POPULATION DYNAMICS OF THE SYCAMORE (Drepanosiphum platanoidis Schrank)

by

Frances Antoinette Wade, B.Sc. (Hons.), M.Sc.

A thesis submitted for the degree of Doctor of Philosophy of the University of London, and the Diploma of Imperial College of Science, Technology and Medicine.

Department of Biology, Imperial College at Silwood Park, Ascot, Berkshire, SL5 7PY, U.K.

August 1999

1 THESIS ABSTRACT

Populations of the sycamore aphid Drepanosiphum platanoidis Schrank (Homoptera: ) have been shown to undergo regular two-year cycles. It is thought this phenomenon is caused by an inverse seasonal relationship in abundance operating between spring and autumn of each year. It has been hypothesised that the underlying mechanism of this process is due to a plant factor, intra-specific competition between , or a combination of the two. This thesis examines the population dynamics and the life-history characteristics of D. platanoidis, with an emphasis on elucidating the factors involved in driving the dynamics of the aphid population, especially the role of bottom-up forces.

Manipulating host plant quality with different levels of aphids in the early part of the year, showed that there was a contrast in aphid performance (e.g. duration of nymphal development, reproductive duration and output) between the first (spring) and the third (autumn) aphid generations. This indicated that aphid infestation history had the capacity to modify host plant nutritional quality through the year. However, generalist predators were not key regulators of aphid abundance during the year, while the specialist parasitoids showed a tightly bound relationship to its prey.

The effect of a fungal endophyte infecting the host plant generally showed a neutral effect on post-aestivation aphid dynamics and the degree of parasitism in autumn. However, density of oviparae was suppressed on endophyte infected plant material. This, together with premature leaf fall of endophyte leaves, has the capacity to curtail the autumn aphid population, which may result in low aphid density in the following spring.

In autumn, third generation aphids gave rise to a female-biased sex ratio. Oviparae were always laid first in the progeny sequence followed, after a brief reproductive pause, by males. Nymphal development time was longer in oviparae than in males, and was accelerated on good host plant quality. Improved nutritional status of the host induced the production of a greater proportion of males, reduced development time of sexual morphs, and increased their survival. In addition, the potential fecundity of oviparae was a function of maternal body weight and host quality.

2 Oviparae mainly laid their eggs in clusters on the rough bark surfaces of tree trunks and branches, away from the more exposed smooth surfaces of twigs and terminal buds. Fierce intra-specific competition among oviparae for optimal overwintering sites was seen at high density. Egg mortality was a function of time in hibernation, with the greatest mortality rates occurring in mid to late winter. March temperature was shown to be important in determining the time of egg hatch and bud burst. The preceding autumnal and spring aphid densities were positively related.

These findings suggest that while host plant food quality was a key factor in determining aphid abundance, density-independent seasonal and meterological variables were the over-riding components driving the population dynamics of D. platanoidis.

3 DECLARATION

The work described in this thesis is entirely my own, except for the inputs from the following people in the chapter below:

Chapter 5 - Aspects of overwintering in the sycamore aphid.

Dr. Simon Leather permitted the use of field data on sycamore aphid eggs collected during the winters of 1982 to 1991 at Bush N.R.S., Edinburgh, Scotland. He also allowed data on the number of sycamore aphid eggs, time of fundatrix appearance and bud burst recorded during the winters of 1992 to 1997 at Silwood Park, Berkshire, England to be used. In addition, Mr. James Culverhouse provided computerised data on the aerial March temperatures at Silwood Park during 1993 to 1998. However, all the data handling and analyses were performed by me.

Signed C-I2 A A c's-c'L-12- Frances Antoinette Wade (Student)

Confirmed Dr. Simon Leather (Supervisor)

Prof. Charles Godfray (Supervisor)

4 To A.M.D.G., my Parents, the late Professor Kinmonth and staff of St. Thomas'.

"It is not the critic who counts; not the man who points out how the strongman stumbles, or uhere the doer ofdeeds could actually haw done better. The credit belongs to the man who is actual& in the arena, whose face is marred by dust and sweat and blow uho strives valiantly; who errs and comes short again and again; Because there is not on&about error and shortcomings; but who does actually strive to do the dee4 ub knows great enthusiasm, the great devotion, who spends himself in a worthy cause; who at best knows in the end high achie-cement and who, at the uvrst, if he fails, at least he fails while daring greatly. So that his place shall never be with those cold and timid souls who know neither victory nor defeat." - Taken from "For him read her" by Theodore Roose-celt.

5 ACKNOWLEDGEMENTS

I would like to thank my two supervisors, Dr. Simon R. Leather and Prof. H.C.J. Godfray, for their technical support and rapid feedback during this project, and to the Natural Environment Research Council (N.E.R.C.) for funding this studentship (Grant No. GT4/95/1 80/T).

My gratitude is expressed to the `Aphid/Drosphila Research Group' for informal discussions, and especially Christine Muller, Lex Kraaijeveld and Donald Quicke for stimulating feedback. Recognition should be made towards Mike Cammell, Mick Crawley and particularly Jim Hardie for their inspiration and help. The comments from Nigel Barlow of AgResearch New Zealand were also greatly appreciated. Thanks also go to Frank van Veen, Darren Greene, Mark Fellows and Milton Mendota for their snippets of advice and moreover their entertaining company over the last three years.

My sincere appreciation is owed to Dominik Wodarz whose patient help with manuscript reading and special friendship has been invaluable. To Nicholas Jarraud for his constant support and tea break chats throughout this Ph.D. — it shall never be forgotten. To Patricia Walker, Kingsley Shadish and Mike Haynes for their friendship, kindness and constructive advice.

To Aunty Pauline and Aunty Anne in reception for their stalwart encouragement and sense of humour. To Carole Collins and Anne-Marie Sarney for always being willing to help quickly with administrative duties and in any other way they can. To Peter Muller, Sim Adamson and Martin Couchman for their computer support and technical advice. Not forgetting Martin Parsons for his constant stream of witticisms while I was student warden at Silwood Park, and Sharon Ives also in the accommodation office offering great cheer.

Last, but by no means least, special thanks to go to my parents who have given me constant support and love throughout this project.

6 CONTENTS

Abstract 2 Declaration 4 Acknowledgements 6 Contents 7 List of Figures 9 List of Plates 13 List of Tables 14 Chapter 1: General Introduction Introduction 15 Dynamics of phytophagous forest populations 15 Theory of population regulation 17 Dynamic behaviour of phytophagous forest insect populations 19 Factors affecting population dynamics of phytophagous forest 20 The study organism 27 Overall aims of the thesis 28 Thesis outline 28 Chapter 2: Population dynamics and life history responses of Drepanosiphurn platanoidis (Schrank), on saplings with different initial aphid densities Abstract 30 Introduction 31 Methods & Materials 34 Results 39 Discussion 51 Conclusion 59 Chapter 3: Does endophytic fungal infection affect the dynamics of the sycamore-aphid-parasitoid interaction? Abstract 60 Introduction 61 Natural history 64 Methods & Materials 65 Results 70 Discussion 79 Conclusion 85

7 Chapter 4: Reproductive sequencing and the sex ratio of the sycamore aphid Abstract 87 Introduction 88 Methods & Materials 92 Results 97 Discussion 111 Conclusion 118 Chapter 5: Aspects of overwintering in the sycamore aphid Abstract 119 Introduction 120 Methods & Materials 124 Results 129 Discussion 148 Conclusion 154 Chapter 6: General Discussion 156 References 164 Appendix: Preliminary pesticide screening 183

8 LIST OF FIGURES

Fig. 2-1. The effects of season on the natural aphid population trends of Drepanosiphum platanoidis between autumn 1996 and spring 1997 40

Fig. 2-2. Correlation between the spring and the autumn aphid populations 40 Fig. 2-3. The effect of four aphid load treatments on cumulative aphid densities in spring and autumn 1997 41 Fig. 2-4. The effect of four aphid density treatments in spring 1997 on the 1997autumnal populations of Drepanosiphum platanoidis 42 Fig. 2-5 i: The numerical response of anthocorids to spring aphid densities 44 Fig. 2-5 ii: The numerical response of anthocorids to autumnal aphid densities 44 Fig. 2-5 iii: The numerical response of coccinellids to spring aphid densities 44

Fig. 2-5 iv. The numerical response of coccinellids to autumnal aphid densities 44 Fig. 2-5 v: The numerical response of to spring aphid densities 45

Fig. 5 vi: The numerical response of arachnids to autumn aphid densities 45 Fig. 2-5 vii: The numerical response of syrphid larvae to spring aphid densities 45

Fig. 2-5 viii: The numerical response of syrphid larvae to autumn aphid densities 45

Fig. 2-5 ix: The numerical response of total primary parasitoids to spring aphid densities 45

Fig. 2-5 x: The numerical response of total primary parasitoids to autumn aphid densities 45 Fig. 2-6 i: Mean duration of the first, second and third instar nymphal development times for generations two and three of Drepanosiphum platanoidis in relation to aphid density treatment 47 Fig. 2-6 ii: Mean duration of fourth instar nymphal development time in three parthenogenetic aphid generations in relation to aphid density treatment 47 Fig. 2-7 i: Pre-larviposition period in three parthenogenetic aphid generations in relation to aphid density treatment 48

Fig. 2-7 ii. Reproductive duration in three parthenogenetic aphid generations in relation to aphid density treatment 49

Fig. 2-7 iii. Post-reproductive duration in three parthenogenetic aphid generations in relation to aphid density treatment 49 Fig. 2-8. Total mean fecundity in three parthenogenetic aphid generations in relation to aphid density treatment 50 Fig. 2-9: Rate of leaf fall on forty saplings in relation to spring aphid density 51

Fig. 3-1: The six treatments used in the endophyte experiment: three treatments in the presence And three in the absence of the parasitoid Aphelinus flavus 67 Fig. 3-2: The design of the enclosed endophyte experiment: arranging six treatments (replicated three times) according to a Latin block system within the field insectary 68 Fig. 3-3: Mean time of appearance of third generation Drepanosiphym platanoidis nymphs after aestivation on Acer pseudoplatanus saplings in relation to acerinum infection and/or Aphelinus flavus presence 72 Fig. 3-4: Total abundance of Drepanosiphum platanoidis on Acer pseudoplatanus saplings in relation to Rhytisma acerinum infection and/or Aphelinus flavus presence 72

9 Fig. 3-5: Abundance of early (instars 1-3) and late (instar 4) nymphs of Drepanosiphum platanoidis on Acer pseudoplatanus saplings in relation to Rhytisma acerinum infection and/or Aphelinus flavus presence 73

Fig. 3-6: Abundance of adult parthenogenetic virginoparae of Drepanosiphum platanoidis on Acer pseudoplatanus saplings in relation to Rhytisma acerinum infection and/or Aphelinus flavus presence 74

Fig. 3-7: Abundance of Drepanosiphum platanoidis adult sexuals on Acer pseudoplatanus saplings in relation to Rhytisma acerinum infection and/or Aphelinus flavus presence 74

Fig. 3-8: Number of Aphelinus flavus mummies as a proportion of available Drepanosiphum platanoidis nymphs (instars 1-3) on Acer pseudoplatanus saplings in the presence and absence of Rhytisma acerinum infection 75

Fig. 3-9: Mean time of appearance of third generation Drepanosiphum platanoidis nymphs after aestivation on Acer psuedoplatanus saplings in relation to Rhytisma acerinum infection and/or Aphelinus flavus presence within mixed culture treatments 76

Fig. 3-10: Total aphid abundance of Drepanosiphum platanoidis on Acer pseudoplatanus saplings within mixed cultures in relation to Rhytisma acerinum infection and/or Aphelinus flavus presence 76

Fig. 3-11: Abundance of early (instars 1-3) and late (instar 4) stage Drepanosiphum platanoidis nymphs within mixed cultures in relation to Rhytisma acerinum infection and/or parasitoid presence 77

Fig. 3-12: Abundance of Drepanosiphum platanoidis adult parthenogenetic virginoparae on Acer pseudoplatanus saplings within mixed cultures in relation to Rhytisma acerinum infection and/or Aphelinus flavus presence 78

Fig. 3-13: Abundance of Drepanosiphum platanoidis adult sexuals on Acer pseudoplatanus saplings within mixed cultures in relation to Rhytisma acerinum infection and/or Aphelinus flavus presence 78

Fig. 3-14: Number of Aphelinus flavus mummies as a proportion of available Drepanosiphum platanoidis nymphs (instars 1-3) of D. platanoidis within mixed cultures in the presence and absence of Rhytisma acerinum infection 79

Fig. 4-1: The average number of oviparae and males produced per day by Drepanosiphum platanoidis sexuparae on Acer pseudoplatanus saplings under natural field conditions 97

Fig. 4-2: Mean day of appearance of Drepanosiphum platanoidis oviparae on twelve Acer pseudoplatanus saplings under natural field conditions 98

Fig. 4.-3: Mean day of maternal switch occurrence in the reproductive sequence of Drepanosiphum platanoidis sexuparae on twelve Acer pseudoplatanus saplings under natural field conditions 98

Fig. 4-4: Mean day of appearance of Drepanosiphum platanoidis males on twelve Acer pseudoplatanus saplings under natural field conditions 98

Fig. 4-5: Mean maternal switch duration of Drepanosiphum platanoidis sexuparae on twelve Acer pseudoplatanus saplings under natural field conditions 99

Fig. 4-6: Mean number of ovipara and males born per sexupara of Drepanosiphum platanoidis on Acer pseudoplatanus saplings under natural field conditions 99

Fig. 4-7: Proportion of oviparae and males of Drepanosiphum platanoidis surviving to reach adulthood on Acer pseudoplatanus saplings under natural field conditions 100

Fig. 4-8: Mean development time of oviparae and males of Drepanosiphum platanoidis on Acer pseudoplatanus saplings under natural field conditions 100

Figs. 4-9 a & b: The average number of oviparae and males produced per day by Drepanosiphum platanoidis sexuparae on Acer pseudoplatanus saplings with poor nutrition (a) and good host nutrition (b) in greenhouse conditions 101

Fig. 4-10 a: The effect of host plant nutrition on the date of first appearance of Drepanosiphum platanoidis oviparae on Acer pseudoplatanus saplings in greenhouse conditions 102

Fig. 4-10 b: Mean day of appearance of Drepanosiphum platanoidis oviparae on Acer

10 pseudoplatanus saplings under poor and good nutrition in greenhouse conditions 103

Fig. 4-11 a: The effect of host plant nutrition on the maternal switch date of Drepanosiphum platanoidis sexuparae on Acer pseudoplatanus saplings in greenhouse conditions 103

Fig. 4-11 b: Mean date of maternal switch of Drepanosiphum platanoidis sexuaparae on

Acer pseudoplatanus saplings under poor and good nutrition in greenhouse conditions 103

Fig. 4-12 a: The effect of host plant nutrition on the date of first appearance of Drepanosiphum platanoidis males on Acer pseudoplatanus saplings in greenhouse conditions 104

Fig. 4-12 b: Mean day of appearance of Drepanosiphum platanoidis males on Acer pseudoplatanus saplings under poor and good nutrition in greenhouse conditions 104

Fig. 4-13 a: The effect of host plant nutrition on the maternal switch duration of Drepanosiphum platanoidis sexuparae on Acer pseudoplatanus saplings in greenhouse conditions 104

Fig. 4-13 b: Mean duration of the maternal switch of Drepanosiphum platanoidis sexuparae on Acer pseudoplatanus saplings under poor and good nutrition in greenhouse conditions 105

Fig. 4-14: Mean number of ovipara and males born per sexupara of Drepanosiphum platanoidis on Acer pseudoplatanus saplings under poor and good nutrition in greenhouse conditions 105

Fig. 4-15: Mean proportion of oviparae and males of Drepanosiphum platanoidis surviving to reach adulthood on Acer pseudoplatanus saplings under poor and good nutrition in greenhouse conditions 106

Fig. 4-16: The development time of Drepanosiphum platanoidis oviparae and males on Acer pseudoplatanus saplings under poor and good nutrition in greenhouse conditions 107

Fig. 4-17: The adult lifespan of Drepanosiphum platanoidis oviparae and males on Acer pseudoplatanus saplings under poor and good nutrition in greenhouse conditions 107

Fig. 4-18: The effect of plant nutritional status on body weight of Drepanosiphum platanoidis oviparae feeding on Acer pseudoplatanus 108

Fig. 4-19: Relationship between adult body size and the number of ovarioles within an ovipara of Drepanosiphum platanoidis on poor and good nutrient quality Acer pseudoplatanus saplings 109

Fig. 4-20: Relationship between adult body size and the number of eggs per ovariole in oviparae of Drepanosiphum platanoidis on poor and good nutrient quality Acer pseudoplatanus saplings 110

Fig. 4-21: Relationship between adult body size and the number of immature eggs as a proportion of the total eggs per ovipara of Drepanosiphum platanoidis on poor and good nutrient quality Acer pseudoplatanus saplings 1 1 1

Fig. 5-1 a: Mean abundance of Drepanosiphum platanoidis eggs on Acer pseudoplatanus trunk bark in relation to year and aspect 130

Fig. 5-1 b: Mean abundance of Drepanosiphum platanoidis eggs on Acer pseudoplatanus branches in relation to year, aspect and surface position 131

Fig. 5-1 c: Mean abundance of Drepanosiphum platanoidis eggs on Acer pseudoplatanus terminal twig sections in relation to year and aspect 131

Figs. 5-2 a: Survival of Drepanosiphum platanoidis eggs from November to February on Acer pseudoplatanus trunks in the winters of 1996 and 1997 132

Fig. 5-2 b: Survival of Drepanosiphum platanoidis eggs from November to December on Acer pseudoplatanus trunks in the winters of 1996 and 1997 132

Fig. 5-2 c: Survival of Drepanosiphum platanoidis eggs from December to January on Acer pseudoplatanus trunks in the winters of 1996 and 1997 133

Fig. 5-2 d: Survival of Drepanosiphum platanoidis eggs from January to February on Acer pseudoplatanus trunks in the winters of 1996 and 1997 133

Fig. 5-3 a: Percentage survival of Drepanosiphum platanoidis eggs from November to February

11 on Acer pseudoplatanus trunks in relation to year and aspect in the winters of 1996 and 1997 134 Fig. 5-3 b: Percentage survival of Drepanosiphum platanoidis eggs from November to December on Acer pseudoplatanus trunks in relation to year and aspect in the winters of 1996 and 1997 134 Fig. 5-3 c: Percentage survival of Drepanosiphum platanoidis eggs from December to January on Acer pseudoplatanus trunks in relation to year and aspect in the winters of 1996 and 1997 135 Fig. 5-3 d: Percentage survival of Drepanosiphum platanoidis eggs from January to February on Acer pseudoplatanus trunks in relation to year and aspect in the winters of 1996 and 1997 135 Fig. 5-4 a: Survival of Drepanosiphum platanoidis eggs from November to February on Acer pseudoplatanus branches in the winters of 1996 and 1997 136 Fig. 5-4 b: Survival of Drepanosiphum platanoidis eggs from November to December on Acer pseudoplatanus branches in the winters of 1996 and 1997 136 Fig. 5-4 c: Survival of Drepanosiphum platanoidis eggs from December to January on Acer pseudoplatanus branches in the winters of 1996 and 1997 136 Fig. 5-4 d: Survival of Drepanosiphum platanoidis eggs from January to February on Acer pseudoplatanus branches in the winters of 1996 and 1997 137 Fig. 5-5 a: Percentage survival of Drepanosiphum platanoidis eggs from November to February on Acer pseudoplatanus branches in relation to year, aspect and locale in the winters of 1996 and 1997 138

Fig. 5-5 b: Percentage survival of Drepanosiphum platanoidis eggs from November to December on Acer pseudoplatanus branches in relation to year, aspect and locale in the winters of 1996 and 1997 138

Fig. 5-5 c: Percentage survival of Drepanosiphum platanoidis eggs from December to January on Acer pseudoplatanus branches in relation to year, aspect and locale in the winters of 1996 and 1997 138 Fig. 5-5 d: Percentage survival of Drepanosiphum platanoidis eggs from January to February on Acer pseudoplatanus branches in relation to year, aspect and locale in the winters of 1996 and 1997 139 Fig. 5-6: Mean level of oviposition per ovipara on three types of bark substrate 140 Fig. 5-7 a: Total number of Drepanosiphum platanoidis eggs laid in an 8 cm2 quadrat by oviparae under low and high density treatment 141

Fig. 5-7 b: Mean number of Drepanosiphum platanoidis eggs laid by an ovipara in low and high density treatments 141 Fig. 5-8: Overwintering survivorship of Drepanosiphum platanoidis eggs on Acer pseudoplatanus in relation to two initial egg density treatments 142 Figs. 5-9 a & b: Egg counts of Drepanosiphum platanoidis on Acer pseudoplatanus in November/December of each year at Bush N.R.S., Edinburgh and Silwood Park, Berkshire 143

Figs. 5-10 a & b: Egg counts of Drepanosiphum platanoidis at time N, and N,_,1 at Bush N.R.S., Edinburgh and Silwood Park, Berkshire 144

Figs. 5-11 a & b: Between-year dynamics of Drepanosiphum platanoidis on Acer pseudoplatanus using initial number of eggs laid at the beginning of winter and the subsequent spring peak in aphid density at Bush N.R.S., Edinburgh and Silwood Park, Berkshire 145 Fig. 5-12 a. Effect of bud burst (stage 3 flush) timing on Drepanosiphum platanoidis population change over winter (between autumn and spring) at Silwood Park, Berkshire 146 Fig. 5-12 b. Impact of Acer pseudoplatanus bud burst timing on Drepanosiphum platanoidis hatch success at Silwood Park, Berkshire 146 Fig. 5-12 c. Synchrony between timing of Drepanosiphum platanoidis egg hatch and Acer pseudoplatanus stage 3 bud burst at Silwood Park, Berkshire 147

Fig. 5-13. Effect of temperature on the timing of Acer pseudoplatanus bud burst (stage 3 flush) and Drepanosiphum platanoidis fundatrices emergence at Silwood Park, Berkshire 147

12 LIST OF PLATES

Plate 2-1: Clip-caged Drepanosiphum platanoidis on an Acer pseudoplatanus Sapling 37

Plate 3-1 a: Healthy sycamore (Acer pseudoplatanus) leaf 71

Plate 3-1 b: Rhytisma acerinum infected sycamore leaf showing black tar spots 71

Plate 3-2: Field insectary at Silwood Park used for the endophyte manipulation experiment 71

Plate 3-3: The set up of an experimental treatment block containing four sycamore saplings 71

Plate 4-1: Different morphs of Drepanosiphum platanoidis on a leaf of Acer pseudoplatanus 96

13 LIST OF TABLES

Table 1-1: Response characteristics of various biotic factors to the density of phytophagous insect populations, and their effectiveness as density regulators 18

Table 2-1. Mean natural abundance of logged cumulative spring and autumn aphid populations on ten sycamore saplings over a two year monitoring period 39

Table 2-2: Effects of spring aphid density treatments on autumn aphid abundance in 1997 41

Table 2-3. Effects of aphid density and treatments on cumulative generalist and specialist predator abundance during spring and autumn 1997 43

Table 5-1: Probability of survival of Drepanosiphum platanoidis eggs on Acer pseudoplatanus trunks in relation to year, aspect and initial egg density throughout the winters of 1996 and 1997 134

Table 5-2: Probability of survival of Drepanosiphum platanoidis eggs on Acer pseudoplatanus branches in relation to year, aspect, position and initial egg density thoughout the winters of 1996 and 1997 137

Table 5-3: Percentage survival of Drepanosiphum platanoidis eggs during the whole winter period and at each monthly interval in relation to trunks and branches of Acer pseudoplatanus during 1996 and 1997 139

Table 5-4: Distribution of Drepanosiphum platanoidis eggs laid on rough bark of Acer Pseudoplatanus 140

Table 5-5: Inter-anuual trends of overwintering Drepanosiphum platanoidis populations on Acer pseudoplatanus over 10 and 6 year periods in two locations of the British Isles 143

14 Chapter 1 General Introduction

General Introduction

This thesis is concerned with the population dynamics and abundance of the sycamore aphid, Drepanosiphum platanoidis Schrank (Homoptera: Aphididae) - that is, the patterns, processes and factors involved in regulating aphid numbers and population density. Particular attention is paid towards exploring the mechanisms driving the two- year cycle of this tree-dwelling aphid, and the role of bottom-up forces on the aphid at the individual and population level.

To begin the thesis with a general review of the current state of knowledge regarding aphid biology and that of population dynamics theory is inappropriate, since these topics have been thoroughly covered in other publications (for aphids see Kennedy & Stroyan, 1959; Dixon, 1973c, 1977; 1985; Blackman, 1974; Minks & Harrewijn, 1987, 1988, 1989; Moran, 1992; Kindlmann et al., 1994; Hales et al., 1997; for population dynamics theory see Andrewartha & Birch, 1954; Nicholson, 1954; Milne, 1957; Holling, 1965; Murdoch, 1970; den Boer & Gradwell, 1971; Varley et al., 1973; May, 1974; Hassell, 1976; Anderson & May, 1978; Anderson et al., 1979; Onstad, 1988; Royama, 1992; Cappuccino & Price, 1995; Begon et al., 1996; Dempster & McLean, 1998). In addition, the two topics have been combined in studies of aphid population dynamics (e.g. Dixon, 1969, 1975a, 1979, 1990c; Chambers et al., 1985; Wellings et al., 1985; Wellings & Dixon, 1987; Kindlmann & Dixon, 1996; Sequeira & Dixon, 1997; Dixon & Kindlmann, 1998). Instead, this introductory chapter concentrates on the population dynamics, in particular cyclicity, of forest insects and the potential factors affecting their spatial and temporal abundance. Next the study organism is discussed, and finally the overall objectives stated and the individual chapters of the thesis introduced by means of short outlines. Since each chapter of the thesis was written as a manuscript for publication in scientific journals, they individually stand alone as pieces of research.

Dynamics of phytophagous forest insect populations:

Populations of organisms are never truly stable, but rise from some low density and then fall to approximately their original size. Theoretically, they may exhibit stable

15 Chapter 1 General Introduction equilibrium points, stable cyclic oscillations between two population points, stable cycles, quasi-periodic oscillations or a regime of aperiodicity (May, 1974, 1976). Hassell et al. (1976) demonstrated that the majority of natural insect populations show monotonic damping (the most stable kind of equilibrium behaviour), while complex dynamics such as damped oscillations, stable limit and aperiodic chaos are rarely found in nature. Insect pests usually exhibit gradient, cyclic, or irruptive patterns (Berryman & Stark, 1985; Furuta, 1976; Rafes, 1978); on the other hand, endemic or rare species approximate stable equilibrium (Wallner, 1987).

In this review, the population dynamics of phytophagous forest insects are focussed upon because: (i) the study organism is a deciduous tree-dwelling insect (Blackman & Eastop, 1994); (ii) sylvan pests cause economic damage (e.g. black pine aphids induce bark splitting and occasionally tree death at high infestations - Shaw, 1983) and reduced amenity value of trees (e.g. winter caterpillars cause forking and defoliation of beech - Brown, 1953); and (iii) the forest ecosystem represents a plant environment that remains relatively constant (in contrast to agricultural systems), due to its slow growth and turnover rates, which enables the investigation into the stability of rapidly changing insect populations, as well as the stabilising and destabilising forces (Berryman, 1988).

The question of how insect herbivore populations are maintained at low densities, and why they occasionally increase to extremely high densities, has intrigued both theoretical and empirical ecologists (Berryman et al., 1987). Early theoretical viewpoints were typified by the 'Great Debate', between Andrewartha & Birch and Nicholson at the 1957 Cold Spring Harbour Symposium on Quantitative Biology, that was principally concerned with the relative importance of physical (density- independence) versus biological forces (density-dependence) in regulating insect abundance (Berryman et al., 1987; Turchin & Taylor, 1992). Following this time of controversy (see Bakker, 1964), a number of long-term ecological studies were undertaken on certain important forest insects: e.g. woodwasp in Australia (Madden, 1988), spruce budworm in Canada (Hardy et al., 1983), winter moth in England (Varley et al., 1973), autumnal moth in Finland (Haukioja, 1980), sycamore aphid in Scotland (Dixon, 1990c), larch budmoth in Switzerland (Baltensweiler et al., 1977), Douglas-fir tussock moth (Brookes et al., 1978), gypsy moth (Campbell & Sloan, 1978), and several bark (Berryman, 1973, 1979, 1982) in North America. These studies yielded a large body of descriptive field based data on forest insect population fluctuations, and

16 Chapter 1 General Introduction provided an empirical foundation for an emerging theory of herbivore population regulation (e.g. Berryman, 1978a; Isaev et al., 1984).

Theory of population regulation:

Population regulation has been given different definitions even in the most recent literature. To some ecologists, regulation is 'the process whereby a population returns to its equilibrium' (Varley et al., 1973; Dempster, 1983; Sinclair, 1989). To many others however, regulation simply means long-term persistence and fluctuations within limits, with the lower limit >0' (see for example Mountford, 1988; Murdoch & Walde, 1989), a definition that essentially equates regulation with persistence. Despite these semantic difficulties, it has been acknowledged that regulation requires a feedback of density on population growth, so that any deviation (up or down) of population size away from the equilibrium density is countered by an increase or decrease in mortality, or dispersal (Myers, 1988b; Berryman et al., 1987; Dempster & McLean, 1998).

Two basic sets of variables have been recognised (Huffaker, 1958; Fretwell, 1972; Varley et al., 1973; Royama, 1977) to affect population density: (i) density-dependent variables that are part of a feedback process which respond to changes in the density of the population, and (ii) density-independent variables that set the framework within which the feedback processes operate and affect the level of population regulation. In the first case, interactions between herbivores and their host plants, predators, parasitoids and pathogens (bacterial, fungal or viral) can all potentially be involved in negative feedback regulation because they often respond to herbivore density and then feed back to negatively affect herbivore reproduction and survival. In the second case, exogenous factors such as climate, soil and landform cannot be considered as part of the regulatory process per se because they do not change in response to population density; nevertheless, they do play crucial roles in determining the level at which population regulation occurs (Berryman et al., 1987).

Density-dependent response characteristics, particularly their speed of reaction, determine the effectiveness of various biotic variables in regulating herbivore population densities (Table 1.1). Fast acting regulatory forces may be seen when vertebrate predators rapidly respond to changes in prey density by quickly switching their feeding preference to the more abundant species. While the voracious appetites of

17 Chapter 1 General Introduction vertebrate predators can render annihilating consequences when prey densities are relatively low; at high prey densities, satiated predators no longer respond to prey density changes (Holling, 1965; Hassell, 1978). Slow acting forces (i.e. delayed density-dependence — see Royama, 1977) may be exemplified by it taking more than one insect generation for plants to re-grow tissue removed by herbivores, for induced defensive phytochemicals to be produced or to be reabsorbed, for predators to reproduce (Berryman et al., 1987; Myers, 1988a). Berryman et al. (1987) differentiates between fast (first-order) and slow acting (second-order) regulatory processes using the R- function (Rt = Nt+t/Nt) or replacement curve (cf. reproductive curves of Ricker, 1954; see also Morris, 1963; Holling, 1973). This single-species model incorporates the combined feedback of all the density-dependent variables (listed in Table 1.1) to determine how reproduction and survival of individual herbivores change with the density of its population.

Table 1-1: Response characteristics of various biotic factors to the density of phytophagous insect populations, and their effectiveness as density regulators. Modified from Berryman et al. (1987) and Myers (1988b).

Density- Response Effectiveness in population regulation dependent characteristics variable Low herbivore density High herbivore density Plant. 1. Rapid induced defence. Stabilising. May suppress outbreaks. 2. Delayed induced Slow recovery keeps Foliage deterioration at high defence. density low. density may cause cycles. 3. Foliage depletion. Does not occur at low Food reduction leading to insect herbivore densities. starvation may suppress outbreaks and increase susceptibility to pathogens. 4. Tree death. Keeps density low. Slow recovery — long cycle period.

Invertebrate I. Reproductive num- Effective if adapted to Sometimes effective in termin- predators & erical responses. find hosts and reproductive ating outbreaks. May generate parasitoids. 2. High fecundity. responses are rapid. herbivore cycles, by delayed 3. Relatively short response, due to generation generation spans span being > than that of prey.

Vertebrate 1. Rapid sigmoid Can potentially keep Ineffective because of weak Predators. functional response. density low. reproductive response and/or 2. Immigration. satiation at high prey densities. 3. Learning to identify May create outbreak thres- abundant prey and holds and amplify population switch feeding growth rates. preference. 4. Large appetites.

Pathogens. 1. Rapid numerical Usually ineffective because Often very effective in term- responses. of density thresholds for inating outbreaks. Possibly 2. Infection & virulence infection and/or virulence. prolongs reduction of host often dependent on fecundity and vigour. May herbivore density. synchronise herbivore cycles. 3. Other stresses may induce infection.

18 Chapter 1 General Introduction

Dynamic behaviour of phytophagous forest insect populations:

Using the framework of population regulation theory, the dynamic behaviour of phytophagous forest insect populations may be quantitatively determined by: (a) the form (uni- or bimodal) of the R-function (see Berryman et al., 1987), (b) time-delays in the regulatory processes (see May, 1973; Berryman, 1978b; Myers, 1988b), and (c) the favourability of the environment for the insect and its regulatory agents as dictated by density-independent variables (Berryman et al., 1987). From this, three principal patterns of numerical behaviour emerge:

(i) Populations that are regulated by fast-acting negative feedback processes (i.e. showing unimodal or shallow bimodal R-functions) and are relatively stable (i.e. exhibit constrained fluctuations around their equilibrium positions). The equilibrium densities are determined by density-independent factors, which vary in time or space. For instance, in forest environments that are consistently unfavourable for the insect and/or favourable for the density-independent agents, the insect population will oscillate around a very low mean density. This kind of behaviour is perhaps typical of the majority of forest herbivores, and is exemplified by those insects that never become serious pests (Ohmart et al., 1983; Larsson & Tenow, 1980).

(ii) Populations that are regulated by delayed (second order) negative feedback processes and exhibit regular cycles of abundance in certain environments. Cycles will be amplified in environments that are more favourable for the phytophagous insect, because this will increase the slope of the R-function and the equilibrium population density (Berryman et al., 1987). The length of the feedback delay may be influenced by environmental conditions, as seems to be the case in the autumnal moth-birch system (Haukioja, 1980). In addition, density independent forces (e.g. extremely cold winters) may synchronise local populations over large areas (`Moran Effect') so that they all start cycling in unison (Chitty 1967; Berryman, 1981; Barbour, 1990; Royama, 1992, 1997; Myers, 1998). A few forest insect populations, as well as those of other herbivores (e.g. snowshoe hares (Boonstra et al., 1998)), usually occur with constant periodicity of around ten years (Keith, 1963). At least 18 species of in the north temperate zone display population cycles with 8-11 year periodicities: where insect densities increase over 5-7 years, remain high for 1-2 years, and decline over 2-3 years (Myers, 1988b, 1990, 1998). Shorter population cycles are readily seen in a

19 Chapter 1 General Introduction number of aphids on deciduous trees e.g. lime (Dixon, 1971c), pecan (Liao & Harris, 1985), sycamore and Turkey oak (Dixon, 1990e; Dixon & Kindlmann, 1998) occurring with yearly periodicity, while moderate periodicities are to be found in the large pine aphid showing a 4-5 year cycle in abundance (Kidd, 1990) similar to that found in non- insect taxa (e.g. voles with 3-4 year cycles, Batzli, 1996; sockeye salmon with 4 year cycles, Levy & Wood, 1992).

(iii) Populations that remain at low densities for long periods of time, under the influence of fast-acting negative feedback processes, but occasionally explode to very high densities due to positive feedback processes (e.g. escape from predators, Southwood & Comins, 1976) that create the unstable outbreak thresholds. Eruptive herbivores are characterised by bimodal R-functions that can have three (or more) equilibrium points. Outbreak behaviour can be triggered by a density independent variable (e.g. improved weather) and/or population eruptions spreading from local epicentres to spread over large forested regions (e.g. mountain pine (Berryman, 1982); Swaine sawfly (McLeod, 1979), Siberian Dendrolimus moth (Isaev et al., 1974); phasmatids (Readshaw, 1965)). Following eruptive outbreaks, short-time delays may cause populations to linger around their upper equilibrium before collapsing (e.g. gypsy moth (Campbell & Sloan, 1978)), whilst long-delays may generate cycles at a high density equilibrium before crashing into the lower domain of attraction. This behaviour may result from host death or a high rate of natural enemy attack during the outbreak phase (e.g. eastern spruce budworm (Morris, 1963)). Although insect epidemics can have dramatic consequences for the forest ecosystem, "oubreak species consistute a minor proportion of all phytophagous insects, and may represent the exception rather than the rule" (Morris, 1964).

Factors affecting the population dynamics of phytophagous forest insects:

From reviewing the voluminous literature on phytophagous insects from varied forest habitats, a number of abiotic (weather, sites) and biotic (host plant quality, natural enemies) factors have been identified to affect the quality of individuals and hence their population dynamics.

Weather:

20 Chapter 1 General Introduction

Uvarov (1931) and Andrewartha & Birch (1954) are perhaps the best early known proponents of weather and climate as controlling factors of insect abundance — poor weather suppresses insect populations. Graham (1939), Wellington (1952) and Greenbank (1963) recognised the potential of good weather for improving conditions for insects and developed the idea that several years of favourable weather could increase insect survival and fecundity leading to population outbreak.

Insect outbreaks have been reported to follow periods of drought (Il'inskii et al., 1975; Thomson & Shrimpton, 1984; Thomson et al., 1984; Mattson & Haack, 1987), high sunspot activity (Benkevich, 1972), decreased storminess and increased atmospheric circulation (Benkevich, 1972), or combinations of drought and excess moisture (Hain, 1979; Maceljski & Balarin, 1974; Mattson & Addy, 1975). However, severe precipitation and/or winds can be extremely detrimental to phytophagous insect populations (Dixon, 1979). Temperature can have important affects on insect populations by causing asynchrony between insect and host (Dixon, 1976a; Day, 1984; Leather et al., 1993) or predator and insect prey (Inozmetsev, 1976), mortality to overwintering stages due to density-dependent development lag (McClure, 1983), selective mortality that shifts population genetics (Campbell, 1964), or appearance of refuge areas for insect survival (Stark, 1959).

While environmental stresses (e.g. drought, temperature and precipitation fluctuations) have been recorded prior to insect outbreaks, its mode of action is open to some speculation. Current evidence suggests that climatic/weather stress may affect insects indirectly by reducing phenotypic variability (via metabolic and physiological responses) in the host plant population (Lorimer, 1980). For example, a psyllid population erupted when drought caused elevated amino acid levels in eucalyptus (White, 1969; see also White, 1974); populations of two Scandinavian geometrids erupted as a consequence of lowered levels of anti-herbivory compounds in birch foliage caused by cold summer temperatures (Niemeld, 1980; see also Rhoades, 1979); and excessive rainfall during the wet season produced new flushes of vegetative growth on widely scattered Bunchosia trees, promoting outbreaks of the Costa Rican chrysomelid Uroplata (Young, 1979). This increased homogeneity in host traits may result in rapid selection of insect phenotypes that can reproduce successfully on these plants; when this occurs, an outbreak ensues (Lorimer, 1980). Thus, it is likely that a

21 Chapter 1 General Introduction combination of environmental conditions (`Climate Release Hypothesis') and insect quality are required for population outbreaks (Myers, 1998b).

Site:

Optimal sites (`primary foci') for survival of certain insects have been associated with forests growing under adverse weather, habitat and host conditions (Chugunin, 1949; Il'inskii et al., 1975). These sites are characterised by soils that are unsuitable for a given species, sharply changing humidity conditions, or vulnerability to drought or human disturbance (Rudnev, 1972). Outbreaks of forest defoliators are usually first noted in open forests that are subject to drought, and are associated with poor growing sites such as ridgetops, upper slopes, or deep sands (Stark, 1959; Il'inskii et al., 1975; Heikkenen, 1981; Stoszek et al., 1981; Kemp & Moody, 1984). Disturbances such as fire and clear-cutting have increased the extent, frequency and severity of outbreaks associated with such sites (Blais, 1983). In addition, insect outbreaks in forests commonly occur in stands that have passed peak efficiency in biomass production (Mattson & Addy, 1975), or bog areas, which had fewer small mammal predators (Hanski & Otronen, 1985). Futhermore, the amplitude of cycles can be modified by altitude (Baltensweiler et al., 1977) and/or latitude (Haukioja, 1980).

Spatially well-defined habitats within forest ecosystems provide highly selective survival conditions for insects during inhospitable periods, and may be the potential source of future epidemics (Il'inskii et al., 1975; Berryman, 1978a). The spatial extent of a refuge (analogous to 'epicentre', 'reserve' or 'focus') varies depending upon the requirement of the insect involved: it can be limited to part or all of a tree (Raffa & Berryman, 1983), or can cover several hectares or more (Larsson & Tenow, 1984). Refuges not only provide escape from natural enemies (Rafes, 1978; Smith, 1985), but can act as repositories of highly nutritious food (see White, 1984). In fact, certain epidemic pests could persist during latent periods in these epicentres of good nutritional resources in otherwise inhospitable environments.

Host plant quality:

Active tree responses to herbivory have been proposed as important factors regulating phytophagous insect population dynamics (Rhoades, 1985; Haukioja & Neuvonen, 1987; Neuvonen et al., 1988; Karban & Myers, 1989; Haukioja, 1980, 1991; Karban &

22 Chapter 1 General Introduction

Niiho, 1995). Following herbivore attack, trees have been noted to have adverse effects on several aphids (Dixon, 1971a; Barlow & Dixon, 1980; Kidd, 1985; Lewis, 1987) and lepidoptera (Wallner & Walton, 1979; Schultz & Baldwin, 1982). The deterioration in foliar quality after herbivore damage had been linked to increases in secondary plant compounds that are detrimental to insect growth, survival and fecundity (Rhoades, 1983, 1985; Rossiter et al., 1988) or reductions in the nutrient quality, lower nitrogen and higher fibre content (Benz, 1974; Fischlin & Baltensweiler, 1979; Baltensweiler, 1984). These alterations make plants more resistant to further herbivory (i.e. induce resistance or defence — Haukioja & Honkanen, 1997) and can be classified into two forms: rapid (RIR) and delayed (DIR) induced resistance (Haukioja, 1982). The effects of RIR (i.e. stabilising negative feedback) are experienced by the same generation inflicting the damage; while the effects of DIR (i.e. destabilising negative feedback) are realised by subsequent insect generations (Haukioja & Neuvonen, 1987; Karban & Myers, 1989). The delayed recovery in foliage quality (i.e. 'relaxation time-delay' - see Berryman et al., 1987) associated with DIR could explain both the continued decline in insect numbers and quality (see 'Maternal Effects Hypothesis' — Rossiter, 1991, 1995; Ginzburg & Taneyhill, 1994) that are characteristic of phytophagous forest insect populations undergoing multi-annual cycles and outbreaks (Myers, 1988a; Haukioja, 1980; Haukioja & Honkanen, 1997). Although evidence for DIR has been questioned (see Myers, 1988b; Haukioja, 1990b and references therein), the strongest support comes from deciduous trees, particularly mountain birch and European larch, where these systems display regular cycles (Benz, 1974; Ruohomaki et al., 1992).

Paradoxically, an increase in foliage quality after herbivore damage (referred to as `resource manipulation' (Craig et al., 1986), 'induced amelioration' (Haukioja et al., 1990) or 'induced susceptibility' (Karban & Niiho, 1995) may also contribute to insect cycles and outbreaks, irrespective of whether induced susceptibility operates with or without time lags (Haukioja & Honkanen, 1997). However, studies illustrating better insect performance on previous herbivore attacked trees (e.g. Niemela et al., 1984; Roland & Myers, 1987; Williams & Myers, 1984) have in the majority of cases been plagued by non-random choice of experimental trees (e.g. Roland & Myers, 1987; Williams & Myers, 1984) and statistical problems, similar to those in early studies reporting induced resistance (see Haukioja, 1990b).

23 Chapter 1 General Introduction

Insect populations may respond to other environmental stresses apart from herbivore- induced responses. White (1974, 1978, 1984) found that insect outbreaks were associated with droughts, root damage from water-logging, nutrient poor soils, plant pathogen infection, and a variety of other stresses, which caused plants to depress their resistance and enrich their tissues with nitrogen (conditions advantageous to the survival and growth of young insects). Rhoades (1979) considered stress to be a mechanism for reducing the production of "defensive" chemicals of trees and therefore for improving foliage quality for phytophagous insects. However, if any pattern exists, stress seems to increase the concentration of at least some types of secondary chemicals (Mattson & Haack, 1987; Myers, 1988a). Wallner (1987) found that nutrient stress from poor soils and following defoliation rearranged the plant carbon/nutrient balance resulting in increased production of carbon-based allelochemicals (also see Tuomi et al., 1988).

Natural enemies:

For most phytophagous insects, enemies take many forms: from polyhedrosis viruses to parasitoids and micorsporidia to mammals (Strong et al., 1984). Evidence suggests that natural enemies may have a minor role in population dynamics of rare, endemic, and gradient insects (Rafes, 1978), but a major role in those of cyclic and eruptive pests (Berryman, 1978; Il'inskii et al., 1975; Kolomiets et al., 1972). Examples of the interactions of parasitoids, predators and pathogens with forest insects are plentiful in the literature, however the scope of this literature review permits only a few illustrative examples.

Predators

Vertebrate and predators have been implicated in the control of numerous forest insects (Dahlsten et al., 1977; Furuta, 1982; Rolling, 1959; Il'inskii et al., 1975; Kolomiets et al., 1972; Takekawa et al., 1982). However, since their rate of reproduction is limited, predators usually cannot compensate fast enough to suppress eruptions of multivoltine insects. For example, the sycamore aphid population increases so rapidly that it saturates the feeding response of its predators (Dixon, 1970c; Dixon & Russel, 1972). Also, the mountain pine beetle, an eruptive pest, is preyed upon to a higher proportion by a clerid beetle in latent than in epidemic infestations, thus suggesting that it may be important in maintaining sparse populations (Amman, 1984).

24 Chapter 1 General Introduction

In the case of the two sawflies on pine, endemic species, polyphagous pupal and cocoon predators maintained populations at equilibrium levels (Bauer, 1985).

Parasitoids

A disproportionate amount of research concerning parasitoids has been directed to endemic insects undergoing outbreaks, rather than those that maintain a steady state (Southwood, 1975). Parasitoids have a dominant influence on the birch-alder casebearer in outbreaks than on latent populations (Pschorn-Walcher, 1980), but a minor role in western spruce budworm gradations (Campbell, 1981). Parasitoids can exhibit a delayed response to their prey (e.g. larch budmoth, DeLucchi, 1982). Less vigorous prey are more parasitised than the more active larvae (Wellington, 1957). Often parasitoids congregate in high density patches (e.g. winter moth larvae, Hassell, 1982). They have been identified as important contributors to mortality in low to moderate Douglas-fir moth populations (Dahlsten et al., 1977), and in maintaining low densities of European spruce sawfly for many years (Neilson & Morris, 1964). In addition, specialist and generalist native parasitoids have a major role in preventing successful establishment of invading migrant insect herbivores (Il'inskii et al., 1975). However, the degree of parasitism may be impaired by predators and hyperparasitoids, as is the case in the sycamore system (Hamilton, 1973, 1974). The interaction of the parasitoids and the host plant has an important effect on the population dynamics of phytophagous insects (Lawton & McNeil, 1979; Price et al., 1980; Rhoades, 1983; Godfray, 1994). It is very likely that this tri-trophic relationship is more important for epidemic pests, which are better adapted to host-plant variability, than for endemic and gradient insects (Wallner, 1987).

Pathogens

The action of entomopathogens is more obvious at high population densities where more contact occurs among individuals, and the unfavourable effects of overcrowding are evident (Viktorov, 1971). Horizontal transmission of viral, bacterial, protozoan and fungal infections usually requires high host densities or host aggregating behaviour (May, 1983), since they are more likely to contact refuges of pathogens. However, transmission can occur at low host density, and thus the density dependent relationship between pathogen and host is not always the case (see Benz, 1987 for a review).

25 Chapter 1 General Introduction

Anderson & May (1980) analysed host-pathogen mortalities for 28 insects and concluded that pathogens contribute wholly or partially to the regulation of their host populations. They presented evidence that 5-12 year population cycles of temperate forest insect pests were attributable to a nucleopolyhedrosis virus (acting singly or in combination with a microsporidian protozoan) or a granulosis virus. However, Berryman (1996) considers viruses acting alone to be a destabilising factor, but when interacting with density-related parasitoids they can create cyclic population dynamics in many forest Lepidoptera. Grobler et al. (1962) found the Entomophthora to be of limited scope in the overall suppression of the woolly pine needle aphid, whereas Zelinskaya (1980) reported that four microsporidian agents are major regulators of gypsy moth populations: they increased larval mortality, egg sterility and lowered fecundity. Martignioi & Schmid (1961) found that high mortality from disease results in the selection of resistant individuals that remain when an epizootic subsides. Sublethal effects of disease may reduce size and reproductive capacity of individuals for several subsequent generations (characteristic of cyclic populations undergoing the decline phase), but the relationship can vary depending on the dose and time of infection (Perelle & Harper, 1986).

Benz (1987), in reviewing the literature, found several studies where food quality (leaf age, toughness, species of food plant) and quantity influenced the susceptibility of insects to virus; virus infection was more likely when leaf quality was poor or in short supply. White (1974) proposed that the sudden outbreak of disease in high density looper caterpillar populations could be associated with reduced nitrogen content of food following defoliation. The cyclic behaviour of the larch budmoth has also been attributed to foliage quality, with the period of outbreak set by the insect-virus dynamics and the amplitude of the outbreak cycle set by the insect-plant relationship (May, 1983). Studies on viral and microsporidian infections of the larch budmoth suggest that the wide amplitude stable-limit cycles are probably driven by long delays built into the infection and transmission process (Anderson & May, 1980). The host-pathogen system is complicated by long latent stages during which the pathogen exists in the external environment, protected from photolysis by the soil of temperate forests (see Hosteller & Bell, 1985; Kaupp & Sohi, 1985 for reviews).

From reviewing the various factors, phytophagous insect populations of forests are regulated more by heritable traits (see Wallner, 1987) and profound plant influences,

26 Chapter 1 General Introduction and to a lesser extent by natural enemies. Gradient, cyclic and eruptive insects respond quickly to host plant availability and susceptibility, and are tracked by natural enemies. Predators are most important in regulating or maintaining low insect densities, whereas entomopathogens tend to be most effective in reducing dense populations.

The study organism:

All the chapters in this thesis use Drepanosiphum platanoidis (Schrank), the sycamore aphid, as a model insect species to study population dynamics in the field. This species was chosen for five reasons. First, it is ubiquitous and therefore readily available for study in the field. Second, it has already been intensively studied, and much is known about its reproductive biology and ecology (Dixon, 1963; 1977; 1979). Third, migrating air-borne D. platanoidis undergoes intriguing cyclic behaviour, in that its annual abundance tends to peak every alternate year (Dixon, 1990c, Dixon & Kindlmann, 1998). Fourth, D. platatoidis adults are relatively long lived such that there are few generations in the year, and that all adults are alatae (except oviparous females in autumn — Dixon, 1969) with the first, second and third parthenogenetic generations not overlapping to a large extent (Chambers, 1979), thus making it easy to work with. Fifth, this species is autoecious and holocyclic (Blackman & Eastop, 1994), thereby making it a relatively uncomplicated organism. Autoecy involves the aphid living and feeding solely on a single host (Blackman, 1974; Dixon, 1973c, 1985); whereas holocycly involves the production of a generation of sexual morphs at the end of each parthenogenetic season (Moran, 1992).

Of the four species of aphid that live on sycamore (i.e. Periphyllus testudinaceus, P. acericola, Drepanosiphum platanoidis, D. acerinum), D. platanoidis is the largest and has been called the sycamore aphid (Dixon, 1985). The annual lifecycle of D. platanoidis commences with eggs hatching in spring (late March) just prior to bud break (Dixon, 1976a, 1977). The fundatrices (first generation aphids) move on to and feed actively on growing buds (Dixon, 1976a). The progeny of the first generation are larviposited upon the maturing leaves and complete their development in early summer. When adult, these second generation aphids are characterised by reproductive aestivation that may last up to 8 weeks (Dixon, 1963, 1966). Reproduction begins in late summer, and alate males and apterous oviparae (egg-laying females) first appear in the progeny of the third generation. The proportion of sexual individuals gradually increases in subsequent generations (Dixon, 1971b). After mating, the oviparae lay

27 Chapter 1 General Introduction overwintering eggs in the crevices in the bark prior to leaf fall (Dixon, 1976a). All individuals, except the males and oviparae, are parthenogenetic, viviparous and alate (Wellings, 1981). The populations are characterised by three peaks each year — spring, summer and autumn (see Dixon, 1979 - Fig. 5.2) corresponding to the first, second and sexual generations respectively. Population density varies considerably between years (Dixon, 1970b, 1990c; Dixon & Kindlmann, 1998) and seasonally within each year (i.e. spring and autumn peaks seem to be inversely related in size) (Dixon, 1970b).

Overall aims of the thesis:

Several factors have already been identified in the natural control of sycamore aphid abundance, in addition to a variety of self-regulatory intraspecific processes (Dixon, 1970a, 1979; Chambers et al., 1985; Wellings et al., 1985). These include plant quality (Dixon, 1966, 1970a, 1975a, 1979; Wellings & Dixon, 1987; Dixon et al., 1993), predators (Dixon, 1970c; Dixon & Russel, 1972) and parasitoids (Hamilton, 1973, 1974, Collins et al., 1981). However, studies examining plant-herbivore interactions in the sycamore system have mainly considered bottom-up control by examining soil nutrients or intrinsic physiological plant condition (e.g. senescence); while those using natural enemies have predominantly used laboratory conditions and/or some uncontrolled field observations. Therefore, the aims of this thesis were:

1. To determine the driving mechanism(s) for the within (i.e. between spring and autumn) and between (i.e. between autumn and spring) year dynamics, involved in the two year D. platanoidis cycle, through aphid density manipulations in the field;

2. To assess the affect of host nutritive quality on the autumnal populations of sexuparae and sexual morphs in the field and by glasshouse rearing; and

3. To address the impact of a fungal endophyte on a specialist tri-trophic interaction under field conditions.

Thesis outline:

First, the sycamore aphid is placed in perspective with other phytophagous insects found within the forest ecosystem. Chapter 1 then goes on to discuss the theory of population regulation and dynamic behaviour, particularly cyclic phenomenon,

28 Chapter 1 General Introduction occurring in forest insects. Chapter 2 looks at within-year aphid dynamics. It reports field experiments examining the effects of spring aphid density on autumnal density. This involved knocking the sycamore aphid cycle out of phase, using aphid density manipulations, from the time of egg hatch until aestivation. Chapter 3 examines the effects of a specialist plant fungus on sycamore saplings, and observing how it alters aphid-parasitoid dynamics. Chapter 4 investigates the impact of host plant quality on the production of sexual morphs. Here, the reproductive sequence of sexuparae is recorded for the first time under field conditions. In addition, the role of host plant quality is noted in relation to egg production via its effect upon oviparae size. Chapter 5 looks at between-year aphid dynamics. The survival of overwintering sycamore aphid eggs is recorded in relation to spatial distribution and initial density on the tree. The natural location of eggs are made by observation only, but the effects of egg density used experimental manipulation of oviparae in autumn. The temporal abundance of overwintering eggs, recorded in Scotland and England, is also included. In addition, the synchronisation between temperature with bud burst and egg hatch with bud burst is noted over a 6-year period. Finally, chapter 6 brings the various chapters together by discussing the findings in terms of aphid population dynamics and host plant quality.

29 Chapter 2 Population dynamics and life history responses to aphid manipulations

Population dynamics and life history responses of Drepanosiphum platanoidis (Schrank), on saplings with different initial aphid densities

ABSTRACT

1. The sycamore aphid, Drepanosiphum platanoidis (schrank), has been shown to undergo a two-year cycle. 2. In this paper a plant factor hypothesis is tested to explain this cycle. 3. Field population and cage experiments were carried out in which starting aphid (fundatrix) densities were manipulated on 40 sycamore saplings. 4. Both population dynamics and individual aphid performance were monitored in weekly census counts and daily/alternating day clip-cage observations. Here, aphid response variates (e.g. aphid development time and fecundity) were used to bioassay plant quality. 5. The numerical response of aphidophagous predators were also observed on the same 40 saplings while conducting the weekly census for assessing aphid abundance. 6. Within-year dynamics of aphids showed positive correlations between the spring and autumn populations, irrespective of the initial aphid densities. 7. During spring, caged aphids showed rapid nymphal development and high fecundity on augmented (i.e. heavy aphid load) saplings compared with uninfested saplings. By autumn, caged aphid performance was poor on heavily infested saplings compared with the lightly infested saplings. 8. Aphids kept in clip-cages during spring lent support to the plant-stress hypothesis. Positive responses to stress were manifested as an increase in insect performance (i.e. shortened insect growth/development; enhanced fecundity; increased post- parturition duration and longevity of survival) and insect abundance. They were shown to occur as a result of experimental induced stress. However, by the end of the season, it was envisaged that aphids caused deterioration in its host via injection of exogenous substances from their saliva and/or exhausted foliar resources, and damage to the vascular system causing phloem occlusion. 9. This suggested temporal modification of host plant conditions, imposed by preceding aphid generations, seems to imply that bottom-up control is important to the subsequent population dynamics of D. platanoidis. 30 Chapter 2 Population dynamics and life history responses to aphid manipulations

10. All five predator categories monitored showed positive numerical responses towards increasing aphid densities during spring, but by autumn only territorial arachnids and the specialist primary parasitoids showed a tightly bound relationship with their prey abundance. 11. The rate of leaf abscission, on the 40 saplings, was not influenced by spring aphid infestation history. 12. It was concluded that these experiments demonstrate that aphid induced changes in host plant quality do play a role on within year density dependent effects, and are complementary to the intra-specific processes demonstrated by other authors. However, while spring stressed plant conditions were found to negatively affect aphid performance at the end of the year, this plant mediated density dependence was not overcompensating and was not of itself sufficient to explain the cycle.

INTRODUCTION

Chemical changes occur in woody and herbaceous plants following insect feeding (Edwards et al., 1990). However, there has been much controversy on the importance of these changes in the host plant, and whether they lead to modified feeding conditions for phytophagous insects (Larsson, 1989). It has been suggested that certain biochemical changes in 'stressed' [e.g. drought, waterlogging, herbivore damage] plants can result in improved insect performance (White, 1974; Rhoades, 1979). White (1984) showed that plants respond with increased concentrations of soluble amino acids irrespective of the type of stress. Alstad et al. (1982) assert that insects responding to host stress with a certain change in life history performance, (e.g. increased survival, fecundity, development or growth rate) have an immense potential for altering population development, provided that the magnitude of the response is great enough (Fowler & Lawton, 1985).

The complexities involved in plant-insect interactions have not been well documented for tree-dwelling aphids (e.g. Kidd et al., 1990), despite the more numerous studies on sap-suckers on non-woody species (e.g. van Emden, 1972; Weibull, 1987). Previous controlled experiments, dealing with sap-sucking insects on woody plants, have essentially evaluated stress effects by measuring aphid densities rather than responses of individual insects (see Larsson, 1989), although Major (1990) attempts to bridge this gap by conducting mean relative growth rate and population growth experiments. However, changes in insect performance at the individual level can be counteracted at 31 Chapter 2 Population dynamics and life history responses to aphid manipulations the population level by processes such as density-dependent , migration and intra-specific competition (Begon et al., 1990). Although the studies reviewed by Larsson (1989) took into account responses at the population level, it should be emphasised that they did not take any population effects other than intra-specific ones into consideration. Until now, no experimental studies have comprehensively dealt with `plant stress' effects on the performance of the individual tree-dwelling aphids (via life table and survivorship studies — see Royama, 1997), and also at the population level in which these regulatory factors have been considered. Furthermore, many studies (e.g. Furuta & Sakamoto, 1984; Furuta et al., 1984; Furuta, 1987; Crawley & Akhteruzzaman, 1988; Hunter, 1990, 1992; Watt & MacFarlane, 1991) although recognising the importance of synchrony between herbivore and host plant, do not consider that tree phenology may be altered by, or in response to insect attack. Indeed, good synchronisation between fundatrices and sexual morph production of the aphid, Periphyllus californiensis, and their host is crucial to achieve a large population size (Furuta, 1987).

The sycamore aphid, Drepanosiphum platanoidis, is a well studied monophagous aphid in terms of population density (e.g. Wellings et al., 1985; Dixon, 1990c) and response to host quality (e.g. Dixon, 1966, 1975a; Chambers et al., 1985; Wellings & Dixon, 1987). Its full life-cycle has been described by Blackman & Eastop (1994). One distinguishing feature of D. platanoidis is that second generation virginoparae (fundatrigeniae) experience a long delay after the final moult before reproduction commences. This phase has been termed reproductive diapause by Dixon (1963) or aestivation (Dixon, 1979), and can be seen to last five weeks, but may vary between one to nine weeks from year to year (Chambers, 1979). Aestivation has been shown to be a response to the harsh conditions of summer such as the poor quality of food available to aphids when feeding on mature leaves (Mordvilko, 1908; Dixon, 1975b) and high summer temperatures (Dixon, 1970a, 1985). Aestivation induction and duration is also influenced by the crowding of fundatrigeniae during nymphal development (Dixon, 1970a) and after their final moult (Dixon, 1963). This period of aestivation clearly divides the spring and autumn populations of sycamore aphids, and moreover unifies the age structure of the population, so that in June, July and August there are few or no developing nymphs. From about late August/beginning of September, reproduction resumes in response to senescing leaves (Dixon, 1970a, 1985), reducing photo- and

32 Chapter 2 Population dynamics and life history responses to aphid manipulations thermo-period (Dixon, 1973c) to produce the sexuparae that in turn give rise to the sexual generation (Dixon, 1973c; Blackman, 1974).

Long term population dynamics of D. platanoidis show clear cycles which tend to peak alternate years (Dixon, 1990c). The between-year dynamics of this alternating two-year cycle is brought about by the inverse seasonal relationship of aphid abundance within a year, where for instance a high spring peak in aphid numbers is followed by a low autumnal peak, and vice versa in the subsequent year (Dixon, 1970b). Initially, Dixon (1970b) argued that the two-year cycle of sycamore aphid abundance was the consequence of a 'plant factor' operating within the year. In this he stated that aphids have the capacity to affect their future intra-annual generations by modifying the foliar metabolism of their deciduous host saplings while they are developing in spring. Dixon (1979) also suggested that the cumulative effects of aphid numbers may result in the leaf tissues becoming 'conditioned' so that they represent unfavourable resources in the late summer and early autumn, especially if aphid populations have been large in the spring. Subsequently this plant-mediated hypothesis was discarded (Chambers, 1982), in favour of pure intra-specific competition affecting reproduction and length of aestivation in the second generation aphids. Later, Chambers et al. (1985) showed that the effect of crowding was mediated though the host, rather than solely through direct behavioural interactions of the aphids. Wellings & Dixon (1987), however, demonstrated that leaf infestation history had no direct bearing on aphid performance. More recently, Dixon et al. (1993) claimed that both host quality (in terms of foliar amino acids) and intra-specific competition were implicated in driving the seasonal cycle in aphid abundance via their effects on adult weight and hence recruitment.

In this study, the effects of plant quality on D. platanoidis development time, fecundity and other life-history traits are examined. The current study puts forward the `herbivore-induced plant stress hypothesis' (see White, 1974, 1984; Rhoades, 1983; Waring & Cobb, 1992; Watt, 1994) as the main driving mechanism of sycamore aphid dynamics. This hypothesis suggests that feeding-induced changes during spring increases plant susceptibility to aphid infestation (Haukioja, 1990a) in the early part of the year, and thus enhances individual aphid performance. Secondly, the existence of a `nutritional depletion hypothesis' (see Hassell, 1976) during autumn is proposed. It would be brought about by the eventual exhaustion of previously aphid-stressed saplings, thus resulting in a reduction in the aphid densities they can support in future.

33 Chapter 2 Population dynamics and life history responses to aphid manipulations

This plant-driven theory would mean that the bottom-up control of aphid dynamics is subject to the intrinsic regulatory processes of the aphids themselves. Thirdly, this paper argues that plant mediated processes (i.e. 'host quality dependent aphid densities hypothesis') surpass the top-down control of natural enemies. These hypotheses were tested by monitoring individual performance of caged aphids (cf. Raworth, 1984a; Raworth et al., 1984), changes in field aphid populations, and assessing the numerical response of natural enemies (cf. Raworth, 1984b) to aphid-infested host saplings that have previously been manipulated with different aphid densities. The results are discussed in relation to: (i) the role of plant quality for individual performance and for the population dynamics of the sycamore aphid; (ii) the population densities of aphidophagous natural enemies on aphid-stressed sycamores; and (iii) the impact of aphid densities on the rate of leaf fall.

METHODS & MATERIALS

Experimental Design:

A two year study (1996-1997) was conducted within a mixed deciduous woodland situated at Observatory Ridge (OS ref. 5946 1689) within Silwood Park (near Ascot, Berkshire). Forty sycamore saplings were chosen with similar architecture and size, ranging between 2.0 and 2.5 m high, to avoid initial bias prior to sampling. These 40 saplings were assigned to 10 blocks. There was at least a 15 m wide zone between adjacent blocks, and a minimum of 3 m distance between saplings within a block.

In 1996 the background aphid populations levels (referred to later on as control saplings) were observed; and then in 1997, aphid densities were manipulated in the initial part of the year. Four treatments were randomly assigned to the 4 saplings in each block. The treatments were: A. Spraying aphicide from bud-burst until the appearance of first generation alates; B. Spraying aphicide through the whole spring, until second generation (aestivating) alates appeared; The aphicide used was Pirimor® (pirimicarb) at one-third the recommended concentration (1.67g/1), and was sprayed on 10/03/97 and 04/04/97 for treatments A and B, then 28/04/97, 17/05/97 and 27/05/97 for treatment B only. For full details on screening and application of aphicides see Appendix.

34 Chapter 2 Population dynamics and life history responses to aphid manipulations

C. Control. Unmanipulated 'natural' aphid population development throughout the seasons; D. Augmentation. Manipulation of aphid numbers by supplementing natural populations with 20 apterous fundatrices/week up until the second generation alates appeared. The additional apterae, taken from an external mature tree source, were placed randomly within the sapling canopies, which then freely reshuffled to their preferred feeding sites. After the appearance of the second generation adults, population development was allowed to proceed unhindered throughout the following season.

Monitoring Regime:

1. Field censuses of aphid densities: Aphid counts were made at weekly intervals from late March to early December in 1996 and 1997, the number of aphids were estimated by randomised direct counts of 40 leaves (13 low, 13 mid, 13 high and 1 uppermost apical leaf) per sapling in all the treatments. However, the 1996 data was only taken from control saplings (treatment C). Aphid morphs were allocated to three categories: adult alates, nymphs, and sexuals. Two exceptions in the aphid sampling regime occurred at the beginning and towards the end of the monitoring period. Firstly, in the early part of spring in 1996 and 1997, prior to bud burst, fundatrices on buds were counted by assuming that one bud gives rise to 6 leaves (Leather pers. comm.), e.g. one nymphal fundatrix on six buds would be equivalent to one aphid on 36 leaves. From this, the mean aphid load per leaf was calculated, and multiplied by 40 in order to contribute towards the standard sampling unit for calculating the cumulative spring aphid density. Secondly, by late autumn in 1996 and 1997, sexual morphs appearing on senescing saplings with less than 40 leaves were calculated to obtain a mean aphid load per leaf. For instance, five apterous oviparae and one alate male (= 6 sexual aphids) on say 37 leaves would have a mean aphid load per leaf of 0.162. This value would then be multiplied by 40 in order to contribute towards the standard sampling unit for calculating the cumulative autumnal aphid density.

2. Aphidophagous Predators: The presence of the most abundantly occurring natural enemies were noted weekly whilst sampling the 40 leaves per sapling in 1997 for aphids. The numbers of anthocorids, coccinellids, arachnids, syrphid larvae, and total parasitised aphids (mummies) were recorded on the same 40 leaves to assess 35 Chapter 2 Population dynamics and life history responses to aphid manipulations

their possible impact on the spring and autumnal prey population in each of the four treatments respectively. Ants and chrysopids were also monitored, but their relative abundance were low in comparison with the other predators, and were therefore excluded from the analysis.

3. Clip cage experiments: On the same experimental saplings, a 'bioassay' approach

of measuring aphid life stages (i.e. 1st, 2nd & 3rd instar duration, 4th instar duration, pre-larviposition period; reproduction duration; post-reproductive duration) was used, and total fecundity of parthenogenetic virginoparous individuals in relation to host plant quality was assessed. The three caged experiments, taking into account temporal overlap, took place between 16/04/97 to 16/07/97, 14/06/97 to 08/10/97, and 21/09/97 to 23/11/97 which approximately corresponded to the first, second and third parthenogenetic generations. Newly born nymphs, taken from greenhouse aphid cultures, were individually caged using D. Huggett's (pers. comm.) modification of the clip cage design devised by Noble (1958). The cage consisted of a 5 cm diameter petri-dish embedded within a 10 x 8 x 1.2 cm polyfoam `sandwich' framework. Six cages per sapling (Plate 2-1) were used for all 40 saplings, and the three experimental periods. The performance of each aphid was monitored on a daily basis, except during aestivation where monitoring took place every alternate day. When recording reproductive output, cumulative counts of aphids in each cage were made at each sampling time. No attempt was made to remove offspring from the clip cages because it would cause high levels of disturbance amongst sibling nymphs and their mother (Chambers et al., 1985).

4. Rate of leaf abscission: Total leaf counts per sapling were made from the point of bud-burst to total leaf fall on weekly intervals throughout 1997 to assess the impact of different levels of spring aphid infestation on autumnal abscission.

36 Chapter 2 Population dynamics and life history responses to aphid manipulations

Plate 2-1: Clip caged Drepanosiphum platanoidis on an Acer pseudoplatanus sapling.

Data transformation and statistical analyses:

For control saplings in 1996, and treatment with control saplings in 1997, the aphid counts per sapling (`census data') were split into two at the mid-aestivation point (4th July). Before and after this point, aphid numbers were pooled over time for each of the 10 saplings per treatment. This derived variable represents the cumulative number of aphids in 'spring' and 'autumn' and was analysed in GLIM (version 3.77 copyright 1985 Royal Statistical Society, London; Crawley, 1993). Log-transformations were used to normalise the data and to reduce heteroscedasticity prior to analysis. Firstly, in order to assess whether the treatments worked during 1997, two one-way ANOVA analyses were carried out using treatment as a single factor with 4 discrete levels and declaring spring and autumn cumulative aphid densities as the respective response variables. Graphical representation of treatment differences used least significant difference (LSD) bars at the 95% confidence level (Sokal & Rohlf, 1995). Secondly,

37 Chapter 2 Population dynamics and life history responses to aphid manipulations

ANCOVA analyses were performed on the within and between year correlations, using treatment with 4 levels as the factor. The 'spring 1997/autumn 1997' correlation used log spring aphid density as the covariate, whereas the 'autumn 1996/spring 1997' correlation used log autumn aphid density as the covariate. From undertaking ANCOVA analyses, the minimum F ratios and r2 values were calculated for treatment, and then for the covariate (i.e. log spring or autumn aphid density) with treatment already incorporated, after the interaction term was removed from the GLIM model. The maximum F ratio and r2 values were determined by simple regression and ANOVA analyses. Thirdly, to detect whether alternating trends exist between the two years and seasons, log transformed cumulative aphid numbers from control saplings were analysed using a two-way ANOVA model with equal replication. This used year as a factor containing two discrete levels, and season as the other factor with two levels.

The impact of natural enemies on aphid dynamics during 1997 was assessed by derived variable analysis with ANCOVA, and also with regression analysis using GLIM (Crawley, 1993). As for the aphid census analysis, the cumulative number of predators or mummified aphids were split into two parts (i.e. pre-aestivation and post-aestivation period) and pooled over time whilst retaining the 10 replicates of the four treatments. The aphid (prey) derived variables for spring and autumn were used as covariates; the predator abundance, expressed as a cumulative number of each type of predator for spring and autumn separately, was used as the response variable with treatment as the explanatory factor with four levels. Two separate statistical models for each predator class were run. In the first model, the response variable was log spring predator density; and in the second, log autumn predator density. Derived variables with zeros present were transformed by log (Y + 1) according to the method of Sokal & Rohlf (1995).

The aphid clip-cage results were count data, but when caged aphids (pseudo-replicates) were averaged to give the mean values per sapling (true replicates), they were then log transformed. One-way ANOVAs were fitted to each aphid life history stage in GLIM, using the manipulated spring aphid density treatments as the factor with four discrete levels. For purposes of graphical representation, LSDs were calculated using the method by Sokal & Rolf (1995) for planned multiple comparisons among pairs of treatment means based on equal sample sizes in generations one and three. For the unplanned comparisons among pairs of means based on unequal sample sizes of generation two, the GT2-method proposed by Gabriel (1978) was used. Both the LSD

38 Chapter 2 Population dynamics and life history responses to aphid manipulations and GT2-method used 95% comparison intervals, so that means whose intervals do not overlap were significantly different. The level of significance between treatments of a generation were depicted as: * P<0.05, ** P<0.01, *** P<0.001.

Sycamore leaf fall in 1997 was expressed as a cumulative percentage, of the peak number of leaves, for each of the 40 saplings. This was achieved by calculating the point with the maximum number of leaves. Using 30/05/97 as the baseline for comparison, the fraction of leaves at each successive time interval was calculated as: (100 — [No. leaves at the end of each week / maximum no. leaves in May] x [100/1]) = % defoliation by the end of each subsequent week. Graphical intrapolation facilitated a more precise Julian date for 50% leaf fall time (cf. LD50 approach). In addition, the number of D. platanoidis on 40 leaves per sapling was added together for each successive week from 13/03/97 to 04/07/97. This accumulated derived variable for spring aphid density was then log transformed. Both the log spring aphid densities and Julian date for leaf fall per sapling were run through a parametric regression analysis, within the GLIM package, using a normally distributed error structure.

RESULTS

Field censuses of aphid densities — comparisons between years

Spring 1996 to Autumn 1997:

D. platanoidis was significantly more abundant in 1996 compared with 1997 (Table 2- 1). No significant seasonal trend was evident, despite the overall low numbers in spring and high in autumn in 1996, and vice versa for 1997.

Table 2-1: Mean natural abundance of logged cumulative spring and autumn aphid populations on ten sycamore saplings over a two year monitoring period. The cumulative aphid numbers in the ten control saplings did not differ significantly between seasons within each year (seasons: F 1,36 = 2.11, r2 = 0.033, P = n.s.). However, there was a significant difference between the years 1996 and 1997 (years: F1,36 = 25.05, r2 = 0.39, P<0.001). This resulted in a non significant interaction term (seasons x years F 1,36 = 0.88, r2 = 0.014, P = n.s.).

Year Season Mean aphid abundance (±SE) Spring 24.92 ±5.88 1996 Autumn 26.47 ±8.83 Spring 6.96 ±1.71 1997 Autumn 2.99 ±0.73

39

Chapter 2 Population dynamics and life history responses to aphid manipulations

Autumn 1996 and Spring 1997:

The relationship between the cumulative spring population and the cumulative autumn population was positively correlated, although not significantly different for the 10 control saplings (Fig. 2-1).

3 A

97) • 2 A 19 ity A A

hid dens A

ap A ing

r 0 (sp In

-1 0 1 2 3 4 5 In (autumn aphid density 1996) Fig. 2-1: The effects of season on the natural aphid population trends of Drepanosiphum platanoidis between autumn 1996 and spring 1997. Density is expressed as cumulative number of aphids per leaf over 'half the year before and after aestivation for ten control saplings. The relationship was positive, but not statistically significant (F 1, 8 = 2.54, r2 = 0.24, n = 10, P = n.s., Ln y = 0.44Ln(x) + 0.33).

Field census of aphid densities — comparisons within years

Spring 1996 and Autumn of 1996:

The cumulative number of aphids in autumn was positively and significantly correlated with the abundance of the preceding spring on the ten control saplings (Fig. 2-2). The relationship reveals an almost 1:1 ratio between spring and autumn.

5

96) •

19 4

ity ns 3 de id

h •

ap • A

tumn 1 au ( In 0 0 1 2 3 4 5 In (spring aphid density 1996) Fig. 2-2: Within 1996 there was a strong positive seasonal correlation between the spring and the autumn aphid populations (F 1,8 = 28.21, r2 = 0.78, n =10, P<0.001, Ln y = 1.15Ln(x) — 0.62).

40

Chapter 2 Population dynamics and life history responses to aphid manipulations

Spring 1997 and Autumn 1997 — comparing the four aphid density treatments:

The four aphid load treatments on sycamore saplings were significantly different from each other in spring (Fig. 2-3). By autumn only the control and aphid augmented treatments were significantly different from the aphicide sprayed treatments.

3.0 _ ;E; 2.5 - 17) 0 Spring w 2.0 _ Autumn 1.5 - a. co 1.0 _ c7t 0.5 - E 0.0 -

-0.5 I -1.0 A C D Treatment Fig. 2-3: The effect of four aphid load treatments on cumulative aphid densities were significantly different among treatments in spring (F 3,36 = 12.80, r2 = 0.52, P<0.001) and autumn (F 3,36 = 4.50, r2 = 0.27, P<0.01). The confidence bars are one LSD set at the 95% comparison interval for spring and autumn treatments separately. Treatments: A = spraying aphicide in March and April; B = spraying aphicide from March to May; C = control saplings; D = augmented aphid populations.

In 1997, the main effects of treatments on aphid population densities were not significant; also the treatment x log spring density interaction was not significant (Table 2-2). However, the log spring density was significantly influential on the following autumnal aphid densities. Since the treatments involved manipulating spring aphid densities on treatment saplings, there was a strong correlation between treatment and log spring density, which together made the treatments highly significant (Fig. 2-4).

Table 2-2: Effects of spring aphid density treatments on autumn aphid abundance in 1997.

Log Spring Aphid Density 1997 & F-ratio d.f. 1-2 Log Autumn Aphid Response 1997 Treatments (ANCOVA) 0.32 3,32 0.024 n.s. (REGRESSION) 4.50 3,36 0.27 <0.01 Spring Aphid Density (ANCOVA) 5.91 1,32 0.15 <0.025 (REGRESSION) 25.12 1,38 0.40 <0.001 Interaction: Treatments x Spring Aphid Density (ANCOVA) 0.23 3,32 0.017 n.s. Full modcl:Treatments*Spring Aphid Density (ANCOVA) 3.58 7,32 0.44 <0.01

A significant and positive linear trend, although decelerating at higher spring densities, was evident between spring and autumn aphid populations (Fig. 2-4). This positive linear relationship clearly points out that the resulting autumn aphid densities were dependent on the previous aphid densities during spring.

41 Chapter 2 Population dynamics and life history responses to aphid manipulations

2.5 2 rnC) A A 1.5 X A XA x X • ■ A X ■ • X X A 0 ca E -0.5 -1 ca ■ A c -1.5 ■ -2 -2 -1.5 -1 -0.5 0 0.5 1 1.5 2 2.5 3 3.5 In (spring aphid density 1997)

Fig. 2-4: The effect of treatments (i.e. manipulating spring aphid density) on the resulting autumnal populations of Drepanosiphum platanoidis throughout 1997 was significant (F 1,38 = 25.12, r2 = 0.40, P<0.001, Ln y = 0.40Ln(x) - 0.026). Treatments: • A = aphicide in March and April; ■ B = aphicide from March to May; A C = control saplings; X D = augmented aphid populations.

Field census to detect numerical response of predators

Table 2-3 shows all the summary statistics for natural enemy abundance in spring and autumn in 1997. During spring 1997, ANCOVA analyses showed that treatment main effects were significant for the abundance of coccinellids (e.g. Halyzia 16-guttata, Coccinella 7-punctata, Adalia 2-punctata, Calvia 14-guttata) and arachnids (e.g. rotundum, Araneus diadematus, A. quadratus); but when regression analyses were performed between predators and prey, only syrphid larvae (e.g. Syrphus ribessi, Episyrphus balteatus) and parasitoid mummies (e.g. Aphelinus flavus, Monoctonus pseudoplatani, Trioxys cirsii, Dyscritulus planiceps) were significant. The aphid treatments x log spring aphid abundance interaction was only significant for coccinellids and arachnids. Spring aphid density significantly influenced the abundance of all the predators monitored during the spring 1997 census. There was a general tendency of increased predator abundance towards the more aphid infested saplings in spring (Figs. 2-5 i to ix, except 2-5 vii), which were statistically significant in all cases, except for the response of syrphid larvae (Figs. 2-5 vii) in the regression statistics for spring aphid density. Since implementing treatments involved spring aphid density manipulations, both the explanatory variable and the covariate were combined to reveal highly significant numerical responses of anthocorids (e.g. Anthocoris nemorum, A. confusus), arachnids & parasitoid mummies in the ANCOVA full model.

42 Chapter 2 Population dynamics and life history responses to aphid manipulations

By autumn 1997, only the arachnids and primary parasitoids (indicated by mummy abundance) showed statistically significant differences between the four treatments. The numerical response of arachnids to aphid density (see both aphid density ANCOVA and REGRESSION analysis) showed positive correlations in treatments B (aphicide March to May) and C (control), but negative relationships for treatments A (aphicide March to April) and D (aphid augmentation) (Fig. 2-5 vi). Density-dependent parasitism of D. platanoidis was evident in autumn 1997, as with the preceding spring, with clear positive correlations between prey abundance and parasitoid mummies (Fig. 2-5 x). The numerical response of primary parasitoids (see treatments REGRESSION; aphid density ANCOVA & REGRESSION; full model ANCOVA) was the most tightly linked to its prey density of all the natural enemies considered during the current study.

Table 2-3: Effects of aphid density and treatments on cumulative generalist and specialist predator abundance during spring and autumn 1997.

PREDATORS 1997 SPRING AUTUMN F-ratio df r2 P F-ratio df r2 P ANTHOCORIDS Treatments (ANCOVA) 1.56 3,32 0.08 n.s. 0.19 3,32 0.02 n.s. (REGRESSION) 0.96 3,36 0.07 n.s. 0.60 3,36 0.05 n.s. Aphid Density (ANCOVA) 11.24 1,32 0.33 <0.005 2.15 1,32 0.06 n.s. (REGRESSION) 17.73 1,38 0.32 <0.001 3.82 1,38 0.09 n.s. Interaction: (ANCOVA) 0.22 3,32 0.01 n.s. 0.42 3,32 0.03 n.s. Treatments x Aphid Density Full model: (ANCOVA) 3.23 7,32 0.41 <0.01 0.76 7,32 0.14 n.s. Treatments*Aphid Density COCCINELLIDS Treatments (ANCOVA) 38.935 3,32 0.03 <0.001 0.98 3,32 0.08 n.s. (REGRESSION) 0.70 3,36 0.06 n.s. 0.80 3,36 0.06 n.s. Aphid Density (ANCOVA) 449.51 1,32 0.11 <0.001 0.81 1,32 0.02 n.s. (REGRESSION) 6.21 1,38 0.14 <0.025 0.15 1,38 0.004 n.s. Interaction: (ANCOVA) 23.10 3,32 0.02 <0.001 0.19 3,32 0.02 n.s. Treatments x Aphid Density Full model: (ANCOVA) 1.06 7,32 0.19 n.s. 0.51 7,32 0.10 n.s. Treatments*Aphid Density ARACHNIDS Treatments (ANCOVA) 6.43 3,32 0.17 <0.005 0.047 3,32 0.004 n.s. (REGRESSION) 0.62 3,36 0.05 n.s. 0.21 3,36 0.02 n.s. Aphid Density (ANCOVA) 50.42 1,32 0.45 <0.001 4.24 1,32 0.11 <0.05 (REGRESSION) 18.25 1,38 0.32 <0.001 5.13 1,38 0.12 <0.05 Interaction: (ANCOVA) 3.75 3,32 0.10 <0.025 1.30 3,32 0.10 n.s. Treatments x Aphid Density Full model: (ANCOVA) 6.71 7,32 0.59 <0.001 1.29 7,32 0.22 n.s. Treatments*Aphid Density SYRPHID LARVAE Treatments (ANCOVA) 1.89 3,32 0.13 n.s. 1.91 3,32 0.13 n.s. (REGRESSION) 2.93 3,36 0.20 <0.05 1.72 3,36 0.13 n.s. Aphid Density (ANCOVA) 3.65 1,32 0.09 <0.025 0.41 1,32 0.01 n.s. (REGRESSION) 2.21 1,38 0.15 n.s. 0.078 1,38 0.002 n.s. Interaction: (ANCOVA) 0.27 3,32 0.01 n.s. 1.60 3,32 0.11 n.s. Treatments x Aphid Density Full model: (ANCOVA) 1.98 7,32 0.30 n.s. 1.50 7,32 0.25 n.s. Treatments*Aphid Density

43 Chapter 2 Population dynamics and life history responses to aphid manipulations

PARASITOID MUMMIES Treatments (ANCOVA) 1.08 3,32 0.05 n.s. 1.87 3,32 0.12 n.s. (REGRESSION) 5.17 3,36 0.30 <0.005 6.31 3,36 0.34 <0.005 Aphid Density (ANCOVA) 35.80 1,32 0.50 <0.001 7.32 1,32 0.16 <0.025 (REGRESSION) 115.47 1,38 0.75 <0.001 23.55 1,38 0.38 <0.001 Interaction: (ANCOVA) 0.36 3,32 0.01 n.s. 0.061 3,32 0.004 n.s. Treatments x Aphid Density Full model: (ANCOVA) 19.79 7,32 0.01 <0.001 4.78 7,32 0.51 <0.005 Treatments*Aphid Density

The effects of spring and autumn aphid density treatments (• Treatment A = aphicide March to April; II Treatment B = aphicide March to May; A Treatment C = control; X Treatment D = aphid augmentation) on the numerical response of generalist and specific aphidophagous predators during 1997 are graphically represented below.

ANTHOCORIDS ANTHOCORIDS 4 5 A • • 72 4 • 3 0 • 0 • X O 0 • II 1 f/A C13 as in 2 I II III cri 0 a, .= !0‹ 2 - 1 I ■ ■ NA A X - 1-

0 0

-3 -2 -1 0 1 2 3 4 -2 -1 0 1 2 3

In (spring aphid density) In (alum aphid density)

Fig. 2-5 i: The numerical response of anthocorids to spring Fig. 2-5 ii: The numerical response of anthocorids prey densities was significantly correlated (F 1,38 = 17.73, to autumnal prey densities was not significantly related r2 = 0.32, P<0.001, Ln y = 0.40Ln(x) + 1.48). (F = 3.82, r2 = 0.09, P = n.s., Ln y = 0.23Ln(x) + 2.80).

COCCINELLIDS 3.0 COCCINELLIDS 3.5

) ' ity 2.5 X 3.0 • X ity, • A • X A

dens 2.0 2.5 dens id ll 1.5 A *< A llid 2.0 - • All• A • ine II • • A X N ine cc c • ■ • 1.0 ■ A OX A 1.5 oc • • 0( X X co c

11111i M• •< X n 1.0 • •X A ing r

0.5 m e< tu 0.5 (sp

In 0.0 • • • so • >(•< X (au 0.0 In •• 4K A X -0.5 -0 5 -3 -2 -1 0 1 2 3 4 -2 -1 0 1 2 In (spring aphid density) In (autumn aphid density 1997)

Fig. 2-5 iii: The numerical response of coccinellids to Fig. 2-5 iv: The numerical response of coccinellids spring aphid densities was significantly correlated to autumnal prey densities was not significantly related (F 1,38 = 6.21, r2 = 0.14, P<0.025, Ln y = 0.69Ln(x) + 0.47). (F 1,38 = 0.15, r2 = 0.004, P = n.s., Ln y = 0.07Ln(x) +1.36).

44

Chapter 2 Population dynamics and life history responses to aphid manipulations

ARACHNIDS 5 ARACHNIDS • 5 •

) ik4 >A • • •X ity A • • 4 • a ik• X I • X >r dens A X X X y

id MO A A - 3 X X hn ■ • • A • V • A X rac a 2 _ mn

tu •

• (au In

0 0 -3 -2 -1 0 1 2 3 4 -2 -1 0 1 2 3 In (spring aphid density) In (autumn aphid density)

Fig. 2-5 v: The numerical response of arachnids to spring Fig. 5 vi: The numerical response of arachnids to autumn aphid densities was significantly correlated (F138 = 18.25, aphid densities was significantly related (F1,38= r2 = 0.32, P<0.001, Ln y = 0.32Ln(x) + 2.99). 5.13, r2 = 0.12, P<0.05, Ln y = 0.27Ln(x) + 3.58).

SYRPHID LARVAE SYRPHID LARVAE 3.0 3.0 -

2.5 ) 2.5 • X ity • ns c 2.0 A 2.0 a) • • de • 1.5 - • X hid 1.5 +Z • A A A X rp • • X l• >0( rn 1.0- ■ • I■ • sy 1.0

1 mn

■ • tu • Ill OM OCI A 0.5 • X 0.5 (au 0.0 • ma • mi •X>404. X In 0.0 NA A eige>0(•• e• A -05 -0.5 -3 -2 -1 0 1 2 3 4 -2 -1 0 1 2 3 In (spring aphid density) In (autumn aphid density)

Fig. 2-5 vii: The relationship between the number of Fig. 2-5 viii: The numerical response of syrphid larvae syrphid larvae to spring aphid densities was not to autumn aphid densities was not significantly significant (F 1,38 = 2.21, r2 = 0.15, P = n.s., related (F1,38 = 0.078, r2 = 0.002, P = n.s., Ln y = 0.23Ln(x) + 0.55. Ln y = -0.04Ln(x) + 0.64).

PRIMARY PARASITOIDS PRIMARY PARASITOIDS 6

) X 5 4XAlk( ity • • • X ns • de

•A2°• 4 4 ....X •

mmy 3 X g 3 - • A X IX • mu

" • • n c 2 • • L' 2 II A tum • • ■ •

(au • In 0 0 •

-3 -2 -1 0 1 2 4 -2 -1 0 1 2

In (spring aphid density) In (autumn aphid density)

Fig. 2-5 ix: The numerical response of total primary para- Fig. 2-5 x: The numerical response of total primary sitoids to spring aphid densities was significantly correlated parasitoids to autumn aphid densities was (F 1,38 = 115.47, r2 = 0.75, P<0.001, Ln y = 1.07Ln(x) + 2.92). significantly related (F 1,38 = 23.55, r2 = 0.38, P<0.001, Ln y = 0.84Ln(x) + 2.72).

45 Chapter 2 Population dynamics and life history responses to aphid manipulations

Clip cage experiments in 1997

The results of the clip cage experiments show that a transitional change in aphid life history performance was evident between the three parthenogenetic generations (Figs. 2-6 i & ii, 2-7 i to iii, 2-8). In generation one and continuing through generation two, aphids performed best on augmented saplings (treatment D), but by generation three aphid performance was suppressed in treatment D when compared with their respective controls (treatment C). In particular, during spring and summer, D. platanoidis responded strongly, in terms of rapid growth rates (Figs. 2-6 i & ii) and reproductive outputs (Fig. 2-8) at the individual level, to saplings with augmented treatments, whereas their reproduction was low on saplings with previous aphicidal treatments. Aphids in autumn were the most fecund (Fig. 2-8) and completed their development (Figs. 2-6 i & ii and 2-7 i) in the most rapid time of all the three generations. The life history performance of these autumnal aphids can be ranked in the following ascending order of treatments: D, C, A & B. These results are described in more detail as follows.

Nymphal development:

The duration of the first, second and third instar stages differed significantly between treatments B (aphicide March to May) and D (aphid augmentation) in generations two and three (see Fig. 2-6 i). In generation two, initial nymphal development on the long aphicide sprayed saplings (treatment B) took the longest, although the only significant difference was found between treatments B and D. By generation three, the duration of nymphal development reversed treatment positions, where nymphal development on treatment B was the shortest, compared with the other three treatments. These treatment trends were continued for the fourth instar development times during generations two and three (see Fig. 2-6 ii), although treatment differences were not statistically significant. However, treatment effects on generation one forth instar fundatrices were highly significant, with the short sprayed saplings (treatment A) showing the longest and the augmented saplings (treatment D) showing the shortest final instar development times compared with the control.

46 Chapter 2 Population dynamics and life history responses to aphid manipulations

25 *

**

I

0 A BCD A BCD ABCD 1 2 3 Treatment Aphid Generation

Fig. 2-6 i: Mean duration of the first, second and third instar nymphal development times for generations two and three. First generation data was not collected due to unknown hatching times of fundatrices. In generation two, the duration of early instars was longer in treatment B compared to treatment D (F 3.34 = 3.61, r2 = 0.24, P<0.025). In generation three, the reverse was true with treatment B having the shortest early instar duration (F 3, 16 = 5.59, r2 = 0.51, P<0.01). The confidence bars are one LSD set at the 95% comparison interval for each generation separately. Treatments: A = spraying aphicide in March and April; B = spraying aphicide from March to May; C = control saplings; D = augmented aphid populations.

10

ca a, 8 ro *** •— _c 0 6

o 4 r c- >+z C

A BCD A BCD A BCD 1 2 3 Treatment Aphid Generation

Fig. 2-6 ii: Mean duration of fourth instar nymphal development time in three generations. In the first generation, the duration of the forth instar took the longest in treatment A (F 2, 27 = 9.83, r2 = 0.42, P<0.001; treatment B still having aphicide treatment). In the other two generations, there were no significant differences in the duration of the forth instar among treatments (second generation: F 3, 31 = 1.17, r2 = 0.10, P = n.s.; third generation: F 3, 16 = 2.41, r2 = 0.31, P = n.s.). The confidence bars are one LSD set at the 95% comparison interval for each generation separately. Treatments: A = spraying aphicide in March and April; B = spraying aphicide from March to May; C = control saplings; D = augmented aphid populations.

47 Chapter 2 Population dynamics and life history responses to aphid manipulations

Pre-larviposition:

The length of pre-larvipostion period was statistically significant in generations two and three (see Fig. 2-7 i). The long sprayed saplings (treatment B) were markedly different from the control (treatment C) and augmented saplings (treatment D) in both generations, with the aphid response completely reversing positions in all treatments. In the second generation, aphids on the long sprayed saplings (treatment B) demonstrated the longest time to reproduction and aphids on augmented saplings (treatment D) rapidly completed the pre-reproductive phase, but by generation three the converse was true. Generation one, showed no significant differences between the treatments.

20 **

***

A BCD A BCD A BCD 1 2 3 Treatment Aphid Generation Fig. 2-7 i: Pre-larviposition period (i.e. from final moult to onset of reproduction) was different among treatments in generations two and three, but not in generation one (F 2, 27 = 0.10, r2 = 0.0075, P = n.s.; treatment B still having aphicide treatment). In generation two, treatment B showed a significantly larger prelarval duration compared to treatments C and D (F 3, 28 = 5.73, r2 = 0.38, P<0.005); and by generation three, the prelarval duration for treatment B was significantly decreased compared to treatments C and D (F 3, 16 = 10.52, r2 = 0.66, P<0.001). In all the generations, treatment A showed no significant difference in prelarval duration from the other treatments. The confidence bars are one LSD set at the 95% comparison interval for each generation separately. Treatments: A = spraying aphicide in March and April; B = spraying aphicide from March to May; C = control saplings; D = augmented aphid populations.

Reproductive duration:

The reproductive duration showed significant differences between the three treatments in generation one only (see Fig. 2-7 ii). Here, aphids on the control (treatment C) and especially the augmented saplings (treatment D) clearly showed the longest fertility schedule, which contrasted with the shortest reproductive time of the short sprayed

saplings (treatment A). Despite the non significant result for generation three (F3, 16 = 2.624, P<0.10), the LSD bars of treatments B (aphicide March to May) and D (aphid augmentation) showed no overlap.

48 Chapter 2 Population dynamics and life history responses to aphid manipulations

35

30

5-

0 A B C D A B C D A B C D 1 2 3 Treatment Aphid Generation Fig. 2-7 ii: Reproductive duration of parthenogenetic virginoparae was only different among treatments in generation one, with treatment A having significantly the lowest reproductive longevity (F 2, 27 = 8.05, r2 = 0.37, P<0.005; treatment B still having aphicide treatment). The treatments in generation two were not significantly different (F 3, 27 = 0.57, r2 = 0.059, P = n.s.). Although treatments B and D in generation three were different from one another, no significance was found (F 3, 16 = 2.62, r2 = 0.33, P = n.s.). The confidence bars are one LSD set at the 95% comparison interval for each generation separately. Treatments: A = spraying aphicide in March and April; B = spraying aphicide from March to May; C = control saplings; D = augmented aphid populations.

Post-reproductive duration:

The post-reproductive duration, time from last birth to death, was significantly different between treatments in generation one, but not in generations two and three (Fig. 2-7 iii). In generation one, the control (treatment C) and augmented saplings (treatment D) clearly contrasted with the short sprayed saplings (treatment A) which showed low post- reproductive survivorship. The same trend continued in generation two. Generation three showed an inverse trend to that of both preceding generations.

12 **

•CD 10 15 'a 17 >, 8 2 co O. c 6 r;) ° o rri 0_'- 4 o -o 2• 2 _

0 A B C D A B C D A BCD 1 2 3 Treatment Aphid Generation Fig. 2-7 iii: Post-reproductive duration of individual parthenogenetic virginoparac measured as the time bcrween last nymph produced to death for the four treatments in each of the three generations. There were no differences in post- reproductive duration among treatments in generation two (F 3, 27 = 1.68, r2 = 0.34, P = n.s.) and generation three (F 3, 16 = 1.91, r2 = 0.26, P = n.s.). However in generation one, aphids in treatment A had significantly the lowest post- reproductive duration compared to treatment C and D (F 2, 27 = 5.72, r2 = 0.30, P<0.01; treatment B still having aphicide treatment). The confidence bars arc one LSD set at the 95% comparison interval for each generation separately. Treatments: A = spraying aphicide in March and April; B = spraying aphicide from March to May; C = control saplings; D = augmented aphid populations.

49 Chapter 2 Population dynamics and life history responses to aphid manipulations

Total fecundity:

Total fecundity differed significantly between treatments in generations one and three, but there was no significance in generation two (Fig. 2-8). In generation one, augmented saplings (treatment D) showed most reproductive output, whereas the short sprayed saplings (treatment A) showed suppressed fecundity in comparison to the control (treatment C). By generation three, the high progeny output from aphids on the long sprayed saplings (treatment B) contrasted sharply with augmented saplings (treatment D) which yielded the lowest reproductive output. A significant difference in net reproductive output was found between treatments A with D, and treatments B with

C.

40 _ — 0_ 35 _ 0 0 3° ** ▪ 25 - = 2 20 - a- 15 -

▪ 10- c CU a) 5 2 0 A B C D A BCD A BCD 1 2 3 Treatment Aphid Generation

Fig. 2-8: Total mean fecundity of individual parthenogenetic virginoparae expressed as the total number of nymphs produced for the four treatments in three generations. In generation one, aphids in treatment A produced signifcantly fewer nymphs than aphids in treatments C and D (F 2, 27 = 6.69, r2 = 0.33, P<0.005; treatment B still having aphicide treatment). In generation two, there was no difference in total reproductive output among treatments (F 3, 27 = 0.99, r2 = 0.10, P = n.s.); but by generation three, treatment B showed significantly the highest and treatment D the lowest reproductive output (F 3, 16 = 5.39, r2 = 0.50, P<0.01). The confidence bars are one LSD set at the 95% comparison interval for each generation separately. Treatments: A = spraying aphicide in March and April; B = spraying aphicide from March to May; C = control saplings; D = augmented aphid populations.

Rate of leaf abscission under aphid density treatments

No correlation was evident between spring aphid density and the time to 50% leaf fall in 1997 (Fig. 2-9).

50 Chapter 2 Population dynamics and life history responses to aphid manipulations

350

300_ • • • 250 _ • *; : • 4.4"• so. • = • • >, • • • • 4 • s 4- CZ • -0 200 - • - 7( 5 C*51 •; 150 --- 100 _

50-

0 -3 -1 0 1 2 3 4 In (spring aphid density)

Fig. 2-9: Leaf fall expressed as day of 50% leaf loss was not explained by spring aphid density on 40 Acer pseudoplatanus saplings in 1997 (F 1, 38 = 0.33, r2 = 0.0084, n = 40, P = n.s., y = 1.83Ln(x) + 260.32).

DISCUSSION

The clip caged aphid study confirmed the hypothesis that aphids change the quality of their host plant via a feedback mechanism. Early in the year aphids cause induced amelioration in host quality where individual aphid performance is enhanced during spring. By the end of the year, aphids modify the host plant so that deterioration in food quality results, which in turn negatively affects aphid performance. At the population level, the within year aphid densities did not show an inverse relationship between spring and autumn (cf. Dixon, 1970b), but an almost perfect linear trend with a slope of unity. This indicates that there is an extrinsic mechanism (i.e. plant factor) driving the within-year aphid dynamics, and not simply density-dependent processes (e.g. crowding experience) regulating aphid numbers (cf. Chambers, 1982; Chambers et al., 1985).

Intra- and inter-annual population trends:

Within-year natural aphid abundance in 1996 and 1997 showed a positive unidirectional change, irrespective of fundatrix densities at the start of the year; that is, low numbers of aphids in spring lead to low aphid abundance in autumn, and vice versa. This indicates an extrinsic process or plant-mediated mechanism is driving the intra-annual dynamics. With these results contradicting the inverse seasonal pattern of aerial aphid abundance

51 Chapter 2 Population dynamics and life history responses to aphid manipulations

operating within a year (Dixon, 1970b), it seems logical to accept the 'host quality dependence hypothesis'. This hypothesis means that less attractive saplings (in terms of either poor nutritional status, elevated phago-deterrents, low locational apparency to searching aphids — see Feeny, 1976; Rhoades & Cates, 1976; Rhoades, 1983; Srivastava et al., 1988; Edwards, 1989; Dixon & Kindlmann, 1990a; Dixon, 1990c) support a lower abundance of aphids throughout the entire year. That is, low numbers of fundatrices in spring will mean a low level of recruitment in the successive generations, despite migratory fluxes, if aphids occupy a poor quality host. This is confirmed by Dixon (1979) who demonstrated that the number of second generation aphids present at the onset of autumn is generally higher in years when aphids are abundant in spring than when they are scarce. Furthermore, Leather (1996b) has shown in sycamores that a poor tree for herbivores in one year is still a poor tree for herbivores the next year.

The annual aphid abundance results showed that high aphid populations in 1996 were followed by lower aphid numbers in 1997. This could be a consequence of high aphid populations in 1996, accelerating bud burst in the following spring in order to evade the full impact of aphids hatching from overwintering eggs, and hence overloading the host tree in 1997 (S.R. Leather, pers. comm.).

The results (Table 2-1 on p. 39) also show that the natural aphid abundance in autumn 1996 was higher than that found earlier in that year, whilst natural aphid abundance in autumn 1997 was lower than in spring of that year. This implies that a regulatory mechanism was operating on autumnal aphid abundance in opposing ways between these two years, which may contribute to an alternating yearly cycle (see Dixon, 1990c). The changes in autumnal fecundity from one year to the next could be due to differences in weight and/or size (quality) of the aphid; that is, large aphids found on good host plant quality are potentially more fecund (Dixon & Wratten, 1971; Leather & Wellings, 1981; Dixon et al., 1982). Conversely, if host nutrient quality is degraded later in the year, aphid size may be reduced and thus lead to lower herbivore abundance (Hassell, 1976).

In addition, the number of aphids in spring 1997 was positively correlated with the previous autumnal abundance on a sapling basis as expected (see Dixon, 1970b). This implies that high numbers of parthenogens/sexual morphs in autumn would give rise to many overwintering eggs, which in turn lead to a high fundatrix emergence in spring, and vice versa (Dixon, 1976a; Leather et al., 1993).

52 Chapter 2 Population dynamics and life history responses to aphid manipulations

Effects of spring aphid density treatments on life table performance of individually clip- caged parthenogenetic virginoparae over three sequential generations:

The efficacy of herbivore-induced alterations in plant quality on inter-generational insect performance was clearly reflected in the clip cage results. Sycamore aphids isolated in cages on aphid augumented saplings initially performed well in terms of rapid nymphal development, increased reproductive duration and output, short pre- parturition and long post-larviposition length. This lends support for the plant-stress hypothesis, as with the study on other aphids (e.g. Elatobium abietinum on Picea sitchensis — Major, 1990). However, by the end of the year, aphid performance on highly infested saplings was poor in relation to previously lightly infested saplings. Similar results were found in Raworth et al. (1984) field-cage experiment, where aphid density treatments significantly affected final aphid numbers.

It is well known that aphids aggregate to form 'physiological sinks' (see Hughes, 1963; Way & Cammell, 1970; Dixon & Wratten, 1971; Kennedy & Stroyan, 1959; Peel & Ho, 1970) which can eventually induce a severe nutrient drain (Dixon, 1971a, 1973c; Leather, 1988), thus stressing the host plant (Way & Cammell, 1970). Indeed, the results of this study provide evidence that delayed aphid-induced changes in host plant quality determine overcompensating responses in sycamore aphids (Dixon, 1970b). It also supports the work by Barlow & Dixon (1980) that lime saplings appear to show an induced response; this is because aphids reared in autumn, on saplings that were heavily infested in the spring, do worse than aphids reared on control saplings. Moreover, the current study ties in with Dixon's (1971a) claim that infestation history of the host-plant might contribute to the seasonal changes in aphid abundance. High aphid numbers in the previous spring can have a delayed inhibiting effect on the reproductive rate of the aphids present in the autumn, i.e. two or three generations later by reducing the amount of nitrogen salvaged from the leaves in the following autumn (Dixon, 1971a). Studies on the green spruce aphid, Elatobium abietinum, showed that an aphid-induced deterioration in both the quantity and quality of food available in summer is important in causing the collapse of the population (Parry, 1974, 1976). Dixon (1971a) and Kidd (1977) suggest that this reduction in plant quality is more likely to be a metabolic disturbance resulting from the growth inhibitors that aphids inject into the leaves whilst feeding; while Dixon (1973a) asserts that this is not likely to be an anti-aphid response by the tree.

53 Chapter 2 Population dynamics and life history responses to aphid manipulations

These clip cage results contrasted with the findings of Wellings & Dixon (1987) showing that sapling infestation history or intensity had no significant effect on aphid performance. Their conclusions are indeed surprising when one considers the intimate nature of the relationship that parasitic insects have with their hosts, influencing plant physiology through feeding and salivary secretions (Miles, 1972, 1989) and, in turn, responding to subtle changes in plant chemistry (Dixon, 1971a; Kidd et al., 1990). Their study does not consider the 'phloem plumbing' system (Coleman & Jones, 1991), and the patterns of systemically induced resistance operating under the control of plant vasculature (see Jones et al., 1993). Neither do they consider the serious limitations on the vascular flow, via reduced turgor pressure, in response to aphid damaged leaves through multiple stylet puncturing (Peel & Ho, 1970). In addition, Wellings & Dixon (1987) found that no major changes in leaf quality (using foliar amino acids as their stress gauge) occurred as a result of prolonged aphid infestation. Whether their analysis of crude concentrations of whole-leaf extracts represents sap content and resembles chemicals directly experienced by aphids is open to question (Kidd et al., 1990).

Aphid population dynamics and individual aphid life history results together:

High aphid densities can inflict enough damage on their host, via their feeding behaviour (Tedders & Thompson, 1981; Evert et al., 1968), to elicit a plant-stress response (Watt, 1994; but see Dixon, 1975a and Dixon et al., 1993). Since phloem sap normally represents a resource low in nitrogen (Dixon, 1985), stress-induced increases in concentrations of soluble nitrogen should positively influence the growth and reproduction of the sap-suckers to a much higher extent compared with other folivorous insects (Broadbeck & Strong, 1987). Even a comparatively short period of stress may result in substantial changes in population density because of the quick reproductive response of aphids (Larsson, 1989). Apart from the induced amelioration in plant quality, herbivore-induced resistance may result (Gibberd et al., 1988; Wratten et al., 1988; Haukioja, 1990a; Kidd, 1990; Edwards et al. 1992; Gora et al., 1994; Barker et al., 1995), thus leading to a reduction of the nutritional value of the plant to the insect (Haukioja, 1980). Indeed, short-lived insects may encounter delayed induced resistance brought about by damage to the plant caused by an earlier generation (Haukioja & Neuvonen, 1987; Haukioja et al., 1988). This may function in a density-dependent way, and act as a destabilising agent for the insect population (Haukioja & Neuvonen,

54 Chapter 2 Population dynamics and life history responses to aphid manipulations

1987; Karban & Myers, 1989), but may also contribute to regular cyclic fluctuations in densities (Haukioja, 1980). However, the immediate relevance of delayed induced resistance to phloem-feeders seems negligible since the noxious compounds found in leaf tissue are believed to be less abundant in vascular tissue (Raven, 1983).

Although population census counts were made on the same saplings as the caged aphid observations, an important distinction has to be made between the information gathered from each methodological approach. The clip cages minimised extraneous factors by forcing the aphid to experience its host's foliar conditions within the year. Here, the nymphs had the capacity to bring about a sink effect, thus promoting aphid performance. However, fierce intra-specific competition was likely to ensue between the siblings and the mother that were confined to a crowded space (see Dixon, 1979 for details), because progeny removal did not take place at each monitoring session. Intra- specific competition (crowding) for food resources has been shown to reduce growth, development and reproductive rates in the sycamore aphid (Chambers et al., 1985), and result in the development of small aphids (Dixon, 1987) which in turn affects future recruitment (Dixon et al., 1993). On the other hand, both intrinsic and extrinsic forces were altering the total aphid count when the weekly census was conducted. The field census recorded the number of aphids remaining after the operation of density- dependent processes (e.g. predation, immigration, emigration) and density—independent processes (e.g. heavy rainfall, wind). Density-dependent migration (Chambers et al., 1985) and/or predation can affect adult to nymph proportions on the sapling. Due to emigration it is likely that, at high adult densities, the number of nymphs recorded could have been produced by considerably more adults than were monitored. Conversely, predation may reduce nymphal numbers to a lower level than should have been expected for the numbers of adults recorded. Hence, the aphid census results should be carefully interpreted since the impact of predators may mask the clear underlying seasonal trends.

Numerical response of natural enemies to spring aphid-load treatments:

Spring aphid densities were correlated to the abundance of all five natural enemies in a positive manner. This study confirms previous research that plant stressing (e.g. Strauss, 1987) not only increases aphid abundance, but also enhances the numerical responses of adult predators and their larvae (see Neuenschwander et al., 1975; Frazer et al., 1981a; Honek, 1983; Tamaki & Long, 1987). Likewise, Dixon & Russel (1972)

55 Chapter 2 Population dynamics and life history responses to aphid manipulations found a positive correlation of A. nemorum densities to the numbers of D. platanoidis on sycamore. Anthocorids have also shown a positive numerical relationship to the densities of psyllid larvae on pear saplings (Trapman & Blommers, 1992) and on hawthorn (Novak & Achtziger, 1995), and were especially abundant when psyllids had built up to a very high level. On reaching a threshold aphid density, crucial for larval survival (Dixon, 1970c; Wratten & Pearson, 1982), increased density of aphids enhances the quantity of eggs deposited by predators (Wright & Laing, 1980; Ives, 1981a,b); this has been shown for coccinellids on sycamore (Dixon, 1970c). The high dispersal powers of adult ladybirds (Hodek, 1967), and tendency to stay and lay eggs only in areas where available aphid population density is high (Dixon, 1958), means they attack aphids only when prey are very numerous and can support a new generation of larvae (Dixon, 1970c).

However, by autumn only the specialist primary parasitoids and, to a lesser extent, territorial spiders were significantly affected by aphid density. Hofsvang (1990) and Ferguson & Stiling (1996) showed that natural levels of parasitoids have the capacity to inflict higher levels of mortality on aphid populations than natural levels of predators. Indeed, the action of certain predators (e.g. spiders) and parasitoids could be important in determining the degree of overcompensation that occurs when aphids reach a high density.

The numerical responses of the more generalist syrphid larvae, coccinellids and anthocorids remained largely independent of prey levels in autumn. This was probably due to the: (a) less synchronised lifecycle of the generalist predator to prey occurrence (Honek, 1983), (b) predators having a slower development rate relative to that of aphids (Mills, 1982), (c) predators exhibiting little prey specificity (Frazer et al., 1981b), (d) predators being easily satiated at moderate to high aphid densities, and/or (e) generalist predators switching to alternative hosts to supplement their diets in times of aphid rarity to avoid starvation (Dixon & Russel, 1972; Anderson, 1962). In particular, the inefficiency of predators in capturing sycamore aphids may account for the weak response they show to increasing aphid numbers. For example, adult coccinellids (Dixon, 1970c) and anthocorids (Dixon & Russel, 1972) that feed on sycamore aphids find it difficult to capture sufficient aphids to mature all their eggs even though the aphid population may be very high (Dixon, 1958; Dixon & Russel, 1972).

56 Chapter 2 Population dynamics and life history responses to aphid manipulations

It was noticed that when D. platanoidis was abundant in spring, it does not reach very high numbers in the following autumn (Fig. 2-2) despite the third clip cage experiment showing high reproductive output (Fig. 2-8). Since the abundance of predators and parasitoids are determined by the numbers of aphids during spring, natural enemies are likely to be more numerous at the beginning of autumn (e.g. Dixon & Russel, 1972) when nutritional conditions are unfavourable for the aphid (Mordvilko, 1908; Dixon, 1985). Consequently, natural enemies could then have a marked influence on the rate of increase of the field aphid population, by killing colossal numbers of young aphids in autumn when aphid fecundity is low. Anthocorids are at their most voracious (Dixon & Russel, 1972) and primary parasitoids kill aphids (Hamilton, 1973, 1974) when sycamore aphids come out of aestivation. Hence, this results in a more overcompensated response than would otherwise occur if plant factors were operating alone (Dixon, 1979).

Weaknesses of this study:

In the second cage experiment, aphid mortality was greater on the newly 'long sprayed' saplings (treatment B), so that the life history results collected from the surviving individuals could be biased. This may have been due to residual aphicide carry-over effects, despite preliminary insecticidal trials in the laboratory (see Appendix). A more likely explanation is the low host quality and high summer temperatures creating a hostile environment for nymph survival (Dixon, 1985, 1987). Another potential caveat of this study was that aphids of the third clip cage experiment may be placed under the biased effects of only saplings with late leaf fall timing, which may therefore adjust the response of the third generation aphids.

Knowledge of predator interference is not known from the way this study was conducted, since only a 'snapshot' impression of the predators and prey present in the field at the time of weekly monitoring was obtained (see Frazer et al., 1981b). However, previous work does give an indication of what interactions are likely to occur. In the sycamore system, anthocorid nymphs have been shown to feed on chrysopid and syrphid larvae, spider egg sacs, parasitised mummies, and occasionally on Periphyllus aceris besides D. platanoidis (Dixon & Russel, 1972).

Aphid specific natural enemies are frequently under-estimated. Vickerman & Sunderland (1975) reported on the activity and abundance of aphid predators in cereal

57 Chapter 2 Population dynamics and life history responses to aphid manipulations

crops at night and suggested that the nocturnal activities of predators has been largely overlooked. Frazer & Gilbert (1976) observed that coccinellids spend most of their time hidden at the base of plants, and were more frequently seen at higher temperatures, when they were more mobile. Also, the smaller syrphid larvae can be overlooked due to their camouflage, and often were underestimated because more pupae than larvae are found in sample counts (Dean, 1974).

Since this study only looked at crude seasonal abundance of aphidophagous natural enemy types, no consideration has been made for the very variable nature of each specific predator's population densities. Perhaps future field based studies could be extended to take account of the fluctuations in the predator complex composition from year to year, differences between sites, and during a season within the sycamore system in relation to manipulated plant quality.

Aphid infestation history and leaf abscission:

Manipulating spring aphid densities had no impact on the rate and time of leaf fall in saplings. This result was unexpected (see Faeth et al., 1981; Addicott, 1982; Kahn & Cornell, 1983) considering that sap-extraction causes water loss and feeding punctures (Whittaker, 1984). Consequently this may intensify water loss from damaged leaf surfaces. Therefore drought stress may ensue in the undamaged areas and cause early abscission (Whittaker & Warrington, 1985). The outcome of this study sharply contrasts with the work by White (1970) showing that previously infested sycamore saplings retained their leaves for a shorter period than uninfested saplings. Chambers (1979) suggests that high autumnal population densities contribute to an earlier time of leaf fall, especially if densities were high throughout the year. The biological implications of early leaf fall would be the truncation of oviparae growth and maturation time of their eggs, so that fewer eggs are deposited in bark crevices.

Although this study focussed on aphids infesting saplings, there may be more subtle implications for the mature tree, in terms of affecting the reproductive fitness of its host (see Crawley, 1983; Jermy, 1984; Edwards, 1989) via modification of fruiting times, seed quantity and/or quality. Although the rate of leaf fall remained independent of spring infestation history, other elements of host phenology (e.g. bud burst the next year) may be affected. Field observations by Leather (1996b) suggest that those branches that supported high aphid populations in the previous year had accelerated bud

58 Chapter 2 Population dynamics and life history responses to aphid manipulations burst, compared to those branches that had few aphids on them the previous year; thereby evading newly hatched fundatrices.

CONCLUSION

Many long-term studies of tree-dwelling aphids indicate that aphid-induced changes in the host plant are implicated in subsequent changes in aphid numbers (Dixon, 1979). Therefore, attention was paid to defining the role of the plant in this interaction. The conclusion of this study is that the density of aphids is linked to host plant quality. Caged sycamore aphids responded positively to herbivore-burdened saplings in spring. Hence, this result lends support to the 'herbivore-induced stress hypothesis'. However, by autumn time the response was reversed. This suggests that previous heavy infestation caused major nutrient depletion of the host, which led to a reduction in caged aphid development and reproductive performance. By modifying host plant quality, not only do direct effects act upon insect response, but indirect effects occur within a tri- trophic framework (Meyer & Root, 1996). All five natural enemies demonstrated a positive numerical response to increased aphid densities on herbivore-stressed plants, but by autumn only territorial spiders and the specialist primary parasitoid mummies showed high association with its host abundance. In view of the lack of close association between prey abundance and numerical response among the assemblage of predators in the field during autumn, it is suggested that aphidophagous predators are unlikely to be reliable agents for regulating aphid populations. In contrast, the specialist parasitoids showed a closely bound numerical association with its prey, which may exacerbate the delayed overcompensated negative feedback in autumnal aphid abundance imposed by the plant factor. Finally, heavy infestation history did not lead to premature senescence, which in turn did not curtail the longevity of parthenogens/sexual morphs in autumn.

Perhaps further research could follow through the effect of aphid densities over the next year, to see whether leaf phenology (i.e. timing of bud burst/leaf fall) is affected by carry over effects of previous manipulated populations. Often the consequences of herbivorous damage may not be seen earlier than the year after the treatment was conducted (Neuvonen et al., 1988), leading researchers to potentially erroneous conclusions (Haukioja & Honkanen, 1997). In addition, predator exclusion experiments could be carried out over a matter of a few years to measure the impact of specific predators on cyclic dynamics, such as that undertaken by Torgersen (1985).

59 Chapter 3 Endophyte infection and the sycamore-aphid-parasitoid interaction

Does endophytic fungal infection affect the dynamics of the sycamore-aphid-parasitoid interaction?

ABSTRACT

1. Endophytic fungi are diverse and abundant in woody plants. They are generally viewed as plant mutualists since they provide enhanced resistance to herbivores perhaps through induced resistance, nutrient depletion or toxin production. 2. This study investigates a system where an arboreal endophyte has been shown to enhance the abundance of phytophagous insects, as well as to increase the nutrient status of infected foliage (see Gange, 1996). 3. The study examines the post-aestivation population dynamics of the sycamore aphid Drepanosiphum platanoidis in relation to its primary parasitoid Aphelinus flavus and the specialised endophyte Rhytisma acerinum on sycamore Acer pseudoplatanus saplings. 4. With an increase in foliar nitrogen (as suggested by Gange, 1996), this study hypothesised that aphids should show an early release from aestivation and undergo higher rates of post-aestivation reproduction. With an elevation in aphid abundance, parasitoid attack should increase thus driving down the aphid population. It is predicted that within a tri-trophic context, the endophyte should indirectly (via the action of parasitoids) cause a reduction in aphid abundance, which in effect nullifies the positive plant-endophyte-herbi yore association. 5. To test this set of hypotheses, a manipulative field experiment was set up in August 1997, using six treaments within a random test block design, to monitor the autumnal aphid populations on sycamore in relation to the presence or absence of the endophyte and the parasitoids. 6. There was no significant difference in the time to aestivation release between R. acerinum infected and uninfected plants in pure (no-choice) and mixed (dual- choice) culture treatments. 7. There was no significant effect of R. acerinum infection on aphid abundance in most categories (i.e. total aphids, early and late nymphs, adult parthenogens, adult males) and the number of parasitoid mummies in pure and mixed treatments. Only

60 Chapter 3 Endophyte infection and the sycamore-aphid-parasitoid interaction

oviparae abundance was significantly reduced on R. acerinum infected saplings in pure culture. 8. It was suggested that mycotoxins or fungal-induced plant defences were possible factors responsible for reduced oviparae abundance on R. acerinum infected sycamores, while the general lack of aphid choice may be due to insufficient visual sensitivity to colour differences between endophyte infected and uninfected host plants. 9. In sum, the generality of mutualisms between endophyte fungi and their host plants, as indicated by antagonistic interactions with herbivores, is not confirmed by this study. If arboreal endophytes do have an effect then they are possibly more complex and indirect (e.g. altering plant phenology) than was originally thought.

INTRODUCTION

Host plant quality for insect herbivores is profoundly influenced by microbial mutualists associated with the plant (Rabin & Pacovsky, 1985; Clay, 1988; Letourneau, 1988; Hammon & Faeth, 1992). Plant symbionts may exert pernicious effects on herbivores by the production of microbial toxins (Miller, 1986; Prestidge & Gallagher, 1988; Johnson & Whitney, 1994; Dahlman et al., 1991), by induction of plant defences (McIntyre et al., 1981; Karban et al., 1987), or by the influence of microbes on tri- trophic interactions (Dicke, 1988). Alternatively, the presence of microbes may improve plant quality for insect herbivores (Kennedy, 1951; Carruthers et al., 1986; Johnson & Whitney, 1994). Consequently, the extent and importance of micro- organisms in population dynamics and interactions among species is not yet clear, and understanding the ecology of micro-organisms in communities is far from complete (Thompson, 1982; Clay, 1990; Hammon & Faeth, 1992).

Plant symbionts (e.g. fungus, bacteria) may directly or indirectly influence trophic interactions involving host plants, herbivores, and natural enemies of insect herbivores (Hammon & Faeth, 1992; Hunter & Price, 1992; Preszler et al., 1996). Tri-trophic interactions involving microbial symbionts have received relatively little attention, apart from that described by Dicke (1988). Fungal activity in foliar tissue can influence higher trophic interactions through the production of mycotoxins or by the induction of plant allelochemicals (Wilson, 1993). Foliar micro-fungi may produce their own chemical compounds which are directly toxic to herbivores that feed on the host plant

61 Chapter 3 Endophyte infection and the sycamore-aphid-parasitoid interaction

(e.g. Miller, 1986), or indirectly to natural enemies of the herbivore once it has ingested the toxin (e.g. Price et al., 1980). Similarly, allelochemicals can exert positive or negative tri-trophic effects. They can elevate plant defences by reducing herbivorous insect performance (Karban et al., 1987; Butin, 1992; Krause & Raffa, 1992; Wilson, 1995b; Hammon & Faeth. 1992) and/or by attracting natural enemies (Whitman, 1988; Faeth, 1994) such as parasitoids (Vinson, 1975; Price 1981; Elzen et al., 1983; Whitman, 1988). Alternatively, they can diminish plant defences as a result of toxic effects on parasitoids (Campbell & Duffy, 1979; Thorpe & Barbosa, 1986; Bultman et al., 1997). In addition to allelochemistry, micro-organisms may alter host quality by changing nutrition and structure of the plant (see Barbosa & Letourneau, 1988). The induced response, however, may or may not increase resistance of the plant to further herbivore damage (Fowler & Lawton, 1985; Karban & Myers, 1989).

One group of symbiotic micro-organisms usually associated with the aerial portion of plants, living asymptomatically and internally within plant tissues, are the endophytic fungi (Wilson, 1993, 1995a; Saikkonen et al., 1998). Endophytic fungi inhabiting the foliage of woody plants have been far less studied than endophytes of grasses, despite being more diverse and abundant than those in grasses (Petrini, 1986, Petrini et al., 1992). Communities of endophytic fungi are found in all woody gymnosperms and angiosperms that have been examined (Petrini, 1986). The great abundance and diversity of endophytic fungi and other micro-organisms in woody plants provide the potential for a wide variety of direct (via mycotoxins) and indirect (by altering the host plant) interactions between insect herbivores and micro-organisms (Petrini et al., 1992; Wilson, 1993; Faeth & Hammon, 1997a). Endophytic fungi in woody plants, like those in grasses, are also thought to confer resistance to insect herbivores (Carroll, 1986, 1988, 1991; Strong, 1988; Johnson & Whitney, 1994). Since endophytes of woody plants are diverse and have shorter life cycles than their perennial host plants, defence via endophytes is considered a mechanism by which long-lived woody plants could keep pace evolutionarily with shorter generational and, hence, presumably more rapidly evolving invertebrate herbivores (Carroll, 1988). However, accumulating evidence suggests that endophytic fungi interact with each other and with insect herbivores across the spectrum of negative (Butin, 1992; Faeth & Wilson, 1996; Gaylord et a/.,. 1996; Taper et al., 1986; Wilson & Carroll, 1994; Wilson, 1995b) to neutral (Faeth & Hammon, 1996, 1997a, 1997b; Saikkonen et al., 1996) to positive (Johnson & Whitney,

62 Chapter 3 Endophyte infection and the sycamore-aphid-parasitoid interaction

1994; Gange, 1996; Gaylord et al., 1996; Preszler et al., 1996) as is expected on theoretical grounds (Hammon & Faeth, 1992).

Despite the enormous potential for interactions between endophytes and insect herbivores that share the same woody angiosperms as hosts, there have been very few experimental manipulations of endophytes that examine these interactions (except those of Wilson, 1995b; Faeth & Hammon, 1997a,b). Moreover, there have been no manipulative investigations that examine the effect of endophytes on higher trophic level interactions (i.e. herbivore-natural enemy) which take into account their population dynamics in the field. However, existing work suggests that endophyte infected trees and grasses do render detrimental effects on parasitoid performance. Observational work by Preszler et al. (1996) found that survival of leafminers, Phyllonorycter sp., was positively associated with the endophytic fungus Gnomonia cerastis owing to reduced parasitism of larvae on Gambel oak with higher endophyte infections. Similarly, but not in woody plants, the attack rate of Microctonus hyperodae was reduced when the quality of potential host weevils, Listronotus bonariensis, were compromised by confinement on non-preferred Acremonium /o/ii-infected ryegrass, since endophyte presence increased development time and reduced survival in the parasitoid (Barker & Addison, 1996). Furthermore, the presence of the endophyte Acremonium coenophialum on tall fescue in the diet of fall armyworm Spodoptera fruigiperda had a negative impact on the pupal mass of two Euplectrus parasitoids, although development rate and survival remained unaffected (Bultman et al., 1997).

The present study investigates the potential effects of the endophytic fungus Rhytsma acerinum, in sycamore (Ater pseudoplatanus L.), on the population dynamics of the sycamore aphid (Drepanosiphum platanoidis Schr.) and its parasitoid (Aphelinus flavus Thompson). The study of a tree-endophyte-aphid-parasitoid system is of particular interest for two reasons. Firstly, aphids have been shown to be particularly sensitive to endophyte presence in infected grasses (Wilson et al., 1991; Eichenseer & Dahlman, 1992). This study tests whether the patterns found in grass systems can be found for tree endophytes. Secondly, the potential for higher trophic level effects, mediated by fungal interference via host chemistry, remains to be tested in insect populations in a manipulated field experiment.

63 Chapter 3 Endophyte infection and the sycamore-aphid-parasitoid interaction

This study extends the field based observational work by Gange (1996) who suggests that R. acerinum infection gives rise to localised elevation of foliar nitrogen, and enhances the abundance of two types of sycamore aphid, D. platanoidis and Periphyllus acericola, particularly in late summer. The aim was to determine whether R. acerinum infection in sycamore saplings affects the population dynamics of the aphids in the presence and absence of a primary parasitoid. This study presents four hypotheses: with an increase in nitrogen availability to the sycamore aphid there should be (i) an early release from aestivation; (ii) an enhanced reproductive rate in the second and third generation parthenogenetic virginoparae; (iii) an increase in parasitism in the presence of R. acerinum; and (iv) an aphid preference for endophyte infected saplings in a choice situation. To test these predictions, a manipulation experiment was set up in semi- natural conditions using a field insectary to assess the numerical performance of D. platanoidis populations over a 13-week post-aestivation period in relation to the presence or absence of an endophytic fungus and a primary parasitoid in a choice and no-choice situation.

NATURAL HISTORY

The ascomycete fungus, Rhytisma acerinum (Pers.) Fries, forms black stromata known as tar spots on the adaxial surface of the leaves of sycamore, Acer pseudoplatanus L. (Jones, 1925). R. acerinum has been shown to be specific to A. pseudoplatanus (Muller, 1912). This fungus overwinters in the fallen leaves of the previous season, ejecting gelatinous sheathed ascospores from swollen apothecia in suitable weather during May and June (Butler & Jones, 1961). On reaching and successfully infecting host tissue (newly expanded sycamore leaves), there is a vegetative asymptomatic phase of 6-8 weeks (Muller, 1912), at the end of which small black dots surrounded by a yellow perimeter appear on the adaxial leaf surface in July (Leith & Fowler, 1987; Gange, 1996). The black stromata (plate 3-1b) increase in size during the summer, eventually forming the characteristic circular tar spots of approximately 15 mm diameter with an average area of 1.0 cm2 (Jones, 1925; Leith & Fowler, 1987). Bevan & Greenhalgh (1976) report that factors such as wetness of the leaf surface, wind speed and temperature are important for successful infection of a leaf by R. acerinum. Also, a localised source of inoculum is necessary for R. acerinum infection, since tar spots on leaves decrease with height above ground and horizontal distribution is limited within 5-10 m (Leith & Fowler, 1987). Although R. acerinum has been listed in most plant

64 Chapter 3 Endophyte infection and the sycamore-aphid-parasitoid interaction

pathology texts (e.g. Brooks, 1953; Holliday, 1992), and fits Wilson's (1995a) definition of endophyte well, it is still unclear whether and when R. acerinum acts as a symbiont or a pathogen (see Gange, 1996).

The sycamore aphid, Drepanosiphum platanoidis Schrank (Homoptera: Aphididae), is monoecious (i.e. living and feeding on a single host) and is common all over Europe where its host plant grows. Its population biology has been extensively studied (Dixon, 1979, 1990e). This aphid overwinters as an egg, which hatches in early spring (March — April), giving rise to the first generation (fundatrices) which develops on unfurling buds. The second generation feeds on the maturing leaves and enters aestivation (i.e. reproductive diapause during summer) (Mordwilko, 1908; Dixon, 1963, 1966), where virtually only adult parthenogenetic virginoparae are present at the end of June and during July. The duration of aestivation is controlled by prevailing high temperatures (Dixon, 1979), impoverished food quality (Dixon, 1963, 1970a, 1979), maternal experience of fundratrices (Chambers, 1982), spring aphid numbers (Dixon, 1975a), and the density of second generation adults (Dixon, 1970a; Chambers et al., 1985). Reproduction recommences in mid to late August. From early September, parthenogenetic sexuparae begin to give rise to sexual morphs (apterous oviparae and alate males) with the proportion increasing towards the end of the aphid season (Dixon, 1971b).

Aphelinus flavus Thompson (: Aphelinidae) is a specialised solitary endoparasitoid, which parasitises small D. platanoidis nymphs (i.e. mainly first and second instars) (Hamilton, 1969, 1973). Adult male and female A. flavus appear in May and are known to feed frequently on aphid honeydew. The female kills 1 aphid per 1.7 eggs deposited, and parasitises 48 aphids over 27 days (Hamilton, 1973). Oviposition occurs 23 days prior to the appearance of black mummies. Development from egg to adult requires 57 days. The parasitoid has two generations in the season, being divided by the absence of suitable hosts during aestivation, after which the parasitoid becomes active again and continues ovipositing as long as hosts are present. Aphelinus flavus overwinters either as final instar within mummies attached to fallen leaves (Start, 1970), or as free living adults (Hamilton, 1973).

65 Chapter 3 Endophyte infection and the sycamore-aphid-parasitoid interaction

MATERIALS AND METHODS

Preparation:

(i) Tarspot innoculation: Seventy two 12 month old saplings (c. 0.75 m) grown in 20 cm diameter plastic pots, containing John Innes No. 2 loam-based compost, under greenhouse conditions were placed in two outdoor shaded locations in late March 1997. Half were exposed to Rhytisma acerinum infection by placing saplings on caged trays containing tarspot infected leaf litter. During late April to early June, the field inoculation site was kept moist by spraying water twice daily to facilitate spore release.

(ii) Aphid cultures: Adult fundatrices of D. platanoidis, collected from the field during April, were put into culture and reared on potted 15 month old sycamore saplings under a photoperiod of LD 16:8 h at 204-2°C within five 70 cm3 perspex chambers (which represented summer conditions under which aphids enter aestivation).

(iii) Parasitoid cultures: Mummies of the parasitoid Aphelinus flavus were collected from the field during June and July and placed individually into gelatine capsules. Once sexed and hyperparasitoids discarded, the emergent adult parasitoids were placed in three 70 cm3 perspex chambers, each containing aphid infested saplings, under a photoperiod of LD 16:8 h at 20-1-2°C. In late August, adult female parasitoids were transferred individually to 5 x 1 cm glass vials containing cotton-wool moistened with 50:50 honey/water solution ready for release.

Experimental design:

An 8 x 8m field insectary (plate 3-2), constructed of a skeleton metal framework covered in nylon netting, was established in the Deer Park sector of Silwood Park, Berkshire (OS ref 5946 1689). The floor of the insectary was covered with tarpaulin sheeting and anchored with sand bags. An eighteen blocked system was devised to accommodate three replicates for each of six treatments. The eighteen adjacent 6.25 m2 (2.5 m length x 1.25 m wide x 2 m height) blocks were separated by netting walls (Plate 3-3), held down by strips of wood and sand bags. Access to each block was provided by a sealable vent.

66 Chapter 3 Endophyte infection and the sycamore-aphid-parasitoid interaction

The six treatments (Fig. 3-1) were: (A) aphids + tarspot infected and uninfected sycamore saplings; (B) aphids + tarspot infected sycamore saplings; (C) aphids + uninfected sycamore saplings (the control); (D) aphids + tarspot infected and uninfected sycamore saplings + parasitoids; (E) aphids + tarspot infected sycamore saplings + parasitoids; (F) aphids + tarspot uninfected sycamore saplings + parasitoids.

In early August 1997, the sycamore saplings were placed in the field insectary, with four potted sycamores per block, according to a Latin block arrangement (Fig. 3-2). On the 10/08/97, 5 m1/1 Levington's 'Deep Feed' (7-1.3-3.3% NPK solution containing 0.018% Mg with trace elements) fertilizer was added to the soil of each pot. On the 17/08/97, each sapling was inoculated with 10 aestivating aphids. Five days later, 23/08/97, fifteen A. flavus individuals were released amongst the saplings in blocks D to F. Once all the treatments had been set up in this pulse experiment (Crawley, 1993), the experimental arenas were securely sealed, and left until leaf fall.

Fig. 3-1: Six treatments were established: three in the presence and three in the absence of the parasitoid Aphelinus flavus. The mixed cultures (A and D) using Rhytisma acerinum infected and uninfected saplings should show the behavioural choice of aphids and parasitoids, whereas the pure cultures (B, C, E and F) show aphids and/or parasitoids constrained to R. acerinum infected or uninfected treatments. The arrows signify statistical comparisons made between and within treatments.

67

Chapter 3 Endophyte infection and the sycamore-aphid-parasitoid interaction

Monitoring regime:

When the second generation parthenogenetic virginoparae started to reproduce (c. 20/8/96), aphid and parasitoid populations were monitored weekly from 24/08/97 to 16/11/97. Since this period encompassed two post-aestivating aphid generations (the parthenogenetic sexuparae and the sexuals), and included a non-diapausing followed by one potentially winter diapausing generation in the primary parasitoid, records were made of the densities found on each of five leaves per sapling in the following categories: (i) the time of first appearance of third generation nymphs; (ii) total aphids; (iii) instars 1-3 aphids; (iv) fourth instar aphids; (v) adult parthenogenetic virginoparae; (vi) oviparae; (vii) males; (viii) parasitoid mummies. Counting five leaves per sapling was a good estimator for density because most saplings had only 10-18 leaves.

Fig. 3-2: The design of the enclosed experiment involved arranging six treatments (replicated three times) according to a Latin block system within the field insectary (8 m x 8 m x2 m). Four saplings were placed in each of the eighteen blocks (2.5 m x 1.25 m x 2 m) two weeks before the introduction of aphids and parasitoids (Plates 3-2 and 3-3).

Data transformation and statistical analyses:

Since this was a balanced experimental design (Fig. 3-1), planned comparisons between treatments could be made as follows: (i) aphid response to R. acerinum infected and

68 Chapter 3 Endophyte infection and the sycamore-aphid-parasitoid interaction uninfected sycamore (treatment A) in the absence of parasitoids; (ii) aphid response to R. acerinum infected sycamore (treatment B) with uninfected sycamore (treatment C) in the absence of parasitoids; (iii) aphid response to R. acerinum infected and uninfected sycamore (treatment D) in the presence of parasitoids; (iv) parasitoid response to aphid infested saplings that are R. acerinum infected (treatment E) with uninfected sycamore (treatment F); (v) aphid density within mixed cultures in the absence (treatment A) and the presence (treatment D) of parasitoids; (vi) aphid density to R. acerinum infected sycamore in the absence (treatment B) and presence (treatment E) of parasitoids; and finally (vii) aphid density to uninfected sycamore in the absence (treatment C) and the presence (treatment F) of parasitoids.

Prior to statistical analysis, the response variables (categories i-viii) of the 'between- treatment' data set (i.e. all six treatments) were prepared in Excel using pivot tables. The collected data on aphid and parasitoid performance was amalgamated over the thirteen weeks, thus discarding the time factor from the summary data. The number of aphids and parasitoid mummies found per leaf were averaged over the whole sapling (n leaves/sapling = 5), and again over the four saplings (pseudo-replicates), to derive a block (n saplings/block = 4) mean. The averages for each of the three blocks (true replicates) per treatment were used for the analysis. The response variables for the 'within-treatment' data set (i.e. mixed cultures in treatments A and D) were averaged in a similar way. However, the tarspot infected and the uninfected saplings within treatments A and D represented two separate treatments. Thus, the mean number of aphids or parasitoid mummies found per leaf over the two saplings (pseudo-replicates) of a treatment were averaged to form a block average. In turn, the three blocks (true replicates) were averaged to form a treatment mean. Finally, the derived (or cumulative values) response variables were transformed by log (Y+1).

Two-way analysis of variance using normal error structures were performed in GLIM (version 3.77 copyright 1985 Royal Statistical Society, London; Crawley, 1993) on both the between- and within-treatment data sets for each of the derived response variables. For the between-treatment analysis, the factors were 'treatment' with three levels, and `parasitoids' with two levels; whereas the within-treatment analysis, the factors were `treatment' with two levels and `parasitoids' with two levels. To determine the proportion of parasitism, in both the between- and within-treatment data sets, one-way ANOVAs with binomial error structure were performed on untransformed data. The

69 Chapter 3 Endophyte infection and the sycamore-aphid-parasitoid interaction derived number of mummies was expressed as the numerator, and the derived number of nymphal instars 1-3 (n.b. this is the susceptible aphid category to parasitoid attack) was declared as the binomial denominator.

The time of aestivation release, as signalled by the first appearance third generation nymphs, was analysed with the Kruskal-Wallis test (Sokal & Rohlf, 1995) in Excel. Prior to non-parametric analysis the data was transformed by giving time of nymphal appearance a code. Since monitoring was only undertaken every seven days, the first week was given the code one and thereafter each subsequent week was given an accumulative code (i.e. increasing by one each week). For the between treatment comparisons, a total block score (i.e. block = a treatment replicate with 4 saplings) was calculated by determining the time of appearance for each sapling and ranking them accordingly within the non-parametric test. In a similar way, within-treatment comparisons (i.e. mixed cultures A and D) were determined, but the blocks used the cumulated time of appearance score from 2 saplings.

RESULTS

Between-treatment differences

(i) Time of aestivation release:

The timing of third generation D. platanoidis nymph appearance was not significantly different between the six treatments (Kruskal-Wallis: H = 4.52, d.f. = 5, P = n.s.; Fig. 3- 3).

70 Chapter 3 Endophyte infection and the sycamore-aphid-parasitoid interaction

Plate 3-1 a: Healthy sycamore (Acer pseudoplatanus) leaf.

Plate3-1b Rhydstra orPrinum infected Plate 3-3: The set up of an experimental 1=1=1 block containing four sycarnoie leaf showing black tar spots. sycamore saplings.

Plate 3-2: Held insectary at Silwood Park used for the endophyte manipulation experiment

71

Chapter 3 Endophyte infection and the sycamore-aphid-parasitoid interaction

4.0 - parasitoids + parasitoids

sa)rs- 3.5 ix) c T 0 o 3.0 T (i) 2.5 2 2.0

o 1.5 T E ° tO 1.0 c > 0.5 2 ca 0.0 A Treatment

Fig. 3-3: Mean time of appearance of third generation Drepanosiphym platanoidis nymphs (± S.E.) after aestivation on Acer pseudoplatanus saplings in relation to Rhytisma acerinum infection and/or Aphelinus flavus presence. Three treatments are shown without parasitoids (A = mixed, B = R. acerinum, C = no R. acerinum) and three treatments with parasitoids (D = mixed, E = R. acerinum, F = no R. acerinum).

(ii) Total aphid abundance:

The total number of D. platanoidis individuals did not significantly differ between

mixed, R. acerinum infected and uninfected cultures (F 2,12 = 2.43, r2 = 0.27, P = n.s.;

Fig. 3-4), or between treatments with and without parasitoids (F 1,12 = 0.28, r2 = 0.016, P = n.s.). The endophyte treatment x parasitoid interaction was also not significant (F 2,12 = 0.26, r2 = 0.030, P = n.s.).

3.0 4t-ts - parasitoids + parasitoids a) co 2.5 o_ T T E co 2.0 Cl E2 o 1.5 T _c E + 1.0 x c -E- coc —I 0.5 a) 2 0.0 A B C D E F Treatment

Fig. 3-4: Total abundance of Drepanosiphum platanoidis (± S.E.) on Acer pseudoplatanus saplings in relation to Rhytisma acerinum infection and/or Aphelinus flavus presence. The means presented (from three block replicates) were for the cumulative average number of aphids over the 13-week experimental period. Three treatments are shown without parasitoids (A = mixed, B= R. acerinum, C= no R. acerinum) and three treatments with parasitoids (D = mixed, E = R. acerinum, F = no R. acerinum).

72 Chapter 3 Endophyte infection and the sycamore-aphid-parasitoid interaction

(iii/iv) Nymphal abundance:

No significant differences were found in early and late D. platanoidis nymphal instars between mixed, R. acerinum infected and uninfected cultures (instars 1-3: F 2,12 = 2.47, r2 = 0.28, P = n.s.; instar 4: F 2,12 = 1.84, r2 = 0.22, P = n.s.; Fig. 3-5), or between treatments with and without parasitoids (instars 1-3: F 1,12 = 0.45, r2 = 0.026, P = n.s.; instar 4: F 1,12 = 0.33, r2 = 0.020, P = n.s.). The endophyte treatment x parasitoid interaction was not significant either (instars 1-3: F 2,12 = 0.13, r2 = 0.015, P = n.s.; instar 4: F 2,12 = 0.28, r2 = 0.034, P = n.s.).

2.5 9E3 - parasitoids + parasitoids a) a> 0- a)E 2.0 (I) El Instars 1-3 _c o II Instar 4 cc 1.5 4-• 0

1 . 0 E c c —1 0.5 a)co 2 0.0 i I i A Treatment

Fig. 3-5: Abundance of early (instars 1-3) and late (instar 4) nymphs of Drepanosiphum platanoidis (± S.E.) on Acer pseudoplatanus saplings in relation to Rhytisma acerinum infection and/or Aphelinus flavus presence. The means presented (from three block replicates) were for the cumulative average number of aphids over the 13-week experimental period. Three treatments are shown without parasitoids (A = mixed, B= R. acerinum, C= no R. acerinum) and three treatments with parasitoids (D = mixed, E = R. acerinum, F = no R. acerinum).

(v) Adult parthenogentic virginoparae abundance:

There were no significant differences in the number of D. platanoidis adult virginoparae between mixed, R. acerinum infected and uninfected cultures (F 2,12 = 1.18, r2 = 0.16, P

= n.s.; Fig. 3-6), or between treatments with and without parasitoids (F 1,12 = 0.11, r2 = 0.0074, P = n.s.). In addition, the endophyte treatment x parasitoid interaction was not significant (F 2,12 = 0.16, r2 = 0.022, P = n.s.).

73

Chapter 3 Endophyte infection and the sycamore-aphid-parasitoid interaction

- 0.9 - parasitoids + parasitoids 1)

+ 0.8

lt (x

du 0.7 Ln f a f ( -o 0.6 o r lea • 0.5 be er m

p • 0.4 nu

ens 0.3

Mean 0.2 henog t 0.1 ar p 0 I 1 11 A B C D E F Treatm ent Fig. 3-6: Abundance of adult parthenogenetic virginoparae of Drepanosiphum platanoidis (+ S.E.) on Acer pseudoplatanus saplings in relation to Rhytisma acerinum infection and/or Aphelinus flavus presence. The means presented (from three block replicates) were for the cumulative average number of aphids over the 13-week experimental period. Three treatments are shown without parasitoids (A = mixed, B= R. acerinum, C= no R. acerinum) and three treatments with parasitoids (D = mixed, E = R. acerinum, F = no R. acerinum).

(vi/vii) Adult sexual morph abundance:

On uninfected plants significantly more oviparae were produced (F 2,12 = 4.98, r2 = 0.42, P<0.05; Fig. 3-7). No significant endophyte treatment effect was found for the

production of males (F 2,12 = 2.04, r2 = 0.24, P = n.s.). Treatments with parasitoids present were not significantly different from those without parasitoids with respect to production of sexual morphs (oviparae: F 1,12 = 0.19, r2 = 0.0079, P = n.s.; males: F 1,12 = 0.21, r2 = 0.012, P = n.s.). The endophyte treatment x parasitoid interaction was not significant (oviparae: F 2,12 = 0.71, r2 = 0.060, P = n.s.; males: F 2,12 = 0.38, r2 = 0.044, P = n.s.).

f 1.2 - parasitoids + parasitoids lea

r d) e 1.0 Oviparae p ls rme ®Males fo 0.8 ns f sexua tra 0.6 o - ber + 1)

m 0.4 nu Ln(x ( 0.2 Mean 0.0 A B C D E F Treatment Fig. 3-7: Abundance of Drepanosiphum platanoidis adult sexuals (± S.E.) on Acer pseudoplatanus saplings in relation to Rhytisma acerinum infection and/or Aphelinus flavus presence. The means presented (from three block replicates) were for the cumulative average number of aphids over the 13- week experimental period. Three treatments are shown without parasitoids (A = mixed, B= R. acerinum, C= no R. acerinum) and three treatments with parasitoids (D = mixed, E = R. acerinum, F = no R. acerinum).

74 Chapter 3 Endophyte infection and the sycamore-aphid-parasitoid interaction

(viii) Proportion of parasitism:

Parasitism was extremely low in each treatment block over the entire 13-week period (range: 0.00 — 0.60 mummies/early instar nymphs), and no significant differences in the proportion of parasitised nymphal instars 1-3 was found between mixed, R. acerinum infected and uninfected treatments (one-way ANOVA: X2 = 0.2807, d.f. = 2, P = n.s.; Fig. 3-8).

0.50 "El a) 4-0- 15 0.40 c 0 cn • fl 0.30 a_ E 2 • 0.20 C a)}, 2 7) 0.10

0.00

D F F Treatment

Fig. 3-8: Number of Aphelinus flavus mummies as a proportion of available Drepanosiphum platanoidis nymphs (instars 1-3) on Acer pseudoplatanus saplings in the presence and absence of Rhytisma acerinum infection. The means (from three block replicates) used in determining proportions were taken from the cumulative average number of instars 1-3 and parasitoid mummies over the 13-week experimental period. Three treatments are shown with parasitoids (D = mixed, E = R. acerinum, F = no R. acerinum).

Within-treatment differences (i.e. mixed cultures of treatments A and D)

(i) Time of aestivation release:

Third generation D. platanoidis nymph appearance was not significantly different between the four treatments in mixed cultures (Kruskal-Wallis: H = 1.22, d.f. = 3, P = n.s.; Fig. 3-9).

75 Chapter 3 Endophyte infection and the svcamore-aphid-parasitoid interaction

3.0 - parasitoids + parasitoids /97 fter

a 2.5 24/08 T

ion t m

fro 2.0 duc

ro ks 1.5 rep T wee to in

( 1.0 n ime t io

t n

iva 0.5 t Mea aes 0.0 A (infected) A (uninfected) D (infected + D (uninfected + parasitoids) parasitoids) Treatment Fig. 3-9: Mean time of appearance of third generation Drepanosiphum platanoidis nymphs (± S.E.) after aestivation on Acer pseudoplatanus saplings in relation to Rhytisma acerinum infection and/or Aphelinus flavus presence within mixed culture treatments. Two treatments are shown without parasitoids (R. acerinum infected and uninfected) and two treatments with parasitoids (R. acerinum infected and uninfected).

(ii) Total aphid abundance:

Aphids did not chose between endophyte infected and uninfected saplings, and no significant difference was found in the total number of D. platanoidis individuals

among treatments (F 1,8 = 2.40, r2 = 0.23, P = n.s.; Fig. 3-10). The total number of aphids was not significantly affected by parasitoid presence when comparing the mixed cultures of treatments A and D (F 1,8 = 0.11, r2 = 0.010, P = n.s.). The endophyte

treatment x parasitoid interaction was also not significant (F 1,8 = 0.11, r2 = 0.011, P = n.s.).

3.5 - parasitoids + parasitoids a) 3.0 47) o.) 2.5 TS E co co a) 2.0 o 1.5

E x 1.0 C C C —I 0.5

2 0.0

A (infected) A (uninfected) D (infected + D (uninfected + parasitoids) parasitoids) Treatment Fig. 3-10: Total aphid abundance of Drepanosiphum platanoidis (± S.E.) on Acer pseudoplatanus saplings within mixed cultures (treatments A and D) in relation to Rhytisma acerinum infection and/or Aphelinus flavus presence. The means presented (for three block replicates) were for the cumulative average number of aphids over the 13-week experimental period. Two treatments are shown without parasitoids (R. acerinum infected and uninfected) and two treatments with parasitoids (R. acerinum infected and uninfected saplings).

76 Chapter 3 Endophyte infection and the sycamore-aphid-parasitoid interaction

(iii/iv) Nymphal abundance:

No significant differences were found in early and late D. platanoidis nymphal instars between infected and uninfected saplings within the mixed cultures (instars 1-3: F 1,8 =

2.87, r2 = 0.26, P = n.s.; instar 4: F 1,8 = 1.77, r2 = 0.18, P = n.s.; Fig. 3-11). The number of nymphs was also not significantly affected by parasitoid presence (instars 1-3: F 1,8 =

0.12, r2 = 0.011, P = n.s.; instar 4: F 1,8 = 0.11, r2 = 0.011, P = n.s.). In addition, the endophyte treatment x parasitoid interaction was not significant (instars 1-3: F 1,8 =

0.051, r2 = 0.0046, P = n.s.; instar 4: F 1,8 = 0.22, r2 = 0.021, P = n.s.).

3.0 - parasitoids + parasitoids 'Cl) V CL E 2.5 a 0 E u) 2.0 Olnstars 1-3

f Minstar 4

o 1.5 ber x 1.0 num

4ro 0.5 Mean 0.0

A (infected) A (uninfected) D (infected + D (uninfected + parasitoids) parasitoids) Treatment Fig. 3-11: Abundance of early (instars 1-3) and late (instar 4) stage D. platanoidis nymphs (± S.E.) within mixed cultures (treatments A and D) in relation to Rlzytisma acerinum infection and/or parasitoid presence. The means presented (for three block replicates) were for the cumulative average number of aphids over the 13-week experimental period. Two treatments are shown without parasitoids (R. acerinum infected and uninfected) and two treatments with parasitoids (R. acerinum infected and uninfected).

(v) Adult parthenogentic virginoparae abundance:

The number of D. platanoidis adult virginoparae was not significantly different between R. acerinum infected and uninfected saplings within mixed cultures of treatments A and

D (F 1,8 = 1.14, r2 = 0.12, P = n.s.; Fig. 3-12). Likewise, the number of adult virginoparae were not significantly affected by parasitoid presence (F 1,8 = 0.0011, r2 = 0.00012, P = n.s.). In addition, the endophyte treatment x parasitoid interaction showed no significant differences (F 1,8 = 0.0050, r2 = 0.00055, P = n.s.).

77

Chapter 3 Endophyte infection and the sycamore-aphid-parasitoid interaction

•a) p 1.50 - - parasitoids + parasitoids m o a)E a) LA 1.25 - 977;"

CU o_ E 1.00 -0 ca 0.75 o -1-x cf) 0.50

C ▪(D 0.25 -

°- 0.00 A (infected) A (uninfected) D (infected + D (uninfected + parasitoids) parasitoids) Treatment Fig. 3-12: Abundance of Drepanosiphum platanoidis adult parthenogenetic virginoparae (± S.E) on Acer pseudoplatanus saplings within mixed cultures (treatments A and D) in relation to Rhytisma acerinum infection and/or Aphelinus flavus presence. The means presented (for three block replicates) were for the cumulative average number of aphids over the 13-week experimental period. Two treatments are shown without parasitoids (R. acerinum infected and uninfected) and two treatments with parasitoids (R. acerinum infected and uninfected).

(vi/vii) Adult sexual morph abundance:

The number of sexual morphs on R. acerinum infected and uninfected saplings were not

significantly different within mixed cultures (oviparae: F 1,8 = 4.66, r2 = 0.34, P<0.10;

males: F 1,8 = 4.28, r2 = 0.33, P<0.10; Fig. 3-13), and their number was also not

significantly affected by parasitoid presence (oviparae: F 1,8 = 0.36, r2 = 0.026, P = n.s.;

males: F 1,8 = 0.20, r2 = 0.015, P = n.s.). The endophyte treatment x parasitoid interaction was also not significant (oviparae: F = 0.56, r2 = 0.041, P = n.s.; males: F

1,8 = 0.35, r2 = 0.027, P = n.s.).

2.5 - parasitoids + parasitoids

El Oviparae IMMales

I

0.0 -1 A (infected) A (uninfected) D (infected + D (uninfected + parasitoids) parasitoids) Treatments

Fig. 3-13: Abundance of Drepanosiphum platanoidis adult sexuals (± S.E.) on Acer pseudoplatanus saplings within mixed cultures (treatments A and D) in relation to Rhytisma acerinum infection and/or Aphelinus flavus presence. The means presented (for three block replicates) were for the cumulative average number of aphids over the 13-week experimental period. Two treatments are shown without parasitoids (R. acerinum infected and uninfected) and two treatments with parasitoids (R. acerinum infected and uninfected).

78 Chapter 3 Endophyte infection and the sycamore-aphid-parasitoid interaction

(viii) Proportion of parasitism:

Average parasitism in each treatment block was low over the duration of the experiment (range: 0.00 — 0.90 mummies/early instar nymphs), and no significant difference in the proportion of parasitism of young nymphs (instars 1-3) was found between R. acerinum infected and uninfected saplings within the mixed cultures (one-way ANOVA: X2 = 0.32, d.f. = 1, P = n.s.; Fig. 3-14).

0.70

9F3 0.60

c o_ 0.50 0 _c o EE 0.40 CL 2 c 0.30 -0 w N 0.20

0.10

0.00 D (infected + parasitoids) D (uninfected + parasitoids) Treatment Fig. 3-14: Number of Aphelinus flavus mummies as a proportion of available Drepanosiphum platanoidis nymphs (instars 1-3) on Acer pseudoplatanus saplings within mixed cultures (treatments A and D) in the presence and absence of Rhytisma acerinum infection. The means (from three block replicates) used in determining proportions were taken from the cumulative average number of instars 1-3 and parasitoid mummies over the 13-week experimental period. Two treatments are shown with parasitoids present on R. acerinum infected and uninfected saplings.

DISCUSSION

Fungal endophyte and insect-herbivore-plant interactions:

This study found no direct effects of the endophyte on aphid population growth in general, and no effect on parasitoid performance which could have indirectly affected aphid growth rates. However, sexual female morphs (oviparae) were significantly reduced on endophyte infected plants. This will probably have important consequences for next year's starting aphid population since a reduction in the number of overwintering eggs, deposited by oviparae, will lead to suppressed fundatrix density in the following spring.

79 Chapter 3 Endophyte infection and the sycamore-aphid-parasitoid interaction

The overall findings of the study do not provide strong support for the notion that endophytic fungi are universally mutualistic in their interaction with host plants by deterring herbivores as earlier studies using grasses (e.g. Clay, 1988, 1990, 1991; Cheplick & Clay, 1988; Strong, 1988; Breen, 1994) and trees (e.g. Carroll, 1986, 1991; Butin, 1992) have shown. Neither do the results agree with those of Gaylord et al. (1996) and Preszler et al. (1996) showing a positive association between endophyte infected trees and herbivorous insects. Instead the results of this study corroborate the findings of Faeth & Hammon, (1996, 1997a,b) using leafininers on Emory oak, Lappalainen & Helander (1997) working with a leaf beetle on mountain birch, and that of Saikkonen et al. (1996) using five insect herbivores on mountain birch and Scots pine, which show a neutral endophyte effect.

The absence of an endophyte effect in this study is noteworthy for two reasons. First, these results disagree with the observational work of Gange (1996) showing a positive correlation between R. acerinum infected foliage and the density of two species of sycamore aphid (Periphyllus acericola and D. platanoidis) during autumn; and the findings of Wulf (1990) showing a negative relationship between the endophyte Diplodina acerina and a gall midge Dasineura vitrina found on the leaves of sycamore. Second, due to the balanced design and manipulative nature of the experimental set-up, together with aphids being sensitive to changes in plant chemistry, it would mean that if this arboreal endophyte was to affect aphid-plant interactions, then it should be most detectable in this field study.

The time of post-aestivation nymphal appearance was unaffected by the presence of the fungal endophyte in both the pure (no-choice situation) and mixed cultures (dual-choice situation); however, there was a trend showing that the two stress factors (R. acerinum infection and parasitoids) cause aphids to start reproducing slightly earlier. This neutral effect was not in accordance with the expectation of the first hypothesis that R. acerinum infection should lead to an early release from aestivation. This discrepancy may reflect two possible problems with the experiment. Firstly, adding fertiliser to potted sycamore saplings may have blurred the effects of the endophyte inducing an increase in foliar nitrogen (see Stuedemann et al., 1986; Davidson & Potter, 1995). Secondly, the resolution of weekly measurements of third generation nymphal appearance was not high enough. Nevertheless, with Gange (1966) showing high total and soluble nitrogen associated with R. acerinum infected leaves, it was anticipated that

80 Chapter 3 Endophyte infection and the sycamore-aphid-parasitoid interaction higher foliar nitrogen should lead to an earlier resumption in post-aestivation activity. Evidence for this process is provided by Dixon et al. (1993) who reported that aphids appear to respond to changes in the total amino acid content of leaves by showing a revival in reproduction with the autumnal rise in total levels. Furthermore, Dixon (1963, 1966, 1975a, 1979) found that aestivation (reproductive diapause) during summer, and reproductive activity of D. platanoidis throughout the growing season was controlled by food quality (soluble nitrogen) and the density of aphids. If R. acerinum had adjusted the length of aestivation, then the density of aphids in autumn would have been influenced, because it would have determined the amount of time for population growth prior to leaf fall (Wellings et al., 1985).

No preference, as depicted by a numerical response, was shown by D. platanoidis between R. acerinum infected and uninfected sycamore saplings in pure and mixed cultures during the thirteen-week experimental period. By using mixed cultures (i.e. placing endophyte infected and uninfected saplings together in the same treatment block), direct aphid choice was assessed. However, due to density-dependent switching or random sequential movement from preferred to less favoured hosts in crowded or nutrient depleted situations (see Dixon, 1969), this result may not have been clear. Therefore, the performance of aphids in pure cultures was used to validate the results from the mixed cultures. Both approaches showed that there was no selective advantage in the sycamore aphid going to infected saplings, and that their lack of preference meant that endophyte presence was unimportant for aphid population growth during the post- aestivation period in autumn. This could again have been influenced by fertilisation of potted saplings, since adding nitrogen may have erased the effect of the endophyte on the aphid. However, Davidson & Potter (1995) argue that fertilisation of the plant does not nullify the effect of the endophyte on plant-feeding, predatory and soil-inhabiting in the cultivated turfgrass (i.e. tall fescue turf) situation.

The lack of aphid choice between endophyte infected and uninfected hosts in this experiment was unexpected (cf. fourth hypothesis), because it has been noted that young immature, old senescing and diseased leaves are nutritionally most beneficial to aphids (Kennedy, 1951; Kennedy & Stroyan, 1959). These leaves often have a more yellow appearance than the deep green of mature, less palatable leaves, and spectral reflectances confirm this (Gates, 1980). Both field and laboratory studies have demonstrated that many aphid species are attracted to yellow, although not all species

81 Chapter 3 Endophyte infection and the sycamore-aphid-parasitoid interaction

are equally sensitive to yellow (Eastop, 1955). The attractiveness to yellow may be explained by spectral reflectivity data which show that yellow surfaces are stronger reflectors of the wavelengths to which aphids are most responsive (530-560 nm) than are green surfaces (e.g. Prokopy & Owens, 1978). However, Prokopy & Owens (1983) believe it unlikely that phytophagous insects use spectral quality as a character during host selection, since spectral reflection and transmission are very similar for all green- leaved plants. The similarity between the spectral specificities of summer and autumn migrants of A. fabae (Hardie, 1989), with their different host requirements, support this idea. Kennedy et al. (1961) also argue that the attraction of aphids to green would serve equally well as an attraction to yellow in returning migrating insects to vegetation.

The absence of an endophyte effect on aphid abundance in each age and morph category (except that of oviparae) within pure and mixed cultures disagrees with the expectation of the second hypothesis that R. acerinum infection leads to increased reproduction via increased foliar nitrogen levels. Lack of clear endophyte effects on phytophagous insect performance has been shown in other studies. Saikkonen et al. (1996) who found that the relative growth rates of the moth Epirrita autumnata and the sawfly Arge fuscinervis remained unaffected by endophyte level in birch; while Faeth & Hammon (1997b) found that Cameraria pupal size, which is often positively correlated with adult female fecundity (e.g. Hough & Pimentel, 1980), was not affected by endophyte infection levels. In addition, the results of this study fit well with the findings of Saikkonen et al. (1996) who showed that the density of Schizolachnus pineti (i.e. total number of aphids, juveniles, pathenogenetic adults, sexual adults and eggs in sexual females) was unaffected by the degree of endophyte infection in Scots pine. A possible reason why arboreal studies generally give no indication of fungal endophyte effects on herbivore performance, when compared with grass systems, may be due to the fungal transmission being horizontal: leaf infection is renewed every year. Hammon & Faeth (1992) expect a strong endophyte-host plant mutualism in fungal endophytes that are transmitted vertically (via parent to offspring). While this may be true for grasses and sedges, deciduous arboreal endophyte infection occurs mostly via airborne spores during the growing season, and thus the relationship with fungus and tree is not so intimate (Saikkonen et al., 1998).

It was interesting to note that aphids on infected saplings produced significantly less oviparae than aphids on uninfected saplings in pure cultures (no-choice situation), and

82 Chapter 3 Endophyte infection and the sycamore-aphid-parasitoid interaction that the same trend was found in mixed cultures (dual-choice situation). This result indicates a negative effect of the endophyte on aphid fitness. Raps & Vidal (1998) found that female diamondback moth progeny reacted more sensitively to endophyte infection of cabbage than male larvae, because female larvae feeding on endophyte infected leaves responded to reduced conversion efficiency of ingested food by increasing their relative consumption rates. The implications of suppressed oviparae density will mean that less overwintering eggs are deposited prior to leaf fall (see Dixon, 1976a). This will result in a lower density of fundatrices (first generation aphids) in the following spring, which in turn will influence the peak size of the spring population (see Dixon, 1970b). In addition, premature senescence and leaf fall, caused by heavy infection by R. acerinum of sycamore leaves (Hudler et al., 1987; Leith & Fowler, 1987), may exacerbate the lower oviparae abundance found on endophyte infected saplings from mid September onwards.

The reason for lower oviparae abundance on endophyte infected foliage may be explained by host plant factors such as nutrition and allelochemistry. Gange (1996) recorded an elevation in total and soluble nitrogen content, which is a major determinant of aphid population increase (McNeill & Southwood, 1978; Broadbeck & Strong, 1987), and greater total carbon, which resulted in a higher carbon/nitrogen ratio in R. acerinum infected leaves. This adjustment of nutrient resources effectively dilutes the concentration of plant nitrogen and decreases its availability to insects (Docherty et al., 1997). Also, carbon accumulation in excess of growth requirement may lead to carbon- based allelochemicals being manufactured (Tuomi et al., 1988). This may lead to retarded growth, development and fecundity of many herbivores (Applebaum et al., 1970). Alternatively, Carroll (1988) proposed that endophytes of woody plants provide a direct defensive role for the host plant by producing a wide range of mycotoxins and enzymes, or by triggering the induction of plant defences in leaves (Darvill & Albersheim, 1984; Haukioja & Neuvonen, 1987) that can detrimentally affect insect herbivore performance (Petrini et al., 1992; Faeth & Wilson, 1996). Although oviparae may be more sensitive than other aphid morphs, the overall lack of significant negative trends in aphid abundance on the endophyte infected saplings may lie with the insect detoxifying allelochemicals, mycotoxins or fungal by-products. For example, the pea aphid Acyrthosiphum pisum has developed a metabolic process that either detoxifies or eliminates plant allomones (Srivastava et al., 1988). It is also possible that endo- symbionts within the insect process the toxic compounds such as canavanine

83 Chapter 3 Endophyte infection and the sycamore-aphid-parasitoid interaction

(Rosenthal, 1983), in addition to their role of nutrient upgraders when aphids are put under dietary stress (Douglas & Prosser, 1992; Douglas, 1993).

The disparity between the results here and that of Gange (1996) does not necessary make either study invalid for three reasons. Firstly, because this study used seventy-two nutrient supplemented potted saplings (17 months old - c. 0.75m) within a field insectary, while Gange (1996) made observations on twenty saplings in situ (8 year old - c. 2.5 m), several differences arise. It is possible that factors such as tree age, size, stand density, tree exposure, edaphic and abiotic conditions could alter insect herbivore performance as well as tree physiology and phenology (reviewed by Schowalter et al., 1986). For example, adding fertiliser to the soil may raise the levels of alkaloids in endophyte infected plants (Bush et al., 1993; Latch, 1993), and within its phytophagous insect (Stuedemann et al., 1986). Secondly, endophytes of woody plants typically display great temporal variability. Faeth & Hammon (1997a) suggested that annual and seasonal variation of fungal endophytes provide scope for numerous and complex interactions with insect herbivores that encounter endophytic hyphae, spores or metabolic by-products. Thirdly, this study subjected sycamore saplings to artificially high levels of the R. acerinum inoculum and therefore had a high tar spot presence per leaf, whereas in Gange's (1996) study the tar spot index was low, relative to that of Leith & Fowler (1987).

Fungal endophyte and parasitism of the insect herbivore:

Contrary to the third hypothesis, that plant endophytes should affect the third trophic level, no significant difference was found in the proportion of early instar D. platanoidis nymphs attacked by A. flavus on endophyte infected and uninfected saplings in both pure and mixed treatment. Similarly, Barker & Addison (1997) found that Acremonium lolii infected ryegrass did not influence the rate of parasitism in weevils confined with naive parasitoids. Nevertheless, the lack of parasitoid response to endophyte treatment was unexpected because the presence of fungi has been shown to influence parasitoid search behaviour (reviewed by Dicke, 1988), similar to odours from insect herbivores and their food plants that enable natural enemies to locate their hosts (Wickremasinghe & van Emden, 1992; Turlings et al., 1993; Du et al., 1996; Demoraes & Mescher, 1999). On the one hand, Madden (1968) reported an attraction by the parasitoid Ibalia leucospoides to fungus (Amylostereum sp.) associated with woodwasp tunnels as a cue to finding its host; while Preszler et al. (1996) observed lower oviposition by parasitoids

84 Chapter 3 Endophyte infection and the sycamore-aphid-parasitoid interaction

(or reduced survival of parasitoids) on oak trees with high frequencies of infection by Gnomonia cerastris. Furthermore, Barker & Addison (1996) and Bultman (1997) found that endophyte infection in ryegrass and tall fescue reduced the suitability of the herbivorous insect as a host for parasitoid development and/or survival.

In addition, the abundance of D. platanoidis in all categorises (i.e. total, nymphal, adult virginoparae and sexuals) remained unaffected by the presence of parasitoids. This was not expected because aphids tend to become numerically suppressed due to parasitoid attack (Start', 1988). Also, aphids are known to adjust their reproductive rate in the presence of hymenopteran parasitoids, since Boenisch et al. (1997) have shown that cereal aphids increase their reproductive output in the presence of hyperparasitoids due to the potential decrease in chances of attack by primary parasitoids, and this mechanism is likely to be due to odours released from the secondary parasitoid. However, the absence of aphid response to endophyte treatment, in the presence and absence of A. flavus, may be due to methodological reasons. The initial number of A. flavus adults being released into the field insectary in late August may be too low to make an impact on its host in each treatment block. Alternatively, A. flavus could have suffered a high degree of mortality due to the time difference between post-aestivation nymph appearance and oviposition time within treatment blocks; however, this possibility may be ruled out since independent observations (see Chapter 2) were made on aestivation release time in the field close by the insectary, so as to carefully synchronise the time between oviposition by parasitoids and third generation D. platanoidis nymph appearance. Finally, some adults of A. flavus may have gone straight into winter diapause (see Hamilton, 1973; and references therein).

CONCLUSION

This study and accumulating research (e.g. Gange, 1996; Gaylord et al., 1996 Preszler et al., 1996; Faeth & Hammon, 1997b) suggest that endophytic micro-organisms may alter interactions of phytophagous insects with their woody host plants, but often in indirect and complex ways. For D. platanoidis, any negative effects of endophytic fungi appear to be indirect, by altering host plant phenology (Leith & Fowler, 1987), rather than direct interactions with herbivores via mycotoxins as has been found for some endophytes of grasses (Clay, 1988; Breen, 1994) and through fungal-induced plant volatiles attracting natural enemies (Turlings et al., 1998). If endophytes have negative effects on sycamore, such as accelerating leaf abscission and reducing

85 Chapter 3 Endophyte infection and the sycamore-aphid-parasitoid interaction photosynthesis (Saikkonen et al., 1998), then any direct or indirect negative effects of endophytes on herbivores may be outweighed by this cost to the host tree.

Perhaps further research could focus on manipulating endophyte infection levels in trees and the intensity of attack by phytophagous insects, in order to assess their relative impacts on host-fitness and aphid life-history parameters, and to clearly establish the nature of the interaction between endophytes and sycamore trees. In addition, the endophyte effect on the parasitoid-aphid interaction could be explored by: (i) examining the parasitoid response to varying endophyte and aphids levels using olfactometer choice trials similar to those used by Vet et al. (1983) and Wickremasinghe & van Emden (1992); and (ii) assessing the indirect impact of the plant endophyte, via host quality, on parasitoid performance such as their rate of survival, time to first mummy appearance, size of parasitoids and their development time (see Barker & Addison 1996).

86 Chapter 4 Reproductive sequencing and sex ratio

Reproductive sequencing and the sex ratio of the sycamore aphid

ABSTRACT

1. Many aphid species have lifecycles which alternate between asexual reproduction during spring and summer, and sexual reproduction in autumn. 2. The reproductive sequence of sexuparae and the fecundity of oviparae of the sycamore aphid, Drepanosiphum platanoidis (Schrank), observed in the field (1996) and in the greenhouse (1997) were examined to elucidate the relationship between host nutrition, sex ratio, maternal size and reproductive potential. This study was performed at the time of year when the transition from parthenogenetic to sexual reproduction occurred. 3. A switch occurred in the reproductive sequence of the sexuparae. The production of apterous sexual females (oviparae) was followed by alate males, after a brief reproductive pause lasting 2.5 - 4.0 days. 4. The switchover period was relatively constant across treatments and years, indicating the importance of the photo- and thermo-periodic influences. The first appearance of oviparae and males remained unaffected by host nutritive status. On poor hosts the commencement of the reproductive sequence (i.e. oviparae first appearance) was a day later, whereas the first appearance of males was a day earlier than those on good nutrition. 5. The primary sex ratio (i.e. proportion of males) was numerically female biased. Under semi-natural field conditions (1996) it was 0.22, but altered to 0.18 and 0.29 under poor and good host nutrition in the greenhouse (1997). 6. In both the field and greenhouse, nymphal development time was significantly longer in oviparae (c. 22 days) than in males (c. 18 days). The time from birth to adult was significantly quicker on good hosts (oviparae: 20 days; males: 17 days) than on poor hosts (oviparae: 23 days; males: 19 days). 7. The proportion of nymphs surviving to the adult moult was not dependent on sex in both the field and greenhouse, but the chances of survival were slightly enhanced under good (88%) compared to poor (83%) host nutritive status.

87 Chapter 4 Reproductive sequencing and sex ratio

8. Adult life-span was significantly longer in (unmated) oviparae (c. 27 days) than in males (c. 17 days), but was not significantly affected by host nutritive status. 9. Improved host condition led to larger body size in oviparae, which gave rise to higher potential fecundity in terms of greater numbers of eggs per oviariole. However, varying host nutrition did not alter the ovariole number. 10. In conclusion, the order of morph production and differences in maturation times ensured the synchronisation of the sexes so that mating and oviposition opportunities were optimised before the onset of harsh winter conditions. The female-bias in the sex ratio maximised the fitness of sexuparae, and reduced the chances of local mate competition by economising in members of the competing sex. The number of eggs/ovariole (i.e. potential fecundity) was a function of oviparae body weight, which in turn was governed by host plant nutrition.

INTRODUCTION

Most species of aphid reproduce both asexually and sexually, with several generations of parthenogenesis between each bout of sexual reproduction (Dixon, 1985; Rispe et al., 1996). The decision to end the production of parthenogenetic offspring must be taken one (if monoecious) or two (if heteroecious) generations in advance of the time of oviposition (Dixon, 1987; Ward & Wellings, 1994). Since this decision ends the aphid clones growth season, it should be postponed as late as is compatible with successful egg-laying, because the insertion of an extra parthenogenetic generation greatly increases the potential fecundity of the clone (Ward et al., 1984; Dixon, 1987; Newton & Dixon, 1987). The switch to the production of the less fecund sexual forms should occur only when the time to leaf-fall is equal to the time required for development, migration, mating and oviposition (Ward et al., 1984; Dixon, 1985; Ward & Wellings, 1994).

External cues, to anticipate the changes in host biology, are often used as triggers to signal the onset of autumn. Leaf-fall is stimulated by falling day-length (Vince Prue, 1975), which is known also to be an important cue for the production of aphid sexual morphs (e.g. Lees, 1959, 1960, 1973; MacGillivray & Anderson, 1964; Blackman, 1975; Kawada, 1987). The number of aphid generations that can be fitted into the time available is dependent on temperature (Ward et al., 1984; Dixon, 1985, 1987).

88 Chapter 4 Reproductive sequencing and sex ratio

Empirical data on various aphid species show that temperature influences the day-length threshold for the end of virginoparae production and induction of one or both sexes (MacGillivray & Anderson, 1964; Dixon & Glen, 1971; Lamb & Pointing, 1972; Tsitsipis & Mittler, 1977a,b; Hand & Wratten, 1985). Some species of aphid e.g. Eriosoma pyricola, Dysaphis devecta and Aphis farinosa respond to changes in plant quality (i.e. produce sexuparae when shoot growth ceases) heralding the onset of adverse conditions (Sethi & Swenson, 1967; Forrest, 1970). Hille Ris Lambers (1960) suggested that root aphids react to photoperiodically-induced changes in the host plant; this was evidenced by the induction of diapause in Erioischia brassicae feeding on the roots of cabbages exposed to short days (Hughes, 1960).

Sexual morphs rarely appear early in the year when the short day conditions of spring are as long as in autumn (Dixon, 1973c, 1985). This is because of the operation of an intrinsic timing mechanism, lacteur fondatrice' (Bonnemaison, 1951) or 'interval timer' (Lees, 1960, 1966). Higher temperatures are known to cause the timer to 'run down' more quickly (Lees, 1960). In the monoecious sycamore and lime aphid, there are two interval timers (Dixon, 1971b, 1972); in the sycamore aphid, one is sensitive to day-length which controls the production of sexual females (oviparae), and the other, insensitive to day-length which controls the production of males. The restraining effect of the interval timer causes a gradual transition from parthenogenetic to gamic reproduction over a period of several generations, particularly under short photoperiods. Incomplete inhibition of the interval timer in autumn permits part of a sycamore aphid population to continue reproducing parthenogenetically in the short days when the senescent foliage of sycamore provides a rich source of food (Dixon, 1963, 1970a). This is particularly advantageous in those autumns when leaf fall is late (Dixon, 1985, 1987).

The switch to production of male embryos in the progeny sequence and to the developmental pathway leading to sexual rather than parthenogenetic females is regulated by neuroendocrine processes that are triggered by environmental factors such as photoperiod (Steel & Lees, 1977; Hardie, 1987). The photocycle acts on the neuroendocrine system of the parthenogenetic mother, triggering physiological processes that include alterations in titres of juvenile hormone (JH) (Hales & Mittler, 1983; Hardie et al., 1985). JH titre controls the behaviour of the X chromosome at the

89 Chapter 4 Reproductive sequencing and sex ratio maturation division of oocytes, where male determination occurs when JH titre is low (Hales & Mittler, 1983, 1987).

Apart from environmental and physiological factors influencing the timing and proportion of sexual morphs in an aphid population, the chromosomal mechanism determines the sex ratio. Aphids show X)(/OX (female/male) sex determination, with multiple X chromosomes in some species (Blackman & Hales, 1986; Blackman, 1987). The diploid sexual females are produced ameiotically (i.e. thelytoky), as in the other parthenogenetic generations of the life cycle (Blackman, 1974, 1987), while the males are produced by reductive mitosis (Orlando, 1974, 1983). Since aphids are diplodiploid with genetic sex determination (Moran, 1993), this gives them a facultative control over their progeny sex ratio (Orlando, 1983).

A mother's ability to pass her genes to future generations will be affected by her offspring sex ratio (i.e. proportion of male offspring) (Heinz, 1996), which in turn, is expected to be under strong selection (Fisher, 1930; Hamilton, 1967; Charnov, 1982). Fisher (1930) argued that natural selection would adjust the sex ratio so as to equalise parental investment (PI) in the two sexes when mating is panmictic. Hamilton (1967) then showed that female-biased sex ratios would be favoured when male siblings compete for matings, in a situation referred to as local mate competition (LMC). This would have the advantage of both increasing a mother's number of daughters and increasing average mating success of her sons, so she maximises the number of her grandchildren (Taylor, 1981). Aphids are known to be a major group in which LMC appears to cause large sex ratio biases (May & Seger, 1985). In aphids, selection may shape the overall brood sex ratio or the sequence and timing of progeny types (Moran, 1993). This may take the form of differential dispersal abilities of the sexes. In almost all aphids, sexual females are wingless, whereas males may be winged, wingless or both occurring within a species (Hille Ris Lambers, 1966; Blackman, 1974). This may lead to situations of either LMC or local resource competition (see Moran, 1993). Other key selection factors are the timing of sex determination, and the relative costs of sons and daughters (Moran, 1993). There have been relatively few attempts to understand sex allocation (i.e. investment of limiting resources in gametes and maternal gonads) in detail in aphids (Foster & Benton, 1992; Ward & Wellings, 1994). These have essentially focussed on the life cycles of heteroecious members of the Pemphigidae, which experience LMC (Yamaguchi, 1985; Kindlmann & Dixon, 1989;

90 Chapter 4 Reproductive sequencing and sex ratio

Foster & Benton, 1992; Moran, 1993). However, the sex allocation in monoecious members of this and other taxa have been ignored (Moran, 1993; Ward & Wellings, 1994).

Mothers appear not only to have direct behavioural control over progeny sex ratios, but also their fecundities over a wide range (Yamaguchi, 1985). Increased maternal weight and/or size has been reported to increase fecundity for many insects (Gilbert 1984), especially aphids (Dixon & Wratten, 1971; Leather & Wellings, 1981; Dixon et al., 1982). In many aphid studies, potential fecundity (reproductive potential, i.e. embryo counts or large embryo counts) is used instead of actual total fecundity (Elliott, 1973; Llewellyn & Brown, 1985). This indicative fecundity estimate is only adequate in aphids where ovulation ceases after adult moult, e.g. D. platanoidis; whereas in aphids which continue to ovulate after the adult moult, e.g. Rhopalosiphum padi, the initial assessments of fecundity can be seriously underestimated (Leather, 1983). In addition, pre- and post-reproductive periods are of great importance in affecting insect fecundity (Taylor, 1975; Kidd & Tozer, 1985; Leather et al., 1985). Nymphal experience of the mother could adjust the fecundity and/or maternal body size; for example, a resource deprived aphid may mature its largest embryos and absorb the smallest thereby sacrificing its potential fecundity to maintain its immediate reproductive rate (Ward et al., 1983a,b). Adult longevity could elongate or curtail the realised fecundity, and therefore complicate the size/potential fecundity relationship (Leather, 1988). In view of the inadequacies of using large embryo counts (Leather, 1988), the number of ovarioles (i.e. tubes containing a series of developing embryos), which are determined before birth (Lees, 1959), has been used as a more reliable indicator of potential fecundity (Wellings et al., 1980; Leather & Wellings, 1981).

The sycamore aphid, Drepanosiphum platanoidis Schrank (Homoptera: Aphididae), is a monoecious aphid spending its whole lifecycle on the sycamore, Acer pseudoplatanus L. It undergoes sexual reproduction (holocycly) once a year in the autumn (Blackman & Eastop, 1994). The sexuparae produce sexuals (males and oviparous females) parthenogenetically. Mating takes place on senescing leaves (Koslowski, 1991) where oviparae mature before they leave to oviposit in crevices of tree bark (Dixon, 1976a). Alate males remain on particular leaves where they increase their fitness by maximising the number of copulations (Kozlowski, 1991). As yet, only Dixon (1971b) has made a start to quantify the progeny sequence, sex ratio and parental investment of this

91 Chapter 4 Reproductive sequencing and sex ratio

monoecious sexuparae in the field and laboratory (Dixon, 1971b). No investigations have assessed the affect of host nutrition on the potential fecundity of the sexual generation of D. platanoidis, despite the effects of host plant nutrition and temperature being examined in the first three parthenogenetic generations (Leather & Wellings, 1981). This is surprising, since the implications of modified sex ratios and oviparae fecundity are far reaching in terms of determining the success of future populations that arise from the overwintering eggs in the following spring. Each aphid egg, therefore, represents an evolutionary individual (Janzen, 1977) which, when emerges as the fundatrix in spring, will cascade parthenogenetically into sub-units later in the year (Dixon, 1985).

This study aims to assess the impact of host plant condition on the reproductive sequence and sex ratio in sexuparae of D. platanoidis and the associated life history traits (e.g. developmental time, nymphal survival, adult longevity) of their progeny. The term 'reproductive sequence' will be used in the same context as Lamb & Pointing, (1975), referring to the order of appearance of sexual morphs and the pattern of birth rates throughout the reproductive period of a virginoparous aphid, sexupara. The study also aims to determine the influence of host nutrition on adult oviparae size and how this in turn affects potential fecundity. For this, two hypotheses are tested: (i) the `guaranteed fitness hypothesis' in which the more abundant oviparae are laid first in the reproductive sequence followed by the faster developing males in order to synchronise mating (and hence oviposition before leaf fall) and to maximise reproductive success; and (ii) the 'maternal size/offspring hypothesis' in which oviparae size provide a positive indicator of reproductive output.

METHODS & MATERIALS

Observations and experiments were undertaken in a deciduous woodland at Silwood Park, Berkshire, (OS ref. 5946 1689) from early September to mid December in 1996 and 1997 respectively. Experimental and aphid rearing work were done within a ventilated greenhouse under ambient temperature and light conditions, so that aphid growth and reproduction would experience the natural progressive changes in autumnal temperature and photoperiod.

92 Chapter 4 Reproductive sequencing and sex ratio

Experimental Designs:

1. Reproductive sequencing in sexuparae and development time of sexuals under field conditions.

Twelve 2.0-2.5 m sycamore saplings in the field were selected. Six clip cages (5 cm diameter petri-dish embedded within a 10 cm x 8 cm x 1.2 cm polyfoam 'sandwich' framework) were attached to each sapling. Randomly chosen newly moulted sexuparae (i.e. 72 third generation parthenogenetic virginoparae), which had been confined to that host since fundatrices, were individually placed into clip cages. During autumn 1996, sexuparae (Plate 4-1) were monitored daily from the time of final moult to death in order to determine: (i) total reproductive output, (ii) order of sexual morph appearance - i.e. reproductive sequence, (iii) number of each sex produced (to derive the sex ratio), (iv) parturition dates, (v) sexual morph nymphal development time - i.e. birth to final moult, and finally (vi) the point in the reproductive sequence the maternal switch occurs and the duration of the reproductive pause. These response variables were used to test for inter- and intra-tree variations under field conditions.

Prior to larviposition by the sexuparae, a set of seven month old potted sycamore saplings in 8 cm diameter polythene pots, grown in John Iimes No. 1 loam-based compost from germination, were arranged on the central bench of a large greenhouse. With the onset of reproduction, three day old sexual morphs were transferred to the greenhouse for maturation, in order to avoid hostile weather conditions dislodging the developing sexual morphs. The rearing regime was based on placing the progeny from the same mother, in a series of mini-blocks (i.e. clip cages) with each cage containing five sexual morphs from the sequential progeny line.

2. Influence of host nutritional quality on reproductive sequencing in sexuparae and development time of sexuals under greenhouse conditions.

In autumn 1997, twenty 19 month old sycamore saplings in 25 cm diameter polythene pots, grown in John limes No. 2 loam-based compost under greenhouse conditions, were selected. Half (n=10) were dosed with 5m1/1 Levington's 'Deep Feed' (7-1.3- 3.3% NPK solution containing 0.018% Mg with trace elements) fertiliser three times between 25/08/97 to 5/09/97 and thus were termed 'good' host nutrient quality; and the

93 Chapter 4 Reproductive sequencing and sex ratio

remaining ten saplings received no nutrient enhancement (i.e. the control) and hence termed as 'poor' host nutrient quality.

Four saplings were placed in each of five blocks in the central bench space of the greenhouse. In each block, two good and two poor saplings were added. Three clip cages were placed separately on the leaves of each sapling (60 cages in total). One randomly chosen second generation parthenogenetic virginoparae emerging from aestivation, obtained from mixed greenhouse cultures, was assigned to each cage, allowed to reproduce and then discarded. On the 26th September 1997, one progeny (a sexupara) from each of the parthenogenetic mothers was randomly chosen, and the isolated sexupara was left in the clip cage until its death. The conditions experienced by the aphid's mother are considered to be more important than the conditions experienced by the nymph during its post-natal development (Dixon & Glen, 1971). With the onset of sexuparae reproduction, the offspring (sexual morphs) were transferred in batches of five individuals to separate clip cages for rearing on the same host as their mothers. The following response variables were recorded daily: (i) total reproductive output, (ii) order of sexual morph appearance, (iii) number of each sex produced, (iv) parturition date, (v) sexual morph nymphal development time, (vi) progeny's adult lifespan, and finally (vii) the duration of the maternal switch mechanism and point it occurs in the reproductive sequence. These variables were used to test for the effects of treatment, block and inter-sapling differences under laboratory conditions.

3. Influence of host plant nutrition on body size and potential fecundity of oviparae under greenhouse conditions.

Oviparae descending from the greenhouse sexuparae reared on 'good' and 'poor' host plant qualities were selected during late October 1997. From the two treatments, using 18 trees each, 4 mature oviparae per tree were randomly taken for dissection. Pre- parturition fecundity was determined by examining the reproductive system of 144 oviparae and recording the number of: (i) immature and mature eggs, (ii) eggs per ovariole, and (iii) total ovarioles per individual. Each live oviparae was placed in aqueous solution under a binocular microscope, illuminated by a cold light source, and dissected according to the method used by Wellings et al. (1980). Each individual ovipara was weighed prior to dissection to +0.1 i=tg accuracy on a Sartorius Micro- balance.

94 Chapter 4 Reproductive sequencing and sex ratio

Data transformation and statistical analyses:

To test whether the field sex ratio differed from the expected 0.5 Fisherian ratio, the proportion of males among the total offspring of sexuparae (i.e. sex ratio) was determined in GLIM (version 3.77 copyright 1985 Royal Statistical Society, London; Crawley, 1993) using a binomial error structure. By fitting the null model, the overall mean and standard error was obtained. Through multiplying the standard error of the sex ratio by the relevant critical value in the student's t table, adding this to the overall mean to obtain the upper confidence limit, and converting from logits to proportional males (Crawley, 1993), the field sex ratio could be compared with the 0.5 ratio. For the plant nutrient manipulation data, the same process was repeated but fitting the 'plant nutrition' (two levels) factor after the null model. In the observational study, the proportion of males among the total offspring of sexuparae (i.e. sex ratio) was analysed in GLIM as a one-way ANOVA with binomial errors using 'sapling' (twelve levels) as the factor to assess whether the sex ratios varied significantly between saplings. For the plant nutrition experiment, testing whether there were fewer males produced on poor hosts, one-way ANOVAs with binomial errors were performed using 'plant nutrition' (two levels) and 'sapling' (twenty levels) as factors.

The day of first female and male appearance and also the timing of the maternal switch were analysed with the Kruskal-Wallis test (Sokal & Rohlf, 1995) in Excel. In each case, the day of first sexual morph appearance was set to a value of one prior to non- parametric analysis. For the observational study, the between sapling variability was tested, whereas in the experimental manipulation both nutrient treatment and sapling differences were analysed.

The duration of the maternal switch in the reproductive sequence of sexuparae was analysed in GLIM using one-way ANOVAs. In the observation field study, 'sapling' (twelve levels) was used as the factor; while the host manipulation experiment used `plant nutrition' (two levels) and 'sapling' (twenty levels) as the factors to test whether good host nutrition elongated the switch-over period and to determine whether between sapling variability was significant.

To assess whether males had the same survival chances of reaching adulthood as oviparae in relation to different trees and/or host plant conditions, one-way ANOVAs with binomial errors, using 'sapling' and 'host nutrition' as factors, were performed on

95 Chapter 4 Reproductive sequencing and sex ratio the observational and experimental data sets. Overdispersion was corrected by dividing the Pearson X2 correction factor by the degrees of freedom in GLIM.

In determining which gender showed the fastest development time and greatest adult longevity, one-way ANOVAs were performed on the observational study using 'sex' (two levels) and 'sapling' (twelve levels) as the factors. In addition, the plant quality experiment performed one-way ANOVAs using 'plant nutrition' (two levels), 'sex' (two levels) and 'sapling' (twenty levels) as the factors. In both cases, the male and female response variables were log (Y + 1) transformed in Excel prior to analysis in GLIM.

The data used in calculating the ovariole number/size and fecundity/size relationships were split into three sections, thus representing three hierarchical response variables: (i) the total number of ovarioles per ovipara; (ii) the total number of eggs per ovariole; (iii) the immature egg proportion over total egg count. Analyses of covariance using normal errors and an identity link were performed on the data for numbers of ovarioles per ovipara and the eggs per ovariole, whereas analysing the immature embryo proportion data used a binomial error structure with a logit link function to ensure linearity. These ANCOVA analyses used 'plant nutrition' (two levels), 'sapling' (eighteen levels) as the factors, and adult oviparae 'body weight' as the covariate. In addition, the impact of host plant nutrition on oviparae body weight was assessed using a one-way ANOVA, by specifying 'plant nutrition' as the factor with two levels.

Plate 4-1: Different morphs of Drepanosiphum platanoidis on a leaf of Acer pseudoplatanus. Left: Two adult oviparae; Middle: Adult sexupara; Right: Adult male.

96

Chapter 4 Reproductive sequencing and sex ratio

RESULTS

Reproductive sequencing in sexuparae and development time of sexuals under field conditions:

A unidirectional change from oviparae to male production was evident in the reproductive sequence (Fig. 4-1), and this was consistent for every sexuparae observed. Oviparae appeared about 18/10/96 (Julian day 291.04+0.31) and males about 9/11/96 (Julian day 312.98+0.64). Oviparae appearance did not significantly vary between saplings (H = 11.25, d.f. = 11, P>0.05; Fig. 4-2), but male appearance did (H = 22.97, d.f. = 11, P<0.025; Fig. 4-4). During the reproductive sequence, sexuparae experienced a reproductive pause in the sexual morph switch over period. The timing of the maternal switch occurred about 7/11/96 (Julian day 310.10+0.63), but varied significantly between saplings (H = 25.90, d.f. = 11, P<0.01; Fig. 4-3). The maternal switch duration lasted 3.67+0.30 days, but did not significantly vary from sapling to

sapling in the field (F 10.37 = 1.21, r2 = 0.25, P>0.05; Fig. 4-5).

6 a)

5- a. .

X ■ onpErm a) ,_ci) 4- ■n . (T) a

"MO 3

4- 0 43 2- SD

1 I 111 cv 2

0- I - c5,1 „.ft 44) ,b4 11. (1, tb rb "..;\ r'3 "9" n" rg' w? rt',P Jui jai cty (i.e 1310'93to 212S) Fig. 4-1: The average number of oviparae (i) and males (0 produced per day by Drepanosiphum platanoidis sexuparae (n = 72) on Acer pseudoplatanus saplings under natural field conditions.

97 Chapter 4 Reproductive sequencing and sex ratio

2.5 _

r>, C 2.3 _ Es 73 c E CI) s- 0 2.0 _ cs_ = 2 • La a, 2a) RI 1., 1.8 -

5 2 — o w + 4- x 1.5 o c ru a 15 1.3 -

1.0 1 2 3 4 5 6 7 8 9 10 11 12 Sapling number

Fig. 4-2: Mean day of appearance (+ S.E.) of Drepanosiphum platanoidis oviparae on twelve Acer pseudoplatanus saplings under natural field conditions.

4.0

1 2 3 4 5 6 7 8 9 10 11 12 Sapling number

Fig. 4.-3: Mean day of maternal switch occurrence (+ S.E.) in the reproductive sequence of Drepanosiphum platanoidis sexuparae on twelve Acer pseudoplatanus saplings under natural field conditions.

4.0

3.5 • a) ..---. U 0 -0 (I) 3.0 • c E co a) 4- 2.5 a to co op f° 41) 04 2.0 •E 2E ‘— 1.5 46 4- ›.., Cr) X C 1 . 0 a E ▪ 0.5 0.0 1 2 3 4 5 6 7 8 9 10 11 12 Sapling number

Fig. 4-4: Mean day of appearance (+ S.E.) of Drepanosiphum platanoidis males on twelve Acer pseudoplatanus saplings under natural field conditions.

98 11,6o =2.14, The numberofprogeny(i.e.sexualmorphs)producedperparthenogeneticsexuparawas Fisherian ratio.Also,thesexratiovariedsignificantlybetweentwelvesaplings(F Chapter 4 on Acerpseudoplatanus Fig. 4-5:Meanmaternalswitchduration(±S.E.)of Fig. 4-6:Meannumber(±S.E.) ofoviparaandmalesbornpersexupara across sexuparaewas0.22+0.05.Theuppersexratioconfidenceboundary0.24 (one-tailed Acer pseudoplatanus biased towardsoviparaeratherthanmaleproduction(Fig.4-6).Thefield

Mean maternal switchdu ration in days (Ln(x + 1) - transformed) t 0.6 2.0 1.6 1.8 1.2 1.4 0.0 0.2 0.4 0.8 1.0 r 71 2 = set atP<0.005),thereforeshowingasignificantdifferencefromthe0.5 0.28, P<0.05). saplings undernaturalfieldconditions.

Mean number born per sexupara( 1 23456789101112 15 10 20 25 30 saplings undernaturalfieldconditions. 5 0

- Ovipare T Sexual morphtype

Sapling number 99 Drepanosiphum platanoidis Reproductive sequencingandsexratio males T Drepanosiphum platanoidis sexuparae ontwelve sex ratio

Chapter 4 Reproductive sequencing and sex ratio

Male nymphal survival rate was not significantly different from that of oviparae (X2 = 1.58, d.f. = 1, r2 = 0.012, P>0.05; Fig. 4-7), and sexual morph survival did not significantly vary between saplings (F 11,108 = 1.22, r2 = 0.13, P>0.05). No significant variability in the proportion of oviparae (F 11, 60 = 0.96, r2 = 0.15, P>0.05) and male (F

10, 37 = 1.06, r2 = 0.22, P>0.05) survival occurred between saplings in the field.

1.0

To0.8= cm x .c a) 5 4-co -E 0.6 o = c w 0 w_C .--E O. 0.4 Q E 2' a_ c 0.2

0.0 Oviparae males Sexual morph type Fig. 4-7: Proportion of oviparae and males of Drepanosiphum platanoidis surviving to reach adulthood on Acer pseudoplatanus saplings under natural field conditions.

Nymphal development time was significantly longer in oviparae than in males (F 1, i ts = 53.57, r2 = 0.31, P<0.001; Fig. 4-8), but sexual morph development time did vary significantly between saplings (F ii, 108 = 27.74, r2 = 0.74, P<0.001). Also, both oviparae (F 11, 60 = 33.56, r2 = 0.86, P<0.001) and male (F 10, 37 = 22.39, r2 = 0.86, P<0.001) development time varied significantly between saplings in the field.

24

)

s 20 ime day t t in

( 1 6 men hs lop mp 1 2 e v l ny a de 8 n xu Mea f se 4 - o

0 oviparae m ales Sexual morph type Fig. 4-8: Mean development time (± S.E.) of oviparae and males of Drepanosiphum platanoidis on Acer pseudoplatanus saplings under natural field conditions.

100 Chapter 4 Reproductive sequencing and sex ratio

Influence of host nutritional quality on reproductive sequencing in sexuparae and development time of sexuals under greenhouse conditions:

A unidirectional switch from oviparae to male production was evident in the reproductive sequence (Figs. 4-9 a & b) of both treatments, and this was consistent for every sexupara in the plant nutrient manipulation experiment.

ao (a) 2.5 >, co E 'a-) 2.0 o. 'a-) Ea) _o 0 1.5 0_ 7 C X • 1.0 Cll 2 ca. 0.5

0.0 --. I [III ME 1 1-i - 11 ,k43 At. cg\ C., A o Co , (1, 93) I, l? (fP rti3 q, (5° ,5° -D\ N ,bNt4 -DP(k -SC) r6bri" n)(b(b ,bt"N Julian days (i.e. 02/10/97 to 07/12/97)

4.0 ( b) 35 -

-c co 3.0 -c -o To 4-, 4-o 0_ 2.5 - .1)' a)E -o ca 2.0 0_ c x 1.5 - c (,)a) • '& 2 o_ 1.0 -

0.5 - 0.0 h 111Ij tx ( rt, riP 151)\, clf e ee naNI‘ 4,,(k op 46 nt r5'b43 Alai days (i.e. 02/10/97 to 07/12/97) Figs. 4-9 a & b: The average number of oviparae (i) and males (1) produced per day by Drepanosiphum platanoidis sexuparae (n = 30 per treatment) on Acer pseudoplatanus saplings with poor nutrition (a) and good host nutrition (b) in greenhouse conditions.

101 Chapter 4 Reproductive sequencing and sex ratio

Oviparae appeared about 16/10/97 (Julian day 288.71+0.98) and 15/10/97 (Julian day 287.97+0.89) on poor and good nutritional hosts respectively. Plant nutritional status did not significantly affect when the first oviparae were born (H = 0.023, d.f. = 1, P>0.05; Fig. 4-10a), but the between sapling variability was significant (H = 42.11, d.f. = 19, P<0.005; Fig. 4-10b). Males appeared about 03/11/97 (Julian day 306.83+0.65) on poor and about 04/11/97 (Julian day 307.67+0.79) on good nutritive hosts. The appearance of the first males was not significantly affected by plant treatment (H = 0.052, d.f. = 1, P>0.05; Fig. 4-12a) but was between saplings (H = 37.98, d.f. = 19, P<0.01; Fig. 4-12b).

Within the reproductive sequence, sexuparae had a reproductive pause where there was a sexual morph switch over period in both plant nutrient treatments. The timing of the maternal switch occurred about 31/10/97 (Julian day 304.13+0.55) and 02/11/97 (Julian day 305.77+0.81) on poor and good host nutrition respectively, which was not significantly different for each nutrient treatment (H = 0.46, d.f. = 1, P>0.05; Fig. 4- 11a), but did vary significantly between saplings (H = 41.19, d.f. = 19, P<0.005; Fig. 4- 1 lb). The maternal switch duration lasted 3.71+0.25 days and 2.83+0.16 on poor and good host plants respectively. The plant nutrient treatment effect was significant (F 1,52

= 8.81, r2 = 0.14, P<0.005; Fig. 4-13a), but the between sapling variability was not (F 19,

34 = 1.34, r2 = 0.43, P>0.05; Fig. 4-13b).

3.0 -

0.0 Poor Good Host nutritional status

Fig. 4-10 a: The effect of host plant nutrition on the date of first appearance of Drepanosiphum platanoidis oviparae (+ S.E.) on Acer pseudoplatanus saplings in greenhouse conditions.

102 Chapter 4 Reproductive sequencing and sex ratio

3.5

3.0 o IO

s

[le 0 0 d) day 2.5 arance e lian p forme 2.0 ap 4 Ju trans 27 arae -

m 1.5 ip fro

+ 1) f ov (x o ing 1.0 t Ln ( tar Day s 0.5

0.0 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Sapling number

Fig. 4-10 b: Mean day of appearance of Drepanosiphum platanoidis oviparae (LP S.E.) on Acer pseudoplatanus saplings under poor (0) and good(®) nutrition in greenhouse conditions.

4.0 _ T. 3.5 coID_ w >, z a)x m i3 3.0 c E c 2.5 _c 5-3 C/1,P- ca 2.0 CNN- "—", E 1.5 c 1.0 _ E 4— ay 0 0.5 0.0 Poor Good Host nutritional status

Fig. 4-11 a: The effect of host plant nutrition on the maternal switch date of Drepanosiphum platanoidis sexuparae (+ S.E.) on Acer pseudoplatanus saplings in greenhouse conditions.

4.0

3.9

arae s

3.8 d) xup day

me 3.7 _ se lian in for h Ju 3.6

ns 0 itc 74 2 tra 3.5

l sw - E:1 1 m

1) 3.4 fro

+ I:1 o (Lr' terna a

ing 3.3 t El Ln(x ( f m tar o s 3.2

Day 3.1

3.0 1 _I 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Sapling number

Fig. 4-11 b: Mean date of maternal switch of Drepanosiphum platanoidis sexuaparae (+ S.E.) on Acer pseudoplatanus saplings under poor (0) and good *nutrition in greenhouse conditions.

103

Chapter 4 Reproductive sequencing and sex ratio

4.0

3.5 s

d) day

3.0 ance me ian l for ear 2.5 74 Ju app 2.0 trans le 2 -

1) 1.5 f ma from + o

ing 1.0 t Ln(x Day ( tar

s 0.5

0 0 Poor Good Host nutritional status

Fig. 4-12 a: The effect of host plant nutrition on the date of first appearance of Drepanosiphum platanoidis males (+ S.E.) on Acer pseudoplatanus saplings in greenhouse conditions.

3.7 s

d)

day 3.6 - rme lian arance fo e Ci ci)j 3.5 143 4) r.:3 74 Ju app trans le 2 - 3.4 _ 1) f ma from +

o 3.3 _ ing t Ln(x Day ( tar s 3.2 _

3.1 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Sapling number

Fig. 4-12 b: Mean day of appearance of Drepanosiphum platanoidis males (+ S.E.) on Acer pseudoplatanus saplings under poor (0) and good ( )nutrition in greenhouse conditions.

1.75 _

1.50

d) h

itc 1.25 s rme fo l sw day

ns 1.00 in

tra terna ion - 0.75 t ra ma + 1) an du 0.50 (x Me Ln ( 0.25

0.00 Poor Good Host nutritional status

Fig. 4-13 a: The effect of host plant nutrition on the maternal switch duration of Drepanosiphum platanoidis sexuparae (+ S.E.) on Acer pseudoplatanus saplings in greenhouse conditions.

104 Chapter 4 Reproductive sequencing and sex ratio

2.0

1.8

1.6

d) ion t 0 1.4 forme dura

h 1.2 ans itc tr

- 1.0 l sw 1) + 0.8 terna Ln(x (

ma s 0.6

day 0.4 Mean in

0.2

00 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Sapling number

Fig. 4-13 b: Mean duration of the maternal switch of Drepanosiphum platanoidis sexuparae (+ S.E.) on Acer pseudoplatanus saplings under poor (El) and good (so) nutrition in greenhouse conditions.

The number of progeny (i.e. sexual morphs) produced per sexupara was biased towards oviparae rather than male production (Fig. 4-14). The sex ratios were 0.18+0.10 and 0.29+0.12 on poor and good nutrition hosts. The upper sex ratio boundary was 0.23 and

0.36 respectively (one-tailed t 29 set at P<0.005), therefore showing a significant difference from the 0.5 Fisherian ratio. The sex ratios of the plant nutrient manipulation experiment significantly differed between both treatments (F 1, 56 = 15.39, r2 = 0.22,

P<0.001), and across saplings (F 19,38 = 2.70, r2 = 0.57, P<0.005).

35

30 _ ara

25 sexup er p

n 20 r bo

r 15 be

10 num n

Mea 5-

0 Oviparae Males Sexual morph type Fig. 4-14: Mean number (+ S.E.) of ovipara and males born per sexupara of Drepanosiphum platanoidis on Acer pseudoplatanus saplings under poor (0) and good (01) nutrition in greenhouse conditions.

105 Chapter 4 Reproductive sequencing and sex ratio

The proportion of survivors reaching adulthood was significantly greater with good

(0.88+0.013) than poor (0.82+0.029) host nutrition (F 1, 109 = 6.59, r2 = 0.057, P<0.025; Fig. 4-15). However, oviparae (0.84+0.022) and male (0.87+0.021) survivors had similar nymphal survival rates (F 1, 109 = 0.38, r2 = 0.0035, P>0.05), and no significant variability in sexual morph survival occurred across saplings in both treatments (F 19, 91 = 1.68, r2 = 0.26, P<0.10). A difference was noticed in the proportion of oviparae (F

56 = 5.08, r2 = 0.083, P<0.05), but not with males (F 1, 52 = 1.77, r2 = 0.033, P>0.05), survival between poor (oviparae: 0.80+0.041; male: 0.84+0.040) and good (oviparae: 0.87+0.015; male: 0.88+0.021) treatments. No significant variability in the proportional survival of oviparae (F 19, 38 = 1.05, r2 = 0.34, P>0.05) and males (F 19, 34 = 0.85, r2 = 0.32, P>0.05) between saplings in the experiment.

1.0 -

Of 0.8 _

• O 0.6 O

0.4

0 to O• 0.2

0.0 Oviparae males

Sexual morph type

Fig. 4-15: Mean proportion (+ S.E.) of oviparae and males of Drepanosiphum platanoidis surviving to reach adulthood on Acer pseudoplatanus saplings under poor (Q) and good (F) nutrition in greenhouse conditions.

Nymphal development time of sexuals took significantly longer with poor (21.01+0.38) than good (18.76+0.38) host nutritional status (F 1, 110— 17.36, r2 = 0.14, P<0.001; Fig. 4- 16). Oviparae (21.53+0.36) took significantly longer to develop than males

(17.95+0.30) (F 1, 110 = 53.61, r2 = 0.33, P<0.001), and nymphal development across

saplings was also significant (F 19, 92 = 7.85, r2 = 0.62, P<0.001). Oviparae (F 1, 56 = 12.87, r2 = 0.19, P<0.001) development time was significantly longer on poor (22.77+0.40) than good (20.38+0.51) nutritional hosts. Male development time was also significantly longer under poor (18.96+0.38) than good (17.14+0.38) nutrition (F

52 = 10.78, r2 = 0.17, P<0.005). Variability between saplings was highly significant on

oviparae (F 19, 38 = 16.50, r2 = 0.89, P<0.001) and male (F 19, 34 = 12.90, r2 -= 0.88, P<0.001) development time in the experiment.

106

Chapter 4 Reproductive sequencing and sex ratio

25 _

'ZI3s as 7:3 20

15 a) E C L 0 10

a) -a 5_ ca a) 2 0 Oviparae males Sexual morph type

Fig. 4-16: The development time of Drepanosiphum platanoidis oviparae and males on Acer pseudoplatanus saplings under poor (El) and good (E) nutrition in greenhouse conditions.

Adult longevity of the sexual morphs was similar on saplings with good (22.53+1.07)

and poor (21.82+0.96) host nutrition (F 1,110 = 0.029, r2 = 0.00026, P>0.05; Fig. 4-17).

Oviparae (27.09+0.80) lived significantly longer than males (16.96+0.74) (F 1, 110 --

86.04, r2 = 0.44, P<0.001), and between sapling variability was also significant (F 19, 92

= 2.40, r2 = 0.33, P<0.005). Oviparae (F 1, 56 = 4.06, r2 = 0.068, P<0.05), but not male

(F 1, 52 = 0.55, r2 = 0.010, P>0.05) adult longevity was significantly elongated under good (oviparae: 28.53+1.09; male: 16.53+1.02) rather than poor (oviparae: 25.54+1.13; male: 17.49+1.08) nutrition. Variability between saplings was significant for oviparae

(F 19, 38 = 9.27, r2 = 0.82, P<0.001) and male (F 19, 34 = 4.67, r2 = 0.72, P<0.001) adult longevity in the experiment.

35 -

) s 30 day

d ( 25 hoo lt 20 du f a 15 o ion t

a 10 dur n 5 Mea 0 Oviparae males Sexual morph type Fig. 4-17: The adult lifespan of Drepanosiphum platanoidis oviparae and males on Acer pseudoplatanus saplings under poor (El) and good (E) nutrition in greenhouse conditions.

107

Chapter 4 Reproductive sequencing and sex ratio

Influence of host plant quality on body size and potential fecundity of oviparae under greenhouse conditions:

The fecundity/size relationship may be determined by several factors, including an increase of ovariole and egg number with maternal body size (Figs. 4-19, 4-20 & 4-21). The body weight of oviparae was significantly lower on poor hosts than those on good

nutritional quality hosts (F 1,142 = 162.43, r2 = 0.53, P<0.001; Fig. 4-18) and ranged

between 0.625-3.087 mg (n oviparae —144) over both treatments.

2.5 - ) mg

( 2.0 ht ig 1.5 - we dy 1.0 bo

0.5 arae ip

Ov 0.0

Poor Good Host nutritional status

Fig. 4-18: The effect of plant nutritional status on body weight of Drepanosiphum platanoidis oviparae (n = 144) feeding on Acer pseudoplatanus.

The 'nutrient treatment x body weight' interaction (F 1,140 = 0.14, r2 = 0.00099, P>0.05) and host nutrition treatment (F 1,14o = 0.25, r2 = 0.0017, P>0.05; Fig. 4-19) did not significantly affect the number of ovarioles within oviparae. Without the bodyweight covariate, the GLIM model showed that good host nutrition enhanced mean ovariole numbers by 0.38; whilst with the bodyweight covariate, the model showed that good host quality led to a decrease in ovariole number by 0.21 per ovipara. The body weight

term caused a significant impact on ovariole numbers (F 1,140 = 5.24, r2 = 0.036, P<0.025), so that a positive relationship existed between ovariole number and oviparae body weight (y = 0.57x + 6.75, n= 144; Fig. 4-19). The 'between-tree' effect had a

significant impact on ovariole number per ovipara (F 17, 126 = 2.76, r2 = 0.27, P<0.001). The number of ovarioles per ovipara occurred mainly in couples, with a few ovarioles lying in the odd integer values (Fig. 4-19). The mean ovariole number for poor and good host nutrition was 7.43+0.14 and 7.81+0.14 respectively. The mean ovarioles:body weight ratio was 7.06+0.19 and 4.21+0.12 for poor and good plant nutrition. Thus, more ovarioles/ovipara per unit body weight were found on poor host nutrition.

108

Chapter 4 Reproductive sequencing and sex ratio

14_ L. a 12 - 0 ON ■

0 L_ 10 .000 ■OWN 4:11 Nom' • • a) 8 *IWO 43>COO1Mlm 0 pp •• IN • • •co 6- a001:10000 0311 III. ■ III 0 0 CO • • 4c-5 4- L _a E 2- Z 0 0.0 0.5 1.0 1.5 20 25 30 3.5 Adult ovipara body weight (mg)

Fig. 4-19: Relationship between adult ovipara body weight and ovariole number per ovipara (--- y = 0.57x + 6.75, r2 = 0.036, n= 144). Relationship between adult body size and the number of ovarioles within an ovipara of Drepanosiphum platanoidis on poor (<> y = 0.92x + 6.40, r2 = 0.026, n = 72), and good (M y = 0.61x + 6.61, r2 = 0.028, n = 72) nutrient quality Acer pseudoplatanus saplings.

There was a significant 'nutrient treatment x body weight' interaction (F 1,140 = 13.18, r2 = 0.028, P<0.001), and a marked difference in the total number of eggs per ovariole

between the two host nutrient treatments (F 1,140 = 21.55, r2 = 0.045, P<0.001; Fig. 4- 20). The number of eggs contained within each ovariole significantly increased with

maternal body weight (F 1,140 = 299.54, r2 = 0.63, P<0.001; y = 2.32x - 0.39). This was shown by a strongly positive relationship, especially in the case of the good host nutrition (Fig. 4-20). Without the body weight covariate, the GUM model showed that good nutritional host plants enhanced the mean egg number per ovariole by 2.40; with the covariate, the model also showed that good hosts led to an increase in egg number by 1.03 per ovariole. The 'between-tree' effect had a significant impact on egg number

(F 17,126 = 1.81, r2 = 0.20, P<0.025). The mean number of eggs per ovariole for poor and good host plant nutrition was 1.96+0.050 and 4.37+0.13 respectively. The eggs per ovariole:body weight ratio was 1.84+0.054 and 2.25+0.051 for poor and good nutrient treatments. Thus, more eggs/ovariole per unit body weight were found on good host nutrition.

109

Chapter 4 Reproductive sequencing and sex ratio

a) 8 II ■ ■ 7 > ■ 0 6 _ a) Q. 5 _ U) lb MIN rn 0) 4 _ a) M I ■ II ME 1E) 3 _ A • i ■ * *t> v • 0 1 1 <,/ et. . . II it) 10 2 *c. .IP. • '::"A-.1 E -4> • * 1_ oC!' z o 0 0.0 0.5 1.0 1.5 2.0 2.5 ao 3.5 Adult ovipara bocly %eight (mg)

Fig. 4-20: Relationship between adult ovipara body size and the number of eggs per ovariole of Drepanosiphum platanoidis on poor (0 - - - y = 0.53x + 1.37, r2 = 0.070, n = 72) and good (MI y = 2.10x + 0.26, r2 = 0.42, n = 72) nutrient quality Acer pseudoplatanus saplings. Relationship between the number of eggs contained within each ovariole with maternal body weight (y = 2.32x - 0.39, r2 = 0.63, n = 144).

A significant negative relationship was evident between adult body weight and the proportion of immature eggs per ovipara (X2 1 = 246.3, r2 = 0.27, P<0.001; y = —0.20x + 0.69; Fig. 4-21). This was reflected in the 'nutrient treatment x body weight' interaction which was also significant (X2 1 = 5.09, r2 = 0.0068, P<0.025). However, the plant nutrition treatments were not significantly different from each other (X2 1 = 0.114, r2 = 0.00015, P>0.05). Without the body weight covariate, the GLIM model predicted that good plant quality led to a decrease in proportion of immature eggs by 0.85 per ovipara; whereas with the covariate, the model predicted that enhanced nutrient quality lead to a 0.037 increase in the proportion of immature eggs per ovipara. The 'between-tree' effect was highly significant (X2 17 = 225.30, r2 =0.30, P<0.001). The mean immature egg proportion for poor and good host plant quality was 0.48+0.021 and 0.29+0.026 respectively. The proportion of immature egg:body weight ratio was 0.47+0.031 and 0.17+0.019 for poor and good host treatments. Thus, more immature eggs per unit body weight were found on poor host nutrition.

110 Chapter 4 Reproductive sequencing and sex ratio

1.0

a) 0.8 _ CTS E a E'5 0.6 - 0 I 46 I— c)a)_ 0 0.4 I *-E 6( 3 o CL 0 0.2 0

0.0

00 0.5 1.0 1.5 20 2.5 35 Adult ovipara Pod), might (mg)

Fig. 4-21: Relationship between adult body weight and the proportion of immature eggs per ovipara (-- y = —0.20x + 0.69, r2 = 0.27, n = 144). Relationship between adult body weight and the number of immature eggs as a proportion of the total eggs per ovipara of Drepanosiphum platanoidis on poor (o y — -0.13x + 0.62, r2 = 0.043, n = 72), and good (1111 y = —0.19x + 0.66, r2 = 0.16, n = 72) nutrient quality Acer pseudoplatanus saplings.

DISCUSSION

The reproductive sequence:

The reproductive sequence of the parthenogenetic sexuparae of D. platanoidis was the same under field and laboratory conditions. Oviparae were always the first-born progeny. The switch between alternative progeny types occurred once and was irreversible. In addition, there was a reproductive pause between sexual female and male production, and the daily birth rate for males was generally less than the birth rate for oviparae.

With oviparae being the first-born progeny in the reproductive sequence, it increases the likelihood that these aphids will add to the overwintering egg population. This is because the oviparae have the advantage of time to mature before leaf fall, and have the opportunity to mate with males produced by other sexuparae or await the maturation of males born later in the same reproductive sequence. In addition, it gives them time to complete oviposition of the overwintering eggs before the onset of adverse winter conditions — a result which agrees with the 'guaranteed fitness hypothesis'. This result corroborates the observations of many workers. In general, a parthenogenetic mother

1 1 1 Chapter 4 Reproductive sequencing and sex ratio

will first produce parthenogenetic and/or sexual females and later produce males (and under some conditions, a series of females continues until death): e.g. Acyrthosiphon pisum (Lamb & Pointing, 1972, 1975; MacKay et al., 1983), Acyrthosiphon svalbardicum (Strathdee et al., 1993), Aphis fabae (Tsitsipis & Mittler, 1977a); Aphis gossypii (Takada, 1988), Cryptomyzus spp. (Guldemond & Tigges, 1992), viciae (Lees, 1959, 1960), Myzus persicae (Blackman, 1972; Hales et al., 1989), Rhopalosiphum padi (Dixon & Glen, 1971; Austin et al., 1996) and Sitobion avenae (Watt, 1984; Hand & Wratten, 1985; Newton & Dixon, 1987).

This pattern of production of sexual morphs depends on the photocycle under which the parthenogenetic mothers develop. Since two different photoperiodic clocks regulate the production of sexual females and of males (Lees, 1989), this imposes a limitation on the reproductive sequence, so that the switch to males becomes irreversible (Lamb & Pointing, 1975; Lees, 1989). These observations suggest that the endocrine processes, which control the induction of males and of sexual females, impose some constraints on sex ratio allocation (Moran, 1993). It is likely that the sex of the embryos is determined by the circulating level of juvenile hormone (Hales & Mittler, 1983, 1987), and that the exact moment at which this falls below a certain critical level determines how many embryos develop into males (Foster & Benton, 1992).

Similar to the findings of Lamb & Pointing (1975), the reproductive pause of sexuparae lasted 3 - 4 days under field and poor nutrient host conditions, whilst it was reduced to 2.5 — 3.0 days on good host nutrition. This was still longer than that found in S. avenae (Watt, 1984). In addition, this study showed that the daily birth rates achieved for oviparae and males were similar under field and good host plant conditions. However, the birth rate of the males was generally lower than that of oviparae under poor and good hosts in the greenhouse. As the sexuparae, that produce males, do not live long afterwards, the reproductive pause and lower male birth rate may account for the lower male production (Newton & Dixon, 1987). This has been attributed to several factors: (i) male embryos taking longer to mature (Lamb & Pointing, 1975) or having a lower growth rate than female embryos (Searle & Mittler, 1981); (ii) a reduction in maternal hormone levels under short photoperiods during which male embryos develop, depressing embryogenesis (Tsitsipis & Mittler, 1977b); or (iii) some embryo resorption contributing to the lower reproductive rate, as observed in male-producing M. persicae (Searle & Mittler, 1981) and (Crema, 1979). However, this is not unique to male development as female embryos are also resorbed in A. fabae, A. pisum

112 Chapter 4 Reproductive sequencing and sex ratio

and M persicae (Searle & Mittler, 1981, 1982). Searle & Mittler (1981) suggest that egg abortion in M persicae may be due to abnormal chromosome behaviour caused by an intermediate titre of hormone during the switchover period.

Primary sex ratio and sex allocation:

As expected, the sex ratios of sexuparae offspring were strongly female-biased in both the field and greenhouse, despite subjecting the sexuparae to poor and good host plant conditions. A female-biased sex ratio is predicted in aphids when LMC occurs (e.g. Yamaguchi, 1985; Foster & Benton, 1992) with the ratio dependent on the density of foundresses (Kindlmann & Dixon, 1989). The female-biased sex ratios are found in species where the sexuals are confined to a single patch and where related males are likely to compete with each other for mates. In spyrothecae where the sexupara gives birth to apterous oviparae and males at the same site, inbreeding is almost certain to occur when population densities are low; thus it is not surprising that a female biased sex ratio is observed (Foster & Benton, 1992). Conversely, in S. avenae, where the males are alate, and the mothers of the female sexuals are also winged, such that inbreeding is unlikely, the sex ratio has no significant bias (Newton & Dixon, 1987). In the case of D. platanoidis, the flight ability and mobility of males suggest that migrations (probably as a means of avoiding inbreeding) are possible, which may change sex ratios between trees, branches and leaves (Kozlowski, 1991). Nevertheless, it seems that males tend to stay on particular leaves with loose aggregations of adult females for a considerable time (Kozlowski, 1991), thus suggesting a case for LMC. Even if females do not select mates actively, their progeny would profit from inheriting the traits of the most effective, i.e. vigorous male. In this way selection would maximise the number of copulations in males, also in the situations where several males are active on a leaf (Kozlowski, 1991). Furthermore, in non host-alternating tree- dwelling aphids, such as D. platanoidis, members of several clones are likely to be present on the same host, so outbreeding may sometimes be common (Ward & Wellings, 1994). However, in autumns with particularly low population densities, full- sib mating may lead to intense inbreeding amongst the sexuals.

Higher sex ratios (i.e. greater proportion of males) were noticeable under good host nutrition in the greenhouse. The increased number of males can be explained in terms of maximising the number of inseminated oviparae, overwintering eggs and thus fitness (Guldemond & Tigges, 1992). This result contrasts with the theoretical 'constant male

113 Chapter 4 Reproductive sequencing and sex ratio hypothesis' (Yamaguchi, 1985) where sexuparae with little parental investment invest it all in sons, but those with more than the threshold level of parental investment invest most, but not all, of the excess in daughters (Foster & Benton, 1992). It also contradicts the empirical findings of Moran (1993) where more males of the non host-alternating aphidine, Uroleucon gravicorne, were produced on deteriorating host plants. Nevertheless, the finding here could be explained by the better host nutrition elongating the reproductive sequence far past the maternal switchover period. With a longer host life, the continuous production of males in late reproductive life, resulting from ovulations in the virtual absence or low titre of JH (Hales & Mittler, 1987), will increase the sex ratio over time. On the other hand, the lower sex ratios associated with poor host conditions would happen if the reproductive sequence was cut short due to serious impoverishment of host nutritional resources, or when leaf fall occurs earlier so curtailing the sequence after a few males are produced. In some cases, aphids may not respond to deteriorating nutritional conditions by producing emigrants, since it is possible that the poor nutrition available to them when feeding on [mature] leaves could inhibit the development of alatae as suggested by Mittler & Kleinhjan (1970). However, it may have been reasonable to expect higher sex ratios on poor hosts, since the production of winged males would facilitate their escape to better quality trees. The production of alate morphs is known to be triggered by factors such as poor host nutrition and/or crowding (e.g. Dixon & Glen; 1971; Kawada, 1987; Kidd, 1990), often occurring prenatally (Forrest, 1970).

As D. platanoidis oviparae are apterous, they spend most of their lives on the plant upon which they are born. This plant (or the one very nearby) will also be the host upon which the overwintering eggs and first spring generations occur. Thus, investment in sexual daughters implies commitment to several generations of future reproduction upon the current host plant, extending into the following season. In contrast, investment in winged sons allows an escape from deteriorating local conditions that is not afforded by production of sedentary daughters. Moran (1993) suggests that these differences in dispersal abilities between the sexes may affect the selection on the sex ratio. In species with winged males and without a winged gynopara morph enforcing dispersal before birth of sexual females, males will disperse more than females. As a result, local resource competition maybe more intense among lines descending from daughters than among lines descending from sons (Clarke, 1978; Moran, 1993).

114 Chapter 4 Reproductive sequencing and sex ratio

The bias in the numerical sex ratio could be explained through the concept of equal parental investment. With a diploid genetic system and autosomal control of the sex ratio, mothers should be equally related to their sons and daughters (Yamaguchi, 1985). However, in this study, like that of Searle & Mittler (1981), a lower male production was clearly evident. This could be partially due to sons being more costly to produce than their sisters, since male embryos take longer to develop and are larger at birth than females (Lamb & Pointing, 1975); for example, S. avenae males are 1.76 times heavier than females at birth (Newton & Dixon, 1987). Thus, if a difference between the numerical sex ratio and the parental investment ratio was apparent (e.g. Yamaguchi, 1985) this could be caused by the sexual dimorphism of body size.

Sexual morph life history traits:

This study revealed that the nymphal development time of oviparae took considerably longer than that of males. Despite the probability of males having lower growth rates during nymphal development than females (Newton, 1986), D. platanoidis like S. avenae adult males were smaller than females thereby contributing towards shorter maturation times. As in A. pisum (Stadler & Mackauer, 1996), the time from birth to adult increased on low nutritive quality hosts. The significance of shorter male development is that by the time the oviparae have reached maturity, winged males which have been produced by the same sexuparae but later in the reproductive sequence, have reached maturity and are ready to mate with the oviparae (Leather, 1993). It is possible that the reproductive pause would also give the later born oviparae a buffered maturation time for when the faster developing males make their appearance. This synchronisation of the sexes is achieved in a similar way in the more sophisticated host-alternating species. Gynoparae, born at the beginning of their parents' reproductive life, migrate to the primary host where they give birth to oviparae; males born after gynoparae arrive on the primary host when the oviparae are mature (e.g. Dixon & Glen, 1971). This series of events agrees in part with the guaranteed fitness hypothesis, through maturation and mating phase synchronisation which gives rise to potentially more overwintering eggs.

No real difference in the proportion of nymphs surviving to adulthood was evident between the sexes in the field and greenhouse, but good host nutrition clearly increased nymphal survivorship. This response was expected because good nutritive status leads to larger adult aphids, which should produce larger nymphs, that are less likely to die

115 Chapter 4 Reproductive sequencing and sex ratio before reproducing than smaller ones found on poor nutrition (Dixon, 1970a). Likewise, Weisser & Stadler (1994) suggest that survival in small offspring may be lower than in large offspring, especially when plant quality is poor.

In contrast to the -carrot aphid Cavariella aegopodii (Dixon & Kundu, 1997), sexual females lived longer than males despite varying host nutrition. With the caged oviparae in this study being unmated, it is likely that they had longer adult lives than mated females occurring in the field. Dixon & Kundu (1997) showed that although males were the smaller of the two sexes, they had larger lipid reserves. As the reserves of the males are not destined for egg production, it is possible that more of the total energy is available for sustaining life. This is particularly important considering that males have to cease feeding in order to find females, whilst the females continue feeding during copulation (Kozlowski, 1991). Nevertheless, male aphids should potentially live longer than females because the fitness of a male depends on the number of females he fertilises, while the fitness of a female depends on her probability of mating and laying eggs (Dixon & Kundu, 1997). Therefore, selection is likely to favour males that spend more time searching, while selection favours females that mature and lay their eggs early in adult life. Unlike many other insects, aphids develop and lay eggs on the same plant, so little time is needed for locating oviposition sites.

Maternal size/offspring relationship:

The results presented here show that enhanced host nutrition increased the size of oviparae, but that the number of ovarioles per oviparae remained independent of host nutritive status. This is consistent with the findings of Leather & Wellings (1981) that ovariole number in the sycamore aphid remained constant on an intra-generational basis irrespective of external stimuli supplied during nymphal experience, such as varied temperatures and host qualities. Other work (e.g. Dixon & Dharma, 1980) reaffirms this, by demonstrating that nutritional differences in the host has no effect on the ovariole number within or between generations, but can result in the development of markedly different sized adults within an aphid species. Wellings et al. (1980) shows that ovariole number is not determined in direct response to food quality in six aphid species within and Callaphidinae. The ovariole constancy allows the aphid to be well suited to the conditions it is most likely to experience during a season (Dixon, 1987; Dixon & Wellings, 1982). Otherwise, on nutritionally poor hosts, aphids with many ovarioles will be less likely to survive to maturity than those with fewer ovarioles,

116 Chapter 4 Reproductive sequencing and sex ratio due to increased total cost of embryo maintenance (Ward et al., 1983a,b; Dixon, 1987). This possibly accounts for why oviparae nymphal survivorship and adult longevity were reduced on poor host nutrition in this study.

In contrast to Wellings et al. (1980), showing that oviparae (n=16) of D. platanoidis fell within the range of 1.0-1.4 mg and typically possessed ten ovarioles, the current investigation showed that oviparae (n=144) had a body weight range of 0.625-3.087mg, with an average of 7.618+0.1421 ovarioles per ovipara. The present study also revealed that in a minority of cases, odd ovarioles were noted within oviparae. This may be due to empty ovarioles not being clearly visible when counts were being made. Another point of disparity between this and other studies on D. platanoidis (e.g. Leather & Wellings, 1980) was that ovariole number increased with larger oviparae body size. This is at odds with the ovariole constancy theme, where ovariole number is dependent on generation, and within a generation the ovariole number is independent of body size (Leather & Wellings, 1980). Although the ovariole differences are programmed with individuals of each aphid generation having a fixed number, Wellings et al. (1980) suggest that this number may vary within a particular range, but that variation within a generation is not determined by differences between clones, but perhaps intra-clonally. However, the mechanism by which it is achieved in not understood. Dixon & Dharma (1980) suggest that the benefit of variability in ovariole number between individuals of the same clone may result in a more effective exploitation of spatial heterogeneity in food quality.

In accordance with the 'maternal size/offspring hypothesis', higher potential fecundities were achieved by oviparae since the total number of eggs/ovariole increased with maternal body weight and host nutrition. Also, more mature eggs were associated with large maternal size; in other words, the proportion of immature eggs declined with increasing maternal weight and with better host nutrition. These observations fit well with those of the first three parthenogenetic generations of D. platanoidis (Leather & Wellings, 1981). Since there is no evidence that ovulation continues after adult moult in D. platanoidis, total embryos/ovariole counts should therefore provide a good estimate of potential reproductive output. Dixon (1985) points out that large aphids have more embryos per ovariole and a higher reproductive rate, especially during the early reproductive phase. Taylor (1975) suggests that this greater initial reproductive rate may be a consequence of large individuals having more mature embryos per ovariole than small individuals at adult moult. However, caution has to be used when

117 Chapter 4 Reproductive sequencing and sex ratio interpreting the proportion of immature embyros, since random adult oviparae were examined. Therefore, a bias may exist in the results because the number of mature embryos are known to be dependent on female age (Stadler & Mackauer, 1996). Female age has also been associated with a concomitant reduction in oocyte chambers (within ovarioles) and a consequent reduction in fecundity (Carlson et al., 1998).

CONCLUSION

The appearance of the first ovipara and male, as well as the timing and duration of the maternal switch, were modified by the nutritional status of its host plant despite being exposed to the same abiotic conditions. The switch from producing females to producing males occurred at a particular time (c. 310 and 305 Julian days in 1996 and 1997 respectively) in D. platanoidis, and was unrelated to the number of females produced earlier in the sequence. If maternal hormone levels control the determination of males as Searle & Mittler (1981) suggest, and the hormonal level changes at a particular time in the development of an aphid, then this will result in a marked bias in favour of oviparae amongst the progeny of sexuparae. Irrespective of host nutrition, the faster development time of the later born males in the reproductive sequence enabled maturation synchronisation of the sexes. Although better host nutrition increased the likelihood of sexuals reaching adulthood, this did not elongate adult lifespan. It can be concluded that fecundity of oviparae was determined by intrinsic factors, controlling the number of ovarioles, and extrinsic factors (e.g. nutrition) influencing maternal size, the number of embryos/ovariole or the rate of embryo maturation.

Further research could focus on whether a correlation exists between D. platanoidis sexuparae weight with that of the progeny sex ratio. Often a negative relationship between sex ratio and size/fecundity of mothers occurs in aphidines (e.g. Newton & Dixon, 1987), suggesting a widespread trend for larger, more fecund mothers to produce more female biased broods (Moran, 1993). Larger size and greater fecundity in aphids are widely associated with better nutrition on particular plants due to environmental variation among hosts (e.g. Service & Lenski, 1982; Moran & Whitham, 1988). In addition, the biomass of the offspring could be determined by measuring the volume of embryos destined to be oviparae and males, or weighing new-born sexual morphs (cf. Newton & Dixon, 1987). This would assess whether the parental investment in the sexes is equal, as should be predicted in the case of random mating species (Fisher, 1930).

118 Chapter 5 Overwintering

Aspects of overwintering in the sycamore aphid

ABSTRACT

1. Spatial and temporal aspects of overwintering in the holocyclic sycamore aphid Drepanosiphum platanoidis (Schrank) were investigated. 2. Observations on spatial distribution on mature trees in this and Dixon's (1976a) study showed that D. platanoidis eggs were principally laid on trunks and branches of Acer pseudoplatanus, especially those of the upper canopy, at considerable distances from the buds. 3. Optimal oviposition sites occurred on rough bark surfaces. Thus trunks with their rough bark, followed by branches with moderate roughness supported more overwintering eggs than the relatively exposed smooth surfaces of twigs and terminal buds. 4. The selection of the optimal overwintering site may involve an array of multitactic responses, which ensure that the insect is protected to some extent from both extremes of the environment and from predators. 5. Observations using individual oviparae showed that hibernating eggs are aggregated on rough bark. Such aggregation benefits the ovipara through the advantages of shortened searching periods, reduced susceptibility to pre-reproductive mortality and increased realised fecundity by laying eggs in quick succession and close together. Whilst for the egg, clustering reduced egg surface exposure to desiccation and attack from natural enemies. 6. Experimental manipulations of varying oviparae densities showed intra-specific competition among oviparae for optimal oviposition sites. This was evident by the reduction in fecundity at a higher density (3.19 eggs/ovipara) than at a lower density (6.77 eggs/ovipara). Furthermore, under low oviparae levels density dependent egg mortality was evident, whereas under high oviparae density the proportion of eggs surviving into February did not show the same phenomenon. 7. Overwintering egg mortality was a function of the length of time spent in hibernation, with the greatest mortality rate occurring late winter. The causes of egg

119 Chapter 5 Overwintering

mortality are unknown, although it varied from 80.4 to 76.9 on trunks and 83.0 to 65.9% on branches in 1996 and 1997. More eggs were deposited in winter 1996, and greater egg mortality occurred than in winter 1997. 8. No statistically significant two-year cycle could be detected in the short-term time- series conducted at Bush N.R.S., Edinburgh and Silwood Park, Berkshire. 9. Autumnal and spring peak aphid densities gave more satisfactory explanation of the between-year dynamics than egg counts made at the outset of winter and spring aphid peak density. 10. In spring, fundatrix emergence always occurred before stage three (see Leather, 1996a) bud burst. March temperatures were important in relation to the timing of aphid egg hatch and bud burst.

INTRODUCTION

Overwintering is one of the most severe tests any insect species in a temperate climate has to endure during its life cycle (Leather et al., 1993). For aphids, photoperiod (Tsitsipis & Mittler, 1977a; MacKay et al., 1983), thermoperiod (Lees, 1959; Tsitsipis & Mittler, 1977b) and/or the quality of the host plant (Sethi & Swenson, 1967; Forrest, 1970) frequently act as cues for the control of overwintering. In addition, internal changes in aphids also determine their response to changes in their environment. Thus aphids which hatch from the overwintering eggs, when day-length is short and temperatures are low, cannot readily be induced to give rise to sexual morphs, but each successive generation can be more readily induced to give rise to sexuals (Dixon, 1973c). This is because of the operation of an intrinsic timing mechanism referred to as a `facteur fondatrice' (Bonnemaison, 1951) or as an 'interval time' (Lees, 1960, 1966).

Numerous and sometimes very complex strategies are employed by aphids during the winter preparation phase, but essentially these fall into two modes. Holocylic overwintering begins with the larviposition of the egg laying females, the oviparae. Anholocyclic overwintering begins when annual plant growth ceases (Leather, 1992). In some cases, aphids employ mixed strategies, depending on genotype, climate or geographic location; e.g. Myzus persicae (Blackman, 1972), and Sitobion avenae (Williams & Wratten, 1987). In the case of S. avenae, oviparae are produced in September and October, but then parthenogenetic virginoparae revert to virginoparous morph production (Williams & Wratten, 1987). This mixed strategy is useful in areas where winters are sometimes mild since this gives the aphid population a vigorous start

120 Chapter 5 Overwintering

in spring (Leather, 1993). If a severe winter is experienced, the anholocyclic population may fail to survive, so then a new population can be established from the holocyclic overwintering eggs (Blackman, 1971). Although anholocyclic viviparous forms of aphids are not as cold-hardy as eggs (Somme, 1969), they are reasonably well adapted to survive the winter temperatures that they are likely to encounter (Leather, 1993). Thus, Griffiths & Wratten (1979) demonstrated that anholocyclic morphs predominantly survive temperatures as low as —5°C, whereas some eggs can survive temperatures lower than —40°C (James & Luff, 1982). The anholocyclic life cycle is a more recent evolutionary development than the holocycle, and has arisen from both the autoecious (without annual alternation between primary/secondary hosts) and heteroecious (host- alternating) groups of aphids (Leather et al., 1988).

The sycamore aphid, Drepanosiphum platanoidis, is autoecious (monoecious) and holocyclic (Dixon, 1970b, 1973c, 1985) where it feeds, mates and overwinters as eggs on Acer pseudoplatanus L. In spring when the buds of sycamore begin to swell, the eggs hatch and the nymphs develop into the winged adults of the first generation, known as fundatrices (stem mothers). These adults are parthenogenetic and give rise to nymphs (fundatrigeniae) which develop into parthenogenetic virginoparae. Two parthenogenetic generations occur in succession until the onset of autumn when the nymphs, produced by sexuparae, develop into apterous egg-laying females and alate males. The sexual morphs mate and the oviparae lay their overwintering eggs (Dixon, 1973c).

Freshly laid aphid eggs are pale yellow to green, but they soon become shiny black, e.g. Aphis fabae (Blackman, 1974), Rhopalosiphum padi (Leather et al., 1993) and Acyrthosiphon svalbardicum (Strathdee et al., 1993). This dark colouration conceals them from predators dependent on vision, e.g. birds, and absorbs solar radiation so that in the spring they warm up faster and hatch (Leather et al., 1993). In addition, the choice of overwintering site can be critical to insect survival (Leather et al., 1993). Aphid eggs are all laid in semi-protected or sheltered sites such as at the base of herbaceous plants e.g. Acrythosiphon pisum (Dunn & Wright, 1955) and S. avenae (Holler, 1990); bud axils e.g. R. padi (Leather, 1980), Pterocallis alni (Gange & Llewllyn, 1988), Aphis .fabae (Way & Banks, 1964); or crevices in tree bark e.g. japonica (Komatsu & Akimoto, 1995) and D. platanoidis (Dixon, 1976a). These sites give some degree of protection from the action of natural enemies

121 Chapter 5 Overwintering and harsh weather conditions (Leather, 1993). Despite the adaptive advantages conferred by the melanic pigmentation and the choice of overwintering site, egg mortality is still very high in most aphid species (Leather, 1983, 1992, 1993). Egg mortalities throughout winter are usually about 70%, although this can vary even within species (Leather, 1990a). The length of winter, not its severity, can markedly influence the number of eggs surviving (Leather, 1980, 1981, 1990b; Leather & Lehti, 1981; Leather et al., 1993) and thus the initial success of the aphid population in the spring (Way et al., 1981; Leather, 1986). Egg mortality has also been shown to occur at a constant rate in mid winter, which Leather (1983) and Hand & Hand (1986) estimate as 3-5% per week. This mortality rate may be greater at the beginning (Gange & Llewellyn, 1988) and end (Leather, 1981, 1990b) of the winter period, perhaps as a direct result of predators or spring frosts (Leather, 1983).

The availability of suitable egg laying sites may be limited. Leather (1990a) demonstrated competition amongst oviparae for overwintering sites, although no account was taken of the ovipositional behaviour of aphids. Leather (1990a) showed that R. padi laid eggs in the axils of tight buds of where there is only space for 10-15 eggs per axil, and buds are in short supply. Thus some eggs are laid on top of other eggs, on the bud or on the bark adjacent to the bud. Such eggs maybe more conspicuous to predators and also more readily dislodged by wind or rain. There is thus fierce intra-specific competition for optimum overwintering sites (Leather, 1990a). This is evident from initial egg mortality being strongly density dependent, whereas total egg mortality is not density dependent (Leather, 1990a). Where only eggs laid in suitable sites remain, then egg mortality reduces to a constant rate of about 3-4% per week (Leather 1983, 1990a). Other aphids, e.g. Pterocallis alni, also show this heavy initial egg loss followed by a constant rate of decline (Gange & Llewellyn, 1988). This high initial loss of eggs is not seen in years when aphid densities are low (Leather, 1990b, 1993).

For the few species of holocyclic aphids that have been studied in detail, a high level of abundance in spring is followed by low numbers in autumn and vice versa (Dixon, 1970b, 1985, 1990c; Dixon & Kindlmann, 1998). This pattern has been observed in one species of host alternating aphid (A. fabae) and three species of autoecious arboreal aphids (D. platanoidis, , Chromoaphis juglandicola) (see Chambers et al., 1985; and references therein). Dixon (1985) suggested that alternating yearly

122 Chapter 5 Overwintering

patterns of abundance in D. platanoidis was due to the inverse relationship between spring and autumn aphid numbers, and a positive relationship between autumn and spring numbers (Dixon, 1985). Furthermore, Dixon (1970b, 1979, 1985) proposed that an overcompensated density-dependent factor acts within years, whereas the between year mortality is inversely density-dependent. The greater the number of aphids present in the autumn, the greater the number of sexuals that develop, and hence the greater the number of eggs laid. The more eggs that are laid in autumn, the more that survive to hatch into fundratrices the following spring (Dixon, 1985, 1970b). What is not known is whether the proportion of D. platanoidis eggs surviving winter is independent of density.

Dixon (1970b, 1979, 1985) suggested that most of the year to year fluctuations in sycamore aphid density were caused by weather factors, especially those acting in autumn and spring. High winds in autumn may dislodge and sometimes kill large numbers of parthenogenetic virginoparae and sexual morphs (Dixon & McKay, 1970), so that less oviparae are available to lay overwintering eggs. In spring, weather can act as a density disturbing factor by affecting the time of bud burst (Dixon, 1970b, 1976a). If eggs hatch in a late bud bursting spring, many of the nymphal fundatrices, which are exposed on the smooth surface of the bud scales, are washed off by heavy rain or eaten by birds before gaining access to the shelter of the highly nutritious unfurling leaves (Dixon, 1970b). However, in warm springs eggs hatch and buds burst earlier than in cool springs, which may give an advantagous start to aphid development and population growth (Dixon, 1987). Therefore, it is crucial to the success of the spring aphid population that the timing of egg hatch and bud burst, like the timing of sexual morph production and egg laying with leaf fall, is closely synchronised (Dixon, 1976b, 1987). Other researchers have also shown that the synchronisation between egg hatch and bud burst is the main factor influencing the survival, performance, and population density of herbivorous insects (Wittler & Waisanen, 1978; Watt & McFarlane, 1991; Hunter, 1990, 1992).

With these studies as background, this chapter aims to examine: (i) where oviparae naturally lay their eggs in relation to mature tree architecture; (ii) egg survival rate in relation to spatial position, micro-habitat type and initial egg density; (iii) whether there is intra-specific competition among oviparae for suitable ovipositional sites; (iv) whether eggs are laid in clusters, and to see if this is due to maternal behaviour (via site

123 Chapter 5 Overwintering preference) or micro-habitat shortage; (v) temporal abundance of overwintering eggs from year to year in two geographical regions to ascertain whether a two-year cycle in D. platanoidis population dynamics exists and what is the mechanism behind the hypothesised alternating pattern; (vi) whether the time of bud burst affects aphid hatch success rate; and finally (vii) whether climate (i.e. aerial temperature) determines the time of bud burst and/or fundatrix emergence date.

METHODS & MATERIALS

Field observations were carried out over ten and six-year consecutive periods in the north west and south east of the British Isles, starting from November 1982 and November 1992 respectively. In addition, more intensive field based observations took place from 26th November 1996 to 21St April 1997 and from 29th November 1997 to 25th April 1998 at Observatory Ridge (OS Ref. 5946 1689) within Silwood Park (near Ascot, Berkshire). The experimental field manipulations and subsequent monitoring also took place at Observatory Ridge between mid/late October 1997 to 24th April 1998.

Experimental Designs:

I. Spatial distribution of hibernating eggs over a mature tree (field observations):

Ten mature sycamore trees at the latter stages of leaf fall were selected from different areas of Observatory Ridge. Each tree (n=10 true replicates) was separated by at least a 30 m distance. The trees were selected on the basis of them possessing four similar sized branches pointing in the direction of the four main cardinal points (i.e. North, South, East & West). On each tree, fixed quadrat boundaries were permanently marked out using colour coded mapping pins in each of three main spatial zones: • Lower trunk (1.5m above the ground): fixed quadrats were set-up on the 4 main compass aspects. A clear acetate sheet with a 12 cm2 square grid, divided into 36 x 2 cm2 sampling units, was placed over the bark. • Mid branch (0.5m away from the branch base): fixed quadrats were established on the top, and undersides of each branch, which pointed to one of the main aspects. A clear rectangular 5 cm2 grid, divided into 25 x 1 cm2 sampling units, was placed over the bark.

124 Chapter 5 Overwintering

• Twigs (from terminal buds): a fixed quadrat was established on the outer most twig section of the four main branches. Again, a rectangular 5 cm2 grid, divided into 25 x 1 cm2 sampling units, was placed over the bark.

Observations were carried out systematically in each numbered sampling unit each month from mid November to April. In the first monitoring session the whole fixed quadrat was surveyed. For this, each overwintering aphid egg was assigned a position within the quadrat by placing the clear acetate grid over the bark. Thereafter only the sub-sample units containing eggs deposited within their boundaries were recorded in each successive monthly census count.

2a. Substrate preference and spatial oviposition patterns (experimental manipulation):

For oviposition to operate freely, an unrestricted egg-laying arena was required. On each of the ten independent mature trees, three 8 cm2 quadrats were established using TanglefootO boundaries. Each of the three quadrats had different substrate types, which constituted three treatments: • Rough creviced bark • intermediate bark texture • relatively smooth bark texture

Fifty oviparae were collected from the field on 21st October 1997, and placed into a small chamber containing freshly cut sycamore leaves. To this culture, ten adult males were introduced to ensure copulation of all the females (n.b. D. platanoidis show female biased sex allocation - see Chapter 4) at 15°C. Two days later, a randomly chosen ovipara was added to each quadrat. If the ovipara fell off or crawled into the sticky perimeter within one hour of starting the experiment, then a replacement was added. The inner 7 cm2 quadrat area was monitored four days later, for each of the three treatments, by placing a transparent acetate 1 cm2 x 49 grid. A record was kept on the overall number of eggs deposited per experimental arena, the number of eggs laid in relation to distance from a crevice, and whether there was any clustering of eggs.

2b. Competition for oviposition sites (experimental manipulation):

Another set of ten mature sycamore trees were selected within the vicinity of Observatory Ridge. Each tree (n=10 true replicates) was separated by at least a 15 m

125 Chapter 5 Overwintering distance. Tree selection was based on the criteria of trees having similarities in trunk girth, grooved bark, and degree of microflora covering.

One month prior to running the experiment, the experimental arenas were kept free of oviparae (which would otherwise give rise to the 'background sycamore egg population') and predators by stapling transparent net topped cartons and sealing the perimeter gaps with an odourless water-based adhesive glue. In early October 1997, two 9 cm2 fixed quadrats were established on each tree by applying two separate square Tanglefoot® perimeters on the north-ease facing trunk bark, with the corners permanently marked by mapping pins that were colour coded according to the treatment given. The sticky boundaries prevented oviparae escaping and/or captured crawling predators.

The field based manipulation of oviparae density commenced on 27th October 1997. Into each fixed quadrat, a low (n=3) or a high (n=20) density treatment using mated oviparae, taken from randomised greenhouse cultures, was introduced and allowed to oviposit unhindered. To avoid vertical bias, the two treatment positions were inversed at every alternate tree running in ascending numerical order. The central point between the two quadrats was situated approximately 1.25 m above the soil surface, and the fixed quadrats were separated by 0.5 m.

When surveying each oviposition arena only the inner 8 cm2 section was monitored to avoid 'edge effects', and a numbered transparent 64 x 1 cm2 grid was used to assess the number of eggs and their relative positions. Initial and successive monitoring time took place at monthly intervals until 24th April 1998, similar to that described in the observational study for natural distribution of overwintering eggs.

3. Temporal abundance of sycamore aphid eggs (overall population trends):

Sycamore egg populations on branches of mature trees were observed from winter 1982 to autumn 1991 at Bush N.R.S., Edinburgh, Scotland; and from winter 1992 to autumn 1997 at Silwood Park, Bershire, England. In mid November (i.e. after leaf-fall and just when the oviparae had died), 5 trees were monitored for the ten year period at Bush

North-east facing trunks were chosen because the bark offered a plentiful supply of aphid-sized cracks, and NE+30° aspects offered most abundant microflora cover. N.E. trunk directions also represented the most favoured oviposition sites, since observations in winter 1996 showed that easterly followed by northerly facing trunk bark had the largest egg load.

126 Chapter 5 Overwintering

N.R.S., and 52 trees for the following six years at Silwood Park. A 2 cm2 template was used for randomised counting over a section of branch bark located between 0 to 2m away from the trunk. This was repeated 30 times for each tree. However, when initial egg density for each winter was low, the sampling effort was increased to reduce the frequency of zeros in the data. In addition, time of stage 3 bud burst (see Leather, 1996a), daily aerial temperatures, weekly counts of aphids found on 40 leaves per mature sycamore prior to aestivation were recorded at Silwood Park.

Data transformation and statistical analyses:

Data were analysed using generalised linear models (Aitkin et al. 1989) and non- parametric tests (Siegel & Castellan, 1956). All parametric statistical analyses were carried out in GLIM (version 3.77 copyright 1985 Royal Statistical Society, London; Crawley, 1993). Using GLIM, full models were fitted initially containing a set of explanatory variables, and for categorical variables, interaction terms. Non significant terms were progressively discarded to derive a parsimonious model containing only significant (P<0.05) terms or interactions.

Spatial distribution of sycamore aphid eggs over mature trees during the winters of 1996/97 and 1997/98 was statistically analysed at three spatial scales: trunk, branch and twig. Trunk and twig data were analysed using two-way ANOVA specifying 'year' with two levels and 'compass aspect' with four levels. A three-way ANOVA was performed on branch data by declaring 'year' with two levels, 'aspect' with four levels and 'position' (i.e. top and bottom facing surfaces) with two levels. The response variable was declared as 'average number of eggs/cm2 laid at the beginning of winter' in each analysis of variance.

The percentage survival of overwintering eggs on trunks and branches2 was determined using angular transformation followed by regression and ANOVA, and ANCOVA analyses. The factors were 'year' with two levels, 'aspect' with four levels, and branch `position' with two levels. In ANCOVA, the full model was fitted (using the initial egg density as the covariate) and individual terms removed. The y variable was the percentage of eggs surviving in the quadrat at each time interval, which was then arcsine transformed. To avoid conflating the data, analyses were performed for the

2 Since twigs supported a sparse initial egg density, statistical analysis of survivorship would be invalid.

127 Chapter 5 Overwintering beginning (November) and end of winter (February), November to December, December to January, and finally January to February, in order to establish the egg survival rate at specific time intervals throughout winter. Only data from November to February were used, since the natural mortality of sycamore eggs during March and April could be confounded by predator activity, or eggs emerging from hibernation early.

The egg-laying patterns of individual oviparae were observed to test whether oviparae distribute their eggs in aggregations because of maternal behavioural preference or rather because of limited suitable ovipositional microhabitats. To determine whether oviparae prefer to oviposit on rough bark surfaces, a one-way ANOVA was performed using 'bark texture' as the factor with three levels. To assess aggregation in overwintering eggs, the number of eggs deposited within each 1 cm2 of the fixed quadrat were collated. The frequencies from 0 to 5 eggs laid per ovipara on each of ten trees were calculated in order to estimate k (i.e. density of the egg cluster) and variance/mean ratio of the negative binomial distribution.

Manipulation of oviparae density examined maternal competition for ovipositional sites. This was assessed by (a) looking at the initial number of eggs deposited on bark under low and high density; and (b) determining the proportion of eggs surviving under high and low densities in February (i.e. before fundatrix hatching time). Firstly, to assess extent of oviposition success between the two oviparae densities, ANOVA using poisson errors was carried out on the total egg counts in each quadrat, by declaring the density treatment as a factor with two levels. To test whether individual oviparae lay more eggs under low density conditions, an ANOVA using normal error variances was performed on the number of eggs deposited at the beginning of winter. Treatment was specified as the factor (with two levels), and the response variable was declared as the number of eggs laid per oviparae per 1 cm2 bark. To derive the response variable, the total number of eggs laid in the 8 cm2 fixed quadrat was divided by 64 to obtain the density per 1 cm2, and again divided by the number of oviparae placed in the experimental arena (i.e. for low density divide by 3, and for high density divide by 20). Secondly, to test whether egg mortality during winter was higher with elevated initial egg densities, regression, ANOVA and ANCOVA analyses were performed using binomial errors. The density treatment was the factor (with two levels), initial total egg

128 Chapter 5 Overwintering count in the quadrat as the binomial denominator (and covariate in ANCOVA), and the total number of eggs in the quadrat surviving in mid February as the response variable.

Temporal distribution of sycamore aphid eggs over the ten and six year period were assessed using the Spearman's Rank-Order Correlation Coefficient, rs (Sokal & Rohlf, 1995). Year was specified as n. With n<10, the critical rs values for significance testing were obtained from Table P in Siegel (1956). To assess whether there were two-year cycles in operation at Bush N.R.S. and Silwood Park, time (X variable) was plotted against the mean number of eggs/cm2 laid in November each year (Y variable). In addition, the logged average number of eggs laid at the beginning of each winter Nt became the X variable, and the following logged winter's eggs at Nt+t became the Y variable. To discover the mechanism for the possible alternating cycle, the relationship between winter egg density and the following summer populations at both locations were examined again using Spearman's Rank Correlation. In the latter instance, prior to the non-parametric analyses, the initial egg density was expressed as the egg count per 1 2 cm . The maximal number of aphids occurring before aestivation (`spring aphid peak density') was derived by calculating the average number of aphids over forty leaves per tree each week, and then dividing by the number of trees. The initial number of eggs per winter and the spring peak aphid density was then log transformed (Y + 1).

To determine whether bud burst affects aphid egg hatch success, a regression analysis was performed using the Julian date of bud burst and mean spring peak aphid density per year. To assess whether bud burst has an effect on the change in population size between log (autumn aphid density + 1) and log (spring aphid density + 1) a k-value (i.e. the difference between autumn and the following spring's population density) was produced; this k-value was then regressed against bud burst Julian date. To determine the synchrony between the Julian date of stage 3 bud burst and timing of egg hatch, a simple regression was used. To ascertain if climate can predict the bud burst date and/or the time of fundatrix appearance, the Julian dates of each Y-variable per year were regressed against the mean daily aerial temperature measured from 7th March each year at Silwood Park.

RESULTS

Spatial distribution and survivorship of sycamore aphid eggs over mature trees:

129 Chapter 5 Overwintering

Initial abundance of eggs on trunks:

Although there were more eggs deposited on trunk bark in 1996 (average eggs/cm2 = 0.28+0.06) than 1997 (average eggs/cm2 = 0.20+0.03), this was not statistically significant (F 1,72 = 1.26, r2 = 0.02, P>0.05; Fig. 5-1 a). The abundance of eggs in relation to aspect followed the ranking: EAST (0.35+0.09) > NORTH (0.28+0.08) >

SOUTH (0.18+0.05) > WEST (0.16+0.05), but differences were not significant (F 3,72 = 1.60, r2 = 0.06, P>0.05). Furthermore, the year x aspect interaction was not significant

(F 3,72 = 0.23, r2 = 0.009, P>0.05).

0.7

0.6 a)

0.5 1-5 8 01996 Z ‘_ 0.4 c\I (1) E .G 1997 0.3 u) 0) 0 cp C O E 0.2 J .c rn a) 0.1

0

North South East West Trunk aspect Fig. 5-1 a: Mean abundance of Drepanosiphum platanoidis eggs (+ S.E.) on Acer pseudoplatanus trunk bark in relation to year and aspect.

Initial abundance of eggs on branches:

The number of eggs deposited on branch bark in 1996 (average eggs/cm2 = 0.048+0.01) was more than double that of 1997 (average eggs/cm2 = 0.023+0.007), and this

difference was significant (F 1 ,87 = 5.13, r2 = 0.04, P<0.05; Fig. 5-1 b). The abundance of eggs laid in relation to branch aspect (EAST: 0.055+0.01 > NORTH: 0.033+0.01 >

SOUTH: 0.025+0.009 > WEST: 0.020+0.007) was not significant (F 3,87 = 2.19, r2 = 0.06, P>0.05). Moreover, the lower branch surfaces (0.055+0.01) supported

significantly higher numbers of sycamore eggs (F 1,87 = 12.1, r2 = 0.11, P<0.001) than the upper branch surfaces (0.016+0.005), and this was consistent throughout the four cardinal points. However, all the interaction terms showed non significant results (year

x aspect: F 3,87 = 0.37, r2 = 0.01, P>0.05; year x locale: F 1,87 = 0.65, r2 = 0.006, P>0.05;

aspect x locale: F 3,87 = 0.26, r2 = 0.007, P>0.05).

130

Chapter 5 Overwintering

0.16 _

0.14 Top

a) ber) 0.12 0Underneath

0.10 Co Novem (

csi ter 0.08 E in

U) f w

o 0.06 _ a) ing

inn 0.04 beg 0.02 _

0.00

1996 1997 1996 1997 1996 1997 1996 1997 NORTH SOUTH EAST WEST Branch aspect/Year

Fig. 5-1 b: Mean abundance of Drepanosiphum platanoidis eggs (+ S.E.) on Acer pseudoplatanus branches in relation to year, aspect and surface position (i.e. upper or lower facing surfaces).

Initial abundance of eggs on twigs:

Terminal twig sections supported similar numbers of sycamore eggs in both 1996 (average eggs/cm2 = 0.019+0.008) and 1997 (average eggs/cm2 = 0.019+0.01), and this

was not statistically significant (F 1,42 = 0.0, r2 = 0.0, P>0.05; Fig. 5-1 c). There were no differences in the way sycamore eggs were distributed between the four aspects (NORTH = 0.017+0.01; SOUTH = 0.033+0.02; EAST = 0.015+0.008; WEST =

0.012+0.009), which was also not significant (F 3,42 = 0.42, r2 = 0.03, P>0.05). This was

also reflected in the non significant year x aspect interaction term (F 3,42 = 0.11, r2 = 0.007, P>0.05).

0.09

r te

in 0.08 w f

o 0.07 ing n 0.06 in 01996

beg 0.05 1997 the t 0.04 id a la 0.03 2 /cm

s 0.02

egg 0.01 No. 0.00 North South East West

Twig aspect Fig. 5-1 c: Mean abundance of Drepanosiphum platanoidis eggs (+ S.E.) on Acer pseudoplatanus terminal twig sections in relation to year and aspect.

131

Chapter 5 Overwintering

Survivorship of eggs on trunks:

Correlating the initial number of D. platanoidis eggs with survivorship showed that strong density dependent survival was occurring at each temporal phase during winter (Table 5-1, Figs. 5-2 a, c & d), except from November to December as revealed by regression analysis (Table 5-1; Fig. 5-2 b).

0

0 10 20 30 40 50 60 Initial no. eggs on trunks in November

Figs. 5-2 a: Survival of Drepanosiphum platanoidis eggs over the entire winter (November to February) on Acer pseudoplatanus trunks in the winters of 1996 and 1997 (F 1,65 = 9.55, r2 = 0.13, P<0.005, y = 0.58x + 15.19).

100 •• • •• •• • • • • _a11) 80 _ ••••••• • • • .• •• • 0a) • • a) 60 • • • • 0 • O • • • ir) 40 E a) • • 2 20

0 0 10 20 30 40 50 60 Initial no. eggs on trunks in Nbverrber

Fig. 5-2 b: Survival of Drepanosiphum platanoidis eggs from November to December on Acer pseudoplatanus trunks in the winters of 1996 and 1997 (F 1,65 = 3.23, r2 = 0.14, P= n.s., y = 8.62Ln(x) + 58.41).

132 Chapter 5 Overwintering

100 • • • • • • 80 • ♦ ca ••• • • • 2 Ws 60 • • RI 0 • •>— s_ • • • • 40 • • ?•,`•' 8 • 0 20

0 .66 0 10 20 30 40 50 60 No. eggs on trunks in December

Fig. 5-2 c: Survival of Drepanosiphum platanoidis eggs from December to January on Acer pseudoplatanus trunks in the winters of 1996 and 1997 (F 1,65 = 5.20, r2 = 0.1135, P<0.05, y = -0.022x2 + 1.83x + 51.42).

100 _ • L' 80 - • . E • • 4—2 15a) 60 U— • > 0 • • 40 - (ti • • 0 • --)cri 20- •• 0 •N 0 10 20 30 40 50 60 No. eggs on trunks in January

Fig. 5-2 d: Survival of Drepanosiphum platanoidis eggs from January to February on Acer pseudoplatanus trunks in the winters of 1996 and 1997 (F 1,65 = 4.54, r2 = 0.12, P<0.05, y = -0.059x2 + 3.082x + 23.34).

ANOVAs and ANCOVAs showed that there was no difference in survivorship between the two years observed (Table 5-1; Figs. 5-3 a, b & c), with the exception of January to February (Table 5-1; Fig. 5-3 d). The survival of eggs in relation to aspect was not significantly different in any of the time intervals (Table 5-1; Figs. 5-3 a to d).

133 Chapter 5 Overwintering

Table 5-1: Probability of survival of Drepanosiphum platanoidis eggs on Acer pseudoplatanus trunks in relation to year, aspect and initial egg density throughout the winters of 1996 and 1997.

WINTER PERIOD/ ANOVA ANCOVA MAIN EFFECT November - February F d.f. rz P F d.f. rz P Year 2.21 1, 65 0.033 n.s. 3.81 1, 62 0.049 n.s. Aspect 1.58 3, 63 0.070 n.s. 1.10 3, 65 0.044 n.s. Initial egg density 9.55 1, 65 0.13 <0.005 9.55 1, 66 0.13 <0.005 November 4 December F d.f. rz P F d.f. r2 P Year 0.076 1, 65 0.0012 n.s. 0.013 1, 62 0.00020 n.s. Aspect 0.59 3, 63 0.027 n.s. 0.48 3, 65 0.022 n.s. Initial egg density 3.23 1, 65 0.047 n.s. 3.23 1, 66 0.047 n.s. December 4 January F d.f. r2 P F d.f. rz P Year 2.19 1, 65 0.033 n.s. 3.50 1, 62 0.049 n.s. Aspect 0.81 3, 63 0.037 n.s. 0.64 3, 65 0.028 n.s. Initial egg density 5.20 1, 65 0.074 <0.05 5.20 1, 66 0.074 <0.05 January February F d.f. rz P F d.f. rz P Year 4.86 1, 65 0.070 <0.05 7.52 1, 62 0.097 <0.01 Aspect 1.68 3, 63 0.074 n.s. 1.29 3, 64 0.050 n.s. Initial egg density 4.54 1, 65 0.065 <0.05 7.27 1, 65 0.095 <0.01

50 - L.' as 2 40 0 from

l U_

a 0 30 - iv a) -0 20 E a) % surv o 10 _

0 96 97 96 97 96 97 96 97 North South East West Trunk Aspect/Year

Fig. 5-3 a: Percentage survival of Drepanosiphum platanoidis eggs (+ S.E.) over the entire winter (November to February) on Acer pseudoplatanus trunks in relation to year and aspect in the winters of 1996 and 1997.

100

r

be 80 -

em from

l 60 - Dec iva to v 40 - ber m % sur 20 Nove

96 97 96 97 96 97 96 97 North South East West Trunk Aspect/Year

Fig. 5-3 b: Percentage survival of Drepanosiphum platanoidis eggs (+ S.E.) from November to December on Acer pseudoplatanus trunks in relation to year and aspect in the winters of 1996 and 1997.

134 Chapter 5 Overwintering

100 -

?' 80 cu E 2 2 co 4- 60 _ o

40 E 8 U a) 20 _ 0

0 96 97 96 97 96 97 96 97 North South East West Trunk Aspect/Year

Fig. 5-3 c: Percentage survival of Drepanosiphum platanoidis eggs (+ S.E.) from December to January on Acer pseudoplatanus trunks in relation to year and aspect in the winters of 1996 and 1997.

70 -

60 ry 50 - brua

l from 40 a Fe iv to

30 - ry

surv 20 %

Janua 10 _

0 96 97 96 97 96 97 96 97 North South East West Trunk Aspect/Year

Fig. 5-3 d: Percentage survival of Drepanosiphum platanoidis eggs (+ S.E.) from January to February on Acer pseudoplatanus trunks in relation to year and aspect in the winters of 1996 and 1997.

Survivorship of eggs on branches:

Survival of eggs was not significantly density-dependent over the whole winter (November to February) (Table 5-2; Fig. 5-4 a) and during the initial period of winter (November to December) (Table 5-2; Figs. 5-4 b). However, density dependent survival was evident in the mid to late phases of winter (December to January and January to February) (Table 5-2; Fig. 5-4 c & d).

135 65.92). Fig. 5-4b:Survivalof pseudoplatanus 3.19x +15.36). on pseudoplatanus Fig. 5-4a:Survivalof 49.67x-0.17). Fig. 5-4c:Survivalof Chapter 5 Acer pseudoplatanus

% survivalfro m % survival from % survivalfrom Z u_ _o 0 a) at) 0 E December to January November to December t 100 - 100 100 60 - 80 - 20 _ 40 - 60 _ 40 - 80 _ 20 20 - 40 - 60 - 80 branches inthewintersof1996and1997(F branches inthewintersof1996 and1997(F 0 - 0

- 0

Drepanosiphum platanoidis branches inthewintersof1996and1997(F • • • Drepanosiphum platanoidis Drepanosiphum platanoidis • Initial no.eggslaidonbranchesinNovember Initial no.eggslaidonbranchesinNovember • • • 2 2 •

1

No. • • • • • • eggs inonbranchesDecember • • • • • 4 4 • • 136 2

eggs overtheentirewinter(NovembertoFebruary) • • eggs fromNovembertoDecemberon eggs fromDecembertoJanuary on 1 1 , , 35 = 35 = 6 6 • • • 10.64, r 3 0.13, r

1 , 35 = 2 =0.0038,Pn.s.,y-0.57x+ 2 =0.27,P<0.005,y-7.55x 0.83, r • • 8 8 • • 4 2

=0.023,Pn.s.,y

Overwintering 10 10 5 Acer Acer 2 +

Chapter 5 Overwintering

100 • •

• 80 E • Tu U- 60 - >_ o • • coL' 40 ,119, c 20 _

0 1 2 3 4 5 No. eggs on branches in January

Fig. 5-4 d: Survival of Drepanosiphum platanoidis eggs from January to February on Acer pseudoplatanus branches in the winters of 1996 and 1997 (F 1,35 = 17.26, r2 = 0.41, P<0.001, y = -8.00x2 + 48.40x + 3.12).

No significant differences in egg survivorship were found between the two years and between the four aspects during the winter period (Table 5-2; Figs. 5-5 a to d). Analysis of variance showed that survival on the undersides of branches (`position') was significantly different from egg survival on the upper branch surfaces during early (November to December) and mid (December to January) winter (Table 5-2; Fig. 5-5 b & c); but when the initial egg density was taken into account (ANCOVA), the branch position was only important for survival in early winter (November to December).

Table 5-2: Probability of survival of Drepanosiphum platanoidis eggs on Acer pseudoplatanus branches in relation to year, aspect, position and initial egg density effects throughout the winters of 1996 and 1997.

WINTER PERIOD/ REGRESSION ANCOVA MAIN EFFECT November - February F d.f. r2 P F d.f. r2 P Year 2.33 1, 35 0.062 n.s. 2.75 1, 31 0.069 n.s. Aspect 0.56 3, 33 0.048 n.s. 0.97 3, 34 0.077 n.s. Position 3.71 1, 35 0.096 n.s. 3.16 1, 35 0.083 n.s. Initial egg density 0.83 1.35 0.023 n.s. 0.83 1, 36 0.023 n.s. November - December F d.f. r2 P F d.f. r2 P Year 2.59 1, 35 0.069 n.s. 1.32 1, 31 0.028 n.s. Aspect 0.61 3, 33 0.052 n.s. 2.82 3, 34 0.18 n.s. Position 5.92 1, 35 0.14 <0.025 6.38 1, 35 0.16 <0.025 Initial egg density 0.13 1, 35 0.0038 n.s. 0.66 1, 35 0.016 n.s. December - January F d.f. r2-- P F d.f. r2 P Year 1.49 1, 35 0.041 n.s. 0.93 1, 31 0.021 n.s. Aspect 0.17 3, 33 0.015 n.s. 0.30 3, 34 0.020 n.s. Position 6.29 1, 35 0.152 <0.025 3.08 1, 35 0.064 n.s. Initial egg density 10.64 1.35 0.23 <0.005 10.64 1, 36 0.23 <0.005 January 4 February F d.f. r2 P F d.f. r2 P Year 1.29 1, 35 0.035 n.s. 0.29 1, 31 0.0057 n.s. Aspect 0.15 3, 33 0.013 n.s. 1.22 3, 34 0.070 n.s. Position 2.59 1, 35 0.069 n.s. 0.28 1, 35 0.0055 n.s. Initial egg density 17.26 1, 35 0.33 <0.001 0.33 1, 36 0.33 <0.001

137

Chapter 5 Overwintering

100 1:2 u_ •O 80 - iu E 60 - a) z0 E 40 0 To2> 20 -

rnz

Top Top Under Under Top Top Under Under Top Top Under Under Top Top Under Under 96 97 96 97 96 97 96 97 96 97 96 97 96 97 96 97 North South East West Aspect/Year/Locale

Fig. 5-5 a: Percentage survival of Drepanosiphum platanoidis eggs S.E.) over the entire winter (November to February) on Acer pseudoplatanus branches in relation to year, aspect and locale in 1996 and 1997.

r be 100 cem De

80 to

ber 60 vem 40 - l No iva

v 20 -

% sur 0

Top Top Under Under Top TopTop Under Under Top Top Under Under Top Top Under Under

96 97 96 97 96 97 96 97 96 97 96 97 96 97 96 97

North South East West Aspect/Year/Locale Fig. 5-5 b: Percentage survival of Drepanosiphum platanoidis eggs (+ S.E.) from November to December on Acer pseudoplatanus branches in relation to year, aspect and locale in 1996 and 1997.

6'3 100

-o 80 _ .0 E 60 - a) a) r21 40 _ 2 To 20 2

O Top Top Under Under Top Top Under Under Top Top Under Under Top Top Under Under 96 97 96 97 96 97 96 97 96 97 96 97 96 97 96 97 North South East West Aspect/Year/Location

Fig. 5-5 c: Percentage survival of Drepanosiphum platanoidis eggs (+ S.E.) from December to January on Acer pseudoplatanus branches in relation to year, aspect and locale in 1996 and 1997.

138

Chapter 5 Overwintering

ai 100 a) 80

60

E 40 2. 20 _

0 I - r Tcp Top Under Under Top Top Under Under Top Tcp Under Under Top Tcp Under Under 96 97 96 97 96 97 96 97 96 97 96 97 96 97 96 97 North South East West Aspect/Years/Locale

Fig. 5-5 d: Percentage survival of Drepanosiphum platanoidis eggs (+ S.E.) from January to February on Acer pseudoplatanus branches in relation to year, aspect and locale in 1996 and 1997.

In sum, D. platanoidis eggs had greater chances of survival in 1997 than 1996; and trunk bark provided more safe hibernating sites than branches in 1996, but this was not the case in 1997.

Table 5-3: Mean percentage survival of Drepanosiphum platanoidis eggs during the whole winter period and at each monthly interval in relation to trunks and branches of Acer pseudoplatanus during 1996 and 1997.

WINTER PERIOD TRUNKS BRANCHES 1996 1997 1996 1997 November to February 19.61 23.06 17.01 34.10 November to December 73.19 75.68 58.33 76.03 December to January 57.53 69.03 43.06 62.82 January to February 28.20 43.06 27.78 45.51

Substrate selection and oviposition patterns by individual oviparae:

Bark substrate significantly influenced the number of eggs laid (F 2, 27 = 6.23, r2 = 0.32, P<0.01; Fig. 5-6). More than twice as many D. platanoidis eggs were laid on the rough bark than on bark of intermediate roughness.

139 Chapter 5 Overwintering

0.30

0.25 -0 6- 2 c0.20

Es 0.15 O c as o 0.10 2 0.05

0.00 ROUGH INTERMEDIATE SMOOTH Bark Type

Fig. 5-6: Mean level of oviposition per ovipara on three types of bark substrate (+ S.E.).

Estimating k (i.e. degree of clustering) of the negative binomial distribution and the variance/mean ratios (Crawley, 1993) showed that the distribution of D. platanoidis eggs over rough bark was highly aggregated (k

Table 5-4: Distribution of Drepanosiphum platanoidis eggs laid on rough bark of Acer pseudoplatanus during November 1997. High aggregation is indicated by k<1.

TREE K VALUE MEAN VARIANCE MEANNARIANCE RATIO 1 0.43 0.51 1.09 2.13 2 60.59 0.02 0.02 1.00 3 2.16 0.24 0.27 1.11 4 0.42 0.12 0.14 1.24 5 0.14 0.16 0.30 1.88 6 0.33 0.27 0.62 2.32 7 0.05 0.14 0.42 2.92 8 0.06 0.06 0.10 1.64 9 0.21 0.53 1.30 2.44 I 10 0.001 0.10 0.09 0.92 •

Density-dependent oviposition and survival:

Significantly more eggs were laid by oviparae under low density than under high density treatment (X2 = 236.30, d.f. = 1, P<0.001; Fig. 5-7 a). In each quadrat, an average of 6.77 and 3.19 eggs per ovipara were laid under low and high density treatments respectively.

140

Chapter 5 Overwintering

80 70 - 60

It 0_ ci) 40 cr) 30 O c 20 a)c'cl 13 2 0

LOA/ HG-I Ouiparae density

Fig. 5-7 a: Total number of Drepanosiphum platanoidis eggs laid in an 8 cm2 quadrat by oviparae under low (n=3) and high (n=20) density treatment (+ S.E.).

Significantly fewer eggs per ovipara were laid under the high oviparae density

compared to low density (F 1,18 = 8.47, r2 =0.32, P<0.01; Fig. 5-7 b). The fecundity of oviparae under high density treatment was 47.14% of the oviparae performance under low density, despite the nearly sevenfold (i.e. x 6.67) difference in the initial oviparae numbers between the two treatments.

0.14

0.12

0.10

0.08

0.06

0.04

0.02

0.00 LON HGH Oviparae density

Fig. 5-7 b: Mean number of Drepanosiphum platanoidis eggs laid by an ovipara in low (n = 3) and high (n = 20) density treatments (+ S.E.).

Regression analysis showed that egg density in November significantly affected the proportion of eggs surviving in February (X2 = 6.64, d.f. = 1, r2 = 0.23, P<0.01; y =

141 Chapter 5 Overwintering

0.0043x - 0.15; Fig. 5-8). ANOVA showed that the mean proportions of eggs surviving low (0.57+0.04) and high (0.50+0.03) density treatments between November 1997 and

February 1998 were not significantly different (X2 = 1.15, d.f. = 1, r2 = 0.005, P>0.05). However, ANCOVA (using the initial number of eggs as a covariate) showed that the proportion of eggs surviving was significantly affected by the two density treatments

(X2 = 5.55, d.f. = 1, r2 = 0.19, P<0.025). ANCOVA also showed that the proportion of eggs surviving into February was significantly affected (X2 = 12.04, d.f. = 1, r2 = 0.42,

P<0.001). The treatment x initial egg density interaction was not significant (X2 = 0.63, d.f. = 1, r2 = 0.022, P>0.05).

0.8 0.7 • co ■ rn II • 0.6 - • 0.) 0 ■ m— a) cti 0.5 ,I•••• ,. • • ■ o • • c 0.4 - o ■ "-Eo '-o .1 0.3 - n rn 2 CL 0.2 _ 0.1 _ 0 0 23 40 63 80 1(X) 120 140 Initial egg mint in Nbv. 1997

Fig. 5-8: Overwintering survivorship of Drepanosiphum platanoidis eggs on Acer pseudoplatanus trunk bark in relation to two initial egg density treatments: • — low density (y = -0.005x + 0.66, r2 = 0.19), Ill - - high density (y = 0.002x + 0.38, r2 = 0.61). Relationship between egg density in November and the proportion of eggs surviving in February (Regression: X2 = 6.64, d.f. = 1, r2 = 0.23, P<0.01; y = 0.0043x - 0.15).

Temporal abundance of sycamore aphid eggs:

To explain inter-year differences involved in the cyclic biennial events of D. platanoidis, winter egg densities were compared. Egg abundance did not significantly alternate every other year at Bush N.R.S. (Spearman's r, = 0.15, n = 10, P>0.05) and

Silwood Park (Spearman's rs = -0.60, n = 6, P>0.05), despite the initial see-saw trends in the first few years of each time series (Table 5-5; Figs. 5-9 a & b). Overwintering D. platanoidis populations did not show density-dependent inward spiralling with a time lag of one year at Bush N.R.S. (Spearman's rs = -0.43, n = 9, P>0.05; Figs. 5-10 a) and Silwood Park (Spearman's rs = -0.20, n = 5, P>0.05; Figs. 5-10 b).

142

Chapter 5 Overwintering

Table 5-5: Inter-anuual trends of overwintering Drepanosiphum platanoidis populations on Acer pseudoplatanus over 10 and 6 year periods in two locations of the British Isles. The number of aphid eggs deposited per cm2 of bark in mid November represents the initial overwintering egg population; and the following spring peak number of aphids/leaf over n trees represents the maximum aphid abundance for each year (S.R. Leather, unpublished).

Location Year Av. No. Julian Julian date Mean aphids/ Julian Mean aphids/ Julian Starting Eggs/cm Date of bud leaf at spring date of leaf at date of deposited of egg burst peak spring autumn peak autumn mid Nov. hatch (stage 3) population peak population peak NRS Bush, 1982/83 0.89+0.15 3.67+0.10 185 Edinburgh. 1983/84 0.013+0.008 - - 7.17+1.85 211 - - 1984/85 0.74+0.25 - - 8.34+1.32 189 - - 1985/86 0.34+0.16 - - 4.58+0.33 188 - - 1986/87 0.4510.21 - - 12.16+1.72 159 - - 1987/88 0.042+0.005 - - 10.06+4.71 199 - - 1988/89 0.18;0.047 - - 1.64-E1 0.33 180 - - 1989/90 2.82+0.79 - - 20.14+5.41 155 - - 1990/91 1.5510.26 - - 2.90;6.79 168 - - 1991/92 0.18+0.061 - - 7.16+1.35 183 - - Silwood 1992/93 0.038+0.022 78 82 4.09+0.81 175 0.63+0.16 294 Park, 1993/94 0.11-1-0.041 69 82 2.33+0.43 174 0.15+0.045 293 Berkshire. 1994/95 0.015+_0.003- 75 96 6.07+0.87 187 0.5f+0.11 299 1995/96 0.09+0.024 93 103 6.34+0.84 185 0.059+0.020 304 1996/97 0.031+0.010 72 89 2.41+0.45 170 0.026+0.019 275 1997/98 0.006+0.002 78 89 2.44+0.47 183 0.42+0.088 250

a) Bush N.R.S. 3.0 ? 2.5 - /crr s 2.0 -

egg 1.5 -

no. 1.0

Mean 0.5 0.0 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 Time (years)

b) Silwood park 0.12 -

? 0.10 - /cm s

g 0.08

eg 0.06 -

no. 0.04 n a 0.02 Me 0.00

1992 1993 1994 1995 1996 1997 Time (years)

Figs. 5-9 a & b: Egg counts of Drepanosiphum platanoidis on Acer pseudoplatanus in November/December of each year at Bush N.R.S., Edinburgh (Spearman's r, = 0.15, n = 10, P = n.s.) and Silwood Park, Berkshire (Spearman's r, = -0.60, n = 6, P = n.s.).

143 Chapter 5 Overwintering

a) Bush N.R.S.

0 0.25 0.5 0.75 1 1.25 1.5 Ln (mean no. egg/cmz, Nt)

b) Silwood Park 0.12

. ..-, 0.10 o T- c Z 0.08 c cusa) "E 0.06 E O —u)C cy) 0.04_ .__I cs) a) 0.02 1993/94 0.00 0 0.02 0.04 0.06 0.08 0.1 0.12 Ln (meal no. eggs/cm2, Nt)

Figs. 5-10 a & b: Egg counts of Drepanosiphum platanoidis on Acer pseudoplatanus at time Nt and NtTJ at Bush N.R.S., Edinburgh (Spearman's rs = -0.43, n = 9, P = n.s.) and Silwood Park, Berkshire (Spearman's rs = -0.20, n = 5, P = n.s.).

The mechanism behind the hypothesised two-year D. platanoidis cycle, was explored by regressing the number of eggs laid at the beginning of winter against peak aphid density in the following spring. No significant relationship was found between these

two variables at Bush N.R.S. (Regression: F 1,8 = 0.60, r2 = 0.07, P>0.05, Ln y = 0.38Ln(x) + 1.83; Fig. 5-11 a) and Silwood Park (Regression: F 1,4 = 0.006, r2 = 0.001, P>0.05, Ln y = 0.35Ln(x) + 1.52; Fig. 5-11 b).

144

Chapter 5 Overwintering

a) Bush N.R.S. 3.5 _ 3.0 0 o)C' 2.5 _ c -c 4 --- 2.0 (r) c • — 1.5 _ 4 co a) a • 1.0 _ 4 0.5 _ 0.0 0 0.25 0.5 0.75 1 1.25 1.5 Ln (mean eggs/cm )

b) Silwood Park 2.5 _ ca a) 2.0 • c3) • N 1.5 _ to • • 4 73 1.0 a) 0- E 0.5-

0.0

0 0.02 0.04 0.06 0.08 0.1 0.12 Ln (mean eggs/cm2-)

Figs. 5-11 a & b: Between-year dynamics of Drepanosiphum platanoidis on Acer pseudoplalanus using initial number of eggs laid at the beginning of winter and the subsequent spring peak in aphid density at Bush N.R.S., Edinburgh (Regression: F 1.8 = 0.60, r2 = 0.070, P = n.s., Ln y = 0.38Ln(x) + 1.83) and Silwood Park, Berkshire (Regression: F i,4 = 0.0064, r2 = 0.0013, P = n.s., Ln y = 0.35Ln(x) + 1.52).

The timing of bud burst each year did not significantly affect the change in aphid

densities from autumn to spring (Regression: F 1,3 = 7.57, r2 = 0.72, P<0.10, y = 0.044x

- 2.78; Fig. 5-12 a) and D. platanoidis egg hatch success (Regression: F 1,4 = 5.70, r2 = 0.59, P<0.10, y = 0.18x - 11.83; Fig. 5-12 b). Also, the synchrony between timing of

egg hatch and bud burst (stage 3) was not significant (Regression: F 1,4 = 5.05, r2 = 0.56, P<0.10, y = 0.76x + 8.62; Fig. 5-12 c).

145 Chapter 5 Overwintering

HO ] >•,-= c 1.75 _ • c a) -0 -o 1.50 • 1:31E :E Q 0_ al 1Z E 1.00 _ sccs, cn co Q75 _ o)(/) • o _ a) 5 Ws QZ _ Y o Qw Efi 90 96 103 106 Eirlirq (nun stazfi3 Alai the)

Fig. 5-12 a: Effect of bud burst (stage 3 flush) timing on Drepanosiphum platanoidis population change over winter (between autumn and the following spring) at Silwood Park, Berkshire (Regression: F 1,3 = 7.57, r2 = 0.72, P<0.10, y = 0.044x - 2.78).

7 _ d 6 a) E ^ 5 _ co ( 4 (D

cci)15 3 CL -4; (J) (ZS 2 1E' CL 1- Q 0 83 85 90 96 100 106 Elthust (rrEen stay: 3 flush Mai the)

Fig. 5-12 b: Impact of Acer pseudoplatanus bud burst (stage 3) timing on Drepanosiphum platanoidis hatch success (i.e. following spring peak in aphid density) at Silwood Park, Berkshire (Regression: F 1,4 = 5.70, r2 = 0.59, P<0.10, y = 0.18x - 11.83). * depicts two data points.

146

Chapter 5 Overwintering

100

) 80_ day ••

lian 60 Ju ( h 40 tc

ha 20_ Egg 0 80 85 90 95 100 105

Bucburst (Julian day of stage 3 flush)

Fig. 5-12 c: Synchrony between timing of Drepanosiphum platanoidis egg hatch and Acer pseudoplatanus stage 3 bud burst at Silwood Park, Berkshire (F t,4 = 5.05, r2 = 0.56, P<0.10, y = 0.76x + 8.62).

There was a significant and negative correlation between the mean aerial temperature

from 7th to 31St March and bud burst time each year (Regression: F 1,4 = 8.67, r2 = 0.68, P<0.05, y = —3.83x + 117.77; Fig. 5-13), as was temperature on hatching time of D.

platanoidis eggs (Regression: F 1,4 = 13.70, r2 = 0.77, P<0.025, y = —4.16x + 107.51; Fig. 5-13).

120 a) co 0 a) 100- ca A E 80 - o a) o X CU 176 60 -0 -0

40 o

a) _05 20 - 2 "0 _c) 0 4 5 6 7 8 9 10 Mean aerial temperature (°) 7th - 31st March

Fig. 5-13: Effect of temperature on timing of Acer pseudoplatanus bud burst (stage 3 flush) and Drepanosiphum platanoidis fundatrices emergence at Silwood Park, Berkshire. Mean March temperature and the time to stage 3 flush bud burst (A— Regression: F 1,4 = 8.67, r2 = 0.68, P<0.05, y = —3.83x + 117.77). Mean March temperature and the time to D. platanoidis egg hatch (• - - - Regression: F 1,4 = 13.70, r2 = 0.77, P<0.025, y = —4.16x + 107.51).

147 Chapter 5 Overwintering

DISCUSSION

Spatial distribution of eggs on tree bark:

The choice of overwintering site is of great importance in determining the survival of both individuals and populations over the winter months. Optimal overwintering sites may provide a number of benefits including protection from wind (i.e. reduction of wind speed and impeding vertical air movement), low temperatures, rapidly changing temperatures, desiccation, predation and other hazards (Leather et al., 1993).

In this study, oviparae preferred to lay their eggs on trunks and then branches, rather than on twigs and terminal buds. Similarly, Dixon (1976a) observed that most eggs of the sycamore aphid were deposited on branches and on trunks. Field observations of egg abundance indicated that oviparae move from their feeding sites, on leaves in late senescence, down the twigs to the lower branches and trunk region. This downward migration is probably under the control of negative phototaxis and, perhaps, is a geotactic response (Hodson, 1937). A negative phototactic response to black objects has been demonstrated in many insects (e.g. Lepidoptera) prior to overwintering. The openings of crevices may appear as black surfaces against light backgrounds and thus

elicit the negative phototactic response. The moth dubitata flies towards black surfaces in autumn, and this response becomes more intense at low temperatures (thermotaxis) (Tercafs & Thines, 1973). However, oviparae migration may not be a phototactic response, but a response to tactile stimuli (thigmotaxis) when entering deep narrow crevices, as is the case of the cereal thrip Limothrips cerealium (Lewis & Navas, 1962). A combination of negative phototaxis and positive thigmotaxis will tend to direct insects into overwintering sites (e.g. in small hollows, bark crevices, under stones and debris) where they are less conspicious to large predators as well as protected from the vagaries of the winter environment.

Egg laying was examined in relation to compass aspect (cardinal point). Most eggs tended to be laid, although not significantly, on the easterly aspect and the least being found on westerly aspects, with these differences in trends being more evident on branches than on trunks. East facing trunks had heavy but patchy microflora, the shaded northerly aspects had intermediate microflora density, the westerly aspects had rare patchily distributed microflora, whereas the sun exposed southerly aspects were essentially bare. Easterly facing aspects receive moderate amounts of sunshine, which

148 Chapter 5 Overwintering represent relatively warm and humid positions for the overwintering egg populations. North-facing aspects are more susceptible to freezing than those south-facing (Leather et al., 1993). The warmer conditions occurring on south-facing sites are attributed to the effect of insolation (Jensen et al., 1970). The disadvantage of westerly facing aspects is that eggs are subjected to prevailing winds and the associated chill factor which also make overwintering sites prone to desiccation. Twigs and buds showed no apparent trend in whether the oviparae preferred any cardinal point direction.

Larger egg numbers were found on the lower branch surfaces than on the upper branch surfaces. In many cases, the undersides of branches provided shaded and damp micro- climates which ensure sufficient moisture in the spring to allow the egg to undergo post- dormancy development in early spring (Danks, 1987; Tauber et al., 1986; Johnsen et al., 1997). On the other hand, fungal (Entomopthora spp.) infection of oviparae and their eggs (Leather, 1992, 1993) may be favoured in damp areas such as the undersides of branches and northerly trunk aspects. Eggs were not laid on upper south facing branches. Desiccation from full exposure to the sun's rays (Leather et al., 1993), coupled with minimal micro-floral bark covering, may pose a negative selection force upon this sub-optimal oviposition site. It has also been shown for other insects that their distribution and abundance in certain micro-habitats may be due to humidity gradients (e.g. Simpson, & Welborn, 1975). Hygrotaxis has been shown to play an influential role in the ladybirds Coleomegilla maculata and Hippodamia convergens, once the overwintering site has been approximately located (Hodson, 1937; Hagen, 1962). Whereas, bumble bee queens, Bombus terrestris, B. pratorum, B. hypnorum, B. agrorum choose their overwintering sites using both humidity and texture as cues (Pouvreau, 1970).

Survival rate in relation to location on the mature tree and initial egg density:

Survival was density dependent at each monthly interval throughout the winter period on trunks and branches, but was only clearly evident from December through to February in both the winters of 1996 and 1997. Aspect did not exert an important impact on survival rates throughout the winter period on trunks and branches. The lower branch microhabitat (i.e. humid and shaded positions) promoted the survival of sycamore aphid eggs in early to mid winter (November to January). Humidity has been shown to be very important in determining egg survival and hatch rates in aphids (Hand, 1983). Low humidity resulted in a 66% reduction in egg survival of S. avenae (Hand,

149 Chapter 5 Overwintering

1983), and has also been shown to increase mortality rates in A. pisum and A. fabae (Bronson, 1935; Way & Banks, 1964). Low humidity may result in egg desiccation and death, especially in the fortnight prior to egg hatch (Way & Banks, 1964). However, relative humidity in spring rarely falls below 50%, and is thus unlikely to cause a significant amount of egg mortality (Leather, 1993).

During 1996 and 1997, mean D. platanoidis egg mortality over the whole winter period was 79% and 74% on trunks and branches respectively. This adds credence to the body of research showing that overwintering egg mortalities were typically in excess of 50% in a number of other aphid species (Dunn & Wright, 1955; Kulman, 1967; Leather, 1983, 1990b; Kidd & Tozer, 1985). Furthermore, the results for D. platanoidis showed that the longer the winter lasts the greater the mortality. This relationship applies to many other species of aphids; for example, R. padi, and Tuberculoides annulatus (Leather et al., 1993). This study found that mortality of overwintering eggs was greater on tree trunks than on lower branch sections, despite the heavily creviced trunk bark, and that mortality was greatest towards the end of winter (January to February) on both trunks and branches. This pattern of egg disappearance confirmed other studies showing egg mortality to be greater at the end than during winter (Leather, 1981; Gange & Llewellyn, 1988). However, this study did not show a heavy initial loss followed by a regular decline in population size at a constant rate through the main winter period (Leather, 1983, 1990a; Hand & Hand, 1986). This difference could be due to monthly sampling rather than weekly sampling technique as used by Leather (1983, 1990a). Nevertheless, mortality of the egg stage in some aphid species has been largely attributed to natural enemies (Leather, 1992, 1993). In early winter, egg mortality in A. fabae was mainly due to Anthocoris nemorum (Way & Banks, 1964), whereas mid-winter mortality was attributed to birds, aphidophagous insects and fungal attack (Leather, 1993). Birds accounted for about 15% of total egg loss and insect predators for about 3% in A. fabae (Way & Banks, 1964). In R. padi, caused 31% of the mortality and birds 15% (Leather, 1981). This is comparable with the alder aphid Pterocallis alni where arthropods caused 50% mortality and birds 10% (Gange & Llewellyn, 1988). By late winter, birds and arthropods are the major mortality factors for aphid eggs (Leather, 1992).

Under the low oviparae density treatment, egg survival was inversely density dependent. Leather (1993) points out that under natural oviparae densities, a heavy initial loss followed by a regular decline in egg population size would be common with

150 Chapter 5 Overwintering aphids that lay their eggs in limited sites, and that high initial egg loss does not occur in years when aphid densities are low. On the other hand, the high oviparae treatment showed density dependent survival. This opposing egg survival phenomenon occurring at low and high oviparae densities could be explained by the low density treatment being synonymous with natural field oviparae levels, but under unnaturally high oviparae density conditions a switch-over mechanism may have come into play. It could reasonably be proposed that under extremely high egg densities, this initial high rate of egg loss is not so obviously seen in D. platanoidis, similar to Metopolophium dirhodum not showing a high initial rate of egg loss on stems (Leather, 1993). Also, oviparae may switch to less favourable oviposition substrate if the preferred sites have been saturated with eggs (cf. Rothschild & Schoonhoven, 1977), so affecting overwintering egg survival rates through the winter.

Competition among oviparae for optimal oviposition sites:

Fewer eggs per ovipara were laid under high density conditions than low density. This reduction in realised fecundity was probably a result of maternal jostling or dislodgement from the experimental arena. Kennedy & Crawley (1967) demonstrated that gentle tactile simulation, crudely resembling that received by one aphid from another, caused settled aphids on leaves to move away but also caused walking aphids to stop and probe; whereas strong stimulation of the same kind can cause settled and walking aphids to move away, but they then settle nearby. While D. platanoidis oviparae may appear gregarious on trunk bark in the field, they are actually repelling one another in order to deposit their overwintering eggs in prime sites. Therefore, in a similar manner to R. padi on Prunus padus (Leather. 1990a), this study illustrates the existence of strong intra-specific competition for the best oviposition sites within a micro-habitat.

Substrate selection and Oviposition Patterns of individual oviparae:

The substrate selection experiment clearly demonstrated that oviparae laid significantly more eggs on rough bark than on smoother bark. Thus, the relatively large number of eggs laid on trunks and the basal parts of a branch was probabily associated with the roughness of the older bark. Folds in the bark at the point where a branch meets the trunk could be particularly favoured sites for oviposition, whereas the moderately rough to smooth surfaces on branches and twigs supporting a low egg density would leave

151 Chapter 5 Overwintering them fully exposed to weather and natural enemies. This crevice selection behaviour also occurs in other overwintering insects. For example, the blackfly Simulium pictipes spends the winter as eggs laid in crevices in rocks at the edges of streams (Kurtak, 1974); and the ladybird, Hippodamia quinquesignata, aggregates in its overwintering hibernation site under rocks and debris on west-facing slopes (Harper & Lilly, 1982).

Individual oviparae lay their eggs in aggregated groups, when provided with a plentiful supply of rough bark. Although many studies have been made on egg clustering in insects (e.g. Godfray, 1994), no studies on egg aggregation in aphids. However, useful parallels may be drawn from egg clustering in butterflies. For adult females, search time for oviposition sites and deposition patterns may favour egg clustering (Stamp, 1980). Females which search for and respond negatively to the presence of eggs, of their own and other species, may require more search time for oviposition sites when the adult population is high due to the increased likelihood of encountering eggs. Alternatively, if the oviposition substrate is patchily distributed or in short supply, clustering may reduce the time locating them and increase realised fecundity (Courtney, 1984). In addition, it may be beneficial to deposit eggs in clusters to avoid parasitism and predation of the adult female whilst ovipositing, or assessing egg load, if the detection distance of natural enemies is large (Emlen, 1973). For eggs, the egg clustering habit may decrease the amount of egg surface exposed to ambient conditions and reduce the possibility of desiccation, which is particularly important for exposed overwintering eggs. It may also decrease the number of eggs left vulnerable to natural enemy attack. For instance, clusters of eggs may have lower rates of parasitism than single eggs if the reproductive capacity of the female parasitoid is low per unit time (Stamp, 1980).

Temporal abundance of aphids over several years:

Consistently more eggs were deposited in 1996 than 1997 in each spatial zone of the tree, but this was only clearly shown with the branch data. This observation fits with the independent study on D. platanoidis egg numbers on mature trees at Silwood Park (Table 5-5). Moreover, from studying the temporal dynamics of egg populations in Bush N.R.S. and Silwood Park, the two-year cycle, as proposed by Dixon (1970b, 1990c), was not clearly shown because there was no density dependent spiralling with a time-lag of one year and no strong alternating pattern over the years in the 'egg density- time plots'.

152 Chapter 5 Overwintering

The mechanism for alternating yearly abundance was not explained by relating egg counts at the beginning of winter and the peak aphid population in the following spring. This weak relationship in between-year population dynamics of D. platanoidis was not expected since other studies on holocyclic aphids have shown a direct relationship. For example, the number of eggs present on P. padus were very closely correlated with the peak population of Rhopalosiphum padi on P. padus in the spring (Leather, 1983), and subsequently with the populations developing on cereals during the summer (Kurppa, 1989). Perhaps it would have been better to use autumn and spring aphid peak densities to demonstrate this relationship (cf. Dixon, 1970b). Dixon (1970b) suggested that a linear relationship would come from high numbers of oviparae, giving rise to an abundance of overwintering eggs, which lead to high fundatrix emergence in the following spring since there would be no migration occurring during this period. The discrepancy between using egg counts and autumn aphid peaks may be due to sampling effort and unitary search. Firstly, eggs may be more difficult to find in crevices than sexuals on open leaves when sampling in the field, and therefore represent the major caveat of this study. Secondly, aphids on leaves are more uniformly distributed than eggs on bark, thereby making autumnal aphid assessment more robust. Egg clustering may lead to erroneous results, due to mis-counting on creviced bark, than aphids on leaves in autumn. Thirdly, the eggs per unit area are not measured equivalently as autumn and spring aphid numbers.

Bud burst, aphid egg hatch success rate and temperature:

Although bud burst timing was positively correlated with the extent of population change (k-value) occurring over the winter and the aphid peak density in spring, these relationships were not significant. The synchronisation between the timing of bud burst and egg hatch was also not significant. However, the positive association between bud burst and fundatrix emergence corresponds with the positive relationship between bud burst and aphid spring peak density. This could indicate that timing of bud burst was important for the success of aphid populations in spring. If there was a large time difference between egg hatch and flushing of buds, the temporal exposure to mortality factors such as desiccation (Way & Banks, 1968), birds and rain (Dixon, 1976a, 1987) would be lengthened, and fewer fundatrices would survive (Dixon, 1976a; Chambers, 1979).

153 Chapter 5 Overwintering

This study consistently showed that egg hatch occurred a few days before bud burst, thus agreeing with Dixon's (1976a) findings. The advantage of aphids hatching at this point is that feeding on the highly nutritious unfurling leaves allows fundatrices to double their adult weight, and produce more numerous and better quality offspring (Wellings et al., 1980) which mature more rapidly and achieve greater body size, than fundatrices emerging after bud burst (Dixon, 1976a).

This study showed that warmer March temperatures significantly induced earlier bud burst and fundatrix emergence time, and confirms Dixon's (1987) suggestion that the rate of development of aphid eggs and tree buds are both temperature-dependent. The implications of egg hatch and bud burst jointly occurring earlier, will mean that more time will be available for aphid population growth in spring before the onset of aestivation. Walters (1987) showed, using a series of field experiments conducted over a period of 6 years, that the temperatures experienced in March play an important role in determining the size of the populations of M persicae present in spring. Further more, the results of this study revealed that aerial temperatures in the last three weeks of March was more crucial than photoperiodic conditions during this time, since bud burst and egg hatching times would have happened at the same time. Although light intensity and duration are important for sycamore flushing, the thermal cue is the critical factor since March temperature causes variation in bud burst from year to year.

CONCLUSION

The spatial variations in tree micro-habitat and temporal variability over years exert a considerable influence on the population dynamics of D. platanoidis. As the ultimate aim is survival over the winter period, it is not surprising that the sites selected are specific and limited. Multitactic cues are likely to direct the movement of newly mated oviparae from feeding sites, on senescing leaves, to their oviposition sites, further down the tree. In evolutionary terms, a two-fold trade-off may have existed between the distance travelled by oviparae (i.e. from twig extremities down the branches/trunks) to: (i) the reproductive fitness (the number of eggs deposited), and (ii) weather exposure and/or predation risk to the advantage of hatching closer to buds. Egg clustering on creviced bark may have evolved in response to selection for increased fecundity, while allowing the ovipara a chance to respond to a scarcity of ovipositional sites by changing cluster size. However, intra-specific competition among oviparae for optimal oviposition sites may not only reduce realised fecundity at higher densities, but may in

154 Chapter 5 Overwintering the process leave many immobile eggs exposed. Overwintering egg mortality of D. platanoidis cannot be regarded as a regulatory factor, the length of the winter can markedly affect the number of eggs surviving and thus the initial success of the aphid population in the spring. From observing D. platanoidis temporal dynamics, no two- year cycle could be detected from the two short-teim time-series conducted. Using initial winter egg counts and spring aphid peak density alone did not provide an adequate explanation of the between-year dynamics. Finally, ambient temperatures during March were influential on the timing of aphid egg hatch and bud burst, which indicates that this density-independent factor was the principal driving force in determining the peak size of sycamore aphid populations in spring.

Perhaps future field observations could make a detailed record of egg predators throughout the whole winter period and calculate their relative quantitative impacts on overwintering sycamore aphid egg populations. Dixon (1976a) has provided a starting point in this direction by monitoring D. platanoidis fundatrices on buds from egg hatch through to bud burst in order to measure the effect of bird predation. In addition, it may be worthwhile testing in the laboratory whether searching time for oviposition sites and deposition patterns are affected by temperature or ovipara age/size/number of mature eggs (cf. Gossard & Jones, 1977). Finally, by determining the initial winter egg density (Way & Banks, 1968; Jones & Dunning, 1972), choice of oviposition sites (Cammell et al., 1978), survival prospect of eggs, fundatrix emergence and bud burst date, together with climate variables such as temperature (Greenbank, 1970; Bommarco & Ekbom, 1995) and rainfall (Way et al., 1992), these key parameter estimates could form a forecasting model to predict potential outbreaks in serious forest and agricultural pests several months in advance (Way & Cammell, 1973; Way et al., 1981, 1992; Leather, 1992, 1993; Leather et al., 1993; Bommarco & Ekbom, 1995).

155 Chapter 6 General Discussion

GENERAL DISCUSSION:

APHID POPULATION DYNAMICS AND HOST PLANT QUALITY

It is known that delayed negative feedback can cause phytophagous forest insect populations to exhibit regular density cycles (Berryman et al., 1987). Negative feedback processes occur when one or several factors increase their impact on the reproduction and survival of individuals in response to increases in population density. Multi-trophic interactions between herbivores and their host plants, predators, parasitoids and pathogens can all potentially be involved in negative feedback regulation. The extent of these interactions are often determined by density dependent processes, as mentioned in Table 1.1. In addition, density-independent factors (e.g. edaphic and climatic variables) can disturb the density-dependent processes regulating the population dynamics of herbivorous insects. However, there is not always a clear distinction between density-dependent and density—independent factors. For example, background phytochemicals may affect herbivores in a density-independent manner, while this insect-plant interaction becomes density-dependent if the release of plant defence substances is determined by herbivore pressure (Berryman et al., 1987).

The overall goal of the thesis was to examine the mechanisms driving the two-year cycle of the sycamore aphid, D. platanoidis. The within-year abundance of this aphid is characterised by an inverse relationship operating between spring and autumn, and this has been thought to be responsible for driving the cycling between the years (Dixon, 1970b). This contrast in intra-annual seasonal density was originally considered to be determined by aphid induced changes in sycamore leaves resulting in a delayed overcompensated negative feedback effect which stabilises aphid numbers (Dixon, 1970b). However, Chambers (1982) rejected this 'plant factor hypothesis' in favour of maternal crowding and aestivation duration as the important factors determining the within-year aphid dynamics. This was in agreement with Dixon (1979) stating that the patterns of changes in the populations were attributed to the action of intra-specific processes (e.g. mortality, migration, reduction in fecundity and reproductive aestivation). Later, Chambers et al. (1985) showed that the effect of crowding was mediated through the host plant, and not directly through behavioural interactions of the

156 Chapter 6 General Discussion

aphids. However, Wellings & Dixon (1987) found that leaf infestation had no direct effect on aphid life history traits. More recently, Dixon et al. (1993) combined the two hypotheses by claiming that both host plant food quality and intra-specific competition were involved in driving the seasonal abundance of aphids.

Chapter 2 re-examined the 'plant factor hypothesis' stated by Dixon (1970b). Through manipulating aphid density on sycamore saplings in the early part of the year, the clip cage results clearly indicated the presence of herbivore-induced alterations in plant quality, since the effects were manifested in the three parthenogenetic generations — i.e. those of spring, summer and autumn. Aphids caged on highly infested saplings in spring initially performed well by showing rapid nymphal development, increased reproductive duration and output, short pre-parturition and a long post-larviposition period. However, by autumn aphid performance on heavily infested saplings was poor in relation to those that were previously uninfested or lightly infested. This response was likely to be due to the aphids in spring affecting the nitrogen metabolism of the foliar tissue so that little nitrogen is translocated from the leaves before abscission, and the food available to aphids is then poor. This results in few and small aphids with a low reproductive rate later in the year (Dixon, 1970a, 1975a). Alternatively, delayed induced resistance (Haukioja, 1990b), brought about by damage to the plant due to an earlier generation, may be responsible for the poor aphid performance in autumn on previously heavily attacked plants. Either way, high aphid density in spring can, through their effect on their host plant, adversely affect the number, size and reproductive rate of the aphids present in the autumn (i.e. two or three generations later).

Chapter 2 also found that there was a contrast in the density relationship between spring and autumn in the natural abundance of aphids in 1996 and 1997 (i.e. aphid abundance in autumn 1996 was higher than that found earlier in that year, whilst aphid abundance in autumn 1997 was lower than in spring that year). This implies that a regulatory mechanism was operating on autumnal aphid numbers in opposing ways between these two years, which may contribute to an alternating yearly cycle (Dixon, 1990c). The changes in autumnal fecundity from one year to the next could be due to differences in weight (quality) of the aphid that are in turn influenced by host nutritional status (Dixon & Wratten, 1971; Leather & Wellings, 1981; Dixon et al., 1982). Thus, if host nutritional quality is degraded later in the year, aphid size will be reduced, so leading to lower herbivore abundance.

157 Chapter 6 General Discussion

Chapter 2 showed that plant stressing (via manipulating aphid load) not only increases aphid abundance, but also enhanced the numerical responses of adult natural enemies and their larvae, which confirms the findings of Neuenschwander et al. (1975), Frazer et al. (1981a) and Honek (1983). Although the predator and parasitoid response was positively correlated to sycamore aphid density in spring, only the specialist parasitoids and territorial spiders showed close numerical relationships to autumnal aphid abundance. Despite this positive association, natural enemies are unlikely to represent a major regulatory force of aphid dynamics. This is because they have slower rates of development relative to that of their prey (Mills, 1982), they are inefficient at capturing prey (e.g. Dixon, 1970c), there is asynchrony between predator and prey life-cycles (Honek, 1983), and there is satiation at moderate to high densities and/or switching to alternative hosts when aphid density is low (Anderson, 1962; Dixon & Russel, 1972). Nevertheless, natural enemies could have a marked influence on the rate of increase of the field aphid population by devouring colossal numbers of young aphids in autumn when aphid fecundity is low (Dixon & Russel, 1972; Hamilton, 1973, 1974). Hence, the action of aphidophages could be important in determining the degree of overcompensation that is mainly dictated by a plant-mediated process.

In addition to herbivore-induced changes in plant chemistry, fungal infection of plant tissue may directly or indirectly affect trophic interactions not only involving host plants, but the herbivores themselves and their natural enemies (Hammon & Faeth, 1992; Hunter & Price, 1992; Preszler et al., 1996). As discussed in chapter 3, post- aestivation aphid dynamics generally remained unaffected by the presence of the endophytic fungus, R. acerinum. That is, there was no early release from aestivation on endophyte infected sycamores, no outright numerical preference for endophyte infected saplings, and no enhanced reproductive rate in the second and third aphid generations, as would be demonstrated by a modified aphid abundance. This overall neutral effect complies with the findings of Faeth & Hammon (1996, 1997 a,b), but contrasts with the work of Gange (1996). In addition, the degree of parasitism was not influenced by the presence of the endophyte. However, of all the age and morphologic aphid categories, only oviparae density was significantly reduced on endophytic plant material. The reason for this could be due to the action of mycotoxins or fungal-induced plant defences (see Clay, 1990; Saikkonen et al., 1996). The biological implications of suppressed sexual female density means that less overwintering eggs will be deposited

158 Chapter 6 General Discussion prior to leaf fall. This in turn will result in a lower fundatrix abundance hatching from the eggs in the following spring. Furthermore, premature senescence and leaf fall is known to be caused by heavy infection of sycamore leaves by R. acerinum (Leith & Fowler, 1987). This may exaggerate the lower oviparae density found on endophyte occupied plant material. This overall neutral finding does not comply with the generality of mutualisms between endophytic fungi and their host plants, as indicated by an antagonistic relationship with phytophagous insects (e.g. Carroll, 1991; Butin, 1992), but instead adds to the variety of effects already observed in other arboreal investigations (Gaylord et al., 1996; Preszler et al., 1996; Lappalainen & Helander, 1997).

In autumn, the change from parthenogenetic to sexual reproduction in holocyclic aphids takes place in response to changes in their abiotic and biotic environment. The onset and determination of sexual morph type is known to be triggered by short day-lengths and low temperatures (Lees, 1966; Dixon, 1971a; Lamb & Pointing, 1975). However, sexual morphs cannot be produced until a certain number of generations have lapsed — the lacteur fondatrice' (Bonnemaison, 1951). Photo- and thermo-periodism may not only affect the aphid directly, via neuroendocrine processes (Hardie, 1987), but may also act on the host plant by inducing senescence (Sethi & Swenson, 1967; Dixon, 1971b). In chapter 4, it was found that the reproductive sequence of the final parthenogenetic generation (sexuparae) produced a predominance of females over males irrespective of plant nutritional status (cf. Foster & Benton, 1992). With a female- biased sex ratio, reproductive success could be maximised via the daughters (see Taylor, 1981; Newton & Dixon, 1988). The progeny sequence of sexuparae was characterised by an irreversible switch from oviparae to males, following a brief reproductive pause, which was also found by Lamb & Pointing (1975). This order of morph production meant the oviparae have time to complete their development, and be ready to mate, by the time the males appear later in the reproductive sequence (cf. Strathdee et al., 1993). This is also facilitated by the pause in reproductive activity during the switch-over period. Although, the commencement of the reproductive sequence was not affected by the host plant quality, the nymphal development of sexuals was significantly faster on hosts of higher nutritional value. Nymphal survival was not sex related, but was enhanced by good host plant quality. Although contrasting with the finding of Dixon & Kundu (1997), adult oviparae lived longer than males. This implies that sexual females have more time for oviposition of overwintering eggs.

159 Chapter 6 General Discussion

Chapter 4 also found that good host plant nutrition increased oviparae body weight, which in turn was positively linked to potential fecundity of oviparae (i.e. number of eggs/ovariole). This agrees with the work of others showing that large individuals are more fecund than small individuals (Dixon, 1970a; Taylor, 1975). The proportion of immature eggs per ovipara was lower on good hosts, which meant that more eggs had a chance to mature when more food resources were available to the mother. However, the number of ovarioles per ovipara remained constant and independent of host plant quality. This supports the finding by Leather & Wellings (1981) showing that sycamore aphid ovariole number was the result of an intrinsic factor. This pre-determined reproductive strategy is generation specific and is therefore programmed to anticipate the seasonal host plant quality (Wellings et al., 1980; Leather & Wellings, 1981). If the aphids with many ovarioles were to feed on nutritionally poor hosts, they would be less likely to survive to maturity than those with fewer ovarioles, as a result of maternal starvation (Ward et al., 1983 a,b).

The life-cycle of the sycamore aphid is brought to a close by the mating of the sexual morphs in late autumn prior to leaf fall. Before the oviparae die at the end of the season, they deposit their eggs on crevices of tree bark (Dixon, 1976a). Chapter 5 found that D. platanoidis eggs were predominantly laid on the rougher bark surfaces of trunks and branches of mature sycamore, away from the smoother terminal buds. Observing the abundance of eggs in relation to tree architecture indicates that oviparae undergo a downward movement away from the senescing leaves, and this is likely to be under the control of a range of cues such as humidity and tactile stimuli. The careful selection of overwintering sites by oviparae was found to increase the chances of egg survival into the following spring, through offering a spatial refuge from the extremes of weather and predators. Overwintering egg mortality was shown to be a function of the length of time in hibernation (cf. Leather, 1990b). Greatest D. platanoidis egg mortality occurred during late winter, which agrees with the findings of Leather (1983) and Thornback (1983).

Chapter 5 also found that hibernating eggs were laid in clusters. This communal response confers advantages both to the ovipara and egg. The oviparous mother could benefit from reduced site searching and oviposition duration. It also increased the realised fecundity of oviparae by laying eggs in quick succession, whilst reducing the pre-reproductive time exposure to mortality due to predators and hostile weather

160 Chapter 6 General Discussion conditions. On the other hand, aggregation reduces egg surface exposure to desiccation and attack from natural enemies (Stamp, 1980). In addition, chapter 5 showed the presence of fierce intra-specific competition among oviparae for optimal ovipostion sites. This was indicated by a reduction in the number of eggs deposited per ovipara under higher density compared to lower density conditions. In contrast to high oviparae abundance, mortality of eggs over the winter was density-dependent at low maternal abundance.

In chapter 5, a two-year cycle trend was observed in the time-series of direct aphid counts on trees in the north-east (Edinburgh) and south (Berkshire) of the British Isles, although this was not statistically significant. This tends to concur with the regular alternating pattern found by Dixon (1990c) and Dixon & Kindlmann (1998) working on the D. platanoidis and Myzocallis boerneri. The relationship between two successive years, in terms of D. platanoidis abundance on sycamore in autumn and the following spring (chapter 2), was positively correlated and agrees with findings of Dixon (1970b). This implies that an autumn with a high density of aphids (i.e. parthenogenetic virginoparae and sexual morphs) produces an abundance of overwintering eggs, which gives rise to many fundatrices hatching in the next spring (Dixon, 1976a; Leather et al., 1993).

Unlike R. padi, D. platanoidis egg hatch was not found to be so closely synchronised with bud burst (Dixon, 1976a; Wellings, 1980; Leather, 1980). In chapter 5, fundatrix emergence always occurred before the stage three (Leather, 1996a) bud burst. This temporal asynchrony permits the newly emerged sycamore aphid to capitalise on the highly nutritious food resources as the host bursts into leaf (Dixon & Wellings, 1982), while allowing the second generation of D. platanoidis to develop on leaves that are still growing (Dixon & Wellings, 1982). Although early hatching time runs the risk of increased mortality from weather and predatory effects, they are compensated for this by being larger and more fecund than the strictly synchronised aphids (Chambers, 1979; Leather & Wellings, 1981). Those that hatch later, although escaping rain and wind exposure, become 'poor quality' aphids (Dixon, 1976a; Wellings, 1980). Chapter 5 showed that temperature determined the time of bud burst in sycamores and egg hatching in D. platanoidis, thus confirming the temperature—dependent suggestion by Dixon (1987). Since temperature in March influences the time of fundatrix emergence, it can therefore play a crucial role in determining the size of the spring aphid

161 Chapter 6 General Discussion populations. Temperature, like food quality and competitor density, are major components influencing growth and development rates, as well as adult aphid body weight (Chambers, 1979; Wellings, 1981). This, together with an intrinsically high reproductive potential of fundratrices (high oviariole numbers — Leather & Wellings, 1981; Dixon & Dharma, 1980), facilitates a rapid increase of the population in spring (Wellings, 1980).

The results from this thesis and previous research (e.g. Dixon, 1979,1987; Dixon & Wellings, 1982) suggest that seasonality and weather are the major driving forces in the determining the within- and between-year aphid dynamics on sycamore. These density- independent variables could be viewed as an ultimate top-down control mechanism, since they can act directly on both the plant quality and aphid itself. Despite these over-riding abiotic forces, the plant-aphid association can be highly interactive, thus allowing a degree of plasticity within the system.

Season is known to influence plant quality, and in turn dictate the abundance of aphids during each year (Dixon, 1970a). Tree foliage is a rich source of food for aphids during bud burst and early leaf development in the spring, and leaf senescence in the autumn; however, during the intervening summer months, the mature foliage is less favourable for aphid growth and development (Dixon & Welling, 1982; Dixon, 1987; Dixon et al., 1993). Seasonality can also directly affect the aphid by causing aestivation in summer when temperatures are high and food quality is low. It can also induce the production of sexual morphs in autumn when photo- and thermal conditions gradually reduce, and foliage start to senesce (Dixon, 1987). Associated with the predictable seasonal trend in habitat quality, each aphid generation has developed reproductive traits in anticipation of the conditions it is most likely to experience, so that it is well adapted (Dixon, 1977; Wellings et al., 1980; Dixon & Wellings, 1982; Leather & Wellings, 1981). Thus, by virtue of short generation time and programmed adaptation to seasonal trends, aphids track changes in habitat quality very closely (Dixon, 1985).

Meterological conditions can also have great influence on the plants and aphids. Both the timing of sycamore aphid egg hatch and bud burst are dependent on temperature, with the eggs hatching and buds bursting earlier in warm than cool springs (Dixon, 1987). Although the time of leaf fall, and thus the end of the season of aphid reproduction, depends mainly on falling day-length (Ward et al., 1984), temperature may also have an effect (Vince Prue, 1975). Temperature is also an important cue for

162 Chapter 6 General Discussion the production of aphid sexual morphs in autumn (e.g. Lees, 1959, 1960; MacGillivray & Anderson, 1964; Blackman, 1975). In addition to phytophagous insect growth and development rates increasing with rising temperature, predators and parasitoids become more effective (Gilbert & Raworth, 1996). Wind and rain can also result in substantial mortality through the dislodgement of aphids, particularly in early spring and late autumn (Dixon, 1976a; 1979). During winter, high air humidity can increase egg mortality when compared to that of drier areas (Kurppa, 1991).

Aphids and the host plant can affect each other throughout the season. Aphids may overcome the constraints of poor host nutrition by aggregating to form an effective `physiological sink', and if the drain for assimilates is sufficiently strong and localised, the plant reacts to it in some respects as if it were a bud (Hill, 1962; Way & Cammell, 1970; Miles, 1989). Dixon (1971a) has shown that if the drain on plant resources imposed by sycamore aphids is substantial enough it can lead to serious plant stress, and in some cases result in tree death (Miles, 1989). Leather (1998) suggested that trees could respond to previous heavy aphid infestation by inducing a change in the sexual orientation of the tree (i.e. from protogynous to protandrous trees), or accelerating bud burst (Leather, 1996b), in order to avoid high levels of herbivory in future. Alternatively, the host plant may respond to aphid feeding through the production of defensive plant chemicals (Haukioja, 1990b; Baldwin, 1994), in response to the influence of salivary components secreted during stylet withdrawal (Miles, 1989). In sum, aphids are not the passive organisms that they are generally portrayed to be, they respond actively to their changing environment.

163 REFERENCES

Addicott, F. T. (1982) Absicission. pp. 369: Univ. Calif. Press, Berkeley, C.A. Aitkin, M., Anderson, D., Francis, B. & Hinde, J. (1989) Statistical modelling in GLIM. Oxford University Press, Oxford. Alstad, D. N., Edmunds, G. F. J. & Weinstein, L. H. (1982) Air pollution and insect populations., Ann. Rev. Ent. 27, 369-384. Amman, G. D. (1984) Mountine pine beetle (Coleoptera: Scolytidae) mortality in three types of infestations. Environ. Entomol. 13, 184-191. Anderson, N. H. (1962) Bionomics of six species of Anthocorids (Heteroptera: Anthocoridae) in England., Trans. R. ent. Soc. Lond. 114, 67-95. Anderson, R. M. & May, R. M. (1978) Regulation and stability of host, parasite population interactions: I. Regulatory processes. J Anim. Ecol. 47, 219-247. Anderson, R. M. & May, R. M. (1980) Infectious diseases and population cycles of forest insects. Science 210, 658-661. Anderson, R. M., Turner, B. D. & Taylor, L. R. E. (1979) Population dynamics. the 20th Symposium of the British Ecological Society, London 1978. Blackwell Scientific, Oxford. Andrewartha, H. G. & Birch, L. C. (1954) The distribution and abundance of . University of Chicago Press, Chicago. Applebaum, S. W., Tadmor, U. & Podoler, H. (1970) The effect of starch and of heteropolysaccharide fraction from Phaseolus vulagaris on development and fecundity of Collosobruchus chinensis (Coleoptera: Bruchidae). Entomol Exp. AppL 13, 61-70. Austin, A. B. M., Tatchell, G. M., Harrington, R. & Bale, J. S. (1996) Adaptive significance of changes in morph production during the transition from parthenogenetic to sexual reproduction in the aphid Rhopalosiphum padi (Homoptera: Aphididae). Bulletin of Entomological Research 86, 93-99. Bakker, K. (1964) Backgrounds of controversies about population theories and their terminologies. Z. Angew. Ent. 53, 187-208. Baldwin, I. T. (1994) Chemical changes rapidly induced by folivory. In Insect-Plant Interactions. pp. 1- 23. (Ed. E. Bernays) CRC Press, Inc., Florida. Baltensweiler, W. (1984) The role of environment and reproduction in the population dynamics of the large bud moth, Zeiraphera diniana Gn. (Lepidoptera: Tortricidae). Adv. Invert. Reprod. 3, 291- 302. Baltensweiler, W., Benz, G., Bovey, P. & Delucchi, V. (1977) Dynamics of large bud moth populations. Ann. Rev. Entomol. 22, 79-100. Barbosa, P. & Letourneau, D. K. (1988) Novel aspects of insect-plant interactions. New York: Wiley. Barbour, D. A. (1990) Synchronous fluctuations in spatially separated populations of cyclic forest insects. In Population dynamics of forest insects. pp. 339-346. (Eds. A. D. Watt, S. R. Leather, M. D. Hunter & N. A. C. Kidd). Intercept, Andover. Barker, A. M., Wratten, S. D. & Edwards, P. J. (1995) Wound-Induced Changes in Tomato Leaves and Their Effects On the Feeding Patterns of Larval Lepidoptera. Oecologia, 101, 251-257. Barker, G. M. & Addison, P. J. (1996) Influence of clavicipitaceous endophyte infection in ryegrass on development of the parasitoid Microctonus hyperodae Loan (Hymenoptera: Braconidae) in Listronotus bonariensis (Kuschel) (Coleoptera: Curculiuonidae). Biol. Control, 7, 281-287. Barker, G. M. & Addison, P. J. (1997) Clavicipitaceous endophytic infection in ryegrass influences attack rate of the parasitoid Microctonus hyperodae (Hymenoptera: Braconidae, Euphorinae) in Listronotus bonariensis (Coleoptera: Curculionidae). Enviro. Entomol. 26, 416-420. Barlow, N. D. & Dixon, A. F. G. (1980) Simulation of Lime Aphid Population Dynamics. Simulation Monographs. pp. 171. Pudoc, Wageningen. Batzli, G. 0. (1996) Population cycles revisited. Trends in Ecology & , 11, 488-489. Bauer, G. (1985) Population ecology of Parida tripunctana Schiff and Nototcelia roborana Den. and Schiff. (Lepidoptera: Tortricidae) - an example of "equilibrium species". Oecologia, 65, 437- 441 Begon, M., Harper, J. L. & Townsend, C. R. (1990) Ecology: Individuals, Populations and Communities.: Blackwell Scientific Publications. Boston, Oxford, London, Edinburgh, Melborne. Begon, M., Mortimer, M. & Thompson, D. J. (1996) Population Ecology: A Unified Study of Animals and Plants. pp. 247. Blackwell Science Ltd., Oxford. Benkevich, V. I. (1972) Gypsy moth outbreaks in the European part of the USSR as related to solar activity, fluctuations, atmospheric circulation, climatic and weather conditions. Proc. 13th Int. Congr. Entomol., Moscow, 1968. 3, 14-15.

164 Benz, G. (1987) Environment. In Epizootiology of insect diseases. (Eds. J. R. Fuxa & Y. Tanada), pp. 177-214. New York, John Wiley. Benz, R. (1974) Naegative Ruckkopplung durch Raum and Nahrungskonkurenz sowie zyklische Veraenderung der Nahrungsgrundlage als Regelprinzip in der Populationsdynamik des Grauen Lerchenwicklers, Zeiraphera diniana (Guenee) (Lep., Tortricidae). Z ang. Ent. 76, 196-228. Berryman, A. A. (1973) Population dynamics of the fir engraver, Scolytus ventralis (Coleoptera: Scolytidae). I. Analysis of population behaviour and survival from 1964 to 1971. Can. Entomol. 105, 1465-1488. Berrryman, A. A. (1978a) Towards a theory of insect epidemiology. Res. Popul. Ecol. 19, 181-196. Berryman, A. A. (1978b) Population dynamics of the Douglas-fir tussock moth (Lepidoptera: Lymantriidae): the time delay hypothesis. Can. Entomol 110, 181-196. Berryman, A. A. (1979) Dynamics of bark beetle populations: Analysis of dispersal and redistribution. Mitt. Schweiz. Entomol. Ges. 52, 227-234. Berryman, A. A. (1981) Population Systems: A General Introduction. Plenum Press, New York. Berryman, A. A. (1982) Population dynamics of bark beetles. In Bark beetles of North American conifers. A system for the study of evolutionary ecology. (Eds. J. B. Mitton & K. B. Sturgeon), pp. 264- 314. Austin: Univ. Texas Press. Berryman, A. A. (1996) What causes population cycles of forest Lepidoptera? Trends in Ecology & Evolution, 11, 28-32. Berryman, A. A. & Stark, R. W. (1985) Assessing the risk of forest insect outbreaks. Z Angew. Entomol. 99, 199-208. Berryman, A. A., Stenseth, N. C. & Isaev, A. S. (1987) Natural regulation of herbivorous forest insect populations. Oecologia 71, 174-184. Berryman, A. A. E. (1988) Dynamics of Forest Population Insects: Patterns, Causes and Implications. Plenum Press, New York and London: Bevan, R. J. & Greenhalgh, G. N. (1976) Rhytisma acerinum as a biological indicator of pollution. Envrion. Pollut. 10, 271-285. Blackman, R. (1971) Variation in the phtoperiodic response within natural population of Myzus persicae (Sulz.). Bull. Ent, Res. 60, 533-546. Blackman, R. (1972) The inheritance of life-cycle differences in Myzus persicae (Sulz.) (Hem., Aphididae). Bull. Ent. Res. 62, 281-294. Blackman, R. (1974) Invertebrate types: Aphids. pp. 175: Ginn & Co. Ltd., London & Aylesbury. Blackman, R. L. (1975) Photoperiodic determination of the male and female sexual morphs of Myzus persicae. J. Insect Phys. 21, 435-453. Blackman, R. L. (1987) Reproduction, cytogenetics and development. In World Crop Pests: Aphids - Their Biology, Natural Enemies and Control, Vol. 2A, pp. 163-195. (Eds. A. K. Minks & P. Harrewijn). Elsevier, Amsterdam. Blackman, R. L. & Eastop, V. F. (1994) Aphids on the World's Trees: An Identification and Information Guide. Natural History Museum, CAB International. Blackman, R. L. & Hales, D. F. (1986) Behaviour of the X-chromosome during growth and maturation of parthenogenetic eggs of Amphorophora tuberculata (Homoptera: Aphididae), in relation to sex determination. Chromosoma 194, 59-64. Blais, J. R. (1983) Trends in the frequency, extent and severity of spruce budworm outbreaks in eastern Canada (Choristoneura fumiferama). Can. J. For. Res. 13, 539-547. Boenisch, A., Petersen, G. & Wyss, U. (1997) Influence of the hyperparasitoid Dendrocerus carpenteri on the reproduction of the grain aphid Sitobion avenae. Ecol. Entomol., 22, 1-6. Bommarco, R. & Ekbom, B. (1995) Phenology and prediction of pea aphid infestations on peas. Mt. J. Pest Manag. 41, 109-113. Bonnemaison, L. (1951) Contribution a l'etude des facteurs provoquant l'apparition des fromes ailees et sexuees chez les Aphidinae. Ann. Epiphytes (C). 2, 1-380. Boonstra, R., Krebs, C. J. & Stenseth, N. C. (1998) Population cycles in small mammals: the problem of explaining the low phase. Ecology, 79, 1479-1488. Breen, J. P. (1994) Acremonium endophyte intractions with enhanced plant resistance to insects. Ann. Rev, Entomol. 39, 401-423. Broadbeck, B. & Strong, D. (1987) Amino acid nutrition of herbivorous insects and stress to host plants. In Insect Outbreaks pp. 347-363. (Eds. P. Barbosa & J. C. Schultz). Academic Press Inc. Bronson, T. E. (1935) Observations on winter survival of pea aphid eggs. J. Econ. Ent. 28, 1030-1036. Brookes, M. H., Stark, R. W. & Campbell, R. W. (1978) The Douglas-fir tussock moth: a synthesis. U.S.D.A. For. Serv, Tech Bull. 1585 . Brooks, T. F. (1953) Plant diseases. Cambridge Univ. Press, Cambridge. Brown, A. (1953) Forestry commission Bulletin No. 20 - Studies on British beechwoods. pp 1-48: HMSO, London.

165 Bultman, T. L., Borowicz, K. L., Schneble, R. M., Coudron, T. A. & Bush, L. P. (1997) Effect of a fungal endophyte on the growth and survival of two Euplectrus parasitoids. Oikos, 78, 170-176. Bush, L. P., Fannin, F. F., Segal, M. R., Dahlman, D. L. & Burton, H. R. (1993) Chemistry, occurrence, and biological effects of saturated pyrrolizidine alkaloids associated with endophyte-grass interactions. Aggr. Ecosyst. Environ. 44, 81-102. Butler, E. J. & Jones, S.G. (1961) Plant pathology. MacMillan, London. Butin, H. (1992) Effect of endophytic fungi from oak (Quercus-Robur L.) on mortality of leaf inhabiting gall insects. Eur. J For. Pathol. 22, 237-246. Cammell, N. E., Way, M. J. & Heathcote, G. D. (1978) Distribution of eggs of the black bean aphid, Aphis fabae, Scop., on the spindle brush, Euonymus europeus L., with reference to forcasting infestations of the aphid on field beans. Plant Pathol. 27, 68-76. Campbell, B. C. & Duffy, S. S. (1979) Tomatine and parasitic wasps: potential incompatibility of plant antibiosis with biological control. Science 205, 700-702. Campbell, I. M. (1964) Genetic variation related to survival in Lepidopteran species. In Breeding pest- resistant trees. pp. 129-135. (Eds. H. D. Gerhold, R. E. McDermott, E. J. Schreiner & J. A. Winieski). Pergamon Press, Oxford. Campbell, R. W. (1981) Evidence for high fecundity amongst certain North American gypsy moth populations. Environ. Entomol. 10, 663-667. Campbell, R. W. & Sloan, R. J. (1978) Numerical bimodality among north american gypsy moth populations. Environ. Entomol. 7, 641-646. Cappuccino, N. & Price, P. W. E. (1995) Population Dynamics: New Approaches and Synthesis. Accademic Press, San Diego. Carroll, G. C. (1986) The biology of endophytism in plants with partcular reference to woody perennials. In Microbiology of the Phyllosphere. pp. 205-222. (Eds. N. J. Fokkema & J. van den Heuvel). Cambridge Univ. Press., Cambridge. Carroll, G. C. (1988) Fungal endophytes in stems and leaves: from latent pathogen to mutualistic symbiont. Ecology 69, 2-9. Carroll, G. C. (1991) Fungal associates of woody plants as insect antagonists in leaves and stems. In Microbial mediation of plant-herbivore interactions. pp. 253-271. (Eds. P. Barbosa, V. A. Krischik & C. G. Jones). Wiley, New York. Carruthers, R. I., Bergstrom, G. C. & Haynes, P. A. (1986) Accelerated development of the European corn borer, Ostrinia nubilalis (Lepidoptera: Pyralidae), induced by interactions with Colletotrichum graminicola (Melanconiales: Melanconiaciaceae), the causal fungus of maize anthracnose. Ann. Entomol. Soc. Am. 79, 385-389. Carlson, K. A., Nusbaum, T. J., Rose, M. R. & Harshman, L.G. (1998) Oocyte maturation and ovariole number in lines of Drosophilia melanogaster selected for postponed senescence. Funct. Ecol., 12, 514-520. Chambers, R. J. (1979) Simulation modelling of a sycamore aphid population. UnpubL Ph.D. Thesis, Univ. E. Anglia, Norwich. Chambers, R. J. (1982) Maternal experience of crowding and duration of aestivation in the sycamore aphid. Oikos, 39, 100-102. Chambers, R. J., Wellings, P. W. & Dixon, A. F. G. (1985) Sycamore aphid numbers and population density: II. Some Processes. J. Anim. Ecol. 54, 425-442. Charnov, E. L. (1982) The theory of sex allocation. Monogr. Pop. Biol. Princeton Univ. Press, Princeton: Cheplick, G. P. & Clay, K. (1988) Acquired chemical defenses in grasses: the role of fungal endophytes. Oikos 52, 309-318. Chitty, C. D. (1967) The natural selection of self-regulatoy beavior in animal populations. Proc. Ecol. Soc. Austr. 2, 51-78. Chugunin, Y. (1949) Focal periodicy of gypsy moth outbreaks. Zool. Zh. 28, 431-438. Clarke, A. B. (1978) Sex ratio and local resource competition in a prosimian primate. Science 201, 163, Clay, K. (1988) Fubgal endophytes of grasses: a defensive mutualism between plants and fungi. Ecology 69, 10-16. Clay, K. (1990) Comparative dempgraohy of three graminoids infected by systemic, clavicipitaceous fungi. Ecology 71, 558-570. Clay, K. (1991) Endophytes as antagonists of plant pests. In Microbial ecology of leaves. pp. 331-357. (Eds. J. H. Andrews & S. S. Hirano). Springer, New York. Coleman, J. S. & Jones, C. G. (1991) A phytocentric perspective of phytochemical induction by herbivores. In Phytochemical induction by herbivores. pp. 3-45. (Eds. D. W. Tallamy & M. J. Raupp). Wiley, New York. Collins, M. D., Ward, S. A. & Dixon, A. F. G. (1981) Handling time and the functional response of Aphelinus thomsoni, a predator and parasite of the aphid Drepanosiphum platanoidis. J. Anim. Ecol. 50, 479-487.

166 Courtney, S. P. (1984) The evolution of egg clustering by butterflies and other insects. Am. Nat. 123, 276- 281. Craig, T. P., Price, P. W. & Itami, J. K. (1986) Resource regulation by a stem-galling sawfly on the arroyo willow. Ecology 67, 419-425. Crawley, M. J. (1983) Herbivory: the Dynamics of Animal-Plant Interactions. Blackwell Scientific Publications, Oxford. Crawley, M. J. (1993) Methods in Ecology - GLIM for Ecologists. pp. 1-379: Blackwell Scientific Publications, Oxford. Crawley, M. J. & Akhteruzzaman, M. (1988) Individual variation in the phenology of oak trees and its consequences for herbivorous insects. Funct. Ecol. 2, 409-415. Crema, R. (1979) Egg viability and sex determination in Megoura viciae (Homoptera: Aphididae). Ent. Exp. Appl. 26, 152-156. Dahlman, D. L., Eichenseer, H. & Siegel, M. R. (1991) Chemical perspectives of endophyte-gras interactions and their implications to insect herbivory. In Microbial mediation of plant-herbivore interactions. pp. 227-252. (Eds. P. Borbosa, V. A. Krischick & C. L. Jones). Wiley, New York. Dahlsten, D. L., Luck, R. F., Schlinger, E. I., Wenz, J. M. & Copper, W. A. (1977) Parasitoids and presators of the Douglas fir tussock moth, Orgyia pseudotsugta (Lepidoptra: Lymantriidae), in low to moderate populations in central California. Can. Entomol. 109, 727-746. Danks, H. V. (1987) Insect dormancy: an ecological perspective. Biological Survey of Canada Monograph series No. 1. Darvill, A. P. & Albersheim, P. (1984) Phytoallexins and their elicitors: a defense against microbial infection in plants. Ann. Rev. Plant Physol. 35, 243-275. Davidson, A. W. & Potter, D. A. (1995) Response of plant feeding, predatory, and soil inhabiting invertebrates to Acremonium endophyte and nitrogen fertilization in tall fescue turf. J. Econ. Entomol. 88, 367-379. Day, K. (1984) The growth and decline of a population of the spruce aphid, Elatobium abietinum, during a three year study, and the changing pattern of fecundity, recruitment and alary polymorphism in a northern island forest. Oecologia 64, 118-124. Dean, G. J. (1974) Effects of parasites and predators on the cereal aphids Metopolophium dirhodum (Wlk.) and Macrosiphum avenae (F.) (Hem., Aphididae). Bull. Ent. Res. 63, 411-422. DeLucchi, V. (1982) Parasitoids and hyperparasitoids of Zeiraphera diniana (Lep: Tortricidae) and their role in population control in outbreak areas. Entomophaga 27, 77-92. DeMoraes, C. M. & Mescher, M. C. (1999) Interactions in entomology: Plant-parasitoid interaction in tritrophic systems. J Entomol. Sci. 34, 31-39. Dempster, J. P. (1983) The natural control of populations of butterflies and . Biol. Rev. 58, 461-481. Dempster, J. P. & McLean, 1. F. G. E. (1998) Insect populations; in theory and in practice. 19th Symposium of the Royal Entomological Society 10-11 September 1997 at the University of Newcastle. Kluwer Academic Publishers, Dordrecht/Boston/London. Den Boer, P. J. & Gradwell, G. R. (1971) Dynamics of populations. Proceedings of the Advanced Study Institute on dynamics of numbers in populations. Pudoc, Wageningen. Dicke, M. (1988) Microbial allochemicals affecting the behavior of insects, , nematodes, and protozoa in different trophic levels. In Novel aspects of insect-plant interactions. pp. 125-163. (Eds. P. Barbosa & D. K. Letourneau). Wiley, New York. Dixon, A. F. G. (1958) The escape responses shown by certain aphids to the presence of the coccinellid Adalia decempunctata (L.). Trans. R. Ent. Soc. Lond. 110, 319-334. Dixon, A. F. G. (1963) Reproductive activity of the sycamore aphid, Drepanosiphum platanoidis (Schr.) (, Aphididae). J. Anim. Ecol. 32, 33-48. Dixon, A. F. G. (1966) The effect of population density and nutritive status of the host on the summer reproductive activity of the sycamore aphid, Drepanosiphum platanoidis (Schr.). J. Anim. Ecol. 35, 105-112. Dixon, A. F. G. (1969) Population dynamics of the sycamore aphid Drepanosiphum platanoidis (Schr.) (Hemiptera: Aphididae) migratory and trivial flight activity." Anim. Ecol. 38, 585-606. Dixon, A. F. G. (1970a) Quality and availability of food for a sycamore aphid population. In Animal Populations in Relation to their Food Resources. pp 271-286. (Ed. A. Watson). Blackwell Scientific Publications, Oxford. Dixon, A. F. G. (1970b) Stabilization of aphid population by an aphid induced plant factor. Nature, 227, 1368-1369. Dixon, A. F. G. (1970c) Factors limiting the effectiveness of the coccinellid beetle, (L.), as a predator of the sycamore aphid, Drepanosiphum plantanoidis (Schr.). J. Anim. Ecol., 39, 739-751. Dixon, A. F. G. (1971a) The role of aphids in wood formation: I. The effect of the sycamore aphid, Drepanosiphum platanoidis (Schr.) (Aphididae), on the growth of sycamore, Acer pseudoplatanus (L.). J Appl. Ecol. 8, 165-179.

167 Dixon, A. F. G. (1971b) The 'interval timer' and photoperiod in determination of parthenogenetic and sexual morphs in the aphid, Drepanosiphum plantanoidis. J Insect Physiol. 17, 251-260. Dixon, A. F. G. (1971c) The role of intra-specific mechanisms and predation in regulating the numbers of the lime aphid, Eucallipterus tiliae L. Oecologia 8, 179-193. Dixon, A. F. G. (1972) The 'interval timer', photoperiod, and temperature in the seasonal development of parthenogenetic and sexual morphs in the lime aphid Eucallipterus tiliae L. Oecologia, 9, 301- 310. Dixon, A. F. G. (1973a) Metabolic acclimatization to seasonal changes in temperature in the sycamore aphid, Drepanosiphum plantanoidis (Schr.), and lime aphid, Eucallipterus tiliae L. Oecologia. 13, 205-210. Dixon, A. F. G. (1973c) Biology of Aphids. In The Institute of Biology's Studies in Biology no. 44. Edward Arnold Ltd., London. Dixon, A. F. G. (1975a) Effect of population density and food quality on autumnal reproductive activity in the sycamore aphid, Drepanosiphum plantanoidis (Schr.). J. Anim. Ecol. 44, 297-304. Dixon, A. F. G. (1975b) Aphids and Translocation. In Encyclopedia of Plant Physiology. New Series. Vol. I. pp. 154-170. (M. H. Zimmermann & J. A. Milburn). Springer-Verlag, Berlin. Dixon, A. F. G. (1976a) Timing of egg hatch and viability of the sycamore aphid, Drepanosiphum platanoidis (Schr.), at bud burst of sycamore, Acer pseudoplatanus L. J. Anim. Ecol. 45, 593- 603. Dixon, A. F. G. (1976b) Factors determining the distribution of sycamore aphids on sycamore leaves during summer. Ecol. Entomol. 1, 275-278. Dixon, A. F. G. (1977) Aphid ecology: life cycles, polymorphism and population regulation. Ann. Rev. Ecol. Syst. 8, 329-353. Dixon, A. F. G. (1979) Sycamore aphid numbers: the role of weather, host and aphid. In Population Dynamics. The twentieth symposium of the British Ecological Society, London. pp. 105-121 (Eds. R. M. Anderson, B. D. Turner, & L. R. Taylor). Blackwell Scientific Publications, Oxford. Dixon, A. F. G. (1985) Aphid Ecology. Blackie & Son, Glasgow, London. Dixon, A. F. G. (1987) Seasonal development in aphids. In World Crop Pests: Aphids - Their Biology, Natural Enemies and Control. Vol. 2A (Eds. A. K. Minks & P. Harrewijn) pp. 315-320. Elservier. Dixon, A. F. G. (1990) Population dynamics and abundance of deciduous tree-dwelling aphids. In Population Dynamics of Forest Insects. pp. 11-23. (Eds. Watt, A.D., Leather, S.R., Hunter, M.D. & Kidd, N.A). Intercept Ltd. Dixon, A. F. G., Chambers, R. J. & Dharma, T. R. (1982) Factors affecting size in aphids with particular reference to the black bean aphid, Aphis fabae. Ent. Exp. Appl. 32, 123-128. Dixon, A. F. G. & Dharma, T. D. (1980) 'Spreading of the risk' in developmental mortality: size, fecundity and reproductive rate in the black bean aphid. Ent. Exp. Appl. 28, 301-312. Dixon, A. F. G. & Glen, D. M. (1971) Morph determination in the bird cherry-oat aphid, Rhopalosiphum padi L. Ann. Appl. Biol. 68, 11-21. Dixon, A. F. G. & Kindlmann, P. (1990) Role of plant abundance in determining the abundance of herbivorous insects. Oecologia, 83, 281-283. Dixon, A. F. G. & Kindlmann, P. (1998) Population dynamics of aphids. Insect Populations in theory and practice. Kluwer Academic, Dordrecht. Dixon, A. F. G. & Kundu, R. (1997) Trade-off between reproduction and length of adult life in males and mating females of aphids. Eur. J. Entomol. 94, 105-109. Dixon, A. F. G. & McKay, S. (1970) Aggregation in the sycamore aphid Drepanosiphum platanoidis (Schr.) (Hemiptera: Aphididae) and its relevance to the regulation of population growth. J. Anim. Ecol. 39, 439-454. Dixon, A. F. G. & Russel, R. J. (1972) The effectiveness of Anthocoris nemorum and A. confusus (Hemiptera : Anthocoridae) as predators of the sycamore aphid, Drepanosiphum plantanoidis. II. Searching behaviour and the incidence of predation in the field. Ent. Exp. Appl. 15, 35-50. Dixon, A. F. G. & Wellings, P. W. (1982) Seasonality and reproduction in aphids. Int. J. Invert. Reprod. 5, 83-89. Dixon, A. F. G., Wellings, P. W., Carter, C. & Nichols, J. F. A. (1993) The role of food quality and competition in shaping the seasonal cycle in the reproductive activity of the sycamore aphid. In Oecologia, 95, 89-92. Dixon, A. F. G. & Wratten, S. D. (1971) Laboratory studies on aggregation, size and fecundity in the black bean aphid, Aphis fabae Scop. Bull. Ent. Res. 61, 97-111. Dixon, T. J. (1959) Studies on oviposition behaviour of Syrphidae (Diptera). Trans. R. Ent. Soc. Lond. 3, 57-81. Docherty, M., Wade, F. A., Hurst, D. K., Whittaker, J. B. & Lea, P. J. (1997) Responses of tree sap- feeding herbivores to elevated CO2. Global Change Biology, 3, 51-59.

168 Douglas, A. E. (1993) The nutritional quality of phloem sap utilized by natural aphid populations. Ecol. Entomol. 18, 31-38. Douglas, A. E. & Prosser, W. A. (1992) Synthesis of the essential amino acid tryptophan in the pea aphid symbiosis. J. Insect Physiol. 38, 565-568. Du, Y. J., Poppy, G. M. & Powell, W. (1996) Relative importance of semiochemicals from first and second trophic levels in host foraging behavior of Aphidius ervi. J. Chem. Ecol. 22, 1591-1605. Dunn, J. A. & Wright, D. W. (1955) Overwintering egg populations of the pea-aphid in East Anglia. Bull. Ent. Res. 46, 389-392. Eastop, V. F. (1955) Selection of aphid species by different kinds of insect traps. Nature, 176, 936. Edwards, P. J. (1989) Insect herbivory and plant defence theory. In Toward a more exact ecology. pp. 275-297. (Eds. P. J. Grubb & J. B. Whittaker). Blackwell Scientific Publications, Oxford. Edwards, P. J., Wratten, S. D. & Gibberd, R. (1990) The impact of inducible phytochemicals on fooed selection by insect herbivores and its consequences for the distribution of grazing damage. In: Phytochemical Induction by Herbivores (Eds. M. J. Raupp & D. Tallamy). John Wiley, New York. Edwards, P. J., Wratten, S. D. & Parker, E. A. (1992) The ecological significance of rapid wound-induced changes in plants-insect grazing and plant competition. Oecologia, 91, 266-272. Eichenseer, H. & Dahlman, D. L. (1992) Antibiotic and deterrent qualities of endophyte-infected tall fescue to 2 aphid species (Homoptera: Aphidiae). Environ. Entomol. 21, 1046-1051. Elliott, W. M. (1973) A method for predicting short-term population trends of the green peach aphid, Myzus persicae (Homoptera: Aphididae) on potatoes. Can. Ent. 105, 11-20. Elzen, G. W., Wiliams, H. J. & Vinson, S. B. (1983) Response by the parasitoid Campoletis sonorensis (Hymenoptera: Ichneumonidae) to chemicals (synomones) in plants: Implications for host habitat location. Environ. Entomol. 12, 1873-1877. Emlen, J. (1973) Ecology: An Evolutionary Approach. Addison Wesley, Reading, Mass. Evert, R. F., Eschrich, W., Medler, J. T. & Alfieri, F. J. (1968) observations on penetration of linden branches by stylets of the aphid caryae. Am. J. Bot. 55, 860-874. Faeth, S. H. (1994) Induced plant reponses: effect on parasitoids and other natural enemies of phytophagous insects. In Parasitoid Communities, pp. 245-260. (Eds. B. A. Hawkins & W. Sheehan). Oxford University Press, Oxford, U.K. Faeth, S. H. & Hammon, K. E. (1996) Fungal endophytes and phytochemistry of oak foliage: Determinants of oviposition preference of leafminers? Oecologia, 108, 728-736. Faeth, S. H. & Wilson, D. (1996) Induced responses in plants: mediators of interactions among macro- and micro-herbivores? In Britsh ecological society symposium, pp. 201-215. (Ed. A. C. Gange). Blackwell, Oxford. Faeth, S. H. & Hammon, K. E. (1997a) Fungal endophytes in oak trees: Long-term patterns of abundance and associations with leafminers. Ecology, 78, 810-819. Faeth, S. H. & Hammon, K. E. (1997b) Fungal endophytes in oak trees: Experimental analyses of interactions with leafminers. Ecology, 78, 820-827. Faeth, S. H., Connor, E. F. & Simberloff, D. S. (1981) Early leaf abscission: a neglected source of mortality for folivores. Am. Nat. 117, 409-15. Feeny, P. (1976) Plant apparency and chemical defence. Recent Adv. Physiochem. 10, 1-40. Ferguson, K. I. & Stiling, P. (1996) Non-additive effects of multiple natural enemies on aphid populations. Oecologia 108, 375-379. Fisher, R. A. (1930) The genetical theory of natural selection. Oxford Univ. Press, Oxford. Fishlin, A. & Baltensweiler, W. (1979) Systems analysis of the large bud moth system. Part I: The large bud moth relationship. Mitt. Schweiz. Entomol. Ges. 52, 273-289. Forrest, J. M. S. (1970) The effect of maternal and larval experience on morph determination in Dysaphis devecta. J. Insect Physiol. 16, 2281-2292. Foster, W. A. & Benton, T. G. (1992) Sex ratio, local mate competition and mating behaviour in the aphid Pemiphigus spyrothecae. Behay. Ecol. Sociobiol. 30, 297-307. Fowler, S. V. & Lawton, J. H. (1985) Rapidly induced defenses and talking trees: The devil's advocate position. Am. Nat. 126, 181-195. Frazer, B. D. & Gilbert, N. (1976) Coccinellids and aphids: A quantitative of the impact of adult ladybirds (Coleoptera: ) preying on field populations of pea aphids (Homoptera: Aphididae). J. Ent. Soc. Brit. Columbia, 73, 33-56. Frazer, B. D., Gilbert, P. M., Ives, P. M. & Raworth, D. A. (1981a) Predator reproduction and the overall predator-prey relationship. Can. Ent. 113, 1015-1024. Frazer, B. D., Gilbert, N., Nealis, V. & Raworth , D. A. (1981b) Control of aphid density by a complex of predators. Can. Ent. 113, 1035-1041. Fretwell, S. D. (1972) Populations in a seasonal environment. Princeton Univ. Press, Princeton:

169 Furuta, K. (1976) Studies on the dynamics of the low density populations of gypsy moth and Todo-fir aphid; Analysis of the environmental resistance factors by artificial host increase method. Bull. Gov. For. Exp. Stn. 279, 85. Furuta, K. (1982) Natural control of Lymantria dispar L. (Lepidoptera, Lymantriidae) populations at low density levels in Hokkaido (Japan). Z Angew. Entomol. 93, 513-522. Furuta, K. (1987) Amounts of favourable feeding materials in spring for the maple aphid, Periphyllus californiensis Shinji, estimated from the phenological relations between the aphid and host trees. J. Appl. Ent. 104, 144-157. Furuta, K., Hashimoto, H. & Iwamoto, N. (1984) The effect of budding and flowering of maple trees on the development of the maple aphid, Periphyllus californiensis Shinji (Homoptera, Aphididae) population. Z. Agnew Entomol. 98, 437-444. Furuta, K. & Sakamoto, N. (1984) Seasonal fluctuation of population density of the maple aphid (Periphyllus californiensis Shinji; Horn., Aphididae). Bull. Tokyo Univ. 73, 97-113. Gabriel, K. R. (1978) A simple method of multiple comparisons of means. J. Amer. Stat. Assn. 73, 724- 729. Gange, A. C. (1996) Positive effects of endophyte infection on sycamore aphids. Oikos, 75, 500-510. Gange, A. C. & Llewellyn, M. (1988) Egg distribution and mortality in the alder aphid, Pterocallis Ent. Exp. Appl. 48, 9-14. Gates, D. M. (1980) Biophysical ecology. Springer. New York: Gaylord, E. S., Preszler, R. W. & Boecklen, W. J. (1996) Interactions between host plants, endophytic fungi, and a phytophagous insect in an oak (Quercus grisea x Q. gambelii) hybrid zone. In Oecologia, 105, 336-342. Gibberd, R., Edwards, P. J. & Wratten, S. D. (1988) Wound-Induced Changes in the Acceptability of Tree-Foliage to Lepidoptera - Within-Leaf Effects. Oikos, 51, 43-47. Gilbert, N. (1984) Control of fecundity in Pieris rapae. I. The problem. J. Anim. Ecol. 53, 581-588. Gilbert, N. & Raworth, D. A. (1996) Insects and temperature - a general theory. Can. Entomol. 128, 1-13. Ginzburg, L. R. & Taneyhill, D. E. (1994) Population cycles of forest lepidoptera - a maternal effect hypothesis. J. Anim, Ecol. 63, 79-92. Godfray, H. C. J. (1994) Parasitoids: Behavioural and Evolutionary Ecology. pp. 473. (Eds. J. R. Krebs & T. Clitton-Brock). Princeton University Press, Princeton, N.J. Gora, V., Konig, J. & Lunderstadt, J. (1994) Physiological defence reactions of young beech trees (Fagus sylvatica) to attack by . Forest Ecology and Management. 70, 245-254: Gossard, T. & Jones, R. (1977) The effects of age and weather on egg laying in Pieris rapae L. J. Appl. Ecol. 14, 65-71. Graham, S. A. (1939) Forest insect populations. Ecol. Monogr. 9, 301-310. Greenbank, D. L. (1963) Climate and the spruce budworm. Mem. Entomol. Soc. Can. 31, 174-180. Greenbank, D. 0. (1970) Climate and the ecology of the balsam woolly aphid. Can. Entomol. 102, 546- 578. Griffiths, E. & Wratten, S. D. (1979) Intra- and inter-specific differences in cereal aphid low-temperature tolerance. Ent. Exp. Appl. 26, 161-167. Grobler, J. H., MacLeod, D. M. & De Lyzer, A. J. (1962) The fungus Empusa aphidis Hoffman parasitic onthe woolly pine needle aphid Schizolachnus piniradiatae (Davidson). Can. Ent. 94, 46-49. Guldemond, J. A. & Tigges, W. T. (1992) Production of sexuals and sex ratios in Cryptomyzus species in relation to dispersal and host alternation (Homoptera: Aphidinae: Aphididae). Entomol. Gener. 16, 257-264. Hagen, K. S. (1962) Biology and ecology of predacious cocchinellidae. Ann. Rev. Entomol, 7, 289-326. Hain, F. P. E. (1979) Population dynamics offorest insects at low levels. North Carolina State Univ. Hales, D. F. & Mittler, T. E. (1983) Precocene causes male determination in the aphid Myzus persicae. J. Insect Physiol. 29, 819-823. Hales, D. F. & Mittler, T. E. (1987) Chromosomal sex determination in aphids controlled by juvenile hormone. Genome 29, 107-109. Hales, D. F., Tomiuk, J., Wohrmann, K. & Sunnucks, P. (1997) Evolutionary and genetic aspects of aphid biology: a review. Eur. I Entomol. 94, 1-55. Hales, D. F., Wellings, P. W. & Parkes, R. A. (1989) Investment in gynoparae and males by Myzus persicae (Sulzer). Funct. Ecol. 3, 727-734. Hamilton, P. A. (1969) The role of hymenopterous parasites in the control of the sycamore aphid. Unpubl. Ph.D. Thesis, University of Glasgow. Hamilton, P. A. (1973) The biology of Aphelinus flavus (Hym.: Aphelinidae), a parasite of the sycamore aphid Drepanosiphum platanoidis (Hemipt. Aphididae). Entomophaga 18, 449-462. Hamilton, P. A. (1974) The biology of Monoctonus pseudoplatani, Trioxys cirsii and Dyscritulus planiceps, with notes on their effectiveness as parasites of the sycamore aphid, Drepanosiphum platanoidis. Ann. Soc. Ent. Fr. 10, 821-840. Hamilton, W. D. (1967) Extraordinary sex ratios. Science 156, 477-488.

170 Hammon, K. E. & Faeth, S. H. (1992) Ecology of plant-herbivore communities: a fungal component. Nat. Toxins 1, 197-208. Hand, S. C. (1983) The effect of temperature and humidity on the duration of development and hatching success of eggs of the aphid Sitobion avenae. Ent. Exp. AppL 33, 220-222. Hand, S. C. & Wratten, S. D. (1985) Production of sexual morphs by the monoecious cereal aphid Sitobion avenae. Ent. Exp. AppL 38, 239-248. Hand, S. C. & Hand, L. (1986) Monitoring of the winter populations of cereal aphids near Wageningen, The Netherlands in 1982-1983. Neth. J. Plant Pathol. 92, 137-146. Hanski, I. & Otronen, M. (1985) Food quality induced variance in larval performance: Comparison between rare and common pine feeding sawflies (Diprionidae). Oikos 44, 43-51. Hardie, J. (1987) Neurosecretory and endocrine systems. In Aphids: Their Biology, Natural Enemies and Control Vol. 2A. (Eds. A. K. Minks & P. Harrewijn). Elsevier, Amsterdam. Hardie, J. (1989) Spectral specificity for targeted flight in the black bean aphid, Aphis fabae. J. Insect Physiol. 35, 619-626. Hardie, J., Baker, F. C., Jamieson, A. D., Lees, A. D. & Schooley, D. A. (1985) The identification of an aphid juvenile hormone and its titre to relation to photoperiod. Physiol. Entomol. 10, 297-302. Hardy, Y.V., Lafond, A., & Hamel, L. (1983) The epidemiology of the current spruce budworm outbreak in Quebec. For. Sci. 29, 715-725. Harper, A. M. & Lilly, C. E. (1982) Aggregations and winter survival in southern Alberta of Hippodamia quinquesignata (Coleoptera: Cocchinellidae), a predator of the pear aphid (Homoptera: Aphididae). Can. Entomol. 114, 303-309. Hassell, M. P. (1976) The dynamics of competition and predation. In The Institute of Biology's Studies in Biology no. 72. pp.1-68. Edward Arnold (Publishers) Ltd., London. Hassell, M. P. (1978) The dynamics of predator-prey systems. pp. 237. Princeton University Press, Princeton, N.J. Hassell, M. P. (1982) Patterns of parasitism by insect parasitoids in patchy environments. Ecol. Entomol. 7, 365-377. Hassell, M. P., Lawton, J. H. & May, R. M. (1976) Patterns of dynamical behavior in single species populations. .1 Anim. Ecol. 45, 471-486. Haukioja, E. (1980) On the role of plant defences in the fluctuation of herbivore populations. Oikos 35, 202-213. Haukioja, E. (1982) Inducible defences of white birch to a geometrid defoliator, Epirrita autumnata. In Insect plant relationships. pp.199-203. (Eds. J. H. Visser & A. K. Minks). PUDOC, Wageningen. Haukioja, E. (1990a) Positive and negative feedbacks in insect-plant interactions. In Population Dynamics of Forest Insects. pp. 113-122. (Eds. A. D. Watt, S. R. Leather, M. D. Hunter & N. A. Kidd). Haukioja, E. (1990b) Induction of defenses in trees. Annu. Rev. Entomol. 36, 25-42. Haukioja, E. (1991) The influence of grazing on the evolution, morphology and physiology of plants as modular organisms. Phil. Trans. R. Soc. Lond. B. 333, 241-247. Haukioja, E. & Honkanen, T. (1997) Herbivore-induced responses in trees: internal vs. external explanations, In Forests and Insects. pp. 69-80. (Eds. A. D. Watt, N. E. Stork & M. D. Hunter). Chapman & Hall, London. Haukioja, E. & Neuvonen, S. (1987) Insect population dynamics and induction of plant resistance: the testing of hypotheses. In Insect Outbreaks. pp. 411-432. (Eds. P. Barbosa & J. Schult). Academic Press, New York. Haukioja, E., Neuvonen, S., Niemeld, P. & Hanhimaki, S. (1988) The autumnal moth Epirrita autumnata in Fennoscandia. In Dynamics of Forest Insect Populations. pp. 167-178. (Ed. A. A. Berryman): Plenum Press, New York. Haukioja, E., Ruohomaki, K. & Senn, J. E. A. (1990) Consequences of herbivory in the mountain birch (Betula pubescence sp. tortuosa): importance of the functional organisation of the tree. Oecologia 82, 238-247. Heikkenen, H. G. 1981 The influence of red pine site quality on damage by the European pine shoot moth. In Hazard dash rating systems in forest insect pest management. pp. 35-44. (Ed. R. Hedden). U.S.D.A., For. Serv. Gen. Tech. Rept. WO-27. Heinz, K. M. (1996) Host size selection and sex allocation behaviour among parasitoid trophic levels. Ecol. Entomol. 21, 218-226. Hendry, G. A. F. & Price, A. H. (1993) Stress indicators: chlorophylls and carotenoids. In Methods of Comparative Plant Ecology: A Laboratory Manual. pp. 148-152. (Eds. G. A. F. Hendry & P. J. Grime). Chapman & Hall, London. Hill, G. P. (1962) Exudation from aphid stylets during the period of dormancy tp bubreak in Tilia americana.(L.). J. Exp. Bot. 13, 144-151. Hill Ris Lambers, D. (1966) Polymorphism in Aphididae. Ann. Rev. Entomol. 11, 47-78.

171 Hille Ris Lambers, D. (1960) Some notes on morph determination in aphids. Entomol. Ber. 20, 110-113. Hodek, I. (1967) Bionomics and ecology of predaceous Coccinellidae. A. Rev. Ent. 12, 79-104. Hodson, A. C. (1937) Some aspects of the role of water in insect hibernation. Ecol. Monogr. 7, 271-315. Hofsvang, T. (1990) Advantages and disadvantages of parasitoids and predators of aphids when used in biological control. Acta Entomol. Bohemoslov. 87, 401-413 Holler, C. (1990) Overwintering and hymenopterous parasitism in autumn of the cereal aphid Sitobion avenae (F.) in northern FR Germany. J. AppL Ent. 109, 21-28. Holliday, P. (1982) A dictionary of plant pathology. Cambridge Univ. Press, Cambridge. Holling, C. S. (1959) Some characteristics of simple types of predation and parasitism. Can. Entomol. 91, 385-398. Holling, C. S. (1965) The functional response of predators to prey density and its role in mimicry and population regulation. Mem. Entomol. Soc. Can. 45, 1-60. Holling, C. S. (1973) Resilience and stability of ecological systems. Ann. Rev. Ecol. Syst. 4, 56-63 Honek, A. (1983) Factors affecting the distribution of larvae of aphid predators (Col., Coccinellidae and Dipt., Syrphyidae) in cereal stands. Z. ang, Ent. 95, 336-345. Hosteller, D. L. & Bell, M. R. (1985) Natural dispersal of bacculoviruses in the environment. In Viral insecticides for biological control. pp. 249-284. (Eds. K. Maramorosh & K. E. Sherman). Academic Press, New York. Hough, J. A. & Pimentel, D. 1980 Influence of host foilage on development, survival and fecundity of the gypsy moth. Environ. Entomol. 7, 97-102. Hudler, G. W., Banik, M. T. & Miller, S. G. (1987) Unusual epidemic of tar spot on Norway maple in upstate New York. Plant Disease, 71, 65-68. Huffaker, C. B. (1958) The concept of balance in nature. Proc. X Int. Cong. Entomol. 2, 625-636. Hughes, R. D. (1960) Induction of diapause in Erioischia brassicae Bouche (Dipt. Anthomyiidae). .1 Exp. Biol. 37, 218-223. Hughes, R. D. (1963) Population dynamics of the cabbage aphid, Brevicoryne brassicae (L.). J. Anim. Ecol. 32, 393-424. Hunter, M. D. (1990) Differential susceptibility to variable plant phenology and its role in competition between two insect herbivores on oak. Ecol. Entomol. 15, 401-408. Hunter, M. D. (1992) A variable insect-plant interaction: the relationship between tree budburst phenology and population levels of insect herbivores among trees. Ecol. Entomol. 16, 91-95. Hunter, M. D. & Price, P. W. (1992) Playing chutes and ladders: heterogeneity and the relative roles of bottom-up and top-down forces in natural communities. Ecology 73, 724-732. Il'inskii, A. I., Evlakhova, M. I., Shvetsova, S., Shvetsova, 0. I. & Andreeva, G. I. (1975) Inspection, calculation, and prognois afforest insects in the USSR. Transi. U.S.D.A. For. Serv. 1980. Inozmetsev, A. A. (1976) The dynamics of the trophic links of forest insectivorous birds and their signiicance in the vertebrate population control. Zh. Obshch. Biol. 37, 192-204. Isaev, A. S., Khlebopros, R. G. & Kondakoc, Y. P. (1974) Regularities of the insect number dynamics. Forestry 3, 27-42. Isaev, A. S., Khlebopros, R. G., Nedorezov, L. V., Kondakov, Y. P. & Kiselev, V. V. (1984) Population dynamics offorest insects. Nauka Publ. House, Novosibirsk, USSR. Ives, P. M. (1981a) Estimation of Coccinellid numbers and movement in the field. Can. Ent. 113, 981- 997. Ives, P. M. (1981b) Feeding and egg production of two species of coccinellids in th laboratory. Can. Ent. 113, 999-1005. James, B. D. & Luff, M. L. (1982) Cold-hardines and development of eggs of Rhopalosiphum insertum. Ecol. Ent. 7, 277-282. Jensen, R. E., Savage, E. F. & Hayden, R. A. (1970) The effect of certain environmental factors on cambian temperature of peach trees. J. Am. Hort. Soc. 95, 286-291. Janzen, D. H. (1977) What are dandelions and aphids? Am. Nat. 3, 586-589. Jermy, T. (1984) Evolution of insect/host plant relationships. Am. Nat. 124, 609-630. Johnsen, S., Gutierrez, A. P. & Jorgensen, J. (1997) Overwintering in the cabbage root fly Delia radicum - a dynamic model of temperature-dependent dormancy and post-dormancy development. J. Appl. Ecol., 34, 21-28. Johnson, J. A. & Whitney, N. J. (1994) Cytotoxicity and insecticidal activity of endophytic fungi from black spruce (Picea mariana) needles. Can. J. Microbial 40, 24-27. Jones, C. G., Hopper, R. F., Coleman, J. S. & Krischik, V. A. (1993) Control of systemically induced herbivore resistance by plant vascular architecture. Oecologia. 93, 452-456. Jones, F. G. W. & Dunning, R. A. (1972) Sugar Beet Pests (3rd Edn.). Bulletin 162, M.A.F.F., H.M.S.O., London. Jones, S. G. (1925) Life-history and cytology of Rhytisma acerinum (Pers.) Fries. Ann. Bot. 39, 41-75. Kahn, D. M. & Cornell, H. V. (1983) Early leaf abscission and folivores: comments and considerations. Am. Nat. 122, 428-432.

172 Karban, R., Adamcheck, R. & Schnathorst, W. C. (1987) Induced resistance and interspecific competition between spider mites and a vascular wilt fungus. Science 235, 678-680. Karban, R. & Myers, J. H. (1989) Induced plant responses to herbivory. Ann. Rev. Ecol. Syst. 20, 331- 348. Karban, R. & Niiho, C. (1995) Induced plant responses to herbivory. Ann. Rev. Ent. 20, 331-348. Kaupp, W. J. & Sohi, S. S. (1985) The role of viruses in the ecosystem. In Viral insecticides for biological control.. pp. 441-465. (Eds. K. Maramorosh & K. E. Sherman). Academic Press, New York. Kawada, K. (1987) Polymorphism and morph determination. In World Crop Pests - Aphids: Their Biology, Natural Enemies and Control Vol. 2A. pp 255-268. (Eds. A. K. Minks & P. Harrewijn). Elsevier, Amsterdam. Keith, L. B. (1963) Wildlife's ten-year cycle. University of Wiscons. Kemp, W. P. & Moody, U. L. (1984) Relationships between regional soils and foilage characteristics and the western spruce budworm (Lepidoptera: Tortricidae) outbreak frequency. Environ. Entomol. 13, 1291-1297. Kennedy, J. S. (1951) Benefits to aphids from feeding on galled and virus infected leaves. Nature 168, 825-826. Kennedy, J. S., Booth, C. 0. & Kershaw, W. J. S. (1961) Host finding by aphids in the fields. III. Visual attraction. Ann. Appl. Biol. 49, 1-21. Kennedy, J. S. & Crawley, L. (1967) Spaced-out gregariousness in sycamore aphids Drepanosiphum platanoidis (Schrank) (Hemiptera, Callaphididae). J. Anim. Ecol. 36, 147-170. Kennedy, J. S. & Stroyan, H. L. G. (1959) Biology of aphids. Ann. Rev. Entomol. 4, 139-160 Kidd, N. A. C. (1977) The influence of population density on the flight behaviour of the lime aphid, Eucallipterus tiliae. Ent. Exp. Applic. 22, 251-261. Kidd, N. A. C. (1985) The role of the host plant in the population dynamics of the large pine aphid, Cinara pinea. Oikos 44, 114-122. Kidd, N. A. C. (1990) A synoptic model to explain long-term population changes in the large pine aphid. In Dynamics of Forest Insects. pp. 317-327. (Eds. A. D. Watt, S. R. Leather, M. D. Hunter & N. A. C. Kidd). Intercept Ltd, UK. Kidd, N. A. C., Smith, S. D. J., Lewis, G. B. & Carter, C. I. (1990) Interactions between host-plant chemistry and the population dynamics of conifer aphids. In Population Dynamics of Forest Insects. pp. 183-193. (Eds. Watt, A.D., Leather, S.R., Hunter, M.D. & Kidd, N.A.). Intercept Ltd. Kidd, N. A. C. & Tozer, D. J. (1985) Distribution, survival and hatching of overwintering eggs in the large pine aphid, Cinaria pinea (Mordv.) (Horn., Lachnidae). Z. ang. Ent. 100, 17-23. Kindlmann, P. & Dixon, A. F. G. (1989) Role of population density in determining the sex-ratios of species that show local mate competition, with aphids as a model group. Funct. Ecol. 3, 311-314. Kindlmann, P. & Dixon, A. F. G. (1996) Population dynamics of a tree-dwelling aphid: Individuals to populations. Ecological Modelling, 89, 23-30. Kindlmann, P., Dixon, A. F. G. & Holman, J. (1994) Critical issues in aphid biology - Proceedings of the 4th international symposium on aphids organized by the Institute of Entomology of the Czech. Academy of Sciences and the Faculty of Biological Sciences of the University of South Bohemia Ceske Budejovice (Czech-Republic), August 30 - September 3, 1993 - foreword. Europ. J. Entomol. 91, 3. Kolomiets, N. G., Stadnitskii, G. V. & Vorontzov, A. I. (1972) The European Pine Sawfly: Distribution, Biology, Economic Importance, Natural Enemies and Control. Amerind, New Dehli. Komatsu, T. & Akimoto, S. (1995) Genetic differentiation as a result of adaptation to the phenologies of individual host trees in the galling aphid .. Ecol. Ent. 20, 33-42. Kozlowski, M. W. (1991) Mating behaviour and high level polygamy in the aphids, Periphillus acericola and Drepanosiphum platanoidis (Homoptera: Drepanosiphidae). Eth. Ecol. & Evol. 3, 285-294. Krause, S. C. & Raffa, K. F. (1992) Comparison of Insect, Fungal, and Mechanically Induced Defoliation of Larch - Effects On Plant Productivity and Subsequent Host Susceptibility. Oecologia 90, 411- 416. Kulman, H. M. (1967) Within-tree distribution and winter mortality of eggs of the woolly pine needle aphid, Schizolachnus piniradiatae. Ann. Ent. Soc. Am. 60, 384-386. Kurppa, S. (1989) Predicting outbreaks of Rhopalosiphum padi in Finland. Ann. Agric. Fenn. 28, 333- 348. Kurrpa, S. (1991) Metereological conditions to forcast of cereal pest occurrence in Finland. EPPO Bull, 21, 495-497. Kurtak, D. (1974) Overwintring of Simulium pictipes Hagen (Diptera: Simuliidae) as eggs. J. Med. Entomol. 11, 383-384. Lamb, R. J. & Pointing, P. J. (1972) Sexual morph determination in the aphid, Acyrthosiphon pisum. J. Insect Physiol. 18, 2029-2042.

173 Lamb, R. J. & Pointing, P. J. (1975) The reproductive sequence and sex determination in the aphid, Acyrthosiphum pisum. J. Insect Physiol. 21, 1443-1446. Lappalainen, J. H. & Helander, M. L. (1997) The role of foliar microfungi in mountain birch - insect herbivore relationships. Ecography, 20, 116-122. Larsson, S. (1989) Stressful times for the plant stress - insect performance hypothesis. Oikos, 56, 277- 283. Larsson, S. & Tenow, 0. (1980) Needle-eating insects and grazing dynamics in a mature Scotch Pine plantation in centran Sweden. Ecol. Bull. (Stockholm) 32, 269-306. Larsson, S. & Tenow, 0. (1984) Areal distribution of a Neodiprion setifer (Hym.: Diprionidae) outbreak on Scots pine as related to stand condition. Holarct. Ecol. 7, 81-90. Latch, G. C. M. (1993) Physiological interactions of endophytic fungi and their hosts - biotic stress tolerance imparted to grasses by endophytes. Agric. Ecosyst.Environ. 44, 143-156. Lawton, J. H. & McNeil, S. (1979) Between the devil and the deep blue sea; on the problem of being a herbivore. Br. Ecol. Soc. Symp. 20, 223-244. Leather, S. R. (1980) Egg survival in the bird cherry-oat aphid, Rhopalosiphum padi. Ent. Exp. Appl. 27, 96-97. Leather, S. R. (1981) Factors affecting egg survival in the bird cherry-oat aphid, Rhopalosiphum padi. Ent. Exp. Appl. 30, 197-199. Leather, S. R. (1983) Forecasting aphid outbreaks using winter egg counts: an assessment of its feasiblity and an example of its application in Finland. Z. ang. Ent. 96, 282-287. Leather, S. R. (1986) Host monitoring by aphid migrants: do gynoparae maximise offspring fitness? Oecologia, 68, 367-369. Leather, S. R. (1988) Size, reproductive potential and fecundity in insects: things aren't as simple as they seem. Oikos, 51, 386-389. Leather, S. R. (1990a) The role of host quality, natural enemies, competition and weather in the regulation of autumn and winter populations of the bird cherry aphid. In Population Dynamics of Forest Insects. pp. 35-44. (Eds. A. D. Watt, S. R. Leather, M. D. Hunter & N. A. C. Kidd). Intercept Press, Andover. Leather, S. R. (1990b) Two case studies: the pine beauty moth and the bird cherry aphid. In Pest, pathogens, and plant comunities. pp. 34-44. (Eds. J. J. Burdon & S. R. Leather). Blackwell, Oxford. Leather, S. R. (1992) Aspects of aphid overwintering (Homoptera, Aphidinea, Aphididae). Entomologia Generalis, 17, 101-113. Leather, S. R. (1993) Overwintering in 6 arable aphid pests - a review with particular relevance to pest management. J. Appl. Entomol. 116, 217-233. Leather, S. R. (1996a) Prunus padus L., J. Ecol. 84, 125-132. Leather, S. R. (1996b) Colonisation and distribution patterns of sycamore aphid on sycamore trees in south-east Britain. Bulletin 27, 214-218. Leather, S. R. & Lehti, J. P. (1981) Abundance and survuval of eggs of the bird cherry-oat aphid, Rhopalosiphum padi in southern Finland. Ann. Ent. Fenn. 47, 125-130. Leather, S. R. & Wellings, P. W. (1981) Ovariole number and fecundity in aphids. Ent. Exp. Appl. 30, 128-133. Leather, S. R., Watt, A. D. & Barbour, D. A. (1985) The effect of host plant and delayed mating on the fecundity and life span of the pine beauty moth, Panolis flammea. (D&S): their possibe influence on population dynamics and pest management strategies. Bull. Ent. Res. 75, 641-651. Leather, S. R. & Wellings, P. W. & Walters, K. F. A. (1988) Variation in ovariole number within the Aphidoidea. J. Nat. Hist. 22, 381-393. Leather, S. R., Walters, K. F. A. & Bale, J. S. (1993) The ecology of insect overwintering. pp. 1-255. Cambridge University Press, U.K. Lees, A. D. (1959) The role of photoperiod and temperature in the determination of parthenogenetic and sexual forms in the aphid Megoura Buckton. I. The influence of these factors on apterous virginoparae and their progeny. J. Ins. Phys. 3, 92-117. Lees, A. D. (1960) The role of photoperiod and temperature in the determination of the parthenogenetic and sexual forms in the aphid Megoura viciae Buckton II. The operation of the 'interval timer' in young clones. J. Insect Phys. 4, 154-175. Lees, A. D. (1966) The control of polymorphism in aphids. Adv. Insect Physiol. 3, 270-277. Lees, A. D. (1973) Photoperiodic time measurement in the aphid Megoura viciae. J. Insect Physiol., 19, 2279-2316. Lees, A. D. (1989) The photoperiodic responses and phenology of an English strain of the pea aphid Acyrthosiphon pisum. Ecol. EntomoL 14, 69-78. Leith, 1. D. & Fowler, D. (1987) Urban distribution of Rhytisma acerinum (Pers.) Fries (tar spot) on Sycarmore. New Phytol. 108, 175-181.

174 Letoumeau, D. K. (1988) Microorganisms as mediators of intertrophic and intratrophic interactions. In Novel aspects of insect-plant interactions. pp. 91-95. (Eds. P. Barbosa & D. K. Letoumeau). Wiley, New York: Levy, D. A. & Wood, C. C. (1992) Review of proposed mechanisms for sockeye salmon population cycles in the Fraser River. Bull. Math. Biol. 54, 241-261. Lewis, G. B. (1987) Regulating interactions between pine aphid colonies (Schizolachnus pineti) and host plant growth. Unpubl. PhD Thesis, Univ. of Wales. Lewis, T. & Navas, D. E. (1962) Thysanopteran populations overwintering in hedge bottoms, grass litter and bark. Ann. App. Biol. 50, 299-311. Liao, H. T. & Harris, M. K. (1985) Population growth of the black-margined aphid on pecan in the field. Agric. Ecosyst. Environ. 12, 253-261. Lichtenthaler, H. K. & Wellbum, A. R. (1983) Determinations of total carotenoids and chlorophylls a and b of leaf extracts in different solvents. Biochem. Soc. Trans. 603, 591-592. Llewellyn, M. & Brown, V. K. (1985) The effect of host-plant species on adult weight and the reproductive potential of aphids../. Anim. Ecol. 54, 639-650. Lorimer, N. (1980) Pest outbreaks as a function of variability in pests and plants. In Restistance to diseases and pests in forest trees. pp. 287-294. Proc. 3rd Int. Workshop Genet. Host-Parasite Interactions For., Wageningen.. Maceljski, M. & Balarin, I. (1974) Factors influencing the population density of the looper Autographa gamma L. in Yugoslavia. Acta Entomol. Yugoslay. 10, 63-76. MacGillivray, M. E. & Anderson, G. B. (1964) The effect of photoperiod and temperature on the production of gamic and agamic forms in Macrosiphum euphorbiae (Thomas). Can. J. Zool. 42, 491-510. MacKay, P. A., Reeleder, D. J. & Lamb, R. J. (1983) Sexual morph production by apterous and alate viviarous Acyrthosiphon pisum (Harris) (Homoptera: Aphididae). Can. J. Zool. 61, 952-957. Madden, J. L. (1968) Behavioural responses of parasites to the symbiotic fungus associated with Sirex noctilio F. Nature, 218, 189-190. Madden, J. L. (1988) Sirex in Australasia. In Dynamics of Forest Insect Populations: Patterns, Causes and Implications. pp. 408-431. (Ed. A. A. Berryman). Plenum Press, New York and London. Major, E. J. (1990) Water stress in sitka spruce and its effects on the green spruce aphid Elatobium abietinum, In Population Dynamics of Forest Insects, pp. 85-104. (Eds. A. D. Watt, S. R. Leather, M. D. Hunter & N. A. Kidd). Intercept Ltd. Martignioi, M. E. & Schmid, P. (1961) Studies on the resistance to virus infection in natural populations of Lepidoptera. J. Invert. Pathol 3, 62-74. Mattson, V. J. & Addy, N. D. (1975) Phytophagous insects as regulators of forest primary production. Science 190, 515-522. Mattson, W. J. & Haack, R. A. (1987) The role of drought in outbreaks of plant-eating insects. Bioscience 37, 110-118. May, R. M. (1973) Stability and complexity in model ecosystems. Princeton University Press, Princeton. May, R. M. (1974) Biological populations with non-overlapping generations: stable points, stable cycles and chaos. Science 186, 645-647. May, R. M. (1976) Simple mathematical models with very complicated dynamics. Nature 261, 459-467. May, R. M. (1983) Parasitic infections as regulators of animal populations. Am. Set 71, 36-45. May, R. M. & Seger, J. (1985) Evolutionary biology - sex-ratios in wasps and aphids. Nature, 318, 408- 409. McClure, M. S. (1983) Population dynamics of a pernicious parasite: Density-dependent vitality of red pine scale. Ecology 64, 710-718. McIntyre, J. L., Dodds, J. A. & Hare, J. D. (1981) Effects of localized infections of Nicotiana tobacum by tobacco mosaic virus on systemic resistance against diverse pathogens and an insect. Phytopathology 71, 297-301. McLeod, J. M. (1979) Discontininous stability in a sawfly life system and its relevance to pest management strategies. In Selected papers in forest entomology from the XV Int. Congr. Entomol. pp. 68-81. (Ed. W. E. Waters). U.S. For. Serv. Gen. Tech. Rept WO-8. McNeil, S. & Southwood, T. R. E. (1978) The role of nitrogen in the development of insect/plant relatioships. In Biohcemical aspects of plant and animal coevolution. (Ed. J. B. Harbome), pp. 77-98. Academic Press, New York. Meyer, G. A. & Root, R. B. (1996) Influence of feeding guild on insect response to host plant fertilization. Ecol. Entomol. 21, 270-278. Miles, P. W. (1972) The saliva of Hemiptera. Adv. Insect Physiol. .9, 183-255. Miles, P. W. (1989) Specific responses and damage caused by Aphidoidea. In Aphids: Their Biology, Natural Enemies and Control. Vol. 2C. pp. 1-21. (Eds. A. K. Minks & P. Harrewijn). Elsevier, Amsterdam.

175 Miller, J. D. (1986) Toxic metabolites of epiphytic and endophytic fungi of conifer needles. In Microbiology of the phyllosphere. pp. 223-230. (Eds. N. J. Fokkema & J. van den Heuvel). Cambridge Univ. Press, Cambridge. Mills, N. J. (1982) Voracity, cannibalism, and coccinellid predation. Ann. Appl. Biol. 101, 144-148. Milne, A. (1957) Theories of natural control of insect populations. Cold Spring Harb. Symp. Quant. Biol. 22, 253-271. Minks, A. K. & Harrewinjn, P. E. (1987) Aphids their biology, natural enemies and control. In World Crop Pests 2A. Elsevier, Amsterdam. Minks, A. K. & Harrewinjn, P. E. (1988) Aphids their biology, natural enemies and control. In World Crop Pests 2B. Elsevier, Amsterdam. Minks, A. K. & Harrewinjn, P. E. (1989) Aphids their biology, natural enemies and control. In World Crop Pests 2C. Elsevier, Amsterdam. Mittler, T. & Kleinhjan, J. E. (1970) Effect of artificial diet composition on wing production by the aphid Myzus persicae. J. Insect Physiol. 16, 833-850. Moran, N. A. (1992) The evolution of aphid life cycles. Annu. Rev. Entomol. 37, 321-348. Moran, N. A. (1993) Evolution of sex ratio variation in aphids., pp. 346-368. In Evolution and diversity of sex ratios in insects and mites (Eds. D. L. Wensch & M. Ebbert). Chapman & Hall, New York. Moran, N. A. & Whitham, T. G. (1980) Population fluctuations in complex life cycles: an example from Pemphigus aphids. Ecology 69, 1214-1218. Mordwilko. (1908) Beitrage zur biologie der pflanzenlause, Aphididae passerini. Biol. Zool. 28, 631-639. Morris, R. F. (1963) The dynamics of epidemic spruce budworm populations. Mem. Ent. Soc. Can. 31, 1- 32. Morris, R. F. (1964) The value of historical data in population research, with particular reference to Hyphantria cunea Drury. Can. Entomol. 96, 356-368. Mountford, M. D. (1988) Population regulation, density dependence, and heterogeneity. J Anim. Ecol. 57, 845-858. Miller, K. (1912) Ueber das biologische Verhalten von Rhytisma acerinum auf verschiedenen Ahornarten. Berichte d. Bot. Ges. xxx, 385-391. Murdoch, W. W. (1970) Population regulation and population inertia. Ecology 51, 497-502. Murdoch, W. W. & Walde, S. J. (1989) Analysis of insect population dynamics. In Toward a more exact ecology. (Eds. P. J. Grubb & J. B. Whittaker). Blackwell, Oxford. Myers, J. H. (1988a) The induced defence hypothesis, does it apply to the population dynamics of insects? In Chemical mediation of coevolution. pp. 530-557. (Ed. K. Spencer). Academic Press, New York. Myers, J. H. (1988b) Can a general hypothesis explain population cycles of forest Lepidoptera? Adv. Ecol. Res. 18, 179-242. Myers, J. H. (1990) Population cycles of western tent caterpillars - experimental introductions and synchrony of fluctuations. Ecology, 71, 986-995. Myers, J. H. (1998) Synchrony in outbreaks of forest Lepidoptera: a possible example of the Moran effect. Ecology, 79, 1111-1117. Neilson, M. M. & Morris, R.F. (1964) The regulation of European spruce sawfly numbers in the Maritime Provinces of Canada from 1933 to 1963. Can. Entomol. 96, 773-784. Neuenschwander, P., Hagen, K. S. & Smith, R. F. (1975) Predation on aphids in California's alfalfa fields. Hilgardia, 43, 78. Neuvonen, S., Hanhimaki, S., Suomela, J. & Haukioja, E. (1988) Early season damage to birch foliage affects the performance of a late season herbivore. J Appl. Entomol. 105, 182-189. Newton, C. (1986) Overwintering strategies of the English grain aphid Sitobium avenae (F.). Unpub. PhD thesis, Univ. East Anglia, UK Newton, C. & Dixon, A. F. G. (1987) Cost of sex in aphids: size of males at birth and the primary sex- ratio in Sitobion avenae (F). Funct. Ecol. 1, 62-70. Newton, C. & Dixon, A. F. G. (1988) The cost from switching from sexual to asexual reproduction. Entomol. Exp. Appl. 47, 283-287. Nicholoson, A. J. (1954) An outline of the dynamics of animal populations. Part I. Aust. J. Zool. 2, 9-65. Niemeld, P. (1980) Dependence of Oporina autumnata (Lep: Geometridae) outbreaks on summer temperature. Rep. Kevo. Subarct. Res. Stn. 16, 27-30. Niemeld, P., Tuomi, J., Mannila, R. & Ojala, P. (1984) The effect of previous damage on the quality of Scots pine foilage as food for Diprionid sawflies. Z. Angew. Entomol. 98, 33-43. Noble, M. D. (1958) A simplified clip cage for aphid investigation., Can. Ent. 90, 760. Novak, H. & Archziger, R. (1995) Influence of heteropteran predators (Het., Anthocoridae, Miridae) on larval populations of hawthorn psyllids (Horn., Psyllidae). J. Appl. Ent. 119, 479-486. Ohmart, C. P., Stewart, L. G. & Thomas, J. R. (1983) Leaf consumption by insects in three eucalyptus forest types in South eastern Australia and their role in short term nutrient cycling. Oecologia 59, 322-330.

176 Onstad, D. W. (1988) Population dynamics theory - the roles of analytical, simulation, and supercomputer models. Ecological Modelling 43, 111-124. Orlando, E. (1974) Sex determination in Megoura viciae Buckton (Homoptera, Aphididae). Monitore Zoologico Italiano (NS.), 8, 61-70. Orlando, E. (1983) Chromosomal abnormalities in male-producing eggs: a study in Megoura viviae (Homoptera: Aphididae). Genetica 62, 55-59. Parry, W. H. (1974) The effect of nitrogen levels in Sitka spruce needles on Elatobium abitetinum (Walker) populations in north-eastern Scotland. Oecologia 15, 305-320. Parry, W. H. (1976) The effect of needle age on the acceptability of Sitka spruce needles to the aphid, Elatobium abietinum (Walker). Oecologia 23, 297-313. Peel, A. L. & Ho, L. C. (1970) Colony size of (Gmelin) in relation to mass transport of 14C-labelled assimilates from the leaves in willow. Physiol. Plant. 23, 033-1038. Perelle, A. D. & Harper, J. D. (1986) An evaluation of the impact of sublethal dosages of nuclear polyhedrosis virus in larvae on pupae, adults, and adult progeny of the fall armyworm, Spodoptera frugiperda. J. Invert. Pathol. 47, 42-47. Petrini, 0. (1986) of endophytic fungi of aerial plant tissues. In Microbiology of the phyllosphere. pp. 175-187. (Eds. N. J. Fokkema & J. van den Heuvel). Cambridge Univ. Press, Cambridge. Petrini, 0., Sieber, T. N., Toti, L. & Viret, 0. (1992) Ecology, metabolite production, and substrate utilization in endophytic fungi. Nat. Tox. 1, 185-186. Pouvreau, A. (1970) Donnees ecologiques sur l'hibernation controlee des reines de bourdons (Hymenoptera, Apoidea, Bombinae, Bombus, Leitr). Apidologie 1, 73-95. Prestidge, R. A. & Gallagher, R. T. (1988) Endophyte fungus confers resistance to ryegrass: Argentine stem weevil larval studies. Ecol. Entomol. 13., 429-435. Preszler, R. W., Gaylord, E. S. & Boecklen, W. J. (1996) Reduced parasitism of a leaf-mining moth on trees with high infection frequencies of an endophytic fungus. Oecologia, 108, 159-166. Price, P. W. (1981) Relevance of ecological concepts to practical biological control. In Biological control in crop production, pp. 3-19. (Ed. N. Papavisas). Allenheld, Osmun, Montclair, New Jersey, USA. Price, P. W., Bouton, C. E., Gross, P., McPheron, B., Thompson, J. N. A. & Weis, A. E. (1980) Interaction among three trophic levels: influence of plants on interactions between insect herbivores and natural enemies. Ann. Rev. Ecol. Syst. 11, 41-65. Prokopy, R. J. & Owens, E. D. (1978) Visual genearlist with visual specialist phytophagous insects: host selection, behaviour and application to management. Ent. Exp. AppL 24, 409-420. Prokopy, R. J. & Owens, E. D. (1983) Visual detection of plants by herbivorous insects. Ann. Rev. Ent. 28, 337-364. Pschorn-Walcher, H. (1980) Population fluctuations and parasitization of the birch-alder casebearer (Coleophora seratella L.) in relation to habitat diversity. Z. Angew. Entomol. 89, 63-81. Rabin, L. B. & Pacovsky, R. S. (1985) Reduced larvae growth of two Lepidoptera () on excised leaves of soybean infected with a mycorrhizal fungus. J. Econ. Entomol. 78, 1358-1363. Rafes, R. M. (1978) Biogeoscoenological theory of the population dynamics of phytophagous forest insects. Mathematical modeling in ecology. Raffa, K. F. & Berryman, A. A. (1983) The role of host plant resistance in the colonisation behavior and ecology of the bark beetles (Coleoptera: Scolytidae). Ecol. Monogr. 53, 27-49. Raps, A. & Vidal, S. (1998) Indirect effects of an unspecialized endophytic fungus on specialized plant - herbivorous insect interactions. Oecologia, 114, 541-547. Raven, J. A. (1983) Phytophages of xylem and phloem: a comparsion of animal and plant sap-feeders. Ann. Rev. Ecol. Syst. 13, 135-234. Raworth, D. A. (1984a) Population dynamics of the cabbage aphid, Brevicoryne brassicae (Homoptera: Aphididae) at Vancouver, British Columbia - II. Development, fecundity, and longevity. Can Ent. 116, 871-878. Raworth, D. A. (1984b) Population dynamics of the cabbage aphid, Brevicoryne brassicae (Homoptera: Aphididae) at Vancouver, British Columbia - IV. Predation by Aphidoletes aphidimyza (Diptera: cecidomyiidae. Can. Entomol. 116, 889-893. Raworth, D. A., McFarlane, S., Gilbert, N. & Frazer, B. D. (1984) Population dynamics of the cabbage aphid, Brevicoryne brassicae (Homoptera: Aphididae) at Vancouver, British Columbia - III. Development, fecundity, and morph determination vs. aphid density and plant quality. Can. Ent. 116, 879-888. Readshaw, J. L. (1965) A theory of phasmatid outbreak release. Autstral. J. Zool. 13. 33-48. Rhoades, D. F. (1979) Evolution of plant chemical defense against herbivores. In Herbivores: Their Interaction with Secondary Plant Metabolites. pp. 3-54. (Eds. G. A. Rosenthal & D. H. Janzen). Academic Press, London.

177 Rhoades, D. F. (1983) Herbivore population dynamics and plant chemistry., In Variable plants in natural and managed systems. pp. 155-220. (Eds. R. F. Denno & M. S. McClure). Academic Press, New York. Rhoades, D. F. (1985) Offensive-defensive interactions between herbivores and plants, their relevance in herbivore population dynamics and ecological theory. Am. Nat. 125, 205-238. Rhoades, D. F. & Cates, R. G. (1976) Toward a general theory of plant antiherbivore chemistry. Recent Adv. Phytochem. 10, 168-213. Ricker, W. E. (1954) Stock and recruitment. J Fish. Res. Bd. Can. 20, 257-284. Rispe, C., Simon, J. C. & Pierre, J. S. (1996) Fitness comparison between colones differing in their ability to produce sexuals in the aphid Rhopalosiphum padi. Ent. Exp. Appl. 80, 469-474. Roland, J. & Myers, J. H. (1987) Improved insect performance from host-plant defoliation: winter moth on oak and . Ecol. Entomol. 12, 409-414. Rosenthal, G. A. (1983) A seed-eating beetle's adaptations to the poisonous seeds. Sci. Am. 249, 164-171. Rossiter, M., Schultz, J. C. & Baldwin, I. T. (1988) Relationships among defoliation, red oak phenolics, and gypsy moth growth and reproduction. Ecology 69, 267-277. Rossiter, M. C. (1991) Environmentally-based maternal effects - a hidden force in insect population dynamics. Oecologia 87, 288-294. Rossiter, M. C. (1995) Impact of life-history evolution on population dynamics: predicting the presence of maternal effects. In Population Dynamics: New Approaches and Synthesis. pp. 251-276. (Eds. N. Cappuccino & P. W. Price). Academic Press, San Diego: Rothschild, M. (1972) Secondary plant substances and warning coloration in insects. In Insect/plant relationships. pp. 59-83. (Ed. H. van Emden). Blackwell, Oxford. Rothschild, M. & Schoonhoven, L. (1977) Assessment of egg load by Pieris brassicae (Lepidoptera: Pieridae). Nature 266, 352-355. Royama, T. (1977) Population persistence and density dependence. J. Ecol. Monogr. 47, 1-35. Royama, T. (1992) Analytical population dynamics. In Population and Community Biology Series 10 . pp. 1-371. (Eds. M. B. Usher & R. L. Kitching). Chapman & Hall, London. Royama, T. (1997) Population dynamics of forest insects: are they governed by single or multiple factors? In Forests and Insects. pp. 37-48. (Eds. A. D. Watt, N. E. Stork & M. D. Hunter). Rudnev, D. F. (1972) Causes of mass reproduction of forest pests and the nature of resistance in plantings. Proc. 13th Int. Congr. Entomol., Moscow 1968. Ruohomaki, K., Hanhimaki, S. & Haukioja, E. (1992) Variability in the efficacy of delayed inducible resistance in mountain birch. Entomol. Exp. Appl. 62, 107-115. Saikkonen, K., Helander, M., Ranta, H., Neuvonen, S., Virtanen, T., Suomela, J. & Vuorinen, P. (1996) Endophyte-mediated interactions between woody plants and insect herbivores? Ent. Exp. Appl., 80, 269-271. Saikkonen, K., Faeth, S. H., Helander, M. & Sullivan, T. J. (1998) Fungal endophytes: A continuum of interactions with host plants. Annual Review of Ecology and Systematics 29, 319-343. Schowalter, T. D., Hargrove, W. W. & Crossley, D. A. (1986) Herbivory in forested ecosystems. Ann. Rev. Entmol. 31, 177-196. Schultz, J. C. & Baldwin, 1. T. (1982) Oak leaf quality declines in respose to defoliation by gypsy moth larvae. Science 217, 149-151. Searle, J. B. & Mittler, T. E. (1981) Embryognesis and the production of males by apterous viviparae of the green peach potato aphid Myzus persicae in relation to photoperiod. J. Insect. Physiol. 27, 145-153. Searle, J. B. & Mittler, T. E. (1982) Embryogenesis and oogenesis in alate virginoparae, gynoperae and oviparae of the aphid Myzus persicae, in relation to photoperiod. J. Ins. Physiol. 28, 213-220. Sequeira, R. & Dixon, A. F. G. (1997) Population dynamics of tree-dwelling aphids: the importance of seasonality and time scale. Ecology, 78, 2603-2610. Service, P. & Lenski, R. E. (1982) Aphid genotypes, plant phenotypes, and genetic diversity: a demographic analysis of experimental data. Evolution 36, 1276-1282. Sethi, S. L. & Swenson, K. G. (1967) Formation of sexuparae in the aphid Eriosoma pyricola, on pear roots. Ent. Exp. Appl. 10, 97-102. Shaw, M. J. P. (1983) Some effects of infestation by the black pine aphid, Cinara cronartii (T and P) (Hemiptera: Aphididae). Proceedings of the fourth entomology congress of the entomology society of Africa, Johannesburg, September 1983, p. 11-17. Siegel, S. (1956) Nonparameteric statistics for the behavioral sciences. McGraw-Hill, New York. Simpson, R. G. & Welborn, C. E. (1975) Aggregations of alfalfa weevils, Hypera postica, convergent ladybirds, Hippodamia convergens, and other insects. Environ. Entomol. 4, 193-194. Sinclair, A. R. E. (1989) Population regulation in animals. In Ecological concepts. (Ed. J. M. Cherrett). Oxford: Blackwell. Smith, H. R. (1985) Wild life and the gypsy moth. Wildl. Soc. Bull. 13, 166-174.

178 Sokal, R. R. & Rohfl, F. J. (1995) Biometry: the principles and practice of statistics in biological reasearch. pp. 850. W.H. Freeman & Co., New York. Somme, L. (1969) Mannitol and glycerol in overwintering aphid eggs. Norske. Ent. Tidsskr. 16, 107-111. Southwood, T. R. E. (1975) The dynamics of insect populations. In Insects, Science and Society. (Ed. D. Pimentel), pp. 151-199. Academic Press, New York. Southwood, T. R. E. & Comins, H. N. (1976) A synoptic population model. J. Anim. Ecol. 45, 949-965. Srivastava, P. N., Lambein, F. & Auclair, J. L. (1988) Nonprotein amino acid-aphid interaction: phagostimulatory effects and survival of the pea aphid, Acyrthosiphon pisum. Ent. Exp. Appl. 48, 109-115. Stadler, B. & Mackauer, M. (1996) Influence of plant quality on interactions between the aphid parasitoid Ephedrus californicus Baker (Hymenoptera: Aphidiidae) and its host, Acyrthosiphon pisum (Harris) (Homoptera: Aphididae). Can. Entomol. 128, 27-39. Stamp, N. E. (1980) Egg deposition patterns in butterflies: why do some species cluster their eggs rather than deposit them singly? Am. Nat, 115, 367-380. Stark, R. W. (1959) Climate in relation to winter mortality of the lodge pole needle miner, Recurvaria starki Free., in Canadian Rocky Mountain Parks. Can. .1 Zool. 37, 753-761. Stary, P. (1970) Biology of aphid parasites (Hymenoptera: Aphididae) with respect to integrated control. pp. 643: W. Junk Publishers, The Hague, The Netherlands. Stary, P. (1988) Aphilinidae. In Aphids, their biology, natural enemies and control. 2B. (Eds. A. K. Minks & P. Harrewijn), pp. 185-188. Elsevier, Amsterdam. Steel, C. G. H. & Lees, A. D. (1977) The role of neurosecretion in the photoperiodic control of polymorphism in the aphid Megoura viciae. J. Exp. Biol. 67, 117-135. Stoszek, K. J., Mika, P. B., Moore, J. A. & Osborne, H. L. (1981) Relationships of Douglas-fir tussock moth defoliation to site and stand characteristics in Northern Idaho. For. Sci. 27, 431-442. Strathdee, A. T., Bale, J. S., Block, W. C., Webb, N. R., Hodkinson, I. D. & Coulson, S. J. (1993) Extreme adaptive life-cycle in a high arctic aphid, Acyrthosiphon svalbardicum, Ecol. Entomol. 18, 254-258. Strauss, S. Y. (1987) Direct and indirect effects of host-plant fertilization on an insect community. Ecology 68, 1670-1678. Strong, D. R. (1988) Endophyte mutualism and plant potection from herbivores. Ecology 69, 132-141. Strong, D. R., Lawton, J. H. & Southwood, R. (1984) Insects on Plants: Community Patterns and Mechanims. Harvard Univ. Press, Cambridge, Mass. Stuedemann, J. A., Wilkinson, S. R., Belesky, C. S., Hoveland, 0. J., Divine, F. N., Thomsen, H. C., McCampbell, W. E., Townsend, W. E. & Ciordia, H. (1986) Efffect of level of fungus and nitrogen fertilisation rate of KY-31 tall fescue on steer performance. J. Anim. Sci. 63, 290-291. Takada, H. (1988) Interclonal variation in the photoperiodic reponse for sexual morph production of Japanese Aphis gossypii Glover (Horn. Aphididae). I Appl. Entomol. 106, 188-197. Takekawa, J. Y., Garton, E. 0. & Langelier, L. A. (1982) Biological control of forest insect outbreaks: the use of avian predators. Trans. North Am. Wildl. Nat. Resour. Conf 47, 393-409. Takemoto, B. K., Bytnerowicz, A. & Olszyk. (1989) Physiological responses of field-grown strawberry (Fragaria x Ananassa Duch.) exposed to acidic fog and ambient ozone. Environ. & Experim. Bot. 29, 379-386. Tamaki, G. & Long, G. E. (1987) Predator complex of the green peach aphid on sugarbeets: expansion of the predator power and efficacy model. Env. Ent. 7, 835-842. Taper, M. L., Zimmerman, E. M. & Case, T. J. (1986) Sources of mortality for a cynipid gall-wasp Dryocosmos dubiosus (Hymenoptera: Cynipidae): the importance of the tannin/fungus interaction. Oecologia 68, 437-445. Tauber, M. J., C.A., T. & Masaki, S. (1986) Seasonal Adoptions in Insects. Oxford University Press, Oxford. Taylor, L. R. (1975) Longevity, fecundity and size, control of reproductive potential in a polymorphic migrant, Aphis fabae Scop. J. Anim. Ecol. 44, 135-159. Taylor, P. D. (1981) Sex ratio compensation in ant populations. Evolution 35, 125-151. Tedders, W. L. & Thompson, J. M. (1981) Histoloical investigation of stylet penetration and feeding damage to pecan foliage by three aphids (Hemiptera (Homoptera): Aphididae). pp. 69-83. In Southeastern Fruit and Tree Nut Research Laboratory, Science and Education Administration, USDA Miscellaneous Publication. Tercafs, R. & Thines, G. (1973) Etude des dedencheurs visuels intervenant lors de la penetration souterraine de Scolopteryx libratrix L. Triphosa dubitata L. (Lepidopterestrogloxenes). Annales de Speleologie 28, 177-181. Thompson, J. N. (1982) Interaction and coevolution. New York: Wiley. Thomson, A. J. & Shrimpton, D. M. (1984) Weather associated with the start of mountain pine beetle outbreaks. Can. J. For. Res. 14, 254-258.

179 Thomson, A. J., Shepherd, R. F., Harris, W. E. & Silversides, R. H. (1984) Relating weather to outbreaks of western spruce budworm, Choristoneura occidentalis (Lepidoptera: Tortricidae) in British Columbia. Can. Entomol. 116, 375-381. Thornback, N. (1983) The factors determining the abundance of Metopolophium dirhodum (Walk.), the rose grain aphid. Unpubl. Ph.D. Thesis, Univ. East Anglia, Norwich. Thorpe, K. W. & Barbosa, P. (1986) Effects of consumption of high and low nicotine tobacco by Manduca sexta (Lepidoptera: Sphingidae) on survival of gregarious endoparasitoid Cotesia congregata (Hymenoptera: Braconidae). J Chem. Ecol. 12, 1329-1337. Torgersen, T. R. (1985) Role of birds and ants in western spruce budworm dynamics. In Recent advances in spruce budworm research. pp. 97-98. (Eds. C. J. Sanders, R. W. Stark, E. J. Mullins & J. Murphy). Can. For. Serv., Ottawa. Trapman, M. & Blommers, L. (1992) An attempt to pear sucker management in the Netherlands. J Appl. Ent. 114, 38-51. Tsitsipis, J. A. & Mittler, T. E. (1977a) Influence of day length on the production of parthenogenetic and sexual females of Aphis fabae at 17.5 °C. Ent. Exp. Appl. 21, 163-173. Tsitsipis, J. A. & Mittler, T. E. (1977b) Influence of temperature and day length on the production of males by Aphis fabae. Ent. Exp. Appl. 21, 229-237. Tuomi, J., Niemela, P., Rousi, M., Siren, S. & Vourisalo, T. (1988) Induced accumulation of foilage phenols in mountain birch: Branch response to defoliation? Am. Nat. 132, 602-608. Turchin, P. & Taylor, A. D. (1992) Complex dynamics in ecological time-series. Ecology 73, 289-305. Turlings, T. C. J., Bernasconi, M., Bertossa, R., Bigler, F., Caloz, G. & Dom, S. (1998) The induction of volatile emissions in maize by three herbivore species with different feeding habits: Possible consequences for their natural enemies. Biological Control 11, 122-129. Turlings, T. C. J., McCall, P. J., Alborn, H. T. & Tumlinson, J. H. (1993) An elicitor in caterpillar oral secretions that induces corn seedlings to emit chemical signals attractive to parasitic wasps. J. Chem. Ecol. 19, 411-425. Uvarov, B. P. (1931) Insects and climate. Trans.EntomoL Soc. Lund 79, 1-247. van Emden, H. F. (1972) Aphids as phytochemists. In Phytochemical Ecology, pp. 25-43. (Ed. J. B. Harborne). Academic Press, London, New York. Varley, G. C., Gradwell, G. R. & Hassell, M. P. (1973) Insect Population Ecology: An Analytical Approach. Blackwell Sci. Publ., London. Vet, L. E. M., van Lenteren, J. C., Heymans, M. & Meelis, E. (1983) An airflow olfactometer for measuring olfactory responses of Hymenoperous parasitoids and other small insects. Physiol. Entomol. 8, 97-106. Vickerman, G. P. & Sunderland, K. D. (1975) Arthropods in cereal crops: nocturnal activity, vertical distribution and aphid predation. J. Appl. Ecol. 12, 755-766. Viktorov, G. A. (1971) Some general principles of insect population density regulation. Proc. 13th Int. Congr. Entomol. Moscow 1968. 1, 573-574. Vince Prue, D. (1975) Photoperiodism in plants. McGraw-Hill, London. Vinson, S. B. (1975) Biochemical coevolution between parasitoids and their hosts. In Evolutionary strategies of parasitic insects and mites. pp. 14-48. (Ed. P. W. Price). Plenum Press, New York. Wallner, W. E. (1987) Factors affecting insect population dynamics: differences between outbreak and non-outbreak species. Ann. Rev. Entomol. 34, 317-340. Wallner, W. E. & Walton, G. S. (1979) Host defoliation: A possible determinant of gypsy moth quality. Ann. Entomol. Soc. Am. 72, 62-67. Walters, K. F. A. (1987) Forcasting the immigration of aphids on potatoe crops. Conf Pests Agricult., Paris. p. 587-594. Ward, A. A. & Wellings, P. W. (1994) Deadlines and delays as factors in aphid sex allocation. Eur. J. Entomol. 91, 29-36. Ward, S. A., Dixon, A. F. G. & Wellings, P. W. (1983a) The relationship between fecundity and reproductive investment in aphids. J. Anim. Ecol. 52, 451-461. Ward, S. A., Leather, S. R. & Dixon, A. F. G. (1984) Temperature prediction and the timing of sex in aphids. Oecologia 62, 230-233. Ward, S. A., Wellings, P. W. & Dixon, A. F. G. (1983b) The effect of reproductive investment on pre- reproductive mortality in aphids. J. Anim. Ecol. 52, 805-813. Waring, G. L. & Cobb, N. S. (1992) The impact of plant stress on herbivore population dynamics. pp. 167-226. In Insect-Plant Interactions. Watt, A. D. (1984) Reproductive strategies of the alate and apterous morphs of the grain aphid, Sitobion avenae. Ent. Exp. Appl. 36, 1-7. Watt, A. D. (1994) The relevance of the stress hypothesis to insects feeding on tree foliage. In Individuals, Populations & Patterns in Ecology. pp. 73-85. (Eds. S. R. Leather, A. D. Watt, N. J. Mills & K. F. A. Walters). Intercept Ltd.

180 Watt, A. D. & MacFarlane, A. M. (1991) Winter moth on Sitka spruce: synchrony of egg hatch and budburst, and its effect on larval survival. Ecol. Entomol. 16, 387-390. Way, M., Cammell, M. & Tatchell, M. (1992) Forecasting based on 20 years data on populations of Aphis fabae in Southern England - the significance of weather. In Monitoring and forecasting to improve crop and environment protection. p.13. Association of Applied Biologists 1992 Presidential Meeting. Ecole Nationale Superieure Agronomique, Rennes, France. 8-10 September 1992. Way, M. J. & Banks, C. J. (1964) Natural mortality of eggs of the black bean aphid, Aphis fabae Scop. on the spindle tree, Euonymus europaeus L. Ann. App. Biol. 54, 255-264. Way, M. J. & Cammell, M. (1970) Aggregation behaviour in relation to food utilization by aphids. In Animal Populations in Relation to their Food Resources. British Ecol. Soc., Symposium 10: pp. 229-247. (Ed. A. Watson). Blackwell. Oxford, London. Way, M. J. & Cammell, M. E. (1973) The problem of pest and disease forecasting - possibilities and limitations as exemplified by work on the bean aphid, Aphis fabae. pp. 933-954. Proceedings 7th British Insecticide & Fungicide Conference (1973). Way, M. J., Cammell, M. E., Taylor, L. R. & Woiwod, 1. P. (1981) The use of egg counts and suction trap samples to forecast the infestation of spring-sown field beans, Vicia faba, by the black bean aphid, Aphis fabae. Ann. appl. Biol. 98, 21-34. Way, N. J. & Banks, C. J. (1968) Population studies on the active stages of the black bean aphid, Aphis fabae Scop. on its winter host Euonymus europaeus L. Ann. Appl. Biol. 62, 177-197. Weibull, J. (1987) Seasonal changes in the free amino acids of oat and barley phloem sap in relation to plant growth stage and growth of Rhopalosiphum padi. Ann. appl. Biol. 111, 729-737. Weisser, W. W. & Stadler, B. (1994) Phenotypic plasticity and fitness in aphids. Eur. J. Entomol. 91, 71- 78. Wellings, P. W. (1980) Qualitative changes in the regulation of sycamore aphid numbers. Unpubl. Ph.D. Thesis, Univ. East Anglia. Wellings, P. W. (1981) The effect of temperature on the growth and reproduction of two closely related aphid species on sycamore. Ecol. Entomol. 6, 209-214. Wellings, P. W., Chambers, R. J., Dixon, A. F. G. & Aikman, D. P. (1985) Sycamore aphid numbers and population density: I. Some Patterns. J. Anim. Ecol. 54, 411-424. Wellings, P. W. & Dixon, A. F. G. (1987) Sycamore aphid numbers and population density: III. The role of aphid-induced changed in plant quality. J. Anim. Ecol. 56, 161-170. Wellings, P. W., Leather, S. R. & Dixon, A. F. G. (1980) Seasonal variation in reproductive potential: a programmed feature of aphid life cycles. J Anim. Ecol. 49, 975-985. Wellington, W. G. (1952) Air-mass climatology in Ontario North of Lake Huron and Lake Superior before outbreaks of the spruce budworm, Choristoneura fumiferana (Clem.) and the forest tent caterpillar, Malacasoma disstria Hbn. Can. J. Zool. 30, 114-127. Wellington, W. G. (1957) Individual differences as a factor in population dynamics: the development of problem. Can. J. Zool. 35, 293-323. White, P. L. (1970) The effect of aphids on tree growth. UnpubL Ph.D. thesis, University of Glasgow. White, T. C. R. (1969) An index to measure weather-induced stres of trees associated with psyllids in Australia. Ecology 50, 905-909. White, T. C. R. (1974) A hypothesis to explain outbreaks of looper caterpillars, with special reference to populations of Selidosema suavis in a plantation of Pinus radiata in New Zealand. Oeclologia 22, 119-134. White, T. C. R. (1984) The abundance of invertebrate herbivores in relation to the availability of nitrogen in stressed food plants. Oecologia 63, 90-105. White, T. D. R. (1978) The importance of a relative shortage of food in animal ecology. Oecologia 33, 71-86. Whitman, D. W. (1988) Allochemical interactions among plants, herbivores and their predators. In Novel aspects of insect-plant interactions. pp. 11-46. (Eds. P. Barbosa & D. K. Letourneau). Wiley. New York. Whittaker, J. B. (1984) Responses of sycamore (Acer pseudoplatanus) leaves to damage by a typhlocybine leaf hopper, Ossiannilssonola callose. Ecology 72, 455-462. Whittaker, J. B. & Warrington, S. (1985) An experimental field study of different levels of insect herbivory induced by Formica rufa predation on sycamore (Acer pseudoplatanus). III. Effects on tree growth. J. Appl. Ecol. 22, 797-811. Wickremasinghe, M. G. V. & van Emden, H. F. (1992) Reactions of adult female parasitoids, particularly Aphidius rhopalosiphi, to volatile chemical cues from the host plants of their aphid prey. Physiol. Entomol. 17, 297-304. Williams, C. T. & Wratten, S. D. (1987) The wintr development, reproduction and life span of viviparae of Sitobium avenae (F.) (Hemiptera: Aphididae) on wheat in England. Bull. Ent. Res. 77, 19-34.

181 Williams, K. S. & Myers, J. H. (1984) Previous herbivore attack of red alder may improve food quality for full webworm. larvae. Oecologia 63, 166-170. Wilson, D. (1993) Fungal endophytes: out of sight but should not be out of mind. Oikos 69, 379-384. Wilson, D. (1995a) Endophyte - the evolution of a term, and clarification of its use and definition. Oikos 73, 274-276. Wilson, D. (1995b) Fungal endophytes which invade insect galls: insect pathogens, benign saprophytes, or fungal inquilines? Oecologia 103, 255-260. Wilson, D. & Carroll, G. C. (1994) Infection studies of Discula quercina, and endophytes of Quercus garryana. Mycologia, 86, 635-647. Wilson, A. D., Clement, S. L. & Kaiser, W. J. (1991) Survey and detection of endophytic fungi in lolium germ plasm by direct staining and aphid assays. Plant Disease, 75, 169-173. Wittler, J. A. & Waisanen, L. A. (1978) The effect of differential flushing times among trembling aspen clones on tortricid catarpilar populations. Environ. Entomol. 7, 139-143. Wratten, S. D. & Pearson, J. (1982) Predation of sugar beet aphids in New Zealand. Ann. Appl. Biol. 101, 178-181. Wratten, S. D., Edwards, P. J. & Winder, L. (1988) Insect herbivory in relation to dynamic changes in host plant-quality. Biological Journal of the Linnean Society, 35, 339-350. Wright, J. E. & Laing, J. E. (1980) Numerical response of coccinellids to aphids in corn in southern Canario. Can. Ent. 112, 977-988. Wulf, A. (1990) Ueber die Bedeutung von Diplodina acerina (Pass.) Sutton and anderen Blattpilzen als Antagonisten der Fenstergallmuecke Dasineura vitrina Kffr an Bergahorn (Ater pseudoplatanus L.). Nachrichtenblatt Deut. Pflanzenschtzd 42, 97-102. Yamaguchi, Y. (1985) Sex ratios of an aphid subject to local mate competition with variable maternal condition. Nature 318, 460-462. Young, A. M. (1979) Notes on population outbreak of the beetle, Uroplata sp. (Coleoptera: Chrysomelidae) on the tree Bunchosia pilosa (Malpighiaceae) in Costa Rica. J. NY Entomol. Soc. 87, 289-298. Zelinskaya, L. M. (1980) The role of microsporidia in the population dynamics of the gypsy moth (Porthetria dispar L.) in forest plantations in the lower Cis-Dneiper region. Vestn. Zool. 1, 57- 62.

182 APPENDIX Preliminary pesticide screening

INTRODUCTION

To select the most suitable pesticide for eradicating aphids in spring 1997, a number of chemicals with known aphicidal activity were compared, and the most effective was chosen for the main experiments. Both phytotoxicity and aphicidal effects were examined.

MATERIALS & METHODS

Forty-eight 7 month old sycamore saplings, conditioned for 6 weeks in a controlled temperature room at 18°C, ambient RH and 16h light:8h dark, were treated with the following chemicals: Cypermethrin (Ambush C®, Zeneca); Rotenone (Liquid Denis®, Bio); Dimethoate (Atlas Dimethoate 40®, Atlas); Malathion (Liquid Malathion®, Murphy); and Pirimicarb (Pirimor®, ICI). These were sprayed onto the leaves until run- off using a Mardrive linear-track sprayer at a flow-volume of 2501 ha-1 with an average speed of 0.92 (range 0.88-0.99) ms' under 3 bar pressure, at the manufacturers recommended dosage for controlling aphids and at two further concentrations (one third and one fifth). Three replicate saplings for each pesticide concentration were used, and then compared to control saplings sprayed with pure water. The performance of each pesticide was assessed by means of: (i) an aphid bioassay; (ii) monitoring the chlorophylls and carotenoids content of the plants before spraying with the pesticide and 36 days after.

The aphid bioassay involved placing 5 adult parthenogenetic virginoparae on each treated plant and monitoring the number of offspring produced over five days. In total, 7 different batches of aphids were placed consecutively on the same treated plant. Leaf chlorophyll (chl a, b and total) and carotenoid concentrations were extracted with diethyl ether, measured spectrophotometrically (Unican SP1800 Ultraviolet), and expressed on a dry weight basis according to the methods of Takemoto et al. (1989) and Lichtenthaler & Wellburn (1983). The loss of photosynthetic pigments was used as an indicator of plant stress (see Hendry & Price, 1993). 183 DATA TRANSFORMATION AND STATISTICAL ANALYSIS

For the pesticide selection experiment, analysis of covariance (ANCOVA) was performed on the phyto-chemistry data using GLIM (version 3.77 copyright 1985 Royal Statistical Society, London) by specifying the response variable to be the logarithm of the ratio of final to initial photosynthetic pigment content, with time as the covariate, and pesticide (six levels) and concentration (three levels) as the two factors. Analysis of the bioassay data was done similarly, but with logged nymphal output as the response variable.

RESULTS & CONCLUSIONS

The biochemical analysis revealed no significant differences in the photosynthetic pigment levels between treatments over time (see Table 1 and Figs. 1 a, b & c). It can be concluded that observing the changes in photosynthetic pigments in sycamore leaves was not a reliable method of assessment for looking at plant stress.

Table 1: Biochemical effects of pesticides on Acer pseudoplatanus foliage:

PHOTOSYNTHETIC STATISTICAL TERMS F RATIO P PIGMENT Chlorophyll a Chemical.concentration F 12, 40 = 0.925 n.s. interaction Chemical F 4, 52 = 0.259 n.s. Concentration F 3, 52 = 1.667 n.s. Chlorophyll b Chemical.concentration F 12,40= 1.120 n.s. interaction Chemical F 4,52 = 0.365 n.s. Concentration F 3,52 = 1.980 n.s. Total carotenoids Chemical.concentration F 12, 40 = 0.237 n.s. interaction Chemical F 4, 52 = 0.044 n.s. Concentration F 3, 52 = 1.247 n.s.

184

\. , 'a„'" ,'? ‘r q' .,, ;-..9-, ,,-..-.- -4_, ..9-1/-4'o'''' 6 " - 0 6 d- e ...,. , .3. oe, .(\ e C`. o6., 0

Fig. 1 a: Change in chlorophyll a content of sycamore leaves subjected to different aphicide treatments.

2:0 1

1C0

e hang c ll b hy

hlorop -an f c o

io -403- t tiC e fo. •S.‘ti ,r ,`‘ ,r sP :..') h Ng, ,ES., s. E9.i - 6(s 6'''' ,C;(' cfs ,e • e

Ra -(k 4. cs ‹-,14 EP ik e ' z • N., • 4 , 00 e <,----, -,v, cs-, , -,9 ec' e' 0. ,i, e e <' q

Fig. 1. b: Change in chlorophyll b content of sycamore leaves subjected to different aphicide treatments.

185

oo a) D) CIS 50

7C/ 0 C a) O co _50 0 0 -100 ca

-150 .,, ,,,,,,s, , ..,9c, ..,„. e, ts •a`';, .P • ,,- ..!' &`) 0 er cP' • e e cf .cs s ,,c` q)\S'. e e' cP e e e ' \ .- . „..,0 . cc,,F, . ,e, e .. o. El. \-\\`‘b' \` \-\`?f. o-N (p 1\ ' ' c)(‘ 6\ cfs '0' '0\ '0'. (0 c, rhc, 5, c.c.;\ cf;\ o° 0 0 GCP GC)Cs 06' cy ce C.I• <, cy nJ (0'1ClCl cPc‘ 5,6' <,) cpc eG , ‘h • ‘.._ \ _`h (Cs' e \.. cpcCs e '?.-.cp Treatr Ten

Fig. 1 c: Change in total carotenoid content within sycamore leaves subjected to different aphicide treatments.

Since biochemical analysis did not indicate plant stress, the selection of an appropriate pesticide was based solely on the bioassay data. Given the selection criterion (i.e. short- term eradication of aphid infestation), bioassay results led to the choice of pirimor at

one-third the concentration (ANCOVA: pesticide x concentration x time interaction F12,

380 = 4.19, P<0.001). This is because it eliminated aphids over 12 days after just one initial spray, and thereafter its inhibitory effect gradually diminished (Figs. 2 e). Derris and Malathion at one-third the recommended dosage were also suitable, but their aphidical properties operated over an even shorter time-span (see Figs. 2 a, b, c, d & e).

186 Control 70 l .1/5 Recommended Dosage rva

te 60 N. I. 1/3 Recommended Dosage in

Recommended Dosage

ime 50 t r e

p 40 rn

bo 30 hs mp 20 ny

No. 10

0 02/10/96 07/10/96 12/10M 17/10% 22/10/96 27/10/96 01/11/96 Time (5 day intervals)

Fig. 2 a: The effect of Ambush concentration on the reproduction rate of parthenogenetic virginoparae of Drepanosiphum platanoidis on Acer pseudoplatanus saplings.

Ccrtrcl l

rva 1/5 Reccrrrrendscl

te • • 1/3 REccrrrrEni3d rhsay in •••••Recarrreted &say ime t er p born hs mp ny No.

Tirre (5 day intervals)

Fig. 2 b: The effect of Derris concentration on the reproduction rate of parthenogenetic virginoparae of Drepanosiphum platanoidis on Acer pseudoplatanus saplings.

187 BD

70 l Contrd 1/5 Reccrrmandal ['tray

terva 60

in 1/3 Reaa rimnda_11-1-Ray

50 RLua u I a rld asazfi time

er

p 40 rn bo 30 hs mp 20 ny

No. 10

0 02/10% 07/10S 12/109 17/10% 22/10/% 27/10'96 01/11/96 Tirre (5 day intervals)

Fig. 2 c: The effect of Dimethoate concentration on the reproduction rate of parthenogenetic virginoparae of Drepanosiphum platanoidis on Acer pseudoplatanus saplings.

250 l rva te in ime t r e p rn bo

hs mp ny No.

02/1CV96 07/10/96 12/10/96 17/10/96 22/10/96 27/10/96 01/11/96 Time (5 day intervals)

Fig. 2 d: The effect of Malathion concentration on the reproduction rate of parthenogenetic virginoparae of Drepanosiphum platanoidis on Acer pseudoplatanus saplings.

188 the sycamoresaplings. was effectiveovertheshort-termindeterringaphidinfestation,timingofpesticidere- preferred toaformalisedandregularsprayingregime,therebyonlywhen application dependsonwhenaphidsstarttoalighttheirhost.Thisapproachwas conditions. Althoughpirimicarbatone-thirdtherecommendedconcentration(1.67g/1) The practicalitiesofwhentospraypesticidesinthefield,issubjectweather necessary, tominimiseside-effects(i.e.phyto-toxicorphyto-enhancement)renderedon Drepanosiphum platanoidis Fig. 2e:TheeffectofPirimorconcentrationonthereproductionrateparthenogeneticvirginoparae

No. nymphs born per time interval 140 160 180 100 120 60 80 40 20 0 02/10/96

on Acer pseudoplatanus 09/10/96

Time (5dayintervals) 16/10/96 189 saplings.

23/10/96

30/10/96