The Evolutionary Genetics of Sexual Size Dimorphism in the Cricket Allonemobius Socius
Total Page:16
File Type:pdf, Size:1020Kb
Heredity (2007) 99, 218–223 & 2007 Nature Publishing Group All rights reserved 0018-067X/07 $30.00 www.nature.com/hdy ORIGINAL ARTICLE The evolutionary genetics of sexual size dimorphism in the cricket Allonemobius socius KM Fedorka1, WE Winterhalter1 and TA Mousseau2 1Department of Biology, University of Central Florida, Orlando, FL, USA and 2Department of Biological Sciences, University of South Carolina, Columbia, SC, USA In recent years, investigations into the evolution of sexual phenotypic change in the next generation. We found that the size dimorphism have moved from a simple single trait, sexes differed significantly in their selection gradients as well single sex perspective, to the more robust view of multi- as several of their genetic parameters. Our predictions of variate selection acting on both males and females. How- next-generation change indicated that the within-sex genetic ever, more accurate predictions regarding selection correlations, as well as the between-sex genetic correlations, response may be possible if some knowledge of the should play a significant role in sexually dimorphic evolution underlying sex-specific genetic architecture exists. In the in this system. Specifically, the female size response was striped ground cricket, Allonemobius socius, females are the increased by approximately 178% when the between-sex larger sex. Furthermore, body size appears to be closely genetic correlations were considered. Thus, our predictions associated with fitness in both males and females. Here, we reinforce the notion that genetic architecture can produce investigate the role that genetic architecture may play in counterintuitive responses to selection, and suggest that affecting this pattern. Employing a quantitative genetic even a complete knowledge of the selection pressures approach, we estimated the sex-specific selection gradients acting on a trait may misrepresent the trajectory of trait and the (co)variance matrix for body size and wing evolution. morphology (that is, either a long-winged flight-capable Heredity (2007) 99, 218–223; doi:10.1038/sj.hdy.6800985; phenotype or a short-winged flightless phenotype) to predict published online 2 May 2007 Keywords: G-matrix; heritability; genetic correlation; selection gradient; microevolutionary change Introduction Early investigations into the maintenance of female- biased SSD centered on this latter hypothesis (Clutton- Sexual size dimorphism (SSD) or the difference in mean brock and Harvey, 1978; Berry and Shine, 1980; Gilbert size between the sexes is widely observed throughout the and Williamson, 1983). Unfortunately, these studies animal kingdom. Adaptive evolution of SSD is generally generally ignored selection on male body size, which explained through three hypotheses (Hedrick and can contribute equally to the dimorphic pattern. Further- Temeles, 1989). First, dimorphism via sexual selection more, other selective pressures may be acting differen- may occur when individuals of a particular size in one tially on female body size over time, which can sex are differentially successful via mate choice or mate considerably modify the relationship between size and competition (Andersson, 1994). Slatkin (1984) suggested fitness. Thus, research over the past decade has focused that dimorphism may also be maintained through purely on generating a more robust view of selection by ecological mechanisms. In this model the sexes occupy examining the interplay between differing selective different fitness optima for a certain trait, creating pressures acting on both sexes to estimate the net dimorphic niches. These different niches may originate selective force (Price, 1984; Ferguson and Fairbairn, through direct competition for resources (where di- 2000; Preziosi and Fairbairn, 2000). However, even a morphism itself is adaptive), or through divergent complete knowledge of selection has limits when ecological roles. Finally, dimorphism may be maintained attempting to predict next-generation phenotypic through differential reproductive selection. The most change. common reproductive model, the fecundity advantage To this end, some knowledge of the underlying sex- model, suggests that large female size is strongly specific genetic architecture for body size is needed. The associated with a greater reproductive output (Shine, reasons for this are manifold. First, body size can only 1988). respond to selection if sufficient genetic variation exists in the population (Roff, 1997). Second, a dimorphic response in size is possible only if the genetic correlation Correspondence: Professor KM Fedorka, Department of Biology, between the sexes for the homologous trait is less than University of Central Florida, 4000 Central Florida Blvd, Orlando, one, indicating the existence of sex-specific genetic FL 32816, USA. E-mail: [email protected] variation (Lande, 1980; Via and Lande, 1985; Ashman, Received 15 November 2006; revised 22 March 2007; accepted 23 2005). Third, genetic correlations within each sex may March 2007; published online 2 May 2007 modify the magnitude and/or direction of the body size Genetics of size dimorphism KM Fedorka et al 219 response (Roff, 1997). Fourth, even if ample genetic Cages were 10 Â 10 Â 8 cm and contained ground cat variance for body size exists and genetic correlations are food and a carrot slice (provided in excess), dampened less than one, the overall genetic architecture may still cheesecloth (water source and oviposition material) and constrain evolution by limiting the amount of genetic strips of brown paper towel for cover. Every 2 days, the variance that exists in the multivariate direction of food, carrot and paper towel were replaced. Cages were selection (Blows and Hoffmann, 2005). Therefore, just kept in a constant environment at 281C and a 12:12 (L:D) as the dynamics between differing selective pressures are photoperiod provided by a Percival incubator (Boone, important in determining net selective force, the IA, USA). Eggs were collected and offspring were reared dynamics between selection and the underlying genetic in the same environment as the parental generation. structure may provide valuable insights regarding the Upon adult eclosion the sexes were separated and evolution of a dimorphic trait (Ashman, 2005). maintained as virgins at low density (4–5 per cage). In the striped ground cricket, Allonemobius socius, Age of experimental crickets was 1072 days post females are the larger sex. Furthermore, body size eclosion. appears to be significantly associated with both male To estimate selection on each sex, we regressed femur mating success and female reproductive success length (estimate of body size) and wing morph (long- (Fedorka and Mousseau, 2002a, b). As with most sexually and short-winged coded as 1 and 0, respectively) against dimorphic systems, however, little is known regarding relative fitness. Characters were standardized to a mean the role that genetic architecture plays in dimorphism. To of zero and unit variance and relative fitness was this end, we employed a quantitative genetic design, calculated as orelative ¼ oindividual/oaverage (Fedorka and coupling sex-specific selection gradients with h2-matrices Mousseau, 2002a). Fitness estimates were based on (standardized form of the G-matrix comprised of trait mating success for males and reproductive success for 0 heritabilities and additive genetic correlations) for body females. Both univariate (b X) and multivariate (bX) size and wing morphology (that is, either a long-wing selection gradients were calculated for each sex-specific flight-capable phenotype or a short-wing flightless trait. Multivariate gradients were estimated via a phenotype). This allowed us to first estimate the amount standardized partial regression, allowing us to disen- of additive genetic variance for body size; second, tangle direct from indirect selection (Lande and Arnold, compare the sexes for differences in genetic parameters; 1983). third, test for significant correlations between the sexes To assess male mating success and female reproduc- for homologous traits; and fourth, model the response to tive success, two unrelated males were placed into a selection in order to assess the potential influence that mating arena (100 mm Petri dish) along with a randomly genetic architecture may have on dimorphic evolution. chosen, unrelated female. Each male (n ¼ 220) was tested on average five times against a different male with a new Methods virgin female used for each trial (n ¼ 1067). Thus, the male estimate of fitness was based on the proportion of Selection gradients (b) and additive genetic variance and successful mating attempts and female fitness was based covariances (G) form the basis of modeling microevolu- on the number of eggs produced. These estimates of tionary change in accordance with quantitative genetic fitness do not consider selection associated with dis- theory (Lande and Arnold, 1983; Roff, 1997). Thus, the persal of the long-winged morph. However, dispersal evolutionary trajectory of body size may be significantly selection is likely to be minor compared with the influenced by its genetic associations with other traits reproductive selection acting on male mating success under selection. Previous work suggests that wing and female fecundity. To help control for the potential morphology represents a potentially important corre- influence of mating experience, males were tested lated character. Evidence from other cricket systems against other males who shared an equal number of suggests