A Nutritional Profile of the Social Wasp Polistes Metricus

Journal of Insect Physiology 56 (2010) 42–56

Contents lists available at ScienceDirect

journal homepage: www.elsevier.com/locate/jinsphys

A nutritional proﬁle of the social wasp Polistes metricus: Differences in nutrient levels between castes and changes within castes during the annual life cycle

Timothy M. Judd a,*, Roxane M. Magnus a, Matthew P. Fasnacht b a Department of Biology, Southeast Missouri State University, Cape Girardeau MO, 63701, USA b Department of Chemistry, Southeast Missouri State University, Cape Girardeau MO, 63701, USA

ARTICLE INFO ABSTRACT

Article history: In wasps, nutrition plays a vital role for colony cohesion and caste determination. However, there is no Received 5 July 2009 baseline data set for the nutritional levels of wasps during the different stages of the colony cycle. Here Received in revised form 30 August 2009 we examined the levels of carbohydrates, lipids, protein, Ca, Cu, Fe, K, Mg, Mn, Na, and Zn in the wasp Accepted 9 September 2009 Polistes metricus at different stages of the wasp’s lifecycle. Individuals were collected at the following stages (1) spring gynes, (2) foundress colonies, (3) early worker colonies, (4) late worker colonies, (5) Keywords: emerging reproductives (gynes and males), (6) early fall reproductives, and (7) late fall reproductives. All Polistes eggs, larvae, pupae and adults were analyzed for their nutritional content to determine if there were any Nutrition differences between the nutrient levels in the different castes and how these nutrients changed within a Caste Cations caste during its lifetime. The results show there are differences in macro and micronutrient levels Macronutrients between the reproductive females and workers during development. Gynes showed changes in nutrient levels during their lifetime especially as they changed roles from a solitary individual to a nesting queen. Males also showed distinct nutritional changes during their lifetime. The implications for these nutritional differences are discussed. ß 2009 Elsevier Ltd. All rights reserved.

1. Introduction differences in gene expression within the developing castes (Evans and Wheeler, 2001; Kunieda et al., 2006; Wheeler et al., 2006). Understanding the nutritional needs and the nutritional state of Nutritional differences extend into adulthood in many of these an organism is important for understanding its physiology, devel- species. Reproductives tend to have higher lipid and storage opment, and behavior. Social insects provide a unique opportunity to proteins than workers in Hymenoptera (Peakin, 1972; O’Donnell, study the effects of nutrition on physiology because different 1998; Judd, 2006; Toth et al., 2009). Presumably these nutrients are functional components of a social insect colony have different important for ovarian growth, egg production and survival during nutritional needs much like different organs within an individual. In diapause and colony initiation. Much work has shown that addition, a single individual in a colony may change roles during its nutrients are distributed unequally through colonies such that lifetime which is potentially accompanied by physiological and individuals that need particular nutrients obtain more than those developmental changes (Robinson, 2009). As a result, the nutritional that do not (Peakin, 1972; Brian and Abbott, 1977; Sorensen and requirements may also change as well. Thus, a full understanding of Vinson, 1985; Sorenson et al., 1985; Hunt et al., 1987; Cassill and thenutritionalrequirementsandstates of individuals withinacolony Tschinkel, 1996, 1999a,b; Weeks et al., 2004; Hrassnigg and would provide insight into various levels of the biology of individuals Crailsheim, 2005; Judd, 2005; Judd and Fasnacht, 2007). within a colony and the biology of a colony as a whole. Behaviorally, workers of colonies can be influenced by the levels It has been known for a long time that nutrition plays a major of internal nutrient stores (Schultz et al., 1998; Blanchard et al., role in caste determination in many social insects (Wheeler, 1986; 2000; Judd, 2006) and down regulation of fatty acid synthesis (Toth Winston, 1987). In ants and honeybees the caste is influenced by et al., 2005). Workers with lower levels of nutrients are more likely the level of nutrition at a critical period during larval development. to forage (Blanchard et al., 2000; Kunieda et al., 2006). In addition, Termites (Roisin, 2000) and social aphids (Fukatsu and Ishikawa, the level of nutrition has been shown to influence colony cohesion 1992) also appear to have a nutritional component for caste in Polistes (Hunt and Dove, 2002; Karsai and Hunt, 2002). Polistes determination. These nutritional differences appear to trigger metricus larvae offer amino acid rich saliva to the adults (Hunt, 1991, 1994). If this saliva is artificially drained the colony will collapse (Hunt and Dove, 2002). Thus, nutrition is an important * Corresponding author. Tel.: +1 573 651 2365; fax: +1 573 651 2382. component for the maintenance of colonies in some eusocial E-mail address: [email protected] (T.M. Judd). insects.

Surprisingly, apart from a few ant studies (Vinson, 1968, then initiate their own colony (Reeve, 1991). Little is known about Solenopsis invicta; Stein et al., 1990, Tschinkel, 1993, Pogonomyr- the nutritional fluctuations during the solitary phase of the Polistes mex badius; Tschinkel, 1998, Pheidole ceres, Judd, 2006), we know life cycle. Such information would bring new insight to the very little about natural nutritional levels and nutritional behavior of Polistes during the solitary phase of its life cycle. fluctuations of individuals within social insect colonies over the Our study was carried out to increase our understanding of the course of the colony cycle. The absence of fundamental data such role of nutrients in the regulation of social behavior and the as these has been discussed in several works (Tschinkel, 1991; nutritional needs of the reproductives during their solitary stage of Arrese et al., 2001; Hunt, 2007; Robinson, 2009). These papers life cycle. In this study, we looked at the nutritional content of suggest that much of the work with social insects has been focused individuals of P. metricus during different stages of the wasp’s life on testing specific theoretical ideas (Tschinkel, 1991), or for insects cycle, creating a nutritional profile for P. metricus. We included in general, investigating the roles of specific proteins or pathways collections from the pre-nesting period, three different phases of (Arrese et al., 2001). This has been especially true with the nesting and two collections post-nesting. Individuals were analyzed increasing advances in genomics research (Arrese et al., 2001). for their total carbohydrates, lipids, proteins, calcium (Ca), copper Recently, several new hypotheses have been put forth such as the (Cu), iron (Fe), potassium (K), magnesium (Mg), manganese (Mn), ‘‘diapause ground plan hypothesis’’ (Hunt and Amdam, 2005, Hunt, sodium (Na), and zinc (Zn). Thus, we were able to compare the same 2007) and the ‘‘dominance-nutrition hypothesis’’ (Markiewicz and caste during different stages of the life cycle and different castes O’Donnell, 2001; Molina and O’Donnell, 2008) emphasizing the within and between different stages of the life cycle. role of nutrition in the evolution and regulation of social behavior. At this juncture, the need for ‘‘nutritional profile’’ studies, the 2. Methods mapping out of natural nutritional levels in colonies at different stages of their life cycle, has become more evident than ever to 2.1. Study site complement the important advances in genomics and physiology to allow for a better understanding of the role of nutrition in the The study site was at Kelso Wildlife area owned by Southeast dynamics of social insect colonies. Missouri State University. The site consists of two adjacent fields Most nutritional ecology studies on insects have focused on the surrounded by woodlands. A wide path connects the fields to each levels of carbohydrates, proteins and lipids in the diets of other. The fields are close to a dirt road which has a number of organisms themselves. Other nutrients such as metal cations species of flowers blooming along the road side. Males and gynes could also have a potential effect on colony dynamics and caste frequented this spot during the fall (see below). determination. The role of micronutrients in the physiology of insects especially social insects has been largely neglected (Cohen, 2.2. Nest boxes 2004). Studies on non-social insects have found that several ions such as Ca, K, Na, Cu and Zn are important for growth, ovarian Nest boxes were 15.24 cm Â 15.24 cm Â 15.24 cm boxes con- development, and fecundity in insects (McFarlane, 1974, 1976, structed from wood. They were of similar construction as reported 1991). Judd and Fasnacht (2007) found that micronutrients are not in Gamboa et al. (1992). Two screws held the top of the box in distributed equally throughout the colony. Indeed, in most cases place. They were open in the front with a chicken wire screen to the micronutrients end up in the caste that has growing tissue. In protect the colony from bird predation. The box was mounted with addition, there is evidence that the levels of manganese differ baling wire onto a wooden stake about 1.3 m from the ground. A between pollen and nectar honey bee foragers (Ben-Shahar et al., total of 69 boxes were placed on the edge of the fields in the study 2004). This suggests that there may be a significant role for site. micronutrients in the regulation of caste determination but this has yet to be explored more fully. 2.3. Surveys and marking Primitively eusocial organisms such as Polistes provide an excellent model for nutritional profile studies because the life of a Colonies were monitored once a week and checked for number colony is restricted to one season. Therefore, it is possible to of individuals, number of cells and the relative amount of pupal measure the changes in the nutritional structure of a colony from cells in the nest. Foundresses were marked on with yellow Testorsß the initial foundress to the resulting reproductives. Polistes paint during the early nesting period. Early workers were marked colonies go through distinct phases including a worker production with red paint after the early worker collection took place (see phase and a reproductive production phase (Reeve, 1991). Thus, it below). is not only possible to tease apart the roles of the adults but the differences between developing workers and developing repro- 2.4. Collections ductives. The diapause ground plan hypothesis (Hunt and Amdam, 2005; Hunt, 2007) suggests that worker destined larvae will be All colonies used in this study were single foundress colonies. nutritionally different than reproductive destined larvae and a Wasps were collected during the following phases:Spring gynes: nutritional profile study would be an ideal place to start looking for Individual spring gynes were collected about a km away from the differences between the two groups. A nutritional profile will field site to avoid collecting potential foundresses in the nest boxes. identify nutritional differences between the castes and may give A total of 20 individuals were collected between March 20, 2007 and further clues to the underlying physiological differences between April 4, 2007. Foundress colonies: 10 nest boxes were collected prior worker and reproductive caste development. Such information to worker emergence on May 21, 2007. The single foundress and her could identify new areas to explore using physiological and nest containing several cells with pupae. Early worker colonies: 10 genomic studies. nest boxes were collected containing early workers were collected In addition to the social phase, there is a solitary phase to the on June 26, 2007. Late workers colonies: 10 nest boxes were Polistes life cycle that is generally understudied. Gynes and males collected containing late workers and laden with pupal cells were emerge from the colony but then eventually lead a more solitary collected on August 15, 2007. In many cases, a few early males had existence (Reeve, 1991; Hunt, 2007). Males defend territories in emerged.Emerging reproductives: Due to an invasion by a raccoon, order to mate with females (Polak, 1993, 1994) and females only five colonies could be collected from the field site on August 21, eventually must procure their own nourishment, overwinter and 2007. Seven colonies were collected from a nearby residence and a 44 T.M. Judd et al. / Journal of Insect Physiology 56 (2010) 42–56 nearby farm (August 26 and 28, 2007 respectively). Nests were Bradford assay was used to determine the total protein (Bradford, brought back to the lab. All workers and larvae were removed 1976). The fluorescamine assay is too sensitive for samples as large leaving the pupae to develop. Nests were placed in a plastic as Polistes (Bo¨hlen et al., 1973). Thus, each individual had all three container with a mesh covering to prevent escape. The nests were nutrients measured. checked daily and any emerged adult was placed in a microcen- Individuals were homogenized in dii water. Smaller samples trifuge tube and frozen at À80 8C. September males and gynes: 40 such as eggs and small larvae were homogenized in 150 mlof males and 40 gynes were collected from the study site on September water. Midsized samples were homogenized in 300 ml of water 13 and 16, 2007 using an insect net. Upon capture the individual was and large samples such as large larvae, prepupae, pupae and adults placed in a vial and placed in a cooler. Once the specimens were were homogenized in 500 ml of water. The different amounts were brought back to the lab they were frozen at À80 8C. October males used to make sure the nutrients in the smaller samples would be and gynes: 24 males and 12 gynes were collected from the study site detected. All samples were then placed in a centrifuge at and a nearby field between October 9 and 21 in the same manner as 14,000 Â g for 2 min. the September collections. An early cold front prevented more individuals from being collected. Once the specimens were brought 2.6.1. Protein back to the lab they were frozen at À80 8C. The Bradford assay (Bradford, 1976) was used to determine Nest collections took place at night to insure all of the total protein. 1 ml of Bradford reagent was placed in a 1.5 ml individuals were present. Large 2 or 3 gallon Ziplockß Storage microcentrifuge tube. A 50 ml of aliquot sample was placed in the bags were placed over the entire nest box and once removed from solution, vortexed and allowed to develop for 5 min. Absorption the stake, the nest box was sealed in the bag and placed in a cooler. was measured at 595 nm and compared to 2.7, 5.35, 10.5, 23.0, This insured no wasps escaped. At the lab, the nest boxes were kept 41.5, 53.5 mg Bovine serum standards. Results were adjusted based in a refrigerator until they were sorted. All nests were sorted on the percentage used from the total sample. within 24 h of collection. All adults were placed in individually labeled 2 ml microcen- 2.6.2. Carbohydrates trifuge tubes. All eggs, larvae, prepupae, pupa and adults were Total carbohydrates was measured using the anthrone assay removed from the nest and placed in individual labeled micro- (Van Handel, 1985a). A 50 ml aliquotof sample was placed in a 1.5 ml centrifuge tubes. Half of the prepupae, pupae, eggs, workers and microcentrifuge tube. 12.5 ml of an 18% NaSO4 solution was added males, if present, from each colony were selected for the followed by 1.25 ml of anthrone reagent. The sample was vortexed micronutrient assay and the other half for the macronutrient and then heated at 100 8C for 12 min and then allowed to cool to assay. Larvae were sorted by size and then divided equally between room temperature. The absorption was measured at 625 nm and the two assays such that each assay had a similar size distribution compared to 10, 25, 50, 75, and 100 mg glucose standards. Results of larvae represented. Because there was only one queen (or were adjusted based on the percentage used from the total sample. foundress) per nest, half of the queens were assigned to the micronutrient assay and the other half the macronutrient assay. 2.6.3. Lipids Individuals destined for the micronutrient assay were stored at Once the first two assays were completed 75, 150, 250 mlof1:1 À20 8C. Individuals destined for the macronutrient assay were chloroform:methanol was added to the tube for small, medium and stored at À80 8C. Meconium (larval feces) pellets from a small large samples (see above) respectively. The mixture was vortexed to sample of pupal cells were also preserved, dried, weighed, and insure the lipids dissolved and then centrifuged at 14,000 Â g for analyzed in the micronutrient assay. In the late colonies several 2 min. Total lipid was measured using the phosphovanillan assay. A meconium pellets in the pupal cell were divided by an exoskeleton 50 ml aliquot was taken from the chloroform layer and placed in a discarded during a molt. This suggested that the bottom layer of 2 ml microcentrifuge tube and allowed to dry. Once dried, 0.2 ml of feces was from one individual and the top from a second. We concentrated sulfuric acid was added, vortexed, and heated at 100 8C presumed that the bottom layer came from a worker destined for 10 min. After heating, the sample was transferred to a glass tube individual and the top from a reproductive destined individual. and 3 ml of phosphovanillan reagent (Van Handel, 1985b)was added to the sample and allowed to develop for 30 min. The 2.5. Measurements absorption was measured at 525 nm and compared to 18, 45, 72, and 90 mg corn oil standards (Wheeler and Buck, 1992). Results were All individuals had their wet weight measured. The head width adjusted based on the percentage used from the total sample. (HW), pronotum width (PW) and wing length (WL) was measured for all adults. Paint used for marking was removed with a 2.6.4. Caloric content paintbrush dipped in xylene. The xylene was allowed to evaporate The caloric content per individual was estimated using the before storage. Pollen was removed from adults when present. following values: 4.19 cal/mg for protein, 4.2 cal/mg for carbohy- Pupa had their HW and PW measured, and prepupae and larvae drates, and 9.5 cal/mg for lipids (Platt and Irwin, 1973). had their HW measured. The sex was determined for adults, pupae, prepupae and most of the late instar (4th and 5th) larvae. The sex 2.7. Determination of total ions determination of the prepupae and larvae was done using the features described by Cotoneschi et al. (2007). Although in their Each individual insect was placed in an acid-washed test tube study they cleared the larvae prior to analysis, the features they with 0.5 ml of trace metal grade concentrated nitric acid, HNO3 described were visible without any manipulation. (Fisher) and placed on a block heater. Each sample was digested by Individuals designated for the micronutrient assay were dried allowing the nitric acid to boil approximately 30 minutes. After each to constant mass at 45 8C and had their dry mass determined. sample was digested, 0.2 ml of a 10 ppm scandium solution (Spectrum Plasma Emission Standard Grade), which was used as 2.6. Determination of total protein, carbohydrates, and lipids an internal standard, and 1.3 ml of deionized water (Millipore Synergy 185) were added to make a final solution volume of Total protein, carbohydrates and lipids were determined using approximately 2.0 ml. Any undigested lipids were separated from a modified version of the technique described by Judd (2006). the sample solution by adding 1.5 ml of HPLC grade chloroform Instead of the fluorescamine assay described in Judd (2006), the (Fisher). The aqueous layer of each sample was measured using ICP- T.M. Judd et al. / Journal of Insect Physiology 56 (2010) 42–56 45

OES (PerkinElmer Optima 3000 DV) to determine the micronutrient 2.8.5. Comparison of adults content of each insect. Emission for each ion was detected at the There were a total of 12 different classifications for adults: following wavelengths: Ca 317.933 nm; Cu, 324.754 nm; Fe, spring gynes, foundresses, early queens, late queens, early workers, 238.204 nm; K, 766.491 nm; Mg, 279.079 nm; Mn, 257.610 nm; late workers, emerging gynes, September gynes, October gynes, Na, 589.592 nm; Sc, 361.384 nm; and Zn, 213.856 nm. Concentra- emerging males, September males, and October males. Because tions were determined using the calibration curve method after there were not an equal number of classifications for the different normalizing the emission intensity of each metal to that of scandium castes (i.e. there were only two worker classes but three male from the same sample. Standard solutions (0.1–10 ppm) each classes) a single class variable (caste) was used in the MANOVA containing 1 ppm Sc were made from a stock solution (Spex (proc glm in SAS). A Tukey Honest Significant Test was used to Certiprep Instrument Calibration Standard 2) containing 100 ppm compare each group. Ca, Cu, Fe, K, Mg, Mn, Na, and Zn. Concentration measurements for all eight elements were obtained for each individual sample. 2.8.6. Worker vs. reproductive meconium The meconium was divided into four groupings (1) early nest, 2.8. Data analysis (2) mid nest, (3) late nest bottom, (4) late nest top. A single class variable ‘‘origin’’ was used in the MANOVA (proc glm in SAS). A 2.8.1. Within colony comparisons Tukey Honest Significant Test was used to compare the groupings. During the colony collections, there were different individuals present on the colony. The first collection lacked workers and the 2.9. Comparisons of morphology and nutrient levels final collection had males. Therefore the data set was unbalanced. All data were log transformed because the levels of each nutrient were 2.9.1. Larvae very different. Principle component analysis (PCA) was used as a Larvae of Polistes go through five instars. Head width is a good data reduction tool because many individuals came from the same indication of instar stage. A stepwise regression model was used to colony and were not independent. A canonical discriminate function compare the nutrient levels of larvae with their head width and the analysis (CDA) by collection period was run to determine if there are collection period. This approach was used because the data set did differences between the different types of individuals within a not meet the assumptions of a MANCOVA. In addition to allowing a colony during each collection period. Caste was the class and the comparison between instars, this analysis allowed us to compare nutrients (micro or macro) were the variables. Micronutrients and worker destined larvae (first two collections) and reproductive macronutrients were analyzed separately. Caloric values were also destined larvae (last collection). analyzed separately because the values were dependent on the three macronutrient measurements. For each colony, an average caloric 2.9.2. Pupae value was calculated per caste (and per sex when sex was used in an Two PCAs were run, one for the morphological characters HW analysis) to control for within colony effects. and PW and the other for the nutritional content. A canonical correlation analysis (CCA) was used to determine any correlation 2.8.2. Nutrient comparisons between size and the nutritional levels in individuals. Males and In the following three analyses, micronutrients and macro- females were run separately and micro and macronutrients were nutrients were analyzed separately. All data were log transformed run separately producing four analyses in total. because the levels of each nutrient were very different. Principle component analysis (PCA) was used as a data reduction tool 2.9.3. Adults because in some collection periods several individuals came from Adults were compared using the same procedure as pupae the same colony and were not necessarily independent. In other except WL was included in the morphological PCAs. collection periods individuals, such as spring gynes, were not part of a colony so a colony variable could not be used in MANOVA 2.9.4. Workers models. In all cases where a PCA and MANOVA analysis was Wing wear is a good indication of the amount of foraging done performed, 95% confidence intervals were also generated as a by a worker. A MANOVA (proc glm in SAS) was used to compare the second analysis (Judd, 2006). results of PCA analyses for nutritional content for workers and their wing wear. Macronutrients, calories, and micronutrients 2.8.3. Comparison of eggs were analyzed separately. Eggs from the three collection periods were analyzed using a PCA followed by a MANOVA (proc glm in SAS) using collection 3. Results period (period) as a class variable. A Tukey Honest Significant Test (THSD) was used to compare collection periods. The results presented here are a summary of the findings based on the statistical procedures mentioned above. Tables of the PCA and 2.8.4. Comparison of larvae, prepupae, and pupae CCA results as well as more details concerning the interpretation of Each of the three developmental stages (stage) was divided into the statistical results can be found in the Supplementary data. the following categories (castes): workers, gynes and males. Only Hereafter we will use the term ‘‘life stage’’ to include developmental the last two instars of larvae were included because they were the stages (eggs, larvae, prepupae, and pupae) as well as all adult castes only individuals we could determine the sex using morphological (worker, female reproductives, and males). characters. Individuals whose sex could not be determined were excluded from the analysis. Individuals from the early and mid 3.1. Within colony comparisons season collections were considered worker destined and individuals from the late collection were considered gyne and male 3.1.1. Macronutrients destined. Only one male larva was discovered in the mid season There were no significant differences between the life stages collection and was not included in the analysis. within foundress colonies. In mid season colonies and late season For both nutrients a PCA was run followed by a MANOVA (proc colonies, eggs had higher protein and lipid content per unit mass glm in SAS) which compared developmental stage and caste. Ad hoc than any other life stage (Fig. 1). Larvae had higher levels of protein comparisons were done using the Tukey Honest Significant Test. and lipids per unit mass than other life stages except eggs in late 46 T.M. Judd et al. / Journal of Insect Physiology 56 (2010) 42–56

Fig. 3. The means and 95% conﬁdence intervals for levels of protein (gray), carbohydrates (white), and lipids (dotted) for Polistes metricus eggs collected from foundress nests (early; N = 3 eggs), early worker nests (mid; N = 80 eggs), and late worker colonies (late; N = 38 eggs). No signiﬁcant differences were found between each nutrient in eggs collected during each of the time periods.

All of these nutrients were high in eggs and K, Mg, and Ca were moderately high in larvae. Of the remaining life stages, Ca was higher in queens and all elements were found in the lowest amounts in Fig. 1. Results of the canonical discrimant analysis for the three macronutrients for (a) mid season colonies and (b) late season colonies. Eg: eggs, Lv: larvae, Pp: prepupae and pupae with the possible exception of K in pupae. prepupae, Pu: pupae, Qu: queens, Wr: workers, Ml: males. No significant effects were seen for early season colonies. 3.2. Eggs season colonies (reproductive destined larvae) but not mid season 3.2.1. Macronutrients colonies (worker destined larvae, Fig. 1). There were no significant differences in macronutrient levels between workers or reproductive destined eggs (Fig. 3). 3.1.2. Micronutrients In the foundress colonies, Cu levels were highest in eggs and 3.2.2. Calories next highest in foundresses (axis 1, Fig. 2a). Both Mg and K were Due to the limited number of colonies with eggs in the late high in eggs and larvae, but low in all other life stages (axis 2, season, the Kruskal–Wallis test was used to compare the average Fig. 2a). When examined further, K was moderately high for both caloric values for the eggs. Eggs from the mid season nests had a prepupae and pupae and lowest in the foundresses. Mn was found significantly higher caloric value than early or late season eggs to be highest in foundresses followed by larvae (axis 2, Fig. 2b). (H = 8.62, p < 0.05). The mid season eggs were reproductive Prepupae had the lowest levels of Mn in foundress colonies. destined eggs while the eggs in the late season nests were either In mid season colonies Mn was highest in the queens, worker laid eggs or eggs laid by a nutritionally drained queen. In moderately high in workers and larvae, and low in prepupa, several of the late season nests many cells contained multiple eggs pupae, and eggs (axis 1, Fig. 2c). suggesting workers were laying eggs in some cases (Chapman In late season colonies, Na (axis1, Fig. 2d) showed the strongest et al., 2009). Overall, reproductive destined eggs had a higher effect while Ca, Fe, K, and Mg showed less of an effect (axis 2, Fig. 2d). caloric value than worker destined eggs (Fig. 4).

Fig. 2. Results of the canonical discrimant analysis for the three micronutrients for (a and b) early season colonies, (c) mid season colonies, and (d) late season colonies. Eg: eggs, Lv: larvae, Pp: prepupae, Pu: pupae, Fd: foundresses, Qu: queens, Wr: workers, Ml: males. T.M. Judd et al. / Journal of Insect Physiology 56 (2010) 42–56 47

3.3. Larvae

3.3.1. Macronutrients There was a signiﬁcant drop in levels of protein (F = 261.79, p < 0.0001, Fig. 6a), carbohydrates (F = 50.61, p < 0.0001, Fig. 6b), and lipids (F = 101.99, p < 0.0001, Fig. 6c) per unit mass as larvae grew. Late season larvae also had signiﬁcantly higher lipid content per unit mass than the larvae from the early and mid collections (F = 9.85, p = 0.002, Fig. 6c). Thus, reproductive destined larvae have higher lipid content than worker destined larvae throughout their development.

Fig. 4. The means and 95% confidence intervals of number of calories in eggs collected from foundress nests (early; N = 3 eggs), early worker nests (mid; N =80 3.3.2. Calories eggs), and late worker colonies (late; N = 38 eggs) of the wasp Polistes metricus. The caloric value of the reproductive larvae increased Different letters indicate significant differences based on the Kruskal–Wallis Test. significantly faster than early workers (F = 48.62, p < 0.0001, Fig. 7) and late workers (F = 4.63, p < 0.032, Fig. 7). 3.2.3. Micronutrients The overall MANOVA showed a significant difference between 3.3.3. Micronutrients micronutrient content between collections (Wilks’ l = 0.68, The levels of Ca per unit mass decreased as the larvae grew

F = 2.73, dfNum = 16, dfDen = 208, p = 0.0006). The only difference (F = 37.73, p = 0.0001, Fig. 8a). Levels of Ca were signiﬁcantly found between the groups of eggs was that worker destined eggs higher in larvae from the early period than the late period (F = 5.94, had lower copper (p = 0.003, THSD) than reproductive destined p = 0.016, Fig. 8a) and decreased faster in the mid period than the eggs (Fig. 5f). late period (F = 4.57, p < 0.034, Fig. 8a). This, reproductive destined

Fig. 5. The means and 95% conﬁdence intervals for levels of (a) Ca, (b) Cu, (c) Fe, (d) K, (e) Mg, (f) Mn, (g) Na, and (h) Zn for Polistes metricus eggs collected from foundress nests (early; N = 3), early worker nests (mid; N = 78), and late worker colonies (late; N = 35). 48 T.M. Judd et al. / Journal of Insect Physiology 56 (2010) 42–56

Fig. 7. Scatter diagram depicting the relationship between head width and calories in Polistes metricus larvae collected from pre-worker nests (open circle, dotted line), early worker nests (black squares and black line), and late colonies (gray triangles and gray line). The ﬁrst two collections represent worker destined larvae while the last collection represents reproductive destined larvae.

higher than the late collections (F = 24.04, p < 0.0001 Fig. 8f). The levels of Mn increased significantly more in larvae from the mid collection period than the larvae from the late period (F = 4.63, p = 0.033, Fig. 8f). Overall, worker destined larvae accumulated Mn at much higher rates than reproductive destined larvae but not mid season larvae. There was no overall relationship between head width and levels of Mg or Zn per unit mass in P. metricus larvae. The late season larvae seemed to retain more Mg than the early season larvae because they had significantly smaller slope (F = 18.61, p = 0.0001, Fig. 8e). Unlike Mg, early season larvae had a significantly higher increase in Zn levels per unit mass than late season larvae (F = 3.62, p < 0.0001, Fig. 8h). Thus, it appears that reproductive destined larvae retained less Zn and more Mg than early worker destined larvae. Levels of Cu per unit mass decreased significantly (F = 27.26, p < 0.0001, Fig. 8b) while Fe (F = 14.75, p = 0.0002, Fig. 8c) and Na (F = 25.26, p < 0.0001, Fig. 8g) increased significantly in all larvae as they grew. There were no significant differences in levels of these cations between the three collection periods.

3.4. Comparison of larvae, prepupa, and pupae

3.4.1. Macronutrients The overall MANOVA model showed significant effects of Fig. 6. Scatter diagrams depicting the relationship between head width and (a) l protein, (b) carbohydrates, and (c) lipids in Polistes metricus larvae collected from developmental stage (stage; Wilks’ = 0.86, F = 5.56, dfNum =6, pre-worker nests (open circle, dotted line), early worker nests (black squares and dfDen = 428, p = 0.0001) and caste (male, gyne, or worker; Wilks’ black line), and late colonies (gray triangles and gray line). The first two collections l = 0.73, F = 12.07, dfNum =6, dfDen = 428, p = 0.0001), but no represent worker destined larvae while the last collection represents reproductive significant interaction between stage and caste (Wilks’ l = 0.91, destined larvae. F = 1.61, dfNum = 12, dfDen = 566, p = 0.084). Lipids in pupae were significantly higher per unit mass than larvae (p < 0.0001, THSD) or prepupae (p = 0.015, THSD; Fig. 9a). larvae had a significantly lower decline in Ca levels per unit mass Protein was higher in worker destined brood but lipids were lower suggesting they retained more Ca than worker destined larvae. in worker destined brood than gyne or male destined brood The levels of K per unit mass decreased significantly in the (p < 0.0001, THSD; Fig. 9b). Carbohydrates did not differ sig- larvae as they grew (F = 15.00, p = 0.0002, Fig. 8d). However, this nificantly between developmental stages or castes. drop was significantly greater in the larvae from the early (F = 4.90, p = 0.023) and mid (F = 8.96, p = 0.0032) collection periods than the 3.5. Calories late period (Fig. 8d), meaning reproductive destined larvae retained more K than worker destined larvae. The overall model showed a significant effect of developmental

There was a significant increase in levels of Mn per unit mass in stage (Wilks’ l = 0.22, F = 391.42, dfNum =2, dfDen = 221, the larvae as head width increased (F = 37.19, p < 0.0001, Fig. 8f). p = 0.0001; Fig. 9a). There was a significant drop in caloric value The levels of Mn in the early period collections were significantly from larvae to pupae (p < 0.0001, THSD). T.M. Judd et al. / Journal of Insect Physiology 56 (2010) 42–56 49

Fig. 8. Scatter diagrams depicting the relationship between head width and (a) Ca, (b) Cu, (c) Fe, (d) K, (e) Mg, (f) Mn, (g) Na, and (h) Zn in Polistes metricus larvae collected from pre-worker nests (open circle, dotted line), early worker nests (black squares and black line), and late colonies (gray triangles and gray line). The ﬁrst two collections represent worker destined larvae while the last collection represents reproductive destined larvae.

3.6. Micronutrients

The overall MANOVA showed signiﬁcant interactions of stage

(Wilks’ l = 0.15, F = 36.99, dfNum = 16, dfDen = 376, p = 0.0001), caste (Wilks’ l = 0.54, F = 8.42, dfNum = 16, dfDen = 376, p = 0.0001), and caste/stage (Wilks’ l = 0.54, F = 4.00, dfNum = 32, dfDen = 694.9, p = 0.0001). Ca and Zn showed a signiﬁcant decrease from larvae to pupae

(pCa < 0.0016, THSD, pZn = 0.039, THSD, Fig. 10a and h) but no difference between castes. There were no significant effects of stage or caste on Cu according to the MANOVA analysis. However, the confidence intervals indicate worker larvae had higher levels of Cu than any other caste (Fig. 10d). Mn, Mg, and Cu showed similar patterns in which levels were high in worker destined larvae and decreased during pupation (Fig. 10b, e and f). The drop was significant for Mn (p < 0.0001, TSHD) and Mg (p = 0.01, THSD) according to both analyses (MANOVA and confidence intervals) but only significant for Cu according to the confidence intervals (Fig. 10b). Levels of these nutrients were higher in worker larvae than reproductive larvae Fig. 9. The means and 95% confidence intervals for levels of protein (light gray), carbohydrates (white), lipids (dotted) and calories (dark gray) for Polistes metricus (Fig. 10b, e and f). Of the three, Mn appears to have the strongest (a) larvae (N = 89), prepupae (N = 45), and pupae (N = 96) and (b) males (N = 66), difference between the worker destined and reproductive destined gynes (N = 59) and worker (N = 105) immatures. larvae (Fig. 10f). Levels of Mg and Cu remained constant in both 50 T.M. Judd et al. / Journal of Insect Physiology 56 (2010) 42–56

Fig. 10. The means and 95% conﬁdence intervals for levels of (a) Ca, (b) Cu, (c) Fe, (d) K, (e) Mg, (f) Mn, (g) Na, and (h) Zn in Polistes metricus late instar larvae, prepupae, and pupae of males, gynes and workers. Sample sizes for each category are the following: male larvae N = 12, gyne larvae N = 20, worker larvae N = 32, male prepupae N = 7, gyne prepupae N = 9, worker prepupae N = 35, male pupae N = 39, gyne prepupae N = 26, worker pupae N = 24.

males and gynes (Fig. 10b and e) but levels of Mn dropped during dfDen = 88, p = 0.03). Males with wider heads had lower levels of pupation in these castes (p < 0.002, Fig. 10f). carbohydrates and higher levels of protein per unit mass (CCA,

Na and Fe showed similar patterns to those observed in Mn, Mg, Wilks’ l = 0.43, F = 4.36, dfNum =6,dfDen = 50, p = 0.0013). There and Cu. Worker destined larvae had higher levels of Na (p = 0.0009, was no relationship between morphological characters and caloric TSHD) and Fe (p = 0.0001) than gyne destined larvae (Fig. 10c and values or micronutrient levels for female or male pupae. g). However, gynes destined individuals showed a significant increase in levels of Na (p = 0.002, TSHD) and Fe (p = 0.0001, THSD) 3.8. Adult morphology and nutrition per unit mass during pupation (Fig. 10c and g). Males only showed a significant increase in levels of Na (p = 0.01, TSHD) per unit mass 3.8.1. Females during pupation (Fig. 10g). Worker destined individuals showed a Females with shorter wings tended to have greater amounts of slight decrease Fe during pupation (confidence intervals) but their carbohydrates and proteins per unit mass (CCA, Wilks’ l = 0.55,

Na levels remained the same (Fig. 10c and g). F = 10.73, dfNum =9, dfDen = 345.74, p = 0.0001). Females with K only showed a signiﬁcant effect of caste but it was the larger pronotum and head widths had higher levels of Mn, Ca, opposite from what was observed for the other elements. Gynes and Cu and lower levels of Na and Mg per unit mass (CCA, Wilks’ and males had signiﬁcantly higher levels of K than workers l = 0.685, F = 2.41, dfNum = 24, dfDen = 415.34, p = 0.0003). (p = 0.001, THSD, Fig. 10d). 3.8.2. Males 3.7. Pupae morphology and nutrients Males with shorter wings had higher lipid and protein levels per

unit mass (CCA, Wilks’ l = 0.454, F = 14.52, dfNum =9, As head width increased in female pupae the levels or dfDen = 340.87, p = 0.0001). In the micronutrient analysis, all carbohydrates and lipids per unit mass decreased while the levels morphological measurements were found to be correlative. Larger of protein increased (CCA, Wilks’ l = 0.73, F = 2.47, dfNum =6, males had higher levels of Zn and lower levels of Na, Fe, and Mg T.M. Judd et al. / Journal of Insect Physiology 56 (2010) 42–56 51

(CCA, Wilks’ l = 0.609, F = 3.18, dfNum = 24, dfDen = 409.54, p = 0.0001).

3.8.3. Worker wing wear There was no relationship between worker wing wear and the levels of macronutrients or calories. The overall MANOVA analysis was signiﬁcant for micronutrients (Wilks’ l = 0.33, F = 2.88, dfNum = 24, dfDen = 151.42, p = 0.0001) and showed that workers with moderate to high wing wear had higher levels of Mn than workers with no or little wing wear (p < 0.05). Thus, workers that foraged more had higher levels of Mn.

3.9. Adults: comparisons between nutrients in adults

3.9.1. Overall results The overall MANOVA model of the effects of caste on Fig. 12. The means and 95% confidence intervals for estimated caloric values of macronutrients was significant (Wilks’ l = 0.519, F = 3.18, adult Polistes metricus female reproductives, workers, and males collected at dfNum = 33, dfDen = 893.4, p = 0.0001). The overall model for different times during the life cycle. Sample sizes are the following: emerging gynes micronutrients was significant for caste (Wilks’ l = 0.80, N = 45, September gyne N = 21, October gyne N = 7, Spring gyne N = 10, foundress N = 5, early queen N = 3, late queen N = 3, early worker N = 21, late worker N =48, F = 10.26, dfNum = 88, dfDen = 1917.7, p < 0.0001). emerging male N = 123, September male N = 19, October male N = 12. 3.9.2. Changes in gynes There was a significant increase in carbohydrates and a significant decrease in lipids from emerging gynes to spring gynes protein levels recovered (Fig. 11b). These results suggest that (1) (p = 0.021, THSD). It appears that the major drop in lipids happened gynes gained protein after emerging from diapause and (2) as the gynes overwintered (between October gynes and spring foundresses lose protein as they produce workers and then regain gynes, Fig. 11a). Caloric values seemed to drop slightly at this time some when workers (or larvae) are available. as well (see confidence intervals, Fig. 12). Protein levels decreased Levels of K dropped significantly from emergence to September in gynes from emergence to September (p < 0.0001) and then in gynes (Fig. 13d) while levels of Cu (p < 0.001, THSD) and Mn increase significantly from October gynes to spring gynes (p < 0.0001, THSD) increased during this time (Fig. 13b and f). The (Fig. 11b). Protein levels then dropped significantly when gynes levels of Mn seemed to remain constant throughout the rest of life were founding colonies (p = 0.02). Once females became queens of the females. Levels of Cu in October measurements were not significantly different than Cu levels upon emergence. This result suggests there was a slight drop in Cu levels in gynes from September to October. In the spring, Cu significantly increased from spring gynes to early queens (p < 0.001, THSD, Fig. 13b). This high level of Cu was maintained in queens for the rest of the nesting season. According to the confidence intervals, the levels of Ca also increased from foundresses to early queens (Fig. 13a). Both Na and Mg showed more gradual changes over the lifetime of the gynes. Na showed a gradual increased in gynes from emergence to October (p = 0.0008, Fig. 13g). There was continuous drop in Mg levels from the time of emergence to the founding of a colony in reproductive females (p = 0.019, THSD, Fig. 13e). Levels of Mg appear to increase from the foundress to early queens (confidence intervals).

3.9.3. Changes in males There was a significant increase of carbohydrates and a significant decrease in lipids from emergence to October in males (p = 0.017; Fig. 11a). Males also showed a significant decrease in caloric value from emergence to October (p = 0.038, THSD; Fig. 12). Thus, males are continuously losing energy throughout the fall. In addition to energy, males also had a significant drop in protein levels during their lifetime (p = 0.0024, Fig. 11b). A similar pattern for Cu, Mn, and K was seen in males as seen in gynes. Levels of K dropped from emerging males to September males (Fig. 13d) while levels of Mn (p < 0.0001, THSD) and Cu Fig. 11. The means and 95% confidence intervals for levels of (a) lipids and (p = 0.001, THSD) increased (Fig. 13b and f). Cu seemed to decrease carbohydrates and (b) protein in adult Polistes metricus reproductive females slightly in males in October as in gynes. Males showed a significant (gynes, foundresses and queens), workers, and males collected at different times increase in levels of Zn from emergence to September and a during the life cycle. Reproductive females and males are presented in the order significant decrease from September to October (p = 0.03, THSD; they appear in their life time. Sample sizes are the following: emerging gynes N = 45, September gyne N = 21, October gyne N = 7, spring gyne N = 10, foundress Fig. 13h). Based on the confidence intervals, there were similar N = 5, early queen N = 3, late queen N = 3, early worker N = 21, late worker N = 48, patterns in Fe and Ca (Fig. 13a and c), however; the fluctuation in emerging male N = 123, September male N = 19, October male N = 12. Ca was weak. 52 T.M. Judd et al. / Journal of Insect Physiology 56 (2010) 42–56

Fig. 13. The means and 95% conﬁdence intervals for levels of (a) Ca, (b) Cu, (c) Fe, (d) K, (e) Mg, (f) Mn, (g) Na, and (h) Zn in adult Polistes metricus reproductive females (gynes, foundresses and queens), workers, and males collected at different times during the life cycle. Reproductive females and males are presented in the order they appear in their life time. Sample sizes are the following: emerging gynes N = 44, September gyne N = 21, October gyne N = 7, spring gyne N = 10, foundress N = 5, early queen N = 3, late queen N = 1, early worker N = 21, late worker N = 50, emerging male N = 121, September male N = 20, October male N = 13.

3.9.4. Early vs. late workers higher in emerging gynes than late workers (p < 0.0001, THSD; Carbohydrates and lipids per unit mass, were significantly Fig. 11b). Early workers also had higher protein levels than higher in early workers than late workers (p = 0.0007, THSD; September and October gynes (psept = 0.0002, poct = 0.025, THSD). Fig. 11a). Protein (p < 0.0001, THSD) and Mn (p = 0.012, THSD) was Interestingly, late workers had higher caloric values than emerging higher in early workers based on the MANOVA analyses but these and October gynes (Fig. 12). Emerging gynes had higher levels of K results were not supported by confidence intervals (Fig. 11b, and lower levels of Mn than both groups of workers (Fig. 13d and f). Fig. 13f). According to the confidence intervals, Mg levels per unit Spring gynes had significantly higher protein levels than both mass were higher in early workers (Fig. 13e). groups of workers but lower caloric values than late workers (Fig. 11b). Mg levels per unit mass were higher in early workers 3.9.5. Workers vs. reproductives than spring gynes and foundresses (confidence intervals; Fig. 13e). There were two stages in adult reproductive females’ lives Spring gynes had higher Mn levels than late workers and this when they differed nutritionally from workers, during emergence difference was also true for foundresses and queens (p < 0.0001, and during the spring. Emerging (p = 0.0001, THSD) and September THSD; Fig. 13f). gynes (p = 0.01, THSD) had lower carbohydrates and lipids than Males had lower levels of all macronutrients than early workers early workers per unit mass (Fig. 11a). Protein levels were lower in (p 0.0053; Fig. 11a and b). October males had lower levels of emerging gynes than early workers (confidence intervals) but protein per unit mass than late workers (confidence intervals; T.M. Judd et al. / Journal of Insect Physiology 56 (2010) 42–56 53

Fig. 11b) and lower caloric values than both groups of workers important role in homeostasis in insects as well as possibly (p < 0.0001, THSD; Fig. 12). As with emerging females, males also regulating levels of stored lipids, carbohydrates and proteins had higher levels of K and lower levels of Mn than workers (Fig. 13d (Bhattacharya and Kaliwal, 2005a,b,c). and f). In addition, Mn levels in September and October males were lower than early workers (p 0.0025, THSD, Fig. 13f). 4.2. Differences between workers and reproductives during development 3.9.6. Gynes vs. males Nutritionally, there was very little difference between emerging Reproductive destined larvae had higher levels of lipids and K, males and gynes. Emerging gynes had higher levels of K and gained calories faster than workers throughout their devel- (p = 0.0001; Fig. 13d). Spring gynes had higher levels of carbohy- opment. Thus, the reproductives are storing more energy as they drates (p = 0.054, THSD) and proteins (p < 0.0001, THSD) than develop. Females can use these stores as adults during over- males but lower levels of lipids (p = 0.054, THSD; Fig. 11a and b). Cu wintering and for egg production the following year, while males was significantly lower in males than foundresses and queens can use this energy for territorial defense. (Fig. 13b). Mn levels in males were lower than spring gynes, Potassium is associated with only a few known enzymes foundresses, and queens (p < 0.0001, THSD; Fig. 13f). (Frausto da Silva and Williams, 2001). One clue to the possible reason for the difference in levels of K between workers and 3.10. Meconium reproductives during development comes from research with silkworms. Several studies have noted that adding K to the diet of There were no significant differences in micronutrient content silkworms increased the levels of glycogen, protein, and lipids in per unit mass in the meconium from worker and reproductive the fat body (Bhattacharya and Kaliwal, 2005b,c; Kochi and destined larvae. Kaliwal, 2006). Bhattacharya and Kaliwal (2005b) also found that the addition of KCl to the diet of silkworms had a minimal effect on 4. Discussion trehalose levels in the haemolymph but did raise the haemolymph protein levels. This suggests that K is primarily involved in the Overall, there were distinct differences of nutrient levels storage capacity of the fat body and may regulate its protein levels. between the different castes in P. metricus colonies. The major In this study, total lipid was high in reproductive destined larvae trends include (1) eggs had larger amounts of macronutrients and but protein and carbohydrates were not. However, this study several micronutrients per unit mass than the rest of the life stages. looked at total nutrient levels, and so the exact location of the (2) Queens (foundresses) had different levels of micronutrients per nutrients in the wasps was not measured. Other studies have found unit mass than the rest of the life stages. (3) There was little that reproductive Polistes do have higher storage proteins than difference between reproductive destined eggs and worker workers (Hunt et al., 2003) and larger fat bodies (O’Donnell, 1998; destined eggs, however; reproductive destined larvae had different Toth et al., 2009). Thus, the higher levels of K in reproductive nutrient levels than worker destined larvae. (4) Gyne destined destined larvae may promote an increase in levels of nutrients they larvae were not different than male larvae. (5) Both gynes and store. males showed fluctuations in macro and micronutrients during Magnesium also influences the production of trehalose in their lifetime. (6) There were differences in nutrient levels between silkworms (Murphy and Wyatt, 1965), and MgCl also increases the adult gynes and workers. Thus, there are distinct nutritional levels of glycogen, lipids and proteins in the fat body (Bhattacharya differences between workers and reproductives, and nutritional and Kaliwal, 2005c). Murphy and Wyatt (1965) found that Mg is an levels within adult individuals change over time. important cofactor for the formation of trehalose, and at low levels it increases the efficiency of glycogen synthase. Although Mg levels 4.1. Eggs were slightly higher in 4th and 5th instar worker larvae, there was very little difference between the different castes during larval Most studies have suggested that workers and reproductive growth and pupation. The formation of trehalose and glycogen development is adjusted by differences in nourishment during may be equally important for all castes at this stage. Indeed, the larval feeding (Wheeler, 1986; Hunt, 1991; O’Donnell, 1998), but levels of carbohydrates did not differ between castes during no one has looked at the nourishment of eggs. Although there were development. It is possible that Mg may be involved in regulating no differences between individual nutrients in worker and carbohydrate formation, whereas K is more important for enzymes reproductive destined eggs, the caloric value of reproductive involved in lipid storage. destined eggs was higher. The queen may possibly influence caste One minor trend was that Ca was retained more in reproductive development by manipulating the total nourishment of eggs or destined larvae than worker destined larvae. This difference was more likely, she may simply have more nutrients available with lost at the prepupa stage. Calcium has been found to be important workers doing most of the foraging, increasing nutrient availability for ovarian development in the moth Heliothis virescens (Pszczo- for the colony. One potential difficulty is that the eggs are lowski et al., 2008) and important for fecundity of the house cricket potentially losing macronutrients such as lipids and glycogen as Acheta domesticus (McFarlane, 1991). Thus, the retention of Ca in they develop (Sloggett and Lorenz, 2008). We did not know the age reproductive destined immatures may be related to the production of the eggs when they were collected but there is no difference in of gonads, however, all castes lost Ca during metamorphosis so the means and variances within individual nutrients. The levels of there may be an alternative reason for this difference. micronutrients should not change in the eggs during development Unlike the other nutrients, Mn was found to be higher during (although their use by the embryo may change) because the egg is worker destined larvae than reproductive destined larvae. Bowen self-contained and elements do not break down. (1950) examined Mn levels of different castes in several species Eggs had high levels of protein, lipids, K, and Mg. The high levels Vespidae. Although it was unclear when his collections took place, of protein and lipids are not surprising, because both nutrients are his findings for four species of Vespula were similar to what we needed to supply the developing embryo with materials and found for P. metricus. The larvae (presumably worker destined) energy to grow. Mg is important for phosphate reactions, showed an increase in Mn as they matured, and both queens and especially ATP, and Mg plays an integral part of the glycolysis workers had high levels of Mn. Pupae and young queens had lower pathway (Frausto da Silva and Williams, 2001). K plays an levels of Mn. The males also had low levels of Mn, but again it is 54 T.M. Judd et al. / Journal of Insect Physiology 56 (2010) 42–56 unclear as to their age. Bowen (1950) also measured Mn levels do not incorporate Zn (Phillips, 1970). The levels of zinc may be from a colony of Polistes fuscata pallipes (= P. metricus) and found tracking the reproductive abilities of the males. Indeed, we found similar levels of Mn in pupae, adult males, and adult females larger males, presumably with larger testes, have higher levels of (whether they were workers or reproductives is unknown) as our Zn in this study. study. His results suggest that the pattern of Mn found in this study All adults showed an increase in copper from emergence to may be a common pattern in all Vespidae. Ben-Shahar et al. (2004) adulthood. One group of enzymes that require Cu as part of protein showed that in honeybees levels of Mn increase in foragers. This structure, the phenoloxidases, may give some insight as to why Cu was due to an up regulation of the gene malvolio, a Mn transporter. was higher in adults than immatures. These enzymes are It is possible that malvolio is up regulated in worker destined larvae important in the production of melanin and in protein cross- and adults in Polistes. Toth et al. (2007) found that malvolio is linking when the cuticle hardens (Winzerling and Law, 1997). equally active in all adult female wasps. Two findings from this Indeed, low Cu levels have been found to cause albinoism in the study, (1) all adults had high levels of Mn and (2) workers that cricket A. domesticus (McFarlane, 1974, 1976). Thus, although the foraged more had higher levels of Mn, support this finding from a melanin is produced in the pupal form, much of the hardening nutritional perspective. The difference between the castes during takes place after adults emerge. Adult Polistes cannot fly during the larval stage suggests that malvolio may be involved in the caste first day after emergence presumably due to incomplete scler- determination of Polistes. Interestingly, the levels of Mn did not otization (Hunt, 2007). Perhaps the acquisition of Cu is necessary vary in the meconium. This suggests that reproductive destined for this to occur. Levels of copper plateau after the emergence of larvae are giving up their Mn to fellow colony mates. In the ant, workers and males, and levels of Cu only increase in reproductive Myrmica punctiventris, workers have high levels of Mn while the females once they become queens. Perhaps queens are collecting larvae have low levels. In this case, the larvae were most likely copper to give to the eggs. Melvin (1931) noted that significant passing the Mn to the workers (Judd and Fasnacht, 2007). A similar levels of Cu are present in egg cases of Blatta orientalis and the inner phenomenon may be happening in P. metricus. cocoon of Samia cecropia. These observations suggest that Cu may One interesting find is that there was little difference between be needed for egg case and cuticle formation. Male crickets also the nutritional levels of developing male and female reproductives. accumulate high levels of Cu in the testes relative to females This suggests that the larvae are not fed differently by the workers. (McFarlane, 1974, 1976). Thus, it is possible the accumulation of Cu Kin selection theory would predict that workers would preferen- in male Polistes at the beginning of their adult life enhances tially feed females over males. However, there is no evidence for spermatogenesis. McFarlane (1976) also found that Zn enhanced preferential feeding in P. metricus based on this study. the effects of Cu in crickets. Magnesium levels were high in overwintering gynes and in 4.3. Change in adults over their lifetime eggs. As mentioned above, Mg is important for glycolysis and the formation of trehalose (Murphy and Wyatt, 1965). Overwintering One major change in macronutrient levels in gynes was gynes using their stored lipids and storage proteins for survival between the fall to spring. The drop in lipids is probably due to whereas and eggs are using their stored nutrients for energy and the use of lipid stores as energy reserves. The fact that spring gynes growth. The decrease of Mg levels in the foundresses suggests that had the higher levels of proteins than the fall gyne or queen stages they are giving much of their stored Mg to the eggs. This may also suggests that the gynes are gathering proteins before they nest. explain the faster decrease in Mg per unit mass in early worker Gynes emerge at least a month before they begin nesting and what larvae. Indeed, Mg has been found to have a positive influence on they are doing during this time period is not well established. We the fecundity of Bombyx mori (Nirwanni and Kaliwal, 1995). Once propose that they are actually hunting for food to build up protein the worker caste begins to bring food back to the colony the queen stores for oogenesis before nesting. In a small pilot study, we can replenish her Mg supply. dissected the digestive tracts of two spring gynes and did find There are two possible reasons to explain the peak of Na in the small solid pieces in the gut suggesting they are consuming fall and the decrease of Na during the spring and summer. First, something. The nature of the material is unknown. sodium has been found to affect female fecundity and longevity in Males have two major events that affect their nutritional levels, the butterfly Thymelicus lineola (Pivnick and McNeil, 1987) and spermatogenesis and mating. Males spend much of their time cricket A. domesticus (McFarlane, 1991). In butterflies, Na is defending flowers (personal observation) and pursuing females. collected at the adult stage. Puddling males collect Na and transfer The loss of lipids during this study probably represents the loss of much of it to females during mating (Pivnick and McNeil, 1987; stored energy reserves, and this is supported by the loss of caloric Lederhouse et al., 1990; Smedley and Eisner, 1995, 1996; Beck value. Because they are defending flowers, males have access to a et al., 1999). Gynes may be collecting Na as they feed on nectar carbohydrate source, and that would explain the rise in during the fall. If the Na is important for fecundity, than the carbohydrate levels. In the earlier part of the season, males may females must be unable to replenish their Na during the spring and rely more on their lipid reserves. The less time they have to spend summer once the colony phase begins. This appears to be an feeding the more time they can devote to reproduction and interesting contrast to Mg and protein which are apparently territorial defense. Males in several species devote much of their recovered by the queens. time and energy to reproduction (Prestwich, 1994; Hunt et al., The second possible explanation for the observed pattern is that 2004) and defending territories (Frost and Frost, 1980). sodium has been found to be necessary for the production of The loss of protein and zinc in males over their lifetimes may be trehalose and glycerol in organisms preparing for winter. Several related to the loss of resources for spermatogenesis during studies have revealed the importance of Na in regulating cold reproduction. Studies on zinc-deprived male mammals have tolerance in insects (Kristiansen and Zachariassen, 2001; Zachar- shown a reduction in spermatogenesis (Hidiroglou and Knipfel, iassen and Kristiansen, 2003; Zachariassen et al., 2004). Thus, the 1984; Wallace et al., 1984; Ueda et al., 1991). Indeed male increase in Na in the fall may be necessary to survive the winter reproductive tracts tend to be one of the organ systems that are and replenishing the lost Na following winter might not be high in zinc (Frausto da Silva and Williams, 2001). Zinc is an necessary. Ashcraft (2009) found a similar pattern in Temnothorax important regulator of cell division (Beyersmann and Haase, 2001), curvispinosus in which queens had higher levels of Na than and it is incorporated into the outer fibers of sperm tails in workers, and the levels of Na were highest for both castes during mammals (Wallace et al., 1984). However, sperm tails of Drosophila the winter months. T.M. Judd et al. / Journal of Insect Physiology 56 (2010) 42–56 55

4.4. Differences between castes Ashcraft, B., 2009. A Study of the Seasonal Cation Content Changes in Naturally Occurring Colonies of the Ant Temnothorax curvispinosus. Southeast Missouri State University, Cape Giradeau. It was interesting to note that there were no stark differences in Beck, J., Muhlenberg, E., Fiedler, K., 1999. Mud-puddling behavior in tropical the levels of micronutrients, carbohydrates, or lipids of workers butterflies: in search of proteins or minerals? Oecologia 119, 140–148. Ben-Shahar, Y., Dudek, N.L., Robinson, G.E., 2004. Phenotypic deconstruction reveals and gynes. The latter result was surprising based on previous involvement of manganese transporter malvolio in honey bee division of labor. studies that did find differences between lipid levels of gynes and Journal of Experimental Biology 207, 3281–3288. workers (O’Donnell, 1998; Toth et al., 2009). One explanation for Beyersmann, D., Haase, H., 2001. Functions of zinc signaling, proliferation and the difference is that most of these studies focused on the fat differentiation of mammalian cells. BioMetals 14, 331–341. Bhattacharya, A., Kaliwal, B.B., 2005a. Biochemical content of the fat body and bodies whereas this study looked at total lipids. Workers could haemolymph of silkworm (Bombyx mori L.) larvae fed with mulberry leaves have had lipids in their gut or were temporarily storing lipids to fortified with mineral magnesium chloride (MgCl2). Philippine Agricultural feed larvae. Lipid levels decreased in overwintering gynes, but Scientist 88, 359–362. Bhattacharya, A., Kaliwal, B.B., 2005b. The biochemical effects of potassium chloride were not completely drained. on the silkworm, (Bombyx mori L.). Insect Science 12, 95–100. The only occasion where proteins were higher in gynes than in Bhattacharya, A., Kaliwal, B.B., 2005c. Synergetic effects of potassium and magne- workers was in the early spring. Workers are continuously hunting sium chloride on biochemical contents of the silkworm, Bombyx mori L. Caspian Journal of Environmental Science 3, 1–7. for food and the queens are helping to feed the larvae. Larvae are Blanchard, G.B., Orledge, G.M., Reynolds, S.E., Franks, N.R., 2000. Division of labour also giving nutrients back to adults through salivary secretions and seasonality in the ant Leptothorax albipennis: worker corpulence and its (Hunt, 2007). Thus, it is not too surprising that adults in the same influence on behaviour. Animal Behaviour 59, 723–738. Bo¨ hlen, P., Stein, S., Dairman, W., Udefriend, S., 1973. Flourometric assay of colony would have similar nutritional levels as they depend on the proteins in the nanogram range. Archives of Biochemistry and Biophysics same food sources. This is different from larger colonies in which 155, 213–220. the queen is dependent on the workers for food and thus, there are Bowen, V.T., 1950. Manganese metabolism of social Vespidae. Journal of Experi- mental Zoology 115, 175–205. distinct differences in macromolecule levels between workers, Bradford, M.M., 1976. A rapid and sensitive method for the quantitation of micro- queens and males. gram quantities of protein utilizing the principle of protein-dye binding. Analytical Biochemisty 72, 248–254. Brian, M.V., Abbott, A., 1977. The control of food flow in a society of the ant Myrmica 5. Conclusions rubra L. Animal Behaviour 25, 1047–1055. Cassill, D.L., Tschinkel, W.R., 1996. A duration constant for worker-to-larva tro- Overall, there were strong nutritional differences between the phallaxis in fire ants. Insectes Sociaux 43, 149–166. Cassill, D.L., Tschinkel, W.R., 1999a. Information flow during social feeding in ant developing worker destined larvae and reproductive destined societies. In: Detrain, C., Deneubourg, J.-L., Pasteels, J.M. (Eds.), Information larvae. However, males and female reproductive destined larvae Processing in Social Insects. Birkhaüser-Verlag, Basel, pp. 69–82. did not differ. Thus, during development one pathway may lead to Cassill, D.L., Tschinkel, W.R., 1999b. Regulation of diet in the fire ant, Solenopsis invicta. Journal of Insect Behavior 12, 307–328. a worker and another to a reproductive. In addition, nutrient levels Chapman, N.C., Beekman, M., Oldroyd, B.P., 2009. Several workers lay eggs in the in adults do change during their lifetime. Many of the micro- same brood cell in queenless honey bee (Apis mellifera) colonies. Insectes nutrient changes are possibly tied to reproduction in males and Sociaux 56, 103–105. Cohen, A.C., 2004. Insect Diets Science and Technology. CRC Press, New York. females. A few nutrients were found to rise in reproductive females Cotoneschi, C., Scognamiglio, F., Scala, C., Cervo, R., Strassmann, J.E., Turillazzi, S., from the founding phase to the queen phase. Queens might be able 2007. Larval sex identification in the paper wasp Polistes dominulus (Vespidae, to resupply once the colony is functioning allowing them to Hymenoptera). Insectes Sociaux 54, 132–135. Evans, J.D., Wheeler, D.E., 2001. Gene expression and the evolution of insect increase their reproductive output. polyphenisms. Bioessays 23, 62–68. In this paper, we suggest possible physiological areas in which Frausto da Silva, J.J.R., Williams, R.J.P., 2001. The Biological Chemistry of the the nutrients may be involved. The hypotheses for the function of Elements, 2nd edn. Oxford University press, New York. Frost, S.K., Frost, P.G.H., 1980. Territoriality and changes in resource use by sunbirds the micronutrients are based on what little we know about at Leonotis leonurus (Labiatae). Oecologia 45, 1432–1439. micronutrient function in insects. The information presented here Fukatsu, T., Ishikawa, H., 1992. Soldier and male of an eusocial aphid Colophina may be particularly useful as advances in genome-based research arma lack endosymbiont: implications for physiological and evolutionary reveals which enzymes are key players in physiological and interaction between host and symbiont. Journal of Insect Physiology 38, 1033–1042. developmental pathways. Our goal was to offer a place to start Gamboa, G.J., Wacker, T.L., Duffy, K.G., Dobson, S.W., Fishwild, T.G., 1992. Defense when examining the role of these nutrients in social wasps and against intraspecific usurpation by paper wasp cofoundresses (Polistes fuscatus, other social insects. Hymenoptera: Vespidae). Canadian Journal of Zoology 70, 2369–2372. Hidiroglou, M., Knipfel, J., 1984. Zinc in mammalian sperm: a review. Journal of Dairy Science 67, 1147–1156. Acknowledgements Hrassnigg, N., Crailsheim, K., 2005. Differences in drone and worker physiology in honeybees (Apis mellifera). Apidologie 36, 255–277. Hunt, J., Brooks, R., Jennions, M.D., Smith, M.J., Bentsen, C.L., Bussie`re, L.F., 2004. WewouldliketothankJoshHagerforhishelpwiththecollections. High-quality male field crickets invest heavily in sexual display but die young. We would like to thank John and Carol Kraemer, and Schwartz Bros Nature 432, 1024–1027. Hunt, J.H., 1991. Nourishment and evolution of the social Vespidae. In: Ross, K., Farm for letting us collect nests on their property. Thanks to James Mathews, R. (Eds.), The Social Biology of Wasps. Comstock Publishing Associ- Hunt, John Scheibe, James Robins, John Lawrence and two ates, Ithaca, pp. 426–450. anonymous reviewers for their helpful comments. This project Hunt, J.H., 1994. Nourishment and evolution in wasps sensu lato. In: Hunt, J.H., Nalepa, C.A. (Eds.), Nourishment and Evolution in Insect Societies. Westview was funded by a GRFC grant by Southeast Missouri State University. Press, Boulder, pp. 211–245. Hunt, J.H., 2007. The Evolution of Social Wasps. Oxford University Press, New York. Appendix A. Supplementary data Hunt, J.H., Amdam, G.V., 2005. Bivoltinism as an antecedent to eusociality in the paper wasp genus Polistes. Science 308, 264–267. Hunt, J.H., Buck, N.A., Wheeler, D.E., 2003. Storage proteins in vespid wasps: Supplementary data associated with this article can be found, in characterization, developmental pattern, and occurrence in adults. Journal of Insect Physiology 49, 785–794. the online version, at doi:10.1016/j.jinsphys.2009.09.002. Hunt, J.H., Dove, M.A., 2002. Nourishment affects colony demographics in the paper wasp Polistes metricus. Ecological Entomology 27, 467–474. Hunt, J.H., Jeanne, R.L., Baker, I., Grogan, D.E., 1987. Nutrient dynamics of a swarm- References founding social wasp species, Polybia occidentalis (Hymenoptera, Vespidae). Ethology 75, 291–305. Arrese, E.L., Canavoso, L.E., Jouni, Z.E., Pennington, J.E., Tsuchida, K., Wells, M.A., Judd, T.M., 2005. The effects of water, season, and colony composition on foraging 2001. Lipid storage and mobilization in insects: current status and future preferences of Pheidole ceres [Hymenoptera: Formicidae]. Journal of Insect directions. Insect Biochemistry and Molecular Biology 31, 7–17. Behavior 18, 781–803. 56 T.M. Judd et al. / Journal of Insect Physiology 56 (2010) 42–56

Judd, T.M., 2006. Relationship between food stores and foraging behavior of Pheidole Schultz, D.J., Huang, Z.Y., Robinson, G.E., 1998. Effects of colony food shortage on ceres (Hymenoptera: Formicidae). Annals of the Entomological Society of behavioral development in honey bees. Behavioral Ecology and Sociobiology 42, America 99, 398. 295–303. Judd, T.M., Fasnacht, M.P., 2007. Distribution of micronutrients in social insects: a Sloggett, J.J., Lorenz, M.W., 2008. Egg composition and reproductive investment in test in the termite Reticulitermes flavipes (Isoptera: Rhinotermitidae) and the ant aphidophagous ladybird beetles (Coccinellidae: Coccinellini): egg development Myrmica punctiventris (Hymenoptera: Formicidae). Annals of the Entomological and interspecific variation. Physiological Entomology 33, 200–208. Society of America 100, 893–899. Smedley, S.R., Eisner, T., 1995. Sodium uptake by puddling moths. Science 2701, Karsai, I., Hunt, J.H., 2002. Food quantity affects traits of offspring in the paper wasp 1816–1818. Polistes metricus (Hymenoptera: Vespidae). Environmental Entomology 31, 99– Smedley, S.R., Eisner, T., 1996. Sodium: a male moth’s gift to its offspring. Proceed- 106. ings of the National Academy of Sciences of the United States of America 93, Kochi, S.C., Kaliwal, B.B., 2006. The effects of potassium bromide on biochemical 809–813. contents of the fat body and haemolymph of crossbreed races of the silkworm, Sorensen, A.A., Busch, T.M., Vinson, S.B., 1985. Trophallaxis by temporal subcastes in Bombyx mori L. Caspian Journal of Environmental Science 4, 17–24. the fire ant, Solenopsis invicta, in response to honey. Physiological Entomology Kristiansen, E., Zachariassen, K.E., 2001. Effect of freezing on the transmembrane 10, 105–111. distribution of ions in freeze-tolerant larvae of the wood fly Xylophagus cinctus Sorenson, A.A., Busch, T.M., Vinson, S.B., 1985. Control of food influx by temporal (Diptera, Xylophagidae). Journal of Insect Physiology 47, 585–592. subcastes in the fire ant Solenopsis invicta. Behavioral Ecology and Sociobiology Kunieda, T., Fujiyuki, T., Kucharski, R., Foret, S., Ament, S.A., Toth, A.L., Ohashi, K., 17, 191–198. Takeuchi, H., Kamikouchi, A., Kage, E., Morioka, M., Beye, M., Kubo, T., Robinson, Stein, M.B., Thorvilson, H.G., Johnson, J.W., 1990. Seasonal-changes in bait pre- G.E., Maleszka, R., 2006. Carbohydrate metabolism genes and pathways in insects: ference by red imported fire ant, Solenopsis invicta (Hymenoptera, Formicidae). insights from the honey bee genome. Insect Molecular Biology 15, 563–576. Florida Entomologist 73, 117–123. Lederhouse, R.C., Ayres, M.P., Scriber, J.M., 1990. Adult nutrition affects male virility Toth, A.L., Bilof, K.B.J., Henshaw, M.T., Hunt, J.H., Robinson, G.E., 2009. Lipid stores, in Papilio glaucus L. Functional Ecology 4, 743–751. ovary development, and brain gene expression in Polistes metricus females. Markiewicz, D.A., O’Donnell, S., 2001. Social dominance, task performance and Insectes Sociaux 56, 77–84. nutrition: implications for reproduction in eusocial wasps. Journal of Compara- Toth, A.L., Kantarovich, S., Meisel, A.F., Robinson, G.E., 2005. Nutritional status tive Physiology A: Neuroethology Sensory Neural and Behavioral Physiology influences socially regulated foraging ontogeny in honey bees. Journal of 187, 327–333. Experimental Biology 208, 4641–4649. McFarlane, J.E., 1974. The functions of copper in the house cricket and relation of Toth, A.L., Varala, K., Newman, T.C., Miguez, F.E., Hutchison, S.K., Willoughby, D.A., copper to vitamin E. Canadian Entomologist 106, 441–446. Simons, J.F., Egholm, M., Hunt, J.H., Hudson, M.E., Robinson, G.E., 2007. Wasp McFarlane, J.E., 1976. The influence of dietary copper and zinc on the growth and gene expresion supports an evolutionary link between maternal behavior and reproduction of the house cricket. Canadian Entomologist 108, 387–390. eusociality. Science 318, 441–444. McFarlane, J.E., 1991. Dietary-sodium, potassium and calcium requirements of the Tschinkel, W.R., 1991. Insect sociometry, a field in search of data. Insect Sociaux 38, house cricket, Acheta domesticus (L.). Comparative Biochemistry and Physiology 77–82. A: Physiology 100, 217–220. Tschinkel, W.R., 1993. Sociometry and sociogenesis of colonies of the fire ant Melvin, R.O.Y., 1931. A quantitative study of copper in insects. Annals of the Solenopsis invicta during one annual cycle. Ecological Monographs 63, 425–457. Entomological Society of America 24, 485–488. Tschinkel, W.R., 1998. Sociometry and sociogenesis of colonies of the harvester ant, Molina, Y., O’Donnell, S., 2008. A developmental test of the dominance-nutrition Pogonomyrmex badius: worker characteristics in relation to colony size and hypothesis: linking adult feeding, aggression, and reproductive potential in the season. Insectes Sociaux 45, 385–410. paperwasp Mischocyttarus mastigophorus. Ethology Ecology and Evolution 20, Ueda, H., Kayama, F., Mori, N., Doi, Y., Fujimoto, S., 1991. Effects of dietary zinc 125–139. deficiency on protein secretory functions of the mouse testis. Archives of Murphy, T.A., Wyatt, G.R., 1965. The enzymes of glycogen and trehalose synthesis in Histology and Cytology 54, 401–410. silk moth fat body. The Journal of Biological Chemistry 240, 1500–1508. Van Handel, E., 1985a. Rapid-determination of glycogen and sugars in mosquitos. Nirwanni, R.B., Kaliwal, B.B., 1995. Effect of ferrous and magnesium sulphate Journal of the American Mosquito Control Association 1, 299–301. supplementation on some commercial characters of Bombyx mori L. Bulletin Van Handel, E., 1985b. Rapid-determination of total lipids in mosquitos. Journal of of Sericulture Research 6, 21–27. the American Mosquito Control Association 1, 302–304. O’Donnell, S., 1998. Reproductive caste determination in eusocial wasps (Hyme- Vinson, S.B., 1968. Distribution of an oil, carbohydrate and protein food source to noptera: Vespidae). Annual Review of Entomology 43, 323–346. members of imported fire ant colony. Journal of Economic Entomology 61, 712– Peakin, G.J., 1972. Aspects of productivity in Tetramorium caespitum. Ekologia Polska 714. 20, 56–63. Wallace, E., Calvin, H.I., Salgo, M.P., Dennis, J.E., Ploetz, K., 1984. Normal levels of Phillips, D.M., 1970. Insect sperm: their structure and morphogenesis. The Journal zinc and sulfhydryls in morphologically abnormal populations of spermatozoa of Cell Biology 44, 243–277. from moderately zinc-deficient rats. Gamete Research 9, 375–386. Pivnick, K.A., McNeil, J.M., 1987. Puddling in butterflies: sodium affects reproduc- Weeks, R.D.J., Wilson, J.T., Vinson, S.B., James, W.D., 2004. Flow of carbohydrates, tive success in Thymelicus lineola. Physiological Entomology 12, 461–472. lipids, and protein among colonies of polygyne red imported fire ants, Solenopsis Platt, T., Irwin, B., 1973. Caloric content of phytoplankton. Limnology and Oceano- invicta (Hymenoptera: Formicidae). Annals of the Entomological Society of graphy 18, 306–310. America 97, 105–110. Polak, M., 1993. Competition for landmark territories among male Polistes cana- Wheeler, D.E., 1986. Developmental, and physiological determinants of caste in densis (L.) (Hymenoptera: Vespidae): large-size advantage and alternative social Hymenoptera: evolutionary implications. The American Naturalist 128, mate-acquisition tactics. Behavioral Ecology 4, 325–331. 13–34. Polak, M., 1994. Large-size advantage and assessment of resource holding potential Wheeler, D.E., Buck, N., Evans, J.D., 2006. Expression of insulin pathway genes in male Polistes fuscatus (F.) (Hymenoptera: Vespidae). Animal Behaviour 48, during the period of caste determination in the honey bee, Apis mellifera. Insect 1231–1234. Molecular Biology 15, 597–602. Prestwich, K.N., 1994. The energetics of acoustic signaling in anurans and insects. Wheeler, D.E., Buck, N.A., 1992. Protein, lipid and carbohydrate use during meta- American Zoologist 34, 625–643. morphosis in the fire ant, Solenopsis xyloni. Physiological Entomology 17, 397– Pszczolowski, M.A., Olsen, E., Rhine, C., Ramaswamy, S.B., 2008. Role for calcium in 403. the development of ovarial patency in Heliothis virenscens. Journal of Insect Winston, M.L., 1987. The Biology of the Honey bee. Harvard University Press, Physiology 54, 358–366. Cambridge Mass. Reeve, H., 1991. Polistes. In: Ross, K., Mathews, R. (Eds.), The Social Biology of Winzerling, J.J., Law, J.H., 1997. Comparative nutrition of iron and copper. Annual Wasps. Comstock Publishing Associates, Ithaca, pp. 99–148. Review of Nutrition 17, 501–526. Robinson, E.J.H., 2009. Physiology as a caste-defining feature. Insectes Sociaux 56, 1–6. Zachariassen, K.E., Kristiansen, E., 2003. What determines the strategy of cold- Roisin, Y., 2000. Diversity and evolution of caste patterns. In: Abe, T., Bignell, D.E., hardiness? Acta Societatis Zoologicae Bohemicae 67, 51–58. Higashi, M. (Eds.), Termites: Evolution, Sociality, Symbiosis, Ecology. Kluwer Zachariassen, K.E., Kristiansen, E., Pedersen, S.A., 2004. Inorganic ions in cold- Academic Publishers, Dordrecht, pp. 95–120. hardiness. Cryobiology 48, 126–133.