UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Influences of weather and temperature on disease : implications for the population dynamics of western… Frid, Leonardo 2001

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

Item Metadata

Download

Media
831-ubc_2002-0081.pdf [ 1.84MB ]
Metadata
JSON: 831-1.0099608.json
JSON-LD: 831-1.0099608-ld.json
RDF/XML (Pretty): 831-1.0099608-rdf.xml
RDF/JSON: 831-1.0099608-rdf.json
Turtle: 831-1.0099608-turtle.txt
N-Triples: 831-1.0099608-rdf-ntriples.txt
Original Record: 831-1.0099608-source.json
Full Text
831-1.0099608-fulltext.txt
Citation
831-1.0099608.ris

Full Text

INFLUENCES OF WEATHER AND TEMPERATURE ON DISEASE: IMPLICATIONS FOR THE POPULATION DYNAMICS OF WESTERN TENT CATERPILLARS by LEONARDO FRID B.Sc. The University of British Columbia 1999 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in THE F A C U L T Y OF GRADUATE STUDIES (Department of Zoology) We accept this as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA November 2001 © Leonardo Frid, 2001 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of " Z o o ^^yj The University of British Columbia Vancouver, Canada Date Mo* /5 2oo) ABSTRACT Studies on insect population outbreaks have failed to identify a single, general cause for population cycles. However, few studies have considered how various factors may together cause periodic outbreaks. Here I consider how disease and weather might interact to cause population cycles of western tent caterpillars, Malacosoma californicum pluviale (Dyar). A multiple regression analysis indicates that nuclearpolyhedrovirus (NPV) prevalence in western tent caterpillar populations increases with both host density and hours of sunshine during the larval feeding season. I explore the potential impact of climatic variability on population dynamics by modifying an existing model of caterpillar-disease dynamics to include uncorrelated stochastic variation in transmission rates between seasons in a manner that mimics the relationship between variability in spring weather and viral infection. Including stochasticity in transmission creates cyclic dynamics over a broader range of parameter space than in the deterministic model. Furthermore, simulated patterns of host density are more realistic when stochasticity in transmission rates is included in this model with outbreaks that vary among years in timing, amplitude and duration. Western tent caterpillars hatch in the early spring when temperatures are cool and variable. They compensate for suboptimal air temperatures by basking in the sun. To determine the impact of basking behaviour on the disease ecology of western tent caterpillars I examined the environmental determinants of larval body temperature and the effects of temperature on larval susceptibility to NPV. In the field, larval body temperature was determined by ambient temperature, solar irradiance and larval stage. The pathogenicity (LD5o) of NPV was not influenced by temperature but the incubation time of the virus declined asymptotically as temperature increased. Basking behaviour of larvae and the associated increase in body temperature can strongly influence growth and development but has little impact on infection by NPV. Thus, an effect of temperature on the pathogenicity of NPV is not the mechanism by which disease prevalence is higher during sunnier years. Wanner temperatures attained by basking may increase the number of infection cycles in sunny springs but do not protect larvae from viral infection. iv TABLE OF CONTENTS ABSTRACT ii TABLE OF CONTENTS iv LIST OF TABLES • v LIST OF FIGURES vi ACKNOWLEDGEMENTS viii DEDICATION ix 1. GENERAL INTRODUCTION 1 1.1 Introduction 1 1.2 Literature Cited 3 2. THE EFFECT OF RANDOM VARIATION IN TRANSMISSION RATES ON DISEASE DYNAMICS: A POTENTIAL ROLE FOR WEATHER IN POPULATION CYCLES 5 2.1 Introduction 5 2.2 Materials and Methods 8 2.3 Results 12 2.4 Discussion 13 2.5 Literature Cited 16 3. THERMAL ECOLOGY OF WESTERN TENT CATERPILLARS AND THEIR INTERACTION WITH NUCLEOPOLYHEDROVIRUS 26 3.1 Introduction 26 3.2 Materials and Methods 27 3.3 Results 33 3.4 Discussion 34 3.5 Literature Cited 37 4. GENERAL CONCLUSIONS 49 V LIST OF TABLES Table 2.1. A N O V A summary of multiple regression model of disease prevalence as a function of sunlight hours during the month of May and population density. 20 Table 3.1. A N O V A summary for model of environmental factors influencing WTC larval body temperature. 42 vi LIST OF FIGURES Figure 2.1. Population density and NPV prevalence at four populations of western tent caterpillars in southwestern British Columbia between 1975 and 1998. 21 Figure 2.2. The relationship between sunlight and NPV prevalence for twelve years of data on four populations of WTC in British Columbia and between relative transmission and the fraction of the population infected during the second disease cycle in the stochastic model. 22 Figure 2.3. The relationship between population density and NPV prevalence at four populations of WTC in British Columbia and between population density and the fraction of the population infected during the second disease cycle in the stochastic model. 23 Figure 2.4. Long term dynamics of the Dwyer et al. (2000) model with heterogeneity of 0.86, heterogeneity increased (C=l.03) and heterogeneity increased (C=l .03) and stochastic variation in transmission rates incorporated in the model (s=0.9). 24 Figure 2.5. The relationship between the period of outbreaks or the autocorrelation coefficient of logi 0 N t and logioNt+period, and heterogeneity in host susceptibility for the deterministic model and the stochastic model. . 25 Figure 3.1. Predicted larval body temperature as a function of ambient temperature. 43 Figure 3.2. Larval growth and development rates as a function of temperature. 44 Figure 3.3. Mean pupal weights for larvae reared at 7 different temperatures. 45 Figure 3.4. LD 5 0 ' s of larvae inoculated with NPV at seven temperatures. 46 Figure 3.5. Mean time to death for larvae dying of NPV infection as a function of rearing temperature. 47 Figure 3.6. Theoretical virus yield as a function of temperature. 48 ACKNOWLEDGEMENTS I am grateful to my supervisor, Judy Myers, for her logistic, intellectual and moral support. It is an honor to contribute to the great body of work that she and her students have accomplished. The members of my committee, Lee Gass, Sally Otto, Bernard Roitberg and Murray Isman, have improved my work tremendously through their advice and comments. I hope I have done justice to their great effort. I thank Michael Doebeli, Rina Freed, and Robert Lessard for discussions and feedback on modeling. Rina patiently helped me through the earlier stages of programming and her encouragement and friendship throughout this project have kept me sane and happy. Greg Dwyer provided helpful comments on an earlier version of chapter one. Dawn Cooper was a constant source of advice on working with viruses in the lab. Jenny Cory gave helpful comments on experimental design. Ilia Hartasanchez, Jessamyn Manson and Jessica Ware were,very helpful with their assistance in the laboratory. Matt Thomas planted the seed for this project during a discussion back in 1999. This project was funded by scholarships from the Natural Science and Engineering Research Council of Canada (NSERC), the Entomological Society of Canada, and the Entomological Society of British Columbia, to L. Frid and by an NSERC operating grant to J. Myers. To my family Alejandro and Dianna for inspiration and perspective Esther and Samuel, The Frid Foundation for Art and Science X As they near maturity, tent caterpillars often disperse throughout the treetops, and, during outbreaks, they collectively pass huge quantities of faecal pellets that create an audible plinking as they rain down from the canopy (T.D. Fitzgerald). Love the organisms for themselves first, then strain for general explanations, and, with good fortune, discoveries will follow. If they don't, the love and the pleasure will have been enough (E. O. Wilson). NPV could be more of a pest than the insects it infects, particularly if disease contributes to the reduction of host populations below densities required to sustain other natural enemy populations (L. D. Rothman). I sympathize with you, but do not find it possible to meddle in the private life of an insect when science does not require it (Nabokov). 1 1. GENERAL INTRODUCTION 1.1 Introduction Population cycles have captured the curiosity and imagination of ecologists since the early 1900's (Elton 1927). A good example is the fluctuation of western tent caterpillar populations in S.W. British Columbia (Wellington 1965; Myers 1988; Myers 2000). Western tent caterpillar population dynamics were studied in detail between the sixties and seventies by Wellington (1960; 1964; 1965) and since the mid seventies by Myers (2000). As with other studies of cycling populations, investigations of western tent caterpillar population cycles ask two questions: What are the factors that cause population density to change and why do outbreaks occur periodically? In his treatise on animal ecology, Elton (1927) suggests that in animal populations, "The chief cause of fluctuating numbers is instability in the environment." However, many ecologists have questioned how seemingly regular fluctuations could be caused by random environmental variation. These ecologists have used deterministic models, observations and experiments to argue that there must be some biotic component to periodic population fluctuations (Dwyer et al. 2000). Long-term population data suggest that viral disease is one of the biotic factors driving population cycles in western tent caterpillars; but that its importance varies among years and populations (Kukan and Myers 1999; Myers 2000). Based on observations that sunlight plays an important role in the ability of tent caterpillars to thermoregulate (Knapp and Casey 1986; Casey et al. 1988; Joos et al. 1988), that weather may determine the success of tent caterpillar colonies (Wellington 1964), that population fluctuations of forest Lepidoptera appear to be synchronized by weather (Myers 1998) 2 and that temperature plays a role in insect-pathogen interactions (Caruthers et al. 1992; Ribeiro and Pavan 1994; Fialho and Schall 1995; Thomas and Jenkins 1997; Blanford et al. 1998), I hypothesize that changing weather patterns may also account for some of the variability observed in the disease dynamics of western tent caterpillars. In chapter one I examine long-term data on disease prevalence and population density with respect to two weather variables, temperature and sunlight. I also present a model that incorporates stochastic variation in transmission rates in a manner that might be caused by climatic variation. Using these techniques I ask two questions: (1) is there any evidence that climate plays a role in the disease dynamics of western tent caterpillars? (2) If so, what are the implications for long-term population dynamics? In chapter two I examine the potential mechanisms that might lead weather to influence disease dynamics. I present the results of field observations of tent caterpillar thermoregulatory behaviour and laboratory experiments on the effects of temperature on tent caterpillar growth, development, susceptibility to disease and thermoregulatory behaviour. Using these techniques I ask the following questions: (1) what environmental variables influence tent caterpillar body temperature? (2) What are the effects of temperature on tent caterpillar fitness both in the presence and absence of viral disease? (3) Do tent caterpillars respond behaviourally to infection such that they increase or reduce their probability of survival? Further research is needed to determine the mechanisms behind the effect of weather on disease dynamics. However, this work suggests that climatic variation, even when it has no periodic component, may play a significant role in shaping the population fluctuations of western tent caterpillars and other Lepidoptera. 1.2 Literature Cited Blanford, S., M . B. Thomas and J. Langewald, 1998. Behavioural fever in the senegalese grasshopper, Oedaleus senegalensis, and its implications for biological control using pathogens. Ecological Entomology, 23: 9-14. Caruthers, R. I., T. S. Larkin, H. Firstencel and Z. Feng, 1992. Influence of thermal ecology on the mycosis of a rangeland grasshopper. Ecology, 73: 190-204. Casey, T. M . , B. Joos, T. D. Fitzgerald, M . E. Yurlina and P. A. Young, 1988. Synchronized group foraging, thermoregulation, and growth of eastern tent caterpillars in relation to microclimate. Physiological Zoology, 61: 372-377. Dwyer, G., J. Dushoff, J. S. Elkinton and S. A. Levin, 2000. Pathogen-driven outbreaks in forest defoliators revisited: Building models from experimental data. The American Naturalist, 156: 105-120. Elton, C , 1927. Animal Ecology. Sidgwick Jackson, London. Fialho, R. F. and J. J. Schall, 1995. Themal ecology of a malarial parasite and its insect vector: consequences for parasite's transmission success. Journal of Animal Ecology, 64: 553-562. Joos, B., T. M . Casey, T. D. Fitzgerald and W. A. Buttemer, 1988. Roles of the tent in behavioral thermoregulation of eastern tent caterpillars. Ecology, 69: 2004-2011. Knapp, R. and T. M . Casey, 1986. Thermal ecology, behaviour and growth of Gypsy Moth and Eastern Tent Caterpillars. Ecology, 67: 599-608. Kukan, B. and J. H. Myers, 1999. Dynamics of viral disease and population fluctuations in Western Tent Caterpillars (Lepidoptera: Lasiocampidae) in Southwestern British Columbia, Canada. Environmental Entomology, 28: 44-52. Myers, J. H., 1988. Can a general hypothesis explain population cycles of forest Lepidoptera? Advances in Ecological Research, 18: 179-242. Myers, J. H., 1998. Synchrony in outbreaks of forest Lepidoptera: a possible example of the Moran effect. Ecology, 79: 1111-1117. Myers, J. H. , 2000. Population fluctuations of the western tent caterpillar in southwestern British Columbia. Population Ecology, 42: 231-241. Ribeiro, H. C. T. and O. H. O. Pavan, 1994. Effect of temperature on the development of baculoviruses. The Journal of Applied Entomology, 118: 316-320. Thomas, M . B. and N . E. Jenkins, 1997. Efffects of temperature on growth of Metarhyzium flavoviridae and virulence to the variegated grasshopper, Zonocerus variegatus. Mycological Research, 101: 1469-1474. Wellington, W., 1960. Qualitative changes in natural populations during changes in abundance. Canadian Journal of Zoology, 38: 289-314. Wellington, W., 1964. Qualitative changes in populations in unstable environments. Canadian Entomologist, 96: 436-451. Wellington, W., 1965. An approach to a problem in population dynamics. Quaest Entomol, 1: 175-185. 5 2. THE EFFECT OF RANDOM VARIATION IN TRANSMISSION RATES ON DISEASE DYNAMICS: A POTENTIAL ROLE FOR WEATHER IN POPULATION CYCLES 1 2.1 Introduction Despite many studies on the population dynamics of forest Lepidoptera, researchers have failed to discover a dominant cause for periodic outbreaks and declines (Hunter and Dwyer 1998). Cyclic population dynamics requires a delayed density-related suppression of population growth following the initial population decline (Hutchinson 1948). This delayed recovery could be caused by a variety of factors individually or in combination. Potential causes for cycles include: predation, parasitization, maternal effects, pathogens, host plant defenses, and weather (Myers 1988). Biologists studying outbreaking Lepidoptera need a way to distinguish among these mechanisms. Beginning with the work of Lotka (1925) and Volterra (1926), mathematical models have been used to study the role of various agents that generate population cycles (reviewed by Myers 1988). Even though the authors consider different influences on population dynamics, they all incorporate a delayed density-related factor and often obtain similar predictions (Hunter and Dwyer 1998). This similarity between the predictions of models makes it difficult to use them to distinguish among proposed mechanisms. To resolve the conundrum of competing hypotheses, Myers (1988) and Hunter and Dwyer (1998) have called for studies that combine an empirically derived ' In this chapter "we" refers to my supervisor, Dr. Judy Myers, and myself. I conducted the analysis, modelling and writing with guidance and feedback from Dr. Myers. 6 understanding of underlying mechanisms with a theoretical framework. Models alone are insufficient. Dwyer et al. (1997, 2000) successfully combined mechanistic field experimentation and mathematical modeling to understand how heterogeneity in host susceptibility to disease might affect the population cycles of gypsy moths, Lymantria dispar. Heterogeneity in gypsy moth susceptibility to nuclear polyhedrovirus (NPV) results in nonlinear transmission rates with respect to pathogen density. Considering heterogeneity in host susceptibility, seasonality and the delay between host infection and death, the Dwyer et al. (2000) model produces cyclic population dynamics under conditions of moderate host heterogeneity in susceptibility to the virus. However, observed levels of heterogeneity in susceptibility span a range that could result in either cycling or a stable equilibrium. By considering disease in the absence of other factors, the model may have omitted other important processes. A common pitfall in studies of the causes of population cycles is that they often consider only one factor. Single factor hypotheses exclude the possibility of interactions among factors (Chamberlain 1897; Hilborn and Stearns 1982). Various authors have called for research that considers the interaction of multiple factors in driving population cycles of forest Lepidoptera (Bowers et al. 1993; Hunter and Dwyer 1998; Dwyer et al. 2000). In this study we modify the Dwyer et al. (2000) model to consider how two factors, stochastic variability in weather and a viral disease, may interact to cause periodic outbreaks in populations of western tent caterpillars, Malacosoma californicum pluviale (Dyar). 7 The biology of western tent caterpillar-NPV interactions is similar to that of gypsy moths as described by Dwyer et al. (2000). Caterpillars have one generation per year and hatch in the spring in synchrony with bud break. Larvae feed for 6-8 weeks on a variety of hosts including red alder, Alnus rubra; hawthorn, Crataegus monogyna; and crab apple, Malus diversifolia, and then pupate. Adults emerge after two weeks and mate. Females lay between 100 and 250 eggs in a single egg mass. Larvae develop in the egg and diapause until the following spring (Kukan and Myers 1999). Larvae may become infected with NPV when hatching from contaminated egg masses or while feeding on contaminated foliage. After 10-14 days, the virus kills infected caterpillars; the epidermis ruptures and virus occlusion bodies are released into the environment. Western tent caterpillars have cyclic population dynamics in southwestern British Columbia (Myers 1993, 2000). NPV plays an important role in these populations, but the relative importance of this disease varies between years and populations (Kukan and Myers 1999; Myers 2000, see Figure 2.1). This suggests that disease transmission rates may be influenced by factors other than host and pathogen densities. These factors could occasionally decouple caterpillar density and NPV transmission and result in lower or higher levels of infection than predicted by population densities alone. One of the factors that may influence NPV transmission rates is temperature. Temperature can influence the infection of caterpillars by NPV (reviewed in Benz 1987). Both the mortality rates of hosts and the production rates of NPV can be influenced by temperature (Ribeiro and Pavan 1994; Reichenbach 1985). Temperature can trigger or halt the expression of latent virus (Mohamed et al. 1985). 8 Another mechanism by which higher body temperatures could lead to higher transmission of NPV is by increasing movement and feeding rates. Casey (1976) found that feeding rates in two desert caterpillars (Sphingidae) increased with body temperature. Higher feeding and movement rates could lead to a higher probability of ingesting virus and thus higher transmission rates (Beisner and Myers 1999). To explore the influence of weather on the disease dynamics of western tent caterpillars, we examine long term trends between weather and disease prevalence in four field populations of western tent caterpillars in southwestern British Columbia, and modify the Dwyer et al. (2000) model accordingly to include uncorrelated stochastic variation in transmission rates between seasons. 2.2 Materials and Methods 2.2.1 Long Term Field Observations: To determine how disease prevalence in a given year was related to host population density and environmental variation, we used data from four populations of western tent caterpillars collected over 12 years in southwestern British Columbia, Canada. Caterpillars were collected by Kukan and Myers (1999) from these four populations and brought back to the laboratory where they were monitored for NPV using a DNA-dot blott hybridization assay or by rearing them for 7 days and observing the number of virus mortalities. The proportion of late stadium field-collected caterpillars carrying NPV is an index of disease prevalence in the field (Kukan and Myers 1999, Figure 2.1). We evaluated, by multiple regression, the linear dependence of yearly average disease prevalence on hours of sunlight in May at Vancouver International Airport (YVR), mean temperature for the month of May at Y V R , the number of tents counted at each site that year (an index of population density) and all two way interactions between the above variables. May is the month of peak larval development. The Environment Canada weather station is 5 to 50 km from the populations surveyed so we assume that weather conditions at Vancouver International Airport reflect those at the four populations surveyed. We selected the best regression model by backwards elimination of the variable with the least significance until all variables in the model were significant (a=0.05). To meet the assumptions of normality and homogeneity of variance we transformed population density (logio) and disease prevalence (arcsin(x)'/2). We tested for periodicity in 37 years of sunlight data (1961-1998) using Fisher's kappa statistic, which tests the null hypothesis that values are drawn randomly from a normal distribution against the alternative hypothesis that there is a periodic component in the data (SAS 2000). 2.2.2 The Disease Dynamics Model: Our model is based on a simple modification to the gypsy moth, NPV dynamics model published by Dwyer et al. (2000), which consists of partial differential equations for within-season disease dynamics and an iterative model for between-generation dynamics. Within a season we use Dwyer et al.'s (1997) ordinary differential equation approximation to the "short epidemic" model: dS(t) = -m{t)P{t)S{t) dt C1) 10 dm{t) =-C2 P{t)m2 {t) dt (2) dl(t) = m(t)P(t)S(t) - m(t - T)P(t - T)S(t - T) dt (3) dP(t) = m(t - T)P(t - T)S(t -T)- iiP(t) dt (4) Where S is the density of susceptible individuals in a population; m is its mean transmission rate (the probability that a caterpillar will encounter, ingest and die from the virus); / is the density of infected individuals in the population; and P is the density of pathogen (cadavers) in the population. The parameters are C, the coefficient of variation of the distribution of host susceptibility (heterogeneity in susceptibility); T, the incubation period of the virus, which is the time required for a lethally infected caterpillar to die; and |J., the decay rate of the pathogen over time. We modified the within season model to incorporate year to year variation in weather, which may be important for NPV dynamics (see below). At the beginning of each season we created a new variable, relative transmission (e). We generated this variable by raising e to the power of a random, normally distributed number with a mean of 0 and a standard deviation of s. Thus, e was random, log-normal and uncorrelated over time with a mean of 1. Each season, the initial transmission rate in the model was linearly related to relative transmission. m o = em (5) where m is the average initial mean transmission rate over all seasons. 11 We simulated this model numerically in Microsoft Visual Basic using Euler's method for 150 days each season. At the end of the season we calculated the fraction of the population that had become infected using: So S final F = ^ (6) To describe changes in population density from one season to the next we used the Dwyer et al. (2000) model: Nr+l=WT(l~F) (7) Z T + 1 = 0 V T F + ^ T ( 8 ) where N and Z are the densities of caterpillars and pathogen at the beginning of the season each year (N=S0 and Z=P0), F is the fraction of caterpillars infected that year, A, is the fitness of the surviving caterpillars, $ is a complex parameter that determines both pathogen survival from year to year and the relative susceptibility of hatchlings, and y determines the year to year survival of pathogens that are not ingested. Pathogen (Z or P) is measured in units of cadavers. While Dwyer et al. (2000) conducted an extensive exploration of parameter space in their deterministic model; we focus on simulations at different levels of heterogeneity in host susceptibility to the virus. Heterogeneity is of particular interest because empirical estimates of this parameter fail to reject either hypothesis of stable equilibrium or population cycles (Dwyer et al. 2000). Since gypsy moth biology is so similar to that of western tent caterpillars (univoltine generations, nine year periodicity in population 12 cycles and batch oviposition), we held other parameters fixed at the levels estimated by Dwyer et al. (2000) to be most realistic for gypsy moths. We analyzed the model output to determine cycle period and autocorrelation coefficients using the JMPIN (SAS 2000) time series analysis platform. At each level of heterogeneity we ran the stochastic model 10 times until the behavior of the model remained constant (i.e. no longer damping). We determined the period as the lag that had the highest, statistically significant autocorrelation coefficient. 2.3 Results 2.3.1 Long Term Field Observations: The pattern of host density and NPV prevalence over 23 years at four populations of western tent caterpillars is shown in Figure 2.1 (Kukan and Myers 1999). In these four populations we found a significant positive relationship between population density and the proportion of caterpillars collected from the field that were infected with NPV (prevalence). We also found a significant positive relationship between the hours of sunlight in May, when the caterpillar larvae were active and developing rapidly, and prevalence (Table 1, Figures 2.2a and 2.3a). Similar relationships between relative transmission rates, population density and NPV prevalence at the end of the second disease cycle are observed in the stochastic model (Figures 2.2b and 2.3b). There was no significant effect on prevalence by either temperature or any two-way interactions between variables (P>0.05). There was no significant periodic component in 37 years of sunlight data at Y V R (Kappa = 2.70, P=0.81). 13 2.3.2 The Disease Dynamics Model: The Dwyer et al. (2000) model demonstrates cyclic dynamics when heterogeneity in susceptibility (C) is sufficiently low. Adding uncorrected stochastic variation in transmission rates to the model broadens the range of heterogeneity over which cycles are observed. In particular, under conditions that would normally be stabilizing, the stochastic perturbation to transmission allows for the persistence of cycles (Figure 2.4). Below a threshold of heterogeneity in susceptibility (C), cycles occur in the absence of stochastic variation in transmission rates (Dwyer et al. 2000). However, above this threshold, the dynamics of the system approach equilibrium unless stochastic variation in transmission rates is included (Figure 2.5). As heterogeneity is increased further, the dynamics eventually become stable, even with stochasticity (Figure 2.5). 2.4 Discussion While studies of insect population dynamics have focused on single factor causes of periodic outbreaks, various authors have called for a more comprehensive investigation of the interaction among multiple factors driving population cycles (Bowers et al. 1993; Hunter and Dwyer 1998; Dwyer et al. 2000). Here we consider the interaction between stochastic variation in transmission, as might be caused by variation in weather, and host-pathogen dynamics. Using numerical simulations of a mathematical model and long-term field observations of disease dynamics, we show that stochastic variation in transmission from one year to the next can contribute to the maintenance of cyclic dynamics in a system that would otherwise go to equilibrium. The climatic variable of interest is sunlight, which increases larval body temperatures (Casey 1976; Knapp and 14 Casey 1986; Casey et al. 1988; Fitzgerald 1995; Frid and Myers, unpublished data) and may influence the virulence of NPV and its rate of transmission. Incorporating between-season stochastic variation in transmission rates into the Dwyer et al. (2000) model broadens the range of parameter space over which cycles are observed; specifically it maintains cycles at levels of caterpillar heterogeneity in susceptibility to NPV that would otherwise be stabilizing. In a model published by Berryman (1986) a stochastic weather variable had a similar effect and allowed for cycles in a density-dependent model of blackheaded budworm dynamics that otherwise resulted in oscillations damping to equilibrium. This finding is significant, since the Dwyer et al. (2000) model exhibits cycles only when heterogeneity in susceptibility is moderately low. Confidence intervals of heterogeneity obtained from gypsy moth laboratory and field data (0.002 to 4.94, Dwyer et al. 2000) span a range that would result in cycles or equilibrium. Heterogeneity in host susceptibility to the virus determines whether rates of infection are best predicted by the stochastic variation in transmission or by cyclical host pathogen dynamics. At high levels of heterogeneity, autocorrelation in caterpillar density (N) is low and rates of infection are mostly explained by random variation in transmission. In contrast, autocorrelation in N is high at low levels of heterogeneity, and random variation in transmission has almost no influence on disease dynamics. Our data on the long-term prevalence of the disease suggest the answer is somewhere in between, with variation in hours of sunshine explaining some of the variation in disease prevalence over time (Figure 2.2), but with populations continuing to demonstrate cyclical dynamics (Kukan and Myers 1999) and density dependent prevalence (Figure 2.3). While our data are only correlational, they do suggest that sunlight, or a correlate of sunlight, is as 15 important as population density for the transmission of the virus. Further experiments are necessary to test this hypothesis. Sunlight may be important in other Lepidopteran-NPV interactions. Miyashita (1964) noted that periodic outbreaks of the oriental tussock moth, Euproctis flava, typically follow years of lower than average sunlight in the spring. Although there is no information on disease prevalence in E. flava, it is possible that lower transmission of NPV at lower body temperatures may also be involved. Myers (1998) notes that outbreaks of forest Lepidoptera across the northern hemisphere are often associated with cool springs and unusually high precipitation. This is consistent with our hypothesis of lower levels of irradiance leading to lower transmission rates of virus and thus resulting in greater caterpillar population growth. In other insect orders, weather also plays a key role in disease dynamics. In the case of grasshoppers, elevated body temperatures actually reduce the virulence of fungal pathogens (Caruthers et al. 1992; Thomas and Jenkins 1997). Grasshoppers may even exhibit behavioral fevers when infected (Blanford et al. 1998). Models incorporating environmental variables such as sunlight and temperature have been found to accurately predict the occurrence of epizootics of fungal diseases in grasshoppers (Caruthers et al. 1992). In the gypsy moth-NPV system, which is the basis for the Dwyer et al. (2000) model, it is unlikely that sunlight per se plays an important role in disease dynamics. Unlike tent caterpillars, gypsy moths are thermal conformers, and their body temperature tends to track ambient temperature regardless of sunlight (Knapp and Casey 1986). 16 However other stochastic perturbations, such as the density of small mammal predators (Elkinton et al. 1996), may have a similar effect. Stochastic perturbations have been traditionally viewed as simply blurring otherwise deterministic dynamics (Blarer and Doebeli 1999) or as a force synchronizing separate populations (Myers 1998). In a model of temperature effects on a diapausing insect, Blarer and Doebeli (1999) show that environmental noise can also result in periodic outbreaks that are qualitatively different from the behavior of a purely deterministic model. Our model similarly results in periodic cycles under conditions that would normally lead to equilibrium, and in outbreaks that vary among years in timing, amplitude and duration. Studies such as these suggest that stochasticity may play an important role in driving population cycles. 2.5 Literature Cited Beisner, B. E. and J. H. Myers. 1999. Population densities and the transmission of virus in experimental populations of the western tent caterpillar (Lepidoptera: Lasiocampidae). Environmental Entomology 28(6): 1107-1113. Benz, G. 1987. Environment. Pages 177-214 in J. R. Fuxa and Y . Tanada, eds. Epizootiology of insect diseases. Wiley, New York. Berryman, A. A . 1986. On the dynamics of blackheaded budworm populations. Canadian Entomologist 118: 775-779. Blanford, S., M . B. Thomas and J. Langewald. 1998. Behavioural fever in the senegalese grasshopper, Oedaleus senegalensis, and its implications for biological control using pathogens. Ecological Entomology 23: 9-14. Blarer, A . and M . Doebeli. 1999. Resonance effects and outbreaks in ecological time series. Ecology Letters 2(3): 167-177. Bowers, R. G., M . Begon and D. E. Hodgkinson. 1993. Host pathogen population cycles in forest insects? Lessons from simple models reconsidered. OIKOS 67: 529-538. Caruthers, R. I., T. S. Larkin, H. Firstencel and Z. Feng. 1992. Influence of thermal ecology on the mycosis of a rangeland grasshopper. Ecology 73(1): 190-204. Casey, T. M . 1976. Activity patterns, body temperature and thermal ecology in two desert caterpillars (Lepidoptera: Sphingidae). Ecology 57: 485-497. Casey, T. M . , B. Joos, T. D. Fitzgerald, M . E. Yurlina and P. A. Young. 1988. Synchronized group foraging, thermoregulation, and growth of eastern tent caterpillars in relation to microclimate. Physiological Zoology 61(4): 372-377. Chamberlain, T. C. 1897. The method of multiple working hypotheses. Journal of Geology 5: 837-848. Dwyer, G., J. Dushoff, J. S. Elkinton and S. A. Levin. 2000. Pathogen-driven outbreaks in forest defoliators revisited: Building models from experimental data. The American Naturalist 156(2): 105-120. Dwyer, G., J. S. Elkinton and J. P. Buonaccorsi. 1997. Host heterogeneity in susceptibility and disease dynamics: Tests of a mathematical model. The American Naturalist 150(6): 685-707. Elkinton, J. S., W. M . Healy, J. P. Buonaccorsi, G. H. Boettner, A. M . Hazzard, H. R. Smith and A . M . Leibhold. 1996. Interactions among gypsy moths, white-footed mice and acorns. Ecology 77: 2332-2342. Fitzgerald, T. D. 1995. The Tent Caterpillars. Cornell University Press, Ithaca and London. Hilborn, R. and S. C. Stearns. 1982. On inference in ecology and evolutionary biology: the problem of multiple causes. Acta Biotheoretica 31: 145-164. Hunter, A. F. and G. Dwyer. 1998. Outbreaks and Interacting Factors: Insect population explosions synthesized and dissected. Integrative Biology 1: 166-177. Hutchinson, G. E. 1948. Circular causal systems in ecology. Annals of the New York Academy of Sciences 50: 221-246. Knapp, R. and T. M . Casey. 1986. Thermal ecology, behaviour and growth of Gypsy Moth and Eastern Tent Caterpillars. Ecology 67(3): 599-608. Kukan, B. and J. H. Myers. 1999. Dynamics of viral disease and population fluctuations in Western Tent Caterpillars (Lepidoptera: Lasiocampidae) in Southwestern British Columbia, Canada. Environmental Entomology 28(1): 44-52. Lotka, A. J. 1925. Elements of Physical Biology. Reprinted in 1956 by Dover Publications, New York. Miyashita, K. 1964. A note on the relationship between outbreaks of the oriental tussock moth Euproctis flava and weather conditions. Research in Population Ecology 6: 37-42. Mohamed, M . A., H. C. Coppel and J. D. Podgwaite. 1985. Temperature and crowding effects on virus manifestation on Neodiprion sertifer (Hymenoptera: Diprionidae) Larvae. The Great Lakes Entomologist 18(3): 115-118. Myers, J. H. 1988. Can a general hypothesis explain population cycles of forest Lepidoptera? Advances in Ecological Research 18: 179-242. Myers, J. H. 1993. Population outbreaks in forest Lepidoptera. American Scientist 81: 240-251. Myers, J. H. 1998. Synchrony in outbreaks of forest Lepidoptera: a possible example of the Moran effect. Ecology 79(3): 1111-1117. Myers, J. H. 2000. Population fluctuations of the western tent caterpillar in southwestern British Columbia. Population Ecology 42: 231-241. Reichenbach, N . G. 1985. Response of the western spruce budworm to temperature and dose of a virus, a growth regulator, and an organophosphate. Entomologia experimentalis et aplicata 38: 57-63. Ribeiro, H. C. T. and O. H. O. Pavan. 1994. Effect of temperature on the development of baculoviruses. The Journal of Applied Entomology 118: 316-320. SAS Institute Inc. 2000. JMPIN 4.02. Cary, NC, USA Thomas, M . B. and N . E. Jenkins. 1997. Efffects of temperature on growth of Metarhyzium flavoviridae and virulence to the variegated grasshopper, Zonocerus variegatus. Mycological Research 101((12)): 1469-1474. Volterra, V. 1926. Fluctuations in the abundance of a species considered mathematically. Nature 118: 558-560. 20 Table 2.1: A N O V A summary of multiple regression model of disease prevalence (arcsinVx ) as a function of sunlight hours during the month of May and population density (logio(number of tents)). Source DF Sum of Squares Mean Square F P r 2 Regression 2 1.18 0.59 8.57 0.0013 0.38 Error 28 1.93 0.069 Total 30 3.12 Effect Tests Sunlight 1 0.38 0.38 5.56 0.026 0.17 Logio(N) 1 0.38 0.38 5.51 0.026 0.16 21 i i 0 1 i ' I i 1975 1980 1985 1990 1995 1975 1980 1985 1990 1995 2000 Y e a r O o' 3 (D O (D Q . Figure 2.1: Population density (hollow circles) and NPV prevalence (solid circles) at four populations of western tent caterpillars ((a) Westham Island, (b) Galiano Island, (c) Mandarte Island and (d) Cypress Mountain) in southwestern British Columbia between 1975 and 1998. Based on data published by Kukan and Myers (1999). 22 1 "2 0.8 A Relative Transmission [Ln(e)] Figure 2.2: The relationship (a) between sunlight and NPV prevalence (±SE) for twelve years of data on four populations of WTC in British Columbia (Prevalence = -33.58 + 0.26 x hours of sun in May, r2=0.5, PO.01) and (b) between relative transmission and the fraction of the population infected during the second disease cycle each season in a two hundred year run of the stochastic model (k = 5.5, C = 1.03, s = 0.9, y= 0, m = 0.01, |X = 0.01, T= 10, ()>= 15, Prevalence = 0.21 + 0.078(fn e), r2=0.17, PO.01). The fraction infected was calculated as the sum of all new infections beginning at t = T and ending at t = 2T (the second disease cycle). 23 1 -r ~o £ 0.8 -o <u "£ 0.6 -Scaled Population Density ( log 1 0N) Figure 2.3: The relationship (a) between population density and NPV prevalence at four populations of WTC in British Columbia (Prevalence = -0.037 + 0.15 x log 1 0 N, r2=0.3, P<0.005) and (b) between scaled population density and the fraction of the population infected during the second disease cycle each season in a two hundred year run of the stochastic model (A, = 5.5, C = 1.03, s = 0.9, y= 0, m = 0.01, fi = 0.01, T = 10, <|> = 15,. Prevalence = 0.036 + 0.063xlogi0N, 1^=0.11, PO.001). The fraction infected was calculated as the sum of all new infections beginning at t = T and ending at t = 2T (the second disease cycle). 24 100 110 120 130 140 Time in Years Figure 2.4: Long term dynamics of the Dwyer et al. (2000) model with (a) parameter values of^ = 5.5, C=0.86, s = 0, <|>= 15,7=0, T= 10, m = 0.01, and p. = 0.01, (b) heterogeneity increased (C=1.03) and (c) heterogeneity increased (C=1.03) and stochastic variation in transmission rates incorporated in the model (s=0.9). Stochasticity can maintain cycles at higher levels of heterogeneity. Hollow circles indicate caterpillar density and solid circles indicate the fraction of the population infected. 25 10 3 6-I 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 Heterogeneity (C) Figure 2.5: The relationship between (a) the period of outbreaks (+SE) or (b) the autocorrelation coefficient of logioN, and logioN t + p e rj 0 ti (±SE), and heterogeneity in host susceptibility for the deterministic model (hollow circles) and the stochastic model (solid circles). A, = 5.5, s = 0.5, <]) = 15, y= 0, T = 10, m =0.01, and |4. = 0.01. Points beyond the dotted lines are not significantly different from zero. 26 3. THERMAL ECOLOGY OF WESTERN TENT CATERPILLARS AND THEIR INTERACTION WITH NUCLEOPOLYHEDROVIRUS 1 3.1 Introduction Western tent caterpillars undergo cyclical population dynamics with outbreaks occurring every 8-11 years (Myers 1993; Myers 2000), and nucleopolyhedrovirus (NPV) plays an important role in driving these outbreak dynamics (Kukan and Myers 1999). Various authors have noted that, in other species, larval susceptibility to NPV is sensitive to temperature (Stairs and Milligan 1979; Kobayashi et al. 1981; Johnson et al. 1982; Mohamed et al. 1985; Reichenbach 1985; Ribeiro and Pavan 1994). Long term population and climatic data suggest that temperature may also play an important role in the interaction between western tent caterpillars and NPV. There is a positive correlation between NPV prevalence and hours of sunlight during May, the month of peak larval development (Figure 2.2). Tent caterpillars thermoregulate by basking gregariously in the sun and thus increasing their body temperature (Knapp and Casey 1986). Temperature, in turn has an important influence on the physiology and ecology of insects (Heinrich 1981; Heinrich 1993). This influence extends to various important processes including: growth (Petersen et al. 2000), development (Gilbert and Raworth 1996) and susceptibility to pathogens (Caruthers et al. 1992; Ribeiro and Pavan 1994; Thomas and Jenkins 1997). Although increasing body temperature is likely to be beneficial to tent caterpillars in terms of ' In this chapter "we" refers to my supervisor, Dr. Judy Myers, and myself. I conducted the experiments, analysis and writing with guidance and feedback from Dr. Myers. 27 growth and development, higher temperatures could be detrimental if they promote the process of viral replication. Furthermore, by reducing variance in body temperature, thermoregulation may reduce variance between individuals in disease susceptibility. Heterogeneity in disease susceptibility has a stabilizing effect on disease dynamics (Dwyer et al. 1997; Dwyer et al. 2000). If certain environmental conditions, such as sunlight, are more conducive to thermoregulation, climate could influence the variance in larval susceptibility to disease and thus the stability of disease dynamics. Speed to kill (Johnson et al. 1982), pathogenicity and yield per host (Ribeiro and Pa van 1994), are three important viral parameters that may be influenced by temperature. To explore the potential influence of temperature on disease dynamics of western tent caterpillars and NPV, we reared larvae in the laboratory at seven constant temperatures ranging from 18 to 36° C. At each of these temperatures we measured the pathogenicity (LD5o) of NPV, its incubation time, and the growth and development rates of tent caterpillar larvae. We also estimated theoretical pathogen yield per host, based on speed to kill and larval growth rates. We relate our results to the temperature preference of control and infected larvae in a thermal gradient and to the influence of climatic variables on larval body temperature. 3.2 Materials and Methods Field observations on body temperatures of a total of 96 western tent caterpillar larvae were collected over eight occasions between May 12th and June 16th 2000 at three sites in the vicinity of Vancouver BC: Westham Island, the University of British Columbia Campus, and Cypress Mountain. To avoid pseudo-replication we measured only one larva per group. On each occasion we measured ambient temperature (Ta), 28 larval body temperature (Tb), irradiance, the developmental stage of the larva and the size of the group with which it was associated. We measured T a and Tb (° C) using a digital thermometer with a thermistor on a wire probe (Fisher Scientific). T a was measured in the open, two meters above the ground and 1 m away from the focal larva. Tb was measured by placing the thermistor against the dorsal side of the larva. Knapp and Casey (1986) found no difference between this technique and the more intrusive technique of stabbing the larva with a thermistor encased within a hypodermic needle. We measured irradiance (photon |imols m"2 s"') using a photometer (LI-COR, LI - 189) on the most direct path of natural light to the larva. We used a hand-held ruler to estimate the average diameter of a larval cluster and thus to estimate group size area (cm2). We used multiple linear regression with backwards elimination to determine which independent variables significantly influenced larval Tb. We ensured that the assumptions of homogeneous variance and normality of residuals were met. To determine the shape of the function between ambient temperature and body temperature we used the cubic spline with normal errors, a form of nonparametric regression that makes no apriori assumptions about the shape of the function (Schluter 1988). A l l cubic splines in this paper (see below) were generated using glmsWIN 1.0 (Schluter 2000). Standard errors and confidence intervals were estimated with 1000 bootstrap replications. Results for all cubic splines are reported with figure legends stating: N , the number of individual observations; X the smoothing parameter; and the effective number of parameters, which is a description of the complexity of the function (Schluter 2000). X was chosen in order to minimize the generalized cross validation score (GCV). This 29 value of X approximately minimizes the sum of squared deviations between the estimated and true functions (Schluter 1988). Larvae for all experiments were obtained from egg masses collected in February and March 2000 and 2001 at Westham Island or at the University of British Columbia campus. The egg masses were stored at 4° C for a maximum of 3 months. To eliminate any potential contamination with NPV, we washed egg masses in a 6% solution of sodium hypochlorite for approximately 2 minutes until most of the spumaline coat was dissolved (Fitzgerald 1995). They were then thoroughly rinsed with running water, set out to dry and left to hatch at room temperature in 300mL paper cups with plastic lids. The time interval between removal from the refrigerator and hatching was about 5 days. Before use we kept larvae in the laboratory at room temperature and supplied them daily with a diet of field collected red alder leaves. Leaves were washed with a 10% bleach solution and rinsed thoroughly with water. We kept leaves in floral water picks to maintain their freshness. Cups were cleaned of faecal pellets and dry or decaying leaves as needed. To measure growth and development rates we reared 20 larvae from 3 r d stadium to pupation at each of the following temperatures: 18, 21, 24, 27, 30, 33, and 36° C. Larvae were reared individually in 200mL cups and fed red alder leaves in water picks. They were weighed prior to being placed in the growth chambers, again 5 to 6 days later and at pupation. Growth rate for each individual larva was measured between the first two weightings as: weight final weight i n i t i a l gr = In \time - l [1] 30 Development rate was measured as the reciprocal of days between the third stadium and pupation. Individual growth, development and pupal weight curves were fit as functions of temperature using cubic splines with normal errors (Schluter 1988; Schluter 2000). We used sex as a categorical covariate for development rates and pupal weights. We did not use sex as a covariate for growth rate because not all larvae survived to pupation, the first stage at which they can be sexed. We measured larval temperature preference as indexed by body temperature on a radiant-light thermal gradient in the spring of 2000 and on a dark thermal gradient in the spring of 2001. The light gradient consisted of a cardboard platform (50x58cm) with a 60W light bulb in the center. We individually placed 18 larvae inoculated with a virus dose of 10000 occlusion bodies (OB's) and 17 control larvae on an arbitrary location on the gradient. Larvae were allowed to roam in the area for five minutes, after which we measured the larval body temperature (Tb) using a thermocouple connected to a digital thermometer. As with field measurements, the thermocouple was placed against the dorsal side of the larvae until the temperature stabilized (about 2 seconds). The dark thermal gradient consisted of a metal tray with a hot plate under one end and an ice pack under the other. The tray was covered with foil to block out light but with enough space to allow larvae to move. This design removes the confound of light preference inherent in the previous experiment. We individually placed 32 larvae inoculated with 10000 OBs of virus and 32 controls at the center of the gradient and left them for 10 minutes before measuring their Tb as described above. We tested for differences in the temperature preference of infected vs. uninfected larvae using t-tests. 31 To measure the effect of temperature on virus pathogenicity we used lethal dose 50 (LD50) analysis. We reared third stadium larvae at seven temperatures (18, 21, 24, 27, 30, 33, and 36° C). At each temperature we fed larvae one of seven doses of NPV: 0 (controls), 236, 1462, 2863, 5025, 11175, and 35188 OBs. These doses are based on quantifications of dilutions that were produced to approximately increase exponentially from one level to the next. We prepared virus dilutions from frozen larvae killed by virus in the lab the previous year. OB's were isolated through two series of centrifugations: slow (1000 rpm for 30 seconds) to remove non-virus particles, and fast (14 000 rpm for 20 minutes) to form the virus pellet. The pellet was resuspended and subsequent dilutions were quantified with a haemocytometer. We replicated each temperature and dose combination twice. A replicate consisted of 8-10 larvae from different egg masses. Larvae were infected upon reaching third stadium. Larvae were starved for 24 hrs before being fed 5u.L of dLL.0 with virus solution on a leaf disk. We did not use larvae that failed to eat more than 75% of their disk. We then monitored larvae for virus mortality and fed them daily a diet of red alder leaves in floral water picks. Due to the delay between host infection and death, virus mortality occurs in cycles. The first cycle is due to the experimental inoculation, whereas later cycles are due to infection from the release of infective OB's at the death of the initially infected larvae. We measured L D 5 0 ' s from the mortality that occurred during the first disease cycle. We defined the end of the first disease cycle as the day when mortality decreased to a minimum before increasing again to begin the second cycle. We transformed 32 indispensable mortality, (the difference between mortality at each dose and control mortality), in each replicate to probit values using: probit = In mortality 1- mortality We then regressed probit mortality against log 10 (dose). The L D 5 0 is the dose at which the predicted probit is 0 (i.e. mortality = survival). We calculated 95% confidence intervals around the predicted LD 5o for each temperature following the method in (Zar 1996): [2] yx [3] where x is dose, y is probit mortality, b is the slope of the relationship, t is to.os <2)v, Sb is the standard error of the slope, s y x is the standard error of the estimate, and n is the number of replicates. For each replicate we also calculated the mean time to death from viral infection (incubation time) during the first disease cycle. We fit a cubic spline with normal errors to virus incubation time as a function of temperature (Schluter 1988; Schluter 2000). We determined theoretical yield by using predicted growth rate (gr) and incubation time (td) in the following equation: y = e g r x t d [4] This assumes that infected larvae at each temperature grow at the same relative rate as healthy larvae, that virus volume is proportional to larval size and that it grows exponentially over time in each tissue. This measure may not be accurate if larvae die at the same virus load, regardless of temperature. 33 3.3 Results In the field all variables measured, except group size, significantly influenced larval body temperature (Table 1). As ambient temperature, irradiance and larval stadium increased so did larval body temperature. The effect of stadium is not confounded by time of the growing season as we observed various developmental stages during each excursion. Larval body temperature was as high as 21°C above ambient. There was a sigmoid-relationship between T b and T a (Figure 3.1). As would be expected for a behaviorally thermoregulating organism, Tb plateaus at higher ambient temperatures (Blanford et al. 1998; Blanford and Thomas 2000). Larval growth rate increased curvilinearly with temperature (Figure 3.2a). Development rates increased linearly between 18 and 30°C, increased at a lower rate to 33°C and then decreased at 36° C (Figure 3.2b). Once larvae pupated, females were larger than males (Figure 3.3). No females survived to pupation at 36° C. In general there were fewer females than males at pupation. We do not know what the primary sex ratio was but the higher number of males may have been caused by differential survival. Lower female survival may have been associated with their longer developmental time and thus prolonged exposure to uncontrolled mortality factors, such as bacterial infection (Figure 3.2b). Pupal weights as a function of temperature were described with a cubic spline with sex as a categorical covariate (Figure 3.3). The largest pupae occurred between 24 and 30° C. There was no significant difference in the temperature preference (mean ± S.E.) of infected (24.8 ± 1.3 °C) and uninfected (26.2 ± 1.3 °C) larvae in a "radiant light" 34 thermal gradient (t=0.80, DF = 33, P=0.43), or in the dark thermal gradient, (26.9 + 1.1 °C infected vs. 27.6 ± 1.1 °C uninfected, 1=0.41, DF = 62, P=0.69). There was no mortality during the first disease cycle at any temperature in the controls. We found no evidence that temperature influences the pathogenicity (LD50) of NPV (Figure 3.4), but it strongly influenced the incubation time of the virus (Figure 3.5). As temperature increased, time to death decreased asymptotically. Our estimates of theoretical virus yield per larva increase significantly between 18 and 21°C and then remain relatively constant decreasing slightly to 36° (Figure 3.6). 3.4 Discussion The body temperature of insects is a complex function of behaviour and the abiotic environment (Knapp and Casey 1986; Casey et al. 1988; Caruthers et al. 1992; Lactin and Johnson 1998). Body temperature in turn determines the outcome of various physiological and ecological processes (Heinrich 1981; Heinrich 1993). Insects capable of thermoregulating within certain environmental constraints must compromise among these processes (Knapp and Casey 1986). Our experiments demonstrate that both the development of western tent caterpillars and their interaction with NPV are sensitive to temperature. In turn, behaviour and the environment influence larval body temperature. Therefore, both insect thermoregulatory behaviour and the abiotic environment have implications for the outcome of population and disease dynamics. Growth and development rates in insects tend to increase linearly with temperature within the range of temperatures normally experienced in the field (Gilbert and Raworth 1996) and our results agree. The ratio of growth to development rates 35 determines adult size and maximum potential reproductive output. Small changes in this ratio can result in large changes to adult size of some insects (Gilbert and Raworth 1996). For western tent caterpillars the optimal temperatures for maximizing pupal size are between 24 and 33° C. Thermal sensitivity in insects can respond rapidly to natural selection (Huey et al. 1991). The relatively constant pupal size of western tent caterpillars between 24 and 33° C may indicate an adaptation to the highly variable environment in which western tent caterpillars develop. Tent caterpillars can exploit sunlight to thermoregulate (Knapp and Casey 1986; Casey et al. 1988; Joos et al. 1988) and the body temperature of western tent caterpillars is a positive function of irradiance (Table 1). By elevating their body temperature when it is sunny, tent caterpillars can reduce the time to pupation while still maintaining maximal pupal size. The body temperatures of non-basking larvae will be close to ambient temperatures, which for the months of April, May and June reach an average daily maximum of 12.7, 16.3, and 19.3°C (Environment Canada 2001). Therefore for much of the developmental period, larvae will be at suboptimal temperatures. Overcast conditions and associated colder temperatures will increase the developmental period but apparently will not greatly reduce pupal size. Unlike growth and development of larvae, the pathogenicity of NPV was insensitive to temperature within the range explored. This is inconsistent with various studies. High temperatures (30 and 35° C) induced latent NPV infections in Neodiprion setrifer larvae, whereas intermediate temperatures (25° C) did not (Mohamed et al. 1985). Temperatures above 38° C inactivated NPV in Galleria mellonella larvae (Stairs and Milligan 1979). Between 17 and 37° C, larval mortality from NPV infection in Diatraea 36 saccharalis increased with temperature (Ribeiro and Pavan 1994). In the silk worm, Bombyx mori, high temperatures (35° C) inhibited the development of NPV infections (Kobayashi et al. 1981). The L D 5 0 o f NPV to Choristoneura occidentalis larvae decreased with temperature between 15 and 30° (Reichenbach 1985). The lack of a temperature effect in this system may not be unusual. Reichenbach's (1985) confidence intervals for LD 5o's overlap at all temperatures. Other studies only found an effect at temperatures greater than 30° C, which are higher than are normally experienced by the host (Stairs and Milligan 1979; Kobayashi et al. 1981; Mohamed et al. 1985). In contrast, the pathogenicity of NPV to western tent caterpillars was insensitive to temperatures within the range that is normally experienced in the field. The insensitivity of NPV LDso's to temperature does not exclude the possibility that temperature may influence disease dynamics. Both the incubation time of the virus and its theoretical yield were sensitive to temperature variation, particularly over the range of 18-24° C, which is typical for the developmental period of larvae. The incubation time of the virus could determine how many secondary disease cycles occur within a season (Rothman 1997). Secondary disease cycles can infect a large proportion of the host population even when initial infection rates are low. In biological control with pathogens, secondary cycles can emulate the environmental persistence of chemical insecticides (Thomas et al. 1995). Shorter virus incubation times could also mean reduced damage caused by pests, depending on how feeding rates are influenced by temperature. Larvae did not respond to infection by changing their temperature preference as indexed by body temperature on a thermal gradient. This contrasts with the behaviour of 37 some grasshoppers infected with fungal disease (Blanford et al. 1998). In the case of western tent caterpillars, however, any change in temperature preference would not influence susceptibility (LD 5 0 ) to the virus. Further studies are necessary to determine the role of sunlight in the disease dynamics of western tent caterpillars. Based on our L D 5 0 data (Figure 3.4), it is unlikely that field observations of higher disease prevalence during sunnier years are caused by changes to virus pathogenicity. However, an increase in the number of disease cycles during sunnier years could influence the occurrence of epizootics of NPV in field populations of western tent caterpillars. Temperature and other factors influencing the development rates of Lepidopteran larvae and their pathogens may be important in determining the onset of disease epizootics 3.5 Literature Cited Blanford, S. and M . B. Thomas, 2000. Thermal behavior of two Acridid species: effects of habitat and season on body temperature and the potential impact on biocontrol with pathogens. Environmental Entomology, 29: 1060-1069. Blanford, S., M . B. Thomas and J. Langewald, 1998. Behavioural fever in the senegalese grasshopper, Oedaleus senegalensis, and its implications for biological control using pathogens. Ecological Entomology, 23: 9-14. Caruthers, R. I., T. S. Larkin, H. Firstencel and Z. Feng, 1992. Influence of thermal ecology on the mycosis of a rangeland grasshopper. Ecology, 73: 190-204. Casey, T. M . , B. Joos, T. D. Fitzgerald, M . E. Yurlina and P. A. Young, 1988. Synchronized group foraging, thermoregulation, and growth of eastern tent caterpillars in relation to microclimate. Physiological Zoology, 61: 372-377. Dwyer, G., J. Dushoff, J. S. Elkinton and S. A. Levin, 2000. Pathogen-driven outbreaks in forest defoliators revisited: Building models from experimental data. The American Naturalist, 156: 105-120. Dwyer, G., J. S. Elkinton and J. P. Buonaccorsi, 1997. Host heterogeneity in susceptibility and disease dynamics: Tests of a mathematical model. The American Naturalist, 150: 685-707. Environment-Canada, 2001. Canadian Climate Normals, http://www.msc-smc.ec.gc.ca/climate/climate_normals/show_stations_e.cfm?province=BC, Fitzgerald, T. D., 1995. The Tent Caterpillars. Cornell University Press, Ithaca and London. Gilbert, N . and D. A. Raworth, 1996. Insects and temperature - A general theory. The Canadian Entomologist, 128: 1-13. Heinrich, B., 1981. Ecological and Evolutionary Perspectives. Pages 235-302 in B. Heinrich (ed.). Insect Thermoregulation. John Wiley and Sons, New York. Heinrich, B., 1993. The Hot Blooded Insects. Harvard University Press, Cambridge, Massachusetts. Huey, R. B., L. Partridge and K. Fowler, 1991. Thermal sensitivity of Drosophila melanogaster responds rapidly to laboratory natural selection. Evolution, 45: 751-756. Johnson, D. W., D. B. Boucias, C. S. Barfield and G. E. Allen, 1982. A temperature dependent developmental model for a nuclearpolyhedrosis virus of the velvetbean caterpillar, Anticarsi gemmatalis (Lepidoptera: Noctuidae). Journal of Invertebrate Pathology, 40: 292-298. Joos, B., T. M . Casey, T. D. Fitzgerald and W. A. Buttemer, 1988. Roles of the tent in behavioral thermoregulation of eastern tent caterpillars. Ecology, 69: 2004-2011. Knapp, R. and T. M . Casey, 1986. Thermal ecology, behaviour and growth of Gypsy Moth and Eastern Tent Caterpillars. Ecology, 67: 599-608. Kobayashi, M . , S. Inagaki and S. Kawase, 1981. Effects of high temperature on the development of nuclear polyhedrosis virus in the silkworm, Bombyx mori. Journal of Invertebrate Pathology, 38: 386-394. Kukan, B. and J. H. Myers, 1999. Dynamics of viral disease and population fluctuations in Western Tent Caterpillars (Lepidoptera: Lasiocampidae) in Southwestern British Columbia, Canada. Environmental Entomology, 28: 44-52. Lactin, D. L. and D. L. Johnson, 1998. Environmental, physical and behavioural determinants of body temperature in grasshopper nymphs (Orthoptera: Acrididae). The Canadian Entomologist, 130: 551-577. Mohamed, M . A. , H. C, Coppel and J. D. Podgwaite, 1985. Temperature and crowding effects on virus manifestation on Neodiprion sertifer (Hymenoptera: Diprionidae) Larvae. The Great Lakes Entomologist, 18: 115-118. Myers, J. Ft., 1993. Population outbreaks in forest Lepidoptera. American Scientist, 81: 240-251. 40 Myers, J. H., 2000. Population fluctuations of the western tent caterpillar in southwestern British Columbia. Population Ecology, 42: 231-241. Petersen, C , H. A. Woods and J. G. Kingsolver, 2000. Stage-specific effects of temperature and dietary protein on growth and survival of Manduca sexta caterpillars. Physiological Entomology, 25: 35-40. Reichenbach, N . G., 1985. Response of the western spruce budworm to temperature and dose of a virus, a growth regulator, and an organophosphate. Entomologia experimentalis et aplicata, 38: 57-63. Ribeiro, H. C. T. and O. H. O. Pavan, 1994. Effect of temperature on the development of baculoviruses. The Journal of Applied Entomology, 118: 316-320. Rothman, L. D., 1997. Immediate and delayed effects of a viral pathogen and density on tent caterpillar performance. Ecology, 78: 1481-1493. Schluter, D., 1988. Estimating the form of natural selection on a quantitative trait. Evolution, 42: 849-861. Schluter, D. 2000. Estimating fitness functyions using the cubic spline, glmsWTN 1.0. http://www.zoology.ubc.ca/~schluter/splines.html. Stairs, G. R. and S. E. Milligan, 1979. Effects of heat on nonoccluded Nuclear Polyhedrosis Virus (Baculovirus) from Galleria mellonella Larvae. Environmental Entomology, 8: 756-759. Thomas, M . B. and N . E. Jenkins, 1997. Efffects of temperature on growth of Metarhyzium flavoviridae and virulence to the variegated grasshopper, Zonocerus variegatus. Mycological Research, 101: 1469-1474. Thomas, M . B., S. N . Wood and C. J. Lomer, 1995. Biological control of locusts and grasshoppers using a fungal pathogen: the importance of secondary cycling. Proceedings of the Royal Society of London, Series B, 259: 265-270. Zar, J. H., 1996. Biostatistical Analysis. Prentice Hall, Upper Saddle River, New Jersey. 42 Table 3.1: A N O V A summary for model of environmental factors influencing WTC larval body temperature (model r2=0.82). Source DF Sum of Squares Mean Square F Ratio P(F) Regression 3 4493.1277 1497.71 135.3090 <0.0001 Irradiance 1 772.9947 69.8354 <0.0001 Ta 1 1036.3937 93.6319 O.OOOl Stadium 1 144.9263 13.0932 0.0005 Error 92 1018.3306 11.07 Total 95 5511.4583 43 Figure 3.1: Predicted larval body temperature as a function of ambient temperature (+ 95% CI) fit with a cubic spline (N = 96, X = 1.4, effective number of parameters = 3.9). The solid straight line is where Tb = T a. 44 t 1 r 24 28 32 Temperature (C) t r 24 28 32 Temperature (C) Figure 3.2: Larval growth (a) and development (b) rates (mean ± 95% CI) fit with cubic splines as a function of temperature (growth: N = 136, A- = 3.1, effective number of parameters = 3.4; development: N = 89, A, = -0.6, effective number of parameters = 5.7). Hollow circles indicate male development rates and solid circles indicate female development rates. 45 0.6 0.1 H 1 1 1 1 1 ' 16 20 24 28 32 36 Temperature(C) Figure 3.3: Mean pupal weights (± SE) for larvae reared at 7 different temperatures (solid circles are for females and empty circles are for males). The solid line indicates the cubic spline fit to the data (± 95%CI) using sex as a categorical covariate (N = 89, X = 2.8, effective number of parameters = 3.2). 46 Figure 3.4: L D 5 0 ' s (± 95% CI) of larvae inoculated with NPV at seven temperatures. 47 0 4 1 1 1 1 r-16 20 24 28 32 36 Temperature (C) Figure 3.5: Cubic spline fit of mean time to death (+ 95% CI) for larvae dying of NPV infection as a function of rearing temperature (N = 82, X = 0, effective number of parameters 5.2). 48 12 0 H 1 1 1 1 r-16 20 24 28 32 36 Temperature (C) Figure 3.6: Theoretical virus yield (+ 95% CI) as a function of temperature, relative to larval weight at the time of infection. 49 4. GENERAL CONCLUSIONS It seems, therefore, that the food relationships of animals result in the numbers being controlled in the majority of cases by carnivorous enemies, but that when disturbances in the environment cause a sudden acceleration in the rate of increase of some smaller species, its enemies no longer act as an efficient control. Actually the numbers of an animal are ultimately very often controlled by organisms smaller than itself, by parasites, which produce epidemics Elton's words, written three quarters of a century ago, could not summarize my work more accurately. Both biotic and abiotic forces regulate animal populations, but changes in the environment can shift the relative importance of these forces. In the case of western tent caterpillars, observational data suggests that while disease prevalence is density dependent, sunlight may also influence disease dynamics. Including environmental stochasticity in transmission, in a way that may be caused by sunlight, into a model of disease dynamics broadens the range of parameter space over which cycles are observed. Random noise can lead to non-random, periodic population dynamics in host-virus populations that could otherwise remain constant. Sunlight is an important determinant of caterpillar body temperature so I examined the effects of temperature on the interaction between tent caterpillars and nucleopolyhedro virus (NPV). I found no evidence that temperature influences the pathogenicity of NPV; therefore I can reject this as a mechanism behind the observed pattern of higher disease prevalence during sunnier years. Temperature did, however, influence the growth and development rates of caterpillars, the speed to kill by virus and the theoretical yield of the virus. Future numerical and experimental studies should further explore the role that these factors play in disease and population dynamics. 1 Elton, C , 1927. Animal Ecology. Sidgwick Jackson, London. 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0099608/manifest

Comment

Related Items