Morphological Differences between Reproductive and Non-reproductive Females in the Social Wasp Mischocyttarus consimilis Zikán ( Hymenoptera : Vespidae )

Insect societies are characterized by the division of labor between the reproductive and non-reproductive castes, which is essential for the success of their societies (Robinson, 1992). In advanced eusocial species, castes are determined pre-imaginally and there is a clear physical divergence between queen and worker (Bourke, 1999). However, in less-advanced eusocial species, morphological divergences between females are less pronounced, and the factors affecting caste differentiation are not well known (Smith et al., 2011; Murakami et al., 2009; 2013; Jandt et al., 2014; Montagna et al., 2015). Nevertheless, evidence suggests that the morphological attributes, such as body size and ovarian state, can play an important role in establishment of the reproductive division of labor in independent-founding (IF) polistine wasps (Tibbetts & Dale, 2004; Tibbetts, 2006; Cervo et al., 2008). Abstract Phenotypic divergence plays an important role in establishment of the reproductive division of labor among castes in eusocial insects; however, little is known about this subject in social wasps. We tested the hypothesis that alpha foundressesin colonial foundations of the independent-founding polistine wasp Mischocyttarus consimilis Zikán show greater body size and ovarian development than auxiliary and solitary foundresses. The hypothesis was also tested that females produced in the late post-emergence stage (gynes) are larger than those produced at the beginning (workers) of the colony cycle of this species. Our results showed that the body size of alpha foundresses was significantly greater than that of auxiliary foundresses, but did not differ from that of solitary foundresses. In addition, the alpha had greater ovarian development than the auxiliary and solitary. We also demonstrated that gynes were significantly larger than workers. These results therefore suggest that the phenotype of females emerging in colonies of M. consimilis can vary with the progress of the colony cycle, and that body size and ovarian state are closely linked to the function performed by the foundress. Sociobiology An international journal on social insects


Introduction
Insect societies are characterized by the division of labor between the reproductive and non-reproductive castes, which is essential for the success of their societies (Robinson, 1992).In advanced eusocial species, castes are determined pre-imaginally and there is a clear physical divergence between queen and worker (Bourke, 1999).However, in less-advanced eusocial species, morphological divergences between females are less pronounced, and the factors affecting caste differentiation are not well known (Smith et al., 2011;Murakami et al., 2009;2013;Jandt et al., 2014;Montagna et al., 2015).Nevertheless, evidence suggests that the morphological attributes, such as body size and ovarian state, can play an important role in establishment of the reproductive division of labor in independent-founding (IF) polistine wasps (Tibbetts & Dale, 2004;Tibbetts, 2006;Cervo et al., 2008).
IF polistine wasps are an important model for study of the division of labor in less-advanced eusocial species.Colonies of IF polistine wasps can be initiated individually (haplometrosis), when a reproductive female unaccompanied by other females starts the colony; or by association (pleometrosis), when two or more reproductive females start the colony (West-Eberhard, 1969;Reeve, 1991).In haplometrosis, the foundress performs all the intra-and extra-nidal tasks until the workers emerge (Jeanne, 1972;Gadagkar, 1991).In pleometrosis, the colony tasks are shared among the foundresses, although the reproductive function is performed by one female, who dominates the others (Röseler, 1991;Keeping, 1992;Prezoto et al., 2004).Hence, females with a greater capacity for dominance are more likely to perform the colony reproductive function (Cervo et al., 2008;Zanette & Field, 2009).Since physical attributes are linked to the capacity for dominance, it has been suggested that body size could influence the reproductive strategies in pleometrotic foundations (Fukuda et al., 2003;Tannure-Nascimento et al., 2008).
In tropical regions, colonies of IF polistine wasps are started at any time of the year, so different colonial stages can occur in the same population in the same period.This nesting cycle pattern is termed asynchronous (Poltronieri & Rodrigues, 1976;O'Donnell & Joyce, 2001;Giannotti, 1998;Torres et al., 2011).The workers are produced at the beginning of the post-emergence stage, and they are mainly dedicated to defending the colony and foraging for food and nest material (Jeanne, 1972;Torres et al., 2011).In contrast, gynes are produced in the late post-emergence stage and generally invest in their own reproduction, contributing little to the maternal colony (Jeanne, 1972;Torres et al., 2011).It is therefore possible that the conditions for the development of immature vary with the progress of the colony cycle, due to fluctuations in the work force, which may be reflected in differences in the body size of emerging females (O'Donnell, 1998;Schmidt et al., 2012).
Mischocyttarus consimilis Zikán is an IF polistine wasp belonging to the subgenus Phi, one of the largest groups in the tribe Mischocyttarini, with the most extensive geographic distribution.In M. consimilis, the caste system is based on the division of tasks between the reproductive and non-reproductive females, in which the behaviors of oviposition, physical dominance, and nest presence characterize the reproductive caste, while behaviors such as alarm and resource foraging characterize the non-reproductive caste (Torres et al., 2012).In order to investigate morphological differences among castes in M. consimilis, we measured the body size and ovarian development of foundresses from haplometrosis and pleometrosis foundations, as well as the body size of females emerging in different phases of the colony cycle.

Field procedures and data collection
Colonies of M. consimilis were collected from the eaves of vacant houses situated in rural areas in the municipality of Dourados (22º13' S, 54º48' W) in Mato Grosso do Sul, from March 2011 through February 2012.To prevent the loss of individuals, colonies were collected in the early evening when foraging activity had ceased.All the individuals collected were killed by freezing (-20°C), prior to any manipulation procedure.
A total of 59 foundresses from 29 foundations were evaluated for body size and ovarian development.The foundresses were classified as: alpha (female performing the reproductive function in a pleometrotic foundation); auxiliary (each female subordinate to the alpha); and solitary (female of a haplometrotic foundation).Alpha foundresses were identified by direct observation of the oviposition behavior.We encouraged oviposition by the alpha by removing a larva with the aid of tweezers, leaving an empty cell in the comb.Our study did not require an ethics statement; however, we endeavored to minimize the stress caused by manipulation.
Additionally, we evaluated the body sizes of a total of 90 newly-emerged females (<24 h old), from 47 colonies.The newly-emerged females were classified as: worker (female created only by the foundresses); intermediate (female created by foundresses and workers); and gyne (female emerging in the colony reproductive phase).The reproductive phase in colonies of M. consimilis occurs in the final post-emergence, when males appear among the offspring (Torres et al., 2011).

Measurements of body size and ovarian development
The measurements of body size, obtained according to Gunnels (2007), were: head height; maximum distance between the inner margins of the eyes; minimum distance between the inner margins of the eyes; mesonotum length; mesonotum width; thorax length; mesopleura height; tibia length; and forewing length.For analysis of ovarian development, the gasters were fixed by immersion in Dietrich's solution, stored in 70% ethanol, and dissected in a petri dish containing physiological saline solution.The ovarian development index was obtained for each female, based on the mean length of the six largest oocytes (Sledge et al., 2004).All measurements of body size and ovarian development were obtained with the aid of a Stemi 2000C stereomicroscope (Carl Zeiss Microscopy, Oberkochen, Germany) fitted with a micrometer eyepiece.

Statistical analyses
All statistical analyses were carried out using R Statistical Environment software (R Development Core Team, 2013), with significance level of 0.05.Morphometric data were subjected to principal component analysis (PCA), and the first PCA component was used as the body size index (Gunnels, 2007).In all cases, the data for body size were normally distributed (Shapiro-Wilk test, P>0.05) and showed homogeneity of variance (Bartlett's test, P>0.05), so the differences among the categories of females were evaluated by parametric tests, including analysis of variance (one-way ANOVA) and Tukey's test (HSD) for multiple comparisons.The ovarian development data were not normally distributed (Shapiro-Wilk test, P<0.05), so differences among groups were assessed by nonparametric tests, including the Kruskal-Wallis and Steel-Dwass test for multiple comparisons.

Results
The PCA with the morphometric data for the foundresses recovered most of the variation among the groups, with the first principal component explaining 82.9% of the total variation and being positively associated with all the measured variables.The second principal component explained an additional 6.7% of the variation.Considering the first principal component as the explanatory variable in the analysis of variance, it was demonstrated that the foundresses differed in body size (one-way ANOVA, F 2.56 =9.26, P<0.001, Fig 1).The PCA performed with the morphometric data for the newly-emerged females recovered most of the variation among the groups.The first principal component was positively associated with all the variables and explained 75.8% of the total data variation.The second principal component explained an additional 5.6% of the variation.Considering the first principal component as the explanatory variable in the analysis of variance, significant differences in body size were found among the categories of newly-emerged females (one-way ANOVA, F 2.87 =6.86,P<0.01,Fig 4).Comparison of the groups showed that the gynes were significantly larger than the workers (Tukey's HSD, P<0.01,Fig 4) and the intermediates (P=0.02,Fig 4).However, the body size of the workers did not differ significantly from that of the intermediates (P=0.79,Fig 4).

Discussion
The results showed that foundresses of M. consimilis performing the reproductive function in pleometrotic foundations were significantly larger than their auxiliary.Several studies have shown that in pleometrotic foundations of IF polistine wasps, the colony reproductive function is aggressively disputed among potential females, and larger females have increased chances of subduing smaller competitors (Cant & Field, 2001;Tannure-Nascimento et al., 2008).Under these conditions, the reproductive organization is highly dependent on a dominance hierarchy, with larger females occupying higher positions in this hierarchy (Turillazzi & Pardi, 1977;Cervo et al., 2008).Larger females are therefore more likely to assume colony reproductive functions, suggesting that attributes established during immature development are relevant for the adult behavioral phenotype (Zanette & Field, 2009).Recent studies have shown that preimaginal caste determination can occur in IF polistine wasps, and that subtle morphological differences can be decisive in establishing the reproductive division of labor among foundresses (Fukuda et al., 2003;Judd et al., 2015;Montagna et al., 2015).These results provide evidence that body size affects the reproductive organization in pleometrotic foundations of M. consimilis.
It was also shown that the solitary foundresses were similar in size to the alpha.In a study with M. consimilis, Torres et al. (2011) demonstrated that haplometrotic foundation is the most common strategy adopted by the foundress.Our results suggest that large females with high reproductive potential invest in direct reproduction using the strategies of either fighting for the alpha position in pleometrotic foundations, or founding colonies individually.In addition, individual foundation could be an alternative strategy for large females who refrainfrom disputing the reproductive function in pleometrotic foundations.Nevertheless, these hypotheses need to be evaluated, considering the origins of the females that found colonies alone.
Alpha foundresses had a significantly greater degree of ovarian development, compared to the auxiliary.Furthermore, a significant positive correlation was found between body size and ovarian development, when the foundresses were evaluated together.These data suggest that the degree of ovarian development is closely relatedto both body size and the function performed by the foundress.Other studies have shown that foundresses performing reproductive functions (i.e., alpha and solitary), must invest more in ovarian development (Turillazzi & Pardi, 1977;Dropkin & Gamboa, 1981;Fukuda et al., 2003;Tannure-Nascimento et al., 2008).Moreover, auxiliary foundresses performing activities with high energy costs, such as foraging for resources, are likely to undergo ovarian regression (Keeping, 2000;Cervo et al., 2008).
It was also found that gynes were significantly larger than workers and intermediates, in agreement with other studies that evaluated differences in the body size of offspring females in IF polistine wasps (West-Eberhard, 1969;Haggard & Gamboa, 1980;Miyano, 1983).For example, females of Polistes fuscatus (Fabricius) produced in the late post-emergence stage, which were possibly gynes, were larger than workers produced at the beginning of this stage (West-Eberhard, 1969).Since the adult body size in holometabolous insects is fixed before emergence, it is possible that in IF polistine wasps it could be influenced by pre-imaginal trophic factors (Miyano, 1998;Karsai & Hunt, 2002).Differential feeding is known to be the main factors influencing caste determination in the honey bee, Apis mellifera, where larvae fed a highly nutritious diet (based on royal jelly) develop into queens (Winston, 1987).The effect of differential feeding in caste determination has also been demonstrated in the bumblebee, Bombus terrestris, where larvae that have access to greater amounts of food emerge as queens (Cnaani et al., 2002).In the case of IF polistine wasps, Judd et al. (2010) showed that the development of larvae of Polistes metricus into either workers or gynes was significantly influenced by the levels of nutrients in the hemolymph, including lipids, carbohydrates, and proteins, and that the differences were associated with possible variations in nutritional quality.In studies of the same species, Judd et al. (2015) showed that an increase in the quality of larval nutrition led to the development of larger females.In M. consimilis, workers and gynes emerge in different phases of the colony cycle, so the conditions for development of the immature are different (Torres et al., 2011).These findings suggest a link between body size and colonial stage, which needs to be further investigated.
However, in IF polistine wasps, other factors are associated with the caste ontogeny, which vary as the colony grows.For example, larvae of P. fuscatus developing in the pre-emergence stage are subjected to higher levels of vibration, compared to larvae that develop in the later postemergence stage (Suryanarayananet al., 2011).The vibration is produced by adults using their antennas to touch the walls of cells occupied by larvae.Adult females subjected to high levels of vibration in the larval stage show a worker-like phenotype, while those subjected to low levels of vibration show a gyne-like phenotype (Suryanarayanan & Jeanne, 2008;Suryanarayanan et al., 2011).This form of manipulation of the caste ontogeny is essential for the success of the colony, because it allows adults to control the caste production throughout the colony cycle.
The results presented in this study suggest that individual attributes such as body size and ovarian development are linked to the function performed by the foundresses in M. consimilis.In addition, it is clear that gynes are larger than workers, suggesting that factors associated with the pre-imaginal caste determination vary with the development of the colony.Knowledge of the caste ontogeny in IF polistine wasps sheds light on individual attributes that increase the chances for direct reproduction, in the context of disputes, and also helps to understand the conditions under which the castes are produced in less-advanced eusocial wasps.

Fig 4 .
Fig 4. Variation in body size index (mean ± SD) for newly-emerged females in colonies of Mischocyttarus consimilis.Different letters indicate statistical significance (P<0.05).