Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Factors affecting the mortality of winter moth in the lower mainland of British Columbia Horgan, Finbarr G. 1993

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

Item Metadata


831-ubc_1993_fall_horgan_finbarr.pdf [ 8.22MB ]
JSON: 831-1.0098855.json
JSON-LD: 831-1.0098855-ld.json
RDF/XML (Pretty): 831-1.0098855-rdf.xml
RDF/JSON: 831-1.0098855-rdf.json
Turtle: 831-1.0098855-turtle.txt
N-Triples: 831-1.0098855-rdf-ntriples.txt
Original Record: 831-1.0098855-source.json
Full Text

Full Text

FACTORS AFFECTING THE MORTALITY OF WINTER MOTH INTHE LOWER MAINLAND OF BRITISH COLUMBIA.byFINBARR GABRIEL (Hons.), University College, Cork, 1989A THESIS SUBMITTED IN PARTIAL FULFILMENT OFTHE REQUIREMENTS FOR THE DEGREE OFMASTER OF SCIENCEinTHE FACULTY OF GRADUATE STUDIESDEPARTMENT OF ZOOLOGYWe accept this thesis as conformingto the required standard.THE UNIVERSITY of BRITISH COLUMBIAJune 1993© Finbarr Gabriel Horgan, 1993In presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Columbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives. It is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.(Signature)Department of ^ZOOLOGYThe University of British ColumbiaVancouver, CanadaDate^2nd JULY 1993DE-6 (2/88)ABSTRACTPopulations of winter moth, Operophtera brumata (L.), were monitored at four sitesin Richmond, British Columbia, between 1989 and 1992. Populations peaked in 1990 onboth blueberry and birch and declined with an eventual population crash in 1992.Parasitism by Cyzenis albicans (Fall.) fluctuated between years at each site. Parasitismreached its highest levels on both birch (ca. 55%) and blueberry (ca. 35%) in 1991. Pupalpredation was the most important stage specific mortality factor throughout the four years.In 1992, the year of population crash, larval mortality was high. Trends in "death of pupaedue to unknown causes" were linked to larval mortality and are suggested to result mainlyfrom poor foliage quality. An unusually early spring in 1992, may have led to the observedincreases in these two mortality factors. The incidence of viral or other diseases among thepopulations are low.Generally, pupal predation peaked in 1990 (ca. 90%) and then declined.Pterostichus spp., Amara spp., Harpalus affinus and subsoil beetle larvae are implicated asimportant predators. The levels and trends in predation were similar at sites with verydifferent assemblages and abundances of beetles. There were no differences in theabundances of beetles at each site between 1991 and 1992. However, many of theimportant predatory species declined in 1992. An examination of the possible interactionsbetween C. albi cans and generalist predators is made. The increased range of pupal sizes inthe soil, due to the presence of C. albi cans may be a mechanism for inducing a numericalresponse among generalist predators. However, simultaneous population declines at siteswith low levels of C. albicans indicate that winter moth outbreak and decline in NorthAmerica may be induced by a number of different factors.1 1TABLE OF CONTENTSABSTRACT^ iiTABLE OF CONTENTS^ i iiLIST OF TABLES viLIST OF FIGURES ixACKNOWLEDGEMENTS x i vDEDICATION^ xvCHAPTER 1GENERAL INTRODUCTION1.1 Winter moth^ 11.2 Winter moth parasitoids^ 61.2.1 Cyzenis albicans 71.3 Population dynamics of winter moth^ 101.4 Objectives^ 14CHAPTER 2MORTALITY OF WINTER MOTH IN BIRCH STANDS ANDBLUEBERRY PLOTS IN THE LOWER FRASER VALLEY OFBRITISH COLUMBIA.2.1 Introduction^ 162.2. Procedures 21Study sites 21Sampling procedures^ 23Data Analysis 26Spread of winter moth and C. albi cans^ 282.3 Results^ 282.3.1. Damage to blueberry and birch^ 282.3.2. Population densities at Richmond 35Early instars^ 35Prepupae 37Adults 39Fecundity 392.3.3. Mortality on blueberry and birch^ 42Parasitism^ 42Disease 45Pupal mortality 45Soil mortality 46K-factor analysis^ 47Refugia against parasitism^ 492.3.4 Winter moth and C. albicans spread^ 572.4 Discussion^ 62Parasitism - host plants^ 62Habitat refuge^ 65Larval mortality 67Disease 71Pupal mortality 73CHAPTER 3PREDATION OF WINTER MOTH PUPAE IN THE LOWER FRASERVALLEY OF BRITISH COLUMBIA.3.1 Introduction^ 753.2 Procedures 79Study sites 79Pitfall trapping^ 79Subsoil traps 83Field planting of pupae 85Arena studies 863.3 Results^ 883.3.1 Trends in predation^ 883.3.2 Predator assemblages 903.3.3 Predatory beetles 953.3.4 Subsoil traps 973.3.5 Timing of attack^ 1003.3.6 Predation on different pupal types^ 1003.3.7 Seasonal abundance of predatory beetles 1083.3.8 Annual abundance of predatory beetles 1083.3.9 Distribution of sizes in the winter moth-Cyzenis pupalcomplex^ 1163.4 Discussion 120Generalist predators^ 120Cyzen is-predator interactions^ 128CHAPTER 4GENERAL DISCUSSIONOutbreak decline^ 136Cyzenis-generalist predator link^ 145Continued control of winter moth 149ivLITERATURE CITED^ 153APPENDIX 1^ 166APPENDIX 2 168APPENDIX 3 169APPENDIX 4^ 170APPENDIX 5 172APPENDIX 6 173APPENDIX 7^ 175VLIST OF TABLESTable 2.1.Mean pupal weights of winter moth, Operophtera brumata, reared on apple, birchand blueberry or switched between hosts. 'N' is the number of individuals rearedthrough to pupation. All larvae were taken from wild populations 34Table 2.2.Mean pupal weights of winter moth, Operophtera brumata, from four field sites atRichmond, B.C. during 1991 and 1992 and estimates of fecundity based onpublished relationships between weight and fecundity (eggs/female) 41Table 2.3.Estimates of soil mortality (from emergence traps), and mortality due to pupalpredation (from tethers) at four sites in Richmond for two years at birch sites andone year at blueberry sites. The difference is attributed to mortality of prepupae onthe ground and of adults after emergence and to death of healthy pupae in the soil.All estimates are presented as percentages 47Table 2.4.Winter moth early instar larval densities (per cluster) at Richmond sites during 1991and 1992, with associated levels of parasitism by C. albicans and death ofprepupae due to unknown causes from 1990 to 1992. Host plants are indicated.Brackets indicate standard errors of larval estimates and sample size for parasitismand "unknown" mortality estimates. Asterix indicates that pupae were taken fromdrop trays, otherwise pupae were reared from collected prepupae 58Table 2.5.Winter moth early instar larval densities (per cluster) at Lower Mainland sitesduring 1992, with associated levels of parasitism by C. albicans and death ofprepupae due to unknown causes from 1990 to 1992. Host plants are indicated.Brackets indicate standard errors of larval estimates and sample size for parasitismand "unknown" mortality estimates 59Table 2.6.Winter moth early instar larval densities (per cluster) at Vancouver sites during1991 and 1992, with associated levels of parasitism by C. albicans and death ofprepupae due to unknown causes from 1990 to 1992. Host plants are indicated.Brackets indicate standard errors of larval estimates and sample size for parasitismand "unknown" mortality estimates 60Table 3.1.Habitat characteristics at nine sites in Richmond. Studies on winter moth pupalpredation have been carried out at these sites between 1989 and 1992. Densities ofbirch and hemlock are presented as the number of plants per 50m 2 . Percentagev icover of salal, labrador tea and heather are also presented. For undergrowthblueberry, percentage cover is presented while for blueberry plots numbers ofplants per 50m2 are presented (indicated by an asterix). Ten randomly chosen plotsof 50m2 were sampled at each site^ 81Table 3.2.Subsoil trap types used in experiments at Richmond in 1991 and 1992^ 83Table 3.3.Predatory beetles trapped at four sites in Richmond B.C. during 1991 and at sixsites in 1992. The Shannon-Weiver Diversity Index (H') and Species Dominance(14, where J' is Pielou's Evenness Index), are presented for each site (Ludwigand Reynolds 1988). B.L.'s are beetle larvae and S's are staphylinids  91Table 3.4.Ground beetle predators of the winter moth-Cyzenis pupal complex. `+' indicatesthat predation was observed, `-' indicates predation did not occur and `n.t.'indicates that no trials were undertaken. Naked pupae are pupae with thehibernaculum removed 96Table 3.5.Mean numbers of ants and beetle larvae caught in exclusion traps during thesummer of 1991. Three trap types were used, unbaited (control), baited 0.1mmand baited lmm subsoil traps. Results are from ten traps at each site 99Table 3.6.Mean numbers of ants and beetle larvae caught in exclusion traps during thesummer of 1992. Four trap types were used, unbaited (control), Cyzenis baited(fly), winter moth baited (pupa) and hibernaculum baited (cocoon) lmm subsoiltraps. Results are from five traps at each site 99Table 3.7.Changes in the proportions of different beetle species in 1991 and 1992, at fourfield sites in Richmond. 'Total numbers' indicates the numbers of beetles caught inpitfall traps between late June and early September of those years. PT =Pterostichus spp., CG = C. granulatus, CN = C. nemoralis, MS = mediumstaphylinids, AA = A. aurata, AL = A. littoralis, BLC = Carabus spp. larvae, BL =beetle larvae, HA = H. affinus, CF = Calathus fuscipes and HR = H. rufipes 117Table 3.8.Weights (+ standard errors) of winter moth pupae at two birch sites in 1991 and1992 and at two blueberry sites in 1991, with corresponding Cyzenis pupal weightsat each site  119viiTable 3.9.Proportions of pupae in different size categories from the winter moth-Cyzenispupal complex at Richmond, with suggested proportions for the same populationsin the absence of parasitism (i.e. taking only the proportions of winter moth pupaeof each size category from each sample 122Table 4.1.Synopsis of characteristics of winter moth outbreaks and the factors attributed totheir declines^  137Table 4.2.Information on outbreaks of winter moth from seven studies, with information onthe success of Cyzenis albicans in parasitizing of the populations^ 139Table 4.3.History of outbreaks of the winter moth from seven regions, with indications of thelevels of parasitism (mainly due to Cyzenis) and soil mortality presented aspercentages and k-values. Corresponding densities of winter moth are alsoindicated. 'H' indicates highest levels, 'I,' indicates lowest levels and '0' indicatesoutbreak populations. An asterix indicates that values for pupal predation arepresented rather than soil mortality  142viiiLIST OF FIGURESFigure 1.World wide distribution of the winter moth, Operophtera brumata (L.). Datesindicate probable introduction events into North America; to Nova Scotia in the1930's and inset, to the western United States in the 1950's, Vancouver Island inthe 1970's and the Lower Fraser Valley in the 1980's (see also Chapter 2) 3Figure 2.1.Study sites in the lower mainland. Solid squares indicate sites used in life tablestudies, open squares indicate winter moth sampling sites. Vancouver sites (1-8); 1.University of British Columbia, 2. West 9th and Alma, 3. West 13th and Trimble,4. Spanish Banks, 5. Locarno Beach, 6. Jericho beach, 7. West 35th andGranville, 8. West 61st and Granville. Burnaby sites (9-10); 9. Burnaby lake, 10.Simon Fraser University. Richmond Sites (11-17); 11. Knight street, 12.Department of National Defense lands, 13. Richmond Nature Park, 14. Birch standI, 15. Blueberry plot I, 16. Blueberry plot II, 17. Blueberry plot III. WesthamIsland (18) and Delta sites (19-20); 19. Deas Island, 20. 112t Street (north). Tensampling sites at Ladner and Chilliwack are not shown.^ 22Figure 2.2.Stages in the development of highgrowth blueberry (Vaccinium corymbosum)flower clusters (a) and leaf clusters (b). See text for details^ 29Figure 2.3.Stages in the development of birch leaves (Betula papyrifera). See text for details ^ 32Figure 2.4.Densities of early instar larvae per branch in 1989 and 1990 (a), and per bud in1991 and 1992 (b), at four sites (open circles = BBI, open squares = BBII, closedcircles = BI and closed squares = RNP) in Richmond, B.C.. All standard errors areincluded but are negligible 36Figure 2.5.Densities of prepupae per square meter at (a) birch sites (dark shading = BI, lightshading = RNP) and (b) blueberry sites (dark shading = BBI, light shading =BBII) at Richmond, B.C. bars indicate standard errors 38Figure 2.6.Densities of healthy winter moth pupae from 1989 to 1992 with correspondingadult densities from 1991 and 1992 at (a) two birch sites (dark shading = BI, lightshading = RNP) and (b) two blueberry sites (dark shading = BBI, light shading =BBII) at Richmond, B.C., bars indicate standard errors 40ixFigure 2.7.Mortality of winter moth at birch sites in Richmond. Closed circles indicate pupalpredation (kpup) estimated from tether experiments, open circles indicate mortalitydue to C. aibicans (kpara). Closed triangles indicate larval mortality (kiarv) andopen triangles indicate death of prepupae due to unknown causes (kprepu), data arefrom BI, and RNP. Total generation mortality (K) is also presented 43Figure 2.8.Mortality of winter moth at blueberry sites in Richmond. Closed circles indicatepupal predation (kpup) as estimated from tether experiments, open circles indicatemortality due to C. albi cans (kpara). Closed triangles indicate larval mortality(kiarv) and open triangles indicate death of prepupae due to unknown causes(kprepu), solid squares indicate total generation mortality (K). Data are from BBIand from BBII 44Figure 2.9.K-values of pupal predation plotted against the total densities of pupae in the soil.Closed circles indicate BI, closed triangles indicate RNP, open circles indicate BBIIand open triangles indicate BBI. Trends indicate the occurrence of weak temporaldensity dependence at each site 48Figure 2.10.K-values of C. albicans parasitism plotted against the total densities of pupae in thesoil, an indicator of prepupal densities. Closed circles indicate BI. Closed trianglesindicate RNP, open circles indicate BBII and open triangles indicate BBI. Trendsindicate a weak delayed density dependence at the blueberry sites, but no trends areapparent for birch sites 50Figure 2.11.Mortality of prepupae due to C. albicans (K parasitism) plotted against prepupaldensities per leaf cluster as estimated from total pupae in drop trays divided by theaverage number of buds per unit area, estimated visually, at each site. Resultsindicate disproportionately low levels of parasitism on blueberry (open circles =BBI, open squares = BBII, closed squares = BI and closed circles = RNP) 51Figure 2.12.Percentage of winter moth larvae at each instar a) on blueberry at three samplingdates and b) percentages of winter moth larvae in birch sites at each instar over foursampling dates. Solid bars indicate larvae in the canopy, shaded bars indicate larvaein the undergrowth. Standard errors are included. There is no significant differencein the development of larvae on blueberry or birch or in either canopy orundergrowth.  53Figure 2.13.Occurrence of winter moth larvae on flower or leaf clusters at three blueberry sitesin 1992. Open circles indicate densities on flower clusters and closed circlesindicate densities on leaves. Error bars are included. There is an apparent shift fromforaging on flowers to foraging on leaves as the season progresses. Data was log+1transformed and analyzed with a two sample t-test; * = P = 0.05, ** = P = 0.005 54Figure 2.14.Progression of damage to birch leaves at four different densities of winter mothlarvae over three sampling dates. Densities of early instar larvae per cluster arepresented at the top. Bars indicate the proportion of leaves in each damage class,with associated standard errors. Leaves with one hundred percent damage generallyhad died as a result of larval attack  55Figure 2.15.Highest incidences of parasitism by C. albi cans encountered at seven regions in thelower mainland, over the four year study period. The year in which parasitism washighest is presented in brackets. Birch was the host plant at each site except atRichmond where blueberry was also sampled (as indicated), and Westham Islandwhere both birch and crabapple were sampled 61Figure 3.1.Predation study sites at Richmond Open circles indicate bog sites, open squaresindicate blueberry sites and closed squares indicate birch sites. The years in whichstudies were conducted at each site are presented. Highway 99 is indicated by anencircled'99'  80Figure 3.2.Subsoil predator trap used in predation studies at Richmond B.C. ^ 84Figure 3.3.Beetle exclosure used in predation studies at Richmond B.C.^ 87Figure 3.4.Levels of predation on the winter moth-Cyzenis complex. Predation estimated fromtethered pupae at a) Three blueberry sites, b) two bog sites (squares) and two birchsites (circles). Bars indicate standard errors 89Figure 3.5.Cluster diagram (average linkage) of six sites based on the similarities of theirpredatory beetle faunas. Pitfall catches are from May 14 until December 3, 1992.Note the similarity coefficients do not exceed 0.05.  94Figure 3.6.Predation of tethered pupae at a) three blueberry sites, b) two bog sites and c) twobirch stands in Richmond B.C., during the summer of 1990. Bars indicate standarderr  01xiFigure 3.7.Cumulative predation of pupae at two birch sites a) RNP and b) BI, during thesummer of 1992. Open circles indicate predation of small Cyzenis pupae (<0.01g),open squares indicate large Cyzenis (0.01-0.020, closed squares indicate smallwinter moth pupae (0.01-0.02g) and closed circles indicate predation on largewinter moth pupae (0.02-0.03g). Bars indicate standard errors 102Figure 3.8.Cumulative predation of pupae at two blueberry sites, a) BBI and b) BBII, duringthe summer of 1992. Open circles indicate predation of small Cyzenis pupae(<0.01g) and closed circles indicate predation on large winter moth pupae (0.02-0.03g). Bars indicate standard errors. 103Figure 3.9.Comparisons of the weekly predation levels of small Cyzenis pupae (open circles)and large winter moth pupae (closed circles) at two blueberry sites, a) BBI and b)BBII in Richmond B.C. 105Figure 3.10.Comparisons of the weekly predation levels of small Cyzenis pupae (opentriangles), large Cyzenis pupae (open circles), small winter moth pupae (closedtriangles) and large winter moth pupae (closed circles) at two birch sites, a) RNPand b) BI in Richmond B.0 106Figure 3.11.Seasonal abundance of Four predators at BBI, a) A. littoralis, b) A. aurata, c) H.affinus and d) beetle larvae from pitfall traps. Solid circles indicate abundances in1991, open circles indicate abundances in 1992 and bars indicate standard errors 109Figure 3.12.Seasonal abundances of four predators at BBII, a) C. granulatus, b) P.melanarius,c) C. nemoralis and d) Carabus spp. beetle larvae from pitfall traps. Solid circlesindicate abundances in 1991, open circles indicate abundances in 1992 and barsindicate standard errors  110Figure 3.13.Seasonal abundances of Pterostichus spp. at a) BI, b) BIa, c) RNP, and d) atRNPI, with abundances of C. granulatus at e) BI and 0 BIa. Solid circles indicateabundances in 1991, open circles indicate abundances in 1992. Bars indicatestandard errors  111Figure 3.14.Summer abundance of predatory beetles at four filed sites in Richmond. Solid barsare abundances of beetles per trap in 1991, shaded bars are abundances of beetlesper trap in 1992. Standard errors are presented  113xiiFigure 3.15.Changes in the abundances of a) Pterostichus spp., b) C. nemoralis, c) C.granulatus, d) staphylinids, e) Carabus spp. larvae and f) beetle larvae excludingCarabus spp.. Solid bars are 1991 abundances, shaded bars are 1992 abundances.Standard errors are presented.  114Figure 3.16.Changes in the abundance of three predators at BBI in Richmond, between 1991and 1992^ 115Figure 3.17.Distribution of pupal weights in the winter moth-Cyzenis pupal complex atRichmond. Key to plot: 1 = BI 1991 winter moth, 2 = BI 1991 Cyzenis, 3 = RNP1991 winter moth, 4 = RNP 1991 Cyzenis, 5 = BBI 1991 winter moth, 6 = BBI1991 Cyzenis, 7 = BBII 1991 winter moth, 8 = BBII 1991 Cyzenis, 9 = BI 1992winter moth, 10 = BI 1992 Cyzenis, 11 = RNP 1992 winter moth, 12 = RNP 1992Cyzenis. Asterix indicate outlying points, circles indicate far outlying points  118Figure 3.18.Effects of parasitism by Cyzenis on the overall sizes of pupae in the winter moth-Cyzenis pupal complex at Richmond at a) BI, b) RNP and c) at BBII in 1991.Black bars indicate Cyzenis pupae, grey bars indicate winter moth pupae. 121ACKNOWLEDGEMENTSI am glad to have this opportunity to thank the many people who have helpedwith various aspects of this dissertation. Firstly, I thank Dr. Judy Myers forintroducing me to the winter moth and for her support and supervision throughoutthe study. Thanks to my parents and family who have always been a constantsource of support and encouragement. Thanks to all my committee members; DrSheila Fitzpatrick, Dr. Charles Krebs and Dr. Goeffrey Scudder.This study would not have been possible but for a number of people whoallowed access to their properties; thanks to Mr. Watanabe, Mrs. Schultz, Mr.Peters, Mr. Edwards and Mr. Shaw. Thanks also to the staff at Richmond NaturePark, Reifel Bird Sanctuary and the Department of National Defence at Richmond.Thanks to Mrs. Rosy van Meel who began collecting data in 1989 and helped me toget started in 1991. Thanks also to Dr. Jens Roland for forwarding manuscripts andfor helpful discussions.A number of good friends deserve special mention for helping to makefieldwork so much fun; thanks to Connie Hrymack, Anna Lindholm, NaomiRichardson, Alain Drouin, Rafi Balien, John Markham, Susan Senger, NickHorgan and Galib Rayani. I am grateful to a number of friends for advice anddiscussion and for helping with various drafts of the thesis; thanks to Dr. CarlosGalindo-Leal , Dr. Laura Lazo, Kathy Craig, Ron Saimoto, Regina Saimoto,Jordan Rosenfel, Lorne Rothman and especially Naomi Richardson and SamirAouadi. Thanks to Alister Blachford for all things computorial. I also wish to thankDr. Denis Chitty and Dr. Jamie Smith for words of encouragement that were muchappreciated. I am very grateful to the many people involved with InternationalHouse and the Department of Zoology, who have helped make my time inVancouver memorable. This thesis is dedicated to the street children of Bogota andalso to my parents.xivThis thesis is dedicated to thestreet kids of Bogota.X VCHAPTER 1GENERAL INTRODUCTIONOne of the most frequently cited examples for the success of classical biologicalcontrol is that of the winter moth, Operophtera brumata (L.), in Canada, controlled by itsparasitoid, Cyzenis albicans (Fall.) (Embree 1971, DeBach 1974, Hassell 1978, Embreeand Otvos 1984, Murdoch a al. 1986). There has been extensive monitoring of both themoth and parasitoid populations in eastern Canada in the 1950's and 1960's (Cuming1961, Embree 1965b, 1966), and in western Canada in the 1980's (Roland 1986a, 1986b,1988, 1992). Long-term population studies have also been carried out on endemicpopulations in Britain (Hassell 1969a, 1969b, Varley et al. 1973). Because of this, thesetwo species have formed a basis for much of the development of modern ideas inpopulation dynamics theory (Varley and Gradwell 1968, 1970, 1973, Varley et al. 1973)and host-parasitoid interactions (Hassell 1968, 1980, 1987, Hassell and May 1974, May1978). In this study, examination of the mortality factors acting on a new post-introductionoutbreaking population of winter moth in Richmond, British Columbia, will be made in thelight of the above information. In particular, I address the question of why C. albicans hasbeen necessary to bring about winter moth population decline in Canada.This first chapter offers a brief introduction to the winter moth and its parasitoid andpresents a background for understanding winter moth population dynamics.1.1 Winter mothDistribution: The winter moth, 0. brumata, is one of five old world Operophteraspecies. It is distributed throughout central Europe, east to the Volga and the Caspian Seaand in the south to the Caucus and the Black Sea. The species also occurs on Iceland, the1Faroe Islands, Great Britain and Ireland. It ranges from Northern Scandinavia (thoughabsent from the extreme north) to the Mediterranean. Winter moth also occurs around theAmur, in Ussuri in southeastern Siberia, and on Hondo in Japan (Fig. 1) (Tenow 1972,Skou 1986). The main areas of mass occurrence of winter moth include regions borderingthe North Sea and the Baltic, southern England, northern France, Holland, Belgium,northern Germany, Poland, the Baltic states, Denmark and Sweden (Tenow 1972).Winter moth was recently introduced to the Faroe Islands, either from the BritishIsles or northern Europe (there is no phytosanitary control of material brought fromDenmark, Scotland, Norway and Iceland). Interestingly, the moth does not exist on someof the larger southern islands of the Faroes (i.e. Sandoy and Suduroy) which are oppositeprevailing winds (Koponen 1985). Winter moth has also recently become established inNorth America on both the east and west coasts (Hawboldt and Cuming 1950, Smith 1950,Ferguson 1978, Gillespie et al. 1978); this will be addressed in more detail in Chapter 2.The life cycle: Operophtera brumata is one of a small number of geometrids whoseadults are adapted for winter existence. The male has a wingspan of about 23 - 31mm. Thewings are grayish brown, the forewing having numerous dark, wavy, transverse lines. Thehind wings have one or two indistinct transverse lines though these may sometimes belacking. There may be colouration differences between populations (pers. obs.). Femalesare brachypterous with very short wing remnants of about 2mm in length. The body isgrayish, brownish or grayish black. The wings are steel gray with a black transverse bandnear the termen.The adults usually emerge in October and November, though they may emergeearlier, (i.e. September) in more northerly regions and later, (i.e. December) during mildwinters or in southern regions (see Wylie, 1960a). Alma (1969) describes male flight2Figure 1. World wide distribution of the winter moth, Operophtera brumata(L.). Dates indicate probable introduction events into North America; to NovaScotia in the 1930's and inset, to the western United States in the 1950's,Vancouver Island in the 1970's and the Lower Fraser Valley in the 1980's (seealso Chapter 2).3activity in Britain and suggests a temperature threshold for flight at between 5 and 5.5°C.Females usually emerge about a week after the males. After the females emerge, they climbadjacent trees and mate during this ascent, usually at about 1 - 2 meters from the ground.Female sex pheromone attracts males and operates at unusually low temperatures (4 -150C), exhibiting an upper response limit which coincides with the lower response limit ofother reported moth sex pheromone systems (usually 15 - 17°C) (Roelof et al. 1982).Females lay up to 300 eggs (but generally 70- 200 eggs) in cracks on the bark, andunder lichens and mosses. The eggs are laid singly or in small groups and are ovoid, butsomewhat flat, 0.7 - 0.8mm long and 0.5mm wide. The surface of the egg is reticulated byminute ridges separated by shallow pits. Newly laid eggs are light green in colour, laterchanging to orange and finally deep brown or black before larval eclosion. Eggs overwinterwith a suggested developmental threshold of about 3.9 - 4°C (Embree 1970, Kimberlingand Miller 1988). The time for egg development depends on the region. In southern Italythe egg stage lasts only about two months whereas in northern Europe it lasts about eightmonths. These differences can be partially attributed to temperature differences betweenregions, but there are also intrinsic differences between the populations (Wylie 1960a,Holliday 1985). Authors disagree on whether there is a diapause, but there again appear tobe differences between regions (Wylie 1960a, Bonnemaison 1971, Holiday 1985).Kimberling and Miller (1988) suggest that responses to chilling of winter moth eggsindicate that there is a diapause even where other authors had discounted it (see also Tauberand Tauber 1976). The eggs are extremely cold tolerant with supercooling points estimatedto be between -310C (Macphee 1967, Hale 1989) and -37°C (Hale 1989).In British Columbia larval eclosion is from mid-March till April. There are fivelarval instars. The first instar is the main dispersal stage (some second instars may alsodisperse). Dispersal is by means of parachuting on a silk thread. The degree and distance of4dispersion depends on temperature and climatic conditions as well as the condition of thehost plant foliage on larval eclosion (Edland 1971, Holliday 1977). The larval stage lastsfrom March till May or June and the larvae then spin to the ground to pupate.Pupation occurs about 1 - 2 cm below the soil surface where a cocoon is spun. Thelength of pupation varies between regions. In Italy, for example, pupation can last for eightmonths while in northern Europe it lasts for about three months, so there are intrinsicpopulation differences (Wylie 1960a, Holliday 1983). In British Columbia, pupation isfrom early to mid-May with adult emergence in mid November, a period of about sixmonths.Host plants and damage: Although the larvae are largely monophagous, the species ispolyphagous primarily on deciduous but also on coniferous trees (Wint 1983). Mrkva (inWint 1983) indicated that winter moth feeds on about 160 species of deciduous trees andshrubs. Wint (1983) indicates that the species is known to feed on plants from 14 differentplant families.It is generally accepted that oak (Quercus spp.) is the primary host (Varley andGradwell 1958, Feeny 1970, Hunter 1990). Other important hosts include: apple, (Malusspp.), pears (Pyrus spp.), plums and blackthorn (prunus spp.), hazel (Corylus avellanaeL.), beech (Fagus sylvatica L.), hawthorn (Crataegus mono gyna Jacq.), etc. (Wint 1983).In southern Europe beech trees are an important host, while in central Europe, willows(Salix spp.) are more important (Topp and Kirsten 1991). In North America, importanthosts include: oak, apple, birch (Betula spp.), highbush blueberries (Vacciniumcorymbosum L.), raspberries (Rubus spp.) and commercial filberts (Corylus spp.)(Embree 1965b, Gillespie et al. 1978, AliNiazee 1986, Fitzpatrick et al. 1991a).5In Scotland, considerable attention has recently been directed to winter moth as apest of Sitka spruce, (Picea sitchensis (Bong) Carr) (Stoaldey 1985, Hunter et al. 1991).Previously populations built up on birch or on heather (Calluna vulgaris L.) and thenswitched to the Sitka spruce. Now however, it appears that the outbreaks build up directlyon the spruce. The larvae cause substantial damage and distorted growth of the spruce trees(Watt and Macfarlane 1991).1.2 Winter moth parasitoidsA number of parasitoids have been recorded from the eggs, larvae and pupae ofwinter moth in Europe (see Sechser 1970a, 1970b, Wylie 1960b). On the two occasionswhen winter moth was introduced to Canada it went into post-introductory outbreak(Embree 1965b, Gillespie et al. 1978). This implicated natural enemies as responsible formaintaining population equilibrium in Europe, and so parasitoids were introduced fromEurope to Canada in an attempt at biological control. Only two of the introducedparasitoids, Cyzenis albicans (Fall.) and Agryponflaveolatum (Gray.), became established(Cuming 1961, Roland 1988).The extent of native parasitoids attacking winter moth in Nova Scotia, beforeintroduced parasitoids became established, is unknown. However, in British Columbia, anumber of native parasitoids, normally parasitic on Bruce's spanworm, Operophterabruceata (Hu1st), were recorded from winter moth. The most important was an unnamedPhobocampe sp. (Ichneumonidae) (Gillespie and Finlayson 1981, Humble 1984). Threefurther species, Trichistus crassus Tow. and Tow., Agrypon provancheri (Dalla Torre) andCyzenis pullala (Tsnd.), also parasitised winter moth. By 1981, there was ca. 1.5%parasitism of winter moth larvae, mainly attributed to these native parasitoids, although C.albi cans was already present in the population at that time (Humble 1985b).6A number of native pupal parasitoids have also been recorded from winter moth inBritish Columbia. Coccugomimus hesperus Tow. and Tow. and Buathra dorsicarinata(Pratt) were found to parasitise 4% of the winter moth-spanworm pupal complex inVictoria in 1981, and 10% in 1982. These species most commonly attacked winter moth. Afurther species, Cratichneumon sp., preferred Bruce's spanworm (Humble 1985a).Adults of C. pullala and C. albi cans are almost morphologically indistinguishable.Furthermore, A. provancheri is morphologically very similar to A. flaveolatum. Thesewere once regarded as synonymous species (Gillespie and Finlayson 1981, Barron 1989).These species similarities may cause some difficulty in assessing the success of thebiological control. However, it is apparent that the native parasitoids were not capable ofpreventing winter moth outbreak, or of causing outbreak decline, since it was not until C.albicans and A. flaveolatum became established that outbreak decline occurred andpopulation stability was reached.1.2.1 Cyzenis albicansCyzenis albicans has largely been implicated in the successful biological control ofwinter moth in Canada (Hassell 1980, Embree 1991). This Tachinid fly is widelydistributed over the range of the winter moth. It is found from central Sweden to centralSwitzerland but probably does not occur in the more southern regions of winter mothdistribution (i.e. Sicily) (Wylie 1960b). The distribution of C. albi cans in Japan and easternAsia has not been discribed.The life cycle: The life cycle of C. albicans is synchronized with that of the wintermoth. Intrinsic differences between regions in the timing of adult emergence correspond7with winter moth larval hatch in those regions (Wylie 1960b). Adult females emergeapproximately seven days after males. Mating occurs during daylight. Following mating,the adult females feed on flowers, sap fluxes and honey due for about four weeks (Embreeand Sisojevic 1965). By the end of that time the winter moth are in their fifth instar and theparasitoid females have their full complement of eggs, about 1000-2000 (Hassell 1980).Laying occurs over a period of 8 - 16 days. In the field, maximum egg populations occur 7- 14 days after the beginning of oviposition.Shiny black microtype eggs of approximately 0.17mm in length are laid on foliagenear caterpillar feeding damage. Eggs are concentrated on the lower portions of the trees(Embree and Sisojevic 1965). Embree and Sisojevic (1965) found ca. 80% of eggs are laidimmediately beside caterpillar feeding damage and predominantly on the underside ofleaves (see also Wylie 1960b). Because of the mode of oviposition there is a significantlyclumped distribution of eggs. Parasitism is correlated with host density, where hostdensities are high (Embree and Sisojevic 1965, Hassell 1980). Cyzenis albicans respondsto damage from a number of defoliators on different host plants. Parasitism may be moreefficient on oak (Embree 1966, Roland 1986a, 1986b, 1990a), but Hassell (1980) foundno difference between the levels of parasitism on oak and those on hawthorn, Crataegusoxyacanthoides Thuill., blackthorn, Prunus spinosa (L.), or hazel, Corylus avellana (L.).However, C. albi cans is more successful in parasitizing winter moth on oak than on apple(Malus spp.) (Embree 1966, Holliday 1977, Roland 1986a, 1986b, 1990a).Wylie (1960b) and Hassell (1980) have described the mode of parasitism in detail.The eggs are viable for about eight weeks. Each egg contains a fully formed first instarlarva. The larva is ingested and hatches in the mesenteron of the host, penetrates the gutwall and lodges in the salivary gland causing a swelling of the surrounding tissue. Thisfirst instar lasts as long as it takes the host to pupate. There is no molting until after host8pupation. After host pupation, the first instar larva makes a respiratory funnel that opens tothe exterior on the ventral side of the pupa near the thoracic-abdominal junction. It feedswith its posterior spiracles inserted in the funnel until only the empty pupal case remains.When mature the parasite larva withdraws its spiracles from the funnel, reverses itself andforms a puparium inside the empty host pupa. The pupa goes into diapause and adultsemerge the following spring.Cyzenis albi cans is known to develop to maturity from Triphosa dubitata (L.) andEupithecia pimpinellota (Hbn.) in the wild, and has been reared to maturity fromOperophtera fagata (Scharf.), Rhyacionia buoliana (Schiff) (Olethreutidae) and Galleriamellonella (L.) (Pyralidae) in laboratory studies (Wylie 1960a, Embree and Sisojevic1965). Embree (1965a) estimated that 27% of Bruce's spanworm in Nova Scotia wereparasitized by C. albicans. However, a number of common defoliators will not support C.albicans to maturity, i.e. Cosmia trapezina (L.), Erannis defoliana (Clerk.), Alsophilapometoria (Han.), Malacosoma americanum (F.), and Pseudexentera cressonianaClem.(Embree and Sisojevic 1965, Hassell 1969a).Mortality: Mortality of adult C. albicans has not been studied in detail. Adults areprobably eaten by a number of bird species. The parasite Perilampus ruficornis (Fab.) anda fungus Entomophthora muscae (F.) Fris. also attack adults but the extent of this in natureis unknown (Wylie 1960b). Hassell (1969a) indicated that the key factor in C. albicanspopulation dynamics was total egg mortality. This results from the high reproductivepotential of the adults with resultant low levels of host parasitization. Synchronization withthe host is important in determining the extent of this mortality. To survive, an egg must beswallowed by a winter moth larva or other suitable host, each of which supports only asingle parasitoid to pupation. Eggs laid after host pupation will not survive. Likewise, eggsconsumed by second or third instars generally rupture and do not survive. Furthermore,9consumption of the eggs by defoliators, other than suitable hosts, represents a considerablesource of mortality. Of the total number of parasitoid eggs found on foliage at Oak Hill,Nova Scotia in 1962, 16% occurred beside feeding damage caused by P. crossonionia.This resulted in a reduced effectiveness of C. albicans (Embree and Sisojevic 1965).Pupal death is also an important source of mortality. In general soil mortality of C.albi cans at Whytham Wood, was greater than the corresponding soil mortality of wintermoth. Hyperparasitism is a significant source of soil mortality for C. albicans. In Britain,Phygadeuom dumetorum (Gray.) has been implicated in causing significant mortality of C.albicans in some years (Hassell 1969a). In Victoria, two native hyperparasites,Phygadeuon sp. and Villa (Hemipenthes) catulinia Coq., were parasitizing up to 12% of C.albicans in 1981 (Humble 1985b). However, the main cause of soil mortality is pupalpredation (Hassell 1969a); this will be discussed in more detail in subsequent chapters.1.3 Population dynamics of winter mothThe most extensive documentation of winter moth population dynamics has beencompiled by Tenow (1972). He collected records from various sources dating back to the1890's and indicated that winter moth in the Scandes go through outbreaks ofapproximately 9 - 10 year periodicity. These outbreaks were associated with outbreaks ofanother Geometrid species, the autumnal moth Oporinia autumnata Bich..In the 1960's Varley and Gradwell (1968, 1970) compiled life tables of a wintermoth population at Whytham Wood near Oxford, England. Analysis of these life tablesindicated two important mortality factors. The first was termed 'winter mortality' and wasfound to be the key factor controlling the population. This factor included all mortalityincurred from egg laying through to pupation and thus consisted of mortalities from larval1 0predation, starvation, dispersal losses and viral or microsporidian diseases. However, theyattributed most of the winter mortality to asynchronies between bud burst and egg hatch.The second mortality factor of importance was 'soil mortality'. This factor was attributedmainly to pupal predation and was found to be density dependent thus representing amechanism of population regulationl (Varley and Gradwell 1960, 1963a, 1963b, 1968,1971, 1973, Varley et al. 1973).When winter moth was first discovered in eastern Canada, a number of parasitoidswere introduced to control it. The Tachinid parasitoid, Cyzenis albicans (Fall.) was foundto be highly efficient at parasitizing larvae at high populations while an introducedIchneumonid species, Agrypon flaveolatum (Gray.) was more efficient at low winter mothpopulations (Embree 1965b, 1966, 1971). This result was surprising, since in NovaScotia, C. albicans was parasitizing up to 80% of the winter moth larvae on oaks, while inEngland parasitism was more usually at about 2% (see Roland 1988). Pupal predationreceived little consideration in Nova Scotia. Small mammals were suggested as beingresponsible for about 30% of predation (Embree 1965b).When winter moth was discovered in western Canada, the two parasitoids werequickly released (Williamson 1981a, 1981b, 1981c, 1984, 1986, and see details in Chapter2). As in Nova Scotia, parasitism increased to high levels, about 70%, after which thewinter moth population began to decline. However, Roland (1986a, 1988, 1990b)suggested that it was not C. albi cans that was actually responsible for the decline, but thatincreased death of pupae in the soil initially caused the population reduction. Onreanalysing Embree's data from Nova Scotia, Roland (1986a, 1988) found that EmbreeI Recently there has been much debate concerning the regulatory role of pupal predation. DenBoer (1986, 1988) maintains that density dependence is not a prerequisite of populationregulation and therefore, that mortality of winter moth pupae is not regulatory. Others e.g.Latto and Hassell (1987) and Poethke and Kirchberg (1987) have argued against den Boer. Inthis thesis I will regard pupal mortality as regulatory.11had overestimated mortality due to C. albicans by including in his mortality estimates, deathdue to nematodes and other parasites. When these other parasites were excluded fromanalyses, it appears that mortality due to parasitism by C. albicans did not increaseproportionately to the total mortality. Furthermore, Roland (1988) highlighted the fact thatdecreases in defoliation had been similar in areas with very different incidences ofparasitism. He therefore suggested that soil mortality had, in fact, been more important thanparasitism due to C. albicans in bringing about winter moth population decline in NovaScotia.Roland (1986a, 1988, 1990b) has suggested that the parasitoid may have beenimportant in instigating the decline of winter moth by inducing soil mortality in both NovaScotia and British Columbia. In Nova Scotia, winter moth populations were high for about25 - 26 years before the parasitoids were introduced. About four years after theintroduction parasites increased and spread, and the decline in winter moth commenced. InVictoria a similar pattern emerged. Winter moth was present at high levels, for only about 6years before being identified (it was initially mistaken for the native Bruce's spanworm,Operophtera bruceata (Hulst)) and soon after its parasitoids were released. Four yearslater, population reduction began to be noticed. Could C. albicans have caused thesedeclines through interactions with generalist predators?Larvae in Nova Scotia were found to be free of viruses before the introduction ofthe parasitoid, but a virus was found soon after and spread throughout the population(Embree 1966). In Victoria, in 1978, virus was used in experiments to control winter mothand in the hope of starting an epizootic. The virus was found to control the moth to a smallextent, but an epizootic did not occur (Cunningham eta!. 1981). This may have been due toa lack of parasitoids to transmit the virus. Perhaps if the experiment were tried again in thepresence of parasitoids the results would be different.12Roland (1986a, 1988) proposed three mechanisms by which parasitoids couldincrease soil mortality, two of which refer directly to pupal predation. The first mechanismwould cause a numerical response in predators. Since parasitoid pupae are in the groundfrom early summer through to the following spring while healthy winter moth pupaeemerge in winter. Therefore, pupae are available in the soil for an extended period enablingpredator populations to build up over winter. A second mechanism suggests that predatorspreferentially take winter moth pupae over parasitoid pupae, and thus as the population ofavailable healthy pupae in the soil diminishes due to parasitism, the remaining healthypupae become increasingly susceptible to predation. A third suggested mechanism is thatparasitoids increase the transmission of virus or disease causing spores (fungal ormicrosporidian) among larvae. The larvae pupate, but die as pupae thus increasing soilmortality. These mechanisms are not necessarily exclusive.Interestingly, in Canada, a similar Operophtera species, Bruce's spanworm, oftenoccurs at the same sites as winter moth (spanworm is described in Chapter 4). Spanwormhas not received as much attention as winter moth. However, examination of reports fromthe Forest Insect and Disease Survey indicate that population outbreaks do occur. Theseoutbreaks last for similar periods of time to those of the winter moth. Generally, outbreakdecline in spanworm has been attributed to viruses or ground predators (CanadianDepartment of Forestry 1964, 1974, Embree 1966). If spanworm outbreaks are generallycontrolled by generalist ground predators, it is surprising that these are not able to suppresswinter moth outbreaks as well. There may be differences in the interactions betweenpredators in the soil depending on whether or not the prey pupae have parasitoids andviruses. Without these additional mortality agents, soil predators may be insufficient tosuppress outbreaks.131.4 ObjectivesWinter moth was first identified in the lower mainland of British Columbia in 1985.Severe defoliation began to be noticed in 1988 along Highway 99 and at Richmond NaturePark (Wood and van Sickle 1985, 1990, 1991). Losses of blueberries began to occur in thelower mainland due to winter moth infestations (Fitzpatrick et al. 1991a). Although themoth probably arrived from Vancouver Island, I regard this as a second introduction intowestern Canada because of the occurrence of a post-introductory outbreak. However, thereare a few important differences with this introduction. Firstly, the main host plants aredifferent, with birch and blueberry being the major available hosts in the Fraser Valley.Secondly, although it is unknown how the moth arrived here, it is likely that transportationof pupae occurred, since the parasitoid C. albi cans is present, with no record of release.The objective of this research was to investigate the mortalities influencing wintermoth populations in the lower mainland. Winter moth populations at two birch sites andtwo blueberry sites have been monitored since 1989 so that winter moth populationdensities, levels of predation and parasitism at these sites is available for four years. In thisthesis, I will therefore;I) Determine the relative importance of parasitoids, disease and pupal predatorsfor winter moth populations in Richmond (Chapter 2).II) Compare mortality on blueberry and birch, and in particular examine thesuccess of C. albicans in parasitizing winter moth on these two hosts (Chapter 2).III) Investigate predation of winter moth pupae and identify the possible predatorsof the pupae in birch and blueberry sites in the Lower Fraser Valley (Chapter 3).14IV) Investigate the possibility of a link between parasitism by C. albicans andpupal predation (Chapter 3).15CHAPTER 2MORTALITY OF WINTER MOTH IN BIRCH STANDS ANDBLUEBERRY PLOTS IN THE LOWER FRASER VALLEY OFBRITISH COLUMBIA.2.1 IntroductionControl of winter moth, Operophtera brumata (L.), by its parasitoid, Cyzenisalbicans (Fall.), is commonly cited as one of the most important examples for the successof classical biological control (Embree 1971, DeBach 1974, Hassell 1978, Embree andOtvos 1984, Murdoch et al. 1985). Winter moth, native to central and eastern Europe, hasonly recently become established in North America. The moth was first introduced toeastern Canada around the 1930's (Cuming 1961, Embree 1965b, 1966), but it was onlycorrectly identified in 1950 having been mistaken for the native spring cankerworm,Paleacritavernata Peck (Hawboldt and Cuming 1950, Smith 1950). From 1954 to 1961, aprogram of biological control was carried out in eastern Canada with the introduction of sixparasitoids from Europe. Only two of the parasitoids, C. albicans and Agryponflaveolatum (Gray.), became established. The subsequent decline of winter mothpopulations from 1961 to 1963, was attributed to these two introduced parasitoids (Embree1966).The first records of winter moth in western North America are from collectionsmade in 1958. The moth was probably introduced around the 1950's (Ferguson 1978). Itis not known whether this was a completely new introduction or whether individuals hadbeen transported from eastern Canada. There appears to have been at least three separateintroductions into the west. (1) In Washington and Oregon the winter moth has beenpresent since the 1950's (Ferguson 1978). (2) In British Columbia it appears to have been16first introduced to Vancouver Island (Gillespie et al. 1978) with (3) a second, laterintroduction to the lower mainland (Wood and Van Sickle 1985). Winter moth wascorrectly identified on Vancouver Island in 1978, having been apparent at high densitiesonly since the early 1970's (Gillespie et al. 1978). The moth has now spread on VancouverIsland from Sooke to Nanaimo and Mill Bay (Wood and Van Sickle 1985, 1986, Pivnick1988).The main host plants of winter moth on Southern Vancouver Island are Garry Oak(Quercus garryana Douglas) and apple (Malus spp.). Defoliation was particularly severe inthe late 1970's to early 1980's (Gillespie et al 1978, Roland 1986a). The two parasitoids,A. flaveolatum and C. albicans, (from Europe and eastern Canada), were released inVictoria between 1979 and 1982 to control the winter moth (for details on introductions andrecoveries see Williamson 1981a, 1981b, 1981c, 1984, 1986). As in eastern Canada, C.albicans proved to be effective in parasitizing the moth and has been attributed (thoughindirectly) to the decline of the moth on Vancouver Island (Roland 1986a, 1988).In Nova Scotia, although A. flaveolatum was suggested to be important incontrolling low density populations of winter moth (Embree 1966, 1991) it has beensuggested that this species was not crucial to the success of C. albicans (Hassell 1980).Levels of parasitism by A. flaveolatum are low in Victoria and Vancouver (Roland 1986a,1992, and pers. obs.) and will not be discussed.One of the most fascinating paradoxes to arise from these two examples ofbiological control is that of the discrepancies between the role of C. albi cans in Britain andits role in both eastern and western Canada. In Britain, winter moth populations are largelygoverned by the destabilizing effects of 'winter disappearance' (largely dependent onasynchronies in the timing of bud burst and egg hatch) and the stabilizing effects of density17dependent soil mortality (mainly attributed to pupal predation) (Varley et al. 1973).Furthermore, C. albicans plays only a minor role in winter moth population dynamics sincethe levels of parasitism are normally low (about 5%) (Hassell 1969a). However, inCanada, on the two occasions where C. albicans was introduced, the parasitoid appeared toinitiate dramatic crashes in winter moth populations. Parasitism in Nova Scotia had reachedlevels of 70% on oak six years after its introduction, with levels as high as 80% at somesites. On apple, levels of parasitism as high as 50% were recorded in the early 1960's(Embree 1965b). Hassell (1980) has suggested that lower pupal mortality observed inNova Scotia may be the key to understanding the differences in the dynamics andinteractions of the two species on the two continents. In Britain, winter moth pupalpredation is high, and predation of C. albicans is expected to be higher due to the extendedavailability of C. albi cans pupae in the soil (Hassell 1969a, 1969b, 1980). Therefore, inBritain, C. albicans is limited to densities that cause insignificant host mortality. A lowerpupal mortality in Canada may have enabled the winter moth population to reach high levelsand when C. albicans was introduced, low soil mortality allowed it to increase rapidly untilit caused winter moth populations to collapse (Hassell 1980).After the introduction of C. albi cans to Victoria in 1979, winter moth populationsalso went into decline. Parasitism reached levels as high as 84% on oak and 50% on apple(Roland 1986a, 1986b). Roland (1986a), investigating the success of C. albicans,indicated that some years after its initial introduction, winter moth soil mortality began toincrease from levels as low as 10% in 1981 to levels of 96% in 1987 (Roland 1992). Asimilar observation had been made in Nova Scotia with soil mortality increasing after theintroduction of C. albicans from 37% in 1954 - 1959 to 94% in 1961 and 1962 (Embree1965b, Hassell 1969a). Roland's (1986b, 1988) analysis indicated that soil mortality wasactually more important than parasitism in bringing about regulation of the winter mothpopulation in Victoria. Roland reanalysed the Nova Scotia data and suggested that18procedures used by Embree may have masked a similar occurrence there. At the time ofoutbreak collapse in Nova Scotia, soil mortality had also been the most important mortalityfactor in causing winter moth population reduction. Roland (1988) has suggested that bysome unknown mechanism(s), C. albi cans may have led to the observed increases in soilmortality and thus, indirectly, to population decline.The introduction of C. albicans to Nova Scotia may have aided in the transmissionof disease among the population. Viral disease, particularly baculoviral disease, has beenreported from a number of declining populations of forest Lepidoptera including Bruce'sspanworm in Canada and winter moth in Europe (Wellington 1962, Stairs 1966, Tenow1972, Cunningham 1982, and see Myers 1988). Winter moth has been shown to besusceptible to all three genera of occluded insect viruses; cytoplasmic polyhedrous viruses,poxviruses and baculoviruses (Wigley 1976). Wigley (1976) identified nuclear polyhedralvirus (NPV) of the genus Baculovirus, subgroup A, polyhedrosis among an outbreakingwinter moth population at Wistman's Wood in England. The virus was responsible for asmuch as 23% of the observed larval mortality. Feeny (1966 in Wigley 1976) also foundNPV in winter moth at Wytham Wood in England, and NPV has been found in populationsof winter moth on Scots pine (Pinus sylvestris L.) and Sitka spruce (Picea sitchensis(Bong) Carr) in Scotland and England (Wigley 1976).In Nova Scotia, no disease was found to infect winter moth between 1954 and1961. In 1961 a single individual was found with NPV and by 1964 NPV was found to be'generally present' throughout the winter moth distribution (in Embree 1966). Therefore,the appearance of NPV coincides with the introduction of the parasitoids and the decline ofthe winter moth population. Embree (1966) suggests that the origin of the virus in Canadamay be from the related Bruce's spanworm, since the occurrence of the virus coincidedwith that of a very similar virus (Borelinavirus bruceata, in Canadian Department of19Forestry 1964, 1974) causing outbreak collapse in spanworm. Cunningham et al. (1981)investigated the incidence of infection among field populations of winter moth near Victoriaand found no evidence of virus. NPV introduced before the introduction of the parasitoidsin 1979 did not become established. Roland (1986a, 1988) found low levels of virusamong winter moth populations on Vancouver Island after the introduction of C. albicans.It may be that, in the absence of parasitoids there is inefficient viral transmission.Winter moth was probably introduced into the lower mainland of B.C. in the early1980's. In 1985 pheromone traps picked up winter moth on fruit trees at Richmond andTsawwassen (Wood and Van Sickle 1985) and by 1989 moderate to severe defoliation ofbirch was being attributed to winter moth in Richmond and from Ladner to Surrey (Woodand Van Sickle 1990, 1991). Pheromone trapping has indicated that winter moth hasspread throughout the lower mainland and now occurs as far east as Mission. The highestpopulation densities have been recorded at Richmond, Delta and Surrey (Fitzpatrick et al.1991a). The moth is also spreading into Vancouver (pers. obs.). On the mainland, birchBetula spp. and blueberry,Vaccinium corymbosum (L.), are the main hosts withraspberries, Rubus spp., in Ladner also being damaged (Fitzpatrick et al. 1991a).Parasitoids were never released in the lower mainland, but they are present amongwinter moth populations, presumably having been introduced at the same time or shortlyafter the winter moth. This recent introduction offers an opportunity to look at a post-introductory outbreaking population of winter moth, with a co-occurring C. albicanspopulation, in the light of the recent ideas put forward on the mechanisms of control. I setout to observe the population trends of winter moth, at every life stage, on the lowermainland of British Columbia. I compared blueberry and birch as plant hosts and examinedthe mortalities affecting populations on these two hosts. I also investigated the prominence20of disease among the Richmond population. A discussion of the possibility for the successof C. albicans on the two different hosts will be presented.2.2. ProceduresStudy sitesFour field sites in Richmond, B.C., were chosen for detailed life table analysis ofwinter moth. These sites are indicated by solid squares in Figure 2.1. Two of the sites wereblueberry plots on agricultural land. The other two were predominantly birch woodland.For further details of these sites see Chapter 3.Blueberry I (BBL no. 15 in Figure 2.1)This blueberry plot is 50m x 20m and is not commercially harvested. It contains anumber of different varieties of highbush blueberry (Vaccinium corymbosum (L.)) andthere is considerable variation in size among the bushes. The sparse density of bushesallow an undergrowth mainly of grasses and ferns. There are some young birch (Betulapapyrifera var. communata Marsh) interspersed with the blueberry bushes.Blueberry II (BBIL no. 16 in Figure 2.1)This site has an area of about 50m x 20m of highbush blueberries (V.corymbosurn). The blueberries are not commercially harvested and therefore, have becomeovergrown and so block light from the undergrowth. The north side of the site adjoins acommercial blueberry plot, where winter moth are controlled by insecticides. The westernedge of the plot receives more light which allows some undergrowth. Undergrowth at the21Figure 2.1. Study sites in the lower mainland. Solid squares indicate sites usedin life table studies, open squares indicate winter moth sampling sites.Vancouver sites (1-8); 1. University of British Columbia, 2. West 9th andAlma, 3. West 13th and Trimble, 4. Spanish Banks, 5. Locarno Beach, 6.Jericho beach, 7. West 35th and Granville, 8. West 61st and Granville.Burnaby sites (9-10); 9. Burnaby lake, 10. Simon Fraser University. RichmondSites (11-17); 11. Knight street, 12. Department of National Defense lands, 13.Richmond Nature Park, 14. Birch stand I, 15. Blueberry plot I, 16. Blueberryplot II, 17. Blueberry plot III. Westham Island (18) and Delta sites (19-20); 19.Deas Island, 20. 112th Street (north). Ten sampling sites at Ladner andChilliwack are not shown.22western edge consists mainly of briar (Rubus spp.) and some sporadic salal (GaultheriaShallon Pursh). Some small birch (B. papyrifera) are interspersed among the blueberrybushes.Richmond Nature Park (RNP, no. 13c in Figure 2.1)Richmond nature park consists of an area of about 0.705km2 and is of mixed forestvegetation consisting mainly of birch (B. papyrifera) and hemlock (Tsuga heterophylla.(Raf.) Sarg.). The undergrowth includes salal (G. shallon), blueberry (V. corymbosum),and red elderberry (Sambucus racemosa L.). Some parts of the park consist of boggy areaswhere the flora is dominated by Labrador tea (Ledum groenlandicum Oeder) and blueberry.Studies were carried out on the west side of the park.Birch Stand I (BI, no. 14 in Figure 2.1)This is an area of 125m x 50m, consisting of birch (B. papyrifera) woodland. It issurrounded by farmland. The birch trees are of a variety of ages, there is some sumac(Rhus glabra L.) and Hemlock (T. heterophylla) growing amongst the birch. Theundergrowth consists largely of salal (G. shallon) and blueberry (V. corymbosum).Sampling proceduresLarvae (Blueberry)The sampling unit for larvae was a leaf cluster. Leaf-clusters consist of individualbuds in the spring, which develop into leaf or flower clusters in later months. In 1991,buds were collected only on April 1 lth. In 1992, due to a mild winter and early spring and23thus an early bud burst, larval sampling commenced on March 1 lth• During 1992,sampling was carried out at regular weekly intervals. Samples were collected on March1 1 th, 23rd, 31st, April 8th, 15th, and May 5th• Sampling stopped after May 5th, since afterthis time pupation had begun. Estimates of early instar densities are from samples taken onApril 1 lth in 1991 and on March 23rd in 1992.In sampling, the two apical buds of branches were collected. This biased the resultson larval densities within plots because winter moth larvae prefer apical buds (seeAppendix 1), but did not affect between plot comparisons. Apical buds were sampledbecause in later months larval densities are expected to be low, and large numbers of larvaewere required for estimation of parasite and pathogen incidences.The buds/leaf clusters were randomly collected from four transects at eachblueberry site. Eight buds were taken from each of 10 bushes, two from each of thecardinal points; a total of 320 buds were collected at each sampling period. Buds weredissected under a light microscope and the larvae counted. In 1992, data on buddevelopment, larval numbers and instars were recorded. Instars were determined bymeasuring head capsule widths of each larva under a dissecting microscope (see Appendix2).Larvae (Birch)Sampling of birch stands was carried out on April 1 lth in 1991 and on March 1 lth,23rd, 31st, April 8th, 15th and May 5th in 1992. Outer branches of birch were clipped witha telescopic pruner which reached a height of ca. 15m. Branch samples consisted of two totwenty leaf clusters and were taken at each of the cardinal points of ten trees along threetransects. A minimum of 240 buds and leaf clusters were collected for each birch site at24each sampling date. Each leaf cluster consisted of two to five leaves. Successive sampleswere taken from the same transects but not necessarily from the same trees. Samples ofundergrowth were randomly taken along each transect. Larvae were counted from eachsample and data were collected on bud development, larval development and leaf damage.Larvae collected from all sites (blueberry and birch) were kept in individual plasticcontainers in an outside shed and reared through to pupation. After pupation, pupae wereexamined and scored as healthy, deformed, parasitized or dead due to unknown causes.Larvae were fed ample supplies of birch, bluebell)/ or apple from non-infested areas, andleaves were changed every two to four days. A number of the larvae were frozenimmediately and examined for viral or microsporidian disease. All larvae which died duringrearing were also examined for pathogens. Smears were made from dead larvae and stainedwith Naphthalene black in 1991 and with both Naphthalene black and Giemsa, to aid inidentification, in 1992. Examinations for microsporidia were carried out at every stage ofthe life cycle, except eggs in 1991. In 1992, eggs were teased apart, fixed in Methanol andstained in Giemsa. Larvae were stained as above. For adults the abdomens were severed,the contents were then moistened, fixed and stained with Giemsa.Pupae (Blueberry and birch)Pupal densities were estimated by using the pupal drop tray method (Varley andGradwell 1968, Varley et al. 1973). Throughout pupation, drop trays were placed alongthe transects beneath the host plants at each site. Drop trays consisted of plastic trays filledwith sifted peat. Each year trays were set out in late April and collected in early June. Trayswere sifted to find cocoons. Cocoons were opened and the condition of the pupaerecorded. Pupae were scored as healthy, parasitized or dead due to unknown causes. Pupae25were weighed and pupal cremasters were examined to distinguish winter moth pupae fromthose of Bruce's spanworm (Eidt and Embree 1968).Adults (Blueberry and birch)Densities of emerging adults were estimated in 1991 and 1992 by using twotrapping methods. Sticky traps were used in both years at birch sites, but only in 1992 atblueberry sites. These consisted of bands of masking tape tightly wrapped about the tree orbush trunk. Tanglefoot was spread over the bands. Ten traps were placed at each site.Stocking traps were used only in 1991 at birch sites. Twelve traps were set up at each siteand these were used to supplement data from sticky traps. The stocking traps were similarto those used by Embree (1965b) except that the females were captured live in containers atthe top of the trap. These containers had funnels through which females could enter, butrestricted them from leaving the trap. One trap was placed on each tree, in three transects offour trees. Traps were equally divided among the four cardinal points. Stocking traps werenot suitable for blueberry plots because the blueberry stems were too narrow and low.Adult females were collected from the traps at regular intervals. The diameters andcircumferences of the trunks and the canopy areas of the trees were measured. Thenumbers of buds per unit area were visually estimated for each tree or bush. The fecundityof the females was examined by counting the number of oocytes in the body and comparingthis with estimates from a number of published studies.Data AnalysisData were analyzed using 2-way ANOVAs with sites as one of the variables andyears as a second variable. Where necessary, log and arcsine transformations were used to26reduce heterogeneity of variance. Where homogeneity of variance did not satisfy Barlett'stest (P <0.05), nonparametric Kruskal Wallis tests were applied. Following ANOVA,Tukey's HSD test was applied to test for mean seperation. The Mann-Whitney U test wasused to test for mean seperations on all Kruskal Wallis tests. Comparisons of 1991 and1992 were carried out for all life stages and mortalities. Only total pupal densities (prepupaldensities) were compared over three years (1990 - 1992).The magnitude and trends in mortality were demonstrated using k-values. Themortality of eggs and larvae (kiarv) was estimated as the difference between the log of theestimated egg potential each year per m2 and the mean density of pre-pupae entering the soileach year. The egg potential each year was estimated by multiplying the number of adultfemales per m2 that year by the average fecundity of females in the same year. The averagefecundity was estimated by weighing pupae in 1991 and 1992 and using a relationshipfrom Roland and Myers (1987) for pupal weight versus fecundity. The value for 1990 wasestimated by multiplying an estimate of adults per m2 (derived from pupal densities andpupal mortality estimates) by 150, an average fecundity for the species (Embree 1965a).Mortality due to parasitism (kpara) was estimated as the difference between the logof the mean density of prepupae and the log of the density of those which wereunparasitized. An important mortality factor, termed kprepu, was observed each year. Thiswas the difference between unparasitized prepupae dropping to the soil and healthy pupaein the soil, i.e. death due to unknown causes.Pupal mortality (kpupa) was estimated as the difference between the log density ofhealthy pupae in the soil and the density of emerging adults in the winter. For 1989 and1990 estimates of the numbers of adults emerging were derived from predation estimatesusing the tether method (see Chapter 3). These estimates are meaningful since in 1991 and271992 estimates of adult densities by both the tether method and the trapping of emergingfemales corresponded well.Spread of winter moth and C. albicansRandom samples of birch were collected from 30 sites throughout the lowermainland of British Columbia (see Fig. 2.1). Ten sites at Langley and Chilliwack are notshown in Figure 2.1. These included sites at Cultus Lake, Mill Lake, Aldergrove Lake andCampbell River. Collections were made by clipping 40 branch tips from birch trees orlowgrowth bushes. Lowgrowth bushes included crabapple (Malus sylvestris Mill.) , cherry(Prunus spp.), and blueberry (V. corymbosum). Caterpillars were counted and rearedindividually in plastic cups in an outdoor shed. Caterpillars were given ample amounts ofbirch from non-infested trees. Leaves were changed every two to four days. Caterpillarswere reared to pupation and the condition of the caterpillars was determined by examinationof the pupae. Pupae were identified as winter moth or spanworm and scored as eitherhealthy, parasitized or dead due to unknown causes.2.3 Results2.3.1. Damage to blueberry and birchDamage to blueberry: Timing of bud burst is dependent on the variety of blueberry inquestion (Anonymous 1990). At Richmond sites, blueberry buds burst before birch in1992. Five stages of development were identified (see Fig. 2.2), these were similar to earlystages in Massie's system (see Holliday 1977).28Figure 2.2. Stages in the development of highgrowth blueberry (Vacciniumcorymbosum) flower clusters (a) and leaf clusters (b). See text for details.29Stage I: The closed bud stage. Flower buds appear sealed, are generallybrown to red in colour and are between 2 and 3.5 mm in width (la). Leafbuds are much smaller, between 0.5 and 1.5 mm in width (lb). Beforeburst, a greenish tinge appears at the edges of the individual scales, thisoccurs in both bud types.Stage II: The flower buds swell to 4 - 4.5 mm and the scale tips turn pinkAt this stage the scales are not tightly closed (2a). The scales then begin toshrink back and turn brown. Leaf buds swell to about 2mm and do not havea pink tinge (2b).Stage III: Eventually the bud begins to loosen and open out (3a). In leafbuds two outer whorls roll back revealing a central leaf whorl (3b). Thiscentral leaf whorl lengthens and the leaves begin to enlarge (leaf buds cangive rise to from 2 to 5 leaves).Stage IV: For flower buds, this stage is marked by the elongation of theflower stalks and the protrusion of the young green flower petals (4).Stage V: Elongation continues and the flowers turn white and open out.Stage five flowers have a bell shaped appearance and the leaves arecompletely expanded (5). Stages IV and V for leaves are marked by thecontinued expansion of the leaves.Larvae attack Stage II buds, generally tunneling through the developing flowerpetals to eat the developing reproductive parts inside. Because of the tightness of the buds,frass accumulates between flower parts and may initiate fungal growth and thus spoilage of30the flower. At high densities the developing buds can turn completely brown and die.Leaves do not have the usual 'shot-hole' damage (cf AliNiazee 1986). Damaged areas aregenerally larger than discrete holes. Early in the season the caterpillars show a preferencefor flower buds, only later switching to the leaves. There is also a preference for apicalbuds (see Appendix 1).Damage to birch: Five stages in the development of birch leaves were noted (Fig. 2.3),these stages corresponded to stages in Malaisse's (1964) system for beech, having apseudoterminal bud type, i.e. leaves emerge one after the other from the base to the tip ofthe shoot.Stage I: Buds are tightly closed, brown in colour and from 3 to 5 mm longand 2 to 3 mm wide (1).Stage II: The buds lengthen, swell and turn green (2).Stage HI: The leaf scales open and the two outer leaves unfurl. A centralinner whorl may be apparent at this stage (3).Stage IV: The inner whorl begins to lengthen, and eventually the outerleaves (5 - 7mm) fold back. Protective leaf scales can be seen about thecentral whorl. The leaves continue to elongate (4).Stage V: At about 10- 14mm the central leaves begin to open, continue toexpand and the petioles elongate (5).311.2.Figure 2.3. Stages in the development of birch leaves (Betula papyrifera). Seetext for details.32The development of buds varies with site, position of buds within sites andpositions of buds on the individual branches. For example, at Birch site BI, outsidebranches were slower to develop than branches at 30 or 60m from the edge.Damage to birch was visually estimated at each site, within trees and among leaves.In 1991, damage was estimated at about 50% to all leaves at BI and 30% at RNP. Larvaebegin their damage on stage III buds. Within trees, lower branches generally had the mostdamage, presumably through downward larval migration throughout the season. Thisdisagrees with Embree's (1965a) observations of winter moth on oak where larvae werepredominantly on the upper branches (see also Eidt and Embree 1968 and Dubrovin 1990).There is a great variability in the damage encountered among trees and within sites.Differences among trees were most striking at RNP. Undergrowth in 1991 was often100% defoliated and among birches young trees were particularly damaged, presumablydue to their low size, rather than through any differential foliage quality. In 1992 similarpatterns were observed with average damage to birch leaves at about 30%.Growth on Blueberry and Birch: The pupal weights of winter moth on birch andblueberry were compared to investigate the suitability of these hosts for winter moth larvae.Growth on apple was also examined and compared since feeding on apple is known toresult in heavy pupae (Holliday 1977, Roland and Myers 1987). The host plants werefound to have a significant effect on the resulting size of pupae (1-way ANOVA on logtransformed data P <0.001). There was no significant difference in the weights of larvaereared in an outside shed on either blueberry or birch, or on a combination of blueberry andbirch (see Table 2.1). However, larvae reared entirely on apple were significantly largerthan those reared on blueberry and birch (Tukey test, P < 0.001). Therefore, it appears thatbirch and blueberry are less favourable host plants than apple for winter moth, but thatgrowth on both blueberry and birch is equal.33Table 2.1 Mean pupal weights of winter moth, Operophtera brumata, rearedon apple, birch and blueberry or switched between hosts. 'N' is the number ofindividuals reared through to pupation. All larvae were taken from wildpopulations.Host plants temperature^mean weight^Nregime (DC)^(g)Initial host^Final host Apple^ 12 - 20^0.0284^22Birch 12 - 20 0.0210 115Blueberry 12 - 20^0.0185^36Blueberry^Birch^12 - 20 0.0212 42342.3.2. Population densities at RichmondData from 1989 already indicated high densities of larvae at Richmond. Thepopulation appears to have peaked in 1990, then underwent a general decline till 1992when the population collapsed on both blueberry and birch.Early instarsAnalysis of bud counts indicate that larval densities were dependent on site and year(2-way ANOVA of log + 1 transformed data BI, RNP and BBII larvae for 2 years, site: P<0.001, year: P <0.001).On blueberry changes in the year to year densities of early instars were examinedfrom four years of data (1989 - 1992) available for the two sites (Fig. 2.4). From 1989 to1990 a general increase in larval densities was apparent at the blueberry sites. For these twoyears the data are presented as the number of larvae per branch. It is not clear from theavailable data whether there is an increase or decrease in the blueberry populations from1990 to 1991. Per branch counts are expected to be lower than per bud counts becauseonly apical buds were sampled in 1991 and 1992. This suggests that a decline in thepopulations from 1990 to 1991 is most probable. A decline is also suggested from datapresented by Sheppard et al. (1990), where winter moth larval densities of 2.2 per clusterwere recorded at BBII in 1990.Between 1991 and 1992 there has been a decline in the populations. Data for BBIare not available for 1991, but there has been an obvious decrease in the numbers of larvaepresent (pers. obs.). At BBII data were collected for both years and there is a significantdecrease in the larval densities (T-test on log + 1 transformed data, P <0.001).35of early instar larvae per branch in 1989 and 1990 (a), and1992 (b), at four sites (open circles = BBI, open squares == BI and closed squares = RNP) in Richmond, B.C.. Allncluded but are negligible.Figure 2.4. Densitiesper bud in 1991 andBBII, closed circlesstandard errors are i36One-way ANOVA of 1991 data indicate that densities of larvae at BBII were notsignificantly different from RNP (Tukey test, P = 0.108), but were different from BI(Tukey test, P <0.001). In 1992, both Blueberry sites were identical (Tukey test, P = 1)and were similar to densities encountered at RNP (Tukey test, P = 0.2). BI hadsignificantly more larvae than all other sites in 1992 (Tukey test, P <0.001).Two years of data (1991 and 1992) are available for the birch sites (Fig. 2.4b).Local accounts and reports from the Forest Insect and Disease Survey suggest that birchsites also underwent a population decline between 1990 and 1991 (Wood and Van Sickle1990, 1991). Between 1991 and 1992 a decrease in early instar densities is apparent atRNP, but at site BI there appears to be a slight increase in larval densities. Data werecollected at the western side of BI in 1991. In 1992, this side was cut down and sosampling in 1992 was carried out at a new area. Observations of defoliation patterns in1991 indicate however, that the western edge had received considerably less defoliation in1991 than the eastern side (<10% damage to leaves and > 50% respectively).PrepupaeEstimates of prepupae per m2 are taken from pupal drop trays. Data on prepupaldensities are available for all sites for each of the four years (Fig. 2.5). Densities ofprepupae were dependent on site (Kniskal Wallis, P <0.001) and year (P < 0.001).Densities were significantly different at the P <0.001 level (Mann Whitney test) betweenblueberry and birch sites. On blueberry, the densities of prepupae undergo a similar trendto that observed among early instars. The numbers of prepupae at the two blueberry siteswere similar (Mann Whitney, P = 0.235). From 1989 to 1990 there was a major increase inthe populations, but after 1990 the populations underwent a general decline. In 199237100806040200 A40030020010001989^1990^1991^19921 989^1990 ^1991^1992YEARFigure 2.5. Densities of prepupae per square meter at (a) birch sites (darkshading = BI, light shading = RNP) and (b) blueberry sites (dark shading = BBI,light shading = BBII) at Richmond, B.C. bars indicate standard errors38populations at both blueberry sites collapsed. Densities of prepupae at the two birch siteswere significantly different (Mann Whitney, P = 0.038). There was a decrease in densitybetween 1990 and 1991 (Mann Whitney, P = 0.036) with a crash at all sites in 1992 (MannWhitney test, P < 0.001).AdultsDensities of emerging females are available for two years only at the birch sites(Fig. 2.6). There was a significant difference in numbers of emerging adults between thesites (Mann Whitney test, P < 0.001). However, there is no difference between thenumbers of adults emerging in 1991 and the number emerging in 1992 (Mann Whitneytest, P = 0.113). Significant differences were observed between densities of adults at allsites in 1992 (Kruskal Wallis, P <0.001). Blueberry sites were similar in densities in 1992(Mann Whitney, P = 0.19), but differed from birch sites (P < 0.001, Mann Whitney tests).The two birch sites also differed in 1992 (Mann Whitney test, P = 0.036).FecundityPupal size was dependent on site (2-way ANOVA on log transformed data, P <0.001) but not on years (P = 0.398). In 1991, pupae from BI tended to be smaller (Table2.2), although not significantly smaller than those from RNP (Tukey test, P = 0.097).Sufficient data were not available to test differences between birch and blueberry sites. Inboth years (1991 and 1992) pupae were heaviest at RNP. Counting of oocytes fromfemales in 1991 indicated that fecundity was generally lower than expected fromregressions taken from apple fed larvae, but more closely approached estimates from oakfed larvae. Therefore fecundity was estimated from Roland and Myers (1987) relationshipof pupal weight to fecundity for oak fed larvae.391989 1990 1991 199219921989^1990^1991^1992YEARFigure 2.6. Densities of healthy winter moth pupae from 1989 to 1992 withcorresponding adult densities from 1991 and 1992 at (a) two birch sites (darkshading = BI, light shading = RNP) and (b) two blueberry sites (dark shading =BBI, light shading = BBII) at Richmond, B.C., bars indicate standard errors.40Table 2.2. Mean pupal weights of winter moth Operophtera brumata fromfour field sites at Richmond B.C. during 1991 and 1992, and estimates offecundity based on published relationships between weight and fecundity(eggs/female).Site^Year Mean N^fecundity estimatesWeight(g) Roland and^Hale^HollidayMyers (1987) (1989)^(1977)Oak^Apple Apple AppleBI 1991 0.0251 83 128.41 175.55 178.6 171.771992 0.0236 40 114.7 159.8 165.1 157.46RNP 1991 0.0277 103 152.18 202.85 202.0 196.581992 0.0290 39 164.06 216.5 213.7 208.98BBI 1991 0.0224 10 103.74 147.2 154.3 146.021992BBII 1991 0.0263 37 139.38 162.15 189.4 183.221992 0.0214 2 94.60 110.7 145.3 136.4841It is apparent therefore, that each of the four study sites underwent a significantpopulation decline in 1992. This decline was apparent at all life stages of the winter moth,except the adults. There has been no clear trend in fecundity changes during the decline.2.3.3. Mortality on Blueberry and birchParasitismThree parasitoid species were recovered from the Richmond populations. The mostcommon species was identified as C. albicans (identified by J.E. O'Hara: BiosystematicsResearch Centre, Ottawa) and has been present at all four sites since 1989. Agrypon sp.was also present in small numbers at birch sites. This species was found among pupaefrom 1990 to 1992, at levels below 0.5%. Because of its low incidence it has beenexcluded from analysis. A third species, morphologically and behaviourally similar toEphialtes spp. (see Wylie 1960b) was recovered from larvae late in the season at BI. Thelarvae of this parasitoid were observed feeding externally on winter moth larvae. Usuallythree to five larvae were found attached to an individual caterpillar. After the winter mothlarvae had died, the parasitoids pupated. These parasitoids were found in 1991 and 1992.The parasitoid was rare and its effects are also considered to be negligible.Parasitism by C. albi cans fluctuated between years at all sites (see Figs. 2.7 and 2.8and Appendix 3). Parasitism was dependent on site but not on year (2-way ANOVA onarcsine transformed data, P <0.001 for sites and P = 0.061 for years (1991 - 1992). Thelack of any significant difference between years may have been due to small sample sizes in1992. The highest parasitism occurred at the birch sites in 1991 with about 50% of thepupae having parasitoids. Parasitism on blueberry was generally lower than on birch. The423.0-2.0-1.0-0.0BI^ BI3.02.0LARVALJA1.0^ ?Ak^LNKNOW■1Cj"---0C'------°.^ •^.1989 1990 1991 1992 1989 1990 1991 19920.0II'■.^TOTAL K? .%.<11.PREDATIONPARASITISMRNP3.0^TOTAL K2.01.0^PREDATIONP RASITISMyear yearFigure 2.7. Mortality of winter moth at birch sites in Richmond. Closed circlesindicate pupal predation (kpup) estimated from tether experiments, opencircles indicate mortality due to C. albicans (kpara). Closed triangles indicatelarval mortality (kiarv) and open triangles indicate death of prepupae due tounknown causes (kprepu), data are from BI, and RNP. Total generationmortality (K) is also presented.0.01989 1990 1991 19923.02.0RNPLARVAL1.00.01989 1990 1991 199243LARVALPREDATION1.0PARASITISM0.01989 1990 1991 1992BBII3.0TOTAL K2.0^liC61.00.0PREDATIONPARASITISM 1989 1990 1991 1992BBI^ BBI^3.0^ 3.0^2.0^ 2.0TOT,,/1.00.01989 1990 1991 19923. 1990 1991 1992year yearFigure 2.8. Mortality of winter moth at blueberry sites in Richmond. Closedcircles indicate pupal predation (kpup) as estimated from tether experiments,open circles indicate mortality due to C. albicans (kpara). Closed trianglesindicate larval mortality (klarv) and open triangles indicate death of prepupaedue to unknown causes (kprepu), solid squares indicate total generationmortality (K). Data are from BBI and from BBII.44highest levels on blueberry were encountered at BBII with 39% of pupae parasitized. Therewas no significant difference between birch sites each year (Tukey test, P = 0.96) andbetween blueberry sites each year (Tukey test, P = 0.98). Birch sites had significantly moreparasitism than blueberry sites (Tukey test BI vs BBI and BBII, P < 0.005, BII vs BBI, P> 0.005 and BBII, P = 0.018).There was no significant difference in the levels of parasitism encountered betweenblueberry sites over the 4 years, although a general increase in the levels is apparent until1991. On birch levels appeared to peak in 1991. However, in 1990 the levels encounteredare much lower than expected. Between 1991 and 1992 there was a decline in the levels ofparasitism encountered at both sites (P = 0.061).DiseaseExamination of cadavers and live specimens indicated that the incidence of diseaseamong the Richmond population was very low. All sites were negative for viral disease.Microsporidia were not identified from egg, pupal or adult stages at any of the sites.However, at site BBI microsporidian-like bodies were present among late larvae, but onlyduring 1991 (see Appendix 4). It appears therefore, that viral and microsporidia diseaseshave had little effect on the lower mainland populations.Pupal mortalityMany of the larvae which pupated were found to be either deformed or dead. Oftendead pupae were found to have fungal hyphae. However, it is difficult to ascertain whetherpupae had died due to fungal attack, or whether the fungi had simply attacked the pupalcadavers. The levels of dead or deformed pupae differed between sites and years (2-way45ANOVA on arcsine transformed data P = 0.026 for sites, P < 0.001 for years). BBI hadsignificantly more dead pupae than any of the other sites (Tukey test, P <0.05). In 1992,pupal death was high at all sites, but particularly high on blueberry (Tukey test, P <0.001), (see Figs. 2.7 and 2.8 and Appendix 3). Because of these differences betweenbirch and blueberry and because of the very high incidence of pupal death in 1992, itappears likely that death was mainly a result of asynchronies between larvae and leafdevelopment.Soil mortalitySoil mortality was estimated in 1991 and 1992 at birch sites, but only in 1992 at theblueberry sites. Stickybands and stocking traps indicated high mortality of winter mothbetween pupal drop and adult female capture. Tethered pupae (these are described in moredetail in Chapter 3) indicate levels of pupal predation. Estimates for both soil mortality andpupal predation were very similar, indicating that most of the soil mortality was attributableto predation of pupae by generalist predators (see Table 2.3). The estimate of soil mortalityfor BBII in 1992 was much greater than that for pupal predation in the same year.However, this is probably inaccurate. There were also large discrepancies at BBI betweenthe estimates of soil mortality and pupal predation, from emergence traps and tethersrespectively. Difficulties in estimating soil mortality at blueberry sites may be due to thesmall numbers of emerging adults in the winter (see Fig. 2.6). This indicates that at lowpupal densities emergence traps are inefficient.46Table 2.3 Estimates of soil mortality (from emergence traps), and mortalitydue to pupal predation (from tethers) at four sites in Richmond for two yearsat birch sites and one year at blueberry sites. The difference is attributed tomortality of prepupae on the ground and of adults after emergence and todeath of healthy pupae in the soil. All estimates are presented as percentages.SITE BI RNP BBI BBIIYEAR 1991^1992 1991^1992 1992 1992Soil^mortality 90.73^66.96 96.90^90.90 91.07 98.61Pupalpredation90.28^58.75 91.00^82.5 94.32 77.63Difference 0.45^8.21 5.90^8.40 -3.25 20.98K-factor analysisK-factor analyses indicate that pupal mortality (kpup) has been an important sourceof mortality throughout the four years for which data are available. This mortality has beengreater than mortality caused by C. albicans (kpara) at each site and each year. At three ofthe sites, kpup appears to be weakly density dependent (Fig. 2.9). Pupal mortality (kpup)appears to be temporally density dependent (except at BBI), since it is dependent on theyearly pupal densities at each site. However, since only four years of data are available,this trend is weak. Larval mortality (klarv) and mortality due to unknown causes (kprepu)increased in 1992 at each site. This increase was most apparent at the blueberry sites.Estimates of larval densities for BI in 1991 were not feasible since they gave values lowerthan the estimates of prepupae from drop trays. This however indicates that klarv at BI in1991 was probably also low.47Figure 2.9. K-values of pupal predation plotted against the total densities ofpupae in the soil. Closed circles indicate BI, closed triangles indicate RNP,open circles indicate BBII and open triangles indicate BBI. Trends indicate theoccurrence of weak temporal density dependence at each site.48The similarities in the trends of kprepu and klarv suggest a link between these twomortalities. I suggest that this link is foliage quality, since both of these mortalities may beinfluenced by foliage quality (and thus asynchronies of bud burst). The total generationmortality (K) demonstrates similar trends at all four sites, with the lowest values in 1991.although it was likely to have been lower in 1989 when larval survival had not beenestimated. Mortality due to parasitism (kpara) peaked at each site in 1991. In that yearwinter moth densities were already decreasing, so that a decrease in kpara should have beenexpected. Plotting kp ara against prepupal densities (Fig 2.10) gives a counterclockwisetrend for BBL This suggests that kpara has a delayed density dependence, however, onlyfour years of data are not sufficient to examine this fully. For birch sites there are noapparent trends. Greater success of C. albicans in 1991 may reflect better synchronizationof C. albicans oviposition with the appropriate stage of winter moth larvae, or greatersurvival of C. albicans compared with winter moth from the previous year. A possiblemechanism for this could be greater soil mortality of winter moth pupae.Refugia against parasitismAbsolute densities of winter moth pupae are higher for birch sites than for blueberrysites. However, birch trees have more leaf clusters per unit area than blueberry sites. Whenthe levels of parasitism are plotted against the densities of winter moth prepupae per leafcluster it becomes apparent that parasites are more efficient on birch in spite of lowerprepupal densities (Fig. 2.11). I investigated some possible refugia by which winter mothon blueberry may be avoiding parasitism, and examined whether a varied habitat couldoffer a further refuge at birch sites.493E'<ft'^100^200^300^400PUPAL DENSITY/m - 2Figure 2.10. K-values of C. albicans parasitism plotted against the totaldensities of pupae in the soil, an indicator of prepupal densities. Closed circlesindicate BI. Closed triangles indicate RNP, open circles indicate BBII and opentriangles indicate BBI. Trends indicate a weak delayed density dependence atthe blueberry sites, but no trends are apparent for birch sites.500.00^0.01^0.02DENSITY (PREPUPAE/CLUSTER)Figure 2.11. Mortality of prepupae due to C. albicans (K parasitism) plottedagainst prepupal densities per leaf cluster as estimated from total pupae indrop trays divided by the average number of buds per unit area, estimatedvisually, at each site. Results indicate disproportionately low levels ofparasitism on blueberry (open circles = BBI, open squares = BBII, closedsquares = BI and closed circles = RNP).51a) Refuge resulting from foraging behaviourLarval development on blueberry appears slower than on birch (Fig. 2.12), but thiswas not significant. The large proportion of second instars in blueberry on April 15th(1992) is difficult to explain and is probably due to the small sample sizes taken from theblueberry sites. Later in the season there appear to be more larvae on the leaves ofblueberry than on the flowers (see Fig. 2.13) , but this trend was significant at only one ofthe sites (BBII) (2 sample t-test on Log + 1 transformed data, P = 0.005).Analysis of leaf damage indicates that winter moth larvae disperse their damage(Fig. 2.14). Early in the season most of the damage to leaves is below 5%, as the seasonprogresses the modal damage levels increase. The modal proportions of damage differamong sites based on the densities of larvae at each site. Similarly, high density areas havethe highest modal damage levels.b) Habitat refuge, birch canopy versus undergrowthCyzenis albicans adults were active and their eggs were found on the birch leaveson April 8th in 1992. Parasites were recovered from both undergrowth and canopycaterpillars (8% and 12% respectively). Few of the larvae reared in the outdoor shed, fromeither canopy or undergrowth, pupated and therefore it was not possible to adequatelyestimate levels of parasitism.There remains a possibility that the undergrowth may act as a refuge, by supportinglater instar larvae. Figure 2.12b indicates the occurrence of larval instars at four samplingdates. The distribution of instars in the undergrowth closely follows that of the canopy forall dates except April 8th. At one site only (the high density site) high numbers of fifth52(a)I^II^ifi^IV^V I^U^III IV V INSTARS(b)80604020804,1^60Z^4020II^III^IV^ III^IvINSTARSFigure 2.12. Percentage of winter moth larvae at each instar a) on blueberry atthree sampling dates and b) percentages of winter moth larvae in birch sites ateach instar over four sampling dates. Solid bars indicate larvae in the canopy,shaded bars indicate larvae in the undergrowth. Standard errors are included.There is no significant difference in the development of larvae on blueberryor birch or in either canopy or undergrowth.53Figure 2.13. Occurrence of winter moth larvae on flower or leaf clusters atthree blueberry sites in 1992. Open circles indicate densities on flower clustersand closed circles indicate densities on leaves. Error bars are included. Thereis an apparent shift from foraging on flowers to foraging on leaves as theseason progresses. Data was log+1 transformed and analyzed with a twosample t-test; *= P = 0.05, ** = P = 0.005.54Figure 2.14. Progression of damage to birch leaves at four different densities ofwinter moth larvae over three sampling dates. Densities of early instar larvaeper cluster are presented at the top. Bars indicate the proportion of leaves ineach damage class, with associated standard errors. Leaves with one hundredpercent damage generally had died as a result of larval attack.550.60.4020.00606OA02OD0604020.0020.10.003020.10.00.18O 1^$ 10 20 30 40 50 60 70 80 90 100O 1 5 10 20 30 40 50 60 70 80 90 100O 1 5 10 20 30 40 50 60 70 80 90 1000.19^ 0.36O 1 5 10 20 30 40 50 60 70 80 90 100Opol 15O 1 5 10 20 30 40 50 60 70 80 90 100UIO 1 5 10 20 30 40 50 60 70 80 90 100DAMAGEO 1 5 10 20 30 40 50 60 70 80 90 1000.80.6OA020.00.6Li0 040 200.0O 1 5 10 20 30 40 50 60 70 BO 90 1000.51O I 5 10 20 30 40 50 60 70 80 90 100O 1 5 10 20 30 40 50 60 70 80 90 100O 1 5 10 20 30 40 50 60 70 BO 90 1000806OA020.0OA0.4020.003020.1instars were found in the undergrowth but differences were not significant. Therefore, asuggested developmental difference between canopy and undergrowth caterpillars isunlikely.2.3.4 Winter moth and C. albicans spreadWinter moth is continuing to spread throughout the greater Vancouver area. Larvaewere present at 18 of the 20 sampling sites (see Tables 2.4, 2.5 and 2.6). Bruce'sspanworm alone was found at the Langley sites. Larval densities were highest atRichmond, but varied greatly over small distances. At Richmond Nature Park, numberswere lower at the eastern side (11 - 14). The southern areas of Vancouver (south of 41stAvenue), had high densities of winter moth causing moderate damage to ornamental andshade trees. At all other sites (1 - 7) Vancouver, (18) Westham Island, (19) Deas Islandand (20) Delta, numbers were low. Larvae were not found at Burnaby (9 - 10).Parasitism by C. albicans also appears high in Richmond. Figure 2.15 indicates thehighest levels of parasitism encountered throughout the four years of study. Parasitism washighest on birch. Figure 2.15 also includes incidences of parasitism from 1990 forTsawwassen and Ladner. Sampling at these sites was discontinued after 1990 but itappears that the Tsawwassen levels of parasitism are the highest encountered in the lowermainland. In Vancouver, parasitism remains low, at below 10%, with sufficient samplesizes only in the southern areas. No parasites were recovered from northern Vancouver(north of 41st Avenue), Westham island, Deas Island or Delta. However, numbers oflarvae reared through to pupae were very low at these sites.57Table 2.4. Winter moth early instar larval densities (per cluster) at Richmond sites during 1991 and 1992,with associated levels of parasitism by C.albicans and death of prepupae due to unknown causes from 1990to 1992. Host plants are indicated. Brackets indicate standard errors of larval estimates and sample size forparasitism and "unknown" mortality estimates. Asterix indicates that pupae were taken from drop trays,otherwise pupae were reared from collected prepupae.SITE HOST LARVAL DENSITIES PARASITISM DEATH (UNKNOWN)1991 1992 1990 1991 1992 1990 1991 1992KNIGHT ST. BIRCH 59(27) 22(27)D.N.D. BIRCH 0.0112 f0.0079) 0 (6)* 0(5) * 0(3) 0 (6)* 0 (5)* 0(3)BLUEBERRY 0.0119 (0.0118)(RNP) EAST I BIRCH 0.7432 (0.0881) 0.1236 (0.0201) 37 (42) * 36 (42) *(RNP) HWY. 99 BIRCH 0.2318 (0.0302) 11 (19) 68(19)(RNP) WEST I BIRCH 19 (348) * 55 (384)* 37 (103) * 28 (348) * 9 (103) * 36 (103) *(RNP) WEST II BIRCH 0.0111 (0.0064) 0.0218 (0.0093) 0(1) 0(1)BIRCH I BIRCH 0.2500 (0.0377) 0.2805 (0.0062) 24 (218) * 54 (448)* 38 (206) * 15 (218) * 8 (448) * 20 (206) *B.BERRY I BLUEBERRY 0.0926 (0.0176) 12 (93)* 14 (16) * 50 (6)* 18 (93)* 0 (16) * 50(6) *B.BERRY II BLUEBERRY 0.3579 (0.0348) 0.0502 (0.0148) 24 (269) * 36 (89) * 0 (15) * 13 (269) * 2 (89) * 42 (15)*B.BERRY III BLUEBERRY 0.1250 (0.0352)Table 2.5. Winter moth early instar larval densities (per cluster) at Lower Mainland sites during1992, with associated levels of parasitism by C.albicans and death of prepupae due to unknowncauses for 1990 to 1992. Host plants are indicated. Brackets indicate standard errors of larvalestimates and sample size for parasitism and "unknown" mortality estimates.SITE HOST LARVALDENSITIESPARASITISM DEATH (UNKNOWN)1992 1990 1991 1992 1990 1991 1992WESTHAM BIRCH 0.0065 (0.0064) 0(10) 20(10)ISLANDCRABAPPLE 0.0698 (0.0275)DEAS ISLAND BIRCH 0.0124 (0.0114) 0(5) 0(5)DELTA OAKTSAWWASSEN BIRCH 80(20) 0(20)LADNER BIRCH 28(162) 7(162)Table 2.6. Winter moth early instar larval densities (per cluster) at Vancouver sites during 1991 and 1992,with associated levels of parasitism by C.albicans and death of prepupae due to unknown causes. Host plantsare indicated. Brackets indicate standard errors of larval estimates and sample size for parasitism and"unknown" mortality estimates.SHE HOST LARVAL DENSITIES PARASITISM DEATH (UNKNOWN)1991 1992 1990 1991 1992 1990 1991 1992U.B.C. BIRCH 0.0041 (0.0040) 0 (7) 0(2) 50(2)9TH ALMA BIRCH 013Th TREMBLE BIRCH 0.0077 (0.0077)SPANISH APPLE 0.06 (0.0336) 0(2) 0(2)BANKSBIRCH 0 0.0039 (0.0039)LACARNO BIRCH 0.0078 (0.0077) 0(3) 66(3)BEACHJERICO BEACH BIRCH 0.0152 (0.0106) 0(46) 0(1) 48(46) 0 (1)35TH BIRCH 14 (21) 0 (41) 0(21) 61 (41)GRANVILLE49TH OAK BIRCH 8 (49) 6(49)61ST BIRCH 0.1914 (0.0256) 2 (44) 18 (44)GRANVILLEVANCOUVER BIRCH 10 (49) 4(49)Figure 2.15. Highest incidences of parasitism by C. albicans encountered atseven regions in the lower mainland, over the four year study period. Theyear in which parasitism was highest is presented in brackets. Birch was thehost plant at each site except at Richmond where blueberry was also sampled(as indicated), and Westham Island where both birch and crabapple weresampled.612.4 DiscussionWinter moth has reached high levels on both blueberry and birch in Richmond. Thepopulation was already high when this study was initiated in 1989. Since 1990 there hasbeen a decline in the population with an eventual population crash in 1992 on bothblueberry and birch. The decline in 1992 was significant at all life stages except adults.Parasitism - Host PlantsOverall average prepupal densities per leaf cluster on unmanaged blueberry aresimilar to those on birch stands (see Fig. 2.11). Significantly lower levels of parasitismalso occur on blueberry. The possible reasons for this include: i) differences in thesynchrony of host on the different plants with parasitoids, ii) differences in the foragingbehaviour of caterpillars on the different host plants or iii) differences in the chemical cuesfor oviposition of C. albicans on blueberry and birch.A large asynchrony between the occurrence of fourth and fifth instar winter moth andthe oviposition of C. albicans could explain the reduced parasitism of caterpillars onblueberry. Such an asynchrony has been implicated in the lack of success achieved by C.albicans on apple in Nova Scotia when compared to its success on oak (Roland 1986a,1986b). Parasitism of winter moth larvae on oak reached levels as high as 80%, while onapple levels generally remained lower than 5% (only in the early 1960's, did parasitismreach comparable levels (50%) on apple in Nova Scotia [MacPhee et al. 1988]). InVictoria, parasitism on apple was generally higher than in Nova Scotia (ca. 50%), but stilllower than on oak (ca. 70%). Reasons for the discrepancies between the two regions areunknown, but they may be due to climatic differences. Examination of the development rateof caterpillars in 1992 indicated no differences between development on birch and62blueberry (see Fig. 2.12). However, this represents only a single year in which samplesizes were small. Therefore, different development rates on blueberry and birch can not beruled out.Differences in the foraging behaviour of larvae on birch and blueberry could result indifferent levels of parasitism by C. albi cans. Roland (1986a, 1986b) found that C. albicansoviposits differently on oak and apple. In the field he observed more eggs on oak foliagethan on apple foliage in response to similar levels of damage. Also, C. albi cans eggs weremore clumped on oak foliage at all levels of scale measured (among clusters, within treesand among trees). In spite of this, the resultant levels of parasitism on apple and oak werenot different in Victoria. One of the reasons he suggested for this was that caterpillars maybe eating more foliage on apple than on oak.Larval feeding behaviour and the morphology of host plant leaves, as well as themorphology of the entire plant, could have profound influences on the success of C.albicans in parasitizing winter moth larvae. Feeding behaviour differs among winter mothcaterpillars on different host plants. Shot-hole damage described by AliNiazee (1986) forfilberts, is apparent on oak but not on birch or blueberry. These differences are probablydue to differences in the morphological or chemical qualities of the leaves. Therefore,although there may be less eggs on apple leaves there could be a greater chance that themoth larvae will ingest them, or start to feed from the site where the eggs are because of thechemical or morphological condition of the leaf. This may explain discrepancies betweenthe actual number of C. albicans eggs and the resultant levels of parasitism.Caterpillars generally disperse their damage, and thus depart from what would bepredicted from optimal foraging theories (see references in Mauricio and Deane Bowers1990). Two suggestions have been put forward to explain this pattern. Firstly, that63dispersal of feeding damage may help to avoid predators (Heinrich 1979). Secondly, that itmay be a response to changes in plant chemistry as the caterpillars feed (Feeny 1976).Regardless of the reasons for the dispersal of damage, it should result in reducedaggregation of eggs, less likelyhood of a caterpillar ingesting eggs, and possibly lesssuperparasitism. On birch, caterpillars ate small amounts of leaves (< 5%) before movingto new undamaged leaves. As the season progressed, the damage to the individual leavesincreased due to density effects and the larger size of caterpillars. This suggests that thelarvae are visiting previously damaged leaves. This revisiting of damaged leaves isexpected to be more common at high densities. On blueberry this trend was not observedbecause larvae feed of flower buds early in the season, making it difficult to examine theprogression of damage. However, an important observation in 1992 was the switch fromfeeding on blueberry flowers early in the season to feeding on blueberry leaves about earlyto mid April. This was at the time when C. albicans adults were ovipositing. Possiblereasons for this shift include: i) size constraints (at this time flower clusters may be unableto accommodate the larger winter moth instars); ii) sudden physiological changes in theflowers and thus changes in nutritive quality; or iii) avoidance of predation from insects,mainly wasps attracted to the flowers. This shift meant that the larvae moved to foliage thathad received little or no damage and thus were less likely to have C. albicans eggs.The switching behaviour observed on blueberry is more likely to cause a refugeagainst parasitism if C. albi cans oviposition on blueberry is similar to that on oak. If, as onapple, oviposition is not aggregative, then the resulting parasitism from lightly damagedleaves could be high. Because birch and blueberry leaves are morphologically more similarto apple leaves, as opposed to oak (i.e. lack of undulating edges, etc.), I expect that theforaging behaviour on blueberry will be more like that on apple. I suggest, based onphotographs of damage to oak and observations of damage to blueberry, birch and appleleaves, that the amount of damaged surface exposed for oviposition increases faster with64oak leaf damage than with apple. This is because caterpillars appear to avoid the toughveins of oak leaves but generally eat through the veins of apple. Also oak leaves aregenerally larger than apple, birch, and blueberry leaves, so that corresponding amounts ofdamage to oak induce more surface area for C. albi cans to lay eggs.The fact that C. albicans is able to aggregate its eggs in response to levels of damageon oak, but not on apple (Roland 1986a, 1986b), suggests differences in the synomonesreleased from these hosts in response to damage. Sugars at the edges of damaged oakleaves stimulate C. albi cans oviposition (Hassell 1968), and an airborne chemical attractantcauses aggregation (Roland et al. 1989, Roland 1990a). Which cues are present in birchand blueberry is not known. Presumably the contact oviposition stimulant is present in boththese hosts, but an airborne factor may be lacking. If the airborne factor is absent, the lackof oak trees in the vicinity of these study sites may explain why parasitoid levels are so lowin Richmond when compared to studies in Nova Scotia or on Vancouver Island.Furthermore, the morphological complexity of blueberry bushes compared to birch, mightreduce parasitism by decreasing direct contact of C. albicans with damaged leaves.Habitat refugeThe possibility of refugia at birch sites were also investigated; a habitat refuge is themost probable type of refuge. Winter moth larvae are largely monophagous, although thespecies is polyphagous (Wint 1983). However, shifting of host plants can occur amonglarvae (Wint 1983, Roland 1986a and see results section). In the birch woodlands in thisstudy, there was dense undergrowth of salal, red elderberry and blueberry. Of these, onlyblueberry was eaten. The blueberry plots had little or no undergrowth with only limitedpatches of salal. The possibility exists that larvae dispersing downward may move to freshhosts, and therefore have fewer opportunities of ingesting C. albicans eggs. A second65possibility by which undergrowth could act as a refuge is by supporting quickerdevelopment of later instars. Finally, C. albi cans may be more attracted to damage in oneof the hosts over the other, with one host plant acting as a sink for C. albicans eggs and theother creating a refuge for winter moth larvae in mixed stands.Undergrowth was found to be important for early instar larvae since bud-burst wasgenerally sooner in the undergrowth. Early in the season there appeared to be a higherincidence of second instars in the undergrowth than in the canopies (pers. obs.). Latersampling however revealed no differences in the development of undergrowth and canopylarvae (see Fig. 2.12). Sampling on April 8th did show a higher number of fourth and fifthinstars in the undergrowth, but this was not significant and was probably due to downwardlarval dispersal rather than to differences in the development rates of larvae in theundergrowth and canopy. Mass rearing of caterpillars from both undergrowth and canopiesdemonstrated no differences in the incidences of parasitism.It is difficult to suggest why C. albi cans is more successful in parasitizing larvae onbirch than on blueberry. There is no strong evidence to suggest that there are differences insynchronies of larvae and leaf development on either blueberry or birch, or that the mixedvegetation of birch stands presents a refuge. In this study, caterpillar foraging behaviourappears as the most likely determinant for the lack ofC. albi cans success on blueberry. Theswitching behaviour of larvae possibly reduces the success of the parasitoid. A morethorough examination of the morphological and chemical nature of these hosts and a betterunderstanding of the chemical stimuli involved in C. albicans oviposition is needed to helpforecast the success of the parasitoid on these hosts.66Larval mortalityLarval mortality and 'death due to unknown causes' was high in 1992. At three of thefour sites klarv was the greatest mortality factor. At the remaining site (BBI) klarv wassimilar to kpup, the highest mortality factor at that site. A number of factors are responsiblefor larval mortality. These include predation by birds (Embree 1965a, Roland et al. 1986),or arthropods (I have observed predation of winter moth larvae by spiders on numerousoccasions), viral or other diseases (Wigley 1976), starvation (Tenow 1972, Wint 1983) ordispersal to unfavourable sites (Holliday 1977). In Britain, larval mortality was the keyfactor influencing winter moth populations, and resulted mainly from asynchroniesbetween bud burst and egg hatch (Varley and Gradwell 1968, Varley et al. 1973).Many authors have investigated the importance of synchrony between host plants andherbivores (West 1985, Feeny 1970, Hunter et al. 1991, Valentine et al. 1983). On oak,apple and alder, winter moth favour young leaves (Feeny 1970, Holliday 1977, Kikuzawaet al. 1979, Wint 1983). Topp and Kirsten (1991) found that on willows, which continueto produce leaves throughout the summer, larvae that feed on young leaves later in theseason, developed more quickly, but were lighter than those feeding on young leavesearlier in spring. Feeny (1970) analyzed oak leaves during development of both the leavesand the winter moth larvae. Early feeding coincided with maximum leaf protein content andminimal leaf sugar content, demonstrating nitrogen availability as a possible limiting factorto spring feeding. There was also an increase in the amount of oak leaf tannins during thesummer. However, Feeny suggested that leaf toughness was the most important factorinhibiting late season feeding by larvae. Wint (1983) also found leaf toughness to mostconsistently affect pupal weights.67Because of the wide range of winter moth host plants, it is difficult to suggest asingle factor which may inhibit late season development. It is likely that a combination offactors are important. The early spring of 1992, with consistently high temperatures,suggests that leaf development may have been faster that year, and could have hadimportant consequences for the winter moth.Authors working with winter moth on Sitka spruce suggest that nutrient levels andasynchronization of bud-burst and larval hatch has no influence on larval densities (Wattand Macfarlane 1991, Hunter et a/. 1991). Five years of field data showed poorsynchronization of bud-burst with eclosion. Larvae survived for up to four weeks beforebud-burst declining to low densities only after bud-burst (Watt and MacFarlane 1991).These observations may be a result of Sitka spruce being taxonomically a very differenthost plant from most other winter moth hosts that have been studied. Spruce, since it isevergreen, has needles available even before bud-burst. These needles are less preferredand nutritionally less desirable for the winter moth (Hunter et a/. 1991). However, theymay support larvae for some time. Unfortunately, Watt and MacFarlane (1991) do notpresent information on the condition of larvae when the buds did burst, nor has there beenany feeding experiments to investigate survival of larvae on old foliage of spruce. Thoughthe results with Sitka spruce show poor synchrony, they do not undermine the importanceof the phenological relationship between winter moth and its host plant, in this case Sitkaspruce.Kikuzawa et al. (1979) suggest that asynchrony with bud burst is not likely to be animportant factor in winter moth population dynamics on alder, since these havepseudoterminal buds. Birch also has a pseudoterminal bud type. However, the concomitantincreases in larval mortality and in pupal deformities in 1992 on both blueberry and birch,68suggest that there may have been a large asynchrony of larval eclosion and bud-burst onboth hosts in that year.Poor synchrony of larval eclosion with bud burst, as suggested for 1992, implicatestarvation and dispersal losses as important sources of mortality in that year. Edland(1971), in perhaps the most detailed study of winter moth larval dispersion, found that bothfirst and second instar larvae disperse. He implied that first instar larvae were dispersing inresponse to larval densities at individual stands. In stands of infested trees, 80% of the firstinstar larvae dispersed while on individually colonized buds there was no dispersal. Hunter(1990) found from laboratory trials that winter moth dispersal was independent of density.However, in his experiments he used buds that were unopened and so were anunfavourable food source. As such he could not really test for density dependent dispersal.I looked at dispersal of first instars on stage two blueberry buds and found no evidence fordensity dependence. Larvae, however, dispersed at a faster rate from smaller buds,suggesting that food limitation may be the major dispersal stimulus (see Appendix 1).Edmonds (in Cox and Potter 1986) found that daily ballooning was closely related toegg hatch of Douglas-fir tussock moth, which usually had occurred one to two hoursearlier. This may occur in winter moth also, since winter moth has diel hatching (Embree1970). Many authors suggest that newly hatched larvae of a number of species arepredisposed to disperse, even in the presence of preferred foliage (Cox and Potter 1986,McManus 1973, Capinera and Barbosa 1978). This might explain Edland's (1971)observations, with large numbers of larvae dispersing from heavily infested stands, simplybecause greater numbers of larvae were available to disperse. Wint (1983) suggested thatfood deprivation may be important. He found that after one day of starvation larval activitybegins to increase, but eventually declines if starvation continues. In 1992, starvation mayhave been severe especially for early instars due to an early spring. Many of the larvae69which dispersed may have died due to lack of available food after dispersal or due toground predation (see Weseloh 1985). On blueberry stands, mortality of dispersing larvaemay have been high due to a lack of suitable undergrowth. Of the birch sites, RNP is likelyto have had most mortality because of the low tree density (see Table 3.1) (Holliday 1977).The concurrent increases in 'death due to unknown causes' and larval mortality areimportant to note. Death of prepupae due to unknown causes included death due todesiccation, failed pupation, deformed pupae failing to emerge and possibly unidentifiedpathogens. The most common causes of death among prepupae were failure to pupate anddeformities. These imply poor nutrition and thus unfavourable foliage quality. Larvae thatsurvived to prepupae may have lacked nutrients essential to survive pupation. Furthermore,because of lower densities of larvae in 1992 than in the previous year, one should expect anincrease in pupal size. This was not observed. The occurrence of lighter pupae in 1992, inspite of lower densities again indicates poor foliage quality (see Morris 1972,Danthanarayana 1975, Barbosa and Capinera 1977, Heliiivaara et al. 1989). Larger pupaeoccurred only at RNP in 1992, suggesting that the different histories of defoliation or thedifferent ages of the trees at the two birch sites (see Table 3.1) may have had an effect.Changes in leaf quality either due to current or proceeding years of defoliation havebeen noted for many plant hosts, but especially among birch (see references in Haukioja1990). These changes may influence herbivore quality. One such induced response thatmay have important consequences for populations of early feeders is that previousdefoliation can affect the timing of spring bud-burst (Heichel and Turner 1976). For birch,some authors suggest that damage to apical buds (as occurs with winter moth attack)actually has ameliorative effects on subsequent larval feeders (Haukioja 1990). There mayhave been an induced response among the birch at BI, because of extensive defoliation in1990. Whether damage to blueberries induces responses is unknown.70Roland and Myers (1987) found reduced fecundity of winter moth which had fed onapple heavily defoliated in the previous year. Those which had fed on previouslymoderately defoliated plants had increased fecundity. Roland and Myers discounted longterm effects of foliage quality on winter moth fecundity as causing winter moth decline.Kikuzawa et al. (1979), studying winter moth on alder in Japan, suggest that reductions infecundity may have lead to a decline in outbreak there. This reduced fecundity could beattributed to a lack of food in the previous year. Unfortunately, it is difficult to suggestwhether there actually was a decrease in fecundity, because comparisons of adult weightson alder from non-outbreak years were not presented. Laasonen and Laasonen (1987)suggest that there may be induced resistance of birch (B. pendula (Roth.)) to larvae ofOperophtera fagata (Scharf.). This finding was based on observations of increasednumbers of dry or wrinkled larvae as an outbreak in Isosauri (Finland) continued.However, they suggest that other unidentified (annihilating) factors also played a role in theoutbreak decline. It is difficult to suggest what may be occurring on either birch orblueberry in this study. Current year larval mortality shows no indication of being densitydependent (see Figs. 2.7 and 2.8), discounting current year induced responses, and longterm data are not available to examine a delay in density dependence which would resultfrom long term induced responses.DiseaseIn Britain, disease is an important factor responsible for larval mortality (Wigley1976). Circumstantial evidence from Canada has suggested that parasitoids may beresponsible for transmitting disease in winter moth populations (Embree 1966,Cunningham et al. 1981). In Nova Scotia, no disease was apparent among winter mothpopulations until after the introduction of C. albicans and A. flaveolatum. It has been71suggested that the virus appearing in the population of winter moth at that time wastransmitted from the native Bruce's spanworm (Embree 1966). Polyhedrosis virus isoccasionally prevalent among spanworm in both eastern and western Canada. Decline ofspanworm outbreak in the Maritimes and Quebec has been attributed to this virus. TheNPV has been described as a simply embedded (unicapsid) NPV and is classified asBaculovirus subgroup A (Ives and Cunningham 1980), which is taxonomically verysimilar to that of the NPV from winter moth in England (Wigley 1976). It is notunreasonable to suggest that cross infection may have occurred.The presence of parasitoids, intimately associated with the winter moth populations,may be necessary for viral transmission. Winter moth larvae killed by virus release viroidsinto the environment. The parasitoid attracted to host damage could transport the virus onit's body surface and contaminate new areas. One problem with this idea is that C.albicansoviposits when late instars are present. Late instars are more resistant to virus, necessitatinglarger amounts of virus to successfully kill a caterpillar (Wigley 1976). The presence ofearly instars in the population may be important in maintaining the infection and increasingthe quantity of innoculum in the environment. I tried cross contamination experiments in1991 using fourth instar larvae of winter moth and spanworm. I used winter moth NPVfrom Britain and spanworm NPV from eastern Canada at heavy doses. There was noindication of cross infection in spite of high mortality of primary hosts. Lack of sufficientnumbers of spanworm larvae prevented more extensive experimentation (see Appendix 4).The incidence of disease in the lower mainland population is low, and certainlydisease played no role in the 1992 decline. Microsporidia were found at BBI in 1991.Larvae exhibited pronounced symptoms and there was 100% mortality of infected larvae. Anumber of microsporidia (Phylum Microspora, see Poinar and Thomas 1984) have beenfound to infect winter moth. In 1956, Kreig described Thelohonia cheimatobiae from7 2winter moth populations in Germany. Canning (1960) described 2 species,Cytosporo genes operophtera and Orthosoma operophterae, from larvae at Wytham Woodin England. A third species, Nosema wistmani, purr. & Skat. was identified in 1979 (seeCanning et al. 1985a for a review of recent taxonomic changes and species descriptions).Microsporidia are normally transmitted between hosts when spores are ingested (Canningand Barker 1982, Canning et al. 1985b). The microsporidia found in 1991 were notidentified and failed to turn up in eggs, larvae or adults in 1992. It is possible that thisinfection was accidental from another species of Lepidopteran.Pupal mortalityPupal mortality (kpup) has been the most consistent mortality factor over the fouryears of study and at each of the four site. Pupal mortality has been more important thanparasitism in bringing about population decline in Richmond. These results are inagreement with Roland's (1988) observations for Victoria, and with the situation in Britain(Varley et al. 1973). However, the four sites differ in the trends observed in pupalmortality. This is probably because of different predator assemblages at the different sites(see Chapter 3). Predation at all sites, except BBII, appears to be temporally densitydependent (see Fig. 2.19), but there is insufficient data to strongly support thisassumption. This differs from the situation in both Nova Scotia and Victoria, where pupalmortality increased during the decline in outbreak of the winter moth, and was thusinversely density dependent (Embree 1966, Roland 1988). Pupal mortality in the lowermainland may have been high for some years and did not go through such an inversedensity dependent process; possible mechanisms behind this process will be addressed insubsequent chapters.7 3Some differences between the mortality occurring in Richmond and that which occursin Britain, Nova Scotia and on Vancouver Island include a high rate of parasitism (whichdiffers from the situation in Britain), and the lack of a period of inverse density dependentpredation (as occurred in Nova Scotia and Vancouver Island). Therefore, this systemappears to be somewhat midway between that of Britain and those of the other sites inCanada.In 1992 the outbreak declined in Richmond. However winter moth is continuing tospread in the Lower Fraser Valley. Examination of tree bands in southern Vancouverindicate that densities of winter moth are now very high. In the winter of 1992 the numbersof adults per m2, estimated from sticky traps, was double that at BI (18.7 + 6.88m2 vs.9.48 + 2.3m2), the highest level recorded from Richmond. Parasitism by C. albicans is stillvery low, about 3% in 1992. It will be interesting to see whether C. albicans will reachhigh levels of parasitism in southern Vancouver and cause population decline, as Embree(1991) would expect, or whether there will be a population decline with low levels ofparasitism, as has occurred in some areas of Nova Scotia and Vancouver Island (Roland1988) and in Oregon (Kimberling et al. 1986). Levels of predation are not expected to behigh in Vancouver, since the urban environment is not conducive to large numbers ofpredators (i.e. ants, beetles, etc.).74CHAPTER 3PREDATION OF WINTER MOTH PUPAE IN THE LOWERFRASER VALLEY OF BRITISH COLUMBIA.3.1 IntroductionLong term studies by Varley and Gradwell (1963a, 1968, 1973) at Whytham wood,in England, indicated that winter moth soil mortality is density dependent and thus regulateswinter moth populations. This 'soil mortality' included all mortality incurred from the timeof pupal drop until adult females were caught ascending trees in winter. Populations inCanada are also regulated by soil mortality (Roland 1986a, 1988, 1990b, 1992, McPhee etal. 1988, Pearsall 1992 and see Chapter 2). East (1974) found at Whythm Wood a total soilmortality of 81% of which 62% was incurred to the pupae, 14% to prepupae and 5% toadults. Pearsall (1992) suggests that in Nova Scotia only a very small amount of soilmortality (between 3 and 37%) is due to mortality of adults or prepupae. Similar resultswere also found in this study (see Chapter 2). Pupal mortality is therefore most importantand has been attributed mainly to predation by generalist predators.Generalist predators have often been recognized for their potential to regulate preypopulations (Luff 1983). Reasons for this include the fact that they have broad diets whichbuffer fluctuations in the abundance of their prey, and that they are not significantly limitedby the maximum attack rates of their functional responses (Hassell 1966, Hassell and May1986). At Whytham wood a number of studies have been carried out in an attempt toidentify the generalist predators of winter moth. Frank (1967a) using serological andradioactive tracer techniques identified the carabids, Pterostichus madidus (F.) and Abaxparallelopidus (Pill. et Mitt.) and a staphylinid, Philonthus decorus (Gr.), as important.The carabids P. cupreus (L.) and P. melanarius (Ill.) and larvae of the elaterid Athous7 5haemorrhoidalis (F.) were also shown to consume pupae, but were not sufficientlyabundant at the site to be of much consequence. Frank also found small mammals,including the mouse Apodemus sylvaticus (L.) the vole Cleithrionmys glareolus (Schr.)and the shrew Sorex araneus (L.) to take pupae. Unlike invertebrate predators, these werefound to actively search out the winter moth pupae. Frank (1967b) estimated that 39% ofhealthy pupae in the soil were being predated, with P. decorus responsible for most of this(53.6%), and P. madidus and S. araneus responsible for most of the remaining predation.Graphing P. decorus abundance against winter moth population density at WhythamWood showed signs of cycling which is expected from a delayed numerical response(Frank 1967b). Kowalski (1976) demonstrated an aggregative numerical response of P.madidus and a numerical response of P. decorus to winter moth density. However, hesuggested that heavy losses of P. decorus, due to predation by small mammals in thewinter, may have impeded a generation response. This stabilization of the P. decoruspopulation inhibits the occurrence of a delayed density dependent effect promoting stabilityof the winter moth population. Buckner (1969) suggested that small mammals, particularlyS. araneus, were causing much greater mortality than predaceous beetles at WhythamWood. East (1974) however, has refuted this, estimating that carabids were responsible forup to 38%, staphylinids 30%, and small mammals only 4% of the predation at WhythamWood. Small mammals may be more important as predators of carabids and staphylinids(see Parmenter and MacMahon 1988 and Grum 1979) setting the levels of the generalistpredator populations.There are few studies of winter moth predators in Canada. Embree (1965b) estimatedsoil mortality in Nova Scotia at between 37% to 92% over the period of winter mothdecline. He suggested as much as 31% of this was due to small mammal predators (he didnot investigate other predators). Pearsall (1992 and in press) studying pupal predators inNova Scotian apple orchards, found high levels of predation (68.6%), as much as 70% of76which she attributed to beetle predators. She indicated Carabus granulatus (L.), C.nemoralis (Mull.), Pterostichus coracinus and Harpalis rufipes (DeGeer) (all carabidsintroduced from Europe) as likely predators and suggested that there is a functionalresponse of these predators to winter moth density.Roland (1986a, 1988, 1990b) working on Vancouver Island, estimated soil mortalityto be as high as 86% at the University of Victoria research orchard and 95% at Mt. TolmiePark (his oakwood site). Roland (1990b) suggested that beetles of 0.5 - 1.5mm in widthare responsible for the bulk of predation losses (since these could fit through a 2mm mesh).Examination of predated pupae implicated carabid or staphylinid larvae. Three carabids(Pterostichus melanarius, C. nemoralis and Calathus fuscipes (Goeze) and twostaphylinids, Staphylinus aeneocephalus (DeGeer) and Ocypus melanarius (Heer), all ofEuropean introduction were also implicated as important predators due to their abundance atthe sites. He did not investigate the importance of small mammals. Mortality of pupae onVancouver Island was spatially density dependent. However, using beetle removalexperiments, Roland (1986a) suggested that winter moth was not affected by theabundance of beetle predators. This points to a lack of either aggregated or numericalresponses. Mortality of C. albicans pupae was reduced and there were higher incidences ofvirus infected or otherwise dead pupae in the absence of beetle predators. This suggeststhat virus killed or otherwise failed pupae are normally preferentially taken by predators.These results point to differences in the predation occurring between unhealthy and healthywinter moth pupae which may be important in the dynamics of the system.Roland (1986a, 1988) reanalysed Embree's data from Nova Scotia and suggestedthat pupal mortality had been more important in bringing about the winter moth decline thanwas previously suggested. A similar trend had occurred on Vancouver Island. On bothoccasions, pupal mortality had increased rapidly from low levels before C. albicansintroduction, to high levels at the time of winter moth decline. This occurred in spite of 2577years of parasitoid free high populations in Nova Scotia, but only 6 years on VancouverIsland. He has suggested that somehow C. albicans may have influenced soil mortality andbrought about this decline (both in Nova Scotia and on Vancouver Island). Roland (1986a)proposed three mechanisms by which this may have occurred;1) the availability of pupae (C. albicans) over a longer period of time, 10 months asopposed to 5 - 6 months for unparasitized pupae, may allow predators to build up anumerical response to winter moth populations,2) unparasitized pupae may be more likely to be taken by predators, than parasitizedpupae, therefore, as the proportions of parasitized pupae in the soil increase the remainingunparasitized pupae become increasingly susceptible to predation, and3) that parasitoids act as vectors of disease, infecting fifth instar larvae and leading topupal death and thus increasing soil mortality.These mechanisms are not exclusive. The third mechanism could not be tested in Richmondsince the incidence of disease was too low to suggest that it may play any role in populationdecline (see Chapter 2).This part of the study has two main objectives: firstly, to look for possible predatorsof winter moth and C. albicans pupae on the Lower Fraser Valley, comparing predatorsand predation at the blueberry and birch sites. and secondly, to examine the possibilities ofa link between C. albicans parasitism and pupal predation, addressing Roland's first twomechanisms. I will examine patterns of predation in the Lower Fraser Valley from fouryears of data and compare observations here with those of Nova Scotia and VancouverIsland.783.2 ProceduresStudy sitesA total of eight sites have been examined throughout the four years for which data arepresented. In 1989 and 1990, seven sites were studied, these were; BI, RNP, RNPa (a bogsite), BBI, BBIa (a commercially harvested site immediately adjacent to BBI), BBII andDND. In 1991, RNPa and BBIa were not studied and in 1992, two further sites, BIa (a siteimmediately adjacent to BI but with markedly different habitat) and RNPII, were studied,but DND was not studied. BBIa, RNPa and DND will only be mentioned in the sections onfield planting of pupae and trends in predation. These sites are indicated in Figure 3.1 andhabitat descriptions are presented in Table 3.1 (see also Fig. 2.1 and information in Chapter2).Pitfall trappingPitfall trapping was carried out at BI, RNP, BBI and BBII in 1991 and 1992, afurther two sites, BIT and RNPIll, were trapped only in 1992. In 1991, traps were set up onJune 18th and trapping was discontinued on September 4th• In 1992, traps were set outfrom May 14th until the spring of 1993. Each site had 20 traps laid out in two transects of10 and spaced at approximately 15m intervals.Traps consisted of 500m1 plastic goblets. Holes were dug with a core sampler so thatthe goblets fitted neatly into the holes. The earth around the mouth of the traps wascompacted so that no part of the traps were protruding from the soil surface. Slits weremade at the sides of the goblets to drain water. In 1992, 1cm2 wire-mesh was placed overthe mouth of each trap to prevent entry of small mammals (shrews, voles and mice). Themesh was kept in place by wire pegs. The mesh was removed from traps at the birch sites7 9Figure 3.1. Predation study sites at Richmond Open circles indicate bog sites,open squares indicate blueberry sites and closed squares indicate birch sites.The years in which studies were conducted at each site are presented.Highway 99 is indicated by an encircled'99'.80Table 3.1. Habitat characteristics at nine sites in Richmond. Studies on winter moth pupal predation have beencarried out at these sites between 1989 and 1992. Densities of birch and hemlock are presented as the number ofplants per 50m2. Percentage cover of salal, labrador tea and heather are also presented. For undergrowthblueberry, percentage cover is presented while for blueberry plots the numbers of plants per 50m2 are presented(indicated by an asterix). Ten randomly chosen plots of 50m2 were sampled at each site.HABITATSITE Birch %Coversmall medium large Hemlock Salel B.berry lablea heatherBlueberry! 3.1 13 10.2*Blueberry Ia 7.08*Blueberry!! 1.2 5.7 7.08*Department of National 46.3 0.2 sporadic 23.9 47 2Defence landsRichmond Nature Park 0.6 3.8 7.5 23.5 62(Bog) (a)Richmond Nature Park 2.7 10.1 7 0.3 59 8.5Richmond Nature Park II 1 5.5 0.75 43.75 5Birch! 11.6 15.7 3.4 0.1 76.3 6.7Birch I (a) 33.1 35.8 4.1 0.4 25.6 14.5in the autumn to prevent leaf litter from concealing the trap entrances. In 1991 trap catcheswere removed every two weeks. In 1992 trap catches were removed weekly till October7th, and from then at irregular intervals. All predacious beetles were identified to specieslevel and recorded. Spiders, woodlice, milipedes, centipedes and non predaceous insectssuch as Collembola, flies, etc. were not recorded.The number of predatory beetles at each site in each year were analysed using theShannon-Weiver diversity index (1-11 and Frmax) and the species evenness index J'(r=1-17111max) (Ludwig and Reynolds 1988) Also a coefficient of community index wasused (c= 2w/a+b [where a is the total number of individuals in sample a, b is the totalnumber of individuals in sample b, and w is the sum of the lower scores for each species])(Screiber et al. 1987). The Shannon-Weiver index is the proportion of individualsoccurring in the species. The species evenness index is a measure of relative diversity. Itmeasures the expression of species dominance, which is 14. A low value indicates lowdominance, a high value indicates high dominance. The coefficient of community index isused to compare similarities in the faunal composition of communities. Values approachingzero are dissimilar, those approaching one are similar.In 1991, baited pitfalls were set up to examine the possibility of predators respondingto olfactory stimuli to locate winter moth pupae. Traps consisted of goblets with muslinbags of bait, the bait consisted of either 1) 4 moth pupae without the cocoon or, 2) 4 flypupae without the cocoon, suspended from a skewer which lay across the mouth of thetrap. Ten traps, five of each type, were set up at each site. The traps were collected at thesame time as the control traps (unbaited) on three dates (July 3rd, July 17th and August82Subsoil trapsIn 1991 and 1992, subsoil traps (Fig. 3.2) were set out to identify possible subsoilpredators. These traps were specific for small sized predators (mainly beetle larvae) andindicated whether subsoil predators may be responding to olfactory stimuli emitted from thepupae. The basic trap design consisted of a mesh over a mason-jar lid (the trap top), whichwas tightly secured to a petri-dish lid with drain-holes (the trap bottom). A circular sheet ofplastic was coated with tanglefoot and placed in the petri-dish. Bait was suspended byflower wire between the mesh and the tanglefoot. Six trap types were used. In 1991, eachbaited trap had 4 pupae and empty cocoons were not used as bait. In 1992, owing to ashortage of available pupae, baited taps had each only 2 pupae and traps with a 0.01mmmesh were not used. In 1991, each site had 30 traps, 10 of each type. In 1992, only birchsites had traps each with 20 traps, 5 of each type (Table 3.2). Traps were collected weeklyin 1991 and on every second week in 1992.Table 3.2. Subsoil trap types used in experiments at Richmond, in 1991 and 1992.TRAP TYPE MESH SIZE BAIT YEARTYPE I lmm complex lmm Complex of pupae 1991TYPE II 0.01mm complex 0.01mm Complex of pupae 1991TYPE B1 lmm control lmm None 1991/1992TYPE IV lmm moth lmm 0. brumata pupae 1992TYPE V lmm fly lmm C. albicans pupae 1992TYPE VI lmm cocoon lmm Empty cocoon 199283.^w 7 7^•^" "^7 7 7 7 w .4,4A. ...b. •Ao. A, A.^ A. A. A. AL A. A A 4.DRAIN HOLESSUSPENDED BAITTANGLEFOOTSUBSOIL-PREDATOR TRAPlmm MESHFigure 3.2. Subsoil predator trap used in predation studies at Richmond, B.C..84Field planting of pupaePupae were set out at each of the four main field sites from late June until Octobereach year from 1989 to 1992 to estimate the levels of predation at each site. Pupae werealso planted at BBIa, RNPa and DND in 1989 and 1990 and at DND also in 1991. Tethersconsisted of groups of four pupae attached by flower wire to a central wooden skewer.Pupae were spaced at about 10cm intervals along the wire. Scures were placed alongtransects at intervals of about 5m, and cocoons were covered with a layer of humus about3cm deep.From 1989 to 1991, 20 scures were placed at each of the sites. Pupae attached to thescures were not examined and so the pupae are taken to represent the natural winter moth-Cyzenis pupal complex. In 1992, pupae were examined before tethering. A hole was madeat one end of the cocoon and the pupa taken out with a forceps, all pupae were weighed andcategorized. Five types of pupae were set out in the field; 1) small moth pupae (0.01 -0.02g) and 2) small fly pupae (<0.01g) were set out at both blueberry and birch sites, 3)large moth pupae (0.02 - 0.03g), 4) large fly pupae (0.01 - 0.02g), and 5) 20 dead pupae(death due to unknown causes) were set out at each of the birch sites.Tethers were set out after pupation was complete, generally about 1-2 weeks later. In1989, they were set out on June 25th, in 1990 on June 26th, in 1991 on June 29th and in1992, on June 25th. Pupae were examined each week in 1990 and 1992 and only once in1989 and 1991. All pupae were collected on October 7th each year before adult winter mothemergence. In 1992, extra pupae were placed out at the field sites on July 29th, to maintainN > 20.85In 1991 and 1992, pupae were out-planted in beetle exclosures (Fig. 3.3) to estimatethe levels of pupal predation due to subsoil invertebrates other than adult ground beetles.Ten exclosures were set out at each site. The exclosures were similar to the traps used forsubsoil predators. Exclosures consisted of two mason-jar lids covered with lmm mesh, 4pupae were suspended between the lids by flower wire. In 1992, pupae were separated intofly or moth groups. The exclosures were examined weekly in 1991 and every two weeks in1992.Arena studiesIn 1991 and 1992, arena studies were carried out to identify which beetles could eatwinter moth and C. albicans pupae. Arenas consisted of 51. plastic containers which werecovered with a lid of lmm mesh. The containers had 2cm of moistened peat with dry leavesfor beetle cover. Pupae were placed in the containers either as pupae without cocoons at thetop of the peat, or cocooned pupae buried beneath the soil. In 1991, groups of four pupaewere used in each experiment and experiments were replicated five times. In 1992, onlytwo cocooned pupae were used in each arena. Whenever cocooned pupae were being usedthey were only opened for examination of the pupal condition after the trials. Thisprevented biases from disturbance of the pupae or cocoons.Beetles were taken from pitfall traps, fed on canned dog-food for some time and thenstarved for two days before being placed in the arenas. The peat was frequently sprayedwith water since the beetles are very susceptible to desiccation. Beetles were left incontainers for one week.86.^•■^T T^ T 7 •^.;b.W ww&2 WW411. AY.^ .10454. .^. "Pr •,"^ .,^Air^ .a.^A...^.116^A.^ •^•á. 7BEETLE EXCLOSURESUSPENDED BAITMESHFigure 3.3. Beetle exclosure used in predation studies at Richmond, B.C..873.3 Results3.3.1 Trends in predationFigure 3.4a indicates the levels of predation at 3 blueberry sites in Richmond,between 1989 and 1992. Sites BBI and BBII have been discussed extensively in Chapter2. BBIa is a commercial blueberry plot immediately adjacent to site BBI, studies at BBIaceased after 1990.Data from 1989 and 1990 indicate that there was a dramatic increase in predation ofthe winter moth-Cyzenis pupal complex at all three blueberry sites in 1990. This increasecoincides with increases in the winter moth population at that time (see Fig. 2.6b). In 1991,levels of predation at BBI and BBII changed little from levels observed in the previousyear. At both sites a slight decrease is apparent. There was a further decrease in predationlevels at BBII in 1992. This corresponded with the a decline in the winter moth populationthere. At BBI however, pupal predation increased in spite of a decline in the winter mothpopulation.Trends in the predation of the winter moth-Cyzenis pupal complex at two bog sites(RNPa and DND) and two birch sites (RNP and BI) are presented in Figure 3.4b, the twobirch sites having been discussed extensively in Chapter 2. Studies ceased at RNPa (a bogsite at Richmond Nature Park) in 1990 and at DND, a second bog site, in 1991. Both ofthese sites are known to have had very low densities of winter moth (at DND prepupaldensities were as follows: 1989 = 0m-2, 1990 = 0m-2 and 1991 = 2.22m-2, [ls1 =20 pupaldrop trays]: see also Table 2.7). Winter moth may have been kept out of these sites eitherdue to a lack of suitable host plants or due to a very high predation pressure. By 1990,predation at both sites had increased to about 100%. In 1991, there was a slight decline in88BBla BBI19921989^1990 1989 1990 1991^1992 1989 1990 199110090807060BA10090807060RNP- RNPa1989 1990 1989 1990 1991 1989 1990 1991 1992 1989 1990 1991 1992YEARFigure 3.4 Levels of predation on the winter moth-Cyzenis complex.Predation estimated from tethered pupae at a) Three blueberry sitesand b) two bog sites (squares) and two birch sites (circles). Barsindicate standard errors.89the levels of predation at DND. The predators at both these bog sites were not studied indetail. However, ants were abundant at the sites and are presumed to be the main predators.At the two birch sites, similar trends to those occurring at BBII were observed overthe 4 years. Between 1989 and 1990, levels of predation increased, underwent little changein 1991, and declined in 1992. This decline in predation was most dramatic at BI.Changes in the levels of predation at each site (except at site BBI in 1992) reflectedthe changing densities of winter moth. This may be due to different predator assemblages atthe sites. Predation therefore appears to be temporally density dependent at each of the fourmain study sites. This density dependence operates at current year densities, i.e. there isnot a delay (see Figure 2.9). However, similar levels of predation were observed at siteswith very different densities of winter moth pupae. Therefore, predation is not spatiallydensity dependent. Interestingly, in spite of very different densities of winter moth at eachof the sites, the levels of predation were similar each year.3.3.2 Predator assemblagesA total of 951 beetles were caught in pitfall traps in 1991, and 3190 beetles in 1992.Higher catches in 1992 were due to a longer trapping period with more pitfall traps (26740trap days in 1992 compared with 5740 in 1991). Since trapping began earlier in 1992, therewas increased diversity at each site (Table 3.3), this was due to the large numbers ofstaphylinids, beetle larvae and small sized beetles (i.e. Bembidion sp. and Agonum sp.)which were abundant only in May and early June. Table 3.3 presents trapping results fromfour sites over two years, with additional sites, BIT and RNPII trapped in 1992. Non-predatory beetles are not presented in the table. These included Silphidae, Histeridae,Hydrophilidae and Tachinus sp. (Staphylinidae), all of which are attracted to carrion, andElateriodea, Curculoidea, etc., which are herbivorous. Very small staphylinids are also90Table 3.3. Predatory beetles trapped at four sites in Richmond B.C. during 1991 and at sixsites in 1992. The Shannon-Weiver Diversity Index (H1 and Species Dominance (1-I, whereJ' is Pielou's Evenness Index), are presented for each site (Ludwig and Reynolds 1988). B.L.'sare beetle larvae and S's are staphylinids.SPECIES Total number^caughtBB I BBII RNP RNPI B! III!1991 1992 11991 1992 11991 1992 11992 11991 1992 11992Amara aurata 45 81 1 2 6A.littoralis 55 22 5 1A.laevipennus 3Calathus fuscipes 1 4 2Callisthenes wilkerii 1Carabus granulatus 11 7 21 5 8 81 3 7C.nemoralis 11 15 5 1 11 5Harpalus affinus 36 29 9H.rufipes 2 2Pterostichus algidus 3 16 5P.herculaneus 42 82 111 227 342 316P.melanarius 2 10 13 36 2 4 11 8Scaphinotus marginatus 8 60 1 23 88 25 106 104Bembidion sp. 7 6 260 16Agonum sp. 5B.L. 1 (Staphylinidae) 3 1 1 1 1B.L. 2 (Carabidae) 2 3 3B.L. 3 (Elateroidae) 14 12B.L. 5 (Carabidae) 1B.L. 6 (Carabidae) 5 1 1B.L. 7 (Carabidae) 1 6 11 15 20 1 2 3B.L. 8 (Carabidae)B.L. 10 (Lampyridae) 1B.L. 11 2 3 3 2 4B.L. 12 (Lampyridae) 1B.L. 13B.L. 14 (Lampyridae) 2B.L. 15 1 1 1B.L. 16 3S. 3 (Philonthus sp.) 1 12 3 9 3 31S. 4 (Philonthus sp.) 1 1 2S. 5 (Philonthus sp.) 5S. 11 (Philonthus sp.) 1 3 1S.12 7S. 15 (Queduis sp.) 1Total^number^caughtTotal^trap^daysDiversity^index^(H')Species^dominance^(1-r)15514001.370.2019144801.950.186114001.670.0318744802.110.244714000.400.6411643400.900.4851144801.370.5227515400.700.6862344801.330.4956644801.940.3391excluded from the table. These were deemed too small to be functionally capable of feedingon winter moth pupae. This was borne out in arena studies. Notiophilus sp., Leistusferruginosis Mann., and Loricera decempunctata Eschs. have also been excluded. Theseare diet specialists which feed on Collembola and mites (see Hengeveld 1980a, 1980c,Thiele 1977). Therefore, Table 3.3 represents the potential predator population at the sites(Appendix 5 is an expanded version of Table 3.3 and includes all groups that weretrapped).Some distinct differences between the beetle assemblages at the different sites areimmediately apparent. In both 1991 and 1992, the highest diversities were at blueberrysites. Birch sites had low diversities since these were dominated by two species P.herculanieus Mann. and S. marginatus. Fisch.. Among the birch sites, BII and RNPI hadthe greatest diversities. This may be due to less ground cover vegetation at both of thesesites. Vegetation may have inhibited beetle movement at the other birch sites. Diversityindices have often been criticized as being biologically meaningless. Species evennesshowever, is more realistic since it is a comparison of H' against H'max. Species evenness(I') at birch sites varied from 0.33 to 0.68, while at blueberry sites J' varied from 0.03 to0.24, indicating that birch sites are dominated by few species. The most commonlycaptured species was P. herculaneus. This was the most abundant species at all the birchsites. Two other Pterostichus species, P. algidus Lecon., and P. melanarius generallyoccurred with this species, but at much lower numbers. P. melanarius was the onlyPterostichus species occurring at the blueberry sites, and was not as common as those atbirch sites. All three Pterostichus species were active at the sites throughout the year,although they were most abundant before mid-August. Birch sites had higher numbers ofstaphylinids and of the cychrisized forml, S. marginatus.1 Cychrisization refers to the narrowing of the head and thorax of beetles asan adaptation for predation of snails.92Figure 3.5 indicates that in general, sites tended to be different in terms of their faunalcompositions. Site BBI had a very different assemblage of beetles when compared to allother sites. Four beetle species which occurred at BBI appear to be most suited to blueberryhabitat, these include Amara aurata DeJean, A. litteralis Mann., A. laevipennus Kirby, andHarpalus affinus (Schr.) (see Appendix 6 and Lindroth 1961-1969). These species (exceptA. laevipennus) were very abundant at BBI, but were only occasionally captured at theother sites. Scaphinotus marginatus which was common at the other sites was completelyabsent from BBI.The very different assemblages of beetles at BBI and BBII is surprising. Thepredator assemblage occurring at BIM was more similar to those occurring at the birchsites. These included the species L. decempunctata (see Appendix 5) and S. marginatus.The Amara species and H. affinus did not occur at BBII. However, these two blueberrysites represent very different habitats. The blueberries at BBII are completely overgrownand the ground is therefore moist. BBI on the other hand, is a dryer habitat. It isperiodically mowed and so it is more similar to commercial blueberry plots. Most of thespecies found at BBI are favoured by its dryer habitat (Lindroth 1961-1969, Jeanne' 1967).BBII had a diverse array of beetle larvae (see also later sections). Carabid larvae were themost commonly captured larvae, larvae of the larger species (i.e. Carabus spp.) appear tobe surface active. Pitfall trapping is not a good indicator of the abundance of beetle larvae.However, these data clearly suggest a higher incidence of surface active larvae at BBILThe pitfall tapping was not designed to examine abundances of small mammals.However, throughout the course of the trapping period a number of small mammals werecaught. In 1991, numbers were particularly high, a total of 73 small mammals, mainlyshrews, being captured. In 1992, covering the traps with wire mesh reduced the number to29. In 1991, most of the small mammals were trapped on the first week of trapping, 33 atBI and 6 at RNP. Birch stands had more small mammals than blueberry sites, total catches93BBIBBIIRNPRNPIBIBIII^I^I^I^i^i0 0.1 0.2^0.3^0.4^0.5Similarity CoefficientFigure 3.5. Cluster diagram (average linkage) of six sites based on thesimilarities of their predatory beetle faunas. Pitfall catches are from May 14 tillDecember 3, 1992. Note that similarity coefficients do not exceed 0.5.94being BI = 45 and RNP = 14, in 1991 with BI = 2, BIT = 7, RNP = 13 and RNPII = 6, in1992. At the blueberry sites catches were as follows; in 1991, BBI = 5, BBII = 9 and in1992, BBI = 1 and BBII =0. Therefore small mammals, either through predation of pupaeor ground beetles, are potentially very important in the dynamics of the system. Theinfluence of small mammals is expected to be greater at birch sites than at blueberry sites.3.3.3 Predatory beetlesIt is difficult to know which beetles are predators of winter moth pupae. Here anumber of experiments and observations are compiled to suggest which beetles may beimportant.Arena studies were carried out in 1991 and 1992. Only the more abundant beetleswere tested. These included A. aurata, A. littoralis, C. granulatus, C. nemoralis, H.affinus, P. algidus, P. herculaneus, S. marginatus and two common staphylinids. In 1992,only C. granulatus, C. nemoralis and A. littoralis were tested. Arena studies were difficultto undertake because mortality of the beetles in the arenas was high. Furthermore, tostandardize the experiments, beetles were pre-starved for two days. This may have resultedin beetles, in arenas, feeding on winter moth pupae when in the wild they normally wouldnot. The results therefore simply indicate which species are functionally capable ofconsuming winter moth pupae.Three species were found to be facile at consuming pupae (see Table 3.4). Thesewere H. affinus, P. algidus and P. herculaneus. All these species appeared to favour wintermoth pupae, as 50% of winter moth pupae were consumed by H. affinus while only 30%of the C. albicans pupae were taken. P. herculaneus took 100% of the winter moth pupae,but only 20% of the Cyzenis pupae and P. algidus took 50% of the winter moth pupae and30% of the Cyzenis pupae. The larger Carabus species have been noted elsewhere for their95Table 3.4. Ground beetle predators of winter moth-Cyzenis pupal complex.`+' indicates that predation was observed, ' - ' indicates predation did notoccur and `n.t.' indicates that no trials were undertaken. Naked pupae arepupae which have had the hibemaculum removed.BEETLE PREDATORS Naked Naked Cocooned0.brumata C.albicans pupaepupae pupae (complex)Amara aurata n.t n.t +A. littoralis n.t n.t +A. laevipennus n.t n.t n.tCalathus fuscipes n.t n.t n.tCallisthenes wilkerii n.t n.t n.tCarabus granulatus + + _C. nemoralis + + _Carabus sp. larvae _ _ +Harpalis affinus + + +H.rufipes n.t n.t n.tPterostichus algidus + + +P. herculaneus + + +P. melanarius n.t n.t n.tScaphinotus marginatus + _ -Staphylinids (Philonthus spp.) - - -96potential to consume pupae (Roland 1986a, Pearsall 1992). Unfortunately, both specieshad very poor survival in arenas and were not seen to consume cocooned winter mothpupae. Carabus larvae did however consume pupae, but this was only noted on oneoccasion. Similarly the Amara species were found to consume pupae, but this occurredonly on one occasion with A. littoralis (one pupa being consumed), and A. aurata (twopupae consumed). Buried winter moth pupae were taken on both occasions.Scaphinotus marginatus was not successful in handling winter moth pupae. Only onone occasion did an individual eat a pupa. This pupa did not have a cocoon and was notburied. S. marginatus is a snail eater and predation of pupae is unlikely. It may howeverrepresent an important mortality factor for pupating prepupae. All the staphylinids that weretested did not consume winter moth pupae.A number of observations were made in the field of ants and beetle larvae eatingwinter moth pupae, beetle larvae were particularly voracious predators and included thelarvae of staphylinids, Carabus species and Pterostichus species.No differences were observed between the catches from baited pitfalls and controls.In general the mean trap catches from baited traps were lower than means from unbaited.This is because of the low number of baited traps (N=5 for each type vs. N=20 unbaited, ateach sites). There were no significant differences between trap catches within each site oneach collection date.3.3.4 Subsoil trapsExclusion cages indicated that large numbers of pupae are consumed by insects thatcan fit through a lmm mesh. However, the exclusion of larger beetles did reduce theamount of predation. In 1991, 25% of pupae were taken from exclusion cages at BI and9715% at RNP. In 1992, levels of 15% and 20% were found at BI and RNP respectively. Atblueberry sites 40% were taken at BBI and 15% at BBII. Exclusion cages were not set outat the blueberry sites in 1992. These results suggest that exclusion of beetles greater thanlmm wide drastically reduces predation. The highest incidences of predation fromexclusion cages were at BBI (40%) which may be due to the very different predatorassemblage there, possibly with generally smaller beetle larvae.Perhaps the most striking evidence that beetle larvae are important predators of wintermoth-C. albicans pupae comes from 1991 exclusion traps. A number of beetle larvae andoccasionally ants, were caught in these traps (see Table 3.5). There was no significantdifference between the catches at different sites (2-way ANOVA on log +1 transformeddata, P = 0.604). However the differences between the trap types were significant (P =0.01). The lmm baited exclusion traps caught significantly more than either the unbaitedcontrols (P= 0.025, Tukey test) or the 0.1mm baited traps (P= 0.046, Tukey test). Thisdifference was due mainly to more beetle larvae in the baited traps. There was no differencebetween the traps when ant catches were taken alone (2-way ANOVA, P= 0.57), butdifferences were significant for beetle larvae alone (2-way ANOVA, P= 0.005), withsignificantly more beetle larvae in the baited lmm traps than in controls or 0.1mm baitedtraps (Tukey test, P< 0.05 level).In 1992 due to a shortage of pupae, exclusion traps were only set out at the birchsites. There was no effect of either sites or trap types on catches (2-way ANOVA on log+1transformed data, sites P= 0.941, traps P= 0.129). However traps did indicate (see Table3.6) that beetle larvae may be attracted to empty cocoons, as the highest numbers of beetlelarvae were caught in traps baited with empty cocoons.9 8Table 3.5. Mean numbers of ants and beetle larvae caught in exclusion trapsduring the summer of 1991. Three trap types were used, unbaited (control),baited 0.1mm and baited 1mm subsoil traps. Results are from ten traps at eachsite.SITE CONTROLlarvae^antsBAITED^0.1mmlarvae^antsBAITED^1 m mlarvae^antsB! 0 0 0 0 0.7 0.1RNP 0 0.2 0 0 0.6 0BBI 0 0 0.2 0.2 0.1 0.2BBII _0.4 0 0 0 1.1 0.5Table 3.6. Mean numbers of ants and beetle larvae caught in exclusion trapsduring the summer of 1992. Four trap types were used, unbaited (control),Cyzenis baited (fly), winter moth baited (pupa) and hibernaculum baited(cocoon) 1mm subsoil traps. Results are from five traps at each site.SITE CONTROLlarvae^antslmm^flylarvae^antslmm^pupalarvae^ants1 m mcocoonlarvae^antsB! 0 0 0 0 0 0 0.2 0RNP 0 0 0 0 0.2 0 0.8 099This evidence suggests that predatory beetle larvae may respond to olfactory stimuliemitted from the winter moth or C. albicans pupae and that these stimuli may emit from thecocoon itself.3.3.5 Timing of attackTether experiments in 1990 indicate that most of the predation on the winter moth-Cyzenis pupal complex occurred from June to August at each site. Within the first fiveweeks after pupae were set out, predation had reached 80% (Fig. 3.6). After early Augustthe rate decreased.In 1991, tethered pupae were only examined on August 6th. By this time predationhad reached about 50% at blueberry sites and 65% at birch sites. The predation rate in 1991therefore appears to have been considerably slower.In 1992, a similar trend to that of 1990 was observed at all sites (see Figs. 3.7 and3.8). Most of the predation occurred before August (that is, during the first 5 weeks).Interestingly, predation during the first week after tethering in 1992 appears to have beenvery low (less than 10% at blueberry sites) with rates increasing rapidly only on the secondweek. Pupae at the birch sites were not examined after the first week.3.3.6 Predation on different pupal typesIn 1992, the rates of predation were different for different types of pupae. In general,C. albicans pupae were more highly predated, though this was dependent on the site andsize of pupae in question. At RNP (Fig. 3.7a) there was no difference between the levels ofpredation on different pupal types (1-way ANOVA on arcsine transformed data, P =0.546). However, at BI (Fig. 3.7b) there were differences (1-way ANOVA on arcsine100July 10 August 7TIME (DAYS)A1004.4F41C.,.44.1U44WAt■^20—0-- BBI- -A- BBlaa-- BBI IOctober 30B —o— RNPaar — DNDOctober 30July 10 August 7—a-- BIIv- RNPOctober 30July 10 August 7Figure 3.6. Predation of tethered pupae at a) three blueberry sites, b) two bogsites and c) two birch stands in Richmond B.C., during the summer of 1990.Bars indicate standard errors.101. AUGUST 5^ OCTOBER 7100TIMEFigure 3.7. Cumulative predation of pupae at two birch sites a) RNP and b) BI,during the summer of 1992. Open circles indicate predation of small Cyzenispupae (<0.01g), open squares indicate large Cyzenis (0.01-0.020, closed squaresindicate small winter moth pupae (0.01-0.02g) and closed circles indicatepredation on large winter moth pupae (0.02-0.03g). Bars indicate standarderrors.102AUGUST 5 OCTOBER 7^QB10040 -A8060 -40 -20 -• AUGUST. 5TIME. OCTOBER 780 -60 -Figure 3.8. Cumulative predation of pupae at two blueberry sites, a) BBI and b)BBII, during the summer of 1992. Open circles indicate predation of smallCyzenis pupae (<0.01g) and closed circles indicate predation on large wintermoth pupae (0.02-0.03g). Bars indicate standard errors.103transformed data, P = 0.008). Small C. albicans pupae were more highly predated thanlarge winter moth pupae (P = 0.028, Tukey test) and large C. albicans pupae (P = 0.038,Tukey test), but not the small winter moth pupae (P = 0.07, Tukey test). At BBI (Fig.3.8a) there was no difference in the predation of C. albicans and winter moth pupae (T-teston arcsine transformed data, P = 0.741). However, at BBII (Fig 3.8b) C. albicans weresignificantly more predated (T-test on arcsine transformed data, P = 0.008).Looking at predation of pupae across sites, there were no differences in the predationof small C. albicans pupae (1-way ANOVA arcsine transformed data, P = 0.734).Predation of large winter moth pupae was dependent on site (1-way ANOVA on arcsinetransformed data, P = 0.001). There was significantly more predation at BBI (P = 0.005,Tukey test) than BI or BBII (P < 0.005 level, Tukey test), but this was not different fromRNP (P = 0.212). More large flies were taken at RNP than BI (T-test on arcsinetransformed data, P = 0.012), but there were no differences between predation of smallwinter moth pupae at birch sites (T-test on arcsine transformed data, P = 0.105).All these results suggest that different predators are responsible for predation at thedifferent sites and that pupae of different sizes or types (i.e. C. albicans, healthy, etc.) maybe taken by different predators within each site. Cyzenis pupae are more highly predated.Dead pupae were set out on July 29th at the birch sites. These consisted of pupaetaken from pupal drop trays and were deformed or dead due to unknown causes. Withintwo weeks these had received 100% predation.The fact that predation levels are highest early in the season indicate that beetles arethe most likely predators, since these were the most abundant at that time. At BI, BBI andBBII there appears to be a bimodal trend in levels of predation, with levels becoming low atthe beginning of August and increasing again one week later (Figs. 3.9 and 3.10). This1040.2^ b.. s`.. AUGUST 1 OCTOBER 3R,S ..^^.. ..^.0.4^. •••.^:.^ .^ .*0. ..0..^id:0.0AUGUST 1^ OCTOBER 3TIMEFigure 3.9. Comparisons of the weekly predation levels of small Cyzenispupae (open circles) and large winter moth pupae (closed circles) at twoblueberry sites, a) BBI and b) BBII in Richmond B.C..0.80.6105AUGUST 1 OCTOBER 3A^0.8AAa.0.60.40a. 3.10. Comparisons of the weekly predation levels of small Cyzenispupae (open triangles), large Cyzenis pupae (open circles), small winter mothpupae (closed triangles) and large winter moth pupae (closed circles) at twobirch sites, a) RNP and b) BI in Richmond B.C..106suggests that Pterostichus spp. may be important since these also appear to be bimodal inoccurrence (see later). To test this, weekly predation levels were regressed against differentcomponents of the predator complex. For the birch sites, predation levels on large wintermoth pupae, smaller winter moth pupae, large C. albicans pupae and smaller C. albi canspupae were regressed against abundances of 1) Pterostichus spp., 2) Pterostichus spp. andCarabus spp. together, 3) Pterostichus spp., Carabus spp. and S. marginatus and 4) allbeetles together. Of 32 regressions, only one was significant (at site BI all beetles againstpredation of small C. albicans pupae, R2 = 0.491, P = 0.035). At RNP abundances ofcarabids against predation of pupae showed positive relations, but were not significant. Forfly pupae abundances of beetles had no effect. This suggests that beetles may have beenimportant in the predation of winter moth pupae there, but that other factors which do notfollow the abundances of beetles were responsible for fly predation. A similar patternemerged when looking at BI, except for predation of small fly pupae. Although theseresults do not identify the predators, they do indicate that different predators are responsiblefor predation of winter moth and C. albicans pupae, or that there is a different functionalresponse to the different pupal types among predators.Similar regressions were carried out for BBII, regressing weekly C. albicans andpupal predation (excluding the first week) against abundances of 1) Pterostichus spp., 2)Pterostichus spp. and Carabus spp. together, 3) Pterostichus spp., Carabus spp. and S.marginatus , 4) all beetles together and 5) beetle larvae only. Of 10 regressions 2 weresignificant. Pupal predation against abundance of beetle larvae (R2= 0.596, P= 0.023) andfly predation against all beetles caught (R2= 0.550, P= 0.035). This again suggests that thepredators are different for flies and pupae in blueberry sites. Regressions for BBI were notcarried out because sufficient pupae were not available to keep N> 20 for the tethers.1073.3.7 Seasonal abundance of predatory beetlesTrapping was carried out from mid May in 1992 to April of 1993 to estimate theseasonal abundances of important predatory beetles. Figures 3.11 to 3.13 indicate that theseasonal patterns were similar for most of the species. Most of the activity was in the earlymonths of trapping, until mid August or early September, with a decline through winter. AtBBI there were no adult predatory beetles active after September. Declines in the activitiesand abundances of beetles corresponded well with observations from 1991. Similarities inthe occurrence of Amara spp. and H. affinus at BBI during both years suggests that thesespecies may be linked by similar biologies (similar trends have been noted by Holliday andHagley 1978). Beetles at BBII had very similar patterns of seasonal occurrence to those atBBI, except that small numbers of P. melanarius were active through the winter months atBBII. However, there was a noticeable drop in the numbers of P. melanarius caught inlater months. At birch sites the important predatory beetles were Pterostichus species.These were found to be generally abundant throughout the year though the numbers wentinto decline from December to February when there was snow on the ground. At all fourbirch sites there was a general decline in the numbers of Pterostichus in early August. Asimilar decline in 1991 was not observed, but this may be because traps were collected onlyevery two weeks in 1991. The bimodal nature of the abundance curves of Pterostichus spp.suggests that there may be two generations in a year, those adults active early in the yearrepresenting those that emerged from overwintering larvae, and the later ones from larvaeof the previous spring.3.3.8 Annual abundance of predatory beetlesChanges in the abundances of predatory beetles between 1991 and 1992 wereexamined using data collected from the final week of June till the first week of Septembereach year. Only those beetles that were deemed possible predators (see previous sections108A. littoral!. app. larvae0.■AUGUST 22 NOVEMBER 5TIME (DAYS)AUGUST 22 NOVEMBER 5TIME (DAYS) 22 NOVEMBER 5 AUGUST 22 NOVEMBER 5Figure 3.11. Seasonal abundance of Four predators at BBL a) A. littoralis, b) A.aurata, c) H. affinus and d) beetle larvae from pitfall traps. Solid circlesindicate abundances in 1991, open circles indicate abundances in 1992 and barsindicate standard errors.109Figure 3.12. Seasonal abundances of four predators at BBII, a) C. granulatus, b)P.melanarius, c) C. nemoralis and d) Carabus spp. beetle larvae from pitfalltraps. Solid circles indicate abundances in 1991, open circles indicateabundances in 1992 and bars indicate standard errors.110 0 22 NOVEMBER 5AUGUST 22 NOVEMBER 5C0.80.40.0-AUGUST 22 NOVEMBER 50.6-0.4'0.0AUGUST 22 NOVEMBER 5TIME (DAYS)D0.8:0.4-0.0AUGUST 22 NOVEMBER 5.F0.6AUGUST 22 NOVEMBER 5TIME (DAYS)0.4-0.0-Figure 3.13. Seasonal abundances of Pterostichus spp. at a) BI, b) BIa, c) RNP,and d) at RNPI, with abundances of C. granulatus at e) BI and f) BIa. Solidcircles indicate abundances in 1991, open circles indicate abundances in 1992.Bars indicate standard errors.111and later discussion) were examined. These included all Pterostichus spp., Amara spp.,Harpalus spp., all beetle larvae, Carabus spp., all medium to large sized staphylinids,Agonum spp. and Bembidion spp.. These latter three groups were represented by onlyvery few individuals during this trapping period.There was a decline in the total abundance of predatory beetles only at one of the foursites for which two years of data are available (Fig 3.14). However, the changes inabundance at all sites, were not significant (2-way ANOVA on log+1 transformed data,years P= 0.36). The abundances between sites were significantly different (sites P<0.001). BI had significantly more beetles than all the other sites (P< 0.001, Tukey test),BBI had less than BI but more than either BBII or RNP (P< 0.001, Tukey test). Therewere no differences between RNP and BBII (P= 0.594, Tukey test). The apparentincreases of predators in 1992 can be attributed mainly to predatory staphylinids and smallbeetles. These may have had lower numbers in 1991 due to wetter weather conditions, ordue to differences in the trap collection periods, i.e. they may have been eaten by largerbeetles in the traps 1991.When looking at the individual species at each site, there are no apparent trends (Fig.3.15). Pterostichus spp. decreased at birch sites, but increased at blueberry sites. Therewere no significant differences between the abundances of these species in the two years(2-way ANOVA on log+1 transformed data, P = 0.496 for years) Carabus spp. increasedat all sites where they occurred, medium staphylinids increased and large Carabus larvaeincreased at all sites except BBI where they decreased. Catches of the smaller carabids andother beetle larvae were too small to give a good indication of changes in their abundances.Those species which occurred predominantly at BBI, underwent significant declines inabundance (T-test on log+1 transformed data, A. aurata, P = 0.023, T-test, A. littoralis, P= 0.016 and H. affinus, P=0.019) (see Fig. 3.16).112Figure 3.14. Summer abundance of predatory beetles at four filed sites inRichmond. Solid bars are abundances of beetles per trap in 1991, shaded barsare abundances of beetles per trap in 1992. Standard errors are presented.113BI RNP BBI^BBIIAPterostichus app.BI RNP BB1^BBIIDSte phyli n IdsECarabus larvaeRNPC.BBI^BBIICgranulatus81^RNP^BBI^BBIIF^Beetle^larvaeBIBI^RNP^BBI BBII^BI^RNP BBI BBIISITE SITEFigure 3.15. Changes in the abundances of a) Pterostichus spp., b) C.nemoralis, c) C. granulatus, d) staphylinids, e) Cara bus spp. larvae and f)beetle larvae excluding Carabus spp.. Solid bars are 1991 abundances, shadedbars are 1992 abundances. Standard errors are presented.114Figure 3.16. Changes in the abundance of three predators at BBI in Richmond,between 1991 and 1992.115It is apparent therefore that although the total abundances of the beetles changed little,the proportions of individual species changed at each site (see Table 3.7). Some speciesdeclined while other species increased in importance. At the birch sites, the most obviouschanges were with Pterostichus spp.. In 1991, these made up 100% of predatory beetles atRNP and 95.8% at BI. Although high in 1992, the proportions had dropped considerably.At the blueberry sites P. melanarius increased in importance. The Carabus spp. alsodemonstrated significant shifts in the proportions, C. nemoralis particularly underwentdramatic shifts with a decrease at BBI and a large increase at BI. In general however, theshifts in importance were small at blueberry sites. A. aurata, A. littoralis, H. affinus and P.melanarius the more important predators changed little.3.3.9 Distribution of sizes in the winter moth-Cyzenis pupal complexIn a previous section I indicated that the pupal type had an effect on predation.Cyzenis albicans pupae tended to be more highly predated than healthy winter moth pupae.Among C. albi cans pupae the smaller individuals were most highly predated. Chapter 2 hasindicated that winter moth pupal weights change between years, are different at differentsites and are dependent on the host plant. Cyzenis albi cans pupae are always significantlysmaller then their healthy host pupae (see Fig 3.17). Furthermore, the sizes of C. albicanspupae tend to follow the sizes of the host pupae at the particular site (see Table 3.8). This isnot surprising given the parasitoid's life cycle. A two-way analysis of variance of C.albi cans pupal weights indicated that they are dependent on site, but not on year (site; P =0.028 and year; P = 0.259 on log+1 transformed data), in a similar manner to healthywinter moth pupae.Cyzenis albi cans pupae appear to have greater constraints on their size, (the greatestrange I have recorded is a maximum of 0.027g and a minimum of 0.004g). This results ina considerable proportion of the available pupae in the soil being of a small size (i.e. below1 16Table 3.7. Changes in the proportions of different beetle species in 1991 and 1992, at four field sites in Richmond.Total numbers indicate the numbers of beetles caught in pitfall traps between late June and early September ofthose years. PT = Pterostichus spp., CG = C. granulatus, CN = C. nemoralis, MS = medium staphylinids, AA = A.aurata, AL = A. littoralis, BLC = Carabus spp. larvae, BL = beetle larvae, HA = H. affinus, CF = Calathus fuscipesand HR = H. rufipes.Site^Year Total^no. PT CG CN MS AA AL BLC BL HA CF HRBI^9 1 420 95.80 3.18 1.029 2 264 65.77 0.42 19.51 10.29 0.81 1.56 0.42 1.22RNP 91 43 10092 47 78.72 17.02 4.26BBI^9 1 157 1.24 0.66 28.71 35.02 10.81 22.90 0.669 2 93 5.38 8.60 4.30 1.08 29.0 20.4 8.60 3.24 19.40BBII 9 1 55 24.28 12.98 28.02 28.12 5.47 1.339 2 66 38.98 22.99 5.36 31.05 1.620. I-7100 1- 02^3^4^5^6^7^8^9^10^11^12SITES/YEARSFigure 3.17. Distribution of pupal weights in the winter moth-Cyzenis pupalcomplex at Richmond. Key to plot: 1 = BI 1991 winter moth, 2 = BI 1991Cyzenis, 3 = RNP 1991 winter moth, 4 = RNP 1991 Cyzenis, 5 = BBI 1991winter moth, 6 = BBI 1991 Cyzenis, 7 = BBII 1991 winter moth, 8 = BBII 1991Cyzenis, 9 = BI 1992 winter moth, 10 = BI 1992 Cyzenis, 11 = RNP 1992 wintermoth, 12 = RNP 1992 Cyzenis. Asterix indicate outlying points, circles indicatefar outlying points.118Table 3.8. Weights (± standard errors) of winter moth pupae at two birch sitesin 1991 and 1992 and at two blueberry sites in 1991, with correspondingCyzenis pupal weights at each site.Site Year winter^moth(g) Cyzenis(g)BI 1991 25 + 0.8 x 10-3 13 + 0.2 x 10-31992 24 + 1.2 x 10-3 12 + 0.4 x 10-3RNP 1991 28 + 0.7 x 10-3 14 + 0.3 x i0-1992 29 + 1.8 x 10-3 13 ±0.6 x 10-3BBI 1991 22 + 2.2 x 10-3 13 + 0.5 x 10-3BBII 1991 26 + 1.4 x 10-3 14 + 0.8 x 10-31190.02g) (see Fig. 3.18). Without parasitoids, less than 2-6% of the pupae would be below0.01g, when parasitoids are present this can increase to over 16% (see Table 3.9).Furthermore, about 20% of the pupae would be below 0.02g, while with parasitoids thiscould reach 40-50%, depending on the levels of parasitism, and the winter moth host plantconditions. This alteration in the distribution of pupal sizes could have a considerableinfluence on the dynamics of the system, and on the interactions of the predators with thewinter moth.3.4 DiscussionGeneralist PredatorsWinter moth populations in Richmond are heavily influenced by pupal mortality. InChapter 2, soil mortality, estimated as the difference between the number of prepupaedropping to the ground in summer and the number of the adults emerging from the soil inwinter, corresponded well with estimates of pupal predation from outplanted pupae.Studies in Britain and Nova Scotia, indicated that the main predators of winter moth pupaeare carabid and staphylinid beetles with varying importance being attributed to smallmammals (Buckner 1969, Frank 1967a, 1967b, East 1974, Pearsall 1992). In this study, Ihave observed a number of new species as capable of predating winter moth pupae, theseincluded C. granulatus and H. affinus which have been introduced from Europe and A.aurata, A. littoralis, P. algidus and P. herculaneus which are native to western NorthAmerica.Amara spp. and H. affinus were abundant at BBI, the site which is mostrepresentative of commercial blueberry habitat. In arena studies with winter moth pupae,H. affinus was an especially voracious predator and Amara spp., although less successful,did attack the pupae. Therefore, these species may play an important role in the predation of12012BBII 1991 (N = 72)8 -11100.00, 11/1111110114 -BI 1991 (N = 170)0.0300.004^0.01012 ^-9 llllll0.040^0.0500^0 0 0.040 0.05012RNP 1991 (N = 185)8-4 -^a00 604 4010111. 1001000:q9 1111)9^0.040^0.050PUPAL WEIGHT/gFigure 3.18. Effects of parasitism by Cyzenis on the overall sizes of pupae inthe winter moth-Cyzenis pupal complex at Richmond at a) BI, b) RNP and c)at BBII in 1991. Black bars indicate Cyzenis pupae, grey bars indicate wintermoth pupae.121Table 3.9. Proportions of pupae in different size categories from the wintermoth-Cyzenis pupal complex at Richmond, with suggested proportions for thesame populations in the absence of parasitism (i.e. taking only the proportions ofwinter moth pupae of each size category from each sample).Pupal Type Weight RNP1991 1992B!1991 1992BBII1991PARASITIZEDCyzenis (<^0.01) 6.7 7.3 16.1 16.2 3.4Cyzenis (0.01-0.02) 44.5 28.4 39.4 22.4 35.6Cyzenis (>0.02) 3.9 2.4 1.4 1.5 1.7winter moth (<0.01) 1.3 1.6 1.0 2.8 3.4winter moth (0.01-0.02) 6.7 12.7 8.7 11.2 8.5winter moth (0.02-0.03) 21.0 23.9 26.7 35.3 29.8winter moth (>0.03) 16.0 23.8 6.6 10.8 17.6HEALTHYwinter^moth (<0.01) 2.8 2.5 2.4 4.7 5.7winter moth (0.01-0.02) 14.9 20.5 20.2 18.6 14.3winter moth (0.02-0.03) 46.7 38.6 62.0 58.8 50.3winter moth (>0.03) 35.6 38.4 15.4 18.0 29.7Moth 103 39 83 40 37Fly 82 47 87 110 35winter moth pupae. This is surprising since Amara spp. and Harpalus spp. are usuallyconsidered as phytophagous (Lindroth 1968, Johnson and Cameron 1969, Thiele 1977,Holliday and Hagley 1978, Hengeveld 1980c). Hengeveld (1980b) has recently disputedthis. He suggests that the idea is based on few studies, with small numbers of specimenscollected by dubious methods. On examining carabid gut contents, he found no cleardistinction between the diets of Amara or Harpalus and those of carnivorous species.These species may undergo seasonal diet shifts. Comic (1973) found that in July, the timeof maximum oviposition, 20 - 40% of H. affinu.s specimens had insect remains in theirguts, much of these remains were of lepidopteran and dipteran larvae. This correspondswith the time when winter moth pupae are in the ground and receiving most predation.Pterostichus species are perhaps most likely to be efficient predators. A number ofthese have already been implicated in the predation of winter moth in Britain and Canada.Pterostichus species, particularly P. herculaneus, were very abundant at all birch sites, andP. melanarius, an introduced species, was abundant at blueberry sites. Frank's (1967a)serological investigations in Britain indicated that P. melanarius consumes winter mothpupae in the field.Cara bus granulatus and C. nemoralis were common at all sites. It is difficult tosuggest what interactions these have with winter moth pupae. Carabus spp. are perhapsmore likely to feed on soft bodied animals such as slugs, worms and caterpillars, whichthey digest preorally (Thiele 1977, Hengeveld 1980b). Thiele (1977) indicates that Carabusspp. consume most insects they come across, both adult and larvae. But subterraneanpupae may have substantially less predation. This is suggested from Lareau's (1987)observations that Carabus spp. are not found deep in the surface litter. Whether Carabusspecies can handle winter moth pupae in the field is difficult to determine. They were notvery successful in arena studies when compared to either H. affinus or Pterostichus spp.,although Roland (1986a) found no difference between pupal predation by C. nemoralis and123P. melanarius. They are much more likely to be important predators of prepupae or fallencaterpillars. The extent of prepupal predation at Richmond is unknown, but is expected tobe low since pupal predation accounts for most of the soil mortality (see Table 2.3).Hengeveld (1980 a, 1980b) analyzed the guts of 6337 carabids of 24 species andfound that all 24 species were polyphagous to some extent. He grouped his species intothose approaching either generalist or specialist natures. In general, the Harpalinae were thegeneralists and the Carabinae were specialists. The Harpalinae includes species of thegenera, Harpalus, Amara, Pterostichus and Bembidion. The specialist Carabinae includeLeistus, Notiophilus and Loricera. The Carabini (Carabus, Calosoma and Cychrus) couldbe distinguished within the Carabinae, as these specialize on larger prey such as snails,worms and caterpillars. Of the species I have implicated as predators of winter moth inRichmond, only C. granulatus and C. nemoralis are specialists. The others all belong to theHarpalinae. Evans (1977) found that the size of the metatrocanters, the length of the legsand the running speeds were all greater in the Carabinae than in the Haipalinae, the latterhaving a greater ability to push under wedges. Furthermore, the Harpalinae have poorlydeveloped sight and seem to recognize their prey more by olfactory means. Thesecharacteristics suggest that the Harpalinae are more capable of locating and attackingsubsoil pupae. Frank (1967a) however, found no differences between catches of A.parallelopedus and P. madidus in either pitfalls baited with winter moth pupae or control(unbaited) traps. He suggests that if the beetles are capable of detecting the scent of pupae itis not over any long distance. Comparisons of baited and unbaited pitfalls in this studysuggest similar conclusions.The amounts of food that carabids eat is an important consideration in their success atcontrolling winter moth. Generally only anecdotal evidence exists on the amounts of foodcarabids consume (see Hengeveld 1980c). A number of species have been shown to takemultiples of their body weights in food daily (Scherney 1959). Philonthus decorus larvae124were found to require only 6 winter moth pupae in order to successfully mature to adults(Kowalski 1976). Varley (1970) estimated that the mean annual consumption rates ofPhilonthus, Abax and Pterostichus per m2 at Whytham Wood totaled 92Kcal, while thetotal winter moth production was only 8Kcal. Therefore, only when winter moth were verynumerous would it provide more than a small portion of food for predators. The ecologicaleffects of carabids as a group of predators is difficult to assess due to a lack of informationon the amount of each prey category consumed within a defined period of time, and on thetemporal and spatial variation of this pattern of prey consumption (see Hengeveld 1980c).I suggest that beetle larvae have an important role in the predation of winter mothpupae. Beetle larvae as predators have not been addressed in great detail in previous studiesof winter moth predation, and larval feeding in general has received little attention in theliterature, probably because larval instars are difficult to identify and are subterranean(Hengeveld 1980b). There have been no successful analyses of the gut contents of carabidlarvae. Carnivory is however, suggested to predominate. Larvae of P. madidus, P. nigrataand Nebria complata require live prey to develop. Similarly animal material may benecessary for the survival of Amara spp. larvae (Thiele 1977). Larvae of H. affinus havebeen shown to be predaceous as well as phytophagous (Brigg 1965). Beetle larvae mayhave more specialized diets than adults, Abax larvae for example, are specialized onworms. However, extreme specialization like this is probably an exception (Thiele 1977).The high occurrence of beetle larvae in baited subsoil traps in this study is notable.Larvae of a number of different species were captured, these included a number of carabidsand staphylinids. Pterostichus spp. larvae were most common and have been observedfeeding on winter moth pupae in beetle exclosure cages. All species that were caught insubsoil traps were brevimandibular suggesting a polyphagous habit (Thiele 1977). Becauseof the subterranean nature of both the winter moth and the beetle larvae, and because of thehigher numbers in baited subsoil traps, olfactory stimuli are most likely used by larvae in125locating prey. It is likely that these larvae are causing most of the mortality on the pupae.Larvae of Carabus spp. were frequently caught in pitfall traps. These were surface activeand feed on pupae in the arena studies. Because of this epigeal habit, they are unlikely toattack subsoil pupae. Larvae of 14 different species were captured in pitfall traps, themajority of these were caught at site BBII. Of these 14 different species, 5 werelongimadibular. This morphological trait suggests that they are predators of snails (Greene1975). Carabus spp. larvae are brevimandibular.Predation of pupae in lmm mesh exclusion cages varied from 15 - 40% (dependingon the site in question). This is low, but the mesh used in the subsoil exclosures probablyexcluded some of the larger subsoil beetle larvae. Roland's (1986a) use of 2mm meshwould have allowed most larvae to enter his beetle exclusion cages, and there wassignificant mortality on the pupae. Larger larvae may be more important in the totalpredation. Kowalski (1976) noted that larger larvae of P. decorus were more successful inattacking winter moth pupae than were smaller earlier instars. Roland's (1986a) beetleremoval plots also implicate beetle larvae as important predators. In his plots he reduced thenumbers of carabids by 10 - 100 times, staphylinids by 10 - 30 times, but beetle larvae byonly 3 times. There was no significant difference in predation of pupae in either theremoval or control plots. This could suggest that beetle larvae were the most importantcomponent of the predator complex to begin with. It is very difficult to estimate densities ofsubsoil larvae and their responses to winter moth densities have never been addressed. Onemight expect that when winter moth densities are high there is increased survival of thelarvae and thus increased numbers of adults in the following season. However, manyspecies undergo dispersal flights as immature adults (i.e. H. affinus see Holliday andHagley 1978) and dispersal in general by carabicis is efficient (Rivard 1965).Ants were abundant at the bog sites studied in Richmond. Both these sites had veryhigh levels of predation (predation was 100% at both sites in 1990). Ants were also a126noticeable feature of the insect community at site BBI. Since ants are not successfullysampled by pitfall trapping (Williams 1958), it is difficult to compare densities orabundances of ants between sites or years. In 1991 and 1992, I made a survey of antcolonies. At the birch sites only three Formica sp. colonies were located (1 at RNP and 2 atBI). At BBII only one Formica sp. colony was located. BBI had two colonies of Formicasp. and numerous colonies of Myrmica sp.. Because of the dry ground at this site and theemergence of ants from cracks in the soil, it was difficult to distinguish the boundaries ofMyrmica sp. colonies. In 1990,20 wooden boards were left on the ground at site BBI overwinter. In 1991, colonies had established beneath 18 of them. The occurrence of these antsin subsoil traps is not surprising then since they were generally abundant. Myrmica sp.were found feeding on winter moth pupae, but with some difficulty. The larger Formicasp. are probably more successful, but were not observed. The foraging trails of eachFormica sp. colony were observed in 1991. There was no apparent confinement of thetrails to areas of higher winter moth densities. In the incidence where trails left a highwinter moth density area to a lower one, i.e. from BBI to an insecticide sprayedcommercial blueberry field, the foraging trails were noticeably longer, but more study isrequired. Formica spp. are probably more important as predators of prepupae and larvae.The impact of ants on pupae has received little attention in the literature (see Campbell andTorgensen 1964, Torgensen et al. 1983, Kelly and Regniere 1985, Eskafi and Kolbe1990). However, ant predation on larvae and their use in biological control have frequentlybeen addressed (Ives 1984, Harris 1984, Finnegan and Smirnoff 1984 and see referencesin Ito and Higashi 1991).A large number of small mammals mainly shrews and voles were caught in pitfalltraps, at the Richmond sites. Analysis of their gut contents was not carried out since theywere generally partially decayed by the time the traps were emptied. Shrews are known tobe predators of a number of subsoil pupae (Buckner 1966, McLeod 1966, Hanski andParvianen 1985, Kelly and Regniere 1985) including pupae of the winter moth (Frank1271967a, East 1974, Buckner 1969). Buckner (1969) suggested that shrews are responsiblefor most of the pupal predation of winter moth at Whytham Wood, but East (1974) hasrejected this idea. Few of the outplanted pupae in this study showed signs of predation bysmall mammals. I suggest that small mammals were responsible for a very small portion ofthe predation, particularly at blueberry sites.Small mammals are perhaps more important in their affects on the beetle populations(Parmenter and MacMahon 1988, Griim 1979) and are expected to impact a greaterinfluence in winter, since there are fewer alternative food items for them at that time(Buckner 1969, Frank 1967b, Kowalski 1976). This influence of winter active smallmammals on populations of beetles has been implicated in the lack of a delayed densitydependent numerical response of generalist invertebrate predators to winter moth densities.There is a strong preference for heterogeneous habitats among small mammals (Parmenterand MacMahon 1983). This was observed at Richmond also, with six times as many smallmammals caught at birch sites. The effects are therefore expected to be greater at birch sitessince these had the highest numbers of small mammals. Unlike the situation in Britain, highdensities of C. albicans pupae in the soil at the birch sites and at Roland's oak site mighthave been expected to reduce predation pressure on invertebrate predators and allow adelayed numerical response, but this has not been observed. Furthermore, beetlepopulations in orchards at Vancouver Island (Roland 1986a) and at blueberry sites in thisstudy have not presented any evidence to suggest delayed numerical responses in spite offew small mammals.Cyzenis-Predator interactionsThere have been suggestions for a number of different response types of generalistpredators to winter moth pupal densities. In Nova Scotian apple orchards, Pearsall (1993 inPress) has demonstrated an aggregated functional response of predators to winter moth128pupal densities. This has also been suggested by East (1974) for P. madidus. Roland(1986a) found pupal predation to be spatially density dependent, but discounted aggregatedor numerical responses among carabids or staphylinids to changing winter moth pupaldensities. Aggregated responses are difficult to interpret. Hagley and Allan (1988) found asignificant relationship between numbers of P. melanarius and of fifth instar larvae of C.pomonella at one of three orchards which they studied in Ontario. However, at the samesite there was no significant relationship between the numbers of serologically positive P.melanarius and larval densities. The fact that predators, particularly generalist predators,aggregate under trees with higher densities of pupae or larvae might suggest other things.In 1991, I noticed that there were more collembola and soil mites in subsoil traps that hadbeen planted near areas of heavy defoliation at site BI. These may have been feeding fromcaterpillar frass or leaf material under the trees at that time, and are a significant source offood for generalist predators (see Thiele 1977 and Hengeveld 1980a).Temporal numerical responses are subject to similar arguments. Numerical responsesby generalist predators (carabids and staphylinids) to lepidopteran pupae have been noted ina number of studies (Bauer 1983 and see Hassell and May 1986) For winter moth, bothdelayed numerical responses (of P. decorus; Frank 1967b, East 1974) and current yearresponses of predators (P. decorus and P. madidus; Kowalski 1976, 1977) have beenproposed. Buckner (1966) suggests that if small mammals are important, only abehavioural numerical response is likely to be observed. For a delay to arise in thenumerical response, the reproductive output or survival of predators must be affected.Abundance of food can affect reproductive processes. Murdoch (1966) for example, foundAgonum fuliginosum to have a delayed reproduction when food was scarce. However,cases like this are rarely noted and a review by Hengeveld (1980a) indicates that the effectsof food availability on the sizes and stability of carabid populations needs considerablymore attention.129In this study predation of pupae appeared to be temporally density dependent at allsites except BBI, where deviations from the trend occurred only in 1992. There was nodelay in the predation at any site. Interestingly, levels of predation were very similar inspite of very different abundances and assemblages of predatory beetles. Only two years ofdata are available on the abundances of predators at the sites. At three of the four sites,there were increases in the total numbers of predaceous beetles in 1992, when winter mothpupal densities were lower. Looking only at the most important predators however, i.e.Amara spp., H. affinus and Pterostichus spp. there have been general declines inabundance, (P. melanarius is the only exception). If these declines are the result of declinesin winter moth pupae, then it may be affected through the larvae. Little is known about theseasonal occurrence of larvae of the native Amara spp.. Indications are that they hibernateas adults, with the larvae present from early spring until August: similarly H. affinus larvaeare present from June till August (Lindroth 1961 - 1969, Holliday and Hagley 1978). Allthe Pterostichus spp. present at Richmond overwinter as larvae and emerge as adults in theearly spring (although the occurrence of two maxima separated by a minimum abundanceof beetles in late July and early August, suggest that there may be larvae present until Julyor August). Larvae of each species could therefore have increased survival either due to thepresence of winter moth or C. albi cans pupae (Amara and Harpalus) or due to C. albicanspupae alone (Pterostichus). Most of the predation of winter moth at Richmond occurredbefore August. Similar observations have been made in Nova Scotia (Pearsall 1992) and inBritain (East 1974). This was when beetle activity (both adult and larval) was at a peak.Roland's (1986a, 1988) suggestion that the increase in the period of availability of pupae inthe soil due to the presence of C. albicans, leads to a numerical response (his firstmechanism), is therefore only possible for Pterostichus spp. and only at birch sites.In this study tethered C. albi cans pupae were more highly predated than tetheredwinter moth pupae. This occurred at both blueberry and birch sites, although at BBI thedifferences were negligible. Interestingly the appearance of predation differed at each site.130At RNP, both large and small sized C. albi cans were more heavily predated. At BI onlysmall C. albicans were significantly more heavily predated than winter moth. Similarly, atBBII C. albicans pupae were significantly more predated, but at BBI there was nodifference in the predation levels on C. albicans or moth pupae. This suggests that thepredators at the sites, even within the same habitat types (i.e. blueberry or birch), aredifferent. In no case did C. albicans have lower predation. This differs from Roland's(1986a, 1988) observations on Vancouver Island, and discounts the possibility of hissecond mechanism (greater parasitism of unparasitized pupae) operating at Richmond.Differences in the predation rates of different sized C. albicans at BI is notable, andsuggests that some pupae may be more vulnerable to predation due to a smaller size. In anunparasitised population of winter moth only about 20% of the pupae are under 0.02g, atlow population densities this is expected to be even less (see Table 3.9). At Richmond,Vancouver Island and Nova Scotia where levels of parasitism were high, the percentage ofavailable pupae that are below 0.02g rises to about 60%. Since the smaller pupae weremost highly predated, I suggests that some predators may be capable of handling only theseand not the larger pupae. These predators may include the early instars of beetle larvae. Aconsiderably higher proportion of smaller more manageable pupae, as in situations withhigh levels of parasitism, may therefore, allow increased survival of early instar beetlelarvae, or of the larvae of smaller beetles i.e. Agonum or Bembidion spp.. Unfortunately,very little is known about the feeding behaviours of beetle larvae, and I have notinvestigated the toughness of pupal integuments. The pupal integuments of small C.albi cans pupae (< 0.01g) did appear to break more easily than those of similar sized wintermoth. Opening the cocoons to examine the condition of the pupae inside may haveintroduced a bias to the study. The cocoon has a protective role and many predators whichfeed on pupae in this study may not normally be able to penetrate the cocoon. Furthermore,C. albicans pupae are enclosed in the integument of the host winter moth pupae. Onexamining the pupae this integument was normally removed. Although the host pupal131integument is very fragile, there again is a suggestion that it may have been easier for thepredators to attack the C. albicans pupae in this study.Roland (1986a) tended to discount larvae as having a role in the predation of wintermoth pupae on Vancouver Island, since pitfall trapping had revealed that larvae wereabundant only when both the winter moth and C. albi cans pupae were absent from the soil.Surface active beetle larvae also diminished in pitfall traps at the time of winter mothpupation at Richmond. However, the traps did not account for subsoil larvae which werean active component of predation as revealed by subsoil traps at that time. Therefore,increased availability of food for larvae may give rise to increased survival and thusincreased numbers of adults in the subsequent generation. However, this would notnecessarily lead to increased adult abundances or the manifestation of a delayed numericalresponse if the adults are actively dispersing.Roland's (1986a) findings that predation of the C. albi cans pupae was reduced in theabsence of beetles and regressions of weekly pupal predation rates against the differentcomponents of the predator community from this study, suggest that an alternate interactioncould occur. It is possible that the adult beetles may be favoured by the presence of C.albicans pupae. In this study, only three of the regressions were positive, C. albicanspredation against all beetles at BBII and at BI, and winter moth pupal predation againstbeetle larvae at BM. In general however, regressions of beetle abundance showed positiverelations against predation, but there were no trends for C. albicans predation, and alsobeetles in arenas appeared to favour winter moth pupae. It is not clear which component ofthe predator population (i.e. adults or larvae) benefits most from the presence of C.albicans pupae. What these data do indicate, however, is that different components of thepredator complex are responsible for predation on different components of the pupalcomplex. A third mechanism therefore by which C. albicans pupae may have increasedpredation could be by making food available for both adult and larval predators, and thus132increasing total survival or reducing dispersal from the area. This mechanism is somewhatsimilar to Roland's (1986a, 1988) first mechanism except that the effects of C. albicans inthe pupal complex act at the same time as the increased availability of food due to wintermoth pupae. The availability of pupae over winter may or may not be necessary to bringabout an effect.The occurrence of C. albicans pupae in the soil may not have been the only factorbringing about increased predation. On both Vancouver Island and in Nova Scotia, pupalpredation began to rise before levels of parasitism were very high. However, rises inparasitism and predation, were largely synchronous at both locations. It is possible that theconstant proportions of favorable pupae allowed populations to build up over time. Butdeclines in the winter moth populations occurred in very different habitats with differentlevels of parasitism at the same time (Roland 1988, Embree 1991). This would tend tosuggest that something in addition to predators or parasites brings about the declines. Atboth Vancouver Island and Nova Scotia in the years previous to the increases in bothparasitism and predation, there were high levels of larval mortality (that is 1959 at NovaScotia and 1983 at Vancouver Island) presumably, to a large extent due to starvation. Sincehigh larval mortality has been associated with increased mortality of pupae (death due tounknown causes), and predation is higher on dead pupae, this high larval mortality mayhave also initiated predation increases. Dead pupae could therefore, either have supportedadult beetles and reduced their dispersal, or increased survival of beetle larvae, building upa numerical response. Cyzenis albicans may have only played a role in maintaining predatorpopulations after larval mortality decreased or may have accentuated the effects of highlevels of dead pupae in the population. High larval mortality in 1957, may not have had aneffect because C. albicans was still at low levels. This fourth mechanism thereforeincorporates larval mortality into the system, with C. albicans only playing a role inaccentuating the effects.133As was mentioned in Chapter 2, the Richmond situation is very different from thoseof Vancouver Island and Nova Scotia. The parasitoids probably arrived in Richmond at thesame time as the winter moth. A post-introductory outbreak occurred about 1986 to 1988,and parasitoids soon went into increase. Levels of parasitism were already high in 1989,but levels of parasitism have never reached the levels observed on oak in Nova Scotia or onVancouver Island. The decline began in 1990. There was no lag of four years between C.albicans occurrence and winter moth decline, as had occurred in the other two situations.The drastic population decline in 1992 was largely attributed to larval mortality (numbers offemales in 1990 and 1991 have been estimated to be about the same). I found predation tobe temporally density dependent, and found weak evidence for numerical responses ofbeetle predators, with a number of beetle species declining in 1992 under lower wintermoth pupal densities. There was no inverse density dependent increase in predation asoccurred on both Vancouver Island and Nova Scotia. I suggest that these differences can beexplained by looking only at the response of predators to parasitoids. In Nova Scotia,Vancouver Island and Richmond this was density dependent, and may be due to those lifestages that utilize the C. albicans pupae surviving better or dispersing less (depending onthe stage in question) and allowing that stage which predates winter moth to build up.There are still a number of problems with this idea because it is not known what is causingthe mortality. Arena studies are not helpful due to altered behaviours of insects in the arenasand cages (see Hand and Keaster 1967). Serological studies are the only way to havereasonable certainty of what eats pupae in the field. Serological studies are becomingfrequently more common in such studies. These too have their problems and knowledge ofthe background phenomenon which influence the detection of prey items is far fromsatisfactory (Liivei et al. 1985). The use of pitfall traps in this and other studies has givendubious results and does not give any information about larval stages (Greenslade 1964,Holliday and Hagley 1978). It is unlikely that a single species of predator would beresponsible for the mortality and it is not important which predators are involved.134The mechanisms by which C. albicans may have caused increased predation ofwinter moth pupae appear to be more complicated than Roland (1986a, 1988) hadoriginally proposed. Roland's first mechanism, in which an increased period of availabilityof pupae in the soil, due to the presence of C. albicans, allows a numerical response, nowappears unlikely. There was a weak indication of a numerical response among somepredators in Richmond, but in general predators were most abundant in the year of decline.Predator activity was also severely reduced for most of the winter months. This suggeststhat the availability of C. albicans pupae overwinter is probably not important in allowing anumerical build-up of predators. If C. albicans were important in this respect, then it wouldbe most likely to affect the Pterostichus spp., since these were active throughout the year.At the birch sites there was an indication of a numerical response among Pterostichus.Roland's (1986a, 1988) second mechanism suggests that winter moth pupae arepreferentially predated from the pupal complex. This causes increased predation pressureon the remaining winter moth pupae and allows better survival of the fly parasitoid. Thismechanism can be discounted in Richmond since examination of predation on tetheredpupae indicated that it is the C. albicans pupae that are more heavily predated. I haveproposed two further mechanisms. The first suggests that predation on the winter mothpupae is by a different life stage(s) of the predators than that predating the C. albicanspupae. Therefore, a numerical response may be allowed since food is available for all lifestages. A second proposed mechanism incorporates the occurrence of dead pupae in thepupal complex. Dead pupae are most highly predated so that large numbers of dead pupaein the soil allow a build-up of predators. The presence of C. albi cans in the pupal complexmaintains the predator population at a high level until outbreak decline. This lattermechanism might allow predator build-up to occur in areas with different levels ofparasitism of winter moth, but is still dependent on parasitism.135CHAPTER 4GENERAL DISCUSSIONIt has recently become clear that the successful biological control of winter moth inCanada by its parasitoid C. albicans has not been due simply to the parasitoid-hostinteractions, but that other factors including generalist predators have played an essentialrole (see Table 4.1). In North America, outbreaks of winter moth have predominantly beenof the post-introduction type, suggesting that natural enemies are essential in maintainingpopulation equilibrium. These natural enemies include viral diseases and parasitoids, butgeneralist predators are not capable of inhibiting the occurrence of outbreaks afterintroduction. Tenow (1972) indicated that a maritime type climate is important foroutbreaks of winter moth to occur. All the regions in North America where outbreaks haveoccurred have mild winters (see also Chapter 1 under 'winter moth'). Winter temperaturesthat fall below -330C inhibit outbreaks (Wylie 1960b). Tenow (1972) however, gives noindication of the causes of outbreaks.Outbreak declineWinter moth outbreak declines can be attributed to a variety of factors, but three mainfactors predominate (see Table 4.2).Parasitism: Varley et al. (1973) maintained that winter moth populations in England wereregulated by generalist predators of the pupae, and that winter mortality determined thelevels of the population each year. Parasitism in Britain was deemed unimportant in wintermoth population regulation. Unfortunately, the period through which they conducted life136Table 4.1. Synopsis of characteristics of winter moth outbreaks and the factorsattributed to their declines.137CHARACTERISTIC POPULATION SOURCEOutbreaks^with^parasitoids Europe Tenow 1972present Nova Scotia Embree^1991Richmond This studyInverse^density^dependent Nova Scotia Embree^1966increases^in predation Victoria Roland 1988Increases^in predation before Nova Scotia Embree 1966increases^in^parasitism Victoria Roland^1988Richmond This studyReduced fecundity after heavy Hokkaido Kukuzawa et a/. 1979defoliation in the previous year Victoria^(apple) Roland and Myers 1987Richmond This studyFoliage quality important Whytham wood Varley et al. 1973Hokkaido Kikuzawa et al. 1979Decline in parasitism after an Nova Scotia Embree 1966initial^high Victoria Roland 1988Richmond This studyDeclines with low levels ofparasitismOregon Kimberling et al. 1986High larval mortality before the Nova Scotia Embree 1965population^crash Victoria Roland 1988Richmond This studyPupal predation as the main Why tham wood Varley et al.^1973regulator Nova Scotia (apple) McPhee et al. 1988Pearsall^1992Victoria Roland 1988Richmond This studyParasitism as the main regulator Nova Scotia Embree^1966Simultaneous declines in sites Nova Scotia Embree^1966with different host plants Victoria Roland 1988Oregon Kimberling et al. 1986Richmond This study1 3 8Table 4.2. Information on outbreak of winter moth from seven studies, withinformation on the success of Cyzenis albicans in parasitizing thepopulations.139WintermothpopulationNative/^host^plantsintroducedduration^ofoutbreakCyzenisreleasedHighestlevels^ofparasitismreachedColapse^sourceattributed^toWhytham native Quercus robar no outbreak between^native 0-22% Pupal^predation Varley et al.Wood 1950 and 1970: 1973Minor outbreaksfrom 1957 to 1958and in1966Hokkaido native Alnus inokumae 1975-1976 Food shortage^Kikuzawa et al.1979Scotland native Picea sitchensis 1981-? native Stoakley^1985Nova Scotia introduced Quercus rubra & 1950's-1962 1954-1961 61% Parasitism^Embree^1966Malus spp.Victoria introduced Quercus garyana &Malnus spp.1970s-1982 1979-1982 84% Parasitism^and^Roland^1988pupal^predationOregon introduced Corylus avellana,Malus silvestris &?-1983 1982 <4% unknown^Kimberling etal.^1986Prunus cerasiferaRichmond introduced Betula papyfolia &Vacciniumcorymbosum1988-1992 not released 55% Pupal^predation This studyand^larvalmortalitytable analyses at Whytham Wood does not appear to have incorporated an outbreak' , infact the year in which they commenced the life table studies was a year of populationdecline from the highest levels recorded at Whytham Wood (see Appendix 7). Densities ofwinter moth larvae in 1949 (500 L/m2) were about five times higher than those of 1950(112L/m2). Minor outbreaks occurred in 1957 and 1964. In the early 1970's there was anoutbreak at Wistman's Wood, a site adjacent to Whytham Wood, in Oxford. Because levelsof parasitism were not recorded at Wistman's wood and because of the lack of a majoroutbreak at Whytham between 1950 and 1968, we do not know how Cyzenis responds tooutbrealdng densities of winter moth in Europe. Tenow (1972) noted that declines fromoutbreak in the Scandes are often associated with high levels of parasitism. Therefore, thebasic premise that parasitism is low in Britain can be disputed; the fact is we do not knowwhat parasitism is like during outbreaks there. Embree (1966) suggested that theintroduction of C. albicans and A. flaveolatum brought about a decline in Nova Scotia.Parasites however, do not appear to be of sole importance. This is apparent from declinesobserved in different regions with different levels of parasitism, including simultaneousdeclines in 1982-1983 in Victoria and Oregon with about 70% and 4% parasitismrespectfully (see Table 4.2). Furthermore, there have been simultaneous declines ondifferent host plants within the same regions, in spite of differences in the ovipositionresponses of C. albicans on these plants with resulting differential success of parasitism.1 What constitutes an outbreak? The arbitrary nature of this term may cause some problems. Abrief review of the densities reported as outbreaks indicate great variability among studies(see Table 4.3). Much of this variation may be due to the great variety of host plants on whichoutbreaks develop. The best measure of density is the number of larvae per bud, leaf cluster orshoot (leaf clusters and shoots having developed from a single bud). In these terms outbreakdensities are dependent on the host plant, for example Kikuzawa et al. (1979) present 0.3L/budas an outbreak density on alder, while on oak 5L/cluster is an outbreak (Embree 1966). Thenumbers of pupae/m2 is a more satisfactory measure for comparing across studies, since it has atwo dimensional nature, is easily recorded, and has significant biological meaning. Also Roland(1986a) has shown that larval density may vary temporally while pupal density remains morestable as a result of compensatory mortality.141Table 4.3. History of outbreaks of the winter moth from seven regions, withindications of the levels of parasitism (mainly due to Cyzenis) and soilmortality presented as percentages and k-values. Corresponding densities ofwinter moth are also indicated. 'H' indicates highest levels, 'L' indicateslowest levels and '0' indicates outbreak populations. An asterix indicates thatvalues for pupal predation are presented rather than soil mortality.142%^Parasitism K3 %^soil^mortality K3 Year Densities^L/Cluster L/1112 P/M2 Location Source33 0.176 1976 (0) 0.3L/bud 240 (1976) 9 Japan Kilcuzawa eta!.17^(1977) 1979 0 (L) 0 (L) 39.74 (L) 0.22 (L) 1950-1968 6.3 0.27 England Varley et al.22.38 (H) 0.11 (H) 86.51 (H) 0.87 (H) 1973 492 81.31.14-8.8 (0) 0.005-0.04 (0) 81.8-86.51^(0) 0.74-0.87 (0) 1964-1966(0)269 (H) 78.498.42^(1974) 1.80 (1974) 1973-1975 (0) 1465-1845 17 6-2 96 England Wigley^197699.81^(1975) 2.71^(1975)1988 4L/shoot Scotland Hunter eta!.1991 4 (H) 0.018 (H) 1980-1983^(P) 2.5L/cluster^1980 Oregon Kimberling etal.^198661.1^(oak)(H) 0.41 (H) 94 (H) 1.2 (H) 1954-1962^(P) 5L/cluster Nova Scotia Embree(reanalysed^byRoland^1988)84^1985(oak)(H) 0.796 (H) 96 (H) 1.40^1987(oak) 1983-1985^(P) 3.36L/cluster^(oak)19838.96 L/cluster (apple)^1983Victoria B.C. Roland 1992and^1986a65^1984(apple)(H) 0.46 (H) 85.87 (H) 0.85^1984 (apple)<16 (Agrypon) (H) 96.84 (H) 1.5 (H) 1967-1980 0.05-1.4L/cluster Nova Scotia McPhee et al.<10 (Cyzenis) 29.21 (L) 0.15 (L) 1988 0-18.6 0-0.089 98.4-74.5 1.795-0.59 1991 0.17-4L/cluster 7.4-347.7 Nova Scotia Pearsall^199295.8-47.8* 1.376-0.282*36.28^(blueberry) 0.196 96.25* 1.426* 1989-1992^(P) 0.36L/cluster^1991 72.41 Richmond B.C. This Study55.18^(birch) 0.349 98.75* 1.902*0.74L/cluster^1991 378.2Foliage quality: The importance of foliage quality in bringing about population declineshas frequently been omitted from studies of outbreaking winter moth populations. This isin spite of a realization of the importance of synchronies between bud-burst and larvaleclosion (Feeny 1970, Holliday 1977, Hunter 1990, Watt 1987). In this study, an earlyspring may have caused asynchronies between the winter moth egg hatch and a number ofits host plants. High incidences of "death due to unknown causes" as well as low femalefecundities suggest that there may have been asynchronies or some nutrient deficiencies thatreduced the vigour of caterpillars and moths. Hunter eta!. (1991) looking at an outbreakingpopulation on Sitka spruce in Scotland, found no evidence to suggest that nutrientdeficiencies of trees or the timing of bud-burst had any role in determining the course ofwinter moth outbreaks there. Kikuzawa et a/. (1979) on the other hand suggest that it wasreduced quantity of food, and thus a reduction in fecundity that was responsible for the1977 crash on alder in Japan. Simultaneous declines under different levels of parasitismwith different host plants occurred on both the east and west coasts of Canada (Roland1988, Embree 1991). This suggests that weather effects may be important. Weatherpatterns can have effects on the quality of host plants for the moth, and on synchronizationof bud burst and larval eclosion. These in turn might appear as increases in larval mortality,increases in soil mortality, and reduced fecundities. High levels of larval mortality in NovaScotia and Victoria before population declines, may indicate that climatic conditions hadaffected these populations in a similar manner to Richmond in 1992. Therefore, climaticconditions may have been very important in the success of the biological control program.Which particular aspect of weather has not been elucidated. Climatic data should beexamined from all three regions, both before and during the winter moth declines.However, in this study, while winter moth were declining in the Richmond area whereoutbreaks initially occurred in the mid-1980's, populations on the edge of the northernspread into Vancouver were still very high. This argues that the history of the populationhas a dominant effect and may over ride any weather related influences.144Pupal predation: Predation of winter moth pupae by generalist predators is now at theforefront of mechanisms implicated in the control of winter moth outbreaks, and themaintenance of winter moth populations at equilibrium densities (MacPhee et al. 1988,Pearsall 1992, Roland 1992). Data on soil mortality from a number of studies, are morecomplete than those of parasitism or other factors (see Table 4.3). Wigley's (1976) datafrom England, suggest that soil mortality was as high as 99.81% in 1975, a figure higherthan any recorded at Whytham Wood. Table 4.3 indicates that pupal predation is generallyhigh among all high density populations of winter moth studied. It appears therefore, thatsoil mortality can not be ruled out as important in the control of outbreaking populations.An exception is that of the 1976 outbreak in Japan, but at that site pupal densities were verylow, ca. 9/m2. In Canada, ground predators have been the main causal factor in bringingabout the initial declines of the winter moth populations, and have also played a vital role inmaintaining the populations at low densities after the initial suppression (Roland 1992).Cyzenis-generalist predator linkIt has been proposed that the presence of C. albi cans was essential in bringing aboutwinter moth regulation by predators (Roland 1986a, 1988). Therefore, a proper evaluationof this case of biological control requires an understanding of a three-way interaction,namely that of, winter moth, Cyzenis and the generalist predators. The winter moth-Cyzenis interaction has been the subject of a number of papers (Embree 1965, 1966,Embree and Sisojevic 1965, Hassell 1966, 1980). However, the winter moth-generalistpredator and especially the Cyzenis-generalist predator interactions have receivedconsiderably less attention. Roland (1988) has suggested a link between ground predatorsand Cyzenis parasitism such that the parasitoid may have been responsible for initiallyincreasing pupal predation. This was based on the observation that in both Nova Scotia and145on Vancouver Island there was a similar occurrence of events, in spite of theimplementation of biological control after very different periods of high winter mothdensity (see Chapter 1, Chapter 2 and Roland 1988). Roland proposed three mechanismsby which Cyzenis could have caused increases in predation. The first mechanism suggeststhat a longer availability of pupae in the soil due to the presence of Cyzenis pupae over thewinter and early spring, allows a numerical build-up of predators. A second mechanismsuggests that a preference for healthy pupae make winter moth pupae more susceptible topredation when parasites are present in the population. Finally, a third mechanism suggeststhat parasitoids act as vectors of disease and cause increased soil mortality by infectingprepupae. Results from this study indicate that all three of these mechanisms are unlikely tohave occurred in Richmond. Based on observations of differential predation on wintermoth and Cyzenis pupae, and the indication that different components (larvae or adults ofthe same species) of the predator complex differentially take moth and fly pupae, I havepresented a further mechanism. I suggest that the important predators of winter moth pupaerequire Cyzenis pupae for a population build-up, due to the inability of some stage of thelife cycle to consume healthy winter moth pupae. Given the complexity of predatorassemblages involved and the very different assemblages at different sites, it may be that allof these mechanisms have acted to increase predation, but the mechanisms may differbetween sites. At sites with low levels of parasitism, factors other than predation may haveplayed a more important role. At these sites, little is known of the impact of predation (seeTable 4.3).Very similar species of moth endemic to North America often occur at the same sitesas winter moth. In North America there are two native Opheroptera species. The mostimportant is 0. bruceata (Hulst) which has a northern distribution. The other 0.danbyi(Hulst) is endemic to the Pacific Northwest with a more limited distribution (Miller andCronhardt 1982, Troubridge and Fitzpatrick 1993). Two allopatric subspecies of 0.146bruceata (0. bruceata occidentalis [Hu1st] and 0. bruceata bruceata [HulstD can bedistinguished (Troubridge and Fitzpatrick 1993). American authors treat 0. bruceataoccidentalis as a separate species, 0. occidentalis (see Miller 1982, Miller and Cronhardt1982). Operophterabruceata and 0 .brumata appear very closely related, and are known tointerbreed in regions of overlapping distribution (Underhill et al. 1987, Hale 1989,Troubridge and Fitzpatrick 1993), although the extent of this in the wild is unknown (seeFitzpatrick et al. 1991). Compared to winter moth, 0 .bruceata emerges later in the fall, hasa one to two week earlier larval eclosion, and has only four instars (Brown 1962,Kimberling et al. 1986). Males are attracted to the same synthetic pheromone as wintermoth males (Underhill et al. 1987). They differ in the levels of parasitism, virus, etc.,observed among their populations (Pivnick et al. 1988), and they have different species ofparasites though they share a number of common genera (Wylie 1960a, Sechser 1970a,1970b). All the Operophtera species are highly polyphagous. Brown (1962) lists 17 plantsfrom which 0. bruceata feed, these included plants from the genera, Populus, Salix, Acer,Fagus, Malus, Prunus, Betula, Lornicera, Alnus, Amelanchier, Ribes, Rosa, andHolodiscus. Millar and Cronhardt (1982) list 19 plants on which 0. bruceata occidentalislarvae have been found, these included both coniferous and deciduous species such asCorylus, Crataegus, Fraxinus, Oemleria, Physocarpus, Pseudotsuga, Quercus, Rubus,Tsuga, and Symphoricarpos. A number of these plants are shared hosts with winter moth.Given the similarities of these species, they should be expected to display similardynamics. In the Scandes, Tenow (1972) showed similarities in the periodicities andoccurrences of outbreaks of winter moth, 0. fagata and a similar species, Oporinaautumnata Bkh., and indicated that mild winters greatly influence the dynamics of thesespecies, since outbreaks generally coincided with mild winters. Bruce's spanwormundergoes outbreak in Canada. These outbreaks last from 1-5 years, and have a periodicityof about 7 years. Decline has been attributed to viral disease (1964 and 1973 in Quebec)147and to high numbers of ground beetles (in 1976 in Cumberland Co. Nova Scotia)(Canadian Department of Forestry 1964, 1974, 1976). Embree (1965) found that mortalityoccurring between pupal and adult stages of Bruce's spanworm was density dependent. Heestimated mortality at between 15 and 49% (although he suggested that this may be low dueto experimental procedures). It is therefore apparent that generalist predators may beregulating Bruce's spanworm also.It seems likely that these ground predators would readily have moved from takingspanworm pupae to taking winter moth pupae. Why then was the increase in predationdelayed until after Cyzenis became established? It is most probable that these predatorshave been eating the winter moth pupae since the early stages of introduction. However, itis a noted feature of generalist predation that it is not capable of preventing outbreaks fromoccurring (Hance 1987). For some reason the winter moth population must have becomeuncoupled from the spanworm population and went into post-introduction outbreak. Thisreason may be the absence of parasitoids or disease among winter moth.Roland (1992) evaluated the effects of reduced levels of the observed parasitism eachyear on the rate of population change (Rt) at Victoria. He demonstrated that for the wintermoth on Vancouver Island strong regulation by predators allows levels of parasitism tovary greatly without the eruption of the winter moth population. However, if parasitismwere to fall below certain very small levels (less than 0.4% of the observed values eachyear), then the populations would erupt to prebiological control levels. This might explainwhy there was successful biological control over large areas of Nova Scotia and BritishColumbia at the same time, in spite of great spatial variations in the levels of parasitismattained at the sites. The observed increases in predation during the decline phase, arenecessary before predation assumes the regulatory process.148In Richmond populations, predation and parasitism were high. The populations havenot been monitored for long enough to suggest whether pupal predation may be temporallydensity dependent. Pupal predators may regulate the population. However, it was larvalmortality in 1992 that caused the greatest stage specific mortality. This suggests thatalthough regulating the moth, at high moth densities the predators are less efficient. Thelink between parasitism and predation, and between weather and predation, may be thatthese factors had driven winter moth populations to levels low enough that the pupalpredators could again gain efficient control of the population. The dramatic increasesobserved between 1989 and 1990 at Richmond indicate that there may also be some factorcausing increased levels of predation before the predators gain control. The absence of alink might simply have extended the period over which the predator control was gained orA. flaveolatum would have played a more important role, since it could drive thepopulations to a very low density, where Cyzenis alone would have allowed higherequilibria to have developed.Continued control of winter mothSince the introductions of Cyzenis and the eventual winter moth declines, thesituation in Canada has become much like that of Britain. Following the post-introductionoutbreak the moth is apparently regulated by generalist predators (Roland 1992). Amongthese, carabid beetles and beetle larvae appear to be very important. Levels of winter mothpupal predation at Richmond were found not to differ in spite of different habitats anddifferent complexes of predators. Continued control of winter moth should thereforeconcentrate on avoidance of interference and maintenance of suitable habitat for invertebratepredators.149Introduced carabids have become an important faunal component of North Americancultivated land. On Vancouver Island and in Nova Scotia, the most abundant beetles andthose implicated in winter moth predation, were all introduced from Europe. In Richmond,native species predominated in abundance at birch sites, but the majority of the mostcommon species were introduced. In blueberry fields, introduced beetles were again anotable feature with P. melanarius and H. affinus noted as possibly the most importantpredators. These two featured commonly in pest management of unsprayed and sprayedorchards in Ontario, where P. melanarius is noted as the most common carabid (Rivard1974, Hagley 1975, Hagley and Allen 1988). In general introduced species are becoming anotable feature of North American cultivated lands. Finlayson and Campbell (1976)studying agricultural land in the Lower Fraser Valley found Bembidion lampros (Hrbst.) asmall generalized European beetle to be the most abundant species on crop and fallow land.Other numerous species included Harpalus affinus , Calathus fuscipes (Goeze) and Cliviniafosser (L.) all introduced from Europe. Introduced carabids were most numerous oncultivated land and introduced staphylinids were an important component of grasslands(Finlayson and Campbell 1976). Introduced species therefore appear to do well in habitatthat has been affected by man's activities. Similar findings have been made for othergroups (i.e. dung beetles), and it has been suggested that European species have had agreater time to adapt to cultivation, while north American species are mainly forestspecialists (Spence and Spence 1988, Hanslci 1992). The Amara species are also a commonfeature of cultivated land and have been noted as common in all the above mentionedstudies (except Roland's). Amara were very common at blueberry site I (BBI), the sitemost approaching a commercial blueberry site at Richmond. These are native species(though A. erratica (Dfsch.) an introduced species was common in eastern Canada)favoured by dry habitat (Lindroth 1961-1969).150Roland (1992) has suggested that the reason for the successful biological control ofwinter moth on oak and not on apple may have to do with the use of pesticides in orchards.Similarly, Bruce's spanworm is a problem in orchards where pesticides are frequently used(McMullan 1973, Angenilli and Logan 1985). Pearsall (1992) found high levels of wintermoth pupal predation in unsprayed orchards in Nova Scotia. She (1992) did not estimatepredation of winter moth pupae in commercial (insecticide sprayed) orchards. However,these had similar abundances of beetles to unsprayed orchards. The compositions of beetlesat the orchards were very different. In commercial orchards Harpalus rufipes andPterostichus coracina were the dominant species, while in unsprayed orchards Carabusspp. were generally more abundant. A number of insecticides do adversely affect beetleabundances, but many authors suggest that carabids undergo rapid recolonization following'spraying (Rivard 1974, Roland 1986a, Brust et al. 1986, Underwood 1989, Pearsall1992). Unfortunately, each of these studies only used pitfall traps as a way of estimatingbeetle abundances. In orchards beetles are likely to disperse more rapidly due to lessground vegetation and so have a greater chance of being trapped. Lack of other insects (i.e.Collembola) and other food sources, especially worms and slugs may inhibit beetles fromstaying on sprayed land, and this could prevent constant buildup of beetle numbers.In general organo-phosphates and pyrethroids have been recommended againstwinter moth (Sanford and Herbert 1966, Tonks et al. 1978, AliNiazee 1986, Hardman andGaul 1990, Sheppard et al. 1990). In eastern Canada the use of pyrethroids has lead toproblems with the European red mite, and so use of Bt has been proposed. The success ofBt has been limited so that mixtures with conventional pesticides are now recommended(Sanford and Herbert 1966, Hardman and Gaul 1990). Simply allowing pupal predators tobuild up is not feasible for commercial orchards, blueberry plots, etc. because:151i) winter moth is generally not the sole pest. Leaf rollers were a noted feature ofRichmond blueberries. Other pests are not necessarily reduced by predators and sogenerally some insecticide is necessary.ii) Since winter moth larvae attack early buds, damage is often severe. Therefore,biological control is not feasible for winter moth on orchards and particularly on blueberry;andiii) Some of the invertebrate predators may actually damage the crop plants,particularly herbivores such as H. affinus and Amara spp., though all carabids are noted asconsuming vegetation to some degree (Briggs 1957,1966, Thiele 1977).There is however a clear need for a greater emphasis on IPM. A frequent use ofpesticides leads to a depletion of ground beetles and to a dependence on pesticides. Thesituation in forest systems is different. There the control of winter moth has been verysuccessful, with winter moth populations now behaving like populations in endemicsituations. Small outbreaks have occurred in Nova Scotia (Embree 1991), but these are notlikely to cause continued long term defoliation episodes as occurred in prebiological controlsituations. In urban environments where predator populations are expected to be lower,winter moth may be prone to more frequent outbreaks. In Europe, for example, wintermoth has remained as an important garden pest in urban environments (Speight 1979). Insuch situations C. albicans and A. flaveolatum may assume a more important role in thecontrol.152LITERATURE CITEDAgriculture Canada 1991. (Ed. Bousquet, Y.) Checklist of Beetles of Canada and Alaska.publication 1861/EAliNiazee, M.T. 1986. The European winter moth as a pest of filberts: Damage andchemical control. J. Entomol. Soc. Brit. Columbia 83: 6- 12.Alma, P.J. 1969. A study of the activity and behaviour of the winter moth Operophterabrumata (L.)(Lepidoptera: Hydriomenidae). Ent. Mon. Mag. 258 - 265.Angerilli, N.P.D., Logan, D.M. 1985. Early season apple pest management: Control oftwo species of scales (Homo.: Diaspidae) and Bruce spanworm (Lep.: Geometridae) withmethidathion. J. Entomol. Soc. Brit. Columbia 82: 31-35Anonymous 1990. British Columbia Ministry of Agriculture and Fisheries: BerryProduction Guide for Commercial Growers, 1990-1991 ed.Barbosa, P., Capinera, J.L. 1977. The influence of food on development characteristics ofthe gypsy moth, Lymantria dispar (L.). Can. J. Zool. 55: 1424-1429Barron, J.R. 1989. Status of the parasite Agrypon flaveolatum (Gravenhorst)(Hymenoptera, Ichneumonidae), introduced to control the winter moth in Nova Scotia andBritish Columbia. Can. Ent. 121: 11-26Bauer, G. 1985. Population ecology of Pardia tripunctat Schiff. and Notocelia roboronaDen. and Schiff. (Lepidoptera, Tortricidae): an example of 'equilibrium species'. Oecologia65; 437-441Bonnemaison, L. 1971. Particularites de la croissance embryonnaire et de la diapausenymphale chez la chematobie (Operophtera brumata L.). Bull. Soc. Ent. Fr. 76: 123-130Briggs, J.B. 1957. Some experiments on control of ground beetle damage to strawberry.44th Rep. E. Malling Res. Sta. 1955-1956. 142-145Briggs, J.B. 1965. Biology of some ground-beetles (Coleoptera, Carabidae) injurious tostrawberries. Bull. Ent. Res. 56. 79-93Brown, C.E. 1962. The life history and dispersal of the Bruce Spanworm, Operophterabruceata (Hulst),(Lepidoptera: Geometidae). Can. Ent. 94: 1103-1107Brust, G.E., Stinner, B.R., McCartney, D.A. 1986. Predator activity and predation incorn agroecosystems. Environ. Ent. 15: 1017-1021.Buckner, C.H. 1966. The role of vertebrate predators in the biological control of forestinsects. Ann. Rev. Ent. 11: 449 - 470.Buckner, C.H. 1969. The common shrew (Sorex araneus) as a predator of the winter moth(Operophtera brumata) near Oxford, England. Can. Ent. 101: 370-375153Campbell, R.W., Torgersen, T.R. 1983. Effect of branch height on predation of westernspruce budworm (Lepidoptera: Tortricidae) pupae by birds and ants. Environ. Ent. 12(3):697-699.Canadian Department of Forestry 1964. Annual Report of the Forest Insect and DiseaseSurvey 1964. Canadian Forestry Service.Canadian Department of Forestry 1974. Annual Report of the Forest Insect and DiseaseSurvey 1973. Canadian Forestry Service.Canadian Department of Forestry 1976. Annual Report of the Forest Insect and DiseaseSurvey 1975. Canadian Forestry Service.Canning, E.U., Barker, R.J. 1982. Transmission of microsporidia between generations ofwinter moth Operophtera brumata. Proc. Brit. Soc. Parasit. 1981. Parasitology, 84: xiv(abstracts).Canning, E.U., Barker, R.J., Nicholas, J.P., Page, A.M. 1985. The ultrastructure ofthree microsporidia from winter moth, Operophtera brumata (L.), and the establishment ofa new genus Cystosporogenes n.g. for Pleistophora operophterae (Canning, 1960). Syst.Parasitol. 7(3): 213-225Canning, E.U., Barker, R.J., Page, A.M., Nicholas, J.P. 1985. Transmission ofmicrosporidia, especially Orthsoma operophterae (Canning, 1960) between generations ofwinter moth Operophtera brumata (L) (Lepidoptera: Geometridae). Parasitology 90(1): 11-19Capinera, J.L., Barbosa, P., 1976. Dispersal of first instar gypsy moth larvae in relation topopulation quality. Oecologia 26: 53-64.Comic, J.F. 1973. Etude du regime alimentaire de trois especes de carabiques et de sesvariations en verger de pommiers. Ann. Soc. Ent. Fr. (N.S.) 9: 69-87Cox, D.L., Potter, D.A., 1986. Aerial dispersal behaviour of larval bagworms,Tryridoteryx ephemeraeformis (Lepidoptera: Psychidae). Can. Ent. 118: 525-536Cuming, F.G. 1961. The distribution, life history, and economical importance of wintermoth, Operophtera brumata (L.)(Lepidoptera: Geometridae) in Nova Scotia. Can. Ent. 93:135-142Cunningham, J.C., Tonks, N.Y., Kaupp, W. 1981. Viruses to control winter moth,Operophtera brumata (Lepidoptera: Geometridae). J. Entomol. Soc. Brit. Columbia. 78: 1724.Cunningham, W.L. 1982. Field trials with baculoviruses: control of forest insect pests. InMicrobial and viral pesticides (ed. Kurstak, E.) Marsell Dekker, New York.Danthanarayana, W. 1975. Factors determining variation in fecundity of the light brownapple moth, Epiphysis postvittana (Walker) (Tortricidae). Aust. J. Zool. 23: 439-451DeBach, P. 1974. Biological control by natural enemies. Cambridge University Press,London.154Den Boer, P.J. 1986. Density dependence and the stabilization of animal numbers 1. Thewinter moth. Oecologia 69: 507-512Den Boer, P.J. 1988. Density dependence and the stabilization of animal numbers. 3. Thewinter moth reconsidered. Oecologia 75(2): 161-168Dubrovin, U.V. 1990. Peculiarities of the distribution of Operophtera brumata L.(Lepidoptera: Geometridae) in plantations in Vorenezh province. Entomol. Obozr. 2: 281-286 (translated from Russian)East, R. 1974. Predation on the soil dwelling stages of the winter moth at Wytham woods,Berkshire. J. Anim. Ecol. 43: 611 - 626.Edland, T. 1971. Wind dispersal of the winter moth larvae Operophtera brumata L. (Lep.,Geometridae) and its relevance to control measures. Norsk Ent. Tidss. 18: 103-107Eidt, D.C. ,Embree,D.G. 1968. Distinguishing larvae and pupae of the winter mothOperophtera brumata, and the bruce spanworm 0. bruceata (Lepidoptera : Geometridae).Can. Ent. 100 : 536 - 539.Embree, D.G. 1965a. Population studies of Operophtera species, 0. brumata (L.), 0.bruceata (Hulst) and Pseudexentera cressoniana (Clem.) in Nova Scotia. Forest ResearchLaboratory, Moncton, New Brunswick, Internal Report M-2. Unpublished.Embree, D.G. 1965b. The population dynamics of winter moth in Nova Scotia 1954-1962.Mem. Ent. Soc. Can. 46: 1-57Embree, D.G. 1966. The role of introduced parasites in the control of the winter moth inNova Scotia. Can. Ent. 98 : 1159 - 1167.Embree, D.G. 1970. The diurnal and seasonal pattern of hatching of winter moth eggs,Operophtera brumata (Geometridae: Lepidoptera). Can. Ent. 102: 759-768.Embree, D.G. 1971. The biological control of winter moth in eastern Canada by introducedparasites. In: Huffaker, C.B. (ed). Biological Control. Plenum Press, New York. pp. 217-226Embree, D.G. 1991. The winter moth, Operophtera brumata, in eastern Canada, 1962-1988. For. Ecol. Manag. 39: 47-54Embree, D.G., Otvos, LS. 1984. Operophtera brumata (L.) winter moth (Lepidoptera:Geometridae). In Biological control programmes against insects and weeds in Canada 1969- 1980. (Ed.Kelleher,J.S., Hulme,M.A.) Commonwealth Agricultural Bureaux.Embree, D.G., Sisojevic, P. 1965. The bionomics and population density of Cyzenisalbi cans (Fall.) (Tachnidae: Diptera) in Nova Scotia. Can. Ent. 97: 631-639.Eskafi, F.M., Kolbe, M.M.,1990. Predation on larval and pupal Ceratitis capitata (Diptera:Tephritidae) by the ant Solenopsis geminata (Hymenoptera: Formicae) and other predatorsin Guatamala. Environ. Ent. 19(1): 148-153.Evans, M.E.G. 1977. Locomotion in the Coleoptera Adephaga, especially Carabidae. J.Zool. London 181: 189-226155Feeny, P.P. 1970. Seasonal changes in oak leaf tannins and nutrients as a cause of springfeeding by winter moth caterpillars. Ecology 51(4): 565 - 581Feeny, P.P. 1976. Plant apparency and chemical defense. Recent Adv. Phytochem. 10: 1-40Ferguson, D.C. 1978. Pests not known to occur in the United States or of limiteddistribution. U.S. Dep. Agric. Coop. Plant Pest Rep. 3 (48-52): 687-694.Finlayson, D.G., Campbell, C.J. 1976. Carabid and staphylinid beetles from agriculturalland in the Lower Fraser valley, British Columbia. J. Entomol. Soc. Brit. Columbia 73:10-20.Finnegan,R.J., Smirnoff,W.A. 1984. Neodiprion swainei (Middleton), Swaine Jack PineSawfly (Hymenoptera : Diprionidae). In Biological control programmes against insects andweeds in Canada 1969 - 1980. (Ed.Kelleher,J.S., Hulme,M.A.) CommonwealthAgricultural Bureaux.Fitzpatrick, S.M., Troubridge, J.J., Peterson, B. 1991. Distribution of the Europeanwinter moth, Operophtera brumata (L.) and the Bruce spanworm, 0 .bruceata (Hulst), inthe Lower Fraser Valley, British Columbia. J. Entomol. Soc. Brit. Columbia 88: 39 - 45.Frank, J.H. 1967a. The insect predators of the pupal stage of the winter moth, Operophterabrumata (L.)(Lepidoptera: Hydriomenidae). J. Anim. Ecol. 36 : 375 - 389.Frank, J.H. 1967b. The effect of pupal predators on a population of winter moth,Operophtera brumata (L.)(Hydromenidae). J. Anim. Ecol. 36: 611 - 621.Gillespie, D.R., Finlayson, T. 1981. Final-instar larvae of native hymenopterous anddipterous parasites of Operophtera spp. (Lepidoptera: Geometridae) in British Columbia.Can. Ent. 113: 45-55Gillespie, D.R., Finlayson, T., Tanks,N.V., Ross, D.A. 1978. Occurrence of the wintermoth, Operophtera brumata, (Lepidoptera : Geometridae), on southern Vancouver Island,British Columbia. Can. Ent. 110: 223 - 224.Greene, A. 1975. Biology of five species of Cychrini (Coleoptera: Carabidae) in the stepperegion of Southwestern Washington (USA). Melanderia 19: 1-43Greenslade, P.J.M. 1964. Pitfall trapping as a method for studying populations ofCarabidae (Coleoptera). J. Anim. Ecol. 33 : 301 - 310.Hagley, E.A.C. 1975. The arthropod fauna in unsprayed apple orchards in Ontario. II.Some predacious species. Proc. Ent. Soc. Ont. 105(1974): 28-40.Hagley, E.A.C., Allen, W.R. 1988. Ground beetles (Coleoptera: Carabidae) as predatorsof the codling moth, Cydia pomonella (L.)(Lepidoptera: Tortricidae). Can. Ent. 120: 917-925Hale, M.A. 1989. Factors affecting the distribution and survival of an endemic and anintroduced species of Operophtera (Lepidoptera: Geometridae). M.Sc. Thesis, Universityof Victoria.156Hance, T. 1987. Predation impact of carabids at different population densities on Aphisfabae development in sugar beat. Pedobiologia 30(4): 251-262.Hand, L.F., Keaster,A.J. 1967. The environment of an insect field cage. J. Econ. Ent. 60(4) : 910 - 914.Hanski, I. 1992. North temperate dung beetles. in (Hanski, I., and Cambefort, Y. eds)Dung beetle ecology. Princeton University Press, New JerseyHanski, I., Parviainen, P. 1985. Cocoon predation by small mammals, and pine sawflypopulation dynamics. Oikos 45: 125-136Hardman, J., Gaul, S.O. 1990. Mixtures of Bacillus thuringiensis and pyretroids controlwinter moth (Lepidoptera: Geometridae) in orchards without causing outbreaks of mites. J.Econ. Ent. 83 (3): 920 _ 936.Harris, P. 1984. Euphorbia esula-virgata complex, leafy spurge and E. cyparissias L.,cypress spurge (Euphorbiaceae).In Biological control programmes against insects andweeds in Canada 1969 - 1980. (Ed.Kelleher,J.S., Hulme,M.A.) CommonwealthAgricultural Bureaux.Hassell, M.P. 1968. The behavioural response of a tachinid fly (Cyzenis albicans [Fall)(Tachinidae) to its host, the winter moth (Operophtera brumata (L.)). J. Anim. Ecol. 38:627-639Hassell, M.P. 1969a. A population model for the interaction between Cyzenis albicans(Fall.)(Tachinidae) and Operophtera brumata (L.)(Geometridae) at Whytham, Berkshire. J.Anim. Ecol. 38: 567-576Hassell, M.P. 1969b. A study of mortality factors acting upon Cyzenis albicans (Fall.), atachinid parasite of the winter moth, (Operophtera brumata (L.)). J. Anim. Ecol. 38: 329-340Hassell, M.P. 1978. The dynamics of arthropod predator-prey systems. PrincetonUniversity Press, Princeton.Hassell, M.P. 1980. Foraging strategies, population models and biological control; A casestudy. J. Anim. Ecol. 49 : 603 - 628.Hassell, M.P. 1987. Detecting regulation in patchily distributed animal populations. J.Anim. Ecol. 56: 705-713.Hassell, M.P., May, R.M. 1974. Aggregation in predators and insect parasites and itseffects on stability. J. Anim. Ecol. 43: 567-594Hassell, M.P., May, R.M. 1986. Generalist and specialist natural enemies in insectpredator prey interactions. J. Anim. Ecol. 55: 923-940Haulcioja, E. 1990. Induction of defenses in trees. Ann. Rev. Entomol. 36: 25-42Hawboldt, D.G., Cuming, F.G., 1950. Cankerworm and European winter moth in NovaScotia. Dom. Dep. Agric. Sci. Serv. Div. Ent. For. Ins. Invest., Bi-mon. Progr. Rep.6(1): 1-2157Heichel, G.H., Turner, N.C. 1976. Phenology and leaf growth of defoliated hardwoodtrees. In (Anderson, J.F., Kaya, H.K. eds) Perspectives in forest entomology. AcademicPress, New York. pp 31-40.Heinrich, B. 1979. Foraging strategies of caterpillars: leaf damage and possible predatoravoidance strategies. Oecologia 42: 325-337Helitivaara, K., Vaisanen, R., Kemppi, E. 1989. Change of pupal size of Panolis flammea(Lepidoptera; Noctuidae) and Bupalus piniarius (Geometridae) in response to concentrationof industrial pollutants in their food plant. Oecologia 79: 179-183Hengeveld, R. 1980a. Qualitative and quantitative aspects of the food of ground beetles(Coleoptera, Carabidae): a review. Netherlands J. Zool. 30(4): 555-563Hengeveld, R. 1980b. Polyphagy oligophagy and food specialization in ground beetles(Coleoptera, Carabidae) Netherlands J. Zool. 30(4): 564-584Hengeveld, R. 1980c. Food specialization in ground beetles: an ecological or phylogeneticprocess? (Coleoptera, Carabidae). Netherlands J. Zoo!. 30(4): 585-594Holliday, N.J. 1977. Population ecology of winter moth (Operophtera brumata) on applein relation to larval dispersal and time of bud burst. J. App. Ecol. 14: 803-813.Holliday, N.J. 1983. Effects of temperature on winter moth pupae, Operophtera brumata(Lepidoptera: Geometridae). Can. Ent. 115 : 243 - 249.Holliday, N.J. 1985. maintenance of the phenology of the winter moth (Lepidoptera:Geometridae). Ecol. J. Linn. Soc. 25: 221 - 234.Holliday, N.J., Hagley, E.A.C. 1978. Occurrence and activity of ground beetles(Coleoptera: Carabidae) in a pest management apple orchard. Can. Ent. 110: 113-119.Humble, L.M. 1984. Emergence behaviour of Phobocampe sp. (Hymenoptera:Ichneumonidae); a larval endoparasitoid of Operophtera spp. (Lepidoptera: Geometridae).J. Ent. Soc. Brit. Columbia 81: 29-32Humble, L.M. 1985a. Final instar larvae of native pupal parasites and hyperparasites ofOperophtera spp. (Lepidoptera: Geometridae) on southern Vancouver Island. Can. Ent.117: 525-534.Humble, L.M. 1985b. Seasonal activity of ichneumonid pupal parasites of Operophteraspp. (Lepidoptera: Geometridae). J. Ent. Soc. British Columbia. 82: 44-46.Hunter, M.D. 1990. Differential susceptibility to variable plant phenology and its role incompetition between two insect herbivores on oak. Ecol. Ent. 15: 401-408Hunter, M.D., Watt, A.D., Docherty, M. 1991. Outbreaks of the winter moth on Sitkaspruce in Scotland are not influenced by nutrient deficiencies of trees, tree budburst orpredation. Oecologia 86: 62 - 69.Ito, F., Higashi, S. 1991. Variance of ant effects on the different life forms of mothcaterpillars. J. Anim. Ecol. 60: 327-334158Ives, W.G., Cunningham, J.C., 1980. Application of nuclear polyhedrosis virus to controlbruce spanworm (Lepidoptera: Geometridae) Operophtera bruceata, in an attempt atbiological control. Can. Ent. 112(7): 741-744Ives, W.G.H. 1984. Malacosoma disstria Hubner, forest tent caterpillar (Lepidoptera :Lasiocampidae) In Biological control programmes against insects and weeds in Canada1969 - 1980. (Ed.Kelleher,J.S., Hulme,M.A.) Commonwealth Agricultural Bureaux.Jeanne!, R. 1967. Faune de France. Vol. 40 Coleopteres carabiques (2eme partie). KrausReprint Ltd. Nendeln/LiechtensteinJohnson, U.E., Cameron, R.S. 1969. Phytophagous ground-beetles. Ann. Ent. Soc. Am.62: 909-914Kelly, B., Regniere, J. 1985. Predation on pupae of the spruce budworm (Lepidoptera:Tortricidae) on the forest floor. Can. Ent. 117: 33-38.Kikuzawa, K, Asai, T., Higashiura, Y. 1979. Leaf production and the effect of defoliationby the larval populations of the winter mth, Operophtera brumata L. in an alder (Alnusinokumae Mural et Kusaka) stand. Jap. J. Ecol. 29: 111-120Kimberling, D.N., Miller, J.C. 1988. Effects of temperature on larval eclosion of wintermoth, Operophterabrumata. Ent. Exp. Appl. 47(3): 249-254Kimberling, D.N., Miller, IC., Penrose, R.L. 1986. Distribution and parasitism of wintermoth, Operophterabrumata (Lepidoptera: Geometridae), in western Oregon. Environ. Ent.15(5): 1042-1046.Koponen, S. 1985. Herbivorous insects on planted birch in the Faroe Islands. Notul. Ent.65: 119- 122.Kowalski, R. 1976. Biology of Philonthus decorus (Coleoptera, Staphylinidae) in relationto its role as a predator of winter moth pupae [Operophtera brumata] (Lepidoptera:Geometridae). Pedobiologia 16: 233-242Kowalski, R. 1977. Further elaboration of the winter moth population models. J. Anim.Ecol. 46: 471-482.Laasonen, E.M., Laasonen, L. 1987. Birch defoliation caused by larvae of Operophterafagata (Lepidoptera, Geometridae) in Finland in the years 1982-1985. Not. Ent. 67(4):181-186Lareau, M. 1987. Vertical distribution of activity of carabid beetles in a beech forest floor.Pedobiologia 30, 173-178Latto, J., Hassell, M.P. 1987. Do pupal predators regulate the winter moth? Oecologia74(1): 153-155.Lindroth, C.H. 1961-1969. The ground-beetles (Carabidae, excl. Cicindelinae) of Canadaand Alaska. Parts 1-6. Opusc. Ent. x 1 viii + 1192 pp. 1961, Part 2, Suppl. 20: 1-200,1963 Part 3, Suppl. 24: 201-408, 1966, Part 4, Suppl. 29: 409-648, 1968, Part 5, Suppl.33: 649-944, Part 6, Suppl. 34: 945-1192, Part 1, Suppl. 35: i - xlviii159Ltivei, G.L., Monostori ,E., Ando, I. 1985. Digestion rate in relation to starvation in thelarvae of a carabid predator, Poecilus cupreus. Ent. Exp. Appl. 37: 123 - 127.Ludwig, J.A., Reynolds, J.F. 1988. Statistical ecology: a primer on methods andcomputing. John Wiley and sons, New York.Luff, M.L. 1983. The potential of predators for pest control. Agric. Ecosyst. Environ. 10:159-181Macphee, A.W. 1967. The winter moth, Operophtera brumata (Lepidoptera; geometridae),a new pest attacking apple orchards in Nova Scotia, and its coldhardiness. Can. Ent. 99:829 - 834.Malaisse, F. 1964. Contribution a l'etude des hôtries d'Europe occidentale. Bull. Soc.Royale de Bot. Belg. 97: 85-97Mauricio, R., Deane Bowers, M., 1990. Do caterpillars disperse their damage? larvalforaging of two specialist herbivores, Euphydryas phaeton (Nymphalidae) and Pieris rapae(Pieridae). Ecol. Ent. 15: 153-161May, R.M. 1978. Host parasitoid systems in patchy environments: a phenomenologicalmodel. J. Anim. Ecol. 47: 833-843McLeod, J.M., 1966. The spatial distribution of cocoons of Neodiprion swainei Middletonin a Jack pine stand: I. A cartographic analysis of cocoon distribution, with specialreference to predation by small mammals. Can. Ent. 98: 430-447McManus, M.L. 1973. The role of behaviour in the dispersal of newly hatched gypsy mothlarvae. USDA For. Serv. Res. Pap. NE-267McMullen, R.D. 1973. The occurrence and control of the Bruce Spanworm in theOkanagan valley, 1972. J. Entomol. Soc. Brit. Columbia 70: 8-10McPhee, A., Newton, A., McRae, K.B. 1988. Population studies on the winter mothOperophtera brumata (L.)(Lepidoptera: Geometridae) in apple orchards in Nova Scotia.Can. Ent. 120: 73-83.Miller, J.C. 1982. Sampling for larvae of Operophtera occidentalis (Lepidoptera:Geometridae) on cherry. J. Econ. Ent. 75: 1021-1024Miller, J.C., Cronhardt, J.E. 1982. Life history and seasonal development of the westernwinter moth, Operophtera occidentalis (Lepidoptera: Geometridae), in Western Oregon.Can. Ent. 114: 629 - 636.Morris, R.F. 1972. Fecundity and colony size in natural populations of Hyphantria cunea.Can. Ent. 104: 399-409Murdoch, W.W. 1966. Aspects of the population dynamics of some marsh Carabidae. J.Anim. Ecol. 35: 127-156Murdoch, W.W., Chesson, J., Chesson, P.L. 1985. Biological control in theory andpractice. Amer. Nat. 125: 344-366.160Myers, J.H. 1988. Can a general hypothesis explain population cycles of forestlepidoptera? Adv. Ecol. Res. 18: 179-242Parmenter, R.R., MacMahon, J.A. 1983. Factors determining the abundance anddistribution of rodents in a shrub steppe ecosystem: the role of shrubs. Oecologia 59: 145-156Parmenter, R.R., MacMahon, J.A. 1988. Factors influencing species composition andpopulation sizes in a ground beetle community (Carabidae): Predation by rodents. Oikos 52: 350 - 356.Pearsall (1993) Mortality of winter moth (Operophtera brumata L.)(Lepidoptera:Geometridae) in Nova Scotian apple orchards. (in Press)Pearsall, I.A., 1992. Mortality of winter moth populations in Nova Scotian apple orchards.M.Sc. Thesis Dalhousie University.Pivnick, K.A. 1988. Wing colouration difference between the bruce spanwormOperophtera bruceata (Hu1st) (Lepidoptera: Geometridae) and the winter moth Operophterabrumata (L.) on Vancouver Island. Can. Ent. 120(7): 697-698Pivnick, K.A., Barton, D.L., Millar, J.G., Underhill, E.W. 1988. Improved pheromonetrap exclusion of the bruce spanworm Operophtera bruceata (Hulst)(Lepidoptera:Geometridae) when monitoring winter moth Operophtera brumata (L) populations. Can.Ent. 120: 389 - 396.Poethke, H.J., Kirchberg, M. 1987. On the stabilizing effects of density dependentmortality factors. Oecologia 74(1): 156-158.Poinar, G.O., Thomas, G.M., 1984. Laboratory guide to insect pathogens and parasites.Plenum Press, New York.Rivard, I. 1965. Dispersal of ground beetles (Coleoptera: Carabidae) on soil surface. Can.J. Zool. 43: 465-474Roelofs, W.L., Hill, A.S., Linn, C.E., Munwald, J., Jain, S.C., Herbert, H.T., Smith,R,F. 1982. Sex pheromone of the winter moth, a geometrid with unusually lowtemperature precopulatory responses. Science 217: 657 - 659.Roland, J. 1986a. Success and failure of Cyzenis albi cans in controlling its host the wintermoth. Ph.D. Thesis, Dept. Zool. U.B.C.Roland, J. 1986b. Parasitism of winter moth in British Columbia during build up of itsparasitoid Cyzensis albi cans : Attack rate on oak vs. apple. J. Anim. Ecol. 55 : 215 - 234.Roland, J. 1988. Decline of winter moth populations in North America: Direct versesindirect effect of introduced parasites. J. Anim. Ecol. 57 : 523 - 531.Roland, J. 1990. Interaction of parasitism and predation in the decline of winter moth inCanada. In Population Dynamics of Forest Insects. (Ed. Watt, A.D., Kidd, N., Leather,S., Hunter, M.) Intercept Publishers, Andover.161Roland, J. 1990. Parasitoid aggregation: chemical ecology and population dynamics. InCritical Issues in Biological Control. (Ed. MacKauer, M., Ehler, L.E., Roland, J.)Andover Hants: Intercept.Roland, J. 1992. After the decline: what maintains low winter moth density after successfulbiological control? unpuplished.Roland, J., Evans, W.G., Myers, J.H. 1989. Manipulation of oviposition patterns of theparasitoid Cyzenis albicans (Tachinidae) in the field using plant extracts. J. Insect. Behav.2: 487-503Roland, J., Hannon, S.J., Smith, M.A. 1986. Foraging pattern of pine siskins and itsinfluence on winter moth survival in an apple orchard. Oecologia 69: 47-52.Roland, J., Myers, J.H. 1987. Improved insect performance from host plant defoliation:winter moth on oak and apple. Ecol. Ent. 12: 409-414Sanford, K.H., Herbert, H.J. 1966. The influence of spray programs on the fauna ofapple orchards in Nova Scotia. XX. Chemical controls for winter moth, Operophterabrumata (L.), and their effects on phytophagous mites and predator populations. Can. Ent.98: 991-999.Scherney, F. 1959. Der biologishe Wirkungseffekt von Carabiden der Gattung Carabusatif Kartoffelkafer-larven. Verh. 4 Intern. Pflanzenschutz Kongr. Hamburg. 1957, 1:1035-1038Screiber, E.T., Campbell, J.B., Boxler, D.J., Peterson, J.J. 1987. Comparison of beetlescollected from the dung of cattle untreated and treated with Fenvalerate ear tags on pastureand two range types in western Nabraska. Environ. Ent. 16: 1135-1140Sechser, B. 1970a. Der Parasitenkomplex des Kleinen Frostspanners (Operophterabrumata L.)(Lep., Geometridae) unter besonderer Beriicksichtigung der Kokonparasiten.Z. Ang. Ent. 66: 1-35Sechser, B. 1970b. Der parasitenkomplex von 0 .brumata in Kanada sowie biologischeBekampfungsversuche. Z. Ang. Ent. 66: 144-160Sheppard, D.H, Myers, J.H., Fitzpatrick, S., Gerber, H. 1990. Efficiency of deltamethrinand Bacillus turingiensis Berliner ssp. Kurstaki on the larvae of winter moth, Operophterabrumata (L.)(Lepidoptera: Geometridae) attacking blueberry in the lower mainland ofBritish Columbia. J. Entomol. Soc. Brit. Columbia 87: 102-110Skou, P. 1986. The Geometrid moths of northern Europe (Lepidoptera: Drepanidae andGeometridae) in Entomonographvol 6 (Ed. L.Lyneborg) Scandanavian Science Press.Smith, C.C. 1950. Notes on the European winter moth in Nova Scotia. Dom. Dep. Agric.Sci. Serv. Div. Ent. For. Ins. Invest., Bi-mon. Progr. Rep. 6(2): 1Speight, M.R. 1979. Tree pests. 1 winter moth, Operophtera brumata (L.). Arboric. J. 3(490 - 491.Spence, J.R., Spence, D. 1988. Of ground-beetles and men: introduced species and thesynanthropic fauna of western Canada. Mem. Ent. Soc. Can. 144: 151-168162Stairs, G.R. 1966. Transmission of virus in tent caterpillar populations. Can Entomol. 98:1100- 1104Stoaldey, J.J. 1985. Outbreaks of winter moth, Operophtera brumata (Lep., Geometridae)in young plantations of sitka spruce in Scotland. insecticidal control and populationassessment using the sex attractant pheromone. Z.Ang.Ent. 99: 153 - 160Tauber, M.J., Tauber, C.A. 1976. Insect seasonality: diapause, maintenance, terminationand postdiapause development. Ann. Rev. Ent. 21: 81-107Tenow, 0. 1972. The outbreaks of Oporina autumnata Bkh. and Operophtera spp. (Lep,Geometridae) in the Scandanavian mountain chain and northern Finland 1862 - 1968. Zool.Bidrag Suppl. 2: 1 - 107.Thiele,H-U. 1977. Carabid beetles in their environments, a study on habitat selection byadaptations in physiology and behaviour. Springer - Verlag, BerlinTonks, N.V., Everson, P.R., Theaker, T.L. 1978. Efficiency of insecticides againstgeometrid larvae, Operophtera spp., on southern Vancouver Island, British Columbia. J.Entomol. Soc. Brit. Columbia 75: 6 -9Topp, W., Kirsten, K. 1991. Synchronization of pre-imaginal development andreproductive success in the winter moth Operophtera brumata L.. J. Appl. Ent. 111: 137 -146.Torgersen, T.R., Mason, R.R., Paul, H.G. 1983. Predation on pupae of Douglas-firTussock moth, Orgyia pseudotsugata (McDonnough) (Lepidoptera: Lymantriidae).Environ. Ent. 12: 1678-1682Troubridge, IT., Fitzpatrick, S.M. 1993. A revision of the North American Operophtera(Lepidoptera: Geometridae). Can. Ent. 125: 379-397Underhill, E.W., Millar, J.G., Ring, R.A., Wong, J.W., Barton, D., Giblin, M. 1987.Use of a sex attractant and an inhibitor for monitoring winter moth and bruce spanwormpopulations. J. Chem. Ecol. 13: 1319-1330Underwood, A.J. 1989. The analysis of stress in natural populations. Biol. J. Linn. Soc.37: 51-79Valentine, H.T., Wallner, W.E., Wargo, P.M. 1983. Nutritional changes in host leafquality during and after defoliation, and their relation to the weight of gypsy moth pupae.Oecologia 57: 298-302Varley, G.C., Gradwell, G.R. 1960. Key factors in population studies. J. Anim. Ecol. 29:399-401.Varley, G.C. 1970. The effects of natural predators and parasites on winter mothpopulations in England. Proceed. Tall-Timbers Conf. Ecol. Anim. Cont. by Hab. Manag.No. 2 Tallahassee Florida. 103-116Varley, G.C., Gradwell, G.R. 1963a. The interpretation of insect population changes.Proc. Ceylon Ass. Adv. Sci. 18: 142-156163Varley, G.C., Gradwell, G.R. 1963b. Predatory insects as density dependent mortalityfactors. Proc. 16th Inter. Congr. Zool., Washington, D.C. 1: 240Varley, G.C., Gradwell, G.R. 1968. Population models for the winter moth. Symp. RoyalEntomol. Soc. London 4: 132-142Varley, G.C., Gradwell, G.R. 1970. Recent advances in insect population dynamics. Ann.Rev. Ent. 15: 1-24Varley, G.C., Gradwell, G.R. 1971. The use of models and life-tables in assessing therole of natural enemies. In: Hauffaker, C.B. (ed) Biological control. Plenum Press, NewYork., pp. 93-112Varley, G.C., Gradwell, G.R. 1973. Insect Population Biology. Blackwell ScientificPublications, Oxford.Varley, G.C., Gradwell, G.R., 1958. Oak defoliators in England. Pro. 10th Inter. Congr.of Ent. Vol 4: 1956, 133 - 136.Varley, G.C., Gradwell, G.R., Hassell, M.P. 1973. Insect Population Ecology.University of California, Berkley. 212pp.Watt, A.D., MacFarlane, A.M. 1991. Winter moth on Sitka spruce: synchrony of egghatch and budburst, and its effect on larval survival. Ecol.Ent. 16: 387 - 390.Wellington, W.G. 1962. Population quality and the maintenance of nuclear polyhedrosisbetween outbreaks of Malacosomapluviale (Dyar). J. Insect. Pathol. 4: 285-305Weseloh, R.M. 1985. Dispersal, survival and population abundance of gypsy moth,Lymantria dispar (Lep. Lymantriidae), larvae determined by releases and mark-recapturestudies. Ann. Entomol. Soc. Am. 78: 728-735West, C. 1985. Factors underlying the late seasonal appearance of the Lepidopterous leafmining guild on oak. Ecol. Ent. 10: 111-120Wigley, P.J. 1976. The epizootiology of a nuclear polyhedrosis virus of the winter moth,Operophtera brumata L., at Wistman's wood, Dartmoor. D.Phil. Thesis OxfordUniversity, England. 185pp.Williams,G. 1958. Mechanical time sorting of pitfall captures. J. Anim. Ecol. 27 : 26 - 35.Williamson, G.D. 1981a. Insect liberations in Canada. Parasites and predators. 1978.Agriculture Canada Research Branch..Williamson, G.D. 1981b. Insect liberations in Canada. Parasites and predators. 1979.Agriculture Canada Research Branch. 19pp.Williamson, G.D. 1981c. Insect liberations in Canada. Parasites and predators. 1980.Agriculture Canada Research Branch. 19pp.Williamson, G.D. 1984. Insect liberations in Canada. Parasites and predators. 1981.Agriculture Canada Research Branch. 19pp.164Williamson, G.D. 1986. Insect liberations in Canada. Parasites and predators. 1982.Agriculture Canada Research Branch..Wint, W. 1983. The role of alternative host plant species in the life of a polyphagous moth,Operophtera brumata (Lepidoptera: Geometridae). J. Anim. Ecol. 52:439 - 450.Wood, C.S., Van Sickle, G.A. 1985. Forest insect and disease conditions, BritishColumbia and Yukon 1985. Canadian Forest Service.Wood, C.S., Van Sickle, G.A. 1986. Forest insect and disease conditions, BritishColumbia and Yukon 1986. Canadian Forest Service.Wood, C.S., Van Sickle, G.A. 1990. Forest insect and disease conditions, BritishColumbia and Yukon 1985. Canadian Forest Service. BC-X-326Wood, C.S., Van Sickle, G.A. 1991. Forest insect and disease conditions, BritishColumbia and Yukon 1985. Canadian Forest Service. BC-X-334Wylie, H.G. 1960. Insect parasites of the winter moth, Operophtera brumata (L.)(Lepidoptera: Geometridae) in western Europe. Entomophaga 5: 111-129Wylie, H.G. 1960. Some factors that affect the annual cycle of the winter moth,Operophtera brumata (L.)(Lepidoptera: Geometridae) in western Europe. Ent. Exp.Appl. 3: 93 - 102.165Figure 1. Dispersal chamber.APPENDIX 1: LARVAL DISPERSAL AND PREFERENCE FOR APICALBUDS.In 1992, experiments were conducted to examine the dispersal behaviour of winter mothneonates on stage II blueberry buds. I wanted to examine whether dispersal was dependenton i) larval density, bud size and^the position of the buds on the branch.Adult moths were collected during the winter of 1991 from stocking traps and from pupaemaintained in the laboratory. Adults were mated and oviposited on sponge (as in Hale1989). Eggs were kept in an outside shed until they turned brown, and then were held in arefrigerator at 40C until required. Blueberry twigs were collected and pruned so that only 4flower buds remained on each twig. Bud length and width were measured and multiplied togive an index of budvolume. Twigs wereheld in cardboardcontainers as shown infigure 1. WinterTwig^ moth neonates which^ Wind direction^paintbrush, at equaltwig. Four densitieshad hatched during theprevious night, wereapplied with aspacing along thewere set up: 2, 5, 10and 20 per twig(N=5). Twigs wereranked according totheir bud sizes and thetreatments wereTanglefoot^arranged so that budsizes were as equal aspossible. Wind wassimulated with anelectric fan, and theexperiment was set upas a Latin squaredesign. The numbersof larvae in the tanglefoot were recorded at irregular intervals over 7 days. A similarexperiment was set up in the field at BBI. Twigs were chosen and pruned as above.Tanglefoot was applied at the base of each twig to restrict larval movement. Densities of 5,10 and 20 neonates were applied (N=12) Half of the twigs were clipped after 6 hours andthe rest were collected 24 hours later.Results indicated that larval dispersal was not density dependent either in the field or in thelaboratory experiment. Most of the dispersal in the field occurred within the first six hours(Fig. 2). There was much less dispersal in the laboratory than in the field. In thelaboratory, dispersal commenced earlier from high density branches, but the proportiondispersing did not vary among treatments. The dispersal rate tended to be greatest fromtwigs with smaller buds. Apical buds are larger than subapicals (1-way ANOVA, P>0.001[N=59]). First and second buds were not different (Tukey test, P=0.999), and third andfourth buds were not different (Tukey test, P=0.538). However, third and fourth budswere smaller than firsts and seconds (Tukey test, P>0.02). Twigs collected from the field166experiment were examined, the numbers of larvae remaining in the buds were recorded andtested with a )c2 analysis. Significantly more larvae were found on the two apical buds thanon the remaining buds at each density (Table 1).Table 1. Chi-squared analysis of larval distribution on blueberry buds, at three larvaldensities. Bud position refers to the position of the buds on individual twigs, budsnumbered 1 are the apical buds, 2 are subapicals, etc.. ** = P> 0.005, *** = P>0.001.Density Bud position N X2 DF1 2 3 420 23 18 10 8 59 9•9** 310 23 12 2 4 41 26.6*** 35 10 11 2 1 24 13.7** 3pooled66 49 14 17 146 52.4*** 3100^ 200TIME SINCE COLONIZATION (Hrs.)Figure 1. Larval dispersal from stage II blueberry buds under laboratory (solid line) andfield (dashed lines) conditions. Numbers indicate larval densities per twig. Note that theproportions of larvae dispersed from each treatment are similar. For the laboratoryexperiment this is about 20% while for the field experiments it is about 50%.167APPENDIX 2: WINTER MOTH INSTARS300II^III^IV^V200 -100 - 1 1.11.2 . 1. Distribution of head capsule widths among larvae collected at Richmond NaturePark in 1992. Five instars can be distinguished as indicated with roman numerals.168APPENDIX 3: PARASITISM AND PUPAL DEATH OF WINTER MOTH.K-values for parasitism and death due to unknown reasons have been presented in the text,for ease of comparison with other published studies. The following tables presentparasitism and death due to unknown factors as percentages of prepupae.Table 1 Levels of Parasitism at four field sites in Richmond, between 1989 and 1992.Standard errors are presented in brackets and are not available for 1989.SITE 1989 1990 1991 1992BI 43 23.68 (5.36) 53.53 (4.87) 38.31 (5.83)RNP 38 18.73 (2.29) 55.18 (7.44) 36.69 (7.06)BBI 3 12.43 (7.21) 13.79 (9.21) 50 (25)BBII 18 23.73 (3.03) 36.28 (9.71) 0Table 2 Incidence of mortality due to unknown factors among pupae at four field sites inRichmond, between 1989 and 1992. Standard errors are presented in brackets and are notavailable for 1989.SITE,1989 1990 1991 1992BI 10.23 15.11 (5.41) 8.13 (1.52) 19.75 (4.92)RNP 7.77 28.02 (2.72) 8.6 (2.45) 36.69 (7.06)BBI 21.05 18.11 (8.30) 0 50 (25)BBII^_ 3.92^_ 13.36 (18.35) 1.78 (0.99) 41.6 (18.31)169APPENDIX 4: WINTER MOTH DISEASEWinter moth [Operophterabrumata (L.)] of every life-stage were taken from wildpopulations at Richmond and reared in the laboratory in 1991 and 1992. Caterpillars werereared in individual cups and fed ample quantities of fresh birch, apple or blueberry foliage.Larvae were carefully observed for the incidence of viral death. Of 628 larvae reared in1991 and 1021 in 1992, there was no indication of viral disease. All larvae which had diedin the laboratory and a number of extra larvae taken from the field (56 in 1991 and 339 in1992) were smeared and stained. At blueberry site I (BBI) in 1991, 18 fifth instar larvaecollected on May 9th showed symptoms of disease. These individuals had severeconstriction of the first abdominal segments, with the second, third and fourth abdominalsegments also constricted. Larvae were sluggish and stopped feeding before death. All 18larvae died and were smeared and stained (for details of staining see text Chapter 2).Fourteen of the larvae had dark staining ovoid bodies of 2 to 4 gm. Most of these weremicrosporidia. However, the larger bodies may have been cytoplasmic polyhedrosis virus(CPV). In 1991, 69 adults and 120 eggs and in 1992, 27 adults and 28 pupae were stainedand smeared, but microsporidia, cytoplasmic polyhedris virus (CPV) and nuclearpolyhedrosis virus (NPV) were not observed.In 1992, an experiment was carried out to examine the cross-infection of 0. brumata and. bruceata NPV. Viruses were supplied by Dr. J. Cunningham, Forest insect and diseaselaboratory, Saulte St. Marie, Ontario. Larvae from four instars were collected fromRichmond on April 29th and infected with 0. brumata NPV and 0. bruceata NPV on May1st. Larvae were inoculated as follows: 400 PIB (+) 0. brumata NPV per individual andhigh doses of 0. bruceata NPV (the actual PIB counts for the 0. bruceata NPV culture wasnot estimated). Larvae of Bruce's spanworm [0. bruceata (Hulst)] were collected from theOkanagan on May 6th and infected on May 10th with high doses of 0. bruceata NPV.Larvae were fed virus on 5mm diameter birch leaf discs. Results of this cross- infectionexperiment are presented in Table 1. Mortality of the primary hosts was high for bothspecies (90-100% for winter moth and 78.6% for spanworm). The 0. bruceata NPV wasnot a pure culture and there was a high incidence of CPV among the dead spanwormlarvae. Only two (10%) early instar winter moth larvae treated with 0. brumata NPVshowed signs of infection. Both died on the 23rd of May and in each case these had bothCPV and NPV infections. Naturally occurring CPV, has not been recorded from wintermoth, CPV reduces fecundity and causes deformities of adult winter moth, and cross-infection among Lepidopteran species is apparently high. Eleven major CPV's have beendistinguished, several of which are capable of infecting more than one host (see Wigley1976). These results indicate that there could be a low incidence of disease among theRichmond population. Viral cross-infection from 0. bruceata is possible, but is probablyvery limited due to the high amounts of innoculum required and the resistance of the lateinstars.170TABLE 1. Results of 1991 cross-infection experiments. 'N' indicates thenumbers of larvae used in each treatment, 'S' is the number of larvaewhich survived to pupation, '+V' indicates the number of larvae whichwere positively identified as having NPV infection. A number of larvaewhich died for reasons other than viral infection. Possible explanations fortheir deaths are presented ('Other'). 'Wm' = winter moth, 'bs' =spanwonn.N P V_0. brumata 0. bruceata ControlSp. Instar N S +V Other N S +V Other N S +V Otherwm 2 35 0 32 ? 20 8 2 failedpupae/dessica-Lion19 9 03/4 10 1 9 10 5 0 ? 14 12 05 100 10 5 0 0 failedpupae5 0 0 failedpupaebs^5 14 3 11 14 4 0 pupaldeath171APPENDIX 5: PITFALL TRAP CATCHES AT RICHMOND.Table 1. Species trapped at Richmond in 1991 and 1992. Generalist invertebrate predators and beetlelarvae are not included in this table.SPECIES Total number caughtBBI BBII RNP RNPII BI BII1991 1992 1991 1992 11991 1992 1992 [1991 1992 1992Ground-beetlesspecialistsCallisthenes wilkerii 1Carabus granulatus 3 7 21 5 8 37 81C.nemoralis 11 15 5 1 5 11Laevitus ferruginosis 2 1Loricera decempunctata 2 2 4 2 13 1 10 12Notiophilus sp. 4 6 12Dyschinius sp. 1 8 44 12 10 5Dyschinius sp. 2 1 2Scaphinotus marginatus 8 60 1 23 88 25 104 106Carrion feedersNicrophora spp. 1 2 15 1 38 69 73 116 99 51Catharis rufa 2 3 5 6 40 157HydrophilidaeSp. 1 11 15 2 108 32 15Cercyon spp. 4StaphylinidaeS.1 1 85 2 3S. 2 (Philonthus sp.) 2 104 27 46 38S. 3 (Philonthus sp.) 1 12 3 9 3 31S. 4 (Phdonthus sp.) 1 1 2S. 5 (Philonthus sp.) 5S. 6 2 3 3 2 4S.7 1S. 8S. 9 (Tachinus sp) 3S. 10 2 1S. 11 (Philonthus sp.) 1 3 1S.12 7S. 13 (Tachinus sp.) 3S. 14 (Tachinus sp.) 18S. 15 (Queduis sp.) 1AntsMyrmica sp. 265 103 2 2Formica sp. 57 60 5 3 2Dennaptera 2 9 2 2Small Mammals 5 1 9 14 13 6 45 2 7Species Habitat BB B Over-winteringstageBody length PupalpredatorNativeAmara aurata Dejean 1828 Found in ch-y areas with *little vegetation.? 5.6 - 7mm *'''A. litteralis Mannerheim 1843 In open moderately dry *country with rich andincoherent vegetation.Hibernateas adults6.2- 9.3mm *^*Usually pronouncedweed characterfavoured by humanactivities.A. laevipennus Kirby 1837 In open dry country^*with grass or meadowvegetation. Usually onsandy moraine.Hibernateas adults5.8 - 7.1mm ?^*Cal athus fuscipes (Goeze) Found on open ground^*with meadow or weedvegetation, synathropic.Hibernateas larvae.10 - 15mm ?Callisthenes wilkerii LeConte 1852 Xerophilous in dry open^*country sometimes inmeadows and thinforests.? 15- 18mm ?^*Carabus granulatus Linne 1758 Open light deciduous^*^*woods, usually nearwater, often oncultivated ground wheresoil is moist.Hibernateas adults.16- 24mmC. nemoralis Mtiller 1764 Restricted to cultivated^*^*ground.Hibernateas adults.21- 26mm *Harp alas offinus (Schrank) 1781 Open dry country -^*favoured by humanactivities.Hibernateas adults.8.5- 12mm *H. rufipes (DeGeer) 1774 Open dry country^*especially cultivatedfields (mainly seedeater).Hibernateas adultsand larvae10- 16.7mm *Pterostichus algidus Leconte 1852 Less pronounced forestspecies, also occurs inopen country with richvegetation.Hibernateas larvae.12 - 16mm *P. herculaneus Mannerheim 1843 In dense but often dry^*conifer or mixedforests.Hibernateas larvae.13.5 - 17mm *P. melanarius (lliger) 1798 In light forests and open *^*meadows, favourscultivated land andwaste places.Hibernatemainly aslarvae.12- 19mm *Scaphinotus marginatus (Fischer) 1822 Eurytopic, in southern^* Hibernate 11.5- 19mm ?^*regions mainly inforests, in particularnear the margins larvae.APPENDIX 6: GROUND-BEETLE INFORMATION.Table 1. Biological information on ground-beetles from this study (taken from Lindroth 1961-1969 and Agriculture Canada 1991).173Table 1. cont.Leistus ferruginosus Mannerheim 1843^On moderately moist^* ?^7.8 - 9.3mm ?half-shaded ground,usually near runningwater.Loricera decempunctata Eschscholtz 1833 In the vicinity of water.^*^Hibernate 6.7- 8.1mmas adultsNotiophilus sp.^ Forest zerophilus.^varies^6 - 8mm^?^?Bembidion sp. Mostly hygrophilus^* ?^6- 8mm ?Agonum sp.^ Mostly hygrophilus ?^5- 8mm^?^?174APPENDIX 7: WINTER MOTH POPULATION TRENDS•1------1SOIL MORTALITY^-?NI a o a 9-040.-45-1)`<b-o--4`-o - r'› 43 V....1955^1965^1975^1985TIMEFigure 1. Winter moth population data from two oak woods in England. 'A' indicates thepopulation densities of winter moth at Whytham Wood and Wistman's Wood. 'B' indicatesthe mortality due to parasitism by Cyzenis albicans (open circles) and mortality due to soilmortality (closed circles). (Data are taken from Varley, Gradwell & Hassell 1973 andWigley 1976).PARASITISM01945175Figure 2. Winter moth population trends in Nova Scotia. Population trends at Oak Hillare shown in Part 1, with trends at Shefield orchard in Part 2. population densities (A) andmortality due to C. albicans parasitism (open circles) and soil mortality (closed circles) (B)are shown for each site. (Data are taken from Embree 1966, Roland 1988 and McPhee etal. 1988).176200000OAK HILL100000 -1oSOIL MORTALITY1954 1956 1958 19'60 1962 1964 1966I ^/3 ARASITISMTIME195223 B2 - SOIL MORTALITY1 -0PARASITISM- , 1) -v 00001960^1970^1980TIME1990177Figure 3. Winter moth population trends in western North America. Population trends areshown at: 1) Victoria and 2) Oregon and at Richmond on 3) birch and 4) blueberry.Population densities are presented (A) with an indication of parasitism by C. albi cans (opencircles) and pupal predation or soil mortality (closed circles) at the sites (B). Data on pupalpredation is not available from Oregon and parasitism did not exceed a K-value of 0.02.(Data is taken from Kimberling et al. 1986, Roland 1988 and 1992, Sheppard et al. 1990, apersonal communication from the Winter Moth Committee of British Columbia and fromthis study.1781cm5gWa,1>1*7v)•-7>20001000 -AVICTORIA-.....0B SOIL MORTALITYPAR SITISM01980 1982 1984 1986 1988 1990TIMEA20.50.01979 1980 1981 1982 1983>0.0PREDATION0^•^N^.....^I^•^1988 1989 1990 1991 1992 1993TIMEPARASITISM•^:•0 •1988 1989 1990 1991 1992 1993TIME179


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



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


Related Items