Peter Crnokrak, Department of Botany, University of Toronto, Toronto, Ontario, M5S 3B2, Canada. Tel.: (416) 978 5603; e-mail: firstname.lastname@example.org
Wing dimorphism, where some macropterous long-winged (LW) individuals can fly whereas micropterous short-winged (SW) individuals cannot, is common in insects and believed to be maintained in part by trade-offs between flight capability and reproductive traits. In this paper we examine differences in whole-organism respiration rate between wing morphs of the sand cricket Gryllus firmus. We hypothesized that maintenance of the flight apparatus would result in elevated CO2 respired because of the high metabolic cost of these tissues, which, in turn, constrain resources available for egg production in females. As the trade-off involves calling behaviour in males, we predicted no equivalent constraint on organ development in this sex. We found female macropters (particularly older crickets) had significantly higher residual respiration rates than micropters. In males, we found only marginal differences between wing morphs. In both sexes there was a highly significant effect of flight muscles status on residual respiration rate, individuals with functional muscles having higher respiration rates. Both female and male macropters had significantly smaller gonads than micropters. Whole-organism residual respiration rate was negatively correlated with fecundity: macropterous females with high respiration rates had smaller gonads compared with macropterous females with lower respiration rates.
A fundamental tenet of life history theory is that there exist constraints to the evolution of fitness related traits (Stearns, 1989; Roff, 1992). Thus an increase in the mean value of one fitness related trait is counter-balanced by the decrease in another, correlated trait. Such trade-offs may arise as a consequence of a limitation of resources that can be partitioned between the two traits (Pease & Bull, 1988; Stearns, 1989; Roff, 1992). Although numerous studies have documented negative associations between fitness related traits such as longevity and fecundity (for a reviews see Pease & Bull, 1988; Stearns, 1989; Roff, 1992), many have failed to document such associations or have instead found positive correlations where negative ones were predicted (van Noordwijk & de Jong, 1986; Pease & Bull, 1988; Roff, 1992). Such anomalous results are explicable if, as is likely, constraints involve more than two traits. For example, a physiological trade-off may potentially exist in terms of a constraint of resources but may not be manifest as a realized trade-off, because of environmental variation or behavioural modifications that compensate for the internal (physiological) constraint (van Noordwijk & de Jong, 1986; Pease & Bull, 1988). An experimental approach that has been successful in demonstrating the existence of physiological constraints has been to manipulate available resources and measure the associations between traits under the assumption that trade-offs will be manifested under conditions of resource limitation (van Noordwijk & de Jong, 1986; Zera & Brink, 2000).
Because of the ease of scoring differences in life history strategies (and associated differences in fitness related traits) between wing morphs, wing dimorphisms in insects afford researchers a useful system by which to examine and model the evolution of trade-offs. Many species of insect exhibit intrasexual polymorphisms for wing length where the long-winged morph (macropter, LW) is usually capable of flight, whereas the short-winged morph (micropter, SW) is flightless. Macropters, in addition to possessing long wings, also frequently possess large (15–20% of body weight) functional flight muscles (dorso-longitudinal muscles-DLM) and flight fuels (Crnokrak & Roff, 2000). Functional DLM (many species have the ability to histolysis their muscles in times of need; Roff, 1989), have respiration rates that are 10-fold greater than nonfunctional muscles (Zera et al., 1997, 1998). Although macropters can fly, and in doing so avoid unfavourable changes in their habitats (e.g. succession), they are at a significant energetic disadvantage, because of the resources required to maintain the functional flight apparatus (Mole & Zera, 1994; Zera et al., 1994, 1997; Crnokrak & Roff, 2000; Stirling et al., 2001). The diversion of resources for the maintenance of the dorso-longitudinal muscles from reproductive traits such as egg production in females or calling in males, is believed to be the primary factor contributing to substantial differences between wing morphs in these traits (Crnokrak & Roff, 1995, 2000; Roff, 1997). The dimorphism represents an extreme case of the antagonism between life-history traits associated with flight capability and reproduction (the flight–oogenesis syndrome; Johnson, 1969; Harrison, 1980; Roff, 1984, 1986a). Trade-offs between wing morph and reproductive traits have been extensively studied in both females (Roff, 1995; Roff et al., 1997; Zera et al., 1997) and males (Crnokrak & Roff, 1995, 1998a,b, 2000) of the sand cricket, Gryllus firmus. Although well studied with respect to phenotypic and genetic correlations between wing morph and characteristics such as egg production, call duration, flight propensity and wing muscle histolysis there have been relatively few studies on the metabolic basis of the trade-off in G. firmus or any other wing dimorphic insect.
Differences between morphs in terms of fecundity in females and mating potential in males rests, in part, on the assumption that the wing morphs have differential whole-organism respiration rates (Mole & Zera, 1994; Zera et al., 1997). The elevated respiration rate of functional pink DLM in macropters (Zera et al., 1997), is believed to result in a whole-organism elevated respiration rate which in turn results in fewer resources being available for such general functions as daily maintenance and more specific functions as egg production in females or calling (= mate attraction) in males. Previous studies have shown that nonfunctional muscles are colourless because of reduced cytochrome content (indicating a lack of mitochondrial activity: Shiga et al., 1991; Gomi et al., 1995), which in turn contributes to a lower respiration rate of this muscle type (Zera et al., 1997). Nonfunctional muscles can be either underdeveloped or histolysed, previously functional, muscles (Zera et al., 1997). Micropterous crickets typically possess underdeveloped white muscle whereas macropterous crickets possess red functional muscles (Crnokrak & Roff, 2000) which later histolyse into white nonfunctional muscles (Zera et al., 1997). Zera et al. (1997) have shown that underdeveloped white muscle have a significantly lower basal respiration rate compared with histolysed white muscles. These differences result in even greater energetic differences between micropterous and macropterous morphs than would occur if all nonfunctional muscles were metabolically similar. Although macropterous crickets are predicted to have higher respiration rates because of their flight musculature, differences in relative body weights of other organs between morphs may have `compensatory' effects and result in nonsignificant differences in whole-organism respiration rates. Ovarian tissue is known to have a 10-fold lower respiration rate than flight muscle in the closely related G. assimilis (Zera et al., 1998), and because macropteorus crickets have significantly smaller ovaries than micropterous crickets (Roff et al., 1997; Zera et al., 1997), they may have a lower proportion of total body mass composed of relatively metabolically inactive tissue. Macropters could also compensate for the increase in metabolic rate to maintain functional DLM by decreasing the weight of other organs that are relatively metabolically active. Although this remains a possibility, no such compensatory effects were found in G. firmus (Mole & Zera, 1994; Zera et al., 1997) or G. rubens (Mole & Zera, 1993) but may exist in G. assimilis (Zera et al., 1998). Instead, the substantial metabolic needs of functional DLM probably results in measurable differences in whole-organism respiration rates between wing morphs in both sexes. Two studies by Zera and colleagues in G. firmus (Zera et al., 1997) and G. assimilis (Zera et al., 1998), previously documented differential resource allocation in females between wing morphs and differential metabolic activity of body parts (DLM, thorax and ovaries). From these studies they inferred differential whole-organism respiration rates between wing morphs. In this paper, we present a different technique to indirectly estimate insect respiration (respirometry) and examine the effect of sex, wing morph and age on variation in respirometry rate.
Based on the assumptions outlined above, we predict that the maintenance of a functional flight apparatus will result in significant physiological differences between micropterous and macropterous crickets and result in a trade-off between wing morph and reproductive potential. Specifically, using CO2 production as an index of metabolic rate (respirometry used to indirectly measure respiration), we predict: (1) macropterous individuals should show higher average CO2 output than micropterous individuals. We predict that this effect will be found in both females and males. Although our approach in estimating respiration rate is also indirect, it is a different approach than that used by Zera and colleagues, the results are useful for comparison purposes. (2) Differences in whole-organism respirometry rate are predicted to be the result differences in DLM status between wing morphs: crickets with large and functional DLM should have higher CO2 levels than those with small and nonfunctional DLM. (3) A trade-off should exist between respirometry rate and gonad mass in females. Females with large functional DLM and resulting, high respirometry rates, are predicted to be constrained with respect to the amount of resources that can be invested in egg production compared with crickets that do not have functional DLM, and in turn, have smaller gonads. We predict that this trade-off will be particularly pronounced in older females where gonads are substantially larger than in younger females. We tested these predictions using a laboratory population of G. firmus.
Materials and methods
Gryllus firmus is a large (live weight, approximately 0.75 g) ground dwelling cricket found in the American south-east as far north as Connecticut (Alexander, 1968; Harrison, 1985). It is usually found in early successional, sandy areas, which being impermanent, favour the evolution of wing dimorphism (Roff, 1990, 1994b). Individuals in this experiment originated from approximately 40 individuals (20 males, 20 females) from a locality in northern Florida (Roff, 1986b). Stock crickets were maintained in the laboratory at approximately 100–300 breeding adults at a temperature of 25–30 °C for approximately 40 generations before being used for the respirometry experiments. Crickets used for the experiments were maintained in circular 4 L buckets (diameter, 21 cm; height, 15 cm) in incubators set at a photoperiod of 15 h light, 9 h dark and at a temperature of 28 °C with ad libitum food and water at a density of 30 nymphs per bucket. Food consisted of crushed Purina rabbit chow. Water was provided in glass shell vials corked with cotton. On the day of their final ecdysis (day nymphs became adults), individuals were removed from these rearing buckets and placed in individual quarters containing food and water until they were of an appropriate age (see age categories below) to be used in the experiment.
For each adult cricket CO2 output was measured over a 30 min period, following a 10 min acclimation period, using a Sable Systems Tr-3 respirometry apparatus (Sable Systems, Las Vegas, NV, USA) for flow-through measurements (consisting of a Licor 6251 infrared CO2 detector together with a computer-controlled switching device for recording baselines, flow controller and A/D converter board attached to a computer for data recording). The flow rate was maintained at 190 mL min–1, the temperature at 30 (±1) °C and activity monitored using a Sable AD-1 activity detector. The following description of its operation is cited from the technical description provided by Sable. The AD-1′s principle of operation is light-based. The subject animal is contained in a RC-M precision respirometry chamber, which has transparent walls and two reflective metal end-pieces. The AD-1 activity detector encloses the transparent glass envelope of the chamber with two curved reflective surfaces. These reflectors have slight random surface irregularities that randomly disperse reflected light throughout the chamber. The subject animal is therefore enclosed on all sides by reflective surfaces. The intensity of this reflected light at any given point is a sensitive qualitative function of the position of any animal within the chamber. Fluctuations in the reflected light intensity are amplified and converted to voltage fluctuations (zero for no activity, and increasing amplitudes with increasing activity levels). The chamber and reflectors are sealed to eliminate the influence of ambient light fluctuations. The activity detector's own light source is in the near-infrared (about 900 nm wavelength), which has very low intensity in most laboratory environments, further minimizing interference. This light is undetectable to all known arthropods, and is of too short a wavelength and too low a power to cause significant heating effects.
Respirometry rates in 306 crickets were measured. Some crickets eclosed over a weekend and their exact age was not known. However, adults were selected so that we could divide them into two age groups, 1–3 days and 5–7 days. Immediately after respirometry was measured, we preserved each cricket in Bouins fluid in individual glass vials. Crickets were preserved in this manner for an average of 2 months at which time they were dissected to obtain organ weights. Crickets were dissected to assess DLM status and mass, and gonad mass. DLM status was scored on a five point scale: N=normal (muscle texture uniform, firm and dense), PN=partially histolysed/normal (initial signs of histolysis, nonuniform texture in minute amounts), P=partially hisolysed (approximately half of muscle mass shows signs of histolysis), PH=partially histolysed/histolysed (extensive histolysis but there still remain sections of dense functional muscle), H=histolysed (fully degenerated). DLM were carefully removed from the body of the crickets (without attached cuticle or fat), placed on a preweighed glass slide and weighed. Gonads from both males and females were then dissected, being careful to remove all loose eggs contained in the abdomen of gravid females. All weight measurements were recorded to 0.0001 g.
Relative weights of organs
For all of the following tests wing morph (0=SW, 1=LW), sex (0=female, 1=male), and age group (1, 2) were coded as categorical. We found significant differences in total body mass between micropterous (SW) and macropterous (LW) crickets and between younger and older crickets (Table 1). LW crickets had significantly larger body mass than SW crickets (only in younger individuals) and older crickets of both sexes had significantly larger body mass than younger crickets (Table 1). Because our interest is in the relative size of the gonads and flight muscles, we tested the effect of wing morph and age group on organ mass after converting them into a proportion of total body mass (arc-sine square-root transformed proportion of organ weight). We also ran separate analyses for females and males because they exhibit substantial differences in gonad size. We found significant interaction effects between wing morph and age group for both females and males (Table 2) for the ANOVA involving gonad mass. LW females had substantially smaller gonads compared with SW females only when crickets were old (47% reduction in LW females; Fig. 1a). In males, we found significant differences between wing morphs for gonad mass for both age groups (Table 2): LW males had smaller gonads compared with SW males (age group 1: 19% decrease, age group 2: 7% decrease, Fig. 1a). The decrease in difference in gonad mass with age between wing morphs (19 to 7%), is the reason for the significant interaction effect between wing morph and age group in the ANOVA (Table 2). We found significant differences in DLM mass (Table 2) between wing morphs: as predicted, LW crickets had substantially larger flight muscles (% of total body mass) than SW crickets (Fig. 1b) for both sexes. There was a significant decrease with age in DLM mass in LW females, but not males (Fig. 1b).
Table 1. Mean (±SE) total body mass (g) and analysis of variance of the effects of wing morph, age group and sex on total body mass in Gryllus firmus.
Table 2. ANOVA results testing for differences in relative (arc-sine square-root proportion of total mass) gonad mass and dorso-longitudinal muscle mass (DLM), between wing morphs and different age group crickets for females and males separately
Overall, we found significant structural differences in the body composition of male and female crickets, between the two wing morphs and age groups (Fig. 2). An appreciable increase in gonad mass occurs between younger and older females and this is substantially larger in SW crickets (Fig. 2). In contrast, gonad mass does not change in males (Fig. 2). Additionally, the relative weight of the DLM is substantially larger in the LW morph than the SW morph for both sexes and age groups. The greatest differences between wing morphs in terms of relative organ weights are for older females (Fig. 2). These differences are hypothesized to affect the whole organism respiration rates of crickets because the thorax contains metabolically active muscles whereas gonads have relatively low metabolic activity (see Zera et al., 1997).
Associations between respirometry rate and wing morph/DLM status
Differences between wing morphs in total CO2 respired should exist only during the time when the flight muscles are functional in crickets. Therefore, in this analysis the principle variables of interest are wing morph and age group. Crickets with larger total body mass should respire more CO2 since there is more metabolically active tissue. Further, activity level will also affect the amount CO2 respired: the more a cricket moves in the test chamber the greater it has respired CO2. Because of these possible effects, we analysed the data using the `raw' data and by taking into account effects attributable to body size and/or activity. To do the latter we used stepwise multiple regression using Mallows' CP statistic (Venables & Ripley, 1997; p. 220). To normalize the data we log transformed CO2, body mass and activity score. Because the sexes exhibit large differences in organ structure (e.g. gonads), we ran separate analyses for females and males.
Stepwise regression in females gave the best model as one that included all possible terms except for the four-way interaction. This model accounted for 78% of the variance and was overall highly significant (F15,134=31.05, P < 0.00001). However, the regression consisting of the additive terms and the interaction wing morph by age group accounted for 73% of the variance and was also highly significant (F5,144=77.87, P < 0.00001). In males the best model from the stepwise regression included all additive terms and the wing morph by age group interaction (R2=0.79, F5,143=105.0, P < 0.00001). For simplicity of exposition we discuss the effects of wing morph and age group after correcting for the additive effects of log(body mass) and log(activity) by taking residuals from the linear regression of log(CO2) on these two `nuisance' variables.
In females, LW crickets had significantly higher absolute respirometry rates than SW crickets for each age group (Fig. 3a, Table 3). In males, no significant differences were found between LW and SW crickets for absolute respirometry rate but there was a significant interaction with age (Fig. 3c, Table 3). For both sexes, absolute respirometry rate significantly increased with age (Table 3). When we used residual respirometry rate in place of the absolute measure, there were significant interaction effects in both males and females (Figs 3b,d, Table 3). The only instance where LW crickets had higher respirometry rates than SW crickets was for older females (compare Fig. 3b with d). In females, residual respirometry rate decreased with age in SW crickets whereas it slightly increased in LW crickets. The patterns of mean residual respirometry rate involving wing morph and age group in females most likely reflects the patterns of relative body weight composition outlined in the previous section: the relative mass of metabolically active tissue (thorax and DLM) increases with age much more in LW than SW females, although relatively inactive tissue (gonads) increases substantially more in SW females (Fig. 2). In contrast, young SW males (1–3 days), had higher residual respirometry rates compared with LW males (Fig. 3d, Table 3). As in females, in males, residual respirometry rate decreased with age in SW crickets while it increased with age in LW crickets. The substantial decrease with age in residual respirometry rate seen in SW females and not found in SW males, is most likely due to the substantially larger gonads of females compared with males. Additionally, gonads increase in size with age much more in females than males (Fig. 2).
Table 3. ANOVA results for relationships shown in Fig. 3: respiration rate and residual respiration rate as a function of age and wing morph for females and males separately.
The addition of the variable DLM STATUS (see Materials and Methods for categories) in the above regression resulted in a nonsignificant wing morph effect (females: F1,141=0.004, P=n.s.; males: F1,142=0.68, P=n.s); thus significant differences in respirometry rate between LW and SW crickets can be partially attributed to differences in DLM status. In general, crickets with functional DLM muscles had substantially higher residual respirometry rates (females: F4,143=2.981, P < 0.05; males: F4,144=6.605, P < 0.0001; Fig. 4). A closer examination of the relationship between adjusted respirometry rate and flight muscle status reveals that in both females and males the above patterns are found in LW crickets while no SW crickets were found to have functional muscles (Fig. 4). These results indicate that the differences in residual respirometry rates documented between SW and LW crickets (primarily in older females) is, to some degree, the result of the possession of large functional and semifunctional flight muscles in LW crickets (see discussion for qualifications).
Respirometry rate and gonad mass
We tested the hypothesis that whole-organism respirometry rates covary with gonad mass using the same approach as in the previous section. First we ran a stepwise regression to find the best model relating the dependent variable log(gonad mass) to the five variables log(CO2), log(body mass), log(activity), wing morph and age group and their interactions. We then compared this model with that composed of the two `nuisance' variables (body mass and activity) plus the three other variables and their interactions. We tested these effects in each of the sexes separately because, as stated previously, gonads are functionally very different between females and males.
The best stepwise model for females included a large number of interaction terms and accounted for 82% of the variances, whereas the simpler model accounted for 76% of the variance. In males the best stepwise model accounted for 82% of the variance and the simpler model accounted for 79% of the variance. Because the difference between the two models is small, for each sex we analysed using ANCOVA the residual respirometry rate as described in the previous section.
In females, we found a significant three-way interaction between age group × wing morph × residual respirometry rate (F1,141=4.392, P < 0.05). Plotting individual regressions of log gonad mass on residual respirometry rate for the two wing morphs and age groups reveals that the only significant relationship is for older LW females, where gonad mass was substantially larger than for the younger LW females (5–7d; Fig. 5). In young females (both LW and SW), a positive, nonsignificant, relationship exists between log gonad mass and residual respirometry rate (Fig. 5). In older SW females the relationship is negative, but nonsignificant (Fig. 5). Therefore, females with functional flight muscles (LW) exhibit a trade-off between respirometry rate and gonad size, which does not exist in females with nonfunctional flight muscles (SW). Additionally, the trade-off is only found later in life when the gonads in females are packed with eggs, presumably because the constraint of available resources to gonad growth is not manifested until they are large.
In males, no significant effects of wing morph, adjusted respirometry rate or interactions on gonad mass were found, although we did find an effect of age (F1,145=16.027, P < 0.0001). The lack of a realized constraint on gonad formation in males is most likely because of their substantially smaller size compared with older females (age 2). Additionally, the lack of an effect of adjusted respirometry rate on gonad mass in males means that the significantly larger gonads in SW males reported above is the result of other factors than the proposed energetic trade-off involving whole-organism respiration.
Previous studies have documented a cost to flight capability in the wing dimorphic cricket G. firmus involving fecundity and age at reproduction in females and mating probability and call behaviour in males (see introduction). In this study we provide evidence that the differences in gonad size between SW and LW female crickets are associated, in part, with differences in respired CO2 and flight muscle status. In males, as predicted, only differences associated with flight muscle status were observed. We found significant differences in live, whole-organism absolute respirometry rates between the two wing morphs of females, with macropterous individuals having substantially higher respired CO2 levels compared with SW individuals (Fig. 3, Table 3). In males the pattern was different in showing a reversal in the older age group, SW males having a higher respiration rate than LW males.
Comparisons among wing morphs and ages are potentially confounded by variation in body size and activity. When corrected for cricket size and activity, SW individuals showed a decline in residual respirometry rate with age, whereas LW individuals showed an increase (Fig. 3). This pattern reflects differential growth of metabolically active (flight and thorax muscles, flight fuels) and relatively inactive (gonads) tissues (Fig. 2). In females the relative mass of thorax and flight muscles increases with age more in LW than SW individuals, whereas gonad mass increases substantially more in SW females (Fig. 2). SW males have relatively larger gonads, although these do not show a noticeable change in relative size with age. Differences in residual respirometry rate were also correlated with differences in DLM weight and status between wing morphs (Figs 1b and 4, respectively). DLM status explains most of the variation in respirometry rates between the wing morphs (analogous effects are discussed in Cisper et al., 2000). On the other hand, the relationship between DLM status and respirometry rate maybe partially because of the large stores of lipid flight fuels that LW crickets harbour which are energetically expensive to biosynthesize and thus may contribute to higher respired CO2 (Mole & Zera, 1993; Zera et al., 1997, 1998). Crickets with large, functional DLM had significantly higher residual respirometry rates than crickets with nonfunctional DLM (Fig. 4). This relationship was most pronounced in the comparison between wing morphs as micropterous crickets very rarely possessed semifunctional flight muscles and were never found with fully functional DLM. We found a significant negative relationship between whole-organism residual respirometry rate and gonad mass in older (5–7 days) females: in macropterous females, crickets with higher residual respirometry rates had smaller gonads (Fig. 5). The negative relationship in macropterous females is consistent with the hypothesis of a constraint of biomass investment in gonads due to energy needed to maintain functional DLM (and/or flight lipids, see above). In micropterous females, as no crickets were found with functional DLM, no such constraint exists and hence no relationship is predicted and non observed. These effects are only seen in older females: younger females (and males) have substantially smaller gonads and therefore are probably not subject to an allocation constraint. Our analyses illustrate the importance of measuring the relative weights of various organs as opposed to total body mass in crickets as different organs have different metabolic rates (see Zera et al., 1997).
Although we found significant negative correlations between whole-organism respirometry rates and gonad mass in females, these trade-offs were only present in older macropterous females. In crickets, flight muscles form during the last nymphal stage of development, before gonads are large and full of developed eggs (Roff, 1989). This temporal separation of organ development precludes a direct trade-off of resources devoted to organ formation between DLM and gonads (Stirling et al., 2001). The trade-off between macroptery and reproductive traits is therefore primarily a function of having to maintain the DLM (and associated flight apparatus tissues) in a functional state. The trade-off is also most pronounced (differences between wing morphs) when the gonads in females are large and full of eggs (previous study, adults 7–10 days; Roff, 1994c. present study; adults 5–7 days old). In early adulthood, when female gonads are relatively small and devoid of developed eggs, we found a positive but nonsignificant association between respirometry rate and the size of gonads. Previous studies (Roff, 1989; Stirling et al., 2001; Crnokrak et al., 2002) and the present (Fig. 1b) have shown that the flight muscles (DLM) begin to histolyse at around day seven of adult life, which is the time when gonads are full of eggs in females. With the complete histolysis of the flight muscles, the energetic resources devoted to them can be used for other functions as can the stored resources present in the energy rich muscles. Indeed, previous studies have shown that later in life, the fecundity curves in female G. firmus cross and macropters have larger (but not significantly so) mean reproductive outputs than micropters (Roff, 1994c). The separation of functional flight apparatus maintenance and reproduction is a component of the oogenesis flight syndrome found in monomorphically winged migrant insects (for reviews see Dingle, 1985; Rankin et al., 1986). In wing dimorphic insects SW individuals omit the migrant phase whereas LW individuals potentially show the oogenesis flight syndrome, although the migrant phase can be omitted by immediate histolysis of the flight muscles if conditions dictate.
If environmental factors that necessitate movement from one habitat to another can be predicted, then selection will favour those individuals that maintain a functional flight apparatus only during that time when such migration is necessary and the subsequent breakdown of the apparatus when not needed. If environmental factors are not predictable, then selection might favour a mix of flight capability and lowered maximal reproduction (Roff, 1994a). Interestingly, there are a number of locust and cricket species which, when faced with such variable timing of environmental conditions that necessitate flight, have evolved developmental mechanisms that allow for discrete inter-reproductive migration phases more than once during adult life (see Tanaka, 1991 for a discussion). For example, the tambo cricket, Velarifictorus parvus oviposits soon after eclosion, then goes through a flight capable phase (during which time it cannot reproduce) after which it is flightless and reproductive again (Tanaka, 1991). Although the long flight wings are maintained throughout adult life in G. firmus, the temporal separation of flight capability (as evidenced by the presence of functional DLM) and reproduction in this species is similar to other species where the separation of the two is more readily noticeable by the outward changes that take place at specific times in development (e.g. flight wing dealation: Roff, 1989). Evidence suggests that selection generally favours the distinct separation of life history phases as opposed to favouring individuals that mix both flight capability and reproduction (albeit to a lesser degree) at the same time (for a discussion see Fairbairn & Roff, 1990).
An unexpected finding of our analyses was the significant difference in gonad size between SW and LW males. When total body mass was accounted for, micropterous males had on average, 19% larger gonads than macropterous males. This difference was primarily found in younger males. Early investigations on trade-offs to macroptery in male G. firmus found no measurable differences in gonad size (Zera et al., 1997) and a subsequent lack of difference in siring ability between wing morphs (Roff & Fairbairn, 1993). Having found no difference in gonad size in males, Roff and Fairbairn postulated that trade-offs to macroptery might involve mate attraction behaviour, as males expend large amounts of energy to attract females to mate. This hypothesis was later verified in a series of behavioural experiments (Crnokrak & Roff, 1995, 1998a,b; Crnokraket al., 2002). Our present results show that there are morphological, as well as behavioural, differences between micropters and macropters. Although gonads are relatively large for male body size (7.6%), they are substantially smaller than female gonads (14.7% of total body size). Unlike other Orthopteran species such as katydids that produce large energetically expensive nutritive spermatophores (see Gwynne, 1982), G. firmus males provide little more sperm to females. Despite the 19% difference in gonad weight between wing morphs, it is difficult to say if differences exist in the amount of sperm transferred to females during any one given mating bout. Males can mate quite frequently in a given night, as frequently as every 25 min (Roff, pers obs.). Size differences in gonads may affect the latent duration between mating bouts, but such conjecture is beyond the scope of the present analysis. The other interesting finding pertaining to males is that, unlike the relationship in females, we found no negative correlation between gonad size and whole-organism respirometry rate. This was true for both males with and without functional flight muscles. Although it is not surprising that there is a lack of a trade-off between respirometry rate and gonad size in males considering, relative to females, male gonads are very small, we are left having to explain the size differences in males.
Gryllus firmus has now been very intensely studied with respect to trade-offs between wing morphology and life history traits. We now know that there are significant costs to maintaining a functional flight apparatus in both females and males that involve important reproductive traits. We also know that the traits involved in the trade-off (wing morph, fecundity, call behaviour, etc.) have a genetic basis and are also genetically correlated. Correlated responses to selection based on these genetic correlation patterns have successfully predicted trait variation in natural populations (Roff & Fairbairn, 1999; Mostowy, 2001). Environmental factors such as temperature, photoperiod and available resources are also known to affect phenotypic trait expression under laboratory conditions (for a discussion see Crnokrak & Roff, 1998b). The present study establishes a link between the life history and physiological components of the trade-offs. An important element still missing from a more complete understanding of the suite of trade-offs, is the knowledge of how variable environments affect the associations of traits on a physiological level. Our study demonstrates that there are significant and predictable correlations between respirometry rates and reproductive potential, but does not shed light on how and if such relationships change under different environments. As the heritabilities and genetic correlations for the traits of importance in the trade-off are not 1 (Roff, 1986b, 1997; Crnokrak & Roff, 1998a), there is room for environmental `adjustment' of phenotypic trait expression. This adjustment will certainly involve physiological changes (for an example in G. assimilis, see Zera et al., 1998), and hence further study of the relationship between the trade-off as expressed in terms of reproductive potential and the underlying physiological changes will be necessary.
We thank Graham Roff for having measured the whole-organism respirometry rates in the crickets. This study was supported by a grant to D.A.R. from NSERC.