TESTING PLASTICITY OF DEVELOPMENTAL STRATEGY IN THE AQUATIC DIPTERAN LARVAE OF CORETHRELLA APPENDICULATA By ERIK BLOSSER A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2014
2014 Erik Blosser
This dissertation is dedicated to my parents, Max and Alta, and to my fiance Irka, for their support through challenging times.
4 ACKNOWLEDGMENTS I would like to thank my dissertation committee, Phil Lounibos, Steven Juliano, Ken Dodd, Jane Brockmann, and George OMeara for constructive comments, advice and encouragement. I am especially grateful to my advisor, Phil Lounibos, for the chance to work on interesting topics and his reviews and constructive comments through the years. I am grateful to Barry Alto for discussions about statistics and to Steven Juliano for further instruction in statistics through his class. I would like to thank Naoya Nishimura for excellent advice on any and all problems and for assistance with experimental upkeep. Irka Bargielowski, Maria Carrasquilla and Cathy Westbrook were a great aid in performing experiments, colony maintenance and engaging in helpful discussions.
5 TABLE OF CONTENTS page ACKNOWLEDGMENTS .................................................................................................. 4 LIST OF TABLES ............................................................................................................ 7 LIST OF FIGURE S .......................................................................................................... 8 ABSTRACT ................................................................................................................... 10 CHAPTER 1 LITERATURE REVIEW AND BACKGROUND ....................................................... 12 Introduction ............................................................................................................. 12 Complex Life Cycles ............................................................................................... 12 Early Amphibian Models of Metamorphosis ............................................................ 15 Comparison of Amphibian and Arthropod Development ......................................... 16 Modeling Thresholds Based on Food Consumption ............................................... 18 Physiology of Insect Development and Diversity among Species ........................... 20 Development in Aquatic Dipterans .......................................................................... 22 Biology of Corethrella appendiculata ...................................................................... 25 2 TESTING DEVELOPMENTAL PLASTICITY IN AQUATIC LARVAE OF CORETHRELLA APPENDICULATA (DIPTERA: CORETHRELLIDAE) ................. 35 Introduction ............................................................................................................. 35 Materials and Methods ............................................................................................ 38 Model Selection ................................................................................................ 39 Reliability of Prey Recovery Technique ............................................................ 41 Plasticity in Egg Development .......................................................................... 42 Results .................................................................................................................... 43 Larval Development ......................................................................................... 43 Model Selection ................................................................................................ 44 Egg Clutch Size ................................................................................................ 45 Discussion .............................................................................................................. 45 Model Selection ................................................................................................ 45 Larval Development ......................................................................................... 47 Egg Develo pment ............................................................................................. 48 3 EFFECTS OF TEMPERATURE ON DEVELOPMENTAL TIMING AND THRESHOLDS ....................................................................................................... 55 Introduction ............................................................................................................. 55 Materials and Methods ............................................................................................ 58 Colony Maintenance ......................................................................................... 58
6 Early Instar Rearing .......................................................................................... 59 Results .................................................................................................................... 60 Larval Development ......................................................................................... 60 Prey Consumption ............................................................................................ 61 Discussion .............................................................................................................. 61 4 EFFECTS OF PENULTIMATE INSTAR ENVIRONMENT ON DEVELOPMENTAL TIMING AND ADULT FITNESS TRAITS OF CORETHRELLA APPENDICULATA ....................................................................... 71 Introduction ............................................................................................................. 71 Materials and Methods ............................................................................................ 73 Colony Maintenance ......................................................................................... 73 Early Instar Food Trials .................................................................................... 73 Experiment One ............................................................................................... 74 Early instar rearing ..................................................................................... 74 Third instar treatments ............................................................................... 75 Final instar treatments ............................................................................... 76 Experiment Two ............................................................................................... 77 Early instars ............................................................................................... 77 Final instar f ood switching method ............................................................. 78 Head capsule widths .................................................................................. 79 Results .................................................................................................................... 79 Experiment One ............................................................................................... 79 Fourth instar development ......................................................................... 79 Ad ult traits .................................................................................................. 80 Experiment Two ............................................................................................... 80 Head capsule size ...................................................................................... 81 Development time ...................................................................................... 81 Discussion .............................................................................................................. 81 5 CONCLUSIONS AND FUTURE STUDIES ............................................................. 97 Larval Development Strategy in Corethrella appendiculata .................................... 97 Future Studies ...................................................................................................... 102 Carry over E ffects from Early Instars ............................................................. 102 Group Development ....................................................................................... 103 Drying and Predation ...................................................................................... 105 Comparison among Dipteran Species ............................................................ 105 LIST OF REFER E NCES ............................................................................................. 109 BIOGRAPHICAL SKETCH .......................................................................................... 117
7 LIST OF TABLES Table page 2 1 C omparison of model regressions of fourth instar development time vs. inverse food consumption for female and male larvae of C. appendiculata. ....... 50 3 1 ANCOVA for the effects of temperature and sex on development time of fourth instar C. appendiculata larvae with food level as a covariate.. ................. 66 4 1 Test for equal slopes among regressions of development time versus inverse food consumed in fourth instar C. appendiculata larvae.. ................................... 96
8 LIST OF FIGURES Figure page 1 1 The Wilbur Collins model of amphibian metamorphosis predicts developmental plasticity in response to food availability ( figure based on Wilbur and Collins 1973) ................................................................................... 28 1 2 Diagram of growth trajectories fitting one interpretation of the Wilbur Collins model of amphibian metamorphosis. .................................................................. 29 1 3 Growth trajectories predicted for either decreased or increased growth rate in foodswitching experiments based on the Wilbur Collins model (figure based on Alford and Harris 1988 and Twombly 1996) ................................................. 30 1 4 Modified growth trajectories of Hensley alter the Wilbur Collins model by adding a transition point ( star ) to fixed development al time in the later part of the larval period ( figure based on Hensley 1993) .............................................. 31 1 5 B radshaw and Johnsons proposed insect model of metamorphosis modified from the original Wilbur Collins model of amphibian metamorphosis ( figure based on Bradshaw and Johnson 1995) ........................................................... 32 1 6 Alternative models of the relationship of development time to inverse food consumption predict the plasticity or canalization found in an organisms developmental strategy (based on Juliano et al. 2004). ..................................... 33 1 7 Unknown species in the genus Corethrella feeding on a southern toad, Anaxyrus ter restris . ............................................................................................ 34 2 1 Four competing models of development created by Juliano et al. (2004) to predict the relationship of total development time (T0) to the inverse of daily food consumption (1/F). ...................................................................................... 51 2 2 Mean daily food consumption ( SE) of female and male C. appendiculata larvae in the 4th instar. ....................................................................................... 52 2 3 The best fit linear regressions of the relationship between development time and inverse daily food consumption of fourth instar larvae of C. appendiculata. ................................................................................................ 53 2 4 The best fit linear regression relating wing length and the number of eggs developed in an autogenous egg batch. ............................................................. 54 3 1 Hypothetical relationships between development time and inverse food at high temperature and low temperature. ............................................................. 67
9 3 2 Best fit linear regression relating the inverse of daily food consumption to total development time in the fourth instar of C. appendiculata larvae reared at h igh and low temperatures.. ........................................................................... 68 3 3 Best fit linear regression relating the inverse of daily food consumption to total development time in the fourth instar of male and female C. appendiculata larvae. ..................................................................................... 69 3 4 Mean prey consumption in fourth instar C. appendiculata larvae when offered 40 prey per day at high (27C) or low (22C) temperature. ................................... 70 4 1 Best fit linear regression relating the inverse of daily food consumption to total development time in the fourth instar of C. appendiculata la rvae from high and low previous instar food treatments. .................................................... 86 4 2 Best fit linear regression relating the inverse of daily food consumption to total development time in the fourth instar of male and female C. appendiculata larvae. ..................................................................................... 88 4 3 Mean food consumption in fourth instar C. appendiculata offered 40 prey/day following high or low food treatment in the previous instar. ................................ 89 4 4 Best fit linear regressions of the relationship between wing length and the number of eggs developed by C. appendiculata females given hi gh or low food in the penultimate instar. ............................................................................. 90 4 5 Mean number of eggs developed by female C. appendiculata reared at various fourth instar food levels following high or low food treatment in the previous instar. ................................................................................................... 9 1 4 6 Mean head capsule width of C. appendiculata fourth instar larvae following high or low food treatments in the previous instar. ............................................. 92 4 7 Mean female development time of fourth instar C. appendiculata coming from high or low food treatments in the previous instar followed by a fourth instar food switching schedule. .................................................................................... 93 4 8 Mean male development time of fourth instar C. appendiculata coming from high or low food treatments in the previous instar followed by a fourth instar food switching schedule. .................................................................................... 94
10 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy TESTING PLASTICITY OF DEVELOPMENTAL STRATEGY IN THE AQUATIC DIPTERAN LARVAE OF CORETHRELLA APPENDICULATA By Erik Blosser M ay 2014 Chair: L. Philip Lounibos Major: Entomology and Nematology Developmental strategies among insects are known from a relatively few, well studied species. Among these species, differences in the ability to respond to variable environmental factors with changes in development are likely related to qualities of the sp ecies natural habitat. Aquatic dipteran larvae in temporary habitats face variations in food availability and temperature as well as threats from drying and predation which may favor the development of plastic developmental strategies. Developmental plast icity was tested in the aquatic larvae of the predatory midge, Corethrella appendiculata. Using previously developed models comparing development time to food consumption, individual midge larvae were tested for plasticity in developmental thresholds and post threshold periods over a range of food treatment levels. Final instar midges were found to have no plasticity in pupation threshold or post threshold time with varying food levels. Though egg number varied with nutrient availability, no plasticity was found in the reproductive strategy of autogenous egg development. Further testing at differing temperatures found no plasticity of threshold in response to
11 temperature, although developmental time was shortened through higher daily food consumption and shorter post threshold time at higher temperature. The effects of the food environment of a previous instar were found to carry over to affect development time in the final instar. This developmental inertia was found in two separate experiments: one using developmental models and one using a food switching technique. When penultimate instars were given low food treatments, development time was increased during the following final instar. This effect was due to the necessity of consuming more prey before reaching the pupation threshold, though it is unclear whether this was due to an adaptively increased threshold or a reduction in assimilation efficiency. Final instar head capsule size and adult fitness traits (egg number, wing length and egg size) were not af fected by previous instar food conditions. Results suggest that C. appendiculata may be adapted to its semi permanent larval environment through a long post threshold period which allows resource acquisition under favorable conditions and escape under poor conditions Comparison with species from more temporary aquatic habitats could test the extent of these conclusions.
12 CHAPTER 1 LITERATURE REVIEW AND BACKGROUND Introduction Current knowledge of the developmental st rategies of larval insects is largely based on a small number of well studied model species. As expected from lifehistory theory, some species are able to adjust their development to fit the environmental factors they encounter and the differences between species strategies seem to be related to differences in the natural habitats from which they originate. Although aquatic Dipterans, such as midges and mosquitoes, have been studied extensively at the level of average group development, less is known about the factors controlling development in the individual. This research uses the aquatic midge, Corethrella appendiculata, to examine the effects of environmental factors on individual development. The following literature review discusses life history theory relating to complex life cycles and metamorphosis as well as early studies of amphibian development. A summary is given of the variety of developmental strategies currently described in insects and related studies among aquatic Dipterans. The biology of C. appendiculata is discussed including possible connections between plasticity in larval development and adult reproduction. Complex Life Cycles Life history transitions are changes through which an organism passes during its life cycle and include events such as molting, metamorphosis, maturation and oviposition. In many species these transitions display phenotypic plasticity, meaning that the phenotype for a given genotype varies with environmental conditions (forming a reaction norm). Examples of phenotypic plasticity in life history transitions include
13 variation in the timing, body size or nutrients stored at a specific transition. The alternative is a canalized trait which displays the same phenotype in all environmental conditions ( Nylin and Gotthard 1998). Plasticity may be adaptive or nonadaptive. Adaptive plasticity allows an organism to respond to environmental variations in a way that increases the organisms fitness. One example would be a tadpole responding to cues of pond drying by increasing development rate in order to reach metamorphosis faster and escape the larval habitat. In this case, the increased fitness is due to increased likelihood of survival ( Newman 1992, Denver et al. 1998, Rose 2005). Non adaptive plasticity, on the other hand, can result from physiological mechanisms which do not necessarily increase fitness. For example, most organisms will develop faster if they are reared at a higher temperature. If this faster development is only the result of unselected physiological mechanisms such as increased reaction rates of biochemical processes, then the plasticity correlated with temperature could be considered non adaptive ( Newman 1992 Nylin and Gotthard 1998). Distinctions have also been drawn between plasticity in the form of developmental conversion versus phenotypic modulation ( Smith Gill 1983 ). In developmental conversion, the environment induces a change in the gene program governing development which is likely to be an adaptive response. Phenotypic modulation, however, results when developmental rates are altered by the environment without a change in the overall program of development. This may be adaptive or nonadaptive as in the temperature example given above. Animals with complex life cycles often show an ontogenetic niche shift which involves metamorphosis accompanied by a change from one environment to another.
14 This life history transition is found in many organisms including holometabolous insects, amphibians and fish. Some hypotheses explain the evolution of complex life cycles as an adaptation to variable, unpredictable environments ( Wilbur 1980). In this view, a larval environment with high nutrition and survivorship can lead to a high fitness level in the resulting adult, whereas a poor larval environment can be escaped through metamorphosis to the adult environment, increasing fitness. In animals such as amphibians, growth in the juvenile stage after metamorphosis allows some recovery from a poor larval environment, although effects often continue into adult life ( Relyea and Hoverman 2003, Hector et al. 2012). In insects, rigid exoskeletons limit growth in the adult stage. Since adult body size is often correlated with fecundity (partially due to the physical space available for eggs), larval nutrition and timing of metamorphosis can play a major role in determining insect fitness ( Armbruster and Hutchinson 2002). Many factors can affect the quality of the larval environment or lead to variation in larval development. Higher temperatures often lead to decreased development time and size at metamorphosis. This effect is at least partially due to nonad aptive physiological mechanisms ( Chown and Gaston 2010). Food availability or competition can affect the growth rate of larvae leading to complex variations in development time and size. In general, higher food levels and faster growth rates tend to result in a shorter development time and larger size at metamorphosis. Other factors known to cause variation in development in the larval environment include cues of predation risk ( Relyea 2007) and, for aquatic larvae such as amphibians and mosquitoes, cues of habitat drying ( Newman 1992).
15 Early Amphibian Models of Metamorphosis An early model which has influenced much of the research on timing of life history transitions is Wilbur and Collins 1973 model of amphibian metamorphosis. Wilbur and Collins suggested that development was plastic and that the timing of metamorphosis could be adjusted to maximize fitness given the quality of the larval environment. In their model ( F igure 1 1 ), all individuals must reach a minimum size, b, required for metamorphosis. After reaching the minimum size, the fastest growing larvae (in the best environments) will continue growing until reaching a maximum size, b + c, which triggers the initiation of metamorphosis. Larvae in poorer environments with slower growth rates will spend less time in the larval environment after reaching the minimum size, instead escaping to the adult habitat. Between the minimum and maximum weights, the decisi on to initiate metamorphosis depends on whether the recent growth rate, dW/dt, is greater than some function of the current body weight, g. This means that individuals which have higher growth rates will spend more time in the larval environment after passing the minimum size threshold than individuals in poorer larval environments. The model is predicting adaptive developmental plasticity. Figure 1 2 shows a graphical interpretation which fits with the Wilbur Collins model. I ndividuals at various growth trajectories initiate metamorphosis according to some function of their current size and growth rate. However, it is important to note that the Wilbur Collins model does not predict a specific shape for this function. Instead, the shape in Figure 1 2 is similar to a model presented by Teder et al. ( 2008) as a modification of Day and Rowes ( 2002) overhead threshold model, although these authors do not include the variation in post threshold development time proposed by Wilbur and Collins. In the years following Wilbur and Collins publication, researchers
16 have extended their predictions to experiments in which food is switched from a high level to a low level or vice versa at various points in larval development ( Figure 1 3 ). This experimental technique allows the detection fixed and plastic periods in development. Researchers testing this model have often found that signi ficant periods in the late larval stage have a fixed developmental time. Travis ( 1984) work on Hyla gratiosa tree frogs suggested that development time became fixed at an early time in the larval period, although size remained flexible depending on resource levels. Alford and Harris ( 1988) introduced food switching experiments in which larvae were switched from a high food level to a low food level and vice versa at various points in development to test for a loss of developmental plasticity. Their data with Fowlers toads indicated that development was fixed late in the larval period, agreeing wit h Hensley ( 1993), who showed that development in Pseudacris crucifer frogs was fixed in the final third of the larval period, and Leips and Travis ( 1994), who found that several species of tree fr og had fixed development beginning around 60% of the way into the larval period. All of these studies describe the point of transition to fixed development in terms of time or percentage of the larval period completed. Hensley ( 1993) modified predictions from the Wilbur Collins model to include a transition to fixed development ( Figure 1 4 ). Comparison of Amphibian and Arthropod Development Many insects have complex life cycles, undergoing life history transitions similar to those of amphibians. Bradshaw and Johnson ( 1995) were among the earliest researchers to explicitly compare insect development with the early amphibian models. Growth trajectories were predicted based on two competing models: one in which development rate varied continuously with changes in growth (represented by the level
17 of food offered) and one in which development rate was fixed early in larval development. Using the pitcher plant mosquito, Wyeomyia smithii a food switching experiment was designed in which each of the four larval instars was fed either a high (H) or low (L) food treatment, resulting in eight treatments (HHHH, HLLL, HHLL, HHHL, LLLL, LHHH, LLHH, LLLH). Results showed that dev elopment was neither fixed early nor continuously variable, instead displaying a developmental inertia in which food levels in earlier stages affected the duration of later stages, but effects of the earliest stages could be largely overcome. Results ind icated that the mass and time at metamorphosis was largely set by the end of the third instar, but that some growth in the fourth and final instar was necessary for metamorphosis to occur. Food level in the final instar did not create a significant difference in timing or mass at metamorphosis. Bradshaw and Johnson ( 1995) created a new model for insects that is similar to the Wilbur Collins ( 1973) model for amphibians ( Figure 1 5 ). In this model, the decision to initiate metamorphosis between the lower and upper size levels is based on whether the mass added in the final instar (M M0 initial mass of the final instar (M0). These ideas originate from work by Nijhout ( 1975 1981) in which the larvae of Manduca sexta were found to initiate the pathway leading to metamorphosis after reaching some critical weight based on the initial weight of the final instar, which is, in turn, based on the final weight of the previous instar ( Nijhout 1994). In an experiment with similar design to Bradshaw and Johnson ( 1995 ) food levels in the first and final instars of lubber grasshoppers were found to have no significant effect on size at or time to adult molt ( Flanagin et al. 2000). This suggested
18 that developmental time may depend mainly on the middle three nymphal stages and was fixed prior to the final stage. However, a nonsignificant trend for longer development at lower food level in the final instar indicated that this result may be a result of low sample sizes (due to mortality). As in pitcher plant mosquitoes, some food was necessary in the final instar to initiate metamorphosis. A follow up experiment by Hatle et al. ( 2003) focusing solely on food variation in the final instar found large effects on developmental time. Studies from other types of arthropods have varied in results. Twombly ( 1996) found that, in copepods, both development time and size became fixed in the final 40% of the nauplius period (prior to the final s tage). Because the results showed a fixation of size, Twombly suggested that the hard exoskeletons of crustaceans may constrain growth in the final stage. Research on barnacle nauplii ( Hentschel and Emlet 2000) showed a sim ilar fixation of development in the final 40% of the nauplius period (corresponding closely with the start of the final instar), but size remained plastic throughout development in contrast to Twomblys ( 1996) results. Hentschel ( 1999 ) created a new model predicting the results of food switching experiments and the location of the point of transition from resource dependent to resource independent development. This model is based on the empiric al evidence from the barnacle studies and has received some support from studies on the larvae of porcelain crabs ( Howard and Hentschel 2005). Modeling Thresholds Based on Food Consumption The early amphibian models of metamorphosis have also been used to investigate plasticity vs. canalization in other life history transitions including sexual maturity ( Reznick 1990 ) and oviposition. Moehrlin and Juliano ( 1998 ) used food
19 switching experiments to show that time to oviposition in lubber grasshoppers became fixed early in egg development. However, the timing of the transition to fixed development depended on food level indicating that the transition correlates to a particular developmental state or threshold rather than a given time. In a critique of food switching experiments, Day and Rowe ( 2002) pointed out that this design could lead to the false appearance of late fixed development ( due to difficulty in detecting the smaller differences in timing after switches late in development). Juliano et al. ( 2004 ) responded by creating a group of models to differentiate between plastic and canalized developmen t without using the food switching design ( Figure 1 6 ). The models assume that the threshold for the initiation of egg development is based on the level of some storage product in the body (which is correlated with food intake). Individuals can be reared at a range of constant food levels to find a relationship between average food intake per day (F) and development time (To). This relationship is predicted to differ among the models. The four models are: 1) a constant threshold w ith the post threshold time canalized, 2) a constant threshold with the post threshold time variable (Wilbur Collins model), 3) a threshold decreasing with decreasing feeding rate and post threshold time canalized, and 4) a threshold decreasing with age and post threshold time canalized. Juliano et al. ( 2004) tested these models on the plasticity of egg development in lubber grasshoppers and found that the best fitting model assumed a constant threshold with canalized pos t threshold time. This model assumes no adaptive plasticity. The experiments in Chapters 2 through 4 of this dissertation use the models of Juliano et al. ( 2004) to test for developmental plasticity in the larvae of
20 C. a ppendiculata Although the mechanism triggering the initiation of pupation is more likely to be related to a critical weight rather than a particular level of storage proteins (see below), food consumption should be closely correlated with weight under many conditions. This likelihood can be increased by providing larvae with one type of food composed of constant proportions of nutrients. Physiology of Insect Development and Diversity among Species One of the best studied models of insect developmental physiology is the caterpillar, M sexta In this species, final instar larvae must reach a critical weight threshold which triggers the release of a hormonal cascade resulting in pupation after a fixed period of time, which is called the interval to cessat ion of growth or post threshold period. The critical weight threshold is correlated with the size of the instars head capsule and is therefore determined by the amount of food gathered in previous instars. Each instar must reach its own critical weight in order to molt to the next stage ( Nijhout 1981, Nijhout 1994). The physical mechanism triggering the cascade of developmental hormones is likely not the larvas weight itself, but some trait that is closely correlated with weight. Although the mechanism has not been completely worked out, recent work by Callier and Nijhout ( 2011) shows that threshold is likely related to the oxygen delivery rate to body cells. Si nce the tracheal system volume is set during each molt, the continual tissue growth throughout an instar leads to increasing oxygen demand by the body resulting in oxygen limitation after the critical weight has been reached. Further support for this idea comes from experiments raising M. sexta larvae under hypoxic conditions resulting in smaller critical weight and from similar findings in Drosophila
21 melanogaster ( Callier et al. 2013). Although this evidence strongly suggests that oxygen limitation is a factor affecting the hormone trigger, the authors point out that oxygen may be just one of several factors involved. Although M. sexta is the most thoroughly studied insect model, it is difficult to know how typical the development of this species is when compared to other insects. At least three alternative developmental strategies have been discovered in other insect species and more are predicted as researchers investigate more species that have evolved to meet the challenges and uncertainties of their particular habitats ( Nijhout 2008). In M. sexta for instance, final instars must reach a critical weight threshold that cannot be altered by differences in food availability or temperature. Thi s threshold is followed by a post threshold period governed by temperature but not food level. Food availability, therefore, mainly affects the time required to reach the critical weight threshold, whereas temperature has its bigger effect on duration of t he post threshold period ( Davidowitz et al. 2003, Nijhout et al. 2006). In contrast, another well studied model organism, D melanogaster shows plasticity in post threshold period with n utrition level and changes in pupation threshold with temperature. With slower growth under low nutrient conditions, D. melanogaster lengthens its post threshold period ( Layalle et al. 2008) and with increases in temperature the threshold size for pupation is lowered ( Ghosh et al. 2013). Both of these plastic traits are suggested to have possible adaptive value although this has not been shown conclusively. A longer post threshold period may provide larvae time to gather more nutrients when food is scarce and a smaller size threshold at high temperature may have been selected to provide the
22 smaller high temperature body size seen in many organisms. The relation of this plasticity to this species natural larval habitat is unclear. Several insect species show alternative strategies that have a clearer connection to larval habitat and lifestyle. The dung beetle, Onthophagus taurus spends the majority of its larval period in the third and final instar with the brood ball in which it hatches as the only source of food. Shafiei et al. ( 2001) reared larvae at the maximum growth rate and starved them at various points in the final instar, resulting in a pattern of development resembling some of the predictions of the Wilbur Collins model. The beetles needed to reach both a minimum time and mass before pupation was possible. After this point, starvation triggered hormone release, and starved beetles initiated the pathway leading to pupation at an earlier time and at a smaller size than continuously fed control beetles suggesting adaptive developmental plasticity. In a fourth developmental strategy, the nymphs of the true bugs Rhodnius prolixus, Dipetalogaster maximus and Oncopel tus fasciatus are triggered to molt when feeding signals abdominal stretch receptors that a large meal has been taken ( Nijhout 2008). This type of strategy makes sense when a large meal guarantees enough nutrients for development, such as in the blood feeding nymphs of R. prolixus and D. maximus Development in Aquatic Dipterans Due to their connection to human disease, mosquitoes are the most thoroughly studied of the aquatic Dipteran larvae. Research has been particularly concentrated on container dwelling species because of their tractability in experiments and their evolutionary relationship to humans and important human diseases. Studies of larval development have often focused on the development of groups of mosquitoes in an attempt to model population dynamics. One of the earlier and most exhaustive of these
23 studies was a sy stems analysis of Aedes aegypti by Gilpin and McClelland ( 1979 ). These researchers performed a series of experiments varying nutrient levels in the laboratory to create a model of group development called the pupation window m odel. This model proposes a window on the size vs. time axis (a minimum size and time) which must be reached for pupation and predicts the location of this window over a variety of nutrient and temperature levels. The pupation window model has been teste d favorably with another container dwelling mosquito, Aedes triseriatus, ( Carpenter 1984, Walker et al. 1997) and provided the basis for later, more complex models of Ae aegypti populations ( Focks et al. 1993). In a more recent model of reserve dependent growth, Padmanabha et al. ( 2012 ) used concepts related to the life history theory of dynamic energy budgets to predict size, devel opment rate and starvation mortality in Ae. aegypti across a range of temperatures. The effect of environment on individual developmental thresholds in mosquitos is less well studied. Lounibos ( 1979) starved larvae at various points in the final instar of Toxorhynchites brevipalpis to show that a critical weight of about 26 mg was followed by a post threshold period of 7 days that was not affected by food availability. Telang et al. ( 2007) used a similar starvation method to show that Ae aegypti larvae have a critical weight between 2.73.2 mg followed by 1.6 day post threshold period which was unaffected by food level. Results of this study further suggested that, in contrast with M sexta the timing of ecdysteroid hormone release was not critical to the timing of pupation, although both ecdysteroid titre and nutritional status were crucial factors. Chambers and Klowden ( 1990) reared Ae aegypti larvae through all four instars at separate temperatures and found that critical weight was higher in the lower
24 temperature (22 C) than the higher temperature (32 C). This difference could be an example of plasticity in threshold with temperature as found in D melanogaster but may also be the result of raising all four instars in separate treatments. Early instars at higher temperatures would have shortened post threshold periods leading to smaller size entering the final instar and therefore a smaller critical weight threshold. Chambers and Klowden ( 1990) compared levels of glycogen, sugars and soluble lipids in larvae at critical weight at the two temperatures in an attempt to discover whether the storage of any of these nutrients was the r eal trigger of pupation. Since only glycogen levels were similar between the high temperature critical weight and low temperature critical weight, the authors suggested that a glycogen storage threshold might be the real trigger for pupation. Later authors however, have pointed out that this is unlikely the sole trigger of metamorphosis since basic nutrient reserves are needed in other categories as well ( Telang et al. 2007) The few studies reviewed above give some evidence for a critical weight threshold in mosquitoes followed by a post threshold period which is not affected by food availability. In Ae aegypti the threshold differs between temperatures, though the underlying explanation of this is unknown. Bradshaw and Johnson ( 1995) investigated inter instar nutrient effects and found evidence for developmental inertia between instars in W. smithii. There is still much unknown about how envir onmental conditions affect development in these container dwellers. Individual development in mosquitoes is becoming more important to understand as studies connect stressful larval conditions with increased adult susceptibility to viruses ( Alto et al. 2008, Alto and Bettinardi 2013) and increased vertical transmission of viruses to offspring. A better understanding of the
25 effects of environmental factors on individual development may help predict conditions under which larval stress could increase susceptibility to viruses and may help refine population models of group development. Biology of Corethrella appendiculata Corethrella appendiculata Grabham is a small frog biting midge from the family Coret hrellidae which is closely related to Culicidae, the mosquitoes. Larvae are found in treeholes and containers in wooded environments throughout Florida and in Central and South America ( Borkent 2008). The larval stage is comp osed of four instars followed by a free swimming pupa, and the entire life cycle lasts around 33 days ( McKeever and French 1991). The predatory larvae are large enough in the 3rd and 4th instars to eat the early instars of m osquito species such as Aedes albopictus and Ae triseriatus which share their treehole/container habitats in Florida ( Alto et al. 2009). The eggs of C. appendiculata, unlike those of Aedes mosquitoes, cannot survive desiccat ion Larvae of this midge, therefore, tend to be more frequently found in the dark, nutrient rich waters of more permanent treeholes, rather than in the clearer, more temporary waters of shallow holes ( Bradshaw and Holzapfel 1983). This may be due to an oviposition site selection preference, although this hypothesis has not been tested. These midges are often found in treeholes with no other predators, although they may co occur with a larger predatory mosquito, Toxorhynchites rutilus ( Bradshaw & Holzapfel 1983, Lounibos 1983). In the lab, T. rutilus larvae will feed on C. appendiculata ( Griswold and Lounibos 2006). Because of their relatively more permanent habitat and lack of abundant predators, cues of drying and predation are probably not as important in determining larval development time as they are in some
26 species of treehole mosquitoes. Instead, nutrient availability and temperature are likely to be the bigger factors affecting development time in this species. Corethrella appendiculata larvae use a sit and wait predation strategy and can survive several weeks of starvation in the final instar ( B losser et al. 2013). They are abundant in south Florida treeholes ( Lounibos 1983, 1985) and in a long term census were found in densities ranging from a few midges per liter to well over sixty ( Lounibos 1983). Thus, there is likely to be a wide variation of food levels per individual in the field. Adult females are autogenous, developing their first batch of eggs from resources collected in the larval stage. Egg batches develop synchronously, and the largest batches developed by pupae collected from the field numbered around 85 eggs ( Blosser et al. 2013). Adult females have also been captured while bloodfeeding on a frog in the field ( Figure 1 7 ) and females in the lab can take a bloodmeal to supply the nutrients needed for development of a second batch of eggs (Blosser, unpublished). The larvae of this midge have several advantages which make this species an attractive choice for studying developmental thresholds and plasticity. As with many container dwelling mosquitoes, C. appendiculata is easily colonized, and their autogenous egg development removes the need of an adult blood source for colony mainten ance. The predacious nature of the larvae allows the use of discrete prey units in the development models created by Juliano et al. ( 2004) to estimate food thresholds with higher precision than the critical weight thresholds estimated by food switching experiments. Development time in C. appendiculata is extended compared with that of some mosquito species, favoring detectability of environmental effects which might be difficult to observe in mosquitoes with instars lasting only a few days. Finally, the low
27 cannibalistic rates compared with larvae of Toxorhynchites mosquitoes allow the rearing of large numbers together through the early inst ars. This ensures a similar developmental history for larvae used in experiments investigating the later instars and provides higher numbers to increase statistical power. The advantages given above provide a good starting point for investigation of developmental plasticity in container dwelling species. Results from these studies can then be compared across species with similar habitat and life histories to determine the extent to which the resulting strategies are shared.
28 Fig ure 1 1 The Wilbur Collins model of amphibian metamorphosis predicts developmental plasticity in response to food availability ( figure based on Wilbur and Collins 1973). This model postulates a lower limit of size b and an upper lim it of size b + c. Between these sizes, metamorphosis is initiated if the current growth rate is greater than some function of the current size, g.
29 Figure 1 2. Diagram of growth trajectories fitting one interpretation of the Wilbur Collins model of amphibian metamorphosis. Growth trajectories extend from zero age and size to end at the initiation of metamorphosis. All trajectories must cross the minimum size threshold, b, but cannot cross the maximum size, b + c. As indicated by the red arrows, individuals experiencing higher growth trajectories spend more time in the larval environment after surpassing the minimum size threshold. Between b and b + c, metamorphosis is initiated at some curve which is a function of current weight and growth rate. The shape of this function is not specified by the model.
30 Figure 1 3. Growth trajectories predicted for either decreased or increased growth rate in foodswitching experiments based on the Wilbur Collins model (figure based on Alford and Harris 1988 and Twombly 1996). Adaptive plasticity above the minimum size threshold predicts that (A) a food switch from high to low will shorten development time to allow escape of a poor e nvironment and (B) food switch from low to high will prolong development time to take advantage of a high food environment. Circles represent the initiation of metamorphosis.
31 Figure 1 4. M odified growth trajectories of Hensley alter the Wilbur Collins model by adding a transition point ( star ) to fixed development al time in the later part of the larval period ( based on Hensley 1993). These graphs show the revised predictions for foodswitching experiments after a switch from (A) high to low food and from (B) low to high food. Circles represent the initiation of metamorphosis.
32 Figure 1 5. Bradshaw and Johnsons proposed insect model of metamorphosis modified from the original Wilbur Collins model of amphi bian metamorphosis ( figure based on Bradshaw and Johnson 1995). In the insect model, metamorphosis is initiated between b and b+c when the current mass minus the mass at the beginning of the final instar (M M0) has exc of the initial mass of the instar. This is based on the idea of a critical weight from Nijhout ( 1981).
33 Figure 1 6. Alternative models of the relationship of development time to inverse food consumption predict the plasticity or canalization found in an organisms developmental strategy (based on Juliano et al. 2004). Juliano et al. created models of some of the possible patterns of development based on a constant or declining threshold and canalized or flexible post threshold time. These models predict the relationship of the inverse of average food consumed per day (1/F) with the total development time (To).
34 Figure 1 7. Unknown species in the genus Corethrella feeding on a southern toad, Anaxyrus terrestris in Vero Beach, Florida. Photograph by Erik Blosser.
Reprinted with permission from Annals of the Entomological Society of America. CHAPTER 2 T ESTING DEVELOPMENTAL PLASTICITY IN AQUATIC LARVAE OF CORETHRELLA APPENDICULATA (DIPTERA : C ORETHRELLIDAE) Introduction Life history transitions are the changes through which an organism passes during its life cycle and include events such as molting, metamorphosis, maturation and oviposition. In m ost if not all species, these transitions exhibit some form of phenotypic plasticity in which environmental conditions influence the expression of the organisms genotype ( Nylin and Gotthard 1998) Researchers find plasticity in the timing of the developmental event or in the individuals size or nutrient stores at the point of transition ( Whitman and Ananthakrishnan 2009) and there is usually a limited window of time during which environmental inputs can affect these traits ( Hensley 1993 Twombly 1996, Hentschel 1999, Telang et al. 2007) F or instance, Hensley ( 1993) found that the timing of tadpole metamorphosis became unresponsive to varied food levels in the final third of development while Twombly ( 1996) observed the canalizat ion of both development time and body size in the final 40% of copepod development. Bradshaw and Johnson ( 1995) were among the earliest researchers to examine insects for the patterns of fixed vs plastic development that had been previously modeled in amphibian systems ( Wilbur and Collins 1973, Alford and Harris 1988, Leips and Travis 1994) Based on their investigations into the molting and metamorphosis of pitcher plant mosquitoes and on physiological work performed on Manduca sexta ( Nijhout 1981) Bradshaw and Johnson proposed an insect development model mirroring early amphibian models with the exception that metamorphosis in insects could be initiated by the attainment of a mass threshold in the final instar. This critical
36 weight threshold has been studied extensively in M. sexta cat erpillars and is likely only correlated with the actual trigger of the hormonal cascade leading to metamorphosis ( Davidowitz et al. 2003) The trigger is also likely to be closely correlated with a storage protein threshold and therefore the amount of food ingested by the organism. Juliano et al. ( 2004 ) used this relationship to create a set of models ( Figure 2 1 ) which differentiate between plastic and canalized developmental timing patterns based on an organisms total development time and food consumed per day. These four models distinguish between a (1) constant food threshold with canalized post threshold time, (2) constant threshold with flexible post threshold time, (3) threshold that varies with food level followed by canalized post threshold time and (4) threshold that varies with age followed by a canalized post threshold time. Although this technique can be used for any life history transition associated with food intake, Juliano et al. ( 2004 ) tested the developmental plasticity of grasshopper ovipos ition. Under varying food conditions the final instar of the well studied M sexta larva displays a constant critical weight threshold followed by a canalized post threshold time. However, a very small percentage of insects have been examined and, among these, at least three alternative developmental strategies have been identified ( Nijhout 2008) These include (1) stretch receptors which stop growth at a constant size threshold in Oncopeltus, Rhodnius and Dipetalog aster (2) commitment to pupation initiated by means of starvation in Onthophagus ( Shafiei et al 2001 ) and (3) flexible post threshold time based on nutrition level in Drosophila ( Layalle et al. 2008) Nijhout ( 2008) predicted the discovery of many more strategies given the broad range of insect life histories.
37 While the metamorphic trigger has not been settled in most insects, suggestions have included thresholds of storage proteins and tracheal oxygen supply ( Callier and Nijhout 2011) Due to their unique env ironment, aquatic insect larvae are a prime group for the comparison of developmental strategies. Although mosquito larvae are comparatively well studied among the aquatic insects and a critical weight has been recorded in several species ( Lounibos 1979, Chambers and Klowden 1990, Telang et al. 2007) no investigations have yet focused on the plastic vs fixed developmental strategy of individual organisms in this or any other aquatic insect group. Within the mosquitoes, nutritional development models have focused on cohorts ( Gilpin and McClelland 1979, Focks et al. 1993) or inter instar effects ( Bradshaw and Johnson 1995) but an increasing interest in the effects of poor nutrition on individual mosquitoes ( Alto et al. 2008) has elevated the importance of individual development. Furthermore, the possibility of alternative development strategies is supported by occasional results from the mosquito literat ure, such as fixation of development time and size prior to the final instar ( Bradshaw and Johnson 1995) or faster development of larvae in poorer food environments ( Reiskind and Zarrabi 2012) In this study we use the frog biting midge Corethrella appendiculata to examine the developmental strategy of an aquatic dipteran larva. The family Corethrellidae is closely related to the Culicidae with a similar life cycle including an aquatic larva and pupa followed by a bloodfeeding female adult. C. appendiculata larvae are predatory ( Griswold and Lounibos 2005, Kesavaraju et al. 2007) with experimental advantages such as discrete prey allowing for easy measurement of food consumption, ability to be reared in large numbers, and lengthy development time increasing detectability of
38 treatment effects. Ad ditionally, adults of this species produce eggs autogenously but have been collected feeding on a squirrel treefrog in the field (personal observation). It is unknown whether C. appendiculata exhibits a plastic reproductive strategy of facultative autogeny for the first egg batch ( O'Meara 1985) or whether bloodfeeding is used only to produce subsequent batches. In this study we tested the developmental plasticity models developed by Juliano et al. ( 2004 ) using final instar larvae of C. appendiculata. We further investigated the reproductive plasticity of this species through laboratory experiments and field collections. Materials and Methods Rearing an d Larval Development Experiment The laboratory colony of C. appendiculata originated with the collection of 200300 larvae from treeholes at the Florida Medical Entomology Laboratory in 2003 and is maintained with additional supplements of aquatic immatures from treeholes and artificial cont ainers at least twice yearly. The colony is housed in an insectary maintained at 27 1 C, 83% humidity and 14L:10D light:dark cycle. Adults are housed in a 0.1 m3 mesh cage as detailed in Lounibos et al. ( 2008). Eggs w ere synchronized by age by withholding an oviposition cup from the colony for 2 days followed by 24 hours of access to the container. Eggs were removed from the cage and checked every 24 hours for hatching. After 3 days, 210 newly hatched (<24 hours old) l arvae were placed in a metal tray (5 x 25 x 35 cm) with one liter of water. These larvae were reared together through the first three instars, receiving a 0.4 mL scoop of oatmeal/nematode ( Panagrellus sp.) mixture every other day after which the water was stirred to distribute nematodes evenly. Preliminary tests found that this high food level led to rapid, even
39 development, suggesting no shortage of food. All 4th instar larvae, recognized by head capsule width, were removed from the pan daily between day 8 and 13 of development. Of the 210 original larvae, 180 developed into 4th instars during this 6 day period, 8 larvae developed more slowly and 22 larvae died in the early instars. Each day newly molted 4th instars were placed into individual glass vials w ith 13 mL of water and distributed evenly among food treatments placed in an incubator set to 27 ( 0.1 SE) C. Food treatments consisted of 15, 10, 8, 6, 5, 4, 3, 2, 1, or 0 first instar Aedes albopictus larvae per day. If the daily number of new 4th inst ars was not divisible by 10, the extra larvae were assigned to a treatment using a random number generator. Prey were hatched daily by placing approximately 500 Ae. albopictus eggs into a pan with 1 L of water and 0.3 grams yeast/albumin mixture and allowi ng them to eclose and develop for 24 3 hours before use in the experiment. Every 24 ( 3) hours the water was dumped from each experimental vial into a petri dish and rinsed once in an attempt to recover all leftover prey. New water and fresh prey were added to the vial and uneaten prey were counted under a microscope. Dead prey were recorded as either half or whole. Each day treatments were arranged randomly in the incubator using a random number generator and processed using the same random order. Mode l Selection Using the data collected on development time and average daily food consumption, the models developed by Juliano et al. ( 2004 ) were compared to find which of the four proposed developmental strategies best explained the data ( Figure 2 1 ). These models include the (1) constant food threshold with canalized post threshold development time (Constant/Canalized) model = ( 1 ) +
40 in which T0 is the total development time, H is the food threshold necessary to trigger commitment to pupation, 1/F is the inverse of the average daily food consumption and Tp is post threshold time. The model representing (2) a constant food threshold with flexible post threshold time (Constant/Flexible) = ( 1 ) + [ ( + ) ] exchanges Tp for an equation allowing variation of the post threshold time with food level. Here Tpmax is the maximum post threshold time and k is a parameter whose value dictates the effect of feeding rate on post threshold time. If the parameter has no value (k=0), then the equation simplifies to the Constant/Canalized model. The model representing (3) a threshold declining with feeding rate with constant post threshold time (Foodde pendent/Canalized) = ( 1 )+ ( 1 ) + replaces the threshold, H, with a maximum threshold, H0, which is adjusted by feeding rate (1/F) in a relationship governed by the b1 parameter. Finally, a model representing (4) a threshold declining with age and constant post threshold time (Agedependent/Canalized) = [ ] + again replaces the threshold, H, with a maximum threshold, H0, which is then adjusted by TH, the time to the threshold, in a relationship governed by the b2 parameter. Full details of model development are given in Juliano et al. ( 2004). Nonlinear regressions were performed using PROC NLIN (SAS Institute 1998) and the resulting error sums of the squares were used to compare models through t he corrected Akaikes Information Criterion (AICc) which is recommended in preference to AIC ( Anderson 2008, Symonds and Moussalli 2011). Models which have the lowest
41 AICc are considered the most likely to have the best explanation of the data. AIC is preferable to some traditional model testing techniques because it avoids arbitrary significance values and instead ranks models based on the amount of information explained by each ( Symonds and Moussalli 2011). Model parameters were compared between sexes using ANCOVA to compare slopes directly. Data were square root transformed to meet assumptions of normality. Reliability of Prey Recovery Technique In order to test the reliability of the leftover prey counts, a separate control was run for three days in which vials received a known number of Ae des albopictus prey but no predator. Twenty vials were divided into four treatments receiving either 20 live, 20 dead, 5 live and 5 dead, or 40 half (mechanically severed) prey larvae and the 24 hour protocol was followed as described above. In the treatment with 10 liv e prey, recovery averaged 9.93 ( 0.07 SE, n =15), the 10 dead prey treatment averaged 9.80 ( 0.11 SE, n =15), the 5 live and 5 dead treatment averaged 7.37 ( 0.24 SE, n =15) and the 20 half treatment averaged 9.80 ( 0.08 SE, n=15). Recovery rate was, t herefore, high in all except the 5 live5 dead treatment, likely due to consumption of dead larvae by the live ones in this group. Although this effect may have influenced our data at the highest food levels of the main experiment, we estimate the influenc e to be slight due to (1) the presence of a predator and predatory cues leading to the reduction of prey feeding activity ( Kesavaraju et al. 2007a) and (2) low numbers of live prey leftover in any experimental treatment. In the highest food treatment only 28% of vials contained live prey at the time of leftover prey counts with a mean of 2.18 ( 0.22 SE, n=33) live prey per vial.
42 Plasticity in Egg Development Vials containing pupae were checked daily until adult emergence at which point a cotton ball soaked in 10% sugar solution was placed daily on the screen covering the vial. About 5 days (116 4 hours) after emergence adult midges were killed with chloroform, wings were removed and females were dissected to count the developed eggs. This time period allowed females to fully develop an autogenous egg batch and only one individual was found to have partially formed eggs, suggesting that females of this species do not resorb eggs. However, females often dropped a portion of their unfertilized eggs into the water, requiring that vials were also checked for the floating eggs during counts. In a subsample of individuals from each food treatment, 10 eggs from each specimen were photographed under the microscope and measured using a computer imaging program (i Solution lite; AIC, Princeton, NJ). The last four females to emerge in each of the three highest (15/day, 10/day, 8/day) and three lowest (3/day, 2/day, 1/day) food treatments were selected for egg measurement, allowing the comparison of egg length in high vs. low food treatments by t test. Wings were photographed and measured using the same computer imaging software. In the mosquito literature, wing length is commonly assumed to be correlated with body size and is measured from the alula to the wingtip ( Packer and Corbet 1989) As the alula was not easily located in C. appendiculata, wing lengths were measured from a closed cell near the base of the wing to the outer edge of the wingtip excluding the fringe. Pupae of C. appendiculata were collect ed in every month of the year from water filled tires and tree holes in Vero Beach, FL with the goal of estimating the numbers of eggs developed autogenously in the field and determining whether facultative autogeny is present in this species. Fieldcollec ted pupae were placed in individual vials in the
43 same 27 C insectory described above. After emergence adults were fed 10% sugar solution for 4 5 days, after which wings were removed and females were dissected to count eggs. Results Larval Development O f the 162 larvae that received daily prey, 9 died before emergence and one was accidentally killed. Survival was lowest in t he food treatment of one prey/day in which 5 of 18 larvae died prior to emergence. Since the 13 survivors consisted of 10 females and 3 males, the lowest food treatment was excluded in the male analysis to avoid skewing the results towards faster developing individuals. In the control treatment with no prey, none of the predator larvae pupated, and starvation survival times varied from 6 to 44 days with a mean of 24.1 ( 2.7 SE, n=18) days. Male larvae developed significantly faster than females through the first three instars (t152 = 3.22, P < 0.01) with males averaging 9.23 0.16 (SE) days and females averaging 9.99 0.18 (SE) days A Wilcoxon rank sum test verified that these results are robust to violations of normality assumptions (Wilcoxon Twosample Test: Z approximation = 3.24, p = 0.0011, twotailed). Males also developed more quickly through the fourth instar in the highest food treatment (t15 = 2.47, P= 0.03), leading to more rapid overall larval development time in males (16.50 0.82 SE days) relative to females(18.44 0.87 SE days) that was not significantly different (t15 = 1.61, P=0.13). In the lowest comparable food treatment, overall male development time (19.00 0.48 SE days) and female development time (24.83 1.30 SE days) were significantly different (t15 = 5.08, P<0.001). Female larvae consumed significantly mor e prey per day than males in the 15/day (t14=2.90, P= 0.01), 10/day (t14 = 2.80, P=0.01), 8/day
44 (t16=4.64, P < 0.001), 6/day (t16=2.67, P=0.02), and 5/day (t11 = 2.33, P = 0.04) food treatments, though analysis of the results of the lower food treatments s howed no significant differences in prey consumption between sexes ( Figure 2 2 ). Model Selection In both males and females, all parameters were significant in the constant/canalized model which also scored the lowest AICc of the four models ( Table 2 1 ). In both sexes, the addition of the k parameter in the constant/flexible model resulted in an increase in AICc of about 2 without any change in the error sum of squares. The remaining alternative models fooddependent/canalized and agedependent/canalized, explain slightly more of the data, reducing the error sum of squares, but have a higher AICc due to their extra parameters. The weight of evidence, wi, and the evidence ratio show that, for females, t he constant/canalized model is more than 1.6 times as likely as the second best model. For males, the best model is 2.3 times as likely as the second best agedependent/canalized model. The R squared for the best model is 0.803 in females and 0.391 in males. The constant/canalized model ( Figure 2 3 ) was used to estimate parameter values for males and females and direct comparison between sexes was performed by ANCOVA. The AICc model selection procedure was repeated using inverse tr ansformed models developed in Juliano et al. ( 2004) in order to check for sensitivity to normality and homogeneity of variance assumptions. Use of the transformed models confirmed the constant/canalized models as the most likely for both sexes with similar parameter estimates and these results are not given here.
45 Egg Clutch Size All females from all food treatments developed eggs autogenously and the number of eggs correlated closely with the females wing length ( Figure 2 4 ). The female with the smallest wing length contained only partially developed eggs when dissected and was also an outlier with only 23 eggs in development. Since all other females contained fully developed eggs, this point was removed from the regression analysis. Differences between the mea n egg lengths from the three highest and three lowest food treatments were nonsignificant (t22 = 0.03, P = 0.97). Field collections of pupae also revealed high autogeny rates with only 1 out of 133 dissected females failing to produce an autogenous egg batch. Across all months, egg batch sizes generally ranged between the mid 20s and mid 80s with a maximum of 86, a minimum of 9 and a mean of 55.3 ( 1.33 SE, n = 132) eggs per batch. Discussion Model Selection We found good evidence that 4th instar C. app endiculata larvae develop to a constant food threshold followed by a fixed post threshold period during which food levels no longer affect the timing of pupation. A similar result in both males and females adds strength to this conclusion. Although the fo oddependent and agedependent threshold models explained more of the data, the additional information was not enough to overcome the penalty imposed by AICc for added parameters. The k parameter from the constant/flexible model added no information, giv ing the same error sum of squares as the constant/canalized model. Anderson ( 2008) calls this a pretending variable since each additional parameter incurs a cost of about 2 AICc and no additional information has been explained in the data. Models with a pretending variable
46 should be rejected for their simpler versions. Since the AICc is a relative ranking, it is important to calculate a measure of worth of the most likely model since all models may be poor ( Anderson 2008 ). The male R squared value of 0.391 is much lower than the female value of 0.803 but this difference is at least partially a side effect of removing the lowest food treatment from the male analysis. One difficulty encountered in the current experiment was choosing the proper range and spacing of food treatments on the 1/F axis. Uneven spacing was caused by an inability to choose intermediate levels between the 1/day, 2/day and 3/day treatments leading to an overly influential l owest food treatment. This led to a disproportionate effect of removing the lowest male food treatment. Use of smaller prey could alleviate this problem in future experiments. However, the grouping of data at the high food end was beneficial for detecting any flexibility in the post threshold period. In comparison with other insects, the constant/canalized developmental strategy appears similar to that of the well studied M. sexta caterpillar as well as several other species, but at odds with the strategy of Drosophila melanogaster ( Nijhout 2008). However, each species is adapted to its particular life history, and it is unknown how broadly these strategies are shared among evolutionary relatives. Are other midges and mosquito larvae likely to develop in a similar fashion? Some predatory mosquitoes, like species in the genus Toxorhynchites, share with C. appendiculata midges many ecological and developmental features, such as long development time and preference for more perman ent water habitats. However, mosquito species such as Aedes aegypti Ae. albopictus and the predator Psorophora howardii ( Lounibos 2001) show a contrasting strategy, developing quickly in more temporary water habitats and ri sking
47 the dangers of drying or depletion of food. Adaptive plasticity is thought to develop more commonly in organisms facing variable environments as a means of surviving an unpredictable future. For example, a comparative study of spadefoot toad species has shown that the species from the most ephemeral habitat had an earlier development threshold followed by higher plasticity in body size ( Morey and Reznick 2000) Mosquito species from temporary water habitats may be the more likely to show highly plastic developmental strategies. The comparatively long post thres hold period (6.52 and 7.47 days) of C. appendiculata is similar to that of Toxorhynchites brevipalpis estimated to be around 7 days ( Lounibos 1979) but contrasts with the 1.6 day period found in Ae aegypti ( Telang et al. 2007) Correspondingly long pupal durations are found in C. appendiculata (5 6 days) and Toxorhynchites rutilus and T oxorhynchites amboinensis (about 5 days) ( Lounibos et al. 1996) and seem maladaptive from a lifehistory viewpoint in which faster generation times are favored and a non feeding, nonreproductive phase contributes little to lifetime reproduction. However, similar hormones control the events of both late larval and pupal development ( Nijhout 1994 ) and lengthy pupal durations could be caused by a physiological linkage to an adaptively beneficial post threshold period. In any case, the long post threshold periods of these species are likely traits adapted with their more permanent habitat preference. Larval Development Male precedence in molting from the 3rd to the 4th instar has been recorded previously in mosquitoes such as Aedes sierrensis ( Mercer et al. 2008) Protandry, the eclosion of male adults prior to females, has been observed in many insect species and is reported here for C. appendiculata. Although the difference between sexes in overall
48 d evelopment time is not significant at the highest food levels, mean male development time is shorter than that of females in all food treatments and the difference becomes significant at the lower food levels. Since males have lower thresholds, the difference in development time between sexes becomes more pronounced (and therefore significant with small sample size) in the more limiting food environments. A similar result found by Kleckner et al. ( 1995) was the presence of density dependent protandry in Ae. sierrensis meaning that protandry increased at higher mosquito densities. This suggested that protandry was not solely a result of sexual selection for mating but partially due to differences in developmental thresholds due to differing target adult body sizes Although the quantities of food consumed are difficult to record for mosquito larvae grazing on microorganisms, our results of female midges consuming more food daily than males fit with some previous observations of higher female weight gain in mosquito larvae (Brust 1967), though faster male weight gain has been recorded in some cannibalistic species ( Lounibos et al. 1996). These effects of higher food consumption and weight gain among females could be a direct effect of the female delay in entering the 4th instar allowing them to grow to larger size. On the other hand, females may simply have a higher metabolism or growth rate. Comparison of male and f emale head capsule sizes may help answer this question. Egg Development Numbers of eggs developed by females in the lab experiments were closely correlated to wing length as has been observed in a number of mosquito species (e.g., Armbruster and Hutchinson 2002) The lack of a lower limit to autogenous egg development may result from too narrow a range of food treatments in the experiment.
49 However, males in the lowest treatment showed poor survivorship suggesting that food levels were nearing the lower limit and the collections of field pupae showed a broader range of egg batch sizes without loss of autogeny. Al though field pupae were collected throughout the year, no evidence was found of a change in reproductive strategy with season and host availability. Field observations of C. appendiculata midges bloodfeeding on frogs suggest that subsequent egg batches may be produced from blood, although some mosquito species are known to produce multiple autogenous egg batches ( O'Meara et al. 1981) Taken together these r esults suggest that C. appendiculata displays plasticity in batch size but not autogenous development strategy for the first egg batch.
50 Table 2 1. Comparison of model regressions of fourth instar development time vs. inverse food consumption for female and male larvae of C. appendiculata. Following Anderson ( 2008), the weight of evidence, wi, is the probability that the given model best fits the data with the sum of all model probabilities equal to one. The evidence ratio is the likelihood of the best model compared with the current model. Weight of Evidence Model Parameters n RSS AICc evidence, w i ratio R 2 Female constant/canalized H, TP 75 190.10 76.09 0 0.3974 0.803 constant/flexible H, TP, K 75 190.10 78.33 2.23 0.1301 3.05 0.803 Food dependent/canalized H, TP, B1 75 187.00 77.09 1.00 0.2410 1.65 0.806 Age dependent/canalized H, TP, B2 75 187.20 77.17 1.08 0.2315 1.72 0.806 Male constant/canalized H, TP 73 71.04 1.21 0 0.4614 0.391 constant/flexible H, TP, K 73 71.04 3.44 2.23 0.1513 3.05 0.391 Food dependent/canalized H, TP, B1 73 70.64 3.00 1.79 0.1882 2.45 0.395 Age dependent/canalized H, TP, B2 73 70.53 2.89 1.68 0.1992 2.32 0.396
51 T0 1/F T0 1/F Constant/CanalizedT0 = H (1/F) + TPConstant/FlexibleT0 = H(1/F) + [TpmaxF/(K+F)]Food-dependent/CanalizedT0 = b1(1/F)2 + H0(1/F) + TPAge-dependent/CanalizedT0 = [H0/(F-b2)] + TP Figure 2 1. Four competing models of development created by Juliano et al. ( 2004 ) to predict the relationship of total development time (T0) to the inverse of daily food consumption (1/F). The Constant/Canalized model predicts no developmental plasticity while the remaining models predict plasticity in either pupation threshold or post threshold period depending on food availability.
52 Food Offered (prey/day) 0 2 4 6 8 10 12 14 16 Mean Food Consumed (prey/day + SE) 0 2 4 6 8 10 Female Male Figure 2 2 Mean daily food consumption ( SE) of female and male C. appendiculata larvae in the 4th instar. Females eat significantly more food/day than males when offered treatments of 5, 6, 8, 10, or 15 prey larvae per day.
53 Figure 2 3. The best fit linear regressions of the rela tionship between development time and inverse daily food consumption of fourth instar larvae of C. appendiculata. The Constant/Canalized model best explains the relationship of development time to the inverse of food eaten per day (1/F). (A) Females display a threshold of 11.80 ( 0.68 SE) prey followed by a 7.47 ( 0.29 SE) day post threshold period while (B) males have a threshold of 6.45 ( 0.96 SE) prey followed by a 6.33 ( 0.27 SE) day post threshold period. ANCOVA results comparing thresholds directly showed that the female threshold is significantly higher than the male threshold (P = 0.02). Lengths of the pos t threshold periods could not be compared directly with ANCOVA. The lowest food treatment is missing from the male graph due to low survivorship. A) fe male: y = 11.80x + 7.47 B) male: y = 6. 45 x + 6. 52
54 y = 211.95x 205.95 R2= 0.7433Winglength (mm) 1.15 1.20 1.25 1.30 1.35 Eggs 30 40 50 60 70 80 90 Figure 2 4. The best fit linear regression relating wing length and the number of eggs developed in an autogenous egg batch.
55 CHAPTER 3 EFFECTS OF TEMPERATURE ON DEVELOPMENTAL TIMING AND THRESHOLDS Introduction Temperature and nutrition are two of the most important factors affecting the timing of an insect larvas transition to the adult reproductive stage and the physical condition of the resulting adult. In the majority of ectotherms studied, higher temperatures result in faster development and smaller adult size ( Atkinson 1994, Van Der Have and De Jong 1996, Forster and Hirst 2012) although larger size at higher temperatures has been found in a small percentage of cases ( Atkinson 1995). A driver o f this effect is temperatures control of the reaction rates of the biochemical processes involved in development ( Van Der Have and De Jong 1996). However, differing environmental temperatures may select for different sized adults or may alter the outcome of tradeoffs in optimal size vs. development time which are inherent in an insects life history strategy ( Stearns 1992). Therefore, if an organism is able to respond to temperatures adaptively, a part or even majority of the change in developmental strategy with temperature may be explained by this plastic response. Among insects, work with a few model species, such as Manduca sexta, has shown that final instar larvae must reach a size thres hold before initiating the events leading to pupation. Between this threshold and pupation is a length of time called the post threshold period ( Nijhout et al. 2006). Although neither the size threshold nor post threshold period can be affected by final instar nutrition in M sexta larvae, several other insect species have shown plasticity in these two traits based on the amount of food available ( Nijhout 2008 ). The Manduca pupation threshold is likewise unaffected by temperature, but the post threshold period is shortened under higher temperatures
56 causing less food consumption and therefore smaller final body size ( Davidowitz et al. 2003). In the larvae of another model organism, Drosophila melanogaster Ghosh et al. ( 2013) recently found that increases in temperature actually lower the size threshold for pupation, reducing food acquisition and therefore causing smaller adult size. This is the only explanation for smaller final size at high temperature since there is no difference between temperatures in weight change during the post threshold period ( Ghosh et al. 2013). The authors argue that this change in threshold with temperature may be evidence for the adaptive value of a smaller adult at higher temperature, though they admit that the change could also be a direct effect of temperature on th e size sensing mechanism. One important difference between the Manduca and Drosophila studies is that Manduca larvae were reared in the same temperature during the early instars and then separated into different temperatures for the final instar while Dros ophila larvae were raised in different temperatures throughout the entire larval period. Since the pupation threshold is often correlated with larval size at the beginning of the final instar, the authors assumed that any effects from the early instar temp erature treatment would be detected in differences in weight. However, no difference in size was observed among Drosophila larvae entering the final instar suggesting to the authors that the difference in development between Manduca and Drosophila was not simply due to a difference in experimental designs ( Ghosh et al. 2013). In this study we used larvae of a corethrellid midge to test for plasticity in developmental thresholds in response to temperature. Larvae of Corethrella appendiculata Grabham are aquatic occupants of container habitats such as water filled
57 treeholes, tires and cemetery vases. Container inhabitants experience variation in temperature throughout the daily cycle and the seasons but are unable to use behavioral changes in depth to modify body temperature to the extent that inhabitants of larger water bodies can (e.g., planktonic Chaoborus midges in lakes ( Bns and Ratte 1991) ) Among the closely related Culicidae numerous studies support the trend of faster development at higher temperatures ( Brust 1967, Trpis 1972, Rueda 1990, Padmanabha et al. 2011, Reiskind and Zarrabi 2012), although fewer examine the mechanisms underlying this pattern. In one such study, Chambers and Klowden ( 1990) used container dwelling Aedes aegypti larvae to show that the cri tical weight threshold for pupation was lower at higher temperatures. As in the study on D. melanogaster larvae in this experiment were reared through all instars in separate temperature treatments, although the size of larvae entering the final instar was not reported. The resulting difference in critical weight may be explained by the faster development of high temperature early instars leading to smaller sized larvae entering the final instar. Alternatively, a mechanism similar to that of D. melanogaster may exist. Although the majority of studies on culicid development focus on groups of larvae, individual development is increasingly important to understand as evidence builds for the influence of larval stress on viral infection of adults. Recent resear ch shows that larval environmental temperature can affect adult susceptibility to viruses ( Westbrook et al. 2010, Alto and Bettinardi 2013). A previous study of C. appendiculata used the models of Juliano et al. ( 2004) to compare developmental times across a range of food levels to determine if plasticity in response to nutrition exists ( Blosser et al. 2013). Results showed that development time
58 was linearly related to the inverse of food consumed suggesting no evidence for plasticity. Using these results we developed three alternative predictions in order to test for plasticity in development with temperature (Figure 3 1 ). Higher temperature is always assumed to reduce development time, so the post threshold period, which is unaffected by food level, is shorter in all three scenarios. In the case where temperature does not affect the pupation threshold (as in Manduca) the relationship between development time and inverse food will be the same across all temperatures resulting in a graph with parallel lines separated by the difference in y intercepts or lengths of the post threshold period. In the case where colder temperatures cause a higher threshold (as in Drosophila ), the cold temperature will have a steeper slope compared with warm temperature. Causes for this situation could include selection for larger body size at colder temperatures or direct action of temperature on the mechanism sensing size. Finally, a third alternative is that colder temperatures reduce the pupation threshold causing a shallower slope when compared with a warmer temperature. This situation could occur if o ptimal tradeoffs in development time and adult size are affected by temperature. For instance, the longer development time at cooler temperatures may increase risk of death in the larval stage reducing fitness. This could select for any mechanism that woul d reduce the pupation threshold with cooler temperature. Using the predicted relationships between development time and food we tested C. appendiculata larvae for plasticity in pupation threshold with change in temperature. Materials and Methods Colony Ma intenance A colony of C. appendiculata was maintained in an insectary at 27 1C, 80% humidity and a 14:10 hour light:dark cycle. The colony began in 2003 with the collection
59 of several hundred larvae from field sites in Vero Beach, Florida and has been supplemented with collections of one hundred field larvae twice yearly in an attempt to reduce genetic differences between the colony and wild populations. Adults kept in insect cages are provided with a dish of water infused with oak leaves to stimulate ov iposition. Further details of colony maintenance are given by Lounibos et al. ( 2008 ). Early Instar Rearing A dish of oak leaf infused water was withheld from the colony cage for one day and then introduced for 24 hours to stimulate oviposition. The collected eggs were removed from the cage and left covered in the insectary until hatching occurred 3 days later. Two hundred freshly hatched (<24 hours old) C. appendiculata larvae were placed in a metal tray (5 by 25 by 35 cm) with 1 liter of water and reared together to the 4th instar. This tray of larvae was given a 0.2 ml scoop of oatmeal and nematode ( Panagrellus sp.) mixture every day until all larvae reached the 4th instar which was identified by head capsule size. All 200 larvae reached the 4th instar between 7 and 9 days after hatching, and 180 of these were randomly selected for use in the experiment. Temperature Treatments The pan of early instars was checked daily for 4th instar larvae which were removed, placed individually in a glass vial and divided haphazardly between high (27.0 0.2C) and low (21.8 0.3C) temperature treatments (mean SD). Within each temperature, larvae were further divided among four prey treatments consisting of 40, 15, 8, or 5 first instar Aedes albopictus (Skuse) larvae each day. Ten C. appendiculata larvae in each temperature were placed into a control treatment which received no food. In an attempt to narrow age differences among prey larvae, Ae. albopictus eggs, aged one to three weeks old, were hatched using a vacuum pump to synchronize eclosion.
60 This was an improvement over techniques from previous experiments in that all prey were guaranteed to have hatched within a few hours and no food was present thereby avoiding the effects of differing growth rates among individual prey. Every 24 ( 3) hours, vials containing 4th instars were dumped into a clear dish and rinsed once in an attempt to recover and count all uneaten prey. Vials were then refilled with 1 3 ml of distilled water and given a fresh set of prey larvae. Leftover prey were recorded as half or whole and live or dead and the date of pupation of the C. appendiculata larva was recorded. Treatments were arranged randomly in the incubator each day usi ng a random number generator. Results Larval Development Of the 160 larvae fed daily prey, 9 died before reaching the pupal stage and 6 of these were from the lowest food treatment (5 prey per day) at 22C. The result of this treatment was the pupation o f 11 males and 3 females with 6 larvae dying before sex could be determined. Assuming an average 50:50 sex ratio, we reasoned that the dead larvae in this treatment were likely females and removed the lowest food treatment from all female analyses. This av oids biasing the results since the surviving 3 females are likely to have faster development with lower pupation thresholds than the nonsurvivors. None of the 20 control larvae pupated by the end of the experiment. No evidence was found for differences in pupation thresholds at differing temperatures when comparisons were made with ANCOVA. In females, differences in slope (pupation threshold) were not significant ( Table 31 ), but post threshold periods differed significantly averaging 5.56 ( 0.45 SE) days at 27 C and 8.59 ( 0.37 SE) days at 22 C ( Figure 3 2 ) The difference in male thresholds was also not significant and the
61 post threshold period at 27 C (4.31 0.30 SE days) was significantly shorter t han the post threshold period at 22 C (6.90 0.27 SE days). When males and females were compared within temperature treatments, females at 27 C had a significantly higher threshold than males ( Table 31 ) with a female threshold of 55.55 ( 7.01 SE) prey and male threshold of 23.28 ( 4.80 SE) prey. Thresholds between sexes at 22 C were not significantly different but female post threshold time was significantly longer than that of males ( Figure 3 3 ). P rey Consumption Differences in mean daily prey consumption could not be compared in the lower food treatments since predators usually ate all food offered to them. In the highest food treatment, however, females consumed significantly more mean daily prey than males in both the 27 C (t17 = 3.50; P = 0.0028) and 22 C (t17 = 6.19; P = 0.0001) treatments. mean daily consumption was also higher in the 27 C treatment compared to the 22 C treatment in both females (t15 = 5.21; P < 0.0001) and males (t19 = 7.17; P < 0.0001). Total prey consumption during the fourth instar was again higher in females than males at both 27 C (t17 = 4.34; P = 0.0004) and 22 C (t18 = 12.44; P < 0.0001), but larvae in the low temperature treatment consumed more total prey than those at the high temperature in both females (t16 = 6.93; P < 0.0001) and males (t19 = 2.97; P = 0.0079) ( Figure 3 4 ). Discussion These results suggest that C. appendiculata larvae are similar to Manduca in that temperature does not affect the final instar pupation threshold but governs the length of the post threshold period ( Nijhout et al. 2006). At higher temperature, the development time in the fourth instar is shortened by both higher feedi ng rates leading to more rapid
62 attainment of the threshold and a shorter post threshold period, perhaps related to the increased rate of biochemical reactions. If this effect is consistent in earlier instars, then larvae reared at a higher temperature throughout development would enter the final instar at a smaller size likely resulting in a smaller pupation threshold than those raised at cooler temperatures throughout development. The results of this experiment suggest that these three effects: feeding rat e, post threshold period and threshold (set at the beginning of the instar), are the main ways in which temperature affects development time in C. appendiculata. One possible objection to these conclusions could be that the power of the experiment was not strong enough, leading to the selection of the default position which predicts no temperature effect on the threshold. Factors such as the experimental temperature and food ranges need to be large enough to allow for detection of any treatment effects. In the present experiment, the temperature range was only 5 C but led to an approximately 2.5 day (~50%) increase in the duration of pupal development and post threshold periods at the colder temperature. The food range was unfortunately narrowed by larval deaths in the lowest food level treatment. However, this suggests that the range extended from the near lowest food level to the highest, since differences in daily consumption at the highest food level indicate that excess food was available. The significa nt difference in threshold between sexes at 27 C indicates that differences were detectable even with a truncated food range, but the lack of a threshold difference between sexes at 22 C is more difficult to explain. It is also interesting to note that whi le the thresholds of males were not significantly different (P = 0.07) they were close, and the difference could even be considered marginally significant. This was the only
63 comparison that included all four food levels and may therefore have been more statistically powerful. The result would suggest that, if anything, C. appendiculata may have higher thresholds at lower temperatures as in D. melanogaster but that the experimental design was not adequate to detect this. A number of assumptions were included in this analysis which may have affected the robustness of the conclusions. First, the procedures used food consumption rather than weight to measure the pupation threshold. Any factors which affect the correlation between food consumption and weight w ill add error to the results. One possibility is a difference between temperatures in maintenance costs. Since larvae at the colder temperature assumedly have both a lower metabolism and a longer development time, it is unclear if a difference in these cos ts exists. Another assumption of the experiment is that the difference in digestion efficiency between temperatures is negligible. This is likely a source of error since Giguere ( 1981 ), using labelled isotopes, found signifi cant effects of temperature on assimilation efficiency in a chaoborid midge. At most meal sizes, maximum differences in efficiency were only 5% assimilation over an 18 C range, although at the largest meal size the recorded difference was closer to 25%. This effect at the largest meal size was largely due to very low temperatures around 6 C. In the current study, the smaller, 5 C temperature range used and avoidance of very low temperatures may have minimized this effect on assimilation efficiency, although some error was likely introduced. Finally, for practical reasons, both temperature treatments were fed at the same point in time each day. This correlates to different gaps in physiological time between feedings for the two temperatures and may have introduced unknown error into the analysis.
64 Daily food consumption by females was higher than by males, as found in a previous study ( Blosser et al. 2013), though this result was not found by Lounibos et al. ( 2008). Larvae at higher temperatures also consumed more per day, supporting results from a number of organisms including some mosquito larvae ( Rashed and Mulla 1989). Although larvae at higher temperature consumed more daily food, their shortened development time led to lower total food consumption in the final instar, especially among females ( Figure 3 4 ). While final size was not measured in the current study, most e ctotherms are smaller at higher temperatures ( Atkinson 1994 ). The few exceptions to this rule are best explained by seasonal constraints ( Atkinson 1995) which seem unlikely to occur in C. appendiculata which pupates in all months of the year ( Blosser et al. 2013). If C. appendiculata are larger at colder temperatures, this could be explained by the lengthened post threshold period allowing more time for feeding and therefore larger total consumption. This leads to the prediction that larvae in high temperatures at nonlimiting food levels will have consumed more total prey when they pupate than larvae in a colder temperature at the same time. If food is withheld in both temperature treatments just after the threshold has been reached, or within the next few days, high temperature larvae will have consumed more total food and may have an associated larger body size than their cold temperature counterparts. This would only occur in an environment with a surplus of food since limiting food conditions equalizes the daily consumption rates between temperatures. This result, however, could possibly be used in experiments to disentangle the effects of size and temperature. In conclusion, no significant evidence was found for temperaturerelated plasticity in the pupation threshold of the larvae of C. appendiculata. Instead, temperature
65 governs the length of development through its effec t on consumption rate, length of the post threshold period and possibly the threshold set at the beginning of the instar (if these results hold for earlier instars. Larvae at colder temperatures consume more total food than larvae at warmer temperatures si mply as a result of their longer post threshold periods.
66 Table 31. ANCOVA for the effects of temperature and sex on development time of fourth instar C. appendiculata larvae with food level as a covariate. Slopes were first compared by testing the significance of the interaction term before repeating the analysis with the interaction term removed. Independent variable Source df F p Temperature Within females Temperature 1 66.34 < 0.0001 Temperature X food 1 0.58 0.4494 Error df 54 Within males Temperature 1 128.10 < 0.0001 Temperature X food 1 3.36 0.0705 Error df 82 Sex At high temperature Sex X food 1 7.57 0.0081 Error df 54 At low temperature Sex 1 69.64 < 0.0001 Sex X food 1 2.00 0.1624 Error df 57
67 Development Time (days) 1/Daily Food C B A Figure 3 1. Hypothetical relationships between development time and inverse food at high temperature (solid line) and low temperature (dotted line). This comparison allows prediction of the effect of temperature on pupation threshold (slope) ( Blosser et al. 2013). A) Temperature has no effect on threshold. B) Colder temperature increases the pupation threshold (more food is needed to trigger pupation). C) Colder temperature decreases t he pupation threshold (less food is needed for pupation). All three options assume that development is longer at cooler temperatures.
68 1/Daily Food 0.000.050.100.150.20 Development Time (days) 468101214161820 22C 27C A) Males 1/Daily Food 0.020.040.060.080.100.120.14 Development Time (days) 468101214161820 22C 27C B) Females Figure 3 2. Best fit linear regression relating the inverse of daily food consumptio n to total development time in the fourth instar of C. appendiculata larvae reared at high (27C) and low (22C) temperatures. Results of an ANCOVA show that in both (A) males and (B) females the temperature does not significantly affect the pupation threshold (the slope of the line). In both sexes, the post threshold period (y intercept) is significantly longer at the colder temperature.
69 1/Daily Food 0.020.040.060.080.100.120.14 Development Time (days) 4681012141618 Male Female 1/Daily Food 0.020.040.060.080.100.120.140.16 Developmental Time (days) 68101214161820 Male Female A) 27 CB) 22 C Figure 3 3. Best fit linear regression relating the inverse of daily food consumpti on to total development time in the fourth instar of male and female C. appendiculata larvae. Results of ANCOVA show that (A) at 27 C, the female pupation threshold is significantly higher than the male threshold. (B) At 22 C, the thresholds are not signif icantly different between sexes, but the post threshold period is longer in females than males
70 Female 22 C Male 22 C Female 27 C Male 27 C Mean Daily Prey Consumed (+ SE) 22 24 26 28 30 32 34 36 38 A) Daily Food Female 22 C Male 22 C Female 27 C Male 27 C Mean Total Prey Consumed (+ SE) 140 160 180 200 220 240 260 280 300 320 B) Total Food Figure 3 4. Mean prey consumption ( SE) in fourth instar C. appendiculata larvae when offered 40 prey per day at high (27C) or low (22C) temperature. A) Females consume significantly more daily prey than males and larvae in the 27 C treatment consume more daily than those of the same sex in the 22 C treatment. B) Females consum e more total prey during the final instar than males and, within the sex, larvae at 22 C consume more total prey than those at 27 C.
71 CHAPTER 4 EFFECTS OF PENULTIMATE INSTAR ENVIRONMENT ON DEVELOPMENTAL TIMING AND ADULT FITNESS TRAITS OF CORETHRELLA APPENDICULATA Introduction Changes in the larval nutritional environment can determine the timing and size of an insect at pupation as well as the fitness of the resulting adult ( Stearns 1992, Nyl in and Gotthard 1998). Insect species have adapted their developmental strategies to fit the larval environments that they most often encounter, and species developing in variable environments are most likely to have the plastic development necessary to ma tch environmental cues ( Nijhout 2008). The two best studied models of insect development, Manduca sexta and Drosophila melanogaster have recently been shown to have differing developmental strategies under variable nutritional and thermal environments. M. sexta caterpillars enter their final instar with a pupation weight threshold which is unchanged by food level or temperature and is followed by a post threshold period during which food level has no effect on development tim e ( Davidowitz et al. 2003, Nijhout et al. 2006). In contrast, D. melanogaster larvae adjust their pupation weight threshold with temperature and their post threshold period with food avai lability ( Ghosh et al. 2013, Layalle et al. 2008). Previous experiments with the aquatic larvae of the frog biting midge Corethrella appendiculata Grabham have shown that this container dwelling species has a developmental strategy similar to that of M. sexta in the final instar ( Blosser et al. 2013). Since the pupation threshold is set prior to the beginning of the final instar, inter instar effects must be an impor tant factor in the development of C. appendiculata Research on M. sexta caterpillars has shown that lower nutrition in early instars leads to a smaller sized final instar caterpillar (measured by head capsule size) correlated with a lower
72 size threshold for pupation ( Nijhout 1994). If all else is equal, this suggests that larvae with a low food history will reach their pupation threshold more quickly in the final instar than larvae with a high food history under similar final instar conditions. However, a number of studies suggest the existence of a developmental inertia or carryover effects from previous instars during development ( Jones 1993, Bradshaw and Johnson 1995, Flanagin et al. 2000). Bradshaw and Johnson ( 1995) used a foodswitching experiment including all 4 instars of the pitcher plant mosquito to show that final instar food level had no significant effect on development time or final size of the larvae. Instead, these traits were controlled by the nutrition level of earlier instars, although some food was necessary in the final instar. A similar food switching study by Flanagin et al. ( 2000) again found that final instar food level had no effect on the development time of lubber grasshoppers and suggested that the hormonal events triggering the adult molt may already be in motion by the start of the final instar. In contrast, a follow up study focusing solely on the final instar of lubber grasshoppers found a large effect of final instar food level on development time ( Hatle et al. 2003). The experiment in Chapter 1 used the models developed in Juliano et al. ( 2004) to test for plasticity in final instar C. appendiculata larvae and found that development time was linearly related to the inverse of daily food consumed with the slope representi ng the pupation threshold and the y intercept as an estimate for the post threshold period. We used this relationship to compare the thresholds and post threshold periods of larvae reared under identical final instar conditions, but experiencing low or high food treatments in the previous instar. If lower food in the penultimate instar causes a larva to enter the final instar at a smaller size, this may
73 result in a smaller pupation threshold for the low food treatment. A smaller pupation threshold could als o be explained as a plastic response allowing the larva to escape a poor larval environment earlier. If, however, a type of developmental inertia exists in C. appendiculata development, we would expect either a higher food threshold or longer post threshold time in larvae with a history of low nutrition. A final alternative is that the nutrition of previous instars has no effect on the final instar threshold or post threshold period, though this seems unlikely. A second experiment used the results of the first experiment to design a foodswitching protocol for testing some of the findings including differences in thresholds and the length of the post threshold period. This second technique can give support to the results of the first experiment by avoiding some of the assumptions inherent in the model including assumptions of model linearity and identical assimilation across the final instar food treatments. Head capsule widths were also used to determine if a lower food treatment led to lower final instar size. Materials and Methods Colony Maintenance A colony of C. appendiculata originating from field collections from Vero Beach, Florida was maintained in an insectary at 27 1C, 80% humidity and a 14:10 hour light:dark cycle. Larvae were reared in group pans with biweekly prey additions, and adults were housed in an insect cage with a cup of oak leaf infused water as an oviposition stimulant. Details of colony maintenance are given in Lounibos et al. ( 2008). Early Inst ar Food Trials Preliminary experiments were performed rearing C. appendiculata first, second and third instars individually at a range of daily food levels using either nematodes or
74 Aedes albopictus (Skuse) larvae as prey. The purpose of these trials was t o determine the feasibility of rearing early instars and the consistency and range of development times that could be achieved. Adult nematodes could be counted individually under a strong light against a black background. Results of nematode feeding tri als showed that third instar development time averaged between 2.60 ( 0.84 SD; n = 10) days with 20 nematode prey daily and 23.0 ( 8.12 SD; n = 10) days with 1 nematode daily. Second instar development time averaged between 2.25 ( 0.49 SD; n = 10) days at 10 daily nematodes and 5.51 ( 1.65 SD; n = 10) days at 1 daily nematode. First instar C. appendiculata midges were too small to consume adult nematodes, though they may consume young nematodes which are difficult to count individually. Attempts at rear ing first instars under varying levels of oatmeal food resulted in successful molts, but development time varied widely and water clarity became a problem. Since no consistent method was found for rearing first instars and nematode prey were more difficult to manage, the following experiments focused only on the effect of nutrition in the third instar using the larvae of Ae albopictus as prey. High and low food levels were selected to cause a clear difference in third instar development time while minimizi ng variability. Experiment One Early instar rearing An oviposition cup was placed in the adult colony cage for 24 hours to collect eggs of similar age. This period was preceded by a day with no oviposition site available in the cage, increasing the number of females ready to lay eggs during the target period. Eggs were removed from the cage and set aside for three days until hatching occurred. On the third day, approximately 220 first instar C. appendiculata which had
75 hatched within a 24 hour period, were placed in a metal tray (5 by 25 by 35 cm) in 1 liter of water and reared through the first 2 instars. The larval tray was given a daily scoop (0.2 ml) of oatmeal and nematode ( Panagrellus sp.) mixture as food and checked for third instars every 4 hours af ter molting had begun. Third instars were differentiated from second instars by head capsule width, and accuracy was improved by removing all larvae from the rearing tray to compare head capsules during each check. Third instar treatments Third instar larv ae were separated into individual vials and divided randomly between high (15 prey/day) and low (3 prey/day) food treatments. Ten larvae were controls receiving no food to determine if development to the fourth instar was possible without additional food. Of the 223 larvae, 186 molted to the third instar within a 29 hour period and were given their first food treatment together at the end of this period. This meant that larvae used for the experiment encountered a variable amount of time with no food (betw een 0 and 29 hours) prior to the initial third instar meal. Therefore, each collection of third instars was divided evenly between food treatments to avoid any effect of this variability in starvation time on the resulting difference between treatments. Bo th high and low food treatments had 88 replicate larvae in total. The unused larvae consisted of 5 that molted before the 29 hour window, 25 larvae that molted later and 7 larvae that were still 2nd instars a day later. Prey for third instar C. appendicul ata larvae were Ae albopictus larvae that were hatched from eggs over a 2 hour period using a vacuum pump. Egg age was limited between 1 and 3 weeks to minimize potential effects on prey quality. The Ae albopictus colony was initiated with the collection of several hundred larvae from artificial containers in Vero Beach, Florida.
76 Each day, third instar C. appendiculata were removed from their vials, leftover prey were counted, and vials were refilled with 13 ml of distilled water and a fresh set of pre y. Once molting to the fourth instar began, all third instars were checked every 3 hours for molting and fourth instars were removed from prey and assigned to one of 5 food treatments, although the initial meal was not given until the appropriate time of day. Preliminary testing showed that fourth instar C. appendiculata larvae did not eat prey in the first 3 hours after molting, although a few prey were eaten after 4 hours. The rapid removal of prey from freshly molted fourth instars insured that larvae fr om the high food treatment did not get an extra meal which was not available to larvae from the low food treatment. Final instar treatments Fourth instar C. appendiculata were divided randomly into 5 food treatments of 40, 15, 9, 6, and 5 prey per day. As in the third instar, leftover prey were counted daily and fresh water and prey were provided each day. Larvae were checked daily for pupation, and vials with pupae were covered with mesh to prevent adult escape after emergence. The relationship between amount of food consumed and development time was compared between high and low food treatments and between sexes using ANCOVA. All ANCOVAs tested for equality of slopes and ANCOVAs with common slopes were retested with food consumption as a covariate to test for equality of intercepts. Data were transformed to fit assumptions of normality when necessary. Fitness traits of adult midges were compared between treatments to determine if carryover effects from the penultimate instar affected adult fitness in addit ion to development time. After emergence, adults were given daily sugar soaked cotton until the fifth day when eggs were fully developed. Adults were then killed with chloroform,
77 wings were removed and photographed and females were dissected to count the number of eggs developed. During egg counts, it was also necessary to check the vials for floating eggs since some females dropped a portion of their egg batch before the fifth day. A subsample of females was selected for egg size comparisons. Among these females, 10 eggs were selected from each individual and photographed. A computer imaging program (i Solution lite, AIC, Princeton, NJ) was used to measure wing lengths and egg lengths. Wings were measured from a closed cell near the base of the wing to the outer edge of the wingtip excluding the fringe as in Blosser et al. ( 2013). Analysis of covariance was used to compare the relationship of wing length to egg n umber between food treatments w hen slopes were found to be common. T tests were used to compare the numbers of eggs developed and egg size between food treatments. Experiment Two Using the results of Experiment One, a foodswitching experiment was designed to test some predictions of the linear model using a separate technique. This experiment repeated the procedures of Experiment One for the first 3 instars, but switched to low food levels at several points during the final instar to test if the pupation threshold (after which food level does not affect development time) had been reached. This experiment aimed to estimate the approximate length of the post threshold period and to v erify differences in threshold between sexes and between larvae raised in high and low food treatments during the previous instar. Early instars The same methods were used as in Experiment One where approximately 220 first instar C. appendiculata were rai sed together in a pan until reaching the third instar. Out of this pan, 206 larvae molted to the third instar within a 36 hour period and were
78 used for the experiment. The third instars were again divided into high (15 prey/day; n = 98) and low (3 prey/day ; n = 98) food treatments and checked every 3 hours for molting to the fourth instar. Third instars vials were daily given fresh food and distilled water. Final instar food switching method Fourth instar larvae were fed 9 prey daily until the date selected for food reduction at which point larvae were switched to 2 prey daily until pupation. Results from Experiment One predicted a 2.8 day difference in development time between females from the high and low food treatments and a range of 4.04 to 4.22 days as an estimate for the post threshold time of all larvae. Based on these estimates and development times from Experiment One, a food switching schedule was created in which larvae from the high food treatment (15 prey/day in the third instar) were switched f rom 9 prey/day to 2 prey/day on days 4, 7, 9, or 11, or placed in a No Switch group which received 9 prey/day until pupation. Similarly, larvae from the low food treatment (3 prey/day in the third instar) were scheduled to switch on days 7, 9, 11, or 13, or placed in the No Switch group with no switch in food. A total of 196 larvae were divided between these treatments meaning that approximately 20 larvae were in each of the 8 switching groups and the 2 Max Food groups. A control treatment receiving no food contained 10 larvae. During the course of the experiment, we realized that development was proceeding more slowly than in Experiment One. In order to ensure that the thresholds were captured in the results, several food switching dates were delayed by one day. The resulting food switching schedule used for this experiment reduced food on days 4, 7, 9, and 12 for the high food treatment and days 7, 9, 12, and 14 for the low food treatment. Using ANOVA, the mean development times from the food switching grou ps were compared with the mean development time of the No Switch groups to
79 estimate the point at which the larvae crossed the pupation threshold (and food level no longer affected development time). Head capsule widths Head capsule widths of fourth instar larvae were measured using microscope photography (Leica DVM 5000) and computer measurement tools. Width measurements were taken at the widest point of the head capsule at a point in the middle of a row of spines on each side near the base of the head. Mean head capsule widths were compared between males and females and between larvae from high and low previous instar food treatments using t tests. Results Experiment One Development time in the third instar was significantly faster (t126 = 14.66; P < 0.0001) in the high food treatment (3.21 0.41 [SD] days; n = 86) than in the low food treatment (4.67 0.83 [SD] days; n = 87). None of the larvae in the control treatment with no food molted to the fourth instar. Fourth instar development Results of equal slope tests ( Table 41 ) showed that larvae from the low food treatment in the previous instar developed more slowly with a significantly higher pupation threshold than larvae from the high food treatment in both males (P = 0.005) and females (P = 0.03) ( Figure 4 1 ). Males with high food had a threshold of 49.6 ( 4.9 SE) prey and males coming from the low food treatment had a threshold of 63.8 ( 3.6 SE) prey. Females from the high food treatment had a threshold of 82.2 ( 7.1 SE) prey and females with low food had a threshold of 97.6 ( 5.1 SE) prey. Females
80 also had a significantly higher threshold than males in both the high food treatment (P < 0.0001) and low food treatment (P < 0.0001) ( Table 41 Figure 4 2 ). When offered the highest food treatment of 40 prey per day, females in the fourth instar consumed significantly more daily food than males, though there were no significant differences in daily food consumption between larvae from high and low previous instar food treatments ( Figure 4 3 ). Females also consumed significantly more total prey in the fourth instar when compared to males from the same treatment. Fourth instar larvae originating from the low previous instar food treatment consumed more total prey than larvae from the high food treatment ( Figure 4 3 ). Adult traits Results showed no significant differences in adult fitness charact eristics when larvae coming from high and low food treatments were compared. The relationship of wing length to number of eggs developed in females did not differ between high and low food treatments according to the results of ANCOVA analysis ( Figure 4 4 ). There was no difference between high vs low food treatments in the mean number of eggs developed at any fourth instar food level ( Figure 4 5 ). There was also no significant difference in mean egg length between females raised at high (233.27 7.11 [SD] m) and low (234.16 9.29 [SD] m) food level during the third instar (t30 = 0.08; P = 0.94). Experiment Two Development time in the third instar was significantly faster (t182 = 23.84; P < 0.0001) at the high food level (3.40 0.49 [SD] days; n = 102) than at the low food level (5.41 0.69 [SD] days; n = 102). No larvae developed to the next stage in any of the control treatments which did not receive food.
81 Head capsule size Fourth instar head capsul e width did not differ between larvae from high and low food treatments in the third instar. However, head capsule width was significantly smaller in males than in females ( Figure 4 6 ). There was little overlap between sexes in head capsule width. Estimating the smallest 50% of larvae to be males and the rest to be females would result in 92.3% accuracy in the high food treatment and 82.0% accuracy in the low food treatment. Development time Females from both the high and low food third instar treatments had an estimated pupation threshold falling between Day 9 and Day 12 of fourth instar development ( Figure 4 7 ). Females that received no food switch averaged 13.0 ( 1.1 SD) days in the high food treatment and 16.1 ( 1.2 SD) days in the low food treatment. This results in a predicted post threshold period between 1 and 4 days in the high food treatment and between 4.1 and 7.1 days in the low food treatment. Males from the high food treatment had an estimated threshold between Day 4 and Day 7 of fourth instar development ( Figure 4 8 ) while the threshold for males from the low food treatment was estimated between Day 7 and Day 9. High food males averaged 10.9 ( 1.0 SD) days for dev elopment while low food males averaged 12.6 ( 1.3 SD) days. Post threshold period estimates ranged between 3.9 and 6.9 days for high food males and between 3.6 and 5.6 days for low food males. Discussion In both experiments, fourth instar larvae derived from a low food treatment in the previous instar developed more slowly than larvae coming from a high food treatment. This is similar to the developmental inertia found in previous studies ( Bradshaw and
82 Johnson 1995, Flanagin et al. 2000), although less severe since food levels in the final instar were still the more important determinant of development time. The larger inertia effect in previous studies is likely related to the use of all earl y instars rather than only just one and could possibly be increased in the present experiments by a larger gap between high and low food treatments. The longer development time in the current study appears to be due mainly to a larger food threshold neces sary for pupation following low food ( Figure 4 1 ). Why would a larva coming from low nutrition have a higher food threshold? One possibility is that the higher threshold is simply an artefact of the experimental setup. Some previous experiments by other authors and an earlier version of the current experiment used a 24 hour schedule to check for larvae which had molted to the next instar. Under this setup, freshly molted larvae in a high food treatment are left with a partial day t o consume any leftover prey from their previous instar. This unrecorded meal would not be available for larvae in a low food treatment which had already eaten all their food, and could give a false impression of a higher threshold in the low food treatment By checking for molts every 3 hours in these experiments, this explanation can be excluded. Another possible explanation is that the higher threshold is adaptive, allowing the larva to catch up gathering extra nutrients in the final instar to make up for losses in the previous instar. Compensatory growth is known in tadpoles which can switch between slow growth with fast development and fast growth with slow development ( Rose 2005). Tadpoles from a restricted diet can show faster weight gain than controls once returned to normal diet and may reach a similar or larger final size than controls given no diet restriction ( Hector et al. 2012). Some supporting evidence for this
83 explanation is the lac k of difference in wing lengths number of eggs developed and egg size when compared between food treatments ( Figure 4 4 Figure 4 5 ). Despite the reduced feeding during the third instar, larvae from the low food treatment showed no evidence of loss in adult fitness, although compensatory growth can reduce fitness later in life due to increased cellular damage ( Hector et al. 2012). However, in M. sexta at least, a higher threshold is correlated with a larger head capsule and there was no evidence for differences between food treatments in C. appendiculata head capsule widths ( Figure 4 6 ). A third explanation for the apparently larger food threshold in larvae from low food conditions is a change in food assimilation efficiency. If previous instars reared in low food environments reduce their food assimilation efficiency, then more food would be required to reach the same level o f nutrient storage or same critical weight threshold. Evidence of a change in food assimilation efficiency has been found in a number of studies although not all agree about the direction in which it is affected. Some optimal digestion models predict that larvae under low food should increase assimilation efficiency ( Scriber and Slansky 1981) and several studies support this pattern ( Lepczyk et al. 1998, Flanagin et al 2000) In contrast, Bradshaw and Johnson ( 1995) found that when food was enhanced (low food in early instars and high food in later instars), the yield (pupal weight/food weight) was decreased compared with the reverse swit ch. However, these results are difficult to interpret since the majority of food is consumed in the final instar and the differences in yield may simply be a result of differing final instar food treatments. As a further complication, the differences in as similation efficiency between larval instars may confuse some of these previous results ( Sears et al. 2012).
84 If differences in assimilation efficiency explain the apparent difference in thresholds, then it is difficult to explain why there is no difference in wing length and egg production between treatments. Possibly, the difference between the high and low food treatments was not large enough to produce significant differences in adult fitness traits, although the differences in development time were significant. A test of the assimilation explanation could be performed to measure differences in yield (weight gain of larvae per weight of food consumed) following the methods of Bradshaw and Johnson ( 1995). The Corethrella food switching experiment supported several results from Experiment One including the differences in threshold between sexes and the differences in threshold between males from high food and low food treatments ( Figure 4 7 Figure 4 8 ). Although no post threshold period length will satisfy all four estimates, 3 of the 4 estimates agree that the post threshold period is greater than 3.6 days, and the fourth estimate (1 4 days) s uggests that it is closer to 4 days. This rough approximation agrees closely with predictions from Experiment One, although longer periods were predicted from previous experiments ( Blosser et al. 2013). The difference in thr esholds between high and low food treatments within females was not detected by the food switching experiment ( Figure 4 7 ). This was likely caused partly by the lengthening of the gap to 3 days between Day 9 and Day 12 of food reduction. Development in Experiment Two was slower than in Experiment One causing us to adjust the food switching schedule to ensure that all thresholds were captured within the experiment. This delay in development was likely an effect of differences in rearing procedures in the first 2 instars where scoops of nematode/oatmeal mixture may vary in composition from experiment to experiment.
85 Male head capsule sizes were significantly smaller than female head capsules ( Figure 4 6 ). T his result along with results from a previous experiment ( Blosser et al. 2013) gives a fuller picture of the differences in development between the sexes in C. appendiculata Females develop more slowly than males through the early instars, entering the fourth instar later ( Blosser et al. 2013) and at a larger size (measured by head capsule). Females are then able to consume more daily prey than males due to their larger size ( Figure 4 3 ). The tradeoff between earlier molting and developing to a larger size is likely optimized in each sex to reach the target adult size in the most efficient way. In conclusion, larvae coming from the low food treatment in the previous ins tar need to reach a higher food threshold in the final instar than larvae from a high food treatment. No difference between treatments was detected in fourth instar head capsule size, number of eggs developed, wing length, or egg size. It is unknown whether the higher food threshold is an adaptation to compensate for food lost in the previous instar or whether the difference in food thresholds is actually a difference in food assimilation efficiency. However, given the similarity in adult fitness traits bet ween treatments and the fact that food assimilation is theorized to increase rather than decrease with low food availability, the most likely explanation is that C. appendiculata larvae are compensating for previous low food by increasing their final instar threshold. A second foodswitching experiment supported differences in threshold between sexes, differences in threshold between food treatments within males, and suggested a post threshold period of approximately 4 days. Further experiments testing assi milation
86 efficiency may help resolve the mechanism for the developmental inertia seen in these experiments. Figure 4 1. Best fit linear regression relating the inverse of daily food consumption to total development time in the fourth instar of C. appendiculata larvae from
87 high and low previous instar food treatments. Comparison between slopes shows that in both (A) females and (B) males, larvae coming from a low food treatment in the previous instar have a significantly higher threshold (slope) tha n larvae coming from a high food treatment.
88 Figure 4 2. Best fit linear regression relating the inverse of daily food consumption to total development time in the fourth instar of male and female C. appendiculata larvae. Comparison between slopes shows that in both (A) larvae given high food in the previous instar and (B) larvae given low food in the previous instar, females have a significantly higher pupation threshold (slope) than males
89 Female HighMale HighFemale LowMale Low Average Daily Prey Consumed (+SE) 262830323436 A) Daily Foodbaab Female HighMale HighFemale LowMale Low Average Total Prey Consumed (+ SE) 140160180200220240260 B) Total Foodabcd Figure 4 3. Mean food consumption ( SE) in fourth instar C. appendiculata offered 40 prey/day following high or low food treatment in the previous instar. (A) Females have significantly higher daily food consumption than males after both high food (t15 = 3.56; P = 0.003) and low food (t14 = 5.50; P < 0.0001) treatments. There were no significant differences in daily food consumption between females in high and low food treatments (t15 = 1.14; P = 0.27) or between males in high and low food treatments (t14 = 1.01; P = 0.33). (B) Fe males consume significantly more total food during the fourth instar than males after both high food (t15 = 6.69; P < 0.0001) and low food (t14 = 10.12; P < 0.0001) treatments. Larvae receiving low food in the previous instar consume significantly more foo d than larvae from a high food treatment when
90 compared within females (t15 = 2.90; P = 0.01) and within males (t14 = 3.15; P = 0.007). Winglength (mm) 22.214.171.124.251.301.35 Eggs Developed 30405060708090 High Food Previous Instar Low Food Previous Instar Figure 4 4. Best fit linear regressions of the relationship between wing length and the number of eggs developed by C. appendiculata females given high or low food in the penultimate instar. After comparison of slopes revealed no significant difference (P = 0.11), ANCOVA results, using wing length as a covariate, found no significant effect of previous instar food treatment on number of egg developed (F = 0.42, df = 68 P = 0.52).
91 Daily Food Offered in Fourth Instar ( Ae. albopictus prey) 40 15 9 6 5 Mean Eggs Developed (+ SE) 0 20 40 60 80 100 High Food Previous Instar Low Food Previous Instar Figure 4 5. Mean number of eggs developed ( SE) by female C. appendiculata reared at various fourth instar food levels following high or low food treatment in the previous instar. The number of eggs developed is not significantly different at any fourth instar food level when compared between larvae raised in a low or high food treatment during the previous instar: 40 prey/day (t14 = 1.45; P = 0.17), 15 prey/day (t15 = 1.24; P = 0.23), 9 prey/day (t13 = 1.41; P = 0.18), and 6 prey/day (t14 = 0.32; P = 0.75).
92 Figure 4 6. Mean head capsule width ( SE) of C. appendiculata fourth instar larvae following high or low fo od treatments in the previous instar. The mean width is significantly higher in females than in males in both the high food treatment (t80 = 9.91; P < 0.0001) and in the low food treatment (t85 = 8.92; P < 0.0001), but width does not differ significantly between food treatments within females (t97 = 0.19; P = 0.85) or within males (t76 = 1.02; P = 0.31).
93 Day of Food Switch 7 9 12 14 none Average Development Time (days + SE) 0 5 10 15 20 25 30 Day of Food Switch 4 7 9 12 none Average Development Time (days + SE) 0 5 10 15 20 25 30 F = 16.11 df = 43 P < 0.001 F = 16.44 df = 46 P < 0.001A) Female High Food B) Female Low Fooda a a a a b b b b b Figure 4 7. Mean female development time ( SE) of fourth instar C. appendiculata coming from high or low food treatments in the previous instar followed by a fourth instar food switching schedule. Food is switched from 9 prey/day to 2 prey/day on one of 4 different days or larvae are continuously fed 9 prey/day and there is no switch (labelled none). Developmental times were compared among treatments with ANOVA followed by Bonferroni post hoc, multiway comparisons (P < 0.05). Treatments with the same letter are not significantly different. Means which are not significantly different from the continuously fed treatment are assumed to have passed the pupation threshold, after which food level no longer affects development time. Using this method, the
94 threshold is estimated to occur between Day 9 and Day 12 in both (A) females from a hig h food treatment and (B) females from a low food treatment in the third instar Day of Food Switch 7 9 12 14 none Average Development Time (days + SE) 0 5 10 15 20 25 Day of Food Switch 4 7 9 12 none Average Development Time (days + SE) 0 5 10 15 20 25 30 F = 16.55 df = 41 P < 0.001 F = 3.99 df = 43 P = 0.008 A) Male High Food B) Male Low Fooda b b b b a ab b b b Figure 4 8. Mean male development time ( SE) of fourth instar C. appendiculata coming from high or low food treatments in the previ ous instar followed by a fourth instar food switching schedule. Food is switched from 9 prey/day to 2 prey/day on one of 4 different days or larvae are continuously fed 9 prey/day and there is no switch (labelled none). Developmental times were compared among treatments with ANOVA followed by Bonferroni post hoc, multiway comparisons (P < 0.05). Treatments with the same letter are not significantly
95 different. Means which are not significantly different from the continuously fed treatment are assumed to have passed the pupation threshold, after which food level no longer affects development time. Using this method, the threshold is estimated to occur (A) between Day 4 and Day 7 in males from a high food treatment and (B) between Day 7 and Day 9 in males fro m a low food treatment in the third instar.
96 Table 41. Test for equal slopes among regressions of development time versus inverse food consumed in fourth instar C. appendiculata larvae. Slopes are tested between larvae from high and low previous instar food treatments and between male and female larvae. Comparison Source df Type III SS F P High vs low food treatments Within females Previous instar treatment X inverse food consumed 1 15.94 4.76 0.0324 Within males Previous instar treatment X inverse food consumed 1 17.56 8.22 0.0052 Female vs male Within high food treatment Sex X inverse food consumed 1 0.92 24.45 < 0.0001 Within low food treatment Sex X inverse food consumed 1 83.75 29.11 < 0.0001
97 CHAPTER 5 CONCLUSIONS AND FUTURE STUDIES Larval Development Strategy in Corethrella appendiculata Previous studies of insect development have uncovered a variety of developmental strategies among species. The ability to adjust thresholds and development periods to match environmental conditions may or may not be present in a p articular species and often seems related to factors in the species natural habitat. Water filled container habitats are a well studied system in which food levels, temperature and macroinvertebrate density can vary widely, and sources of mortality such a s habitat drying and predation may favor rapid development to the adult stage. The container dwelling midge, C. appendiculata, was used in the present experiments to investigate how variation in environmental factors affects larval developmental strategy. Larvae of C. appendiculata did not vary their pupation threshold or post threshold period with food availability in the final instar ( Figure 2 3 ) though food level in the previous instar did alter the food threshold necessary for pupation ( Figure 4 2 ). Higher temperature resulted in a shorter post threshold period and shorter time to the threshold as a result of higher daily food consumption ( Figure 3 4 ). However, temperature did not affect the amount of food needed to reach the pupation threshold ( Figure 3 2 ). The effect of temperature on previous instars was not tested. These results suggest that the developmental strategy of C. appendiculata is similar to that of the caterpillar Manduca sexta in which the pupation threshold is unchanged by food availability or temperature in the final instar ( Nijhout et al. 2006).
98 The pupation threshold is instead controlled by environmental factors in previous instars, although it is unclear whether or not this strategy is also similar to that seen in M sexta When the larvae of C. appendiculata are given low food in the penultimate instar, they enter the final instar with a similar head capsule width ( Figure 4 6 ) but higher food threshold than those given high food in the previous instar. In contrast, M. sexta larvae given low nutrition in earlier instar s enter the final instar with a smaller head capsule and lower critical weight threshold than larvae from high nutrition environments ( Nijhout 1994). However, these differences in results may be due to differences in methods and the limited food range used for testing C. appendiculata The differences in high and low food levels may not have been large enough to produce significant differences in head capsule size, or alteration of food in multiple instars may be necessary to observe this effect. Additionally, the present study with midges used a food threshold to measure the point at which pupation events were initiated, whereas studies of M. sexta used a critical weight threshold that correlated closely with measured horm onal events. These two estimates of the actual threshold mechanism may have become decoupled making comparison between these species difficult. However, the carryover effect of early instar nutrition on final instar development time is similar to effects seen in other insects ( Jones 1993 Bradshaw and Johnson 1995, Flanagin et al. 2000), though the m echanisms involved are unknown. This contrast between development al patterns of M. sexta and C. appendiculata may be explained by differences in food availability in larval habitats. Caterpillars running low on food may have low prospects of finding more food in the future due to the difficulty of moving to a new host plant. In contrast, midges and mosquitoes liv ing in aquatic container habitats
99 encounter food pulses when a new cohort of mosquito prey hatch or a leaf falls into the container. These midge larvae may have developmental strategies adapted to food pulses which allow larvae to make up for previous peri ods of low food. A number of traits, such as thresholds and post threshold periods, were measured by multiple experiments and can be compared between experiments to determine the consistency of these estimates. Females had a higher pupation threshold in Ch apter 4 results (82.2 [ 7.1 SE] prey) than in Chapter 3 results (55.55 [ 7.01 SE] prey) and males showed a similar trend with a Chapter 4 pupation threshold of 49.6 ( 4.9 SE) prey and a Chapter 3 threshold of 23.28 ( 4.80 SE) prey. Chapter 2 could not be compared since a different size of prey was used. Post threshold periods also varied among experiments with female estimates at 27 C ranging between 4.19 days (Chapter 4) and 7.47 days (Chapter 2) and male estimates between 4.22 days (Chapter 4) and 6.52 days (Chapter 2). These differing results are likely to be at least partly explained by differences between experiments in early instar rearing conditions. Each experiment began with a pan of midge larvae raised in a group through the early instars. Alth ough these pans received a measured scoop of nematode/oatmeal mixture, this mixture was variable in quality among experiments due to oatmeal aging and changes in nematode population size. This effect was clearly seen in Chapter 4 where larvae in Experiment 2 developed more slowly than those of Experiment 1. The differing post threshold periods are more difficult to explain since temperature was similar between experiments (approximately 27 C). However, there is some evidence from the foodswitching experiment in Chapter 4 that suggests that post threshold period may
100 also be affected by early instar food, although this result is not well supported and needs more rigorous confirmation. One question about any study performed in the laboratory is the extent to w hich the results would hold in natural field conditions. A number of factors may have affected the results of the experiments reported above. Although the natural diet of C. appendiculata is not well known, a study of a cooccurring predator in the same co ntainer habitats, Toxorhynchites rutilus showed that a wide variety of prey was consumed ( Campos and Lounibos 2000). A consistent prey unit was necessary for the current experiments, but the variety in nutrient levels among natural prey likely affects developmental timing in the field. Differences between prey species in vulnerability to predation has been shown in this system and is also likely to affect development time under natural conditions ( Griswold and Lounibos 2006, Alto et al. 2009). Additionally, midge larvae in the field encounter daily temperature variation, habitat structure, and differing water composition and are often found in groups rather than singly Although these factors were removed in the laboratory to isolate the effects of nutrition and temperature, they may be important in nature and modify the results found in the laboratory studies. One attempt was made to relate the lab experiments to field conditions by collecting C. appendiculata pupae from field habitat during every month of the year. The resulting wing sizes and the numbers of eggs developed showed that adult characteristics from Chapter 2 fell within the middle to upper size range of th e samples taken from the field. Further experiments in field conditions would provide more realistic results while increasing variability and difficulty in detecting differences between treatments.
101 The difference in developmental strategy between sexes may be explained by two competing hypotheses: differing selection on optimal final body size or differing selection on optimal timing of adult emergence. When reared at high food level, females molted to the final instar significantly later than males (Chapter 2) presumably allowing the collection of more food in the penultimate instar and resulting in significantly larger final instar larvae (as measured by head capsule width; Figure 4 6 ). These larger female larvae were able to cons ume more daily food than males ( Figures 2 2 3 4 4 3 ) reaching the pupal stage later with more total food consumed in the final instar ( Figures 3 4 4 3 ). Under the first hypothesis, tradeoffs between development time and size (correlating to food capacity) may be optimized in each sex to reach its target size most efficiently. Females would require a higher nutrient threshold to finance egg production. This would suggest that protandry in C. appendiculata may be explained mostly as a byproduct of differing optimal developmental schedules. A similar conclusion was drawn for the larvae of Aedes sierrensis based on studies of prot andry under various levels of intraspecific density ( Kleckner et al. 1995). Alternatively, earlier pupation in males may be adaptive by increasing mating opportunities as suggested by research on Toxorhynchites species ( Lounibos 1996) and developed in studies of Lepidoptera ( Wiklund and Solbreck 1982, Wiklund et al. 1991). In this scenario, C. appendiculata males could have evolved their developmental strategy to optimize size given the necessity of an earlier emergence time. This strategy might be more likely if females mate only once in this species, although this is unknown. The experiments presented in
102 this dissertation cannot differentiate between these two explanations, of which, either is possible or a combination of the two. Future Studies Carry over Effects from Early Instars The results from Chapter 4 have several possible explanations which should be explored further. Delayed development following low food in previous instars has been recorded in a number of insects ( Jones 1993 Bradshaw and Johnson 1995 Flanagin et al. 2000) a nd may represent a widespread phenomenon. One possibility is that a delay allows the larva to collect more resources and catchup from the previous instar. This type of compensatory growth has been studied in amphibians ( Hecto r et al. 2012) and has been previously found in flesh flies during egg development. Under low food conditions, flesh flies attempting to reach a reproductive threshold show an adaptive delay during which they can gather more resources to increase egg production ( Hahn et al. 2008, Wessels et al. 2011). Since amphibians showing compensatory growth have hidden costs connected with their accelerated growth rate ( Hector et al. 2012 ), aquatic dipterous larvae exhibiting accelerated growth should also be tested for hidden costs that do not show up in reduced body size and egg production. These costs could include reduced longevity or strength of immune system with possible consequences for vector competence. An alternative explanation for the delay in development observed in Chapter 4 could be reduced food assimilation efficiency or metabolism following low food. This could create a higher food threshold simply because the l arva requires more food to reach the same critical weight required for pupation. This is the reverse of some model predictions of assimilation efficiency in which larvae with low food increase assimilation
103 efficiency ( Scriber and Slansky 1981) but agrees with some evidence from mosquitoes ( Bradshaw and Johnson 1995). Changes in food assimilation could be tested using the technique of Bradshaw and Johnson ( 1995) in which the yield (pupal weight/food weight) was tracked. The final instars of C. appendiculata from low and high food treatments could be reared under identical food regimens with subsamples killed and dried at several time points to compare dry weight gai n over time between treatments. These experiments would also benefit from creating a larger difference between high and low food treatments in the third instar and by including low food treatments in earlier instars. Previous studies with other species ( Bradshaw and Johnson 1995, Flanagin et al. 2000) have shown a much larger carry over effect when using all early instars. This would exaggerate any differences in assimilation efficiency and provide a stronger test for the changes in head capsule size predicted by work on M. sexta that were not observed in the Chapter 4 experiments. Further studies could investigate the carry over effects resulting from rearing multiple early instars at different temperatures rather than focusing solely on the final instar as in Chapter 3. If all instars respond similarly to high temperature, then shorter post threshold periods in the earlier instars should lead to smaller size and a smaller pupation threshold in the final instar under higher temperatures. Although this was not the case in Drosophila melanogaster ( Ghosh et al. 2013), Chambers and Klowden ( 1990) found that the pupation threshold was lower at high temperature in Aedes aegypt i, possibly due to carry over effects from previous instars. Group Development The experiments presented in this dissertation focus on individual development in an effort to isolate the effects of environmental factors and compare results with
104 previously studied model insect species. However, in C. appendiculata and in many container dwelling mosquito species, intraspecific interactions are the norm. Larval development strategies may have evolved to depend on cues from o thers of the same cohort. Additionally, interspecific interactions may affect an individuals efficiency in capturing and devouring prey. For instance, Kesavaraju et al. ( 2007) found that some container dwelling mosquito larvae respond to predation by C. appendiculata by reducing activity to reduce risk of capture. The extent of this reduction in activity is correlated with the concentration of cues from the predator ( Kesavaraju et al. 2007). In groups, these midges may therefore interfere with each other directly (through aggression) or through indirect interference caused by predatory cues creating a type of ratio dependent predation ( Abrams and Ginzburg 2000, Schenk et al. 2005). Alternatively, C. appendiculata may engage in some form of facilitation in which development is enhanced in groups. Intriguingly, this species can be found at high densities in treeholes ( Lounibos 1983) but does not engage in high levels of cannibalism like some other predators from the same habitat. Future studies could focus on the effects of cohort development by first comp aring the development time and size of larvae raised individually with groups of larvae at various densities using a consistent amount of food per individual. If group development can be simply predicted from the results of the previous experiments on indi viduals, then there should be no difference between treatments. If, however, individuals in groups either interfere with or facilitate each other or trigger changes in
105 developmental strategies, then differences in mean development time or size would emerge. Drying and Predation Although food availability and temperature are two of the most important factors affecting development, larvae developing in temporary water habitats are also threatened with habitat drying and the arrival of predators. These factors have not been as thoroughly studied, but research suggests that some organisms are able to respond to these threats by altering developmental timing. The effects of habitat drying on development have been examined in mosquitoes ( Juliano and Stoffregen 1994) but are mo re thoroughly known in spadefoot toads where desiccation stress triggers corticotropinreleasing hormone leading to the acceleration of metamorphosis ( Denver 1997, 1998). Developmental effects due to predation have also been investigated in mosquitoes ( Hechtel and Juliano 1997) and several studies show that C. appendiculata larvae may be at risk from predation of t he co occuring predator, T. rutilus ( Lounibos 1985, Griswold and Lounibos 2006). The effects of both habitat drying and predation could be further tested with individual C. appendiculata larv ae to determine how accelerated development may be accomplished in these situations. Comparison among Dipteran Species The larvae of C. appendiculata face many of the same environmental variations and risks as a number of related species of midges and mosquitoes inhabiting containers or other temporary aquatic habitats. Fluctuation in food availability, competitor density, temperature, and threats of d esiccation and predation combine to create unpredictable habitats, though not to the same extent in all species. For instance, although C. appendiculata larvae face the risk of habitat desiccation, the threat is
106 reduced since this species prefers the darker water of more permanent habitats when compared with mosquitoes such as Aedes albopictus and Aedes triseriatus ( Bradshaw and Holzapfel 1983, Lounibos 1983 Hawle y 1988). The desiccation resistant eggs of these mosquito species often hatch into shallow, temporary collections of rainwater requiring rapid development to avoid death. Adaptive developmental plasticity may be more likely in these species encountering th e very unpredictable habitats into which their eggs hatch. Further evidence of adaptation to larval habitats comes from the comparison of post threshold periods between species from more and less temporary environments. The post threshold periods of C. a ppendiculata (4.0 7.5 days) and Toxorhynchites brevipalpis (around 7 days) ( Lounibos 1979), which share a preference for more permanent containers ( Bradshaw and Holzapfel 1983), are much lon ger than the 1.6 day post threshold period of Ae aegypti ( Telang et al. 2007) that encounters higher risks from drying and predation. The pupal duration is similarly related in these species with Ae aegypti pupae lasting 2 days while C. appendiculata (5 6 days) and Toxorhynchites amboinensis (~5 days) remain pupae for far longer ( Lounibos et al. 1996). Since a long pupal period is unfavorable from a lifehistory perspective, it seems likely that the length of the pupal period is an unselected trait which is tied to another adaptive trait. One possibility is that the duration of the pupal stage is linked to the length of the post threshold period, considering their correlation in the few speci es that can be compared and the similar hormonal mechanism s controlling their development. In this scenario, a long post threshold period could be beneficial to a species like C. appendiculata which faces high unpredictability in food availability and competitor
107 density along with lower risk of habitat desiccation. The evolution of a high food threshold would allow high growth in good food environments, but be detrimental in environments with poor food where the larva may never be able to pupate. In contras t, a relatively low food threshold followed by a long post threshold period still allows a larva to store up reserves in good food conditions, but guarantees pupation even under fairly poor food environments. Since C. appendiculata larvae are unable to var y their pupation threshold in response to nutrient availability, a long post threshold period may serve as an alternative method of taking advantage of high nutrient environments. In that case, a long pupal stage may simply occur as a sideeffect. This hy pothesis needs further testing by comparing the correlations of pupal duration and post threshold period length to habitat permanence across a range of species. A previous comparison of mosquito species across a range of habitat permanence found that speci es from more permanent habitats tended to have better starvation resistance ( Barrera and Medialdea 1996) and a comparison of two predatory mosquito species from temporary and semi permanent habitats found differences in life history strategy correlated with habitat preference ( Lounibos 2001). A more explicit method of testing these ideas would be to focus comparison on closely related species or subspecies which have recently diverged into temporary and permanent habitats. One such pair of subspecies is the M and S molecular form of Anopheles gambiae. In this mosquito, the M form is adapting to more permanent rice field habitat and slowing its larval development while the S form is adapted to te mporary puddles with faster larval development ( Lehmann and Diabete 2008). Experiments with evolving species
108 such as this could help elucidate the connection between habitat permanence and developmental strategy.
109 LIS T OF REFER E NCES Abrams, P. A., and L. R. Ginzburg. 2000. The nature of predation: prey dependent, ratio dependent or neither?. Trends Ecol. Evol. 15: 337 341. Alford, R. A., and R. N. Harris. 1988. Effects of larval growth history on anuran metamorphosis. A m. Nat. 131: 91106. Alto, B. W., and D. Bettinardi. 2013. Temperature and dengue virus infection in mosquitoes: Independent effects on the immature and adult stages. Am. J. Trop. Med. Hyg. 88: 497 505. Alto, B. W., B. Kesavaraju, S. A. Juliano, and L. P. Lounibos. 2009. Stage dependent predation on competitors: consequences for the outcome of a mosquito invasion. J. Anim. Ecol. 78: 928936. Alto, B. W., M. H. Reiskind, and L. P. Lounibos. 2008. Size alters susceptibility of vectors to dengue virus infecti on and dissemination. Am. J. Trop. Med. Hyg. 79: 68895. Anderson, D. R. 2008. Model based inference in the life sciences: A primer on evidence. Springer, New York. Anderson, J. F. 1971. Autogeny and mating and their relationship to biting in the saltmarsh deer fly, Chrysops atlanticus (Diptera: Tabanidae). Ann. Entomol. Soc. Am. 64: 14211424. Armbruster, P. and Hutchinson, R.A. 2002. Pupal mass and wing length as indicators of fecundity in Aedes albopictus and Aedes geniculatus (Diptera: Culicidae). J. Med. Entomol. 39: 699 704. Atkinson, D. 1994. Temperature and organism sizeA biological law for ectotherms? Adv. Ecol. Res. 25: 1 58. Atkinson, D. 1995. Effects of temperature on the size of aquatic ectotherms: exceptions to the general rule. J. Therm. Biol. 20: 61 74. Barrera, R., and V. Medialdea. 1996. Development time and resistance to starvation of mosquito larvae. J. Nat. Hist. 30: 447 458. Blosser, E. M., N. Nishimura, and L. P. Lounibos. 2013. Testing developmental plastic ity in aquatic larvae of Corethrella appendiculata (Diptera: Corethrellidae). Ann. Entomol. Soc. Am. 106: 810 817. Borkent, A. 2008. The frog biting midges of the world (Corethrellidae: Diptera). Zootaxa. 1804: 1456.
110 Bradshaw, W. E., and C. M. Holzapfel. 1983. Predator mediated, nonequilibrium coexistence of treehole mosquitoes in southeastern North America. Oecologia. 57: 239256. Bradshaw, W. E., and K. Johnson. 1995. Initiation of metamorphosis in the pitcher plant mosquito: effects of larval growth hi story. Ecology. 76: 20552065. Brust, R. A. 1967. Weight and development time of different stadia of mosquitoes reared at various constant temperatures. Can. Entomol. 99:986993. Bns, M., and H. T. Ratte. 1991. The combined effects of temperature and food consumption on body weight, egg production and developmental time in Chaoborus crystallinus De Geer (Diptera: Chaoboridae). Oecologia. 88: 470 476. Callier, V., A. W. Shingleton, C. S. Brent, S. M. Gosh, J. Kim, and J. F. Harrison. 2013. The role of reduced oxygen in the developmental physiology of growth and metamorphosis initiation in Drosophila melanogaster. J. Exp. Biol. 216: 43344340. Callier, V., and H. F. Nijhout. 2011. Control of body size by oxygen supply reveals size dependent and size independent mechanisms of molting and metamorphosis. Proc. Nat. Acad. Sci. 108 : 1466414669. Campos, R. E., and L. P. Lounibos. 2000. Natural prey and digestion times of Toxorhynchites rutilus (Diptera: Culicidae) in southern Florida. Ann. Entomol. Soc. Am. 93: 12801287. Carpenter, S. R. 1984. Experimental test of the pupation window model for development of detritivorous insects. Ecol. Model. 23: 257 264 Chambers, G. M, and M. J. Klowden. 1990. Correlation of nutritional reserves with a critical w eight for pupation in larval Aedes aegypti mosquitoes. J. Am Mosq. Control Assoc. 6: 394399. Chown, S. L., and K. J. Gaston. 2010. Body size variation in insects: a macroecological perspective. Biol. Rev. 85: 139169. Christophers, R. S. 1960. Aedes aegy pti (L.) the yellow fever mosquito: its life history, bionomics and structure. Cambridge University Press, New York. Davidowitz, G., L. J. Amico, and H. F. Nijhout. 2003. Critical weight in the development of insect body size. Evol. Dev. 5: 188197. Day, T., and L. Rowe. 2002. Developmental thresholds and the evolution of reaction norms for age and size at lifehistory transitions. Am. Nat. 159: 338350.
111 Denver, R. J., N. Mirhadi, and M. Phillips. 1998. Adaptive plasticity in amphibian metamorphosis: res ponse of Scaphiopus hammondii tadpoles to habitat desiccation. Ecology 79: 18591872. Denver, R.J. 1997. Environmental stress as a developmental cue: corticotropinreleasing hormone is a proximate mediator of adaptive phenotypic plasticity in amphibian metamorphosis. Horm. Behav. 31: 169179. Denver, R.J. 1998. Hormonal correlates of environmentally induced metamorphosis in the Western spadefoot toad, Scaphiopus hammondii Gen. Comp. Endocrinol. 110: 326336. Flanagin, V. L., S. P. Haase, and S. A. Juliano. 2000. Effects of growth rates on development to metamorphosis in the lubber grasshopper, Romalea microptera. Oecologia. 125: 162169. Focks, D. A., D. G. Haile, E. Daniels, and G. A. Mount. 1993. Dynamic life table model for Aedes aegypti (Diptera: Culicidae): analysis of the literature and model development. J. Med. Entomol. 30 : 10031017. Forster, J., and A. G. Hirst. 2012. The temperaturesize rule emerges from ontogenetic differences between growth and development rates. Funct. Ecol. 26: 483492. Ghosh, S. M., N. D. Testa, and A. W. Shingleton. 2013. Temperaturesize rule is mediated by thermal plasti city of critical size in Drosophila melanogaster Proc. Royal Soc. B: Biol. Sci. 280: 1760. Giguere, L. A. 1981. Food assimilation efficiency as a function of temperature and meal size in larvae of Chaoborus trivittatus (Diptera: Chaoboridae). J. Anim. Eco l. 50: 103109. Gilpin, M. E., and G. A. H. McClelland. 1979. System analysis of the yellow fever mosquito Aedes aegypti. Fortschr. Zool. 25: 355388. Griswold, M. W., and L. P. Lounibos. 2005. Competitive outcomes of aquatic container diptera depend on predation and resource levels. Ann. Ent. Soc. Am. 98: 673681. Griswold, M. W., and L. P. Lounibos. 2006. Predator identity and additive effects in a treehole community. Ecology 87: 987 995. Hah n, D. A., L. N. James, K. R. Milne, and J. D. Hatle. 2008. Life history plasticity after attaining a dietary threshold for reproduction is associated with protein storage in flesh flies. Funct. Ecol. 22: 10811090.
112 Hatle, J. D., W. A. Miller, and D. W. Bo rst 2003. Canalization of development and ecdysteroid timing during the last instar in lubber grasshoppers. J. Insect Physiol. 49: 73 80. Hawley, W. A. 1988. The biology of Aedes albopictus J. Am. Mosq. Control Assoc. Supplement 1 : 1. Hechtel, L. J., and S. A. Juliano. 1997. Effects of a predator on prey metamorphosis: plastic responses by prey or selective mortality? Ecology 78: 838851. Hector, K. L., P. J. Bishop, and S. Nakagawa. 2012. Consequences of compensatory growth in an amphibian. J. Zool. 286 : 93101. Hensley, F. R. 1993. Ontogenetic loss of phenotypic plasticity of age at metamorphosis in tadpoles. Ecology. 74: 24052412. Hentschel, B. T. 1999. Complex life cycles in a variable environment: predicting when the timing of metamorphosis shifts fr om resource dependent to developmentally fixed. Am. Nat. 154: 549558. Hentschel, B. T., and R. B. Emlet. 2000. Metamorphosis of barnacle nauplii: effects of food variability and a comparison with amphibian models. Ecology. 81: 34953508. Howard, S. C., and B. T. Hentschel. 2005. Effects of short term food variability on the plasticity of age and size at metamorphosis of porcelain crab larvae. Limnol. Oceanogr. 50: 19601971. Jones, C. J. 1993. Larval growth rates and adult reproduction of Toxorhynchites splendens (Diptera: Culicidae) with restricted dietary intake. Environ. Entomol. 22: 174182. Juliano, S. A., and T. L. Stoffregen. 1994. Effects of habitat drying on size at and time to metamorphosis in the tree hole mosquito Aedes triseriatus Oecologia. 97: 369376. Juliano, S. A., J. R. Olson, E. G. Murrell, and J. D. Hatle. 2004. Plasticity and canalization of insect reproduction: testing alternative models of life history transitions. Ecology. 85: 29862996. Kesavaraju, B., B. W. Alto, L. P. Lounibos and S. A. Juliano. 2007. Behavioural responses of larval container mosquitoes to a sizeselective predator. Ecol. Entomol. 32 : 262 272. Kesavaraju, B., K. Damal, and S. Juliano. 2007. Threat sensitive behavioral responses to concentrations of water borne cues from predation. Ethology 113: 199206.
113 Kleckner, C. A., W. A. Hawley, W. E. Bradshaw, C. M. Holzapfel, and I. J. Fisher. 1995. Protandry in Aedes sierrensis: the significance of t emporal variation in female fecundity. Ecology. 76: 1242 1250. Layalle, S., N. Arquier, and P. Leopold. 2008. The TOR pathway couples nutrition and developmental timing in Drosophila Dev. Cell 15: 568577. Lehmann, T., and A. Diabate. 2008. The molecular forms of Anopheles gambiae: A phenotypic perspective. Infect. Genet. Evol. 8 : 737746. Leips, J., and J. Travis. 1994. Metamorphic responses to changing food levels in two species of hylid frogs. Ecology. 75: 13451356. Lepczyk, C. A., W. H. Karasov, and E. CaviedesVidal. 1998. Digestive responses during food restriction and realimentation in nestling house sparrows ( Passer domesticus ). Physiol Biochem Zool 71: 561 573. Lounibos, L. P. 1979. Temporal and spatial distribution, growth and predatory behav iour of Toxorhynchites brevipalpis (Diptera: Culicidae) on the Kenya coast. J. Anim. Ecol. 48: 213 236. Lounibos, L. P. 1983. The mosquito community of treeholes in subtropical Florida. In Frank, J.H. & Lounibos, L.P. (Eds.). Phytotelmata: Terrestrial Plants as Hosts of Aquatic Insect Communities. Plexus, Medford, New Jersey, pp. 223 246. Lounibos, L. P. 1985. Interactions influencing production of treehole mosquitoes in south Florida. In Lounibos, L.P., Rey, J.R. & Frank, J.H. (Eds.). Ecology of Mosquitoes: Proceedings of a Workshop. Florida Medical Entomology Laboratory, Vero Beach, Florida, pp. 65 77. Lounibos, L. P. 2001. Boom or bust development of a predatory mosquito in temporary aquatic habitats of Florida, USA. Israel J. Zool. 4 7: 433 444. Lounibos, L. P., R. L. Escher, D. Duzak, and E. A. Martin. 1996. Body size, sexual receptivity and larval cannibalism in relation to protandry among Toxorhynchites mosquitoes. Oikos 77:309316. Lounibos, L. P., S. Makhni, B. W. Alto, and B. Kes avaraju. 2008. Surplus killing by predatory larvae of Corethrella appendiculata : prepupal timing and sitespecific attack on mosquito prey. J. Insect Behav. 21: 47 54. McKeever, S., and F. E. French. 1991. Corethrella (Diptera: Corethrellidae) of eastern N orth America: laboratory life history and field responses to anuran calls. Ann. Entomol. Soc. Am. 84: 493 497. Mercer, D., A. Schoergendorfer, and R. Vandyke. 2008. Sexual differences in larval molting rates in a protandrous mosquito (Diptera : Culicidae) species, Aedes sierrensis J. Med. Entomol. 45: 861866.
114 Moehrlin, G. S., and S. A. Juliano. 1998. Plasticity of insect reproduction: testing models of flexible and fixed development in response to different growth rates. Oecologia. 115: 492500. Morey, S ., and D. Reznick. 2000. A comparative analysis of plasticity ln larval development in three species of spadefoot toads. Ecology 81: 17361749. Newman, R. A. 1992. Adaptive plasticity in amphibian metamorphosis: What type of phenotypic variation is adaptiv e, and what are the costs of such plasticity? BioScience. 42: 671678. Nijhout, H. F. 1975. Dynamics of juvenile hormone action in larvae of the tobacco hornworm, Manduca sexta (L.). Biol. Bull. 149: 568 579. Nijhout, H. F. 1981. Physiological control of molting in insects. Am. Zool. 21: 631640. Nijhout, H. F. 1994. Insect Hormones, Princeton University Press. Nijhout, H. F. 2008. Size matters (but so does time), and it's OK to be different. Dev. Cell 15: 491492. Nijhout, H. F., G. Davidowitz, and D.A. R off. 2006. A quantitative analysis of the mechanism that controls body size in Manduca sexta. J. Biol. 5: 1 15. Nylin, S., and K. Gotthard. 1998. Plasticity in life history traits. Annu. Rev. Entomol. 43: 6383. OMeara, G. F., L. Lounibos, and R. Brust. 1981. Repeated egg clutches without blood in the pitcher plant mosquito ( Wyeomyia smithii ) (Diptera, Culicidae). Ann. Entomol. Soc. Am. 74: 6872. O'Meara, G. F. 1985. Gonotrophic interactions in mosquitoes: kicking the bloodfeeding habit. Fla. Entomol.68: 122133. Packer, M., and P. Corbet. 1989. Size variation and reproductive success of female Aedes punctor (Diptera, Culicidae). Ecol. Entomol. 14: 297309. Padmanabha, H., C. C. Lord, and L. P. Lounibos. 2011. Temperature induces tradeoffs between devel opment and starvation resistance in Aedes aegypti (L.) larvae. Med. Vet. Entomol. 25 : 445453. Padmanabha, H., F. Correa, M. Legros, H. F. Nijhout, C. C. Lord, and L. P. Lounibos. 2012. An eco physiological model of the impact of temperature on Aedes aegypti life history traits. J. Insect Physiol. 58: 15971608. Rashed, S. S., and M. S. Mulla. 1989. Factors influencing ingestion of particulate materials by mosquito larvae (Diptera: Culicidae). J. Med. Entomol. 26: 210216.
115 Reiskind, M. H., and A. A. Za rrabi. 2012. Is bigger really bigger? Differential responses to temperature in measures of body size of the mosquito, Aedes albopictus J. Insect Physiol. 58: 9117. Relyea, R. A. 2007. Getting out alive: how predators affect the decision to metamorphose. Oecologia. 152: 389400. Relyea, R. A., and J. T. Hoverman. 2003. The impact of larval predators and competitors on the morphology and fitness of juvenile treefrogs. Oecologia. 134: 596604. Reznick, D. N. 1990. Plasticity in age and size at maturity in male guppies ( Poecilia reticulata): An experimental evaluation of alternative models of development. J. Evolution. Biol. 3: 185203. Rose, C. S. 2005. Integrating ecology and developmental biology to explain the timing of frog metamorphosis. Trends Ecol. E vol. 20 : 129135. Rueda, L. M., K. J. Patel, R. C. Axtell and R. E. Stinner. 1990. Temperaturedependent development and survival rates of Culex quinquefasciatus and Aedes aegypti (Diptera: Culicidae). J. Med. Entomol. 27: 892 898. Schenk, D., L. F. Bersier, and S. Bacher. 2005. An experimental test of the nature of predation: neither prey nor ratiodependent. J. Anim. Ecol. 74 : 86 91. Scriber, J. M., and F. Slansky Jr. 1981. The nutritional ecology of immature insects. Annu. Rev. Entomol. 26 : 183 211. Sears, K. E., Kerkhoff, A. J., Messerman, A., and H. Itagaki. 2012. Ontogenetic scaling of metabolism, growth, and assimilation: Testing metabolic scaling theory with Manduca sexta larvae. Physiol. Biochem. Zool. 85: 159173. Shafiei, M., A. Mocz ek, and H. Nijhout. 2001. Food availability controls the onset of metamorphosis in the dung beetle Onthophagus taurus (Coleoptera : Scarabaeidae). Physiol. Entomol. 26: 173180. Smith Gill, S. J. 1983. Developmental plasticity: developmental conversion versus phenotypic modulation. Integr. Comp. Biol. 23: 4755. Stearns S. C. 1992. The evolution of life histories. Oxford University Press, Oxford and New York. Steffan, W., and N. Evenhuis. 1981. Biology of Toxorhynchites Ann. Rev. Entomol. 26: 159181. Symonds, M., and A. Moussalli. 2011. A brief guide to model selection, multimodel inference and model averaging in behavioural ecology using Aikaike's information criterion. Behav. Ecol. Sociobiol. 65: 1321.
116 Teder, T., T. Tammaru, and T. Esperk. 2008. Dependence of phenotypic variance in body size on environmental quality. Am. Nat. 172: 223232. Telang, A., L. Frame, and M. R. Brown. 2007. Larval feeding duration affects ecdysteroid levels and nutritional reserves regulating pupal commitment in the yellow fever mosquito Aedes aegypti (Diptera: Culicidae). J. Exp. Biol. 210: 854864. Travis, J. 1984. Anuran size at metamorphosis: experimental test of a model based on intraspecific competition. Ecology. 64: 11551160. Trpis, M. 1972. Development and predatory behavior of Toxorhynchites brevipalpis (Diptera: Culicidae ) in relation to temperature. Environ. Entomol. 1 : 537 546. Twombly, S. 1996. Timing of metamorphosis in a freshwater crustacean: comparison with anuran models. Ecology. 77: 18551866. van der Have,T. M., and G. de Jong. 1996. Adult size in ectotherms: temperature effects on growth and differentiation, J. Theor. Biol. 183: 329 340 Walker, E. D., R. W. Merritt, M. G. Kaufman, M. P. Ayres, and M. H. Riedel. 1997. Effects of variation in quality of leaf detritus on growth of the eastern treehole mosquito, Aedes triseriatus (Diptera: Culicidae). Can. J. Zool. 75 : 706 718. Wessels, F. J., R. Kristal, F. Netter, J. D. Hatle, and D. A. Hahn. 2011. Does it pay to delay? Flesh flies show adaptive plasticity in reproductive timing. Oecologia. 165 : 311320. Westbrook, C. J., M. H. Reiskind, K. N. Pesko, K. E. Greene, and L. P. Lounibos. 2010. Larval environmental temperature and the susceptibility of Aedes albopictus Skuse (Diptera: Culicidae) to Chikungunya virus. Vector Borne Zoonot. 10 : 241247. Whitman, D. W., and T. N. Ananthakrishnan. 2009. Phenotypic plasticity of insects: mechanisms and consequences. Science Publishers, Enfield, New Hampshire. Wiklund, C., and C. Solbreck. 1982. Adaptive versus incidental explanations for the occurrence of protandry in butterfly, Leptidea sinapis L. Evolution 36: 56 62. Wiklund, C., S. Nylin, and J. Forsberg. 1991. Sex related variation in growth rate as a result of selection for large size and protandry in a bivoltine butterfly, Pieris napi Oikos 60: 241 250. Wilbur, H. M. 1980. Complex life cycles. Annu. Rev. Ecol. Syst. 11: 6793. Wilbur, H. M., and J. P. Collins. 1973. Ecological aspects of amphibian metamorphosis: Nonnormal distributions of competitive ability reflect selection for facultative metamorphosis. Science. 182: 13051314.
117 BIOGRAPHICAL SKETCH Erik Blosser was born in Newport News, V A and graduated from Denbigh Baptist High School. Following early interests in math ematics and animals, he recei ved his Bachelor of Science in mechanical engineering with a minor in biology from Messiah College in Grantham, Pennsylvania. During his undergr aduate degree, Erik discovered the world of insects through explorations in macrophotography. After graduation, he pursued his interest in entomology by taking jobs with mosquito control in Delaware and Virginia before moving to Florida for graduate resear ch. In his spare time Erik enjoys running, hiking and birdwatching.
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EGCYEFJDU_TDON3O INGEST_TIME 2014-10-03T21:43:05Z PACKAGE UFE0046606_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC