Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The influence of row covers and plant population density on the growth and productivity of bell peppers… Gaye, Mary Margaret 1990

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

Item Metadata

Download

Media
831-UBC_1990_A6_7 G39.pdf [ 7.74MB ]
Metadata
JSON: 831-1.0098134.json
JSON-LD: 831-1.0098134-ld.json
RDF/XML (Pretty): 831-1.0098134-rdf.xml
RDF/JSON: 831-1.0098134-rdf.json
Turtle: 831-1.0098134-turtle.txt
N-Triples: 831-1.0098134-rdf-ntriples.txt
Original Record: 831-1.0098134-source.json
Full Text
831-1.0098134-fulltext.txt
Citation
831-1.0098134.ris

Full Text

THE INFLUENCE OF ROW COVERS AND PLANT POPULATION DENSITY ON THE GROWTH AND PRODUCTIVITY OF BELL P E P P E R S (CAPSICUM ANNUUM L.) By MARY MARGARET GAYE B.Sc.(Agr.), The University of British Columbia, 1986 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE IN THE DEPARTMENT OF PLANT SCIENCE We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA May 1990 ©Mary Margaret Gaye, 1990 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 of The University of British Columbia Vancouver, Canada Date 13-DE-6 (2788) ABSTRACT An interlocking group of studies was conducted to examine the effects of row covers and plant population density on the growth and productivity of field-grown bell peppers. The studies were carried out at the Agriculture Canada Research Station, Agassiz, British Columbia, in 1988. Analysis of variance was used to determine treatment effects on reproductive and vegetative components of yield. A nonlinear regression model was used to define the yield responses. Yield components were assessed, at different stages of plant growth, for their contribution to reproductive and total plant yield variation, using a two-dimensional partitioning technique. The influence of growth and treatments on fruit and whole plant allometry was studied using a best subset multiple regression procedure. Row covers enhanced reproductive and vegetative yield per plant and per land area. Increasing plant population densities resulted in decreasing yield per plant, but increasing yield per land area. Furthermore, the effect of row covers on yield was greater at low population densities than high. The number of nodes was the most important contributor to variation in reproductive and total plant yield. Fruit weight as a proportion of total plant yield was also a major contributor to reproductive yield. The analysis showed the - iii -importance of row covers in the initial stages of growth, and the increasing importance of plant population density as growth proceeded, on yield components and on reproductive and total plant yield. Allometric relationships defining fruit and plant morphology changed during growth, and were also affected by row covers and plant population density. The changes were reflected through the allometric exponent and the allometric coefficient. Significant increases in horticultural yield resulted from the application of row covers and from high plant population densities. The response was quantified in mathematical models. Yield component analysis and allometric analysis of reproductive and total plant yield, proved to be valuable techniques for extending our understanding of plant growth relations resulting from the applied treatments. - iv -T A B L E O F C O N T E N T S ABSTRACT ii LIST OF TABLES viii LIST OF FIGURES xii LIST OF APPENDICES . xv ACKNOWLEDGEMENTS xvi CHAPTER I - INTRODUCTION 1 CHAPTER II - LITERATURE REVIEW 4 2.1 History, production and culture of peppers 4 2.2 Morphological characteristics of bell peppers 5 2.2.1 Nomenclature 5 2.2.2 Root system 6 2.2.3 Shoot system 6 2.2.4 Foliage 8 2.2.5 Flower 8 2.2.6 Fruit 9 2.2.7 Seed 10 2.3 Environmental factors affecting bell pepper development 11 2.3.1 Light 11 2.3.2 Temperature 12 2.3.3 Relative humidity 13 - V -2.4 Cultural influences on development and yield of bell peppers 14 2.4.1 Protected cultivation 14 2.4.2 Plant populations 15 2.5 Analytical techniques for interpreting plant development and yield . . 16 2.5.1 Yield-plant population density relationships 16 2.5.2 Yield component analysis 19 2.5.3 Allometric relationships 20 CHAPTER III - MATERIALS AND METHODS 22 3.1 Experimental design 22 3.1.1 Yield-plant population density study 25 3.1.2 Plant growth study 26 3.2 Data analysis 27 3.2.1 Yield-plant population density models 28 3.2.2 Yield component analysis 28 3.2.3 Allometric evaluation 30 CHAPTER IV - RESULTS 32 4.1 Plant microclimate 32 4.1.1 Soil temperature 32 4.1.2 Air temperature 34 4.1.3 Relative humidity 34 4.2 Yield-plant population density relationships 40 4.2.1 Effects of row cover and plant population density on fruit morphology: ANOVA results 40 4.2.2 Effects of row cover and plant population density on fruit yield: ANOVA results 40 4.2.3 Effects of row cover and plant population density on fruit yield: nonlinear regression models 47 - vi -4.2.4 Effects of row cover and plant population density on plant yield: ANOVA results 53 4.2.5 Effects of row cover and plant population density on plant yield: nonlinear regression models 53 4.2.6 Effects of row cover, plant population density and date of harvest on plant components: ANOVA results and spline regressions 58 4.3 Yield component analysis 61 4.3.1 Temporal trends of yield components 61 4.3.2 Effect of treatments on the contribution of yield components to fruit yield per plant: individual harvests 75 4.3.3 Effect of treatments on the contribution of yield components to fruit yield per plant: combined harvest data 79 4.3.4 Effect of treatments on the contribution of yield components to total shoot dry matter yield per plant: individual harvests . . . 82 4.3.5 Effect of treatments on the contribution of yield components to total shoot dry matter yield per plant: combined harvest data 86 4.3.6 Effect of treatments on the contribution of yield components to fruit yield per unit land area: individual harvests 87 4.3.7 Effect of treatments on the contribution of yield components to fruit yield per unit land area: combined harvest data 89 4.3.8 Effect of treatments on the contribution of yield components to total shoot dry matter yield per unit land area: individual harvests 94 4.3.9 Effect of treatments on the contribution of yield components to total shoot dry matter yield per unit land area: combined harvest data 97 4.4 Allometric relationships 100 4.4.1 Allometric relationships between fruit fresh weight and fruit morphological characteristics as affected by row cover, plant population density and date of fruit harvest 100 - vii -4.4.2 Allometric relationships between fruit fresh weight and fruit morphological characteristics as affected by row cover, plant population density and fruit location 103 4.4.3 Allometric relationships among plant weight and plant components as affected by row cover and plant population density 106 4.4.4 Allometric relationships between shoot weight and plant components as affected by row cover, plant population density and growth 109 CHAPTER V - DISCUSSION 112 5.1 Temperature effects 112 5.2 ANOVA and repeated measurements 113 5.3 Yield-plant population density relationships 114 5.3.1 Fruit characteristics 114 5.3.2 Yield response 114 5.4 Yield component analysis 116 5.5 Allometric relationships 118 CHAPTER VI - SUMMARY OF GROWTH AND YIELD OF BELL P E P P E R S 122 REFERENCES 124 APPENDICES 133 - viii -LIST O F T A B L E S Table 1. Experimental details 24 2. ANOVA F-statistics and EMS: fruit morphology 41 3. ANOVA F-statistics and EMS: early fruit fresh yield per plant 42 4. ANOVA F-statistics and EMS: total fruit fresh yield per plant 43 5. ANOVA F-statistics and EMS: early fruit fresh yield per land area 45 6. ANOVA F-statistics and EMS: total fruit fresh yield per land area 46 7. Nonlinear model statistics for fruit fresh weight 52 8. ANOVA F-statistics and EMS: plant components from yield density study 54 9. Nonlinear model statistics for dry weight plant components 59 10. ANOVA F-statistics and EMS from plant growth study: plant components from plant growth study 60 11. Two dimensional partitioning (TDP) analysis of fruit dry weight per plant: forward analysis of individual harvests 76 12. TDP of fruit dry weight per plant: backward analysis of individual harvests 77 13. TDP of fruit dry weight per plant: forward analysis of combined harvest data 80 - ix -14. T D P of fruit dry weight per plant: backward analysis of combined harvest data 81 15. T D P of shoot dry weight per plant: forward analysis of individual harvests 83 16. TDP of shoot dry weight per plant: backward analysis of individual harvests 85 17. TDP of shoot dry weight per plant: forward analysis of combined harvest data 87 18. TDP of shoot dry weight per plant: backward analysis of combined harvest data 88 19. TDP of fruit dry weight per land area: forward analysis of individual harvests . 90 20. TDP of fruit dry weight per land area: backward analysis of individual harvests 91 21. TDP of fruit dry weight per land area: forward analysis of combined harvest data 92 22. TDP of fruit dry weight per land area: backward analysis of combined harvest data 93 23. TDP of shoot dry weight per land area: forward analysis of individual harvests 95 24. TDP of shoot dry weight per land area: backward analysis of individual harvests 96 25. TDP of shoot dry weight per land area: forward analysis of combined harvest data 98 - X -26. TDP of shoot dry weight per land area: backward analysis of combined harvest data 99 27. Regression coefficients and other statistics for best subset multiple regression models of the allometric relationships between fruit fresh weight and morphological fruit characteristics: as affected by row cover, plant population density and harvest date 101 28. Standard partial regression coefficients for best subset multiple regression models of the allometric relationship between fruit fresh weight and morphological fruit characteristics: as affected by row cover, plant population density and harvest date 102 29. Regression coefficients and other statistics for best subset multiple regression models of the allometric relationships between fruit fresh weight and morphological fruit characteristics: as affected by row cover, plant population density and fruit location 104 30. Standard partial regression coefficients for best subset, multiple regression models of the allometric relationship between fruit fresh weight and morphological fruit characteristics: as affected by row cover, plant population density and fruit location 105 31. Regression coefficients and other statistics for best subset multiple regression models of the allometric relationships between plant dry weight and plant components: as affected by row covers and plant population density 107 32. Standard partial regression coefficients for best subset multiple regression models of the allometric relationship between plant dry weight and plant components: as affected by row cover and plant population density 108 - xi -33. Regression coefficients and other statistics for best subset multiple regression models of the allometric relationships between shoot dry weight and plant components: as affected by row cover, plant population density and growth 110 34. Standard partial regression coefficients for best subset multiple regression models of the allometric relationship between shoot dry weight and plant components: as affected by row cover, plant population density and growth 111 - xii -LIST O F F I G U R E S Figure 1. Shoot systems of bell peppers 7 2. Mean daily soil temperatures as affected by row covers: JD 137 to 188 (16 May to 6 July) 33 3. Mean, minimum and maximum air temperatures as affected by row covers: JD 134 to 188 (13 May to 6 July) 35 4. Diurnal air temperature fluctuation as affected by row covers: JD 166 (14 June) 36 5. Mean diurnal air temperature fluctuation prior to row cover removal: JD 182 to 186 (30 June to 4 July) 37 6. Mean daily air temperatures: JD 134 to 250 (13 May to 6 September). . 38 7. Per cent relative humidity as affected by row covers: (a) JD 166 (14 June) and (b) daily mean, JD 137 to 188 (16 May to 6 July) 39 8. Means and nonlinear regressions representing the relationship between early marketable fresh fruit yield and plant population density for uncovered and covered treatments; yield per plant and yield per m2. 48 9. Means and nonlinear regressions representing the relationship between early total fresh fruit yield and plant population density for uncovered and covered treatments; yield per plant and yield per m 2. . . 49 10. Means and nonlinear regressions representing the relationship between total marketable fresh fruit yield and plant population density for uncovered and covered treatments; yield per plant and yield per m 2. 50 - xiii -11. Means and nonlinear regressions representing the relationship between total fresh fruit yield and plant population density for uncovered and covered treatments; yield per plant and yield per m 2. . . 51 12. Means and nonlinear regressions representing the relationship between total plant dry yield and plant population density for uncovered and covered treatments; yield per plant and yield per m 2 55 13. Means and nonlinear regressions representing the relationship between vegetative plant dry yield and plant population density for uncovered and covered treatments; yield per plant and yield per m 2. . . 56 14. Means and nonlinear regressions representing the relationship between leaf area and plant population density for uncovered and '* covered treatments; yield per plant and yield per m 2 57 15. Spline regressions representing the relationship between shoot dry weight, growth and plant population density for uncovered and covered treatments 62 16. Spline regressions representing the relationship between vegetative shoot dry weight, growth and plant population density for uncovered and covered treatments 63 17. Spline regressions representing the relationship between leaf dry weight, growth and plant population density for uncovered and covered treatments 64 18. Spline regressions representing the relationship between leaf area, growth and plant population density for uncovered and covered treatments 65 19. Spline regressions representing the relationship between number of nodes per plant, growth and plant population density for uncovered and covered treatments 66 - xiv -20. Spline regressions representing the relationship between fruit fresh weight, growth and plant population density for uncovered and covered treatments 67 21. Spline regressions representing the relationship between number of nodes per m 2, growth and plant population density for uncovered and covered treatments 68 22. Spline regressions representing the relationship between leaf area (LA):number of nodes (NN), growth and plant population density for uncovered and covered treatments 69 23. Spline regressions representing the relationship between leaf dry weight (WL):leaf area (LA), growth and plant population density for uncovered and covered treatments 70 24. Spline regressions representing the relationship between shoot dry weight (W):leaf dry weight (WL), growth and plant population density for uncovered and covered treatments 71 25. Spline regressions representing the relationship between vegetative shoot dry weight (WV):leaf dry weight (WL), growth and plant population density for uncovered and covered treatments 72 26. Spline regressions representing the relationship between fruit dry weight (WF):vegetative shoot dry weight (WV), growth and plant population density for uncovered and covered treatments 73 27. Quadratic polynomial regressions representing the relationship between fruit dry weight (WF):shoot dry weight (W), growth and plant population density for uncovered and covered treatments 74 - XV -LIST O F A P P E N D I C E S Appendix I. Soil block media used for seedling culture 133 II. Means of morphological characteristics of fruit 134 III. Means of early fruit yield: kg per plant (fresh weight) 135 IV. Means of total fruit yield: kg per plant (fresh weight) 136 V. Means of early fruit yield: kg per m 2 (fresh weight) 137 VI. Means of total fruit yield: kg per m 2 (fresh weight) 138 VII. Means of plant components; yield density study . . 139 - xv i -A C K N O W L E D G E M E N T S I sincerely thank all those who have assisted me in the pursuit of this endeavour. I am especially indebted to Dr. Peter Jolliffe and Mr. Fred Maurer. These individuals have provided support and encouragement through my undergraduate and graduate years at U.B.C. I thank Dr. Jolliffe for his patient teaching and guidance over the last three years and Mr. Maurer, for his generosity in the conduct of this project. I thank Dr. George Eaton and Dr. John Hall (Agriculture Canada) who provided invaluable help with the analysis of variance. Ms. Maureen Henderson, Ms. Carolyn Schettler and Ms. Franziska Seywerd provided the assistance necessary to conduct the studies; I thank them for this support. I also thank the members of my committee for their thoughtful comments on the thesis. Finally, I am grateful to my family, Doug, Soren and Matthew Halverson, for their moral support and perseverance over the past years. The research was conducted while I was employed under an Agri-Food Regional Development Subsidiary Agreement grant (Project No. 11004) awarded to Mr. Maurer, Agriculture Canada Research Station, Agassiz, B.C. It was also supported by a Natural Science and Engineering Research Council operating grant awarded to Dr. Jolliffe. -1 -C H A P T E R I I N T R O D U C T I O N The relationship between plant population density and crop yield can be described using mathematical models. These models enable the determination of optimum crop population densities for maximum agricultural yields, and they permit the establishment of economic thresholds for crop management. Statistically and biologically valid equations can explain yield variation in response to competitive interference among neighbouring plants. Furthermore, the models can provide a reference for the analysis of interspecific plant competition (Jolliffe, 1988). The pursuit of optimum horticultural yields through the manipulation of the plant environment has led to investigations of the use of agricultural plastics in high plant populations. The use of agricultural plastics to protect field crops and to promote plant growth and development, is increasing in North America. Their use enables the commercial production of many warm-season crops in locations that are otherwise climatically marginal for their culture. For example, studies conducted in the Fraser Valley region of British Columbia have shown that row covers, applied over bell peppers (Capsicum annuum L ) , promote early fruit development and enhance total fruit yield (Maurer and Frey, 1988). Most studies involving agricultural plastics, however, have been concerned primarily with horticultural yield; few studies have considered the progress of growth of the - 2 -whole plant under a protective covering. Similarly, many studies have reported greater productivity from high plant populations densities of bell peppers (Ahmed, 1984; Batal and Smittle, 1981; Metwally et al., 1982; Porter and Etzel, 1982; Stoffella et al., 1984), but few have attempted to fit mathematical models. The determinants of plant growth and development can be investigated through several methodologies. These techniques, applied at the organismal or sub-organismal levels, provide insight into the mechanisms of response to environmental conditions. Plant growth analysis indicates how plant presence, growth rate, persistence of growth, and partitioning of growth contribute to growth and productivity (Jolliffe et al., 1982). Yield component analysis (YCA) can quantify yield variation and identify sources of variation through morphological yield components (Fraser and Eaton, 1983). 'Allometry' connotes interrelationships among different features of an organism. Allometric relationships may change in response to growth and treatments (Jolliffe et al., 1988), and can be linked to plant growth analysis (Jolliffe and Courtney, 1984). Examination of allometric parameters can provide further insight into the relationships among different component parts and the effects of treatments upon these relationships. This study was undertaken to examine the growth and productivity of bell peppers as affected by row covers and plant population density. Mathematical models defining fruit yield-plant population relationships will be developed to express crop response to intraspecific interference. The contribution of yield - 3 -components to yield responses will be detailed using two YCA models. Yield variation in each model will be studied, either independent of time (i.e. growth) or incorporating time, utilizing a two-dimensional partitioning technique. Finally, fruit and whole plant allometry will be interpreted using models reflecting treatment effects as well as plant growth. - 4 -C H A P T E R II L I T E R A T U R E R E V I E W 2.1 History, production and culture of peppers Peppers are a prehistoric fruit of Central and South America (Heiser, Jr. and Smith, 1953). Their cultivation spread widely throughout the world after the European discovery of America. Cultivation in the United States began before 1828. The first cultivated plants, however, were brought from Europe rather than from Central or South America (Heiser, Jr. and Smith, 1953). In 1980 world production of peppers was 6.9 million tonnes (Somos, 1984). The major production areas are Asia, Europe, Africa, North America and South America. In 1980, these areas respectively accounted for 40.8, 31.3, 15.5, 9.9 and 2.5 per cent of the world's production (Somos, 1984). In the United States peppers are produced in Florida, Michigan, New Jersey, North Carolina, Texas and California (Hartmann et al., 1981). Peppers are a minor crop in Canada, and market demands are met by imports from the United States and Mexico. Peppers are a warm-season crop with a high heat-unit requirement (O'Sullivan and Muehmer, 1978). In cooler climates, they are usually transplanted as two month old seedlings (Erwin, 1932). The plants grow best on - 5 -a well-drained, warm, moderately fertile soil. Apart from control of pests, peppers require little attention once established. 2.2 Morphological characteristics of bell peppers 2.2.1 Nomenclature Botanists have been concerned with the classification of peppers since the 16th century (Somos, 1984). In Species plantarum (1753), Linnaeus described two species of peppers, Capsicum annuum and C. frutescens. More recently, domesticated Capsicum has been classified into four species: C. annuum L., C. pubescens R. et P., C. frutescens L. and C. pendulum Willd (Somos, 1984). Botanical traits that distinquish C. annuum L. include long-petiolated leaves; single standing flowers; white or purple-violet spotted corolla; acuminate calyx dentation; cylindrical, greyish violet anthers with a slight incision at the base; and light yellow, flat and slightly curved seeds (Somos, 1984). Attempts at further systematization of C. annuum have been based on the following traits: the position, size, locule number, colour, taste and shape of the fruit (Somos, 1984). Green peppers with bell-shaped fruit, var. latum Erw., were classified by Erwin (1932). The morphology of the pepper plant has been described by Somos (1984). - 6 -2.2.2 Root system The root system of the pepper consists of a tap root with evenly distributed and developed lateral roots. Lateral root production is increased if the main root is damaged, as can occur during transplanting. The root system is located close to the soil surface and accounts for approximately 10 percent of the total dry weight of a mature plant. The contribution to the total weight of a juvenile plant is less (Somos, 1984). 2.2.3 Shoot system The development of the shoot system of the pepper is cymose. Longitudinal growth terminates in a flower, and subsequent shoot development arises from two, and occasionally three, axils under the flower. Kormos and Kormos (1956) described unrestricted and partly restricted shoot systems. The main axis of an unrestricted shoot system exhibits unlimited growth, and side-shoots are lacking or, at most, poorly developed. A partly restricted shoot system is indicative of plants in which the growth of the main axis quickly ceases, and is overtaken by the growth of the side-shoots. The crossing of plants with these characteristics produces new types. Somos (1984) described the following types which are common to bell peppers: a) unrestricted growth of main and lateral shoots (Fig. 1a); b) unrestricted growth of main shoot; restricted lateral shoot growth (Fig. 1b-c); - 7 -Fig. 1. Shoot systems of bell peppers: a) unrestricted growth of main and lateral shoots, b and c) unrestricted growth of main shoot, restricted lateral shoot growth and; d) initial unrestricted growth of main and lateral shoots, restricted growth of both with later development (Somos, 1984). - 8 -c) initial unrestricted growth of main and lateral shoots; restricted growth of both with later development (Fig. 1d); Seven nodes develop on the main axis of the bell pepper. The weight of the shoot and the ratio of leaf weight to stem weight, have been reported to differ in response to cultivar (Somos, 1984). 2.2.4 Foliage The leaf of the pepper is simple, entire in shape and petiolate. Leaf number, area and weight differ among cultivars (Somos, 1984). 2.2.5 Flower The longitudinal growth of the branches of the main shoot and the side-shoots ends with a flower; two or more flowers may develop on the apex of the main axis, and one flower on the apex of lateral branches. The flowers are attached to the shoot by a peduncle, and are bisexual. The relative position of the stigma and the stamens determine the extent of self or cross-fertilization. Self-fertilization, common with large fruit or bell pepper cultivars, occurs when the style is equal in height or lower than the anthers (Kiss, 1969). A plant may produce more than 100 flowers; however, fruit set is inversely related to the number of developing fruit present on the plant (Somos, 1984). Kato and Tanaka (1971) reported that fruits set initially on the main shoot - 9 -and then on the lateral branches. Fruit set on the main shoot was 80 per cent and on the side-shoots, 30 per cent. 2.2.6 Fruit The fruit of the pepper is a berry, and, as it is hollow inside, can be called an inflated berry. It is attached to the stem of the plant by the peduncle which continues inside the fruit as the central placenta. The shape of the peduncle varies among and within cultivars, and determines the position of the fruit: upright, lateral (semi-upright) or pendant (drooping). The shape of the placenta follows that of the fruit: the placentas of bell peppers ('California Wonder' type) are a short conic shape. The number of carpels range from two to five; the larger number of carpels is generally characteristic of bell peppers. The average locule number is thought to be quantitatively inherited (McArdle and Bouwkamp, 1983). Mohd Khir et al. (1987), however, reported temperature effects on locule number. The pericarp varies between 0.9 and 6 mm in thickness (measured in the median line of the fruit between the septae). A thick, fleshy pericarp is a preferred market characteristic for bell peppers. Genetic studies of the fruit apex indicate the presence of a gene controlling pointed versus blunt fruit (McArdle and Bouwkamp, 1983). They suggested that bell peppers have a gene which influences, possibly through an environmental interaction, the gene controlling fruit apex, to produce an inverted - 10 -apex. The fruit base, which develops from the flower receptacle, may be convex, plane, or concave in shape. The inconsistency of results from genetic studies suggests that this trait may also be affected by the environment (McArdle and Bouwkamp, 1983). Fruit width is considered to be quantitatively inherited. Fruit length, however, is most likely influenced by both genetic and environmental factors (McArdle and Bouwkamp, 1983). Temperature effects on fruit length were noted by Rylski (1973). Sinnott (1935), investigating fruit-shape genetics, proposed the existence of a gene responsible for fruit shape. McArdle and Bouwkamp (1983) introduced a shape index (length:width ratio), but failed to substantiate a gene controlling shape. Shifriss et al. (1989) found a small correlation between the fruit shape index and fruit weight, and suggested fruit weight could be selected for, independently of fruit shape. The contribution of fruit dry weight to the total plant dry weight (i.e. harvest index) has been observed to range, among different cultivars, from 22 to 45 percent (Cselotei, 1957; Somos and Sovany, 1964). 2.2.7 Seed Seed of the bell pepper develop primarily on the surface of the placenta, but may be located along the septa. The seed is kidney-shaped, yellow in colour, and weighs from 5 to 7 mg. -11 -The size and development of fruit is influenced by seed number (Dempsey and Boynton, 1965). McArdle (1984) reported a positive correlation of total fruit seed weight with fruit size, fruit length and length:width ratio. Studies concerning bell peppers have found similar correlations between seed number and fruit size (Rylski, 1973). 2.3 Environmental factors affecting bell pepper development 2.3.1 Light The effect of light on the growth and development of peppers is, to a certain extent, dependant on the pepper cultivar. Results obtained from different researchers vary, and are thus difficult to generalize. Plant development is commonly influenced more by photoperiod than by light intensity; the rate of development increases with increasing light (Somos, 1984). Shading (30 to 70 per cent) was observed to reduce leaf area, leaf weight and flower formation (Quagliotti et al., 1974). Polyethylene film (0.2 mm thickness) has been shown to reduce natural light intensity by 25 percent (Gulyas et al., 1970). The film had a positive effect on radiation intensity, however, and fruit yield was greater than from the uncovered control. - 12 -2.3.2 Temperature The main stages of pepper development have been identified as seed dormancy, germination, cotyledonous, vegetative (stem and leaf formation), and generative (fruit formation) (Somos, 1984). The heat requirement for optimum development differs for each developmental stage. The following discussion is concerned with the latter two stages. The optimum temperature (diurnal) for the vegetative phase of pepper development has been reported as 25C in sunny weather and 18-20C in cloudy weather (Somos, 1984). Most reports, however, differentiate between day and night temperatures. The nocturnal heat demand of juvenile plants decreases as the plants mature; optimal growth of juvenile plants occurs at night temperatures of 30C and plants aged 15 weeks, at 15C (Dorland and Went, 1947). Low temperatures (18C day and 15C night) can result in abnormal stamen and gynoecium development (Polowick and Sawhney, 1985). Flowers developing at low temperatures (pre-anthesis) have an elongated stigma which can restrict self-pollination (Rylski, 1973). Furthermore, temperatures less than 16C inhibit flower opening (Ghaleb, 1967). Day temperatures of 25C, and night temperatures of 16 to 21C promote flower opening (Wells, 1967). Fruit set was greatest at temperatures from 16 to 21C whereas temperatures above 32C inhibited fruit set (Cochran, 1936). Excessive temperatures (50C) have been observed to reduce fruit set (Kohm et al., 1983). Optimum yields were achieved at day temperatures of 25C and night - 13 -temperatures of 20C when soil temperatures were maintained at 25C (Filius, 1968). Parthenocarpic fruit have been induced with low post-anthesis temperatures (10C) (Charles et al., 1979; Rylski, 1973; Rylski and Halevy, 1974). Parthenocarpic fruit of Capsicum annuum L. var. longum DC. was reported to be larger than fruit with seeds (Charles et al., 1979), however, parthenocarpic fruit of Capsicum annuum L. var. latum Erw. was smaller than seeded fruit (Rylski, 1973). 2.3.3 Relative humidity Low relative humidity has been reported to suppress flower production and fruit set (Cochran, 1936). High (95%) relative humidity resulted in flower drop but improved seed set (Baer and Smeets, 1978). Seed set and number were improved by high day humidity whereas fruit numbers increased with low humidity by night (Bakker, 1989a). Also, humidity did not affect fruit shape (length:width ratio). Other studies have reported that early and total pepper yields were not affected by humidity (Bakker, 1989b). - 14 -2.4 Cultural influences on development and yield of bell peppers 2.4.1 Protected cultivation The effects of soil mulches on soil moisture conservation, reduction of fertilizer leaching, enhancement of soil temperatures and promotion of crop growth are well documented (Gerald and Chambers, 1967; Porter and Etzel, 1982). Polyethylene row tunnels have also been shown to enhance crop earliness and yield through their effect on the crop microclimate (Pratt et al., 1981; Wolf et al., 1986). Ventilated polyethylene row tunnels have elevated air and soil temperatures, but depressed relative humidity (Maurer and Frey, 1987). Plant exposure to continuous high temperatures (above 30C) or to short durations of temperatures from 40 to 50C can have a deleterious effect on pollination and fruit set (Cochran, 1936; Kohm et al., 1983 and Mohd Khir et al. 1987). Early yields were negatively correlated with maximum air temperatures (as high as 50C) under tunnels, and initial fruit set improved with increased ventilation and opaque tunnel film (Kohm et al., 1983). Enhanced early and total yields of bell peppers grown under ventilated polyethylene row tunnels have been reported, although plants were exposed to short durations of temperatures from 35 to 40C (Maurer and Frey, 1987). Elevated temperatures under polyethylene tunnels have been reported to cause an increased number of locules on bell pepper fruit (Mohd Khir et al., 1987). - 15 -2.4.2 Plant populations Bell pepper development and yield can be significantly influenced by plant population density and arrangement. Marketable yield increased as plant population increased from 2.7 plants m"2 to 4.0 or 6.0 plants m"2 (Batal and Smittle, 1981). Plants were spaced 28 or 41 cm apart in two or three rows per bed. Marketable yield increased with decreasing within-row spacing from 45 cm to 13 cm. (one to three plants per hill) (Stoffella et al., 1984). Also, increasing plant populations of 3.9 plants m"2 to 10.7 plants m"2, spaced 30 to 10 cm apart within-row, resulted in increased marketable yields (Everett and Subramanya, 1984). Plant population density did not affect plant growth characteristics (plant height, branch number, leaf number or flower number) in a study of within-row spacings ranging from 30 to 75 cm (Ahmed, 1984). Marketable yield, however, increased with decreasing plant spacing. Plant height increased and stem diameter decreased with higher plant populations (13 to 51 cm within-row spacings) (Stoffella and Bryan, 1988). Root and stem dry weights were greater at lower population densities. Primary and secondary branch numbers, however, were not influenced by plant population. In another study, dry matter production of leaves, stems and fruit also increased with decreasing plant populations (Srinivas and Hegde, 1984). An increase in leaf area at lower populations was also reported. - 16 -2.5 Analytical techniques for interpreting plant development and yield 2.5.1 Yield-plant population relationships The quantitative relationship between plant population density and crop yield has been described using mathematical models (Jolliffe, 1988; Willey and Heath, 1969). Models that are statistically and biologically valid enable the determination of optimum densities for maximum yields. From this, economic thresholds for crop management can be established. Furthermore, monoculture yield-density models can provide a reference for the analysis of interspecific plant competition (Jolliffe, 1988). Mead (1979), in a review of response models, suggested that three principal equations describing reciprocal relationships are most commonly used. These are: 1. A reciprocal relationship derived by Shinozaki and Kira (1956) and Holliday (1960): W 1 = a + pp (1) where w = yield per plant, p = plant population density and a and p are parameters. 2. Holliday's (1960) reciprocal equation generalized to the family of inverse polynomials by Nelder (1966): - 17 -w"1 = a + pp + TP 2 (2) where a, p and y are parameters. 3. Bleasdale's (1967) 'simplified' version: w e = a + pp (3) where 8 = some positive quantity, usually less than unity. The quantitative response of crop yield per unit land area to increasing plant population density has been described by Holliday (1960) as 'asymptotic' or 'parabolic'. An asymptotic response is characterized by an increasing yield relationship until a limit of relatively constant yield is approached. A parabolic response exists where further increases in population density ultimately result in a loss of yield per unit land area. Holliday (1960) suggested the former response was typical where yield was a function of vegetative growth, and the latter occurred where it was a function of reproductive growth. Equation 1, above, was used for asymptotic relationships, while equations 2 and 3, for parabolic relations. The model of Bleasdale and Nelder (1960) was based on the assumption of an allometric relationship between the weight of a portion of the plant and the weight of the whole, which is represented by 9. Bleasdale (1967) - 18 -described 6 as a constant which caused yield per land area to be asymptotically related to plant population density when 9 = 1 and parabolically related when 8 < 1. Jolliffe et al. (1988) warned against the treatment of the allometric exponent as a constant, as it varied among yield variates, species and stages of growth. They suggested that yield-density relationships should only be combined with the allometric power function if the two equations account for a similar variation in yield, or if either predict the yield without error. Either situation, however, is uncommon. Yield-plant population density relationships have usually been concerned with total plant yield. Yields graded for marketable qualities have not been regarded as biological forms of yield, and hence attempts to fit a yield-plant population density relationships involving graded yields, have generally been avoided (Chapman, 1980). Salter et al. (1980, 1984) used inverse polynomial equations to fit a relationship between the yield of one grade of a crop and plant population density. A model fitting cumulative graded yields was proposed by Chapman (1980); however, it was not biologically meaningful at all densities. Willey and Heath (1969) suggested graded yield relationships may have to remain empirical. - 19 -2.5.2 Yield component analysis Yield component analysis determines the contribution of yield components to variation in yield, and assesses the effects of treatments on yield components. Sequential yield component analysis (SYCA) partitions total variation in yield into components, and analyzes these in a stepwise multiple regression procedure. Yield components are entered into the regression in an assumed chronological order. The method determines the contribution of a yield component to the total sum of squares for yield, both directly and through its relationship with other yield components. This regression procedure has been applied to studies of cranberries (Eaton and Kyte, 1978; Shawa et al., 1981), white clover (Huxley et al., 1979), and beans (Herath and Eaton, 1981). Sequential yield component analysis was advanced when components were also entered into the multiple regression in a reverse chronological order. This technique, 'backward SYCA', combined with 'forward SYCA', enabled a more thorough interpretation of the relationship of yield components to yield, and it has been used in the study of several plant species (Anderson et al., 1986; Bowen, 1983; Bowen and Eaton, 1983; Eaton, Shawa and Bowen, 1983; Lovett Doust et al., 1983; Neilson and Eaton, 1983). A further extension of SYCA resulted with the development of two-dimensional partitioning (TDP). This procedure combined ANOVA and SYCA and enabled a concise description of the sources of yield, and yield component, variability. The method was initially presented in a study of cucumbers (Eaton et - 20 -al., 1986). TDP has since been applied to many berry crops (Baumann and Eaton, 1986; Freeman et al., 1989; McArthur and Eaton, 1988a; McArthur and Eaton, 1988b; Strik and Proctor, 1988) and to tree fruits (Eaton, 1987.). Yield components were entered into the regressions in a forward and a backward order in a study of forage maize (Jolliffe et al., 1990). 2.5.3 Allometric relationships Allometry refers to the "quantitative relationships which exist among different features of an organism as growth proceeds" (Jolliffe et al., 1988) or more simply, "the study of size and its consequences" (Gould, 1966). A simple bivariate power function, brought into general use by Huxley (1932), relates growth of two features (y and z) of an organism: y = az p (4) where parameters a and p represent the allometric coefficient and exponent respectively. The allometric exponent expresses the ratio of relative growth rates of plant characteristics and has often been treated as a constant. Recent studies, however, have shown that the allometric exponent and coefficient are subject to plant growth and treatment effects (Perkkio, 1985; Jolliffe, 1988; Stanhill, 1977,a,b). Their examination can provide further insight into the relationships - 21 -between plant components and the effect of treatments on these relationships. Allometric relationships have been linked to plant growth analysis (Jolliffe and Courtney, 1984). Jolliffe et al. (1988) used a best subset multiple regression procedure to interpret allometric responses to treatments. Morris and Myerscough (1987) demonstrated the effect of plant allometry on plant interference, using Bleasdale's (1966) yield-density model. - 22 -C H A P T E R III M A T E R I A L S A N D M E T H O D S 3.1 Experimental design The experiment was conducted during 1988 at the Agriculture Canada Research Station, Agassiz, British Columbia. Seeds of Capsicum annuum L. cv. 'Ace Hybrid' (Stokes Seeds Ltd., St. Catherines, Ont.) were sown on 25 March in polyethylene flats with 5.5 x 5.5 x 7.0 cm cells containing soil block media (Appendix I), and grown in the greenhouse. On 28 April, raised beds (0.23 m high and 1.1 m wide) were formed on 1.8 m centres by disc ridging and shaping a silt loam soil (Rego Humic Gleysol, pH 6.2). The soil had been prepared with a broadcast incorporated application of 165 N, 165 P 20 5, 165 K 20 and 4.5 B (kg ha'1 respectively) and with 23 t ha"1 cattle manure. A black polyethylene surface mulch, 1.1 mil and 1.5 m in width, was applied to the beds. The plants were transplanted into twin rows (0.5 m between-row spacing) on the beds on 12 May. The experimental design was a split-plot with four blocks. Main plots were uncovered or were covered with polyethylene tunnels. Five population densities, 1.39, 1.85, 2.78, 5.56 and 11.1 plant m"2, formed the subplots. Each subplot was divided into two sections: a 10-plant section for yield measurements and a 26-plant section for plant growth measurements. (A third 10-plant section for demographic measurements, not reported here, was incorporated into density - 23 -treatments of 2.78 plants m"2.) The location of each section was randomized within each plot. Four plants were sampled from each treatment for yield-density measurements, and 12 plants were sampled for plant growth measurements. The remaining plants in each section were guards. Subplot length and plant spacing are listed in Table 1. Row covers, hoop-supported polyethylene tunnels, were applied to the appropriate treatments immediately following transplanting. The tunnels were vented with two rows of slits, each located 15 cm from the top of the tunnel and running parallel to the length of the tunnel. Slits, 15 cm long and 2 cm apart, were cut perpendicular to the length of the tunnel. Row covers were removed on 7 July. Fruit were harvested twice a week, at the mature green stage from 3 August to 21 September. The experiment was overhead-irrigated on 12 August. Insecticides were applied as necessary throughout the season, and included 0,0-diethyl-0-(2-isopropyl-6-methyl-5-pyrimidinyl) phosphorothioate (diazinon); 5,6-dimethyl-2-dimethylamino-4-pyrimidinyl-dimethylcarbamate (Pirimor), and Bacillus thuringiensis Berliner (Dipel). Air and soil temperatures, and atmospheric relative humidity, were monitored for the duration of the experiment with a datalogging microprocessor (21X, Campbell Scientific, Logan, Utah). Thermistor probes (107 (air) and - 24 -Table 1. Experimental details. Treatments Row cover Density Within row Plant number Subplot (Main plot) (Subplot) spacing per subplot length (plants m"2) (m) (m) No row cover Row cover 11.1 0.1 30 1.5 5.56 0.2 30 3.0 2.78 0.4 38 7.6 1.85 0.6 30 9.0 1.39 0.8 30 12.0 11.1 0.1 30 1.5 5.56 0.2 30 3.0 2.78 0.4 38 7.6 1.85 0.6 30 9.0 1.39 0.8 30 12.0 - 25 -107B (soil), Campbell Scientific, Logan, Utah) were placed 22 cm above the soil surface and at 10 cm depth in the soil. Relative humidity probes (207, Campbell Scientific, Logan, Utah) were located 20 cm above the soil surface. Probes were located in the centre of each main plot with the exception of probes measuring air temperatures in row cover treatments. In this case, probes were also located at one end of the tunnel and at a point midway to the centre of the tunnel. Daily minimum, maximum and mean temperatures and relative humidity, and also mean hourly temperature and relative humidity, were computed from readings made every ten minutes. The Julian date has been used in all figures. A Julian day refers to the number of days that have elapsed since 1 January 4713 B.C., commencing at 12 hours Universal Time (UT) (Bishop, 1988). The Julian date for 0 UT 0 January 1988 is approximately 2,447,160.5 (Bishop, 1988). In this thesis, the Julian date refers to the number of days that have elapsed since 31 December 1987 (eg. 1 January 1988 is Julian day 1). 3.1.1 Yield-plant population density study At the time of fruit harvest, the harvest date and fruit location (i.e. node) were recorded. The fruit were graded into the following categories: marketable (80 g or greater and free from disease and insect damage), undersized (<80 g), and cull (diseased or insect damaged). The fresh weight of - 26 -all fruit was recorded. The width, length and lobe number of marketable fruit were also recorded. A destructive harvest was conducted on two plants from each treatment on 27 and 28 September, after the final mature fruit harvest. Plants were removed from the soil and the roots gently washed to remove soil particles. The number of immature fruit, total fruit fresh weight, total leaf area and the total dry weights of leaves, stem (including petioles) and roots, were recorded. Leaf area was determined using a LI-3000 area meter (LICOR Ltd., Lincoln, Neb.). Dry weights were obtained after plant components were oven-dried to constant weight at 70C. Fruit dry weight was estimated from the calculated ratio of dry:fresh weight of fruit harvested in the plant growth study. This ratio, 0.0677, 0.0844 and 0.0706 for fruit graded marketable, undersize or cull respectively, did not significantly differ between treatments within each grade. 3.1.2 Plant growth study A destructive harvest of two plants from each treatment was conducted six times over the duration of the experiment: 9 June, 27 June, 7 July, 27 July, 23 August and 21 September. At each harvest, plants were severed at the soil surface and removed to the laboratory for analysis. The following data were recorded: plant height, leaf area, leaf dry weight, shoot dry weight (excluding fruit). The number of nodes, buds, flowers and fruit and the weight (fresh and dry) of fruit were recorded separately for side shoots and for branches originating - 27 -from the first main node. Leaf area and component dry weights were measured as outlined above. Fruit reaching the mature-green stage were harvested from each plant up to the date of the destructive harvest. Harvest date, fruit location (side shoot or main axis), weight (fresh and dry), width, length and lobe number of marketable fruit were recorded. These fruit were also sectioned longitudinally and each section photocopied. The photocopies were used to determine the inversion depth of the apex and base as well as the area and perimeter of the longitudinal section. A LI-3000 area meter was used to determine area and a map-measurer (Minerva, Switzerland), perimeter. The ratio of area to the square of the perimeter was calculated for use as a shape indicator. 3.2 Data analysis Functional plant growth analysis from the plant growth study and results from the demographic study will be presented elsewhere, as their inclusion is beyond the scope of this document. All measured and derived variables were subjected to a preliminary analysis of variance [GLM procedure of Statistical Analysis System (SAS), SAS Institute, 1985]. The sums of squares were partitioned using orthogonal contrasts. - 28 -3.2.1 Yield-plant population density models Yield-plant population density response was defined using Bleasdale's (1967) model: y 9 = a + pX (5) where y=mean yield per plant, X=plant population. Alpha is an index of plant yield in the absence of competition, and p is an index of the responsiveness of a plant to population density changes. Theta is thought to be related to the utilization of environmental resources in the space accessible to a plant (Watkinson, 1984). To estimate 0, a and p, total plant weights and weights of plant components were separately regressed against plant population density in a nonlinear regression procedure developed using the BMDPAR program (Dixon, 1985). 3.2.2 Yield component analysis Two models were used to evaluate sources of variation in successive morphological components contributing to reproductive and total plant yield per plant and per land unit area: - 29 -y = NN x LA/NN x WL/LA x W7WL x WF/W = WF (6) y = NN x LA/NN x WL/LA x WV/WL x WF/WV x W/WF = W (7) where y is fruit (WF) or shoot (W) dry weight. Yield components included leaf area per node (LA/NN); leaf dry weight per leaf area (WL/LA); vegetative shoot or total shoot dry weight per leaf weight (WV/WL, W/WL); fruit dry weight per vegetative shoot or total shoot dry weight (WF/WV, WF/W); and total shoot dry weight per fruit dry weight (W/WF). A two-dimensional partitioning procedure, described by Eaton et al. (1986), was used to analyze the yield components. Yield components were transformed to natural logarithms and analyzed in the order of their presumed chronological sequence, in a stepwise multiple regression procedure ('forward'). This order was reversed in a 'backward' yield component analysis. In each case, ln(y) was regressed directly against the first component entering the regression while subsequent components were calculated as residuals from a multiple regression on the previously entered components. This stage formed the first dimension of the two-dimensional partitioning. The second dimension resulted from the analysis of variance, used to partition the sources of variation for each component. Data from the six destructive harvests were analyzed individually and also combined, using harvest date as an additional source of variation in the analysis of variance. - 30 -Mean values and standard deviations of yield components for separate treatments, were In-transformed and fitted to growth curves using a cubic spline regression procedure described by Jolliffe and Courtney (1984). 3.2.3 Allometric evaluation Fruit morphology. Potential effects on fruit fresh weight, of time (t), population density (D) and cover (C) were evaluated using the following linear model (Jolliffe etal., 1988): ln(yHn(a)+p0ln(z)+p1tln(z)+p2Dln(z)+^ +p6CDIn(z)+p7tCDIn(z)+Y1ln(t)+Y2ln(D)+Y3ln(C)+Y4ln(tD)+Y5ln(tC) +Y6ln(CD)+Y7ln(tCD)+ln(e) (8) In each case, y was marketable fruit fresh weight and z, a characteristic of fruit morphology (e.g. length, width). Harvest date and fruit node location were used in separate regressions as indices of time. Allometric relationships were evaluated using a best subset multiple regression procedure, developed using the BMDP9R program (Dixon, 1985). This procedure has been described in a previous study (Jolliffe et al., 1988). Linear regressions using least squares were determined in stages, each stage representing the incorporation of an additional independent variable, one of the terms following In(oc') on the right hand side of equation (8), into the model. Mallow's CP (Daniel and Wood, 1971) was calculated at each stage. The - 31 -variable incorporated into the model at each stage was the one which led to the lowest interim Mallow's CP. The eventual best subset multiple regression had the lowest CP value. Plant morphology. Allometric relationships with total dry weight per plant in the plant population density study were also determined using the BMDP9R program. Two factors, population density (D) and cover (C), were considered and were expressed by: ln(y)=ln(a)+p0ln(z)+31Dln(z)+p2Cln(z)+p3CDIn(z)+Y1ln(D)+y2ln(C) +ln(e) (9) Here, the variable y was total plant weight (dry), and a measure of some plant component was represented by z. Equation (8) was also used to express potential treatment effects on allometric relationships in the plant growth study. In this study, y was total shoot weight (dry); z, plant components; and t, date of plant harvest. - 32 -C H A P T E R IV R E S U L T S In this section, microclimatic data will be considered, followed by ANOVA results, nonlinear regression models, and spline regression models concerning the yield-plant population density study and plant growth study. Results from the yield component analysis will be then be presented and finally, allometric relationships will be defined. 4.1 Plant microclimate 4.1.1 Soil temperature Row covers significantly elevated mean, maximum and minimum soil temperatures 1.9C, 2.4C and 1.5C, respectively, above uncovered treatments. Temperature differential between the row cover treatments decreased as the season advanced (Fig. 2). - 33 -Fig 2. Mean daily soil temperatures as affected by row covers: J D 137 to 188 (16 May to 6 July). Measured at 10 cm depth. Treatments are: uncovered, and row covers. - 34 -4.1.2 Air temperature Air temperatures recorded at three different locations under the row covers did not differ significantly (data not presented) and the mean of these data were compared with ambient air temperatures. Row covers significantly elevated mean and maximum air temperatures, over the period of covering, compared with ambient temperatures (Fig. 3). Minimum temperature was significantly higher under the row covers; however, temperature differential was less than 1C. Diurnal temperature fluctuations on the day with the highest recorded maximum temperature, 14 June, is illustrated in Fig. 4. On this day, temperatures under the row covers reached 47C, 12C above ambient. Maximum air temperature under the row covers was greater than 40C on 19 of the 56 days the plants were covered; during this period ambient temperatures were never above 37C (Fig. 3). Mean daily temperatures were not above 27C during the week prior to cover removal (Fig. 5). Ambient temperatures over the entire season are presented in Fig. 6. 4.1.3 Relative humidity On warm days, relative humidity was higher under row covers from 2400 to 0600 hours, but dropped below that of uncovered treatments until 1700 hours (Fig. 7a). The daily mean relative humidity, however, did not differ between treatments, when averaged over the covering period (Fig. 7b). TEMPERATURE, C 0 I i 1 1 I 1 1 1 130 160 170 100 130 160 170 190 JULIAN DAY Fig. 3. Mean (a), minimum and maximum (b) air temperatures as affected by row covers: JD 134 to 188 (13 May to 6 July). Measured at 22 cm above soil surface. Treatments are: uncovered, and row covers. - 36 -T E M P E R A T U R E , C 50 -1 4 0 -3 0 -2 0 -1 0 -0 H 1 1 1 1 1 1 0 4 8 12 16 2 0 2 4 H O U R Fig. 4. Diurnal air temperature fluctuation as affected by row covers: JD 166 (14 June). Measured at 22 cm above soil surface. Treatments are: uncovered, and row covers. - 37 -T E M P E R A T U R E , C 3 0 n 2 6 -2 0 -1 6 -1 0 -6 -0 H 1 1 1 1 1 1 0 4 8 12 16 2 0 2 4 H O U R Fig. 5. Mean diurnal air temperature fluctuation prior to row cover removal: JD 182 to 186 (30 June to 4 July). Measured at 22 cm above soil surface. Treatments are: uncovered, and • row covers. - 38 -T E M P E R A T U R E , C 3 0 i OH 1 1 1 1 1 125 150 176 2 0 0 2 2 6 2 5 0 J U L I A N D A Y Fig. 6. Mean daily air temperatures: JD 134 to 250 (13 May to 6 September). Measured at 22 cm above soil surface. 16 20 24 130 H O U R Fig. 7. Per cent relative humidity as affected by row covers: (a) JD 166 (14 June) and (b) daily mean, JD 137 to 188 (16 May to 6 July). Measured 20 cm above soil surface. Treatments are: 150 170 J U L I A N DAY 100 uncovered, and row covers. - 40 -Relative humidity from 12 May to 7 July ranged from 32% to 99%; mean relative humidity was 70%. 4.2 Yield-plant population density relationships 4.2.1 Effects of row cover and plant population density on fruit  morphology: ANOVA results Fruit length and two dimensional area were not affected by the treatments (Table 2). Row cover treatments accounted for a greater variation in the characteristics of fruit morphology than population density, where it was significant. Plant population density effects were linear, with the exception of a quadratic response for apex depth. Treatment means generally decreased with increasing population density. Interactions between row cover and plant population density were significant for fruit width. Significant treatment means are presented in Appendix II. The statistically significant differences determined between treatment means were usually quantitatively small. 4.2.2 Effects of row cover and population density on fruit yield:  ANOVA results Fruit yield per plant. ANOVA results for early yield, and for total yield, are summarized in Tables 3 and 4 respectively. Treatment means for fruit yields are presented in Appendices III and IV. Row covers resulted in greater marketable Table 2. A N O V A F-statistics and EMS: fruit morphology. Source df Lobe Width Length W:L Base Apex Area y Perimeter P 2 :A number depth depth (A, m 2, (P) (mm) (mm) (mm) (mm) x10"3) (mm) Block 3 0.60 129** 1.06 1.65 0.11 0.17 2.12 2.69 0.56 Cover (C) 1 20.0* 89.6** 0.81 3.27 32.6* 0.34 3.75 18.3* 12.58* Error (a) 3 3.27 0.100 3.12* 2.55 0.97 6.03*** 1.59 0.82 7.36*** Density (D) (4) 2.60 1.77 0.61 1.43 3.64* 3.85* 1.08 1.12 3.32* linear 1 4.39* 3.97 2.23 4.34* 5.37* 6.98* 1.19 1.11 0.00 quadratic 1 0.43 1.98 0.70 2.20 0.63 4.41* 1.07 0.64 8.39** cubic 1 2.82 0.63 0.01 0.03 1.82 1.81 0.99 1.39 0.09 deviation 1 0.13 0.05 0.00 0.01 2.90 0.22 0.21 0.05 0.99 C«D (4) 1.01 4.05* 0.57 0.21 1.10 0.49 0.90 1.03 0.14 C-D, 1 0.00 11.0** 0.92 0.30 1.37 0.03 1.94 2.63 0.20 Error (b) 24 1.91 2.64 4.42*** 3.28*** 1.40 1.87** 3.12*** 3.52*** 2.64*** Error d f 2150 2159 2159 2159 2160 2145 2160 2159 2159 EMS 0.485 51.8 195 0.0195 18.2 71.2 9.8x10"7 993 11.3 y Area and perimeter of a longitudinally sectioned fruit. z degrees of freedom for respective variates. Asterisks denote significance at p=0.05(*), 0.01 (**) and 0.001 (***). E M S denotes error mean square. - 42 -Table 3. ANOVA F-statistics and E M S : early fruit fresh yield per plant (kg).y Source df Marketable 2 Undersized Cull Total Block 3 0.65 0.40 0.19 0.66 Cover (C) 1 57.0** 44.5** 8.76 87.03** Error (a) 3 1.20 1.04 1.61 1.41 Density (D) (4) 26.9*** 2.86* 5.76** 35.4*** linear (L) 1 105*** 7.13* 20.6*** 139*** quadratic 1 1.87 1.83 0.07 2.32 cubic 1 0.18 1.69 1.75 0.08 deviation 1 0.45 0.78 0.61 0.47 C»D (4) 6.36*** 0.80 1.58 7.91** C - D L Error (b) 22.3*** 0.72 5.25* 29.3*** 24 1.37 0.93 1.03 1.38 Error 120 E M S 0.151 0.0117 0.210 0.164 y Early yield is yield from the initial 28 days of harvest. z Marketable fruit > 80g; undersized fruit < 80g and cull fruit, diseased or insect damaged. Asterisks denote values that are significant at p=0.05(*), 0.01 (**) and 0.001 (***). EMS denotes error mean square. - 43 -Table 4. A N O V A F-statistics and E M S : total fruit fresh yield per plant (kg). Source df Marketable 2 Undersized Cull Total Block 3 0.73 0.44 0.75 0.96 Cover (C) 1 5.67 14.5* 5.32 33.1* Error (a) 3 2.27 2.89* 0.84 1.23 Density (D) (4) 111*** 1.66 9.95*** 156*** linear (L) 1 424*** 2.62 32.5*** 600*** quadratic 1 12.5** 2.66 0.0 19.9** cubic 1 5.15* 0.79 3.79 1.64 deviation 1 1.92 0.58 0.49 1.84 C»D (4) 3.33* 0.92 0.62 4.66* C«DL Error (b) 10.0** 0.06 1.42 14.9** 24 0.75 1.69* 0.74 0.72 Error 120 E M S 0.294 0.0245 0.344 0.275 z Marketable fruit > 80g; undersized fruit < 80g and cull fruit, diseased or insect damaged. Asterisks denote values that are significant at p=0.05(*), 0.01 (**) and 0.001 (***). E M S denotes error mean square. - 44 -and total early fruit yields compared with uncovered treatments (Table 3): marketable and total early yields were increased 91% and 103% respectively (Appendix III). Over the entire harvest period, however, row covers enhanced total yield per plant but not marketable yield (Table 4). Early yield (marketable and total fruit) per plant decreased in a linear fashion in response to increasing plant population density (Table 3). Yield response to increasing population density over the entire harvest period tended to decline at a faster rate at high population densities than low (significant quadratic component) (Table 4). The decrease in marketable and total yield resulting from increasing plant population densities was greater from covered than uncovered treatments, at both harvest intervals (significant interactions between row cover and the linear component of population density) (Tables 3 and 4). Fruit yield per land area. Tables 5 and 6 summarize ANOVA results for early and total yield per m 2 respectively. Treatment means are presented in Appendices V and VI. Row covers increased marketable and total early yields 87% and 98% respectively (Appendix V). Over the entire harvest period, however, row covers enhanced total yield but not marketable yield (Table 6). Total yield increased with increasing plant population density for both harvesting periods (Tables 5 and 6). Interactions between row cover and plant population - 45 -Table 5. ANOVA F-statistics and E M S : early fruit fresh yield per land area. y Source df Marketable 2 Undersized Cull Total Block 3 0.29 0.69 0.28 0.20 Cover (C) 1 28.1* 80.8** 3.38 36.0** Error (a) 3 1.90 0.42 1.70 2.18 Density (D) (4) 4.65** 8.44*** 3.89* 9.25* linear (L) 1 17.22** 31.6*** 12.3* 34.4**" quadratic 1 0.88 1.40 1.79 2.25 cubic 1 0.34 0.11 0.50 0.04 deviation 1 0.16 0.62 0.95 0.29 C-D (4) 2.35 4.45** 0.39 2.73 C«DL Error (b) 0.73 17.4*** 0.01 2.44 24 1.38 0.47 0.96 1.28 Error 120 E M S 1.61 0.230 0.377 2.34 y Early yield is yield from the initial 28 days of harvest (kg per m2). z Marketable fruit > 80g; undersized fruit < 80g and cull fruit, diseased or insect damaged. Asterisks denote values that are significant at p=0.05(*), 0.01 (**) and 0.001 (***). E M S is error mean square. - 46 -Table 6. ANOVA F-statistics and E M S : total fruit fresh yield per land area. Source df Marketable2 Undersized Cull Total Block 3 1.16 0.44 1.04 1.84 Cover (C) 1 1.75 10.0 3.44 14.2* Error (a) 3 2.91* 2.60** 0.82 1.47 Density (4) 12.13*** 6.94*** 5.11** 29.7*** linear (L) 41.9*** 25.7*** 15.6*** 108*** quadratic 1 2.93 1.54 1.19 7.40* cubic 1 3.75 0.04 1.99 3.33 deviation 1 0.00 0.51 1.63 0.01 C-D (4) 0.13 1.60 0.47 0.26 C - D L Error (b) 0.11 4.23 0.19 0.35 24 1.28 1.34 0.75 1.00 Error 120 E M S 4.30 0.540 0.513 4.47 2 Marketable fruit > 80g; undersized fruit < 80g and cull fruit, diseased insect damaged. Asterisks denote values that are significant at p=0.05(*), 0.01 (**) and 0.001 (***). EMS denotes error mean square. - 47 -density were generally not significant. The exception was a significant interaction between row cover and the linear component of population density, on undersized fruit (Table 6). 4.2.3 Effects of row cover and population density on fruit yield:  nonlinear regression models The curves shown in Figs. 8 to 11 represent the best fit of equation (5) for fruit yield, per plant and per unit area. Yield per plant declined with increasing population density while yield per unit land area, increased. Yield per unit land area was rarely described by an asymptotic relationship with population density. Estimates of model parameters (a and b) varied in response to row cover treatments (Table 7). Theta estimates were always larger in models for covered treatments than uncovered ones, but the difference was greater in models developed for early fruit harvests compared with total fruit harvests. This is reflected in a distinct difference in the slope of the lines for uncovered and covered treatments of early fruit yield, at low population densities (Figs. 8 and 9 compared with Figs. 10 and 11). Yield potential in the absence of competition (coefficient a scaled by 6) was highest in treatments with row covers. Coefficient b (scaled by 9) indicates responsiveness of yield to changes in population density. The greatest response was observed with early fruit yields from covered .treatments (Fig. 8 and 9). Estimates of b from regressions for total harvest 3.00 - i 6.0-2.25 - I 4.5-1.50 -CC < UJ CL 3 .0 -0.75 < LJ 1.5-0.00 12 0.0-POPULATION DENSITY, plants m" F i g . 8. M e a n s and nonl inear reg ress ions represent ing the relat ionship be tween ear ly marke tab le f resh fruit yield and plant populat ion density for uncove red , , and cove red , , t reatments ; (a) y"°=a + b X and (b) y'°=a + b X conver ted to a land a rea bas is . a 3.00 -b 8.0-, CT) z < a. cc UJ 72 CC u _J < r— O < 2.25 1.50 0.75 • 73 CC u_ _J i< o I— >j a : < 6.0-4. OA 2.0 0 .00 12 0.0 P O P U L A T I O N DENSITY , p l a n t s m " 2 12 Fig. 9. Means and nonlinear regressions representing the relationship between early total fresh fruit yield and plant 4 . 0 CT) < _J Q_ or U J C L 3 or u_ LU _ l m < i — or < 3.0 2.0 1.0 0.0 — 1 12 8.0-j 1 1 6 .0 -< l_J_J or < or L J CL t— 4 . 0 -or u . 111 _j CO j< UJ 2 .0 -or < 0 . 0 - — I 12 POPULATION DENSITY, p lants m -2 F ig . 10. M e a n s and nonl inear regress ions represent ing the relat ionship be tween total marke tab le f resh fruit yield and plant populat ion densi ty for uncove red , , and cove red , , t reatments; (a) y"°=a + b X and (b) y'°=a + b X conver ted to a land a rea bas is . F ig . 11 . M e a n s and nonl inear regress ions represent ing the relat ionship be tween total f resh fruit y ield and plant populat ion densi ty for uncove red , , and cove red , , t reatments; (a) y'°=a + b X and (b) y"°=a + b X conver ted to a land a rea bas is . - 52 -Table 7. Nonlinear model statistics for fruit fresh weight.7 Fruit weight Treatment G a b R S S E M S Early yield: marketable NC 0.0265 1.00 0.00293 6.70 0.0871 C 1.56 -0.230 0.438 18.3 0.237 total NC 0.343 0.965 0.0384 6.93 0.0990 C 1.75 -0.134 0.257 19.5 0.253 otal yield: marketable NC 1.03 0.232 0.125 20.9 0.274 C 1.18 0.0688 0.141 25.6 0.332 total NC 1.10 0.174 0.109 17.4 0.226 C 1.19 0.0841 0.0911 24.3 0.316 z Model formed according to y ^ a + bX where X-variates were expressed in plant m"2 and y was fruit fresh weight per plant. NC and C denote not covered and covered treatments respectively. R S S and E M S denote residual sum of squares and error mean square respectively. - 53 -(marketable and total yield), were generally more consistent between uncovered and covered treatments, than from regressions for early harvests (Table 7). This is illustrated in Figs 8 to 11. Satisfactory fits to nonlinear models for cull or undersized fruit, were not obtained. 4.2.4 Effects of row cover and population density on plant yield:  ANOVA results The application of row covers significantly increased total plant dry weight; plant components were not affected (Table 8). The dry weight of all variates decreased in a linear fashion with increasing population density. Interactions between row covers and plant population density were generally not significant. Significant treatment means are presented in Appendix VII. 4.2.5 Effects of row cover and population density on plant yield:  nonlinear regression models Figures 12 to 14 illustrate the fitted nonlinear regression models for total plant dry weight and plant components. Yield decreased per plant and increased per unit land area, with increasing population density. As with fruit yield, an asymptotic relationship between population density and total plant yield or vegetative yield, was not evident. - 54 -Table 8. ANOVA F-statistics and E M S : plant components from yield density study. Source df Total dry Vegetative Root dry Stem dry Leaf dry Leaf weight plant dry wt weight weight weight area (g) (g) (g) (g) (g) (m2) Block 3 0.59 0.14 0.90 0.29 0.04 0.490 Cover (C) 1 113" 7.73 2.72 8.74 6.07 4.40 Error (a) 3 1.80 2.36 3.66* 0.0585 1.40 2.55 Density (D) (4) 32.5*" 24.8*** 9.60*** 18.4*** 45.2*** 26.7*** linear 1 127*** 94.8*** 35.6*** 69.3*** 175*** 99.9*" quadratic 1 1.17 0.96 1.85 0.69 1.15 1.38 cubic 1 0.25 1.57 0.02 1.70 2.20 2.35 deviation 1 1.34 2.04 0.89 1.82 2.76 3.11 C«D (4) 1.28 1.16 0.440 1.04 1.68 1.30 OD L 1 0.69 2.18 0.34 2.26 2.70 1.41 Error (b) 24 2.17 2.37** 2.96 2.28 2.04 2.21 Error 40 EMS 1468 531 5.30 202 57.8 0.228 Asterisks denote significance at p=0.05O, 0.01 ( ") and 0.001 ("*). E M S denotes error mean square. F ig . 12. M e a n s and nonl inear regress ions represent ing the re lat ionship be tween total plant dry y ie ld and plant populat ion density for uncove red , , and cove red , , t reatments; (a) y"°=a + b X and (b) y"°=a + b X conver ted to a land a rea bas is . F ig . 13. M e a n s and nonlinear regressions representing the relationship between vegetative plant dry yield and plant population density for uncovered, , and covered, , treatments; (a) y ^ a + bX and (b) y ^ a + bX converted to a land a rea bas is . OJ < UJ cc < < UJ 1.2 • 0.9 0.6-0.3 0.0 CN I E cc < ll_ < ~1 12 5.0-4 . 0 -3 . 0 -2.0 1.0-0.0 b 12 -2 POPULATION DENSITY, plants m F ig . 14. M e a n s and nonl inear regress ions represent ing the relat ionship be tween leaf a r e a and plant populat ion densi ty for uncove red , , and cove red , , t reatments; (a) y"°=a + b X and (b) y"°=a + b X conver ted to a land a rea bas is . - 5 8 -Theta was consistently larger in covered treatments and was always greater than one (Table 9). Trends for a and b values were not as consistent as observed in fruit yield models. 4.2.6 Effects of row cover, population density and date of harvest on  plant components: ANOVA results and spline regressions The F statistics for tests of row cover, plant population density and harvest date effects on plant components are presented in Table 10. The linear effects of harvest date accounted for the greatest source of yield variation, with all plant components. Harvest date is indicative of plant growth, so this effect was expected. Except for leaf dry weight and number of nodes, row cover treatments had a greater influence on plant components than population density treatments. In all cases, the effect of population density was primarily linear with a relatively small quadratic contribution. The weight of each component decreased with increasing plant population density. Also, the effect of row covers decreased with increasing population density. Both row cover and population density effects on yield variation increased with later harvest dates. In both cases, the greatest source of variation was accounted for by the linear component of harvest date, and where applicable, population density. The interaction of row cover, population density and harvest date was rarely significant. - 59 -Table 9. Nonlinear model statistics for dry weight of plant components. 2 Fruit weight Treatment 0 a b R S S E M S Total plant dry weight NC 2.87 -21.6 36.3 0.0723 0.0020 C 3.48 5.40 16.5 0.0913 0.0025 Vegetative plant dry wt. NC 1.81 8.37 17.7 0.0250 0.0007 C 2.62 -95.7 126 0.0385 0.0010 z Model formed according to y ^ a + bX where X-variates were expressed in plant m"2. NC and C denote not covered and covered treatments respectively. R S S and E M S denote residual sum of squares and error mean square respectively. - 60 -Table 10. ANOVA F-statistics and E M S : plant components from plant growth study.z Source df Shoot Vegetative Leaf dry Leaf area Number Fruit dry dry shoot weight of nodes weight weight weight (m2) (g) (g) (g) (g) Block 3 0.171 0.542 Cover (C) 1 265"* 217"* Error (a) 3 0.495 0.870 Density (D) (4) 110*** 118"* linear (L) 1 421*** 444*** quadratic (Q) 1 17.1*" 27.4*** C-D 4 5.20** 5.61" C«DL Error (b) 1 17.5"* 19.8"* 24 1.22 1.07 Harvest (H) (5) 1023"* 743*** HL 1 5963*** 3691*** HQ 1 133*** 6.08* Hc 1 16.2*** 4.25* C«H (5) 17.3"* 16.9*** C«HL 1 80.6*** 29.6*** C-HQ 1 4.31* 13.6"* C-Hc 1 0.0119 8.16" D-H (20) 41.4*** 26.2"* DL-HL 1 746*** 490*** DL'HQ 1 42.5*** 3.99* DQ-HL 1 23.7*" 17.5"* C«D'H 20 1.49 1.29 Error 3 150 1.13 1.02 Error 240 EMS 838 164 0.686 0.480 0.180 1.01 122" 205*** 45.6** 310*** 1.39 0.872 3.18 0.249 175*** 95.4*** 92.9*** 81.7*" 663*** 353*** 323*** 315*** 35.8*** 26.5*** 40.7*** 9.11" 7.13" 6.42** 5.95** 3.93* 28.0*" 22.4*** 19.2*** 12.5" 1.03 1.18 0.919 1.31 704*** 495*** 396*** 980*** 3349*** 2198*** 1811*** 4515*** 137*** 232*** 157*" 354*** 9.15" 11.0" 0.273 22.3*" 9.57*" 12.6"* 4.95** 21.1"* 2.81 1.78 0.0969 110*** 25.7*** 33.0*** 3.91* 0.615 14.4*" 24.3*** 19.4*" 3.86 30.9*** 18.0*** 11.1*** 41.8*** 578*" 319*" 193"* 726*** 0.219 5.56* 3.56 72.2*" 11.7* 11.1" 7.33** 21.7*" 1.40 1.67* 0.867 1.61 1.04 1.21 1.43** 1.03 27.5 0.126 3221 411 Asterisks denote significance at p=0.05(*), 0.01 (**) and 0.001 (***). E M S denotes error mean square. Subscripts L, Q and C denote linear, quadratic and cubic trends respectively. - 61 -Spline regressions representing the response of variates, on a land area basis, to row covers, plant population density and growth are illustrated in Figs. 15 to 20. Early growth response to row covers is evident with all plant components. The effect of plant population density on component dry matter yield was consistent; the response increased with plant growth and was inversely related to plant population density. 4.3.0 Yield component analysis 4.3.1 Temporal trends of yield components. Spline regressions of yield components are illustrated in Fig. 19 and Figs. 21 to 27. Yield component regressions based on treatment means per plant or treatment means per unit land area, did not differ. The exception was node number. In this case, the lowest plant population density resulted in the greatest number of nodes per plant but the lowest number of nodes per unit land area (Figs. 19 and 21). Node number also showed a greater response to plant population density than the other yield components. Node number and WF/WV (Fig. 26) illustrate the early influence of row cover treatments as well as the increasing effects of plant population density as growth progressed. 2.25 1.80 • C N I CD 1.35 X O UJ Q: 0 .90 Q I— o o X m 0.45 0 - » 2.25-1 202 1.80 1.35 0.90 0.45 223 244 265 160 181 HARVEST DATE, Julian Day 202 223 244 265 Fig. 15. Spline regress ions represent ing the relat ionship be tween shoo t dry weight , growth a n d plant populat ion for a) u n c o v e r e d , and b) covered t reatments. P lant populat ion dens i t ies a re : , 1.39; , 1.85; — • — , 2.78; , 5.56 and; ,11 .1 plants m ' 2 . B a r s indicate con f idence intervals (P<0.05). CO CO HARVEST DATE, Julian Day Fig . 16. Sp l ine regress ions represent ing the relat ionship be tween vegetat ive shoot dry weight , growth and plant populat ion densi ty for a) uncove red , and b) covered t reatments. Plant populat ion dens i t i es a re : , 1.39; 1.85; — — , 2 .78 ; , 5.56 a n d ; ,11 .1 plants m" 2. B a r s indicate con f idence intervals (P<0.05). 0.300 0.225 CN I c c CD o • 0.150 cc Q < —I 0 .075-0.000 0.300-1 0.225-0.150 0.075-0.000 223 244 265 160 181 HARVEST DATE, Julian Day 202 223 244 265 F ig . 17. Sp l ine regress ions represent ing the relat ionship be tween leaf dry weight, growth a n d plant populat ion densi ty for a) uncovered , and b) cove red treatments. P lant populat ion dens i t ies a re : , 1.39; , 1.85; — 2.78; , 5.56 and ; ,11.1 plants m' 2 . B a r s indicate con f idence intervals (P<0.05). 223 244 265 160 181 HARVEST DATE, Julian Day 202 223 244 26S F ig . 18. Sp l ine regress ions represent ing the relat ionship be tween leaf a rea , growth a n d plant populat ion densi ty for a) uncove red , and b) covered t reatments. Plant populat ion densi t ies are: , 1.39; , 1.85; — — , 2 .78 ; 5.56 a n d ; ,11.1 plants m' 2 . B a r s indicate conf idence intervals (P<0.05). Fig. 19. Spline regressions representing the relationship between number of nodes per plant, growth and plant population density for a) uncovered, and b) covered treatments. Plant population densities are: , 1.39; 1.85; —•— , 2.78; , 5.56 and; ,11.1 plants m'2. Bars indicate confidence intervals (P<0.05). 160 181 202 223 244 265 160 181 202 223 244 265 HARVEST DATE, Julian Day F ig . 20 . Sp l ine reg ress ions represent ing the relat ionship be tween fruit f resh weight ; growth and plant populat ion densi ty for a) uncove red , and b) covered t reatments. Plant populat ion dens i t ies a re : , 1.39; , 1.85; — — 2.78; , 5.56 and ; ,11 .1 plants m' 2 . B a r s indicate con f idence intervals (P<0.05) . 3 0 0 0 n 3000-1 2500 2000 1500-1000-5 0 0 -223 244 265 160 HARVEST DATE, Julian Day 265 F ig . 2 1 . Sp l ine regress ions represent ing the relat ionship be tween number of n o d e s per m 2 , growth and plant populat ion densi ty tor a) uncove red , and b) covered t reatments. P lant populat ion dens i t i es a re : , 1.39; •• 1.85; — — , 2 .78 ; , 5.56 a n d ; ,11.1 plants m ' 2 . B a r s indicate con f idence intervals (P<0.05). o.o H 1 1 1 1 1 o.o-160 181 202 223 244 265 160 181 HARVEST DATE, Julian Day 202 2 2 3 244 265 F ig . 22. Sp l ine regress ions represent ing the relat ionship be tween leaf a r e a (LA):number of n o d e s (NN) , growth a n d plant populat ion densi ty for a) uncove red , and b) cove red t reatments. P lant popula t ion dens i t ies a re : , 1.39; , 1.85; — — , 2 .78; , 5.56 and ; ,11.1 plants m' 2 . B a r s indicate con f i dence intervals (P<0.05). Fig . 23 . Sp l ine regress ions represent ing the relat ionship be tween leaf dry weight (WL): leaf a r e a (LA) , growth and plant populat ion density for a) uncove red , and b) cove red treatments. P lan t populat ion dens i t ies a re : , 1.39; , 1.85; —•— , 2 .78; , 5.56 a n d ; ,11.1 plants m" 2. B a r s indicate con f i dence intervals (P<0.05). HARVEST DATE, Julian Day Fig. 24. Spline regressions representing the relationship between shoot dry weight (W):leaf dry weight (WL), growth and plant population density for a) uncovered, and b) covered treatments. Plant population densities are: , 1.39; , 1.85; — •— , 2.78; , 5.56 and; ,11.1 plants m'2. Bars indicate confidence intervals (P<0.05). 0.0- ) , 1 1 1 1 o . o - l 1 1 1 1 1 160 181 202 223 244 265 160 181 202 223 244 265 H A R V E S T DATE, J u l i a n D a y F ig . 25. Sp l ine regress ions represent ing the relat ionship between vegeta t ive shoot dry weight (WV): leaf dry weight (WL), growth and plant populat ion densi ty for a) uncovered , and b) c o v e r e d t reatments. P lan t popula t ion dens i t ies are: , 1.39; • , 1.85; — • — , 2.78; , 5.56 and ; ,11 .1 p lants m ' 2 . B a r s ind icate con f idence intervals (P<0.05). 160 181 202 223 244 265 160 181 202 223 244 265 HARVEST DATE, Julian Day F ig . 26 . Sp l ine regress ions represent ing the relat ionship be tween fruit dry we igh t (WF) :vegeta t i ve shoo t dry weight (WV) , growth and plant populat ion densi ty for a) uncove red , and b) cove red t reatments. P lant populat ion dens i t ies are: , 1.39; , 1.85; — • — , 2.78; , 5.56 and ; ,11.1 plants m' 2 . B a r s indicate con f idence intervals (P<0.05). a b 0 . 7 5 • 0.75 -i HARVEST DATE, Julian Day Fig . 27 . Quadra t ic po lynomia l regress ions represent ing the relat ionship be tween fruit dry weight (WF) :shoot dry weight (W), growth and plant populat ion densi ty for a) uncovered , and b) cove red t reatments. P lan t populat ion dens i t ies a re : , 1.39; , 1.85; — • — , 2 .78; , 5.56 a n d ; , 11 .1 plants m" 2 . B a r s indicate con f idence intervals (P<0.05). - 75 -4.3.2 Effect of treatments on the contribution of yield components to  fruit yield per plant: individual harvests Two dimensional partitioning (TDP) of total sums of squares assessed the extent to which experimental sources of variation influenced variation in yield components and in fruit yield. The results of the forward and backward analyses are presented in Tables 11 and 12 respectively. Yield components were entered into the stepwise multiple regression procedure in the order presented in the tables (i.e. in column order from left to right). Forward analysis determines the influence of successive components on yield after the effects of chronologically earlier components have been assessed. Significant yield components may affect yield directly, or indirectly through their effect on components entered later in the regression. Backward analysis determines the influence of successive components on yield after the effects of chronologically later components have been assessed. Significant yield components may affect yield directly, or indirectly through a relationship with chronologically earlier components, entered later in the regression. Forward TDP indicated that variation in WL/LA, W/WL and WF/W significantly influenced fruit yield in the first harvest, but no yield components were significantly affected by treatments (Table 11). In the first three harvests, WF/W made the greatest contribution to the total sum of squares for fruit yield (62%, 60% and 60% respectively). WFAA/ was the last component to be entered into the regression; therefore, significant contributions of this component are - 76 -Table 11. Two dimensional partitioning (TDP) of fruit dry weight per plant: forward analysis of individual harvests.2 Harvest Yield components Sum of Yield date Source df NN LA/NN WL/LA W/WL WF/W products WF 160 181 202 223 224 265 Block 3 0 0 1 1 4 -3 3 Cover 1 1* 0 0 0 0 2 3 Error (a) 3 0 0 1 0 1 1 3 Density 4 0 0 2 0 1 1 4 Cover-density 4 0 0 4 1 4 -5 4 Error 64 1 0 22* 4 52 4 83 Total 79 2 0 30* 6* 62* 0 100 Block 3 0 0 0 1* 1 1 3 Cover 1 12* 1* 0 1* 0 11 25* Error (a) 3 1 0 0 0 3 -1 3 Density 4 1* 0 0 4* 8* -9 4 Cover-density 4 0 0 1* 1 2 -3 1 Error 64 5 4 2 6 46 1 64 Total 79 19* 5* 3 13* 60* 0 100 Block 3 1 0 0 1 3 -3 2 Cover 1 13* 1 0 1 1 8 24* Error (a) 3 0 0* 0 1 6* 4 11* Density 4 6* 1* 1* 0 4 -7 5 Cover-density 4 1* 1* 0 0 0 -1 1 Error 64 4 2 1 6 46 -2 57* Total 79 25* 5* 2 8* 60* 0 100 Block 3 0 5 0 1 0 -2 4 Cover 1 3 13* 0 7* 1 28 52* Error (a) 3 1* 1 0 2* 1 -2 3 Density 4 12* 1 1* 2 0 -6 10* Cover-density 4 0 2 0 0 0 1 3 Error 64 9 17 2 14 5 -19 28 Total 79 25* 39* 3 26* 7* 0 100 Block 3 1 1 0 0 0 -2 0 Cover 1 1 0 0 7* 0 5 13* Error (a) 3 1 1 0 0 0 -2 0 Density 4 41* 1 0 1 0 24 67* Cover-density 4 2 0 0 2 0 -3 1 Error 64 19 5 4 12 1 -22 19 Total 79 65* 8* 4 22* 1 0 100 Block 3 2 2 0 0 0 -3 1 Cover 1 3 0 0 1 0* 1 5* Error (a) 3 1 0 0 0 0 0 1 Density 4 53* 1* 0 0 0 15 69* Cover-density 4 0 0 0 0 0 1 1 Error 64 18 4 2 12 1 -14 23 Total 79 77* 7* 2 13* 1 0 100 2 Cells to the right of the df column are percentages of the total sum of squares for WF. Asterisks denote significance at p=0.05(*). 0 denotes less than 0.5% of the total sum of squares. - 77 -Table 12. TDP of fruit dry weight per plant: backward analysis of individual harvests.2 Harvest Yield components Sum of Yield date Source df W7W WTwI WULA VMR M~ Products "WT 160 181 202 223 224 265 Block 3 1 0 r 0 0 1 3 Cover (C) 1 5 0 2* 0 5* -9 3 Error (a) 3 2 0 0 0 0 1 3 Density (D) 4 3 0 1 0 1 -1 4 Cover«density 4 2 0 0 0 0 2 4 Error 64 62 0 5 2 8 6 83 Total 79 75* 0 9* 2 14* 0 100 Block 3 4 0 0 0 0 -1 3 Cover 1 15* 1* 0 0 0 9 25* Error (a) 3 3 0 0 0 0 0 3 Density 4 4 0 0 0 0 0 4 Cover«density 4 2 0 0 0 0 -1 1 Error 64 70 0 0 0 1 -7 64 Total 79 98* 1 0 0 1 0 100 Block 3 3 0 0 0 0 -1 2 Cover 1 14 0 0 0 0 10 24* Error (a) 3 13* 0 0 0 0 -2 11* Density 4 4 0 0 0 1* 0 5 Covefdensity 4 1 0 0 0 0 0 1 Error 64 63 0 1 0 0 -6 57* Total 79 98* 0 1 0 1 0 100 Block 3 3 0 0 2 1 -2 4 Cover 1 16* 0 0 2* 4 30 52* Error (a) 3 1 0 0 0 3* -1 3 Density 4 3 0 0 2* 7* -2 10* Cover«density 4 2 0 0 0 0 1 3 Error 64 35 0 1 6 12 -26 28 Total 79 60* 0 1 12* 27* 0 100 Block 3 2 0 0 0 1 -3 0 Cover 1 3* 0 0 0 3 7 13* Error (a) 3 0 0 0 1* 1 -2 0 Density 4 11* 0 0 1 19* 36 67* Cover-density 4 3 0 0 0 2 -4 1 Error 64 31 0 0 4 18 -34 19 Total 79 50* 0 0 6' 44* 0 100 Block 3 1 0 0 4 1 -5 1 Cover 1 0 0 0 0 2 3 5* Error (a) 3 0 0 0 1 1 -1 1 Density 4 5* 0 0 4 17* 43 69* Cover-density 4 0 0 0 1* 1 -1 1 Error 64 28 1 3 10 20 -38 23 Total 79 34* 1 3 20* 42* 0 100 z Cells to the right of the df column are percentages of the total sum of squares for WF. Asterisks denote significance at p=0.05(*). 0 denotes less than 0.5% of the total sum of squares. - 78 -considered to represent direct effects on fruit yield. The contribution of WF/W declined to 7% in the fourth harvest, and did not significantly contribute to fruit yield in the later harvests. Plant population density treatments affected WF/W only at the second harvest. In the forward TDP analysis, NN, LA/NN, and W/WL contributed significantly to fruit yield variation from the second harvest on (Table 11) . The relative contribution of NN increased with each successive harvest. Where NN was a significant contributor to the total sum of squares for fruit yield, it was influenced by plant population density and, at the second and third harvest, by row covers. The relative contributions of LA/NN and W/WL to fruit yield variation were highest at the fourth harvest (39% and 26% respectively). Backward analysis showed a greater contribution of WF/W to fruit yield compared with the forward analysis; however in this procedure, the indirect influence of chronologically earlier components has not been eliminated (Table 12) . Direct effects of NN on fruit yield were observed at all harvests except the second and third. Row covers were a significant source of variation on NN at the first harvest, while plant population density was significant from the fourth harvest on. Row covers accounted for 25% of the fruit yield variation, at the second harvest: the variation was attributed to W/WL (1%) and to compensation among components (11%, indicated by the sum of products) (Table 11). The contribution of NN was through its influence on chronologically later components, and of WFAA/, earlier components; NN and WFAV did not contribute significantly - 79 -to the total sum of squares when entered last in the regression (Tables 11 and 12). Row covers and plant population density jointly and significantly influenced fruit yield from the fourth harvest on (Table 11 and 12). The effect of row covers declined from 52% to 5% during growth while that of plant population density increased from 10% to 69%. Forward and backward analyses showed that the variation in plant population density, for the last three harvests, was attributed mainly to differences in node number, and then to compensation among other components. The exception to this was at the third harvest, where compensation among components subtracted some variation (6%). 4.3.3 Effect of treatments on the contributions of yield components to  fruit yield per plant: combined harvest data The results of forward and backward TDP are summarized in Tables 13 and 14 respectively. NN, W/WL and WF/W significantly influenced fruit yield in both forward and backward TDP. NN and WF/W, when entered last into the regression, accounted for 1% (Table 14) and 15% (Table 13) respectively, of the total sum of squares for fruit yield. The contribution of these components was high when they were the first components entered into the regression, however, their influence on fruit yield was primarily indirect. In almost all cases, components showed the greatest treatment response to harvest date (Table 13). NN was the most responsive to treatment - 80 -Table 13. TDP of fruit dry weight per plant: forward analysis of combined harvest data. 2 Source df Yield Component Sum of Yield NN LA/NN WULA W/WL WF/W Products WF Blocks 3 0 0 0 0 0 0 0 Covers (C) 1 4* 0 0 0 0 -2 2* Error (a) . 3 0 0 0 0 0 0 0 Density (D) 4 3* 0 0 1* 1* -5 0 C«D 4 0 0 0 0 0 0 0 Error (b) 24 0 0 0 0 1 0 1 Harvest date (H) 5 59* 1* 1* 3* 2* 12 84* C«H 5 2* 2* 0 0 0 -2 2* D«H 20 1* 0 0 0 1 -1 0 C«D«H 20 0 0 0 0 0 -1 0 Error (c) 150 2 1 0 0 4* -2 5 Error 240 2 3 0 0 6 -5 6 Total 479 73* 7* 1* 4* 15* 0 100 z Cells to the right of the df column are percentages of the total sum of squares for WF. Total sums of squares are 823, 53, 14, 72, 3114 and 20 377 respectively. Asterisks denote significance at p=0.05(*). 0 denotes less than 0.5% of the total sum of squares. - 81 -Table 14. TDP of fruit dry weight per land area: backward analysis of combined harvest data. z Source df Yield Component Sum of Yield WF/W W/WL WL/LA LA/NN NN Products WF Blocks 3 0 0 0 0 0 0 0 Covers (C) 1 1 0 0 0 0 1 2* Error (a) 3 0 0 0 0 0 0 0 Density (D) 4 0 0 0 0 0 0 0 C«D 4 0 0 0 0 0 0 0 Error (b) 24 0 0 0 0 0 1 1 Harvest date (H) 5 75* 1* 0 0 1* 7 84* C-H 5 3* 0 0 0 0 -1 2* D«H 20 1 0 0 0 0 0 1 C«D«H 20 0 0 0 0 0 0 0 Error (c) 150 8* 0 0 0 0 -4 4 Error 240 10 0 0 0 0 -4 6 Total 479 98* 1* 0 0 1* 0 100 z Cells to the right of the df column are percentages of the total sum of squares for WF. Total sums of squares are 14 556, 79, 19, 32, 361 and 23 573 respectively. Asterisks denote significance at p=0.05(*). 0 denotes less than 0.5% of the total sum of squares. - 82 -of all components. Harvest date accounted for 84% of the total fruit yield variation; row covers, and row cover by harvest date interaction, each accounted for 2%. 4.3.4 Effect of treatments on the contribution of yield components to  total shoot dry matter yield per plant: individual harvests Forward TDP showed that NN was the greatest contributor to the total sum of squares for yield (73% to 92%), at all harvests (Table 15). LA/NN also contributed significantly, with the exception of the second harvest, and WF/WV contributed after the third harvest. No significant contributions to the total sum of squares from the other variates were observed. Row covers accounted for the greatest proportion of the variation in NN in the first three harvests (44% to 60%), but the effects of row covers were either not significant or were of minor importance in the remaining three harvests (Table 15). The contribution of plant population density effects on NN rose from 7% at the second harvest to 57% at the final harvest. At all harvests, treatment effects accounted for at least 75% of the total yield variation (Table 15). The relative distribution of treatment variation followed that observed for NN; as growth proceeded, plant population density effects rose from 4% to 73% of the total, while row cover effects declined from 71% at the initial harvest to 6% at the final harvest. The variation in treatment effects was mainly attributed to differences in NN, and to compensation among yield components; compensation - 83 -Table 15. TDP of shoot dry weight per plant: forward analysis of individual harvests.7 Harvest Yield components Sum of Yield date Source df NN LA/NN WL7LA WMVL WhAW WM/F products 160 Blocks 3 1 0 0 1 0 0 0 2 Cover(C) 1 44' 3' 0 0 0 0 24 71' Error (a) 3 2 1 0 0 0 0 -2 1 Density (D) 4 2 0 0 0 0 0 2 4 C-D 4 2 1 0 0 0 0 -2 1 Error (b) 24 9 5 0 2 0 0 -5 11' Error 40 13 9 1 4 0 0 -4 10 Total 79 73* 19* 1 7 0 0 0 100 181 Blocks 3 0 0 0 0 0 0 0 0 Cover 1 60* 1 0 0 0 0 14 75' Error (a) 3 2 0 0 0 0 0 0 2 Density 4 7* 0 0 1 0 0 -4 4' C-D 4 1 0 1 0 0 0 -2 0 Error (b) 24 7 0 1 0 0 0 -2 6 Error 40 15 2 2 0 0 0 -6 13 Total 79 92' 3 4 1 0 0 0 100 202 Blocks 3 3 0 0 0 0 0 -1 2 Cover 1 48' 1 0 0 0 0 17 66' Error (a) 3 1 1 0 0 0 0 •1 1 Density 4 21* 1' 0 0 0 0 -7 15* C-D 4 4* 1' 0 0 0 0 -4 1 Error (b) 24 6 1 0 0 0 0 -2 5 Error 40 8 2 0 1 0 1 -2 10 Total 79 91* 7* 0 1 0 1 0 100 223 Blocks 3 0 3 0 0 0 0 -1 2 Cover 1 8 7' 0 1 0 0 27 43* Error (a) 3 4' 0 0 0 0 0 -1 3' Density 4 34* 0 1' 0 0 0 -1 34* C-D 4 1 1* 0 0 0 0 0 2 Error (b) 24 9 3 0 1 2 0 -9 6 Error 20 17 5 1 0 1 1 -15 10 Total 79 73' 19* 2 2 3 1 0 100 224 Blocks 3 1 1' 0 0 0 0 -2 0 Cover 1 2' 0 0 1 1' 0 8 12' Error (a) 3 1 1' 0 0 0 0 -2 0 Density 4 47* 1* 1 0 0 0 20 69' C-D 4 3' 0 0 0 1 0 -3 1 Error (b) 24 12* 3 1 0 3 0 -10 9 Error 20 10 3' 2 1 3 0 -10 9 Total 79 77' 9* 4 2 8* 0 0 100 265 Blocks 3 2 2 0 0 0 0 -1 1 Cover 1 4 0 0 0 0 0 2 6' Error (a) 3 2 0 0 0 0 0 -1 1 Density 4 57' 1* 0 0 0 0 14 73' C-D 4 0 1 0 0 0 0 0 1 Error (b) 24 10' 2 1 1 2 0 -7 9 Error 20 8 3 0 1 3 0 -6 9 Total 79 83' 9* 1 2 5' 0 0 100 z Cells to the right of the df column are percentages of the total sum of squares for W. Asterisks denote significance at p=0.05(*). 0 denotes less than 0.5% of the total sum of squares. - 8 4 -added to variation between row cover treatments and, in almost all cases, subtracted some variation between plant population density treatments. NN was entered last into the backward TDP regression, enabling an assessment of the direct effects of NN on total yield. The analysis confirmed the results of the forward analysis; NN was the greatest contributor to the total sum of squares for yield at all harvests except one (Table 16). The exception was in the initial harvest where there was no significant contribution. In this case, forward TDP results implied an indirect effect of NN on chronologically later components. NN was affected equally by row covers and plant population density at the second harvest, but primarily by plant population density after this harvest (Table 16). W/WF, which showed no significant contribution in the forward TDP analysis, was a significant contributor at all harvests in the backward TDP analysis. This indicates a relationship of W/WF with chronologically earlier components. Backward analysis showed the greatest contributor to the total sum of squares for yield at the initial harvest was WF/WV (84%). This component was not, however, a significant contributor in the forward TDP analysis, suggesting that WF/WV was not a direct contributor to yield variation. All components were directly or indirectly related to yield variation in the third harvest. - 85 -Table 16. TDP of shoot dry weight per plant: backward analysis of individual harvests. Harvest Yield components Sum of Yield date Source df "W/WF—Wh/WV WV/WL WL/LA TORN RN" Products T T 160 Blocks Cover (C) Error (a) Density (D) OD Error (b) Error Total 181 Blocks Cover Error (a) Density OD Error (b) Error Total 202 Blocks Cover Error (a) Density OD Error (b) Error Total 223 Blocks Cover Error (a) Density OD Error (b) Error Total 224 Blocks Cover Error (a) Density C-D Error (b) Error Total 265 Blocks Cover Error (a) Density OD Error (b) Error Total 3 0 3' 1 1 53' 3 0 2 4 1 4* 4 0 2 24 5 8 40 6 12 79 13' 84' 3 1 0 1 3* 0' 3 0 0 4 1 0 4 0 0 24 4 1 40 8 2 79 17* 3 3 1 0 1 2* 1' 3 2' 0 4 1 0 4 0 1 24 6 2 40 5 3 79 17' 7* 3 1 0 1 3' 0 3 0 0 4 • 1 0 4 0 0 24 2 2 40 4 3 79 11* 5* 3 1 0 1 2' 0 3 0 0 4 7' 0 4 2 0 24 9 0 40 10 1 79 31' 1 3 0 0 1 0 0 3 0 0 4 3' 0 4 0 0 24 6 1 40 11 2 79 20' 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 1 0 0 0 0 0 7* 0 0 0 0 0 5' 0 0 0 0 2' 0 0 3 3 3 4 17' 3 9* 0 1* 0 12* 1* 0 0 0 0 3' 0 0 1 0 1* 2 2 1 5 3 2 23' 7' 4' 0 0 3' 1 0 3' 0 0 1 0 1' 4* 0 0 1 0 2 6* 0 3 6 1 6' 24* 0 0 1 1 0 0 0 0 1* 0 0 1 0 0 0 0 0 3* 0 0 2 1 0 8' 0 0 4* 0 0 1 0 0 1 0 0 4' 0 0 1 1 1 3 1 1 9 2 2 23' 0 -1 2 0 17 71' 0 -1 1 0 0 4* 0 -1 1 1 -3 11* 1 -10 10 2 0 100 0 -2 0 13* 49 75* 1 2 3 13' -16 3' 3* -5 0 10 -12 6 10 -17 13 51* 0 100 2* -2 2* 2* 48 66' 0 -1 1 23* -12 15* 3' -5 1 5 -13 5 7 -15 10 42* 0 100 3' -5 2 5' 30 42* 5* -3 3* 19' 9 34* 2 -1 2 9 -14 7 10 -16 10 53' 0 100 2 -4 0 4* 5 12' 1 -2 0 24' 37 69* 3 -4 1 13* -16 9 12 -16 9 59* 0 100 2 -5 1 3' 2 6* 1 -1 1 21' 45 73* 1 -1 1 12* -15 9 10 -25 9 50' 0 100 z Cells to the right of the df column are percentages of the total sum of squares for W. Asterisks denote significance at p=0.05(*). 0 denotes less than 0.5% of the total sum of squares. - 86 -4.3.5 Effect of treatments on the contribution of yield components to  total shoot dry matter yield per plant: combined harvest data The results of forward and backward TDP, analyzing data combined from six independent harvests, are presented in Tables 17 and 18 respectively. Forward TDP showed that all components, excluding WF/WV, significantly contributed to the sum of squares for total yield (Table 17). NN was the greatest contributor to yield variation (88%). Harvest date was a significant source of variation for all significant yield components, whereas only NN was affected by row covers and plant population density. Backward analysis showed that W/WF (72%), WF/WV (12%) and NN (16%) significantly contributed to yield variation (Table 18). The contribution of NN was a direct relationship to yield variation, whereas W/WF most likely expressed a relationship with chronologically earlier components. All yield components were significantly affected by harvest date and NN was also affected by row covers and plant population density. 4.3.6 Effect of treatments on the contribution of yield components to  fruit yield per unit land area: individual harvests In general, the contribution of yield components to fruit yield, analyzed per unit land area, followed the same trends as those determined for individual plants. Forward analysis indicated that WFA/V was the greatest contributor to fruit yield variation in the initial three harvests, after which its contribution was - 87 -Table 17. TDP of shoot dry weight per plant: forward analysis of combined harvest data. z Source Yield components df NN LA/NN WL/LA WV/WL WF/WV Block 3 0 0 0 0 0 Cover (C) 1 5* 0 0 0 0 Error (a) 3 0 0 0 0 0 Density (D) 4 3* 0 0 0 0 C-D 4 0 0 0 0 0 Error (b) 24 0 0 0 0 0 Harvest date 5 71* 1* 1* 1* 0 O H 5 3* 1* 0 0 0 D-H 20 1* 0 0 0 0 OD-H 20 0 0 0 0 0 Error (c) 150 2 1 1 1* 0 Error 240 3 2 1 1 0 Total 479 88* 5* 3* 3* 0 W/WF Sum of products Yield W 0 0 0 0 0 0 1* 0 0 0 0 0 1* 0 -1 0 -1 0 0 15 -3 0 0 -4 -6 0 0 4* 0 2* 0 0 90* 1* 1* 0 1 1 100 z Cells to the right of the df column are percentages of the total sum of squares for W. Total sums of squares are 823, 53, 14, 12, 3273, 16 and 1622 respectively. Asterisks denote significance at p=0.05(*). 0 denotes less than 0.5% of the total sum of squares. - 88 -Table 18. TDP of shoot dry weight per plant: backward analysis of combined harvest data. 2 Source Yield components Sum of Yield df W/WF WF/WV WV/WL WL/LA LA/NN NN products W Block 3 0 0 0 Cover (C) 1 1 0 0 Error (a) 3 0 0 0 Density (D) 4 0 0 0 C-D 4 0 0 0 Error (b) 24 1 0 0 Harvest date 5 55* 8* 0 O H 5 2* r 0 r>H 20 0 0 0 OD-H 20 0 0 0 Error (c) 150 6 1 0 Error 240 7* 2 0 Total 479 72* 12* 0 0 0 0 0 0 0 0 1* 2 4* 0 0 0 0 0 0 0 1* 1 2' 0 0 0 0 0 0 0 1 -2 0 0 0 4* 23 90* 0 0 1* -3 1* 0 0 1* 0 1* 0 0 0 0 0 0 0 3 -9 1 0 0 4 -12 1 0 0 16* 0 100 2 Cells to the right of the df column are percentages of the total sum of squares for W. Total sums of squares are 11 457, 38, 7, 16, 32, 259 and 1622 respectively. Asterisks denote significance at p=0.05(*). 0 denotes less than 0.5% of the total sum of squares. - 8 9 -small or not significant (Table 19). WF/W was affected by row covers in the first harvest and by density in the second. The latter three harvests were dominated by the contribution of NN. Row covers and density significantly influenced NN at all harvests except the first. W/WL was a significant contributor to the sum of squares at all harvests, but the greatest contribution was from the last two harvests. Row covers influenced W/WL at all but the initial harvest (Table 19). Backward analysis indicated NN was a significant and direct contributor to the sum of squares for fruit yield at all harvests but the second and third (Table 20). WF/W was significant at all harvests, but its contribution in the last two harvests was most likely through its relationship with chronologically earlier components (Tables 19 and 20). W/WL contributed directly to the sum of squares from the third harvest on (Table 20). Treatment influences on yield components in forward and backward TDP also followed the same trends those determined for individual plants. The influence of row covers on total fruit yield, significant from the second harvest on, declined steadily after the fourth harvest (Tables 19 and 20). Population density significantly affected fruit yield variation after the third harvest. 4.3.7 Effect of treatments on the contribution of yield components to  fruit yield per unit land area: combined harvest data Forward and backward TDP of fruit yield per unit land area followed the same trends as those established for individual plants (Tables 21 and 22). - 90 -Table 19. TDP of fruit dry weight per land area: forward analysis of individual harvests.7 Harvest Yield components Sum of Yield date Source df NN LA/NN WULA W/WL WF/W products WF 160 Blocks 3 0 0 1 1* 3 -2 3 Cover 1 2 0 3* 0 4* -7 3 Error (a) 3 0 0 1* 0 2 0 3 Density (D) 4 2 0 2* 0 3 -3 4 C-D 4 0 0 3* 0 3 -2 4 Error (b) 24 0 0 7 2 16 8 33 Error 40 0 0 5 3 37 6 50 Total 79 4 0 22* 6* 68* 0 100 181 Blocks 3 0 1 0 2* 1 -1 3 Cover 1 3* 0 0 4* 0 16 23* Error (a) 3 0 0 0 0 3 0 3 Density 4 4* 0 0 2 9* -11 4 C-D 4 0 0 1* 1 2 -3 1 Error (b) 24 0 0 2 4 14 3 23 Error 40 1 4 3 9 32 -6 43 Total 79 8* 5* 6* 22* 61* 0 100 202 Blocks 3 0 0 0 1* 2 0 3 Cover 1 4* 3* 2* 1* 1 11 22* Error (a) 3 0 1* 0 1* 7* 3 12* Density 4 7* 1* 1* 0 5 -11* 3 C-D 4 0 1* 0 0 1 -1 1 Error (b) 24 0 1 2 2 26 2 33 Error 40 1 3 2 3 21 -4 26 Total 79 12* 10* 7* 8* 63* 0 100 223 Blocks 3 0 4* 0 0 0 -1 3* Cover 1 3* 12* 0 3* 0 23 41* Error (a) 3 2* 0 0 1 0 -1 2 Density 4 26* 1 0 2 0 -1 28* C-D 4 0 2 0 1 0 0 3 Error (b) 24 3 5 1 6 2 -8 9 Error 40 6 7 1 9 3 -12 14 Total 79 40* 31* 2 22* 5* 0 100 224 Blocks 3 0 1* 0 0 0 -1 0 Cover 1 1* 0 0 12* 0 8 21* Error (a) 3 0 0 0 0 0 -1 0 Density 4 44* 0 0 1 0 1 47* C-D 4 1* ' 0 0 2 0 -2 1 Error (b) 24 6* 3* 1 8 1 -6 13 Error 40 5 3 1 9 1 -2 17 Total 79 57* 7* 2 32* 2 0 100 265 Blocks 3 1* 2* 0 0 0 -2 1 Cover 1 3* 0 0 3* 0 6 12* Error (a) 3 1 1 0 1 0 -1 2 Density 4 41* 0 0 1 0 -7 35* C-D 4 0 1 0 0 0 0 1 Error (b) 24 6* 2 1 9 0 4 22 Error 40 5 3 2 16 1 0 27 Total 79 57* 9* 3 30* 1 0 100 z Cells to the right of the df column are percentages of the total sum of squares for WF. Asterisks denote significance at p=0.05(*). 0 denotes less than 0.5% of the total sum of squares. - 91 -Table 20. TDP of fruit dry weight per land area: backward analysis of individual harvests.2 Harvest Yield components Sum of Yield date Source df WF/W W/WL WL/LA LA/NN NN products WF 160 Blocks 3 1 0 0 0 1 1 3 Cover (C) 1 2 0 0 0 7* -6 3 Error (a) 3 2 0 0 0 0 1 3 Density (D) 4 10* 0 0 0 16* -22 4 C-D 4 1 0 0 0 1 2 4 Error (b) 24 21 0 0 1 2 9 33 Error 40 27 1 0 1 6 15 50 Total 79 64* 1 0 2 33* 0 100 181 Blocks 3 3 0 0 0 0 0 3 Cover 1 14* 1 0 0 0 8 23* Error (a) 3 3 0 0 0 0 0 3 Density 4 6 1 0 0 1 -4 4 C-D 4 2 0 0 0 0 -1 1 Error (b) 24 24 0 0 0 0 -1 23 Error 40 45 0 0 0 0 -2 43 Total 79 97* 2 0 0 1 0 100 202 Blocks 3 4 0 0 0 0 -1 3 Cover 1 13* 1 0 0 0 7 22* Error (a) 3 13* 0 0 0 0 0 12* Density 4 4 0 0 0 1 -2 3 C-D 4 1 0 0 0 0 0 1 Error (b) 24 35* 0 0 0 0 -2 33 Error 40 27 1 0 0 0 -2 26 Total 79 97* 2 0 0 1 0 100 223 Blocks 3 3 1 0 0 1 -2 3 Cover 1 17* 2* 0 0 1 21 41* Error (a) 3 2 1 0 0 0 -1 2 Density 4 3 2* 1 0 6* 16 28* C-D 4 2 0 0 0 0 1 3 Error (b) 24 14 1 1 0 5 -12 9 Error 40 23 3 2 0 9 -23 14 Total 79 64* 10* 4 0 22* 0 100 224 Blocks 3 1 0 0 0 1 -2 1 Cover 1 1* 16* 0 0 0 4 21* Error (a) 3 0 1 0 0 1 -2 0 Density 4 4* 6* 0 0 26* 11 47* C-D 4 1 1 0 0 2 -3 1 Error (b) 24 5 4 0 0 10 -6 13 Error 40 6 7 0 0 23 -19 17 Total 79 18* 35* 0 0 63* 0 100 265 Blocks 3 0 1 0 0 1 -1 1 Cover 1 0 5* 0 0 1 6 12* Error (a) 3 0 1 0 0 4* -3 2 Density 4 1* 5* 0 0 22* 7 35* C-D 4 0 2 0 0 1 -2 1 Error (b) 24 2 5 0 0 15 0 22 Error 40 5 9 0 0 20 -7 27 Total 79 8* 28* 0 0 64* 0 100 2 Cells to the right of the df column are percentages of the total sum of squares for WF. Asterisks denote significance at p=0.05(*). 0 denotes less than 0.5% of the total sum of squares. - 92 -Table 21. TDP of fruit dry weight per land area: forward analysis of combined harvest data. z Source df NN Yield components LA/NN WL/LA W/WL Sum of Yield WF/W Products WF Blocks 3 0 0 0 0 0 0 0 Covers (C) 1 3* 0 0 0 0 -1 2* Error (a) 3 0 0 0 0 0 0 0 Density (D) 4 9* 0 0 0 3* -12 0 CO 4 0 0 0 0 0 0 0 Error (b) 24 0 0 0 0 1 0 1 Harvest date (H) 5 47* 1* 1* 2* 4* 28 83* C«H 5 2* 1* 0 0 1 -2 2* D«H 20 1* 0 0 0 1* -1 1 C«D-H 20 0 0 0 0 0 0 0 Error (c) 150 1 2 1 1 5* -5 5* Error 240 2 3 1 1* 6 -7 6 Total 479 65* 7* 3* 4* 21* 0 100 z Cells to the right of the df column are percentages of the total sum of squares for WF. Total sums of squares are 921, 53, 15, 67,4954 and 23 573 respectively. Asterisks denote significance at p=0.05(*). 0 denotes less than 0.5% of the total sum of squares. - 93 -Table 22. TDP of fruit dry weight per land area: backward analysis of combined harvest data. 2 Source df Yield components Sum of Yield WF/W W/WL WL/LA LA/NN NN Products WF Blocks 3 0 0 0 0 0 0 0 Cover (C) 1 r 0 0 0 0 1 2* Error (a) 3 0 0 0 0 0 0 0 Density (D) 4 1* 0 0 0 1* -2 0 C«D 4 0 0 0 0 0 0 0 Error (b) 24 1 0 0 0 0 0 1 Harvest date (H) 5 73* 1* 0 0 1* 8 83* C-H 5 2* 0 0 0 0 0 2* D«H 20 1 0 0 0 0 -3 0 C-D-H 20 0 0 0 0 0 0 0 Error (c) 150 8 0 0 0 0 -3 5* Error 240 10 0 0 0 0 -4 6 Total 479 97* 1* 0 0 2* 0 100 2 Cells to the right of the df column are percentages of the total sum of squares for W. Total sums of squares are 14 056, 79, 19, 32, 361 and 23 573 respectively. Asterisks denote significance at p=0.05(*). 0 denotes less than 0.5% of the total sum of squares. - 94 -All components contributed to the sum of squares for fruit yield in forward analysis; however, the greatest contributors were NN and WF7W (Table 21). The contribution of NN was primarily through its influence on chronologically later components, as indicated in the backward analysis (Table 22). WF/W affected fruit yield directly and as a consequence of its relationship with chronologically earlier components. Harvest date had the greatest influence on all significant yield components and on fruit yield. 4.3.8 Effect of treatments on the contribution of yield components to  total shoot dry matter yield per unit land area: individual harvests The results of forward TDP analysis of total yield per unit land area were similar to those from the analysis conducted on an individual plant basis. Forward analysis showed the greatest contribution to the sum of squares for total yield was from NN, followed by the contribution from LA/NN (Table 23). Backward TDP analysis indicated the contribution of NN to total yield was direct, except at the first harvest (Table 24). WVA/VL contributed a large proportion of the total sum of squares from the second harvest on. Forward TPD, however, indicated the effects of WV7WL on yield were most likely through its relationship with chronologically earlier components. LA/NN only contributed at the first harvest. Total yield was influenced by row covers and population density at all harvests. Furthermore, the effects of population density on total yield were - 95 -Table 23. TDP of shoot dry weight per land area: forward analysis of individual harvests/ Harvest Yield components Sum of Yield date Source df "RR DORR WL/LA WV/WL WhAW W/WF Products ~"W~ 160 Block Cover Error (a) Density (D) OD Error (b) Error Total 181 Block Cover Error (a) Density OD Error (b) Error Total 202 Block Cover Error (a) Density OD Error (b) Error Total 223 Block Cover Error (a) Density OD Error (b) Error Total 224 Block Cover Error (a) Density OD Error (b) Error Total 265 Block Cover Error (a) Density C-D Error (b) Error Total 3 1 0 1 31' 0 3 1 1 4 31' • 6' 4 2 1 24 6 3 40 9 5 79 81' 16' 3 0 0 1 32' 0 3 1 0 4 50* 0 4 1 0 24 4 0 40 8 2 79 96' 2 3 1 0 1 29' 1' 3 1 0 4 52* 1* 4 2' 1* 24 4 1 40 5 1 79 94' 5' 3 0 2' 1 6' 6* 3 3' 0 4 50' 0 4 0 1 24 6 2 40 13 4 79 78' 15' 3 0 1* 1 1' 0' 3 1 0 4 61* 0 4 2' 0 24 8' 3' 40 6 3 79 79' 7' 3 2' 2' 1 3' 0 3 1 1 4 53' 1 4 0 1 24 8' 2 40 7 5 79 73' 12' 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 2 0 0 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 1 0 0 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 1 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 1 1 1 2 2 1 3 0 0 0 0 1 1' 0 0 0 0 0 0 0 0 0 1 1 3 2 1 4 3 3 8' 0 0 0 0 1 0 0 0 0 0 0 1 0 0 0 1 2 2 1 2 5 2 5 8' 0 0 1 0 -3 28' 0 -1 1 0 24 61' 0 -2 1 0 -6 4' 0 -12 4 0 0 100 0 0 0 0 11 43* 0 0 1 0 -5 45' 0 -1 0 0 -1 4 0 -4 7 0 0 100 0 0 1' 0 13 43' 0 -1 0 0 -9 44' 0 -2 1 0 -2 3 0 0 7 0 0 100 0 -1 1 0 22 34* 0 -1 2' 0 -2 48' 0 1 2 0 -4 5 1 -14 8 1 0 100 0 -1 0 0 12 15' 0 -1 0 0 3 64' 0 -1 1 0 -6 10' 0 -6 10 0 0 100 0 -3 1 0 8 12' 0 0 2 0 -8 47' 0 1 2 0 3 18 0 -1 18 0 0 100 z Cells to the right of the df column are percentages of the total sum of squares for W. Asterisks denote significance at p=0.05(*). 0 denotes less than 0.5% of the total sum of squares. - 96 -Table 24. TDP of shoot dry weight per land area: backward analysis of individual harvests.7 Harvest Yield components Sum ot Yield date Source dt W/WF WF/WV WV/WL WL/LA LA/NN NN Products W 160 181 202 223 224 265 Block 3 0 2 0 0 0 0 -1 1' Cover (0) 1 1 19' 0 0 0 0 8 28' Error (a) 3 1 1 0 0 1 0 -3 0 Density (D) 4 3* 31' 0 0 0 0 28 62* O D 4 0 1 0 0 1* 0 -1 1 Error (b) 24 6 7 1 0 2 0 -12 4* Error 40 9 11 1 0 2 0 -19 4 Total 79 20* 72* 2 0 6' 0 0 100 Block 3 0 1 1 0 0 1 -3 0 Cover 1 1 0 20' 0 0 0 23 44' Error (a) 3 0 0 1 0 0 0 0 1 Density 4 0 0 14* 0 0 14' 17 45' O D 4 0 0 1 0 1 3* -4 0 Error (b) 24 1 1 5 0 1 8 -13 3 Error 40 2 3 8 2 2 9 -19 7 Total 79 4 5 50' 2 4 35' 0 100 Block 3 0 0 0 1 0 3* -2 2' Cover 1 1' 1* 25' 0 0 1* 15 43' Error (a) 3 1' 0 1 0 0 0 -2 0 Density 4 0 1 7' 0 0 20' 16 , 44' C-D 4 0 0 1 0 0 1' -1 1 Error (b) 24 2 2 5 1 0 4 -11 3 Error 40 2' 3 10 2 0 5 -15 7 Total 79 6' 7' 49' 4 0 34' 0 100 Block 3 1 0 3 0 0 2* -5 1 Cover 1 5' 0 18 0 0 0 11 34* Error (a) 3 1 0 1 0 0 0 0 2 Density 4 1 0 17' 0 0 5* 25 48* C-D 4 1 0 1 0 0 2 -2 2 Error (b) 24 4 1 7 1 0 4 -12 5 Error 40 7 1 9 1 0 7 -17 8 Total 79 20' 2 56' 2 0 20' 0 100 Block 3 0 0 1 0 0 1 -2 0 Cover 1 0 0 14' 0 0 0 1 15' Error (a) 3 0 0 0 0 0 1 -1 0 Density 4 0 0 8' 0 0 24' 32 64' C-D 4 0 0 1 0 0 3 -3 1 Error (b) 24 1 0 4 0 0 10 -5 10 Error 40 1 1 8 1 0 21 -22 10 Total 79 2 1 36' 1 0 60' 0 100 Block 3 0 0 1 0 0 1 -1 1 Cover 1 0 0 5' 0 0 1 5 11' Error (a) 3 0 0 1 0 0 2 -1 2 Density 4 0 0 6' 0 0 21' 21 48' C-D 4 0 0 2 0 0 1 -1 2 Error (b) 24 0 1 7 1 0 21 -11 18 Error 40 0 2 10 0 1 17 -12 18 Total 79 0 3 32' 1 1 63* 0 100 z Cells to the right of the df column are percentages of the total sum of squares for W. Asterisks denote significance at p=0.05(*). 0 denotes less than 0.5% of the total sum of squares. - 9 7 -consistently greater than the effects of row covers. The contribution of population density to yield variation ranged from 44% to 64% (Tables 23 and 24). TDP analyses of variation in yield per plant indicated that density was not a major contributor to yield variation until the fourth harvest (Tables 15 and 16). Analysis per unit land area and per plant showed a steady decline after the third harvest in the contribution of row covers to variation in yield. 4.3.9 Effect of treatments on the contribution of yield components to  total shoot dry matter yield per unit land area: combined harvest data All components except WFA/VV significantly contributed to total yield in forward TDP (Table 25). The greatest contributor was NN (88%). Backward TDP showed a direct influence of NN and its influence on chronologically later components (Table 26). The contribution of WAA/F in the backward analysis was through its relationship with chronologically earlier components. The greatest influence on total yield was harvest date. Population density was a greater contributor to total yield variation than were row covers. - 98 -Table 25. TDP of shoot dry weight per land area: forward analysis of combined harvest data. 2 Source df Yield components Sum of Yield NN LA/NN W U L A WF7WV W/WF NN Products W Blocks 3 0 0 0 0 0 0 0 0 Covers (C) 1 5* 0 0 0 0 0 -1 4* Error (a) 3 0 0 0 0 0 0 0 0 Density (D) 4 12* 0 0 0 0 0 -5 7* C-D 4 0 0 0 0 0 0 0 0 Error (b) 24 0 0 0 0 0 0 0 0 Harvest date (H) 5 63* 1* 1* 1* 0 0 19 85* OH 5 2* 1* 0 0 0 0 -2 1* D«H 20 1* 0 0 0 0 0 0 1* O O H 20 0 0 0 0 0 0 .0 0 Error (c) 150 2 1 1 1 0 0 -4 1 Error 240 3 2 1 1* 0 1 -6 1 Total 279 88* 5* 3' 3* 0 1* 0 100 2 Cells to the right of the df column are percentages of the total sum of squares for W. Total sums of squares are 921, 53, 15, 10, 5443, 17 and 1701 respectively. Asterisks denote significance at p=0.05(*). 0 denotes less than 0.5% of the total sum of squares. - 99 -Table 26. TDP of shoot dry weight per land area: backward analysis of combined harvest data. z Source df Yield components Sum of Yield W/WF WF/WV WV/WL WL/LA LA/NN N N Products W Blocks (B) 3 0 0 0 0 0 0 0 0 Covers (C) 1 r 0 0 0 0 r 2 4* Error (a) 3 0 0 0 0 0 0 0 0 Density 4 0 0 0 0 0 6* 1 7* OD 4 0 0 0 0 0 0 0 0 Error (b) 24 1 0 0 0 0 0 -1 0 Harvest date (H) 5 48* 10* 0 0 0 4* 23 85' OH 5 2* 0 0 0 0 1* -2 1* OH 20 1 0 0 0 0 2' -2 1' O O H 20 0 0 0 0 0 0 0 0 Error (c) 150 5 1 1 0 0 3* -9 1 Error 240 6 2 1 0 0 4 -12 1 Total 479 64* 13* 2* 0 0 21* 0 100 2 Cells to the right of the df column are percentages of the total sum of squares for W. Total sums of squares are 14 056, 40, 7, 16, 31, 361 and 1701 respectively. Asterisks denote significance at p=0.05(*). 0 denotes less than 0.5% of the total sum of squares. - 100 -4.4 Allometric relationships 4.4.1 Allometric relationships between fruit fresh weight and  morphological characteristics, as affected by row cover, plant population density  and date of fruit harvest Tables 27 and 28 indicate the parameter estimates determined from the best subset multiple regressions. Treatments can affect ln(y) (fruit weight) indirectly through changes in the allometric exponent, B, or the allometric coefficient, a, as well as directly through the residual in equation (8). The former is evidenced by significant parameter estimates for B1 7 while the latter two effects, by significant values for y1 7. Direct allometric relationships, independent of treatment effects, are assessed through B0. The biological significance of the regression constant, In(oc'), since it is strongly dependent on the scale of measurement of y, is not clear. Within each model, the relative contribution of each term can be evaluated by the size and sign of the standard partial regression coefficients. R 2 values of best subset regression models for width, area and perimeter were above 0.45 (Table 27). In most cases, parameter estimates were significant. The models showed mostly positive significant effects of row covers on the allometric exponent relating ln(y) and ln(z) (B3) (Table 28). Except for the relationship between ln(y) and fruit width, lobe number, or area, this was the largest, and hence the most important term, in the regression. The best subset - 101 -Table 27. Regression coefficients and other statistics for best subset multiple regression models of the allometric relationships between fruit fresh weight and morphological fruit characteristics: as affected by row cover, plant population density and harvest date. Potential Parameter2 Variate  independent variable Width Length Lobe No. Area Perimeter Base Apex Intercept In(a') 27.1* 11.4* 4.09* -20.5* 38.0* ln(z) Po -0.234 0.553* -0.414 t ln(z) P, 4.64-10-3* 2.07X10-3* -1.17X10"3* 3.90X10"* 5.11X10"* D ln(z) P2 2.57X10"* 1.51X10"* C ln(z) Pa 0.166* 0.203* -5.86X1 G2* 0.113* 0.143* tD ln(z) P< -1.04* tC ln(z) P5 -1.55* CD ln(z) P« tCD ln(z) P, ln(t) YI -3.83* 6.27* •6.11* ln(D) Y2 1.20* -5.03X102* ln(C) -3.50X102* In(tD) 7* -2.52X102* In(tC) Ys 0.115* -1.40* In(CD) Y6 1.03* In(tCD) Y7 -1.25* 5.22* 0.207* 7.03* 1.49X10-4* 2.95X10" 4.12X102* 2.04X10* -0.407* -9.06X102' 1.92X10'2* -4.95X10"2 -0.145* Statistics Mallows CP 6.34 Residual mean square 0.0236 R2 0.475 d.f. 6,2192 4.93 4.82 6.47 0.358 0.0434 0.0195 0.205 0.0354 0.567 5,2193 5,2184 6,2192 6.17 5.64 4.46 0.0203 0.0356 0.0416 0.549 0.209 0.0745 6,2191 5,2193 5,2178 2 parameters according to equation (8). z designates variate; t, Julian harvest date; D, density in plants m"2; C, row cover treatment. Asterisks denote p=0.05. - 102 -Table 28. Standard partial regression coefficients for best subset multiple regression models of the allometric relationship between fruit fresh weight and morphological fruit characteristics: as affected by row cover, plant population density and harvest date. Potential Parameter" Variate  independent variable Width Length Lobe No. Area Perimeter Base Apex Intercept In(a') 128 53.8 19.3 -96.5 180 26.6 33.2 ln(z) Po -0.162 0.561 -0.403 0.312 t ln(z) P, 1.29 0.526 -0.314 1.58 1.86 0.136 D ln(z) P2 0.140 0.111 0.0990 C ln(z) P3 2.92 3.74 -0.323 2.43 3.35 0.441 0.260 tD ln(z) P< -3.09 tC ln(z) P, -3.60 CD ln(z) P« tCD ln(z) Pr ln(t) Yt -0.908 1.49 -1.45 -0.096 ln(D) Y 2 3.54 -0.149 ln(C) Y3 -0.213 In(tD) Y* -0.103 -0.075 -0.146 -0.0570 In(tC) Y5 0.267 -3.25 -0.337 In(CD) Ye 3.88 In(tCD) Y7 -4.68 z parameters according to equation (8). z designates variate; t, Julian harvest date; D, density in plants m'2; C, row cover treatment. - 103 -regression models rarely included terms involving plant population density or interactions of treatments with date of harvest, as components of the allometric exponent. Date of harvest was a minor contributor in some of the models. Only lobe number and base depth showed a direct significant allometric relationship with ln(y) (B0). Terms containing y were generally negative and did not appear as frequently in the models as those including B. Plant population density (y2) showed a strong positive contribution in the relationship between ln(y) and width, however, its contribution in other models was either small or nonexistent. 4.4.2 Allometric relationships among fruit fresh weight and  morphological characteristics as affected by row cover, plant population density  and fruit location Models using fruit location as an additional source of variation had similar R 2 values as those using date of harvest (Table 29). Row cover effects on allometric exponent (B3) were significant in every model, but were not the dominant contributors to the regressions (Table 30). The exception was the relationship between ln(y) and perimeter. Components of the allometric exponent involving plant population density or fruit location were not frequently present in the models. However, the relationship between ln(y) and length showed a strong positive contribution from B 4 but an equally strong negative - 104 -Table 29. Regression coefficients and other statistics for best subset multiple regression models of the allometric relationships between fruit fresh weight and morphological characteristics of fruit: as affected by row cover, plant population density and fruit location. Potential Parameter Variate  independent variable Width Length Lobe No. Area Perimeter Base Apex Intercept ln(z) t ln(z) D ln(z) C ln(z) tD ln(z) tC ln(z) CD ln(z) tCD ln(z) ln(t) ln(D) ln(C) In(tD) In(tC) In(CD) In(tCD) In(a') Pc P, P2 P3 P< Ps Pe Ti Y2 Y3 Y« Ys Ye Y7 -0.774* 1.13* 2.4X10"* 0.165* 1.18* 1.58X102* -1.24* 3.05* 0.178* 0.207* 1.65* -1.61' -1.69* 4.54* 0.288* -1.82X102* 8.37* 0.566* 2.93X10"* 4.87X10'2* 9.68X10'2' -0.851* -0.893* 2.48X10'2* 0.101* -3.15* -1.23* -3.68X10"* 1.17X10"* 0.152* 4.64X102* -0.102* -1.38* 4.10* 0.246* 3.11X10" 2.15X10'2* -5.11X102 4.64* 4.76X102* 7.12X10"* 2.19X10"* -0.169* 0.150* -8.92X10'2 0.879* Statistics Mallows CP Residual mean square R2 d.f. 6.24 0.0238 0.472 6,2174 4.90 0.0364 0.191 5,2175 4.64 0.0433 0.0379 5,2166 6.59 0.0197 0.562 6,2174 8.00 0.0205 0.544 7,2172 6.22 0.0357 0.207 4,2176 6.63 0.0417 0.0732 6,2159 z parameters according to equation (8). z designates variate; t, fruit location; D, density in plants m"2; C, row cover treatment. Asterisks denote p=0.05. - 105 -Table 30. Standard partial regression coefficients for best subset multiple regression models of the allometric relationship between fruit fresh weight and morphological characteristics of fruit: as affected by row cover, plant population density and fruit location. Potential Parameter* Variate  independent variable Width Length Lobe No. Area Perimeter Base Apex Intercept In(a') -3.65 14.4 21.4 39.5 -14.9 19.3 21.9 ln(z) Po 0.529 0.124 0.293 0.552 0.587 0.370 0.170 t ln(z) P, -0.176 0.111 -0.153 -0.161 D ln(z) P* 0.132 0.128 0.105 C ln(z) P3 2.90 3.82 -0.268 2.08 3.56 0.230 0.279 tD ln(z) P« 6.58 0.185 tC ln(z) P5 -6.22 CD ln(z) Pe -0.385 tCD ln(z) PN ln(t) Yi -2.33 ln(D) Y2 3.49 4.99 -2.64 -0.498 ln(C) y3 -3.25 In(tD) Y« 0.0630 -0.099 0.596 In(tC) Ys 0.387 -0.345 In(CD) Ye -4.70 -0.193 In(tCD) Y? 4.23 z parameters according to equation (8). z designates variate; t, fruit location; D, density in plants nrf2; C, row cover treatment. - 106 -contribution from y5. Regression models consistently showed direct allometric relationships that were independent of treatments (significant B 0 terms). Terms including y, 7 generally occurred less frequently in the regression models than those including B0 7. Plant population density (y2) was generally a strong contributor when present in the models. 4.4.3 Allometric relationships between plant dry weight and plant  components as affected by row covers and plant population density The best subset regression models resulted in values of R2>0.88 (Table 31). Parameter estimates, with few exceptions, were significant. Except for the relationship between ln(y) (plant weight) and leaf area, the most dominant contributor in each regression was y2 (Table 32). In each case, the effect on y2 was balanced by an opposing effect of almost equal magnitude on B2. Population density affected the relationship between ln(y) and all variates, excluding leaf weight, via y, (Table 32). Potential effects of the interactions between row covers and plant population density were never included in the best subset models. Direct allometric relationships (B0) were evident in every modeK . - 107 -Table 31. Regression coefficients and other statistics for best subset multiple regression models of the allometric relationship between plant dry weight and plant components: as affected by row cover and plant population density. Potential Parameter Variate  independent Vegetative Root Stem Leaf Leaf Fruit Fruit variable Plant wt. weight weight weight area weight number Intercept In(a') 2.24* 4.43* 2.95* 2.72* 5.75* 2.27* 4.26* ln(z) Po 0.906* 0.716* 0.871* 0.897* 0.613* 0.695' 0.546* D ln(z) P, 0.0200 0.0231' C ln(z) P 2 -0.198* -0.191* -0.216* -0.150* -0.143* -0.138 DC ln(z) P 3 ln(D) Y -0.122* -0.435* -0.176* -0.0793 -0.165* -0.318* ln(C) I2 1.95* 1.20* 1.87* 1.31* 0.295* 1.27* In(DC) Y3 Statistics Mallows CP 4.06 6.00 4.01 3.30 6.00 3.48 4.00 Residual mean square 0.0249 0.0339 0.027 0.0219 0.0282 0.0128 0.0461 R2 0.885 0.845 0.875 0.897 0.871 0.939 0.787 d.f. 4,75 5,74 4,75 3,76 5,74 2,77 4,75 z parameters according to equation (9). z designates variate; D, density in plants m"2; C, row cover treatment. Asterisks denote p=0.05. - 108 -Table 32. Standard partial regression coefficients for best subset multiple regression models of the allometric relationship between plant dry weight and plant components: as affected by row cover and plant population density. Potential Parameter Variate  independent Vegetative Root Stem Leaf Leaf Fruit Fruit variable plant wt. weight weight weight area weight number Intercept In(a') 4.94 ln(z) Po 0.953 D ln(z) P, C ln(z) P. -1.87 DC ln(z) P 3 ln(D) Y, -0.205 ln(C) Y2 2.17 In(DC) Ti 9.77 6.50 6.00 0.601 0.908 1.09 0.300 -0.966 -1.83 -1.10 -0.729 -0.296 1.33 2.08 1.46 12.7 5.01 9.40 0.682 0.807 0.451 0.255 -0.318 -1.03 -0.133 -0.277 -0.533 0.328 1.41 z parameters according to equation (9). z designates variate; D, density in plants m"2; C, row cover treatment. - 109 -4.4.4 Allometric relationships between shoot dry weight and plant  components as affected by row cover, plant population density and growth Squared multiple correlations of the best subset regression models were consistently greater than 0.94 (Table 33). Parameter estimates in all models were, with few exceptions, significant. The effects of population density on the allometric exponent (p\>) were significant in all relationships except those between ln(y) and leaf dry weight or leaf area. Row covers significantly affected the allometric exponent fl33) in all best subset regressions, except in relationships between ln(y) and vegetative shoot dry weight or leaf dry weight. The most important contributors to the regression models defining the relationship between ln(y) (shoot dry weight) and plant height, vegetative shoot dry weight and leaf dry weight, were (30 and $ : (Table 34). The relationships between ln(y) and leaf area and node number were dominated by terms containing B0 and y v - 110 -Table 33. Regression coefficients and other statistics for best subset multiple regression model of the allometric relationship between shoot dry weight and plant components: as affected by row cover, plant population density and growth. Potential Parameter Variate  independent Vegetative Leaf Leaf Node Fruit Fruit variable shoot wt. weight area number weight number Intercept In(ct') -9.70* -13.0* ln(z) Po 0.725* 0.489* t ln(z) P, 1.16X10"3' 2.14X10' D ln(z) P 2 -2.43X10"* C ln(z) P 3 tD ln(z) P< tC ln(z) P 5 CD ln(z) P* tCD ln(z) % ln(t) Yi 1.90* 2.63* ln(D) Y2 -0.194* ln(C) T& 0.108* In(tD) Y< In(tC) Y5 In(CD) Ye 0.207* In(tCD) Y7 Statistics Mallows CP 6.13 6.05 Residual mean square 0.0139 0.0207 R 0.996 0.994 d.f. 5,474 5,474 -23.3* -22.4* -19.9* -29.6 0.434* 0.458* 0.279* 0.674 2.13X10"* 1.36X10°* -8.83X10" -8.12X10"* -7.19X10"* -1.04X10'2 -2.33X102* 2.93X10'2* -2.26X10'2* -6.53X102 0.230* -0.234 -7.45X102* 5.09' • 4.06* 3.66* -5.06 4.40* 6.09 -3.94* 6.65X102 7.98 6.98 6.87 8.00 0.0318 0.0714 0.0832 0.0738 0.991 0.979 0.951 0.958 6,473 6,473 6,352 7,359 z parameters according to equation (9). 2 designates variate; t, Julian harvest date; D, density in plants m"2; C, row cover treatment Asterisks denote p=0.05. -111 -Table 34. Standard partial regression coefficients for best subset multiple regression models of the allometric relationships between shoot dry weight and plant components: as affected by row cover, plant population and growth. Potential Parameter1 Variate  independent Vegetative Leaf Leaf Number Fruit Fruit variable shoot weight weight area of nodes weight number Intercept In (a') -5.27 -7.06 -12.7 -12.2 -15.4 -22.6 ln(z) Po 0.595 0.372 0.326 0.326 0.643 0.772 t ln(z) Pi 0.249 0.402 0.259 0.295 -0.255 D ln(z) p2 -0.0150 -0.074 -0.092 -0.087 C ln(z) Ps 0.033 0.084 -0.110 -0.193 tD ln(z) p« tC ln(z) p5 0.066 CD ln(z) p6 -0.162 tCD ln(z) P 7 -0.037 ln(t) Yi 0.173 0.240 0.465 0.371 ln(D) Y2 -0.080 2.11 ln(C) Y3 0.0290 In(tD) Y4 In(tC) Y5 1.73 2.36 In(CD) Ye 0.102 -2.74 In(tCD) Y7 0.033 z parameters according to equation (9). z designates variate; t, Julian harvest date; D, density in plants m'2; C, cover treatment - 112 -CHAPTER V DISCUSSION 5.1 Temperature and humidity effects The presence of row covers elevated air and soil temperatures, as has been reported in previous studies (Bonanno and Lamont, Jr., 1987; Maurer and Frey, 1987). Continuous air temperatures above 32C, or short-duration temperatures of 40 to 50C, can affect pollination and fruit set (Gerber et al., 1989). In the present study, plants under row covers were often exposed to short periods of potentially deleterious temperatures, but were not exposed to continual temperatures that would be detrimental to fruit set. Relative humidity between 55 and 80% has been reported to not affect fruit set, although flower abortion occurred above 95% (Baer and Smeets, 1978). In the present study, mean relative humidity was 70% and did not differ between treatments. Relative humidity above 95% has also been shown to increase plant growth (Baer and Smeets, 1978). In this study, plants under row covers were exposed to a higher relative humidity than uncovered plants during the early morning hours when light, and consequently photosynthesis, was low. Humidity, in both row cover treatments, was greater than 95% only on days with precipitation. Humidity effects on reproductive or fruit yield did not likely differ between row cover treatments. - 113-5.2 ANOVA and repeated measurements Littell (1989) warned against univariate analysis of experiments comprised of repeated measurements. In the growth study, pairs of plants for each destructive harvest were randomized within a subplot; however, the experimental unit was the entire subplot and not individual plants. Because of this, the individual plants constituted repeated measurements on the same experimental units. Measurements must meet the Huynh-Feldt condition (H-F) (sets of orthogonal contrasts must have equal variances and zero covariances) in order for the univariate analysis of variance to be valid (Littell, 1989). If these conditions are not met, tests involving time and time interactions will generally result in inflated Type I error rates (Littell, 1989). Analytical approaches which avoid problems associated with repeated measurements include multivariate analysis, adjustment to the univariate P values, analysis of data from fitted response curves (Littell, 1989), separate analysis for each harvest, and pooled (averaged) dates (personal communication, G. Eaton,, 1990). The multivariate techniques (step-wise multiple regression and best subset multiple regression) used to analyze the data from the growth study avoided the problems assocated with repeated measurements. F tests for the main effects of row cover or population density are valid; however, inflated F-values for time and time interactions may have resulted if measurements did not meet the H-F condition. - 114 -5.3 Yield-plant population density relationships 5.3.1 Fruit characteristics The results of this study indicate environmental influences on many fruit characteristics. Other studies have shown genotypic effects on locule number, and apex and base shape (McArdle and Bouwkamp, 1983). Genotype-environment interactions most likely control many aspects of fruit shape, and provide an area for further research. Fruit length was not affected by treatments, although previous studies have indicated an effect of low night temperature (18 to 20C) on this characteristic (Rylski, 1973). In the present study, night temperature (represented by minimum temperature) did not differ between row cover treatments and, throughout the growing season, was never above 18C. The index of two-dimensional shape (perimeter^area) was a better indicator of treatment influences than the width:length shape index. 5.3.2 Yield response Row covers enhanced early and total fruit yield, as has been reported in previous studies (Maurer and Frey, 1987; Mohd Khir et al., 1987). Increasing plant population densities resulted in decreasing fruit yield per plant but increasing fruit yield per land area. This response has also been reported for bell peppers (Batal and Smittle, 1981; Stoffella and Bryan, 1988). Row cover - 115 -treatments influenced total plant dry weight but not the proportion of dry weights of plant components. Dry weights of all plant variables, however, decreased in response to increasing plant interference. Nonlinear regressions fit to Bleasdale's model defined yield response to plant interference. Parameter estimates varied in response to treatments and provided a means to assess the effects of plant population density and row cover. The fitted models could also be used to predict yield response, within the limits of the population densities tested. Jolliffe (1988) demonstrated, using Bleasdale's model, the effects of variation in 8 when coefficients a and b were held constant and equal. Bleasdale's model describes an asymptotic yield-population density relationship when 9 = 1, and a parabolic relationship when 9 < 1 (Willey and Heath, 1969). Theta was always greater than one in models concerning total fruit fresh weight and vegetative yield, but was less than one in regressions with early fruit yield from uncovered treatments. The former would be indicative of a asymptotic relationship while the latter would suggest a parabolic one. Holliday (1960) proposed that asymptotic relationships described total or vegetative yields. In the present study, reproductive and vegetative crop yield generally increased with increasing population densities. Thus, higher population densities than those tested might have resulted in a clear asymptotic or parabolic yield response. - 116 -Chapman (1981) attempted to fit yield graded according to a specific criterion but did not find a mathematical model that was consistent in fitting all graded yields and which also predicted biologically meaningful yields. In the present study, satisfactory regressions using Bleasdale's model were not obtained for culled and undersized graded fruit. These grades represented a relatively small proportion of total fruit yield and, as such, the difficulty in fitting was not unexpected. Willey and Heath (1969) also concluded that the relationship between graded yield and population density would probably remain empirical. 5.4 Yield component analysis Yield component analysis, using two dimensional partitioning, enabled a concise interpretation of the effects of treatments on yield components and the contribution of yield components to yield variation. In this study, the analysis of variance performed on a per plant basis showed the importance of row covers in the initial stages of growth, and the increasing importance of plant population density as growth proceeded, on yield components and on yield (fruit and total plant). Plant population density had a greater effect throughout the study period on total yield per unit land area than on yield per plant. Data from the individual harvests were also combined and harvest date was used as an additional source of variation in the ANOVA. The analysis using combined data did not contribute to an understanding of the changes in - 117-treatment effects or yield component contributions over the duration of the study, but it determined the overall treatment and yield component effects. The large contribution of harvest date (i.e. growth) to yield variation was not unexpected. The interactions of harvest date with treatment effects are of greater interest, but the combined analysis failed to isolate the timing of these effects. This could be determined through further partitioning of the sums of squares associated with harvest date. Similar studies analyzed the harvests separately (Jolliffe et al., 1989). Direct component effects on yield variation could only be determined if a component was the last to be entered into the forward or backward regression. At any other location in a forward regression, a component could also affect yield indirectly by influencing chronologically later components. Similarly, components entered into a backward regression at a step prior to the last step could have an indirect effect on yield through a relationship with chronologically earlier components. Simple correlations (Bennett, 1977) and simple linear regressions (Bowen, 1983) have been used to determine the simple relationship between yield and yield components. The contributions of yield components to both fruit and total yield per unit land area or per plant, were similar. The most important contributors to fruit yield variation were fruit weight as a proportion of total plant weight (WF/W) and the number of nodes (NN). NN was the most important contributor to variation in total plant yield. Previous studies have reported that plant population density did - 118 -not affect the branch number of pepper plants (Ahmed, 1984; Stoffella and Bryan, 1988). However, longitudinal branch growth terminates in a flower, and subsequent shoot development arises from two or three axils under the flower (Somas, 1984); an increase in node number is indicative of an increase in branch number and in an increase in the capacity for greater fruit production. TDP analysis of yield component and treatment effects was applied to yield per unit land area and to yield per plant. The analysis showed that fruit yield per plant and total plant yield were directly related to the number of nodes and inversely related to plant population density; the number of nodes and yield per plant increased in response to decreasing plant population density (Figs. 2 and 3). This analysis furthers our understanding of plant population density relationships. The components, however, must also be examined for crop management. Reproductive and vegetative yield per unit land area increased with increasing plant population density; the greatest yield per unit area was therefore obtained from plants which produced the smallest yield per plant. 5.5 Allometric relationships The best subset regressions provided a precise assessment of the effects of treatments on fruit and plant proportions. Significant regression coefficients for S1 7 are indicative of a curvilinear allometric relationships. A linear relationship exists when only 6 0 is significant (Jolliffe et al., 1988). Regressions pertaining to fruit morphology showed that (5 was consistently influenced by treatments. Row - 119 -covers influenced all variates through p whereas the influence of plant population density was less frequent. The response to time was assessed in three ways: one through the harvest date of the fruit; secondly, through the node location of the harvested fruit; and thirdly, via ln(z) which changed during growth. To some extent, each is reflective of the other due to the acropetal nature of flower and subsequent fruit development. Fruit set, however, occurs initially on the main shoot and then on the lateral branches (Kato and Tanaka, 1971). Harvest date may provide a more accurate reflection of environmental conditions during fruit development than fruit location. Moreover, it is possible that the correlation of fruit location and harvest date would decrease with decreasing plant population density as plants at low population densities showed more lateral branch development than plants at high densities. The correlation coefficients from the best subset regressions, using either harvest date or as fruit location as time, were similar. Harvest date, however, influenced p in more regressions than did fruit location. Genetic studies of pepper fruit shape indicate the potential influence of environmental factors on the genes controlling the shape of the apex and base, as well as the length of the fruit (McArdle and Bouwkamp, 1983; Rylski,1973). Fruit width and locule number are considered to be quantitatively inherited (McArdle and Bouwkamp, 1983). Some studies, however, have reported temperature effects on locule number (Mohd Khir et al., 1987). The best subset multiple regressions used in the present study established quantitative allometric - 120 -relationships between fruit weight and either apex, base, length, width or locule number that were influenced by treatments. The treatments, particularly row cover, affected allometric relationships directly via p 0 and through p. Allometric relationships were established between total plant weight and plant components. Stoffella and Bryan (1988) using ANOVA, reported lower shoor.root ratios at higher population densities using. The best subset multiple regression procedure enabled a concise interpretation of allometric relationships including those between the plant and its roots. This method showed direct allometric relationships independent of treatments, and indirect effects of population density through the allometric coefficient. The best subset multiple regressions also determined that allometric relationships defining plant morphology changed as growth proceeded and were also affected by row cover and plant population density. The changes were reflected through the allometric exponent and the allometric coefficient. In most cases, the greatest treatment influence on p or y was plant age. Combining the data and including age in the regressions precludes the ability to isolate the time when treatment effects on plant allometry arose over the duration of the study. Allometry was also shown to be influenced by age and plant population in a study concerning monocultures and mixtures (Jolliffe et al., 1988). In that study, the authors commented on the relative simplicity of regressions for monocultures compared with mixtures. It was suggested this could result from a relative environmental simplicity or from differences in the degrees of freedom - 121 -available for the regressions. However, the degree of simplicity was not consistent between species or between variates for ln(y) (yield). Furthermore, regressions for monocultures were not always less complex than for mixtures. In the present study, the degrees of freedom available for regressions including age were greater than those without, but both regressions included a similar proportion of the potential independent variables. - 122-C H A P T E R VI S U M M A R Y O F G R O W T H A N D Y I E L D O F B E L L P E P P E R S The application of row covers enhanced early and total fruit yield per plant and per land area. Vegetative plant yield was also greater from row cover treatments. Increasing plant population densities resulted in decreasing yield per plant, but increasing yield per land area. Furthermore, the effect of row covers on yield was greater at low population densities than high. Nonlinear regressions using Bleasdale's simplified model adequately described the effects of row cover and of competitive plant interference. The exceptions were graded reproductive yields that were proportionally small. Yield-response was never clearly described by either an asymptotic or a parabolic curve. The number of nodes was the most important contributor to variation in reproductive and total plant yield. Fruit weight as a proportion of total plant yield was also a major contributor to reproductive yield. TDP showed the importance of row covers in the initial stages of growth and the increasing importance of plant population density, as growth proceeded, on yield components and on reproductive and total plant yield. Allometric relationships defining plant morphology changed during growth and were also affected by row covers and plant population density. The changes were reflected through the allometric exponent and the allometric coefficient. Future work should include an analysis at each harvest, as was done using TDP - 123 -analysis of yield components. This would isolate the time when changes in plant allometry occurred. Fruit shape characteristics were affected by row cover and population density treatments. Although the differences in mean values were small, the results confirmed environmental influences on traits that genetic studies suggested were most likely controlled by genetic-environmental interactions. Changes in fruit allometry were affected, directly and through the allometric exponent and allometric constant, by treatments and by fruit location or time of harvest. The application of row covers and high plant population densities resulted in increases in horticultural yield that have significant implications for producers of bell peppers. This is the first known study to use Bleasdale's mathematical model to quantify the reproductive and vegetative response of the crop to increasing plant population densities. Also, it is the first report to use a multiple regression technique to evaluate the effects of row covers and plant population density, on fruit and plant allometry. Yield component analysis furthered our understanding of plant growth relations of bell peppers. - 124 -REFERENCES Ahmed, M.K. 1984. Optimum plant spacing and nitrogen fertilization of sweet pepper in the Sudan Gezira. Acta Horticulturae 143:305-310. Anderson, W . C , W.A. Haglund, G.W. Eaton and J . Fraser. 1986. Sequential yield component analysis of processing peas. HortScience 21:103-105. Baer, J . and L. Smeets. 1978. Effect of relative humidity on fruit set and seed set in pepper {Capsicum annuum L). Neth. J. Agr. Sci. 26:59-63. Bakker, J.C. 1989a. The effects of air humidity on flowering, fruit set, seed set and fruit growth of glasshouse sweet pepper (Capsicum annuum L ) . Scientia Hortic, 40:1-8. Bakker, J.C. 1989b. The effects of day and night humidity on growth and fruit production of sweet pepper (Capsicum annuum L ) . J. Hortic. Sci., 64:41-46. Batal, K.M. and D.A. Smittle. 1981. Response of bell pepper to irrigation, nitrogen, and plant population. J. Amer. Soc. Hort. Sci. 106:259-262. Baumann, T.E. and G.W. Eaton. 1986. Competition among berries on the cranberry upright. J. Amer. Soc. Hort Sci. 111:869-872. Bennett, J.P., M.W. Adams and C. Burga. 1977. Pod yield component variation and intercorrelation in Phaseolus vulgaris L. as affected by planting density. Crop Science 17:73-75. Bishop, R.L ed. 1988. Observer's handbook. The Royal Astronomical Society of Canada. The University of Toronto Press. Ontario. 212 pp. Bleasdale, J.K.A. 1966. Plant growth and crop yield. Annals of Applied Biology 57:173-182. - 125 -Bleasdale, J.K.A. 1967. The relationship between the weight of a plant part and total weight as affected by plant density. J. Hort. Sci. 42:51-58. Bleasdale, J.K. and J.A. Nelder. 1960. Plant population and crop yield. Nature 188:342. Bonanno, A.R. and W.J. Lamont, Jr. 1987. Effect of polyethylene mulches, irrigation method, and row covers on soil and air temperature and yield of muskmelon. J. Amer. Soc. Hort. Sci. 112:735-738. Bowen, P.A. 1983. Growth and yield relations in highbush blueberry. M.Sc. Thesis. University of British Columbia. 71 pp. Bowen, P.A. 1983. The effect of oxygen fumigation of sawdust medium on the yield and yield components of greenhouse cucumbers. Scientia Hortic. 19:131-136. Bowen, P.A. and G.W. Eaton. 1983. Yield component analysis of winter damage and flower buds in highbush blueberry (Vaccinium corymbosum). Scientia Hortic. 19:279-286. Charles, W.B., L. Amirouche and L. Malaoui. 1979. Seedlessness in Capsicum annuum L var. longum DC. (Sendt.). J. Amer. Soc. Hort. Sc. 54:159-161. Chapman, E.A. 1981. Methods of analysing yield from trials in which the produce is graded according to diameter. J. Agr. Sci. Camb. 97:55-68. Cochran, H.L 1936. Some factors influencing growth and fruit setting in the pepper (Capsicum frutescens L ) . N.Y. Agr. Exp. Sta. Mem. No. 190. Cselotei, L. 1957. (Questions of water utilization by some vegetable crops.) Kert. Szol. Fdisk. Eavk. 19:(3) 1:3-85. Cited in A. Somos. 1984. The Paprika. Akademiai Kiado. Budapest. 302 pp. - 126 -Daniel, C. and F.S. Wood. 1971. Fitting Equations to Data: Computer Analysis of Multifactorial Data for Scientists and Engineeers. Wiley-lnterscience, New York. 342 pp. Dempsey, W.H. and Boynton, J.E. 1965. Effect of seed number on tomato size and maturity. Proc. Amer. Soc. Hort. Sci. 86:575-581. Dixon, W.J. 1985. BMDP Statistical Software, University of California, Berkeley. 725 pp. Dorland, R.E. and F.W. Went. 1947. Plant growth under controlled conditions. VIII. Growth and fruiting of the chile pepper (Capsicum annuum). Amer. J. Botany 34:393-401. Eaton, G.W. 1987. Yield components in several apple clones. Fruit Varieties J. 41(2):73-79. Eaton, G.W., P.A. Bowen and P.A. Jolliffe. 1983. Two-dimensional partitioning of yield variation. HortScience 21:1052-1053. Eaton, G.W. and T.R. Kyte. 1978. Yield component analysis in the cranberry. J. Amer. Soc. Hort. Sci. 103:578-583. Eaton, G.W., A.Y. Shawa and P.A. Bowen. 1983. Productivity of individual cranberry uprights in Washington and British Columbia. Scientia Hortic. 20:179-184. Erwin, A.T. 1932. The peppers. Iowa Agr. Exp. Sta. Ames. Bull. No. 293. pp. 120-152. Everett, P.H. 1984. Pepper production as influenced by plant spacing and nitrogen-potassium rates. Citrus and Vegetable Magazine. December pp. 42-47. Filius, I. 1967. (Effect of night air temperature and soil heating in paprika grown under glass.) Kert. Szol. Foisk. Kdzl. 32:99-104. Cited in A. Somos. 1984. The Paprika. Akad6miai Kiado. Budapest. 302 pp. - 127 -Fraser, J . and G.W. Eaton. 1983. Applications of yield component analysis to crop research. Field Crop Abstracts 36(10):787-797. Freeman, J.A., G.W. Eaton, T.E. Baumann, H.A. Daubeny and A. Dale. 1989. Primocane removal enhances yield components of raspberries. J. Amer. Soc. Hort. Sci. 114(1):6-9. Gerber, J.M., W.E. Splittstoesser and G. Choi. 1989. A heat unit system for predicting optimum row tunnel removal time for bell peppers. Scientia Hortic. 40:99-104. Gerald, C.J. and G. Chambers. 1967. Effect of reflective coatings on soil temperatures, soil moisture, and the establishment of fall bell pepper. Agron. J. 59:293-296. Ghaleb Adel Hamadeh. 1967. The effect of environment, nitrogen nutrition and flower sprays on the fruiting of sweet peppers (Capsicum annuum L ) . Cornell Univ. Diss. Abstr. Ann Arbor Mich. Sect. B. 28:1307. Cited in A. Somos. 1984. The Paprika. Akademiai Kiad6. Budapest. 302 pp. Gould, S.J. 1966. Allometry and size in ontogeny and phylogeny. Biol. Rev. 41:587-640. Gulyas, S., E. Takacs, I. Dobos, and I. Horvath. 1970 Wirkung der Spektralzusammensetzung des Lichts auf die Gewebestruktur der Capsicum-, Fagopyrum-, Phaseolus- and Wc/a-Arten. Acta Biol. 16:51-61. Cited in A. Somos. 1984. The Paprika. Akademiai Kiado. Budapest. 302 pp. Hartmann, H.T., W.J. Flocker and A.M. Kofranek. 1981. Plant Science; Growth, Development and Utilization of Cultivated Plants. Prentice-Hall, Inc. New Jersey. 676 pp. Heiser, Jr., C B . and P.G. Smith. 1953. The cultivated Capsicum peppers. Economic Botany 7:214-227. - 128 -Herath, H.M.E. and G.W. Eaton. 1981. Yield component analysis in relation to climatic models in bush beans (Phaseolus vulgaris L ) . J. Nat. Agr. Soc. Ceylon 17-18:1-10. Holliday, R. 1960. Plant population and crop yield. Nature 186:22-24. Huxley, D.M., V.C. Brink and G.W. Eaton. 1979. Seed yield components in white clover. Can. J. Plant Sci. 59:713-715. Huxley, J.S. 1932. Problems of relative growth. London: Methuen. Cited in R.J. Smith. 1980. Rethinking allometry. J. Theor. Biol. 87:97-111. Jolliffe, P.A. 1988. Evaluating the effects of competitive interference on plant performance. J. Theor. Bio. 130:447-459. Jolliffe, P.A. and W.H. Courtney. 1984. Plant growth analysis: additive and multiplicative components of growth. Annals of Botany 54:243-254. Jolliffe, P.A., G.W. Eaton and J . Lovett Doust. 1982. Sequential analysis of plant growth. New Phytol. 92:287-296. Jolliffe, P.A., G.W. Eaton and M.V Potdar. 1988. Plant growth analysis: allometry, growth and interference in orchardgrass and timothy. Annals of Botany 62:31-42. Jolliffe, P.A., A.J.P. Tarimo and G.W. Eaton. 1990. Plant growth analysis: growth and yield component responses to population density in forage maize. Annals of Botany 65:137-149. Kato, T. and M. Tanaka. 1971. Studies on the fruit setting and development of sweet peppers. I. Fruiting behavior. J. Jap. Soc. Hort. Sci. 40:359-366. Kiss, A. 1969. (Flower structure investigations in the international paprika collection of Tapioszele.) Agrobotanika 11:97-104. Cited in A. Somos. 1984. The Paprika. Akad6miai Kiadd. Budapest. 302 pp. - 129 -Kohn, P . C , H.C. Wien and A.J. Pratt. 1983. Effect of film material, ventilation, and mulch on tunnel temperatures, pepper fruit set and yield. Abstr. HortScience 18(4) :589. Kormos, J . and J . Kormos, (-ne). 1956. (Paprika types of determinate growth.) Novdnytermele's 5:1-10. Cited in A. Somos. 1984. The Paprika. Akademiai Kiad6. Budapest. 302 pp. Linnaeus, C. 1753. Species plantarum. Holmiae. Cited in A. Somos. 1984. The Paprika. Akademiai Kiado. Budapest. 302 pp. Littell, R.C. 1989. Statistical analysis of experiments with repeated measurements. HortScience 24:37-40. Lovett Doust, J . , L. Lovett Doust and G.W. Eaton. 1983. Sequential yield component analysis and dry matter allocation in beans, (Phaseolus vulgaris L.). Amer. J. Botany 70:1063-1070. Maurer, A.R. and B.M. Frey. 1987. Response of bell peppers to row covers-1986 results. Agri. Canada. Agassiz Research Station. Tech. Report No. 31. 17 pp. McArdle, R.N. 1984. The relationship of number and position of seeds to fruit size and curvature in Capsicum annuum L. J. Agri. Sci., Camb. 102:509-510. McArdle R.N. and J.C. Bouwkamp. 1983. Inheritance of several fruit characters in Capsicum annuum L. J. of Heredity 74:125-127. McArthur, D.A.J, and G.W. Eaton. 1988a. Cranberry growth and yield response to fertilizer and paclobutrazol. Scientia Hortic, 38:131-146. McArthur, D.A.J, and G.W. Eaton. 1988b. Strawberry yield response to fertilizer, paclobutrazol and chlormequat. Scientia Hortic, 34:33-45. Mead, R. 1979. Competition experiments. Biometrics 35:41-54. - 130 -Metwally, A.M., A.Z. El-Abdin and S. Kastour. 1982. Effect of population density on the yield and quality of pepper. Agri. Res. Review 60(3):165-158. Mohd Khir, I., J.M. Gerber and W.E. Splittstoesser. 1987. Row tunnel effects of growth, yield and fruit quality of bell pepper. Proc. Agr. Plastics Cong. 20:152-158. Morris, E.C. and P.J. Myerscough. 1987. Allometric effects on plant interference. Annals of Botany 59:629-633. Nelder, J.A. 1966. Inverse polynomials, a useful group of multi-factor response functions. Biometrics 22:128-141. Neilson, B.V. and G.W. Eaton. 1983. Effects of boron nutrition upon strawberry yield components. HortScience 18:932-934. O'Sullivan, J . and J.K. Muehmer. 1978. Growing peppers. Ministry of Agric. and Food, Ontario. No. 78-026. 3 pp. Perkkio, J . 1985. The relationship between growth and allometry. J. Theor. Biol. 113:81-87. Polowick, P.L. and V.K. Sawhney. 1985. Temperature effects on male fertility and flower and fruit development in Capsicum annuum L. Scientia Hortic. 25:117-127. Porter, W.C. and W.W. Etzel. 1982. Effects of aluminum-painted and black polyethylene mulches on bell pepper, Capsicum annuum L. HortScience 17:942-943. Pratt, A.J. , P.C. Kohm, and H.C. Wien. 1981. Plastic mulches and plant growing tunnels and some of their effects on temperature control, water conservation and yields of pepper in Arizona and Upstate New York. Proc. Natl. Agr. Plastics Cong. 16:66-78. - 131 -Quagliotti, L , G. Lepori, and P.G. Bigotti. 1974. Responses to solar radiation by two varieties of peppers {Capsicum annuum L ) . Korean J. Breeding 6:29-33. Rylski, I. 1973. Effect of night temperature on shape and size of sweet pepper (Capsicum annuum L ) . J. Amer. Soc. Hort. Sci. 98:149-152. Salter, P.J., D.J. Andrews and J.M. Akehurst. 1984. The effects of plant density, spatial arrangement and sowing date on yield and head characteristics of a new form of broccoli. J. Hort. Sc. 59:79-85. Salter, P.J., I.E. Currah and J.R. Fellows. 1980. Further studies on the effects of plant density, spacial arrangement and time of harvest on yield and root size in carrots. J. Agr. Sci. Camb. 94:465-478. Shawa, A., G.W. Eaton and P.A. Bowen. 1981. Cranberry yield components in Washington and British Columbia. J. Amer. Soc. Hort. Sci. 106:474-477. Shifriss, C , J . Zacks and A. Goldman. 1989. Some notes on the association between fruit dimensions and fruit weight in Capsicum annuum L. Euphytica 43:275-277. Shinozaki, K. and T. Kira. 1956. Intra-specific competition amoung higher plants. VII. Logistic theory of the C-D effect. J. of the Institute of Polytechnics, Osaka 7:35-72. Sinnott, E.W. 1935. Evidence for the existence of genes controlling shape. Genetics 20:12-21. Somos, A. 1984. The Paprika. Akad6miai Kiado. Budapest. 302 pp. Somos, A. and Zs. Sovany. 1964. (Water household in some paprika varieties.) Kert. Szdl. Fdisk. Kozl. 28:1. fasc. 1:49-63. Cited in A. Somos. 1984. The Paprika. Akademiai Kiadd. Budapest. 302 pp. Stanhill, G., 1977a. Allometric growth studies of the carrot crop. I. Effects of plant development and cultivar. Annals of Botany 41:533-540. - 132 -Stanhill, G., 1977b. Allometric growth studies of the carrot crop. II. Effects of cultural practices and climatic environment. Annals of Botany 41:541 -551. Stoffella, P.J. and H.H. Bryan. 1988. Plant population influences growth and yields of bell pepper. J. Amer. Soc. Hort. Sci. 113:835-839. Stoffella, P.J., B.J. Williams, H.H. Bryan, M. Sherry and I. Stough. 1984. Influence of plant population on fruit yield and size of bell peppers. Proc. Fla. State Hort. Soc. 97:143-145. Strik, B.C. and J.T.A. Proctor. 1988. Yield comoponent analysis of strawberry genotypes differing in productivity. J. Amer. Soc. Hort. Sci. 113:124-129. Watkinson, A.R. 1984. Yield-density relationships: the influence of resource availability on growth and self thinning in populations of Vulpia fasciculata. Annals of Botany 53:469-482. Wells, O.S. 1967. The effect of night temperature on fruit set of the pepper (Capsicum annuum). Diss. Abstr. Ann. Arbor Mich. Sect. B. 27:4206. Cited in A. Somos. 1984. The Paprika. Akademiai KiadO. Budapest. 302 pp. Willey, R.W. and S.B. Heath. 1969. The quantitative relationships between plant populations and crop yield. Advances in Agronomy 21:281 -321. Wolfe, D.W., J . Wyland, L.D. Albright and S. Novak. 1986. Effect of row covers on microclimate and yield of tomato and cucumber. Proc. Natl. Agr. Plastics Cong. 19:35-50. Appendix I. Soil block media. 0.113 m 3 fluffed peat 0.0566 m 3 vermiculite 4.3 kg Bentonite 283.5 g hydrated lime 283.5 g dolomite lime 11 ml wetting agent 66.5 ml minor element solution 600 g osmocote (14-14-14) water to moisten Appendix II. Means of morphological characteristics of fruit. Source' Number of lobes Width (W, mm) Length (L, mm) W:L ratio Base depth(mm) Apex deptn(mm) Area (X103) (m2) Perimeter (mm) (Perimeter)2: Area ratio Cover (C) no cover 3.2 74 96 0.79 15.6 15.5 4.82 324 22.4 cover 3.5 76 96 0.81 14.1 15.9 4.96 321 21.3 Significance * ** NS NS * NS NS * * Density (D, plant m2) 74 0.76 14.1 11.1 3.3 99 13.8 4.89 319 21.3 5.56 3.3 74 97 0.78 14.5 14.4 4.76 319 21.9 2.78 3.3 75 96 ' 0.80 - 14.3 . 15.6 4.80 321 22.0 1.85 3.5 76 95 • 0.81 15.4 17.1 4.92 327 22.3 1.39 3.4 75 96 0.80 14.8 15.6 4.99 321 21.2 Significance L* NS NS L* L* L*Q* NS NS D" C-D C D , cc cc cc c 2D 2 cc Significance 3.2 3.0 3.3 3.3 3.2 3.3 3.5 3.4 3.6 3.5 NS 74 75 74 75 73 73 73 76 77 77 L* 99 100 95 96 95 98 95 96 95 96 NS 0.76 0.77 0.80 0.80 0.78 0.77 0.79 0.80 0.82 0.82 NS 14.9 15.9 15.2 16.1 15.5 13.5 13.3 13.6 14.8 14.2 NS 13.1 14.9 15.8 16.4 15.3 14.4 14.1 15.4 17.6 15.8 NS 4.87 4.83 4.70 4.86 4.84 4.90 4.70 4.89 4.97 5.12 NS 325 326 321 328 322 315 314 322 325 322 NS 22.1 22.6 22.5 22.8 21.9 20.6 21.4 21.6 21.8 20.7 NS 1 C. and C , denote not covered and covered respectively. D 1 P D,, D 3, D 4, D 5 denote population densities of 11.1, 5.56, 2.78, 1.85 and 1.39 plants m"2 respectively Asterisks denote significance at p=0.05(*), p=0.01(**) and p=0.001 (***). NS denotes not significant. L, Q and D are linear, quadratic and deviation responses respectively. - 135 -Appendix Means of early fruit yield: kg per plant (fresh weight). Source Marketable (kg) Undersized (kg) Cull (kg) Total (kg) Cover (C) no cover row cover Significance 0.547 1.05 Density (D, plants m"2) 11.2 5.56 2.78 1.85 1.39 Significance C«D C A C A C A C A C A C A C A C A C A C A Significance 0.286 0.506 0.792 1.14 1.27 0.160 0.440 0.608 0.826 0.699 0.413 0.572 0.976 1.46 1.85 C A * * * 0.037 0.154 0.0508 0.0802 0.125 0.0992 1.22 L* 0.018 0.0223 0.0471 0.0352 0.0646 0.0834 0.138 0.203 0.163 0.181 NS 0.103 0.189 NS 0.0636 0.131 0.135 0.166 0.236 0.046 0.128 0.0774 0.114 0.151 0.0807 0.134 0.193 0.219 0.320 NS 0.687 1.40 0.401 0.717 1.05 1.41 1.63 0.225 0.590 0.733 0.974 0.915 0.577 0.844 1.37 1.84 2.35 C A * * * C, and C 2 denote not covered and covered respectively. D, , D 2, D 3, D„, D 5 denote population densities of 11.1, 5.56, 2.78, 1.85 and 1.39 plants m 2 respectively. Asterisks denote significance at p=0.05(*), p=0.01(**) and p=0.001(* N S and L denote not significant and linear response respectively. - 136 -Appendix IV. Means of total fruit yield: kg per plant (fresh weight). Source Marketable Undersized Cull Total (kg) (kg) (kg) (kg) Cover (C) no cover 1.57 0.098 0.167 1.83 row cover 1.88 0.258 0.229 2.36 Significance NS * NS * Density (D, plants m"2) 11.1 0.673 0.112 0.0755 0.860 5.56 1.06 0.158 0.174 1.39 2.78 1.71 0.236 0.196 2.14 1.85 2.53 0.188 0.224 2.95 1.39 2.64 0.196 0.322 3.16 Significance L***Q**C* NS |_***Q** C*D C A 0.644 0.0819 0.0578 0.783 C A 1.01 0.0530 0.174 1.23 c , D 3 1.53 0.126 0.146 1.80 C A 2.42 0.0925 0.189 2.70 C A 2.23 0.137 0.270 2.64 C A 0.702 0.141 0.0933 0.937 C A 1.10 0.263 0.175 1.54 C A 1.88 0.346 0.247 2.48 C A 2.65 0.284 0.258 3.19 C A 3.04 0.256 0.374 3.67 Significance D L* NS NS D L** C, and C 2 denote not covered and covered respectively. D 1 ( D 2, D 3, D 4, D 5 denote population densities of 11.1, 5.56, 2.78, 1.85 and 1.39 plants m 2 respectively. Asterisks denote significance at p=0.05(*), p=0.01(**) and p=0.001 (***). L, Q, and C are linear, quadratic and cubic responses respectively. NS denotes not significant. - 137 -Appendix V. Means of early* fruit yield: kg per m 2 (fresh weight). Source Marketable Undersized Cull Total (kg) (kg) (kg) (kg) Cover (C) no cover 1.68 0.122 0.372 2.18 row cover 3.15 0.562 0.605 4.32 Significance * ** NS ** Density (D, plants m"2) 11.1 3.18 0.563 0.706 4.45 5.56 2.81 0.446 0.728 3.98 2.78 2.20 0.348 0.376 2.93 1.85 2.12 0.183 0.308 2.61 1.39 1.77 0.171 0.328 2.27 Significance L* C»D Significance C*D Q * O D L * * * NS NS y initial 28 days of harvest. C, and C 2 denote not covered and covered respectively. Asterisks denote significance at p=0.05(*), p=0.01(**) and p=0.001 (***). NS, L and Q denote not significant, linear and quadratic responses respectively. - 138 -Appendix VI. Means of fruit yield: kg per m 2 (fresh weight). Source Marketable Undersized Cull Total (kg) (kg) (kg) (kg) Cover (C) no cover 4.92 0.383 0.548 5.85 row cover 5.66 0.975 0.738 7.37 Significance NS NS NS * Density (D, plants m"2) 11.1 7.47 1.24 0.838 9.55 5.56 5.87 0.879 0.969 7.72 2.78 4.75 0.656 0.545 5.95 1.85 4.69 0.348 0.414 5.45 1.39 3.67 0.273 0.448 4.39 Significance L*** L***Q* C-D Significance NS NS NS NS C, and C 2 denote not covered and covered respectively Asterisks denote significance at p=0.05(*), p=0.01(**) and p=0.001 (***). NS, L and Q denote not significant, linear and quadratic responses respectively - 139 -Appendix VII. Means of plant components; yield density study. Total dry Vegetative Root dry Stem dry Leaf dry Leaf Source 2 weight plant dry weight weight weight area (g) weight (g) (g) (g) (g) (m2) Cover (C) no cover 215 101 10.5 54.7 35.7 0.633 row cover 338 123 12.1 70.2 40.6 0.746 Significance ** N S NS NS NS NS Density (D, plant m"2) 11.1 173 60.5 7.17 36.2 17.1 0.352 5.56 230 81.3 9.65 46.4 25.2 0.495 2.78 272 103 11.3 57.0 34.6 0.623 1.85 334 154 14.1 85.4 54.8 0.983 1.39 375 161 14.4 87.2 58.9 0.995 Significance L*** C-D Significance NS NS NS NS NS NS Asterisks denote significance at p=0.05(*), p=0.01(**) and p=0.001 (***). NS and L denote not significant and linear trend respectively. 

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.831.1-0098134/manifest

Comment

Related Items