.91 2.99 5.02 1.65 1.77 4.94 2.58 3.34 5.73  0.623 0.557 0.794 0.619 0.312 0.922 0.638 0.576 0.817  <0.001 <0.001 <0.001 <0.001 0.010 <0.001 <0.001 <0.001 <0.001  62  The slopes of the regression equations of the stem weight/root weight ratios over disease severity (Table 7) were significantly different among clones according to Cunia's (1973) method (p < 0.001) (Appendix 5). This indicates that the degree to which allocation to stems is favoured over roots in different clones varies among the clones. However, these ratios  were  poorly correlated with the average R D S T values of clones (r = 0.042, p = 0.887). This shows that the degree of resistance of clones is not related to their partition ratios. There was also no clear relationship with geographic origin of the clones.  The correlations between dimensional growth parameters  and rust  ratings were not as strong as those involving biomass parameters. However, considerable loss i n volume (D^H) of the main stem (not including branches) was demonstrated for most of the clones. Branching habit varied among the clones. The correlation coefficients between R D S T and volume ranged from 0.17 to -0.84, for R D S T and diameter, from -0.10 to -0.85, and for R D S T and height, correlated  from -0.19 to -0.62. Height growth was the most poorly  with  rust  severity  among  these  parameters.  Volume and  diameter growth of the ramets were not much different from each other in their relations with the disease severity. Significant linear relationships between volume and diameter with R D S T values were obtained for 12 and 13 out of 14 clones respectively (p < 0.080). Only seven significant linear relationships were found for the height growth at the same probability level. Variation  of the slopes of the linear equations  for the three  parameters was large among clones (Table 8, 9 and 10).  growth  63  Table 8. Linear regression equation coefficients and statistical significance of the relationship between volume ( D H ) and relative disease severity (RDST) 2  of each of the 14 clones [ D H = i + bRDST]. 2  Clone  i  b  x  p  l  Cl C2 C3 C4 C5  107.99 173.38 89.42 160.56 112.96  -66.43 -184.09 -93.54 -241.33 -161.46  0.133 0.160 0.335 0.170 0.247  0.114 0.080 0.008 0.071 0.026  11 12 13 14 15 16 17 18 19  49.67 61.72 107.99 76.52 42.92 63.02 95.25 60.02 57.93  -28.69 -38.98 -130.12 -50.22 -49.12 -26.26 -110.18 -63.24 -68.40  0.299 0.231 0.711 0.242 0.404 0.027 0.265 0.286 0.421  0.013 0.032 <0.001 0.028 0.003 0.486 0.020 0.015 0.002  64  Table 9. Linear regression equation coefficients and statistical significance of the relationship between diameter (D) and relative disease severity (RDST) of each of the 14 clones [D = i + b R D S T ] .  Clone  i  b  r  z  p  Cl C2 C3 C4 C5  9.81 10.95 9.07 11.02 10.31  -2.80 -4.78 -4.42 -6.57 -6.49  0.173 0.212 0.296 0.206 0.239  0.068 0.041 0.013 0.044 0.029  11 12 13 14 15 16 17 18 19  8.03 8.38 10.07 9.52 8.11 7.70 9.04 8.31 8.04  -1.63 -1.97 -6.61 -2.67 -4.12 -0.71 -4.57 -3.87 -4.91  0.259 0.212 0.724 0.281 0.376 0.010 0.277 0.262 0.334  0.022 0.041 <0.001 0.016 0.004 0.673 0.017 0.021 0.008  65  Table 10. Linear regression equation coefficients and statistical significance of the relationship between height (H) and relative disease severity (RDST) of each of the 14 clones [H = i + b R D S T ] .  Clone  i  b  r  z  p  Cl C2 C3 C4 C5  110.87 141.80 112.55 129.22 106.86  -11.66 -40.88 -56.95 -49.07 -69.87  0.025 0.058 0.384 0.075 0.226  0.503 0.307 0.004 0.244 0.034  11 12 13 14 15 16 17 18 19  77.02 87.24 109.83 81.44 72.57 102.79 114.71 88.22 93.23  -18.79 -18.36 -29.83 -11.50 -61.57 -17.31 -41.70 -36.88 -39.33  0.135 0.068 0.209 0.037 0.358 0.036 0.162 0.194 0.257  0.112 0.267 0.043 0.417 0.005 0.424 0.078 0.052 0.022  66  4.23  The effect of rust incidence during 1988 on the 1989 growth of ramets  The first visible sign of the extended effects of the rust infection was the appearance of lethal cankers on severely infected ramets in the early spring of 1989. These ramets were later killed, probably by a combination of frost damage and canker fungi (Figure 10). In severely infected ramets, leaf flushing was delayed for a week or so compared to control ramets.  Since  sample size was too small to make a valid comparisons among clones, average R D S T values of dead and surviving ramets of clones with mortality were compared using a t-test. The difference of 0.24 highly significant (p <  in R D S T value was  0.001). This suggests that ramets with high R D S T  values were more likely to die. However, some clones with higher R D S T values did not suffer more mortality than others. Clone 15 was such an example.  0.6 0.5 0.4 r  *  0.3 0.2 0.1  "  *  '  »  *  C2  C3  C4  C5  1 1  *  f  12  13  14  t  t  •: 15  t  *  '  t  16  17  18  19  Clone  Figure 10. Relative disease severity i total (RDST) values of dead and n  surviving ramets of each clone ( + : surviving, o: dead).  68  The cottonwood ramets remaining on the test site in 1989 received no treatment except fungicide protection. The ramets were virtually free of disease at the end of the season. The growth performance of the original control ramets,  however, was better than that of the previously rusted  ramets. The average initial volume increment (AIVI) [volume measured on June 21 in 1989 minus volume measured at the end of 1988] of the former was more than double that of the latter. Covariance analysis on average initial volume increment was conducted to exclude the effect of the volume (Appendix 6). The covariate was, on average,  1988  1.3 times bigger for  the control than the diseased ramets (Table 11). The significant covariate indicated  that  the  growth  of ramets  in the  second  year  was  partly  dependent on their size in the previous year. However, the relationship between average initial 1989  volume increment and 1988  volume within  both the previously rusted and the control group was found to be negative. This was probably because the pots had limited the growth of ramets with bigger sizes. Nevertheless, the covariance analysis showed that the diseased ramets performed significantly poorer than the control ramets. The final volume and total dry weight at the end of the 1989 growing season were difficult to interpret because the larger ramets were pot bound, and the effect of that on the 1989 effects of Melampsora  increment could not be distinguished from the  infection.  69  Table 11. Average relative disease severity (RDST) values i n 1988, volume (cm^)  at the end of 1988, and average initial volume increment (AIVI)  (cm^) i n 1989 for 1988 diseased and control ramets.  Treatment  RDST  Vol.88  AIVI  Diseased  0.242  63.95  3.34  Control  0.026  82.57  7.66  70  4.3  Variation i n Rust Tolerance of Black Cottonwood  4.31  The definition of tolerance  The Melampsora  tolerance  of  individual  black  cottonwood  clones  to  the  rust infection was defined as the relationship between disease  severity (RDST) and the dry weight of ramets.  The growth reduction of  infected ramets is most completely expressed by the character of total dry weight although other yield parameters might also be used.  The  different regression  weight and R D S T  coefficients or  (Table 4) demonstrate  slopes between  total  variation of this trait  dry  among  different cottonwood clones. However, a reasonable comparison of these slopes can only be achieved by converting the total dry weight values into percentage values of their control ramets, since different clones have very different growth rates. Having done so, one ends up with a relationship in which both variables are expressed as proportions; yield as a proportion of control (uninfected) yield and disease severity as a proportion of the total number of leaf-weeks infected for a ramet. Because the control ramets were not absolutely free of disease, the conversion was achieved by dividing the individual  total  dry weight values of ramets by the  intercept  of the  regression line between total dry weight and R D S T for the clone. Then a new regression was built with the quotient and relative disease severity. The quotient was referred to as total dry weight in percentage,  and the  slope of the new regression line was called relative reduction rate (RDR) of  71  total  dry weight.  Table  12 lists the R D R values  of each  clone. T h e  parallelism of slopes of different clones was tested by using the method of Cunia  (1973)  (Appendix 5).  T h e result  showed that  the slopes  significantly different, which indicates that variation of rust  were  tolerance  among black cottonwood clones exists (Table 13).  In view of the fact that stem dry weight is also a yield variable of great interest,  the relative  reduction rates of stem  dry weight  (slope  between stem dry weight in percentage and R D S T ) of the clones except 16 were also calculated (Table 12). Comparison of these slopes (Appendix 7) showed that they are also significantly different (p <  0.001). In other  words, variation i n disease tolerance expressed as the relative reduction rate of stem dry weight also exists.  72  Table  12. Relative reduction rate of total dry weight (RDR),  relative  reduction rate of stem dry weight (RDR'), and products of R D R and R D S T (relative disease severity) of each clone.  Clone  RDR  RDRxRDST  RDR'  Cl C2 C3 C4 C5  -1.569 -1.892 -1.742 -2.715 -2.067  -0.312 -0.229 -0.314 -0.236 -0.353  -1.311 -1.439 -1.408 -2.513 -1.824  11 12 13 14 15 16 17 18 19  -1.164 -1.324 -1.695 -1.046 -1.294 -1.342 -1.839 -1.789 -1.655  -0.377 -0.354 -0.459 -0.400 -0.492 -0.181 -0.318 -0.329 -0.367  -0.893 -0.821 -1.522 -0.798 -1.157  ,  -1.432 -1.287 -1.242  73  Table 13. Parallelism test (Cunia 1973) of relative reduction rates of total dry weight (RDR) of 14 clones.  Source  DF  F u l l model  28  138.257  Without slopes  15  137.647  A d d . of slopes  13  0.610  0.047  252  6.530  0.026  Residual  SS  MS  F  P  1.808  0.042  74  4.32  Relationship between rust resistance and rust tolerance  If the tolerance of the 14 experimental clones is plotted against their resistance (Figure 11), a significant negative correlation results (r = -0.772, p < 0.001); the more susceptible clones were also more tolerant. Given such a  relationship, it is necessary  resistance  and tolerance  to determine  on the yield  the relative  of black  influence of  cottonwood.  This  will  determine whether resistance or tolerance is the more important attribute of a clone. This is done in Figure 11 by comparing the slope of the regression line of tolerance over resistance with an imaginary 'line of equal yield'.  The relationship between tolerance and resistance might be such that an  increase  in  yield  resulting  from  greater  resistance  is  exactly  counterbalanced by a decrease in yield resulting from a corresponding loss in tolerance. Figure 12 describes the relationship between yield and disease severity in clone 12 as:  [1]  Yield (total dry weight) = b x R D S T + i  where Yield of ramets is on a proportional basis (section 4.31), R D S T is the proportion of leaf area infected for each ramet, and b, the slope, (always negative) is the measure of tolerance. The intercept (i) is by definition 1.  The experimental methods were deliberately designed to produce a  75  *y*-0.2811*niss0.4225 y+3.579x-2.444 C4 •  0.05  0.1  0.15  0.2  0.25  0.3  0J5  0.4  Average RDST(x)  Figure 11. Plot of the relative reduction rate of total dry weight (RDR) of each clone against its average relative disease severity (RDST) value to show the linear relationship (solid line) between tolerance and resistance. The dotted line represents the relationship between RDR and RDST in which RDR x RDST is a constant, the value of that constant being selected so that the line is the best fit for the 14 observations.  76  Figure 12. Plot of yield (total dry weight as the proportion of iminfected control) of each ramet in clone 12 over its relative disease severity (RDST) value.  77 set of ramets with a range of disease severities i n each clone. In a normal plantation situation, the ramets of a particular clone would exhibit more nearly the same disease severity. The average disease severity of such a set of ramets would be a measure  of the susceptibility of the clone, to be  compared to that of other clones exposed to the same disease conditions. Since the experiment described in this study had a balanced design, the susceptibility of clones can be assessed by averaging disease levels (average R D S T ) of all ramets. The yield of various clones under a standard set of disease conditions can therefore be estimated as:  [2]  Y  t  = bj x R i + 1  where Y j is the yield on proportional basis of clone i ; bj is the slope (RDR or tolerance) for clone i ; and Rj is average R D S T value (index of resistance) under these conditions.  In Figure 11, bj is plotted against Rj. Hence a 'line of equal yield' is any line for which bpcRj is a constant. For clones lying along such a line, any  decrease  counterbalanced  in  yield  caused  by  reduced  by a corresponding increase  resistance  is  in yield due to  exactly increased  tolerance. The dotted line in Figure 11 has a value of bjxRj, such that the sum of the squared deviations of the experimental clones (the plotted points) from that line is at somewhat  more  a minimum.  important  It can be seen that resistance is a  attribute  than  tolerance.  The  eleven  less  resistant clones all have a smaller yield than that predicted by the line of equal yield, and the three most resistant clones (C2, C 4 and 16) all have a  78  greater yield. The best clone, that is the clone with the smallest reduction in yield due to Melampsora  rust, is the clone with the m a x i m u m bjxRj value.  T h a t turns out to be clone 16 (Table 13).  The effect of resistance and tolerance on yield can be more precisely defined. The black cottonwood populations could be exposed to various levels of disease hazard, the amount and timing of inoculum and the climatic conditions  being the  main  determinants.  The  disease  hazard  can  be  quantified as the average level of disease in a population of hosts. In this experiment, the disease hazard would be the overall average R D S T rating values. The yield of an infected clone (Y) may be estimated as:  [3]  Y = Y  where Y  u  - amount of disease x yield loss per unit of disease  u  is the uninfected yield of the clone. In this study, it was estimated  as the y-intercepts of the regression of total dry weight of ramets on their R D S T (Table 4). The amount of disease is a function of the hazard and the resistance of the clone in question. The dry weight loss per unit of disease is a function of the tolerance and the uninfected yield. Assuming that there is no interaction  between  the  expression  of resistance  and hazard,  then  rewriting [3] gives:  [4]  where  Yj = Y  Y^  and  u i  Y ^ u  + (H  are  o b s  the  /H  b a s e  ) x RDSTi x Tj x Y  infected  and  u i  uninfected yield  of clone  i  respectively; T\ is the tolerance of clone i (always negative); R D S T j is the  79  resistance of clone i and estimates the amount of disease on that clone when the hazard to which the population is exposed to H |  3 a s e  ; and H  0 D S  is the  observed hazard.  Equation [4] can be used as a more complete analysis for evaluating the performance of clones than that presented i n Figure 11, i n which the uninfected yield of a clone is not taken into account. Figure 13 presents the actual yield of clones as total dry weight predicted as a function of disease hazard. The hazard of experiment condition ( H  D a s e  ) is assigned the value of  1. Obviously, Clone C 2 , C 4 and C l are most desirable. The high uninfected yield of these clones is a major attribute. However, among the clones whose uninfected yield ranged from about 30 to 50 grams, it can be seen that the yields of clones depend on the degree of exposure to disease. F o r example, the uninfected yield of Clone 12, 13 and 16 are 45.64, 49.78 and 31.89 (g) respectively (Table 4). A t low disease hazards, Clone 13 has the highest uninfected yield. However, clone 13 is quite susceptible and only moderately tolerant (Figure 11), hence, as disease hazard increases, its yield falls off rapidly. A t median hazard, clone 12 has the highest yield. B u t under severe disease conditions, clone 16, which is the least productive of the three under disease free conditions, becomes the highest yielding clone.  80  Figure  13.  T h e predicted  yield  different disease hazard conditions.  (total dry  weight) of each clone  under  81  4.4  Induced Resistance in Black Cottonwood  Two repeated trials of the systemic induced resistance study failed to demonstrate induced resistance. The number of uredial pustules produced on challenging inoculated leaves  was  nearly constant, and they  were  also  spread evenly on each inoculated leaf. There was no significant difference in uredial pustule production between the rust induced and the control leaves no matter where they were situated (Table 14). The outcomes from the two clones tested were virtually the same. In both trials and both clones, there was  no  significant difference  between  the  average  number  of uredial  pustules on challenged and control leaves. The results seem to suggest that stimulation of one part of a ramet by urediospore inoculation has no effect on resistance of leaves directly above and below the inoculated leaf ten days later. Similar results were obtained using wounding as the inducing agent (Table 15). F o r the local induced resistance study, no significant differences were detected between the rust or wounding induced and the control halves of leaves (Table 15).  Table 14. M e a n number of uredial pustules of Melampsora  occidentalis per  leaf resulting from a challenging inoculation 10 days after an inducing inoculation and probability that means do not differ.  Exp. 1  Clone 12  n  ULl  LL  Induced  8  104  Control  8  P  Clone C 6  Exp. 2 n  ULl  LL  204  8  122  255  113  200  8  117  263  0.463  0.767  0.725  0.669  2  2  Induced  8  207  123  8  221  132  Control  8  219  133  8  216  142  0.530  0.511  0.752  0.480  P 1: upper leaves. 2: lower leaves.  83  Table 15. M e a n number of uredial pustules of Melampsora leaf or per half leaf resulting  oecidentalis per  from a challenging inoculation after an  inducing inoculation or wounding and probability that means do not differ.  Clone 12  B  210  8  220  8  A  Induced  8  Control  8  P  Clone C 6  n  n  1  0.581  n  C3  158  8  93  151  8  100  2  0.435  0.679  Induced  8  190  8  150  8  96  Control  8  186  8  145  8  91  P  0.955  0.720  1: Systemic study with wounding as inducing agent. 2: half leaf study with urediospores as inducing agent. 3: half leaf study with wounding as inducing agent.  0.700  84  DISCUSSION  5.1  The Effects of Melampsora  5.11  The impact of leaf rust  Rust on the Growth of Black Cottonwood  There have been a number of reports on the growth loss of Populus trees due to leaf rust infection. However, this experiment was the first to examine the relationship between rust infection and the yield reduction over a range of disease severities in the Populus-Melampsora  pathosystem. A  linear relationship between the growth of black cottonwood ramets  (as  measured by height, diameter, volume, root dry weight, stem dry weight and total dry weight) and the relative disease severity of Melampsora  rust  was demonstrated. T h e most important of these relationships was between the total dry weight and R D S T values, which indicated that for a single unit increase in disease severity, there would be a corresponding unit of loss in total biomass of the roots and stems. In other words, the impact of rust on biomass production of the cottonwood tree was additive for one season. The hypothesis of a threshold infection level in this pathosystem was rejected. It follows that any treatment at any time that reduces the amount of the rust will decrease the loss i n increment caused by the rust. The control ramets which were sprayed with Cyprex i n this experiment serve as an example. The fungicide was effective in preventing the disease spread, and resulted in light rust on these control plants, and hence limited damage.  85  The impact of leaf rust infection on the yield of black cottonwood is dramatic. It consists of three major components: (1) reduced growth of ramets i n the year of heavy disease; (2) the death of severely diseased ramets during the following winter i n same clones; and (3) reduced growth of ramets in the following year. T h e regression models given i n previous sections have shown the magnitude of reduction as expressed by various growth parameters for each clone i n the first growing season. The results confirm earlier reports of Spires (1976); and Widin and Schipper (1981) of heavy loss in growth of Populus  trees due to Melampsora  rust infection.  The infection by leaf rust also predisposed the cottonwood trees to canker fungi. In this experiment, 14 out of 112 ramets were killed (12.5 percent).  Similar damage was reported by Spires (1976). T h e effect of  Melampsora  rust infection i n 1988 was also reflected i n the reduction of  initial volume increment of diseased ramets i n 1989. It may be that the high reduction i n root weight reflects that reduction of storage materials i n infected cottonwood ramets. '  The leaf rust causes considerable damage to black cottonwood ramets, but the extent of damage varies greatly among clones and among growth parameters. This was i n contrast with the effect of injury or defoliation on poplar trees.  Bassman et al. (1982)  showed that 40 percent artificial  defoliation of hybrid poplars i n summer caused negligible growth impact on the poplar. Stimulated growth activity of the remaining leaves was also detected (Bassman and Dickmann 1982). Rust infection, however, not only  86  destroys the photosynthetic capacity of the occupied area (the chlorophyll deteriorates in these areas), but also acts as a sink to draw more materials from other parts of the plant. This was indicated by the greater percentage of total dry weight loss compared to the percentage  of disease  severity  ratings.  Melampsora  rust infection affected the allocation of photosynthate to  various plant parts. It was evident that photosynthate was preferentially allocated to shoot growth rather growth parameters  measured,  affected by the rust,  than roots in diseased ramets.  the root  dry weight was most  followed by stem  Among severely  dry weight, volume and basal  diameter. The effect on height growth was the least, probably because of its earlier cessation allows it to escape the final pressure from the rust. Similar results of the effects of Melampsora Populus Schipper  rust on height and diameter growth of  trees can be found in the reports of Spires (1974); and Widin and (1981).  T h e reduction  i n volume  and stem  weight  may be  considered as a real loss in the practical sense. T h e loss in root weight, however, may reflect reduction in storage materials which could restrict the further development of the tree and eventually cause losses i n yield.  5.12  The rationale for R D S T rating system  There are two features incorporation  of time  in the R D S T rating system.  as one of factors  that  comprises  One is the the total or  accumulated disease severity. Since the duration of disease is an essential factor of any disease epidemic, it is equally as important as the area  occupied by the  disease  in affecting the  growth of plants. Therefore,  including a time factor in the disease rating system is necessary.  In this study, the accumulated disease severity was expressed by R D S T values. It is actually the ratio of two areas under two curves: the denominator is the area under leaf area progress curve, and the numerator is the area under diseased leaf area progress curve. Unlike the  AUDPC  which gives the same weight to each time of disease assessment (like R D S W in this experiment), R D S T gives equal weight to individual measurements of adjusted leaf number and adjusted disease leaf number. Thus R D S T is the ratio of total number of diseased leaf-weeks to total number of leaf-weeks for the 13 week period of observation. The assertion that for a given clone R D S T is linearly related to yield loss implicitly assumes that each leaf-week of foliage makes the same contribution to total dry weight, independent of leaf position or time of year, and similarly that the loss in yield attributable to a particular area of disease for a particular observation is also constant, and  thus  independent of the  location of that area or the  time  span  considered. These assumptions might be tested by conducting an experiment like the  one described here  under a  variety  of climatic  and cultural  conditions, thus i n effect creating a set of disease progress curves of different shape. The four treatments and the control (Appendix 2) introduce such variation to some extent.  Another feature of the R D S T system is the direct recording of the amount of disease instead of describing it with numeric codes or using a scale. B y doing this, the percent reduction i n yield of cottonwood ramets,  88  then could be directly related and compared to the percentage of disease severity.  T h e assignment  of numeric values  to  arbitrarily determined  amounts of disease infection requires an estimation of the possible impact of that amount of disease on the plant. However, before there is a disease rating scale,  this impact can never be worked out;  thus the  numeric  assignment is largely done from subjective experience and judgement. Thus, the assertion that the relationship between disease severity and yield loss is linear may well be trivial, since the rating scheme may be specifically or deliberately designed to achieve that linear relationship.  There  could be  as  many  types  of numerical assignments  to  a  particular pathosystem as there are researchers who studied it. Comparison of disease severity among these systems is difficult or impossible. Unlike ordinary poplar rust rating systems (Schreiner 1959  and Jokela & Lovett  1976), artificial scores representing certain infected leaf area was avoided in the  RDST  system.  The individual R D S W  values were  simply counted.  Therefore, the rating is the objective recording of the state of the rust levels at the rating time. It achieves the consistency among researchers and even between  pathosystems.  Comparison of disease  severity  from different  sources becomes possible.  The design and judgement of a disease rating system is largely dependent on the purpose of the study. A lot of rating systems used i n crop disease assessment aim at building the best model to predict the potential yield losses (Teng 1985,  1987; Rouse 1988; Campbell and Madden 1990). In  such cases, if a highest r  2  value is generated, the model will be the almost  89 certain be the best. However, since weather and disease epidemics always change from year to year, the date(s) of observation that yield the highest r  z  i n critical-point or multiple-point models may change over the years too.  The consistency of a model can not easily be achieved.  This study was mainly concentrated on the examination of the nature and  extent  reduction  of  of the  relationship between  cottonwood  trees.  Although  disease  severity  RDST  also  and  growth  achieved  high  correlations with the total dry weight for most of the clones, the guiding thought of design the system was not particularly aimed at that. The system need to be as objective as possible. One way to estimate  the  precision of the R D S T is to actually calculate and measure the leaf areas and diseased leaf areas with measuring instrument (Lindow 1983)  and  compare them with visually estimated results. Measurement of total leaf area  can  easily  be  automated,  measurement of disease area.  but  the . difficulty  is  automating  the  90  5.2  Disease Resistance and Induced Resistance  5.21  Variation i n disease resistance  The clonal variation i n rust resistance demonstrated in this study substantiates the results of Hsiang and v a n der K a m p (1985), based on in vitro leaf disc tests, namely that there is wide variation in resistance to Melampsora  occidentalis  in the natural population of black cottonwood in  B . C . The range of R D S T from the most to the least resistant is about fourfold. Variation is present both among the interior and the coastal clones; and the latter, i n general, are more resistant than the former. This is probably because the coastal conditions are more favorable for the rust, resulting in a stronger selection pressure on trees for resistance to the rust. This is supported by the fact that clones from I C H which has similar summer climate exhibited about equal resistance to that of coastal clones. Different susceptibilities of young and mature leaves were observed i n this study  and by Shain  and Miller  (1982).  The possible difference in  susceptibility of cottonwood leaves on juvenile and mature trees were not assessed i n this study. Although variation i n resistance among mature cottonwood trees is significant (Hsiang and v a n der K a m p 1985), a study on correlation i n resistance ranking between juvenile and mature trees will be of interest.  5.22  Induced resistance  91  Contrary to the many reports of induced resistance detected in plants, this experiment found no evidence for local or systemic induced resistance in the  black  cottonwood  and  Melampsora  oecidentalis  pathosystem.  Furthermore, since wounding also failed to induce resistance to the rust, the possibility of using other agents to induce the resistance in black cottonwood on both local and systemic levels is not promising. This result is similar to that reported by Paine and Stephen (1987), which might suggest that the phenomenon of induced resistance is not as common or apparent in forest trees as in herbaceous plants.  Demonstration of the absence of a phenomenon is difficult. It can always  be argued that with different, more  appropriate  methods,  the  demonstration would have been successful. Limited by time and space, this experiment did not create a range of infection or wounding intensities and intervals between the inducing and challenging inoculations in order to detect the induced resistance. However, the method of induction used here resembles closely the sequence of infection events experienced by the host under natural conditions. It was therefore suggested that induced resistance did not play a role in the epidemic studied. Thus, the resistance of a clone can apparently be estimated  equally well from a single inoculation of  experimental materials or from the total amount of disease over the infection period.  accumulated  92  5.3  Disease Tolerance  5.31  The demonstration of variation i n disease tolerance  This study is the first to demonstrate variation in tolerance of black cottonwood clones i n response to the infection by Melampsora measurements  rust. E a r l y  of tolerance were largely dependent on the judgement of  individual researchers (Salmon and Laude 1932; Ellis 1954). Caldwell et al (1958) succeeded i n creating an identical disease level by a well designed experiment  using "isogenic" varieties.  B u t i n many cases,  the disease  severity of compared strains was not precisely identical, but said to be comparable (Clifford 1968). Conclusions on disease tolerance were drawn from an overall view of data from different rust races and different years (Simons 1966). A more useful measurement of tolerance is the slope of yield on disease severity (Lipps and Madden 1989). In this study, a range of rust infection levels of ramets within each clone was induced. Linear regressions for  various  yield  parameters  as  a  function of disease  severity  were  calculated for each clone. T h e slopes of these regression lines then became the measure of tolerance.  The  relationship between disease  severity  (RDST) and total dry  weight was then chosen as the most meaningful among growth parameters measured. T h e significant differences i n relative reduction rate of total dry weight suggest that the variation of rust tolerance between clones of black cottonwood is real. T h e same result was obtained by using stem dry weight  93 as the yield parameter. It appears that clones from the interior are more tolerant than clones from the Fraser Valley.  Total dry weight production of cottonwood ramets depends critically on cultural conditions. Temperature, light intensity, water availability, wind speed, relative humidity and mineral nutrition regimes, among others, will all affect the yield. A l l of these can be experimentally varied, and they will certainly differ from year to year under field conditions. It would be rash to assume that the absolute value of tolerance, as defined in this study, does not depend to some degree  on the experimental conditions. A  further  question is whether the rankings of clone by tolerance dependent on the cultural conditions. The experimental work described in this thesis can not address that question. Lipps and Madden (1989) show substantial variation in tolerance  of wheat  to powdery mildew among years.  Those results,  however, might also in part or whole be a consequence of the  disease  severity measure they used. The relationship between that critical point measure and total disease load might well vary from year to year, and thus account for the apparent variation i n tolerance.  5.32  Disease tolerance versus disease resistance  The nature of disease resistance concerns the ability of a plant to resist and restrict the establishment and spread of pathogens within it. O n the other hand, disease tolerance concerns the capacity of a plant to limit the adverse effects brought about by the pathogens on its growth. Disease resistance, generally, is expressed i n the form of disease severity ratings. It  9 4  differs from disease tolerance i n that it merely reflects the condition of a diseased plant at a certain time or over a time period, without considering the effect on plant growth. Variation in tolerance is of considerable practical interest. Clearly, high tolerance is a desirable characteristic of a commercial crop species. Selection for tolerance  might be as useful as selection for  resistance. One great advantage of using tolerance to reduce yield losses caused by disease  is that it does not impose a  selection pressure  pathogens for greater pathogenicity. However, variation in tolerance  on was  not as great as variation in resistance in this study. It follows that gains of yield in the presence of Melampsora  rust will be more difficult to achieve by  selection for tolerance than for resistance. Also, tolerance is more difficult to measure than resistance.  Rust resistance was inversely correlated with rust tolerance in black cottonwood. The resistant clones are not as tolerant as the susceptible ones. Thus, there are trade-offs between rust resistance and rust tolerance in black cottonwood. The tolerant clones, although badly infected, can still maintain a certain amount of growth. The resistant clones, on the other hand, are less tolerant and have a higher rate of growth reduction. E v e n so, they are resistant (i.e. they do not exhibit as severe disease as susceptible clones), so the effect of the rust is also limited on these clones.  Studies of other  pathosystem  have  not always  shown the  same  phenomenon. For instance, Lipps and Madden (1989) reported that the wheat cultivars that were least resistant  to powdery mildew were  not  necessarily the most tolerant. However, comparison of these results with  95 those from this study need to be cautious, because: (1) there is no statistics associated with Lipps and Madden's results, it is not known whether those slopes among different varieties or years are significantly different or not; (2) The assessment of disease severity in Lipps and Madden's work was based on top 5 leaves of a wheat and did not take plant size into account, thus varieties  or varieties  over different years  with the  same  disease  severity scale may actually contain different amounts of disease; (3) The yield parameter  used to assess tolerance i n Lipps and Madden's work is  grain yield which is different from the total dry weight used i n this study. The allocation of photosynthate to grain could vary among varieties and years. A l l these factors may actually affect the relative ranks in tolerance of wheat  varieties  with  different  susceptibilities,  which  makes  direct  comparison of results from the two studies impossible.  The tolerance  physiological basis and resistance  cottonwood clones  have  for  in this  the study  negative is not  relationship  clear.  It  between  may  be  that  different strategies for responding to the  rust  infection. It may be that resistant clones spend relatively more energy in defence than susceptible ones. It may also be that tolerant clones have stronger  compensating  effects  than  intolerant  ones.  Grain  yield  of  symptomless resistant barleys which were inoculated with a avirulent race of powdery mildew was decreased 7% compared to uninoculated controls (Smedegaard-Petersen  1980).  This  demand in inoculated , resistant growth.  suggests  that  the  increased  energy  plants is sufficient to influence plant  96  These findings have a significant bearing on our understanding of natural selection and the resulting natural populations. One presumes that selection (in the presence of Melampsora)  will proceed towards both greater  resistance and greater tolerance. The clones described i n this study form parts of cottonwood populations that have resulted from such selection. If resistance each  new  and tolerance were inherited independently, one would expect sexual  generation  of the  host  to  exhibit a  range  of both,  independently distributed. Thus within a local population of cottonwood, one would expect no relationship between resistance and tolerance. A t the same time, if one were to compare cottonwoods from a range of local populations exposed to a range  of hazard levels (the  14  clones i n the  experiment  represent such a set), then one would expect clones from high hazard areas to be both more resistant and more tolerant than clones from low hazard areas. However, this study revealed the opposite relationship among the 14 clones. Thus, it may well be that the two phenomena are related at a physiological level. One effect of the apparent relationship between these two attributes resistance  of clones will be the maintenance  and tolerance  in the  of a wider range  population than would be the  of  case if  selection for the two attributes proceeded independently.  The quantification of yield as a function of resistance and tolerance will help to evaluate the best cottonwood clones against the  Melampsora  rust. Clones with maximum products of R D R and R D S T are most desirable. However, the choice of best clone is also affected by the different disease hazards and the yield of a clone under disease free conditions. The analysis presented in section 4.32 provides the logical steps necessary to quantify the  97 effects of resistance, tolerance, and hazard on yield. Although this analysis is based on the yield as total dry weight, a parallel approach using stem dry weight  as  the  variable  of  interest  might  also  be  conducted  since  determination of root weight is often not feasible. The answers, i n terms of the ranks of clonal performance for various hazard levels would then be different, but the procedures are identical. In either case, the effect of infection i n a particular year on the growth in later years has not been considered, although some preliminary results, presented above, show that such effects might be substantial.  It should be pointed out that the evaluation of clones from this study is mainly from the perspective of forest pathology. The leaf rust is just one of numerous factors judgements  that affect  on performance  the  yield of a cottonwood clone. The  of clones are  mostly based on the  results  obtained from one growing season. If the cottonwood ramets were allowed to grow i n following several years and be,infected every year. The results in terms of accumulated effects of the rust would be of potential interest.  98  CONCLUSIONS  T e n conclusions were drawn from this study:  (1)  The 14 western black cottonwood clones used in this study exhibited  considerable variation in resistance to Melampsora  occidentalis  leaf rust,  both within and between the two geographic areas (coast and interior) from which they were collected. Coastal clones were, on average, more resistant than interior clones, and within the Interior, southern clones were more resistant than northern clones.  (2)  Induced  resistance  (in  response  to  Melampsora  inoculation  or  wounding) was not detected i n this pathosystem. Thus, resistance of a clone can legitimately be estimated equally well from a single inoculation of from the total amount of disease accumulated over a period.  (3)  T h e relationship between various yield parameters (total dry weight,  dry weight of tops or roots, and stem volume) and disease severity (as the proportion of the total number of leaf-weeks of foliage infected) can be adequately described by linear equations.  (4)  T h e percentage reduction i n yield (total dry weight) resulting from  infection  by  Melampsora  rust  was  greater  than  the  corresponding  percentage of leaf-weeks of foliage infected in all clones tested. This suggests  99 that infected leaf areas not only fail to photosynthesize but also that they act  as  sinks of photosynthate,  large enough to outweigh the effect of  compensatory photosynthesis in uninfected leaves or leaf parts, if any.  (5)  T h e hypothesis that in this pathosystem there is a critical level of  infection such that any infection below that level (the threshold) will not result in any yield loss, was not supported by evidence. A caution is necessay, however. A threshold over a very narrow range of infection would not have been detected by the methods used.  (6)  Rust infection resulted in the preferential allocation of photosynthate  to top growth. Shoot/root ratios increased rapidly with increasing disease levels i n all clones. Furthermore, that phenomenon was significantly more pronounced i n some clones than in others.  (7)  Additional losses resulting from rust infection included mortality of  severely infected ramets of some clones during the dormant season following the year of infection, and reduced initial growth during the growing season following the year of infection.  (8)  Tolerance of cottonwood to Melampsora  slope  of  the  relationship  between  yield  rust can be described by the (expressed  uninfected controls) and disease severity (expressed  proportion  of  as proportion of the  total number of leaf-weeks infected).  (9)  as  Clones varied significantly in tolerance so defined.  100  (10)  T h e relationship between  tolerance  and resistance  of clones  was  negative. Both phenomena were shown to have an effect on the yield of clones in the presence of Melampsora.  Thus, increases in yield resulting  from greater resistance were partly counterbalanced by decreases in yield due to decreased tolerance. The effect of resistance was somewhat greater than that of tolerance.  The demonstration of disease  tolerance  and its relationship with  disease resistance opened a new area for research i n the black cottonwoodMelampsora complexity  pathosystem. of host  reactions  physiological and genetic ecological  Results  significance  in  to  basis the  of  the  pathogen  study  further  showed  infection. Searches  of these phenomena,  as  natural  should  pathosystem,  well  for as  the the their  continue.  Incorporation of the concept of disease tolerance will certainly be beneficial to breeders in their effort to procure superior trees for maximum wood production.  101  BIBLIOGRAPHY  Anchisi, M . , Gennri, M . , and M a t t a , A . 1985. Retardation of Fusarium wilt symptoms in tomato by pre and post inoculation treatments of the roots and aerial parts of the host in hot water. Physiol. Plant Pathol. 26: 175-183. Anderbrahan, T . , and Wood, R. K . S. 1980. The effect of ultraviolet radiation on the reaction of Phaseolus vulgaris to species of Colletotrichum. Physiol. Plant Pathol. 17: 105-110. Anonymous, 1988. Victoria, B . C .  Biogeoclimatic zone of British Columbia.  MAPS-BC,  Bassman, J . H . , and Dickmann, D . I. 1982. Effects of defoliation i n the developing leaf zone on young Populus x euramericana plants. I. photosynthetic physiology, growth, and dry weight partitioning. For. Sci. 28: 599-612. Bassman, J . H . , Myers, W . , Dickmann, D . I., and Wilson, L . 1982. Effects of simulated insect damage on early growth of nursery-grown hybrid poplars in northern Wisconsin. C a n . J . For. Res. 12: 1-9. Caldwell, R. M . , Craybill, H . R., Sullivan, J . T . , and Compton, L . E . 1934. Effect of leaf rust Puccinia triticina on yields , physical characters, and composition of winter wheats. J . A g r . Res. 48: 1049-1071. Caldwell, R. M . , Schafer, J . R., Compton, L . E . , and Patterson, F . L . 1958. Tolerance to cereal leaf rusts. Science 128: 714-715. Campbell, C . L . , and Madden, L . V . 1990. Introduction to plant disease epidemiology. Wiley and Sons, Toronto, pp. 532. Caruso, F . L . 1977. Protection of cucumber against Colletotrichum lagenarium by Pseudomonas lachrymans. Proc. A m . Phytopathol. Soc. 3:159 (Abstr.). Caruso, F . L . , and K u c , J . 1977a. Protection of watermelon and muskmelon against Colletotrichum lagenarium by Colletotrichum lagenarium. Phytopathology 67: 1285-1289. Caruso, F . L . , and K u c , J . 1977b. Field protection of cucumber, watermelon, and muskmelon against Colletotrichum lagenarium by Colletotrichum lagenarium. Phytopathology 67: 1290-1292. Chiba, S., and Nagata, Y . 1973. Studies on the breeding of Populus maximowiczii clones selected from progenies of intraspecific hybridization. Oji Inst. For. Tree Improv. Tech. 124.  102  Ciesla, E . , Czajka, J . , and Siwecki, R. 1975. Field observations on the resistance of poplars infested by Melampsora sp. Arbor. K o m i . 20: 279-289. Clarke, D . D . 1984. Tolerance of parasitic infection in plants, In Plant disease: infection, damage and loss. Edited by Wood, R. K . S., and Jellis, G . J . Blackwell Scientific Publications, Oxford, pp. 119-127. Clifford, B. C . 1968. Relations of disease resistance mechanisms to pathogen dynamics in oat crown rust epidemiology. P h . D thesis, Lafayette, Ind. Purdue Univ., (Diss. Abstr. 2 9 < 3 > B : 835-836). Cobb, N . A . 1892. Contribution to our economic knowledge of the Australian rusts (Uredineae). Agric. Gaz. N . S. Wales 3: 60-68. C u n i a , T . 1973. D u m m y variables and some of their uses in regression analysis, in Proceedings of I U F R O meeting, V o l . 1 : 1-146. Daly, J . M . , and Sayre, R. M . 1957. Relations between growth and perspiratory metabolism in safflower infected by Puccinia carthami. Phytopathology 47: 163-168. Dean, R. A . , and K u c , J . 1986. Induced systemic protection i n cucumber: time of production and movement of the signal. Phytopathology 76: 966970. Dean, R. A . , and K u c , J . 1987. Rapid lignification in response to wounding and infection as a mechanism for induced systemic protection in cucumber. Physiol. Molecul. Plant Pathol. 31: 69-81. Doubrana, N . S., Dean, R. A . , and K u c , J . 1988. Induction of systemic resistance to anthracnose caused by Colletotrichum lagenarium i n cucumber by oxalate and extracts from spinach and rhubarb leaves. Physiol. Molecul. Plant Pathol. 33: 69-79. Duniway, J . M . , and Durbin, R. O . 1971. Detrimental effect of rust infection on the water relations of bean. Plant Physiol. 48: 69-72. Ellis, R. T . 1954. Tolerance to the maize rust Puccinia Nature 174: 1021.  polysora  Underw,  Fried, P. M . , MacKenzie, D . R., and Nelson, R. R. 1979. Disease progress curves of Erisiphe graminis f. sp. tritici on Chancellor wheat and four multilines. Phytopathol. Z . 95: 151-166. F r y , W . E . 1978. Quantification of general resistance of potato cultivars and fungicide effects for integrated control of potato late blight. Phytopathology 68: 1650-1655. Gallo, L . A . , Stephan, B. R., and Krusche, D . 1985. Genetic variation of Melampsora leaf rust resistance in progenies of crossings between and within Populus tremula and P. tremuloides clones. Silvae Genet. 34: 208-  103  214. Gao, Y . 1981. Infection cycle of Melampsora larici-populina Kleb. and selection of poplar stocks resistant to rust. J . Northeastern For. Inst. 3: 1018. Gaunt, R. E . 1981. Disease Phytopathology 71: 915-916.  tolerance  - an indication of thresholds?  Gessiler, C . , and K u c , J . 1982. Appearance of a host protein in cucumber plants infected with virus, bacteria and fungi. J . E x p . Bot. 33: 58-66. Griffiths, E . 1984. Foliar disease: the damage caused and its effect on yield. In Plant disease: infection, damage and loss. Edited by Wood, R. K . S., and Jellis, G . J . Blackwell Scientific Publications, Oxford, pp. 149-160. Guedes, M . E . M . , Richmond, S., and K u c , J . 1980. Induced systemic resistance to anthracnose in cucumber as influenced by the location of the inducer inoculation and fruiting. Physiol. Plant Pathol. 17: 229-233. Hammerschmidt, R., Nuckles, E . M . , and K u c , J . 1982. Association of enhanced peroxidase activity with induced systemic resistance of cucumber to Colletotrichum lagenarium. Physiol. Plant Pathol. 20: 73-82. Heather, W . A . , Sharma, J . K . , and Miller, A . G . 1980. Physiologic specialization in Melampsora larici-populina Kleb. on clones of poplar demonstrating partial resistance to leaf rust. Aust. For. Res. 10: 125-131. Hoes, J . A . and Dorrell, D . G . 1979. Detrimental and protective effects of rust in flax plants of varying age. Phytopathology 69: 695-698. Horsfall, J . G . , and Barratt, R. W . 1945. A n improved grading system for measuring plant disease. Phytopathology 35: 655 (abstr.). Hsiang, T . , and v a n der K a m p , B . J . 1985. Variation i n rust virulence and host resistance of Melampsora rust on black cottonwood, C a n . J . Plant Pathol. 7: 247-252. James, W . C . 1971. A n illustrated series of assessment keys for plant diseases, their preparation and usage. C a n . Plant Dis. Surv. 51: 39-65. James, W . C . 1974. Assessment of plant diseases and losses. A n n u . Rev. Phytopathol. 12: 27-48. Jenns, A . E . , and K u c , J . 1977. Localized infection with tobacco necrosis virus protects cucumber against Colletotrichum lagenarium. Physiol. Plant Pathol. 11: 207-212. Johnston, C . O . , and Huffman, M . D . 1958. Evidence of local antagonism between two cereal rust fungi. Phytopathology 48: 69-70. Jokela, J . J . 1966. Incidence and heritability of Melampsora  rust in Populus  104  deltoides Bartr. In Breeding pest resistant trees. Edited by Gerhold, A . D . , and et al. Program Press, New York. pp. 111-117. Jokela, J . J . , and Lovett, W . R. 1976. Selection and breeding eastern cottonwood for resistance to foliage disease. Proceedings of the 15th meeting of the Canadian tree improvement association, part 2. 12th lake states forest tree improvement conference I: (Abstr.) Ottawa, Canada; Canadian Forestry Service. 95 pp. K r z a n , Z. 1981. Poplar resistance to infection by the fungus larici-populina in field conditions. Arbor. K o r n i . 26: 123-142.  Melampsora  Kuc, J . , Shockley, G . , and Kearney, K . 1975. Protection of cucumber against Colletotrichum lagenarium by Colletotrichum lagenarium. Physiol. Plant Pathol. 7: 195-199. K u c , J . , and Richmond, S. 1977. Aspects of protection of cucumber against Colletotrichum lagenarium by Colletotrichum lagenarium. Phytopathology 67: 533-536. Lemoine, M . and Pinon, J . 1978. Clonal differences susceptibility to the rusts Melampsora larici-populina populina. Revue For. F r a n . 3: 181-185.  i n poplars i n and M. allir-  Large, E . O , and Doling, D . A . 1963. Effect of mildew on yield of winter wheat. Plant Pathol. 12: 128-130. Lindow, S. E . 1983. Estimating disease severity of single plants. In Symposium on estimating yield reduction of major food crops of the world. Phytopathology 73: 1575-1600. Lipps, P. E . , and Madden, L . V . 1989. Assessment of methods of determining powdery mildew severity i n relation to grain yield of winter wheat cultivars in Ohio. Phytopathology 79: 462-470. Littlefield, L . J . 1969. Flax rust resistance induced by prior inoculation with an avirulent race of Melampsora lini. Phytopathology 59: 1323-1328. Livine, A . 1964. Photosynthesis in healthy and rust affected plants. Plant Physiology 39: 614-621. Matta. A . 1979. Resistenza biologicamente batterichee fungine. Inf. Fitopatol. 8:17-29.  indotta  verso  malattie  Matthews, R. E . F . 1970. Plant virology. N e w York: Academic Press. 778 pp. Mclntyre, J . L . , and Dodds, J . A . 1979. Induction of localized and systemic protection against Phytophthora parasitica var. nicotiana by tobacco mosaic virus infection of tobacco hypersensitive to the virus. Physiol. Plant Pathol. 15: 321-330.  105  Mclntyre, J . L . , Dodds, J . A . , and Hare, J . D . 1981. Effects of localized infection of Nicotiana tabaccum by tobacco mosaic virus on systemic resistance against diverse pathogens and an insect. Phytopathology 71: 297301. M c K i n n e y , H . H . 1923. Influence of soil temperature and moisture on infection of wheat seedlings by Helminthosporium sativum. J . A g r i . Res. 26: 195-218. Mlodzianowski, F . , and Siwecki, R. 1976. Ultrastructure of poplar leaves naturally infected by rust Melampsora larici-populina Kleb. Arbor. K o m i . 21: 375-400. Mlodzianowski, F . Werner, A . , and Siwecki, R. 1978. Germination of Melampsora larici-populina urediospores on poplar leaves. European Journal of Forest Pathology 8: 119-125. Nagel, C . M . 1949. Leaf rust resistance within certain species and hybrids of Populus, Phytopathology 39: 16. Orton, W . A . 1909. The development of farm crops resistant to disease. U . S . Dep. A g r . Yearbook 1908: 453-464. Ostry, M . E . , and McNabb, H . S., J r . 1985. Susceptibility of Populus species and hybrids to disease in the North Central United States. Plant Dis. 69: 755-757. Ostry, M . E . , and McNabb, H . S., J r . 1986. Populus species and hybrid clones resistant to Melampsora, Marssonina, and Septoria. Research Paper, North Central Forest Experiment Station, U S D A Forest Service, N o . N C 272. Peace, T . R. 1962. Pathology of trees and shrubs. Oxford U n i . Press London. 753 pp. Paine, T . D . , and Stephen, F . M . 1987. Influence of tree stress and site quality on the induced defense system of lobololy pine. C a n . J . For. Res. 17: 569-571. Perucca, M . , & M a t t a , A . 1988. Induction of a transitory state of resistance to vascular wilt of tomato by means of different abiotic stresses. Rivista di Patologia Vegetale 22: 116-124. (Rev. of Plant Pathol. 67: 3621). Pojar, J . K . , K l i n k a , K . , and Meidinger, D . V . 1987. Biogeoclimatic ecosystem classification i n British Columbia. For. Ecol. Manage. 22: 119154. Rahe, J . E . , K u c , J . , Chuang, C . M . , and Williams, E . B . 1969. Induced resistance in Phaseolus vulgaris to bean anthracnose. Phytopathology 59: 1641-1645. Raymond, P. J . , Bockus, W . W . , and Norman, B . L . 1985. T a n spot of  106  winter wheat: Procedures to determine host response. Phytopathology 75: 686-690. Rouse, D . I. 1988. Use of crop growth-models to predict the effects of disease. A n n . Rev. Phytopathology 26: 183-201. Saari, E . E . , and Prescott, J . M . 1975. A scale for appraising the foliar intensity of wheat diseases. Plant Dis. Rep. 59: 377-380. Salmon, S. C , and Laude, H . H . 1932. Twenty years of testing varieties and strains of winter wheat. K a n s . A g r . E x p . Sta. Bull. 30. Schafer, J . F . 1971. Tolerance to plant disease. A n n u . Rev. Phytopathol. 9: 235-252. Schipper, A . L . , and Dawson, D . H . 1974. Poplar leaf rust - a problem in m a x i m u m wood fiber production. Plant. Dis. Rept. 58: 721-723. Schreiner, E . J . 1937. Improvement in forest trees, i n U . S . Dept. A g r . Yearbook 1937: 1242-1279. Schreiner, E . J . 1959. Rating poplars for Melampsora leaf rust infection. Forest Research Notes, Northeastern For. E x p . Sta., U S D A For. Serv. No.90. Sequeira, L . 1983. Mechanisms of induced resistance in pants. A n n u a l Rev. Microbiol. 37: 51-79. Shain, L . , and Jarlfors, U . 1987. Ultrastructure of eastern cottonwood clones susceptible or resistant to leaf rust. C a n . J . Bot. 65: 1586-1598. Shain, L , and Miller, J . B. 1982. Pinocembrin: an antifungal compound secreted by leaf glands of eastern cottonwood. Phytopathology 72: 877-880. Shaner, G . 1973. Evaluation of slow-mildewing resistance of K n o x wheat in the field. Phytopathology 63: 867-872. Simons, M . D . 1966. Relative tolerance of oat varieties to the crown rust fungus. Phytopathology 56: 36-40. Simons, M . D . , and Murphy, H . C . 1967. Determination of relative tolerance to Puccnia coronata avenae of experimental lines of oats. Plant Dis. Rept. 51: 947-950. Simons, M . D . 1968. Additional sources of tolerance to oat crown rust. Plant Dis. Rept. 52: 59-61. Siwecki, R., and Przybyl 1981. Water relations in the leaves of poplar clones resistant and susceptible to Melampsora larici-populina. E u r . J . For. Pathol. 11: 348-357. Siwecki, R. and Werner, A . 1980. Resistance mechanism involved in the  107  penetration and colonization of poplar leaf tissues by Melampsora Phytopathol. Medit. 19: 27-29.  rust.  Skipp, R. A . , and Deverall, B. J . 1973. Studies on cross protection in the anthracnose disease of bean. Physiol. Plant Pathol. 3: 299-314. Smedegaard-Petersen, V . 1980. T h e effect of defence reactions on the energy balance and yield of resistant plants. In Active defense mechanisms in plants. Edited by Wood, R. K . S., Plenum Press, New York and London. Spires, A . G . 1974. Control of poplar leaf rust Melampsora in New Zealand. N . Z. J . E x p . A g r . 2: 433-436.  larici-populina  Spires, A . G . 1976. Fungicides for control of poplar leaf rust and effects of control on growth of Populus nigra cv. 'Sempervirens' and P. x euramericana. N . Z. J . E x p . A g r . 4: 249-254. Stakman, E . C , and Harrer, J . G . 1957. Principles of plant pathology. New York: Ronald Press. 581pp. Teng, P. S. 1985. Construction of predictive models. II. Forcasting crop losses. In Advances in plant pathology. V o l . 3: Mathematical modeling of crop diseases. Edited by Gilligan, C . A . Academic Press, London. P P . 179206. Teng, P. S. 1987. Crop loss assessment and pest management. A P S Press, St. Paul. M N , pp. 270. Teng, P. S., and Johnson, K . B. 1988. Analysis of epidemiological components i n yield loss assessments. In Experimental techniques in plant disease epidemiology. Edited by Kranz, J . , and Rotem, J . Spring-Verlag, New York. pp. 179-189. Thielges, B. A . , and Adams, J . C . 1975. Genetic variation and heritability of Melampsora leaf rust resistance i n eastern cottonwood, For. Sci. 21: 278 282. Toole, E . R. 1967. Melampsora medusae causes cottonwood rust in lower Mississippi Valley. Phytopathology 57: 1361-1362. van der Meiden, H . A . , and v a n Vloten. 1958 Roest en schorsbrand als bedreiging v a n de teelt v a n populier. N e d . bosb. Tijdschr 30: 262-273, (For. Abstr. 20: 2018). Widin, K . D . , and Schipper, A . L . , J r . 1976. Epidemiology and impact of Melampsora medusae leaf rust on hybrid poplars. In Intensive plantation culture: five years research. U S D A For. Serv. G e n . Tech. Rep. N C - 2 1 : 6374. Widin, K . D . , and Schipper, A . L . , Jr. 1981. Effect of Melampsora medusae leaf rust infection on yield of hybrid poplars in the north-central United States. E u r . J . For. Pathol. 11: 438-448.  108  Wilkinson. L . S Y S T A T , Inc.  1988.  SYSTAT:  The system  for statistics.  Evanston,  IL:  Wilcox, J . R. and Farmer, R. E . , Jr. 1967 Variation in inheritance of Juvenile characters of eastern cottonwood. Silvae Genet. 16: 162-165. Ziller, W . G . 1974. The tree rusts of western Canada. C a n . For. Serv. Publication, No. 1329. Ottawa. 272 pp.  APPENDICES  Appendix 1. Composition of the fertilizer mix used in the experiment. Amount  Name  Cups(200ml)/per cubic yard of soil mixture  Nutricote  (17,17,17)  12  Gypsum  2.7  Dolomite  2.5  Limestone  2.5  Superphosphate  4  Micromax  1.6  Appendix 2. The average (n = 56) weekly relative disease severity (RDSW) of each treatment (T).  A\SQH  Ill  Appendix 3. The average (n = 4) relative disease severity (RDST) values in log transformation of each treatment of each of the 14 clones.  112  Appendix 4 . Plots of total dry weight of black cottonwood ramets of each clone over their relative disease severity (RDST) ratings.  Clone C l  RDST  Clone C2  114  CJoDe C3  10  Clone C4  116  Clone C5  RDST  117  C ] 0 D f II  RDST  118  Clone 12  Clone 13  120  Clone 14  121  Clone 15  25  r  RDST  Clone 16  50, 45 40 ^  35  "5> £  30  3  25 e  20  e e  15  10'  0.05  0.1  0.15  0.2 RDST  0.25  0.3  035  123  Clone 17  RDST  Clone 16  RDST  Clone 19  RDST  126  Appendix 5. A method of testing parallelism of linear regression lines. The test of parallelism of regression lines was achieved by Cunia's (1973) method of using dummy variables. If several regressions are said to be parallel, then they can be described by a common regression.  The  procedures are illustrated below:  1.  D u m m y variables (Dp were assigned as follows:  D  x  = 1 if clone C l  D j = 0 if clone not C l D  2  =  D  2  = 0 if clone not C 2  D^4 =  1 if clone C 2  1 if clone 19  D^4 = 0 if clone not 19  2.  New independent variables (Vj) were created by multiplying each  independent variable (Xp, by the corresponding dummy variables (Dp:  V  x  = D  x X  V  x  = Di xX  V  2  = D  2  x X  2  = X  V  2  = D  2  x X  2  = 0 if clone not C 2  1  x  2  = X  x  if clone C l  =0 if clone not C l 2  if clone C 2  V14 = D^4 x X^4 = X^4 if clone I9  127  V^4  3.  x X ^ 4 = 0 if clone not 19  =  A new  regression  without  a constant was  fitted using the  new  independent variable and the dummy variables (full model). A s a result, such  an  regression  equations.  The  is  equivalent  dummy variables  to  the  can  be  combination only 0  or  of 1,  14  separate  therefore,  the  coefficient for each new variable is the slope for that clone.  4.  To test if slopes were significant, a regression with dummy variables  was fitted by using 14 intercepts but only one regression coefficient (model without slopes). The sum of squares and degrees of freedom of 'addition of slopes' were obtained by subtracting the corresponding terms of the model without slopes from the full model. The mean squares of 'addition of slopes' (MS^d)  w  a  s  determined accordingly. A n F test was performed by F  M£>add/M!3res> where M S  r e s  is from the full model.  =  128  Appendix 6. Analysis of covariance  of average initial volume  increment  (AIVI) of black cottonwood ramets in 1989, the final volume of a ramet i n 1988 acts as a covariate  Source  MS  DF  F  P  AIVI  1050.858  1  14.519  <0.001  Vol.88  1030.375  1  14.305  <0.001  72.380  95  Error  129  Appendix 7. Parallelism test of relative reduction rates of stem dry weight of 13 clones (16 not included).  Source  DF  SS  MS  F  3.688  F u l l model  26  149.336  Without slopes  14  144.032  A d d . of slopes  12  5.304  0.442  254  27.756  0.109  Residual  P  <0.001  

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.831.1-0101045/manifest

Comment

Related Items