SEARCH

SEARCH BY CITATION

Keywords:

  • dimorphism;
  • geographic variation;
  • life history;
  • response to selection;
  • trade-offs

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Populations, species description and methods of Rearing
  6. Experimental design
  7. General setup
  8. Dorso-longitudinal muscle dissection
  9. Results
  10. Do populations differ in the proportion of macropterous males?
  11. Do populations differ in the call duration?
  12. Do populations differ in the rate of histolysis of dorso-longitudinal muscle in macropterous males?
  13. Is female choice dependent on both relative call duration and population?
  14. Does the trade-off between call duration and dorso-longitudinal muscle mass vary among populations?
  15. Discussion
  16. Acknowledgment
  17. References

Quantitative genetic theory assumes that trade-offs are best represented by bivariate normal distributions. This theory predicts that selection will shift the trade-off function itself and not just move the mean trait values along a fixed trade-off line, as is generally assumed in optimality models. As a consequence, quantitative genetic theory predicts that the trade-off function will vary among populations in which at least one of the component traits itself varies. This prediction is tested using the trade-off between call duration and flight capability, as indexed by the mass of the dorsolateral flight muscles, in the macropterous morph of the sand cricket. We use four different populations of crickets that vary in the proportion of macropterous males (Lab = 33%, Florida = 29%, Bermuda = 72%, South Carolina = 80%). We find, as predicted, that there is significant variation in the intercept of the trade-off function but not the slope, supporting the hypothesis that trade-off functions are better represented as bivariate normal distributions rather than single lines. We also test the prediction from a quantitative genetical model of the evolution of wing dimorphism that the mean call duration of macropterous males will increase with the percentage of macropterous males in the population. This prediction is also supported. Finally, we estimate the probability of a macropterous male attracting a female, P, as a function of the relative time spent calling (P = time spent calling by macropterous male/(total time spent calling by both micropterous and macropterous male). We find that in the Lab and Florida populations the probability of a female selecting the macropterous male is equal to P, indicating that preference is due simply to relative call duration. But in the Bermuda and South Carolina populations the probability of a female selecting a macropterous male is less than P, indicating a preference for the micropterous male even after differences in call duration are accounted for.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Populations, species description and methods of Rearing
  6. Experimental design
  7. General setup
  8. Dorso-longitudinal muscle dissection
  9. Results
  10. Do populations differ in the proportion of macropterous males?
  11. Do populations differ in the call duration?
  12. Do populations differ in the rate of histolysis of dorso-longitudinal muscle in macropterous males?
  13. Is female choice dependent on both relative call duration and population?
  14. Does the trade-off between call duration and dorso-longitudinal muscle mass vary among populations?
  15. Discussion
  16. Acknowledgment
  17. References

Trade-offs are a core concept of evolutionary biology (Futuyma, 1998; Roff, 2002). Two views of trade-offs have developed, one in response to the implementation of an optimality approach to evolutionary adaptation and the other as a result of considerations of the genetic basis of trade-offs. Typically the first viewpoint sees a trade-off as a fixed function along which trait means of a population ‘slide’ (e.g. Mountford, 1968; Schaffer, 1974; Smith & Fretwell, 1974; Roff, 1984), whereas the second sees a bivariate normal distribution, resulting in a correlation between traits (Young & Weiler, 1960). The second viewpoint predicts that the trade-off line, which is the major axis of the bivariate relationship, will itself shift under selection. Thus, the two approaches, in their simplest manifestation, make different predictions about how trade-offs will vary among populations subject to selection. These predictions were recently tested using the trade-off between fecundity and flight muscle weight in the sand cricket, Gryllus firmus. The genetic perspective was supported, the trade-off line showing geographic variation (Roff et al., 2002). In the present paper we extend this analysis to geographic variation in the trade-off between call duration and flight muscle maintenance in male G. firmus. Gryllus firmus males, like the majority of Orthoptera, attract females by ‘holding station’ and calling. For such species a non-calling male can obtain copulations by wandering and coming into contact with females at random (Hissmann, 1990) or by becoming a satellite to a calling male (Rowell & Cade, 1993). The latter behaviour is common among taxa in which calling is a principal means of attraction (Roff, 1996). Wandering and satellite behaviours are alternate modes of encountering females but the primary mode in most crickets (and taxa such as the anura) is calling (Zuk & Simmons, 1997). Experiments with G. firmus have shown that the probability of attracting a female is proportional to the time spent calling (Crnokrak & Roff, 1995, 1998a, b), demonstrating the importance of call duration in this species.

Calling is an energetically expensive activity, substantially increasing metabolic rates (Roff, 1992; Bailey et al., 1993). As a result calling could ‘compete’ for resources, thereby generating a trade-off. We have extensively examined one component of this trade-off, namely that between call duration and the size of the principal flight muscles, the dorso-longitudinal muscles. Flight is a metabolically expensive activity (Hocking, 1953; Dudley, 2000) but even in the absence of flight per se there are substantial ‘overhead’ costs in terms of maintaining the flight muscles, which are metabolically very active (Zera & Denno, 1997; Crnokrak & Roff, 2000, 2002). If flight is not required the organism could either not develop the flight apparatus or histolyse the flight muscles if they have already been developed. Both processes are found in G. firmus.

Like a large number of insects, G. firmus is wing dimorphic, a condition in which one morph is macropterous (long-winged and typically capable of flight at some stage) while the other morph is micropterous (short-winged and flightless: Harrison, 1980; Roff, 1986, 1990, 1994). The existence of wing dimorphism itself suggests that there are significant fitness costs to developing and maintaining the flight apparatus. One obvious competing trait is reproductive activity. In females the overwhelming majority of experiments have shown that the micropterous morph has an earlier age of first reproduction and larger overall egg production than the macropterous morph (reviewed in Roff, 1996). Fewer experiments have been conducted with males but these have also shown a reproductive cost, with macropterous males obtaining fewer copulations than micropterous males (reviewed in Roff, 1996; Langellotto et al., 2000). In G. firmus, micropterous males call more and attract proportionally more females than macropterous males (Crnokrak & Roff, 1995, 1998a, b).

Macropterous males can ultimately reduce their costs of flight muscle maintenance by histolysing the flight muscles. Histolysis of flight muscles has been observed in both monomorphically winged species and in the macropterous morph of dimorphic species (Johnson, 1969, 1976; Dingle, 1996). In particular, it occurs frequently in both female and male G. firmus and is associated with an increased egg production (Roff, 1989; Stirling et al., 2001) and increased call duration (Crnokrak & Roff, 1998a, b).

Two trade-offs between flight ability and call duration in G. firmus can be identified; first, there is an extreme trade-off associated with the dimorphism and secondly there is the trade-off within the macropterous morph between call duration and flight ability as measured by the size of the flight muscles. We shall focus on the latter, continuously distributed trade-off. First, we present an overview of selection acting on the trade-off function under a quantitative genetic framework, a fuller mathematical discussion of which is given in Roff et al. (2002). As noted above, the quantitative genetic viewpoint sees a bivariate normal relationship: thus the phenotypic relationship between two traits, Y and X can be written as

  • image(1)

where μI, σI are the phenotypic mean and SD of trait I (X,Y), and rP is the phenotypic correlation between the two traits. The first set of terms is the intercept of the regression line between the traits and the second term in parentheses is the slope. In the present paper, we shall arbitrarily designate call duration as the dependent variable and flight muscle condition (status or mass) as the independent variable. Assuming that the traits are heritable, directional selection acting on either one or both traits will alter the trait means (μX,μY) because of both direct selection and indirect selection acting through the correlation between the traits. In principle, selection will also alter the SD and the genetic correlation, and hence the phenotypic correlation. Theory and artificial selection experiments indicate that fairly strong selection is required to make significant changes to these parameters (Roff, 1997) and so it is likely that in natural populations these parameters will stay approximately constant (Roff et al., 2002). Therefore, the general prediction is that among different geographic populations there will be variation in the intercept of the trade-off function (Fig. 1), such variation being the result of weak selection and/or drift. Variation in the slope of the function may or may not be found, depending on the strength of selection or effects of genetic drift.

image

Figure 1. Schematic illustration of the predicted change in the trade-off function under weak selection. Shown are two hypothetical bivariate normal distributions which characterize the trade-off function under the quantitative genetic perspective. Under selection the mean trait values change (•) leading to a change in the intercept of the trade-off function but no change in its slope.

Download figure to PowerPoint

In addition to predicting variation in the trade-off function we can also predict how the two traits will vary in relation to the proportion of macropterous individuals in the population. The direct and indirect effects of selection on wing dimorphism and associated traits can be understood using the threshold model of quantitative genetics (Roff, 1986). According to this model, the dimorphism is a phenotypic manifestation of the combined effects of a normally distributed underlying trait, called the liability, and a threshold of sensitivity. At some stage in development, the organism's developmental trajectory is shunted into one of two pathways; individuals lying above the threshold develop into one morph while individuals below the threshold develop into the alternate. Under this model, the heritability of wing morph is actually the heritability of the liability and the observed correlated responses to selection on proportion macropterous are consequences of genetic correlations between the liability, and the associated traits (fecundity, wing muscle mass, call duration etc.). Thus, the mean values of the traits associated with macroptery or microptery in a population of primarily macropterous individuals will not be the same as the mean values of those traits in macropters and micropters from a primarily micropterous population (Roff, 1990, 1997; Roff & Fairbairn, 2001; Fig. 2).

image

Figure 2. Schematic illustration of the prediction that the mean values of the traits associated with macroptery or microptery in a population of primarily macropterous individuals will not be the same as the mean values of those traits in macropters and micropters from a primarily micropterous population. The top curve illustrates the distribution of the liability in a population that is 50% macropterous. Some trait such as call duration is negatively correlated with the liability (i.e. trade-off between morph status and the trait), as shown in the bottom graph. The two middle graphs show the distribution of liability for a population that is primarily micropterous (second from top) and primarily macropterous (third from top). The circles show the mean liabilities for the two morphs under the different distributions. Changing the proportion macropterous is a consequence of changes in the distribution of the liability and hence, as illustrated, the mean liability within each morph changes. A result of this is that the mean call duration also changes. Note that it is possible for the mean call duration of the micropterous morph from a population primarily macropterous (third from top) to be very similar to the mean call duration of the macropterous morph from a population that is primarily micropterous (second from top).

Download figure to PowerPoint

The trade-off between wing morphology and reproductive traits is hypothesized to be primarily a consequence of the allocation of resources between the flight muscles and egg production in females and between the flight muscles and call duration in males (see above). A simple model for the trade-off is Y = X − A, where Y is the reproductive trait, X is the value of the trait in the absence of any allocation to flight capability (i.e. the value for the micropterous morph) and A is the allocation to flight capability (e.g. flight muscle weight at some age). Thus, the trade-off will vary according to the phenotypic and genetic correlations between the liability and the two components X and A. Without loss of generality we shall assume that individuals with liabilities above the threshold develop into macropterous individuals. Dorso-longitudinal muscle mass is positively correlated with the liability, which is negatively correlated with fecundity (for a review see Roff & Fairbairn, 2001). What changes in the trade-off function between fecundity and dorso-longitudinal muscle mass are expected following selection that changes proportion macroptery? From the general derivation given above, we can write the trade-off function as C = a − bXM = (μC + bμM) − bXM, where C is call duration, XM is dorso-longitudinal muscle mass, μC, μM are the respective mean values and b is the absolute value of the slope of the trade-off. Selection for increased proportion macropterous is equivalent to selection for increased liability and hence such selection will lead to a correlated increase in μM and a decrease in μC. Because of the opposing changes in the two mean values it is not possible, without knowing the actual values, to predict the direction of change in the intercept of the trade-off function. Nevertheless, the foregoing does predict that the intercept value of the trade-off function will be a function of the proportion macropterous in the population. There are insufficient populations studied in the present paper to test this prediction, but it has been validated with respect to fecundity (Roff et al., 2002). Most importantly, for the present analysis, call duration of macropterous males is predicted to decline as the frequency of macroptery in the population increases (for numerical analyses see Roff & Fairbairn, 1999, 2001). Because they have little or no commitment to flight capability (A∼0: because of genetic correlations micropterous individuals can possess some flight muscle), micropterous individuals should show relatively little change with mean liability and thus proportion macroptery.

The latter two predictions have been tested in the laboratory by selecting either on proportion macroptery or on fecundity. In the former experiments the predicted correlated change in fecundity in the macropterous females was observed (Roff, 1990). There was a statistically non-significant change in fecundity of micropterous females, which is consistent with the hypothesis that the trade-off derives primarily from the commitment to flight muscles, which micropterous females all but lack (Roff, 1990; Zera et al., 1997). In addition, macropterous males and females showed correlated changes in flight muscle histolysis and flight propensity: selection for increased proportion macroptery decreased the rate of flight muscle histolysis and increased flight propensity, while selection for decreased proportion macroptery produced the opposite correlated responses (Fairbairn & Roff, 1990). In the second series of experiments selection for increased or decreased fecundity produced the predicted correlated changes in proportion macroptery (Roff et al., 1999) and flight muscle histolysis (Stirling et al., 2001). In the present paper, we extend these analyses by examining the relationship between proportion macroptery and the call durations of the two wing morphs. Specifically, we predict that if proportion macroptery varies among populations then so also will the within-morph call duration. Further, mean call duration within morphs should decrease with proportion macroptery.

Populations, species description and methods of Rearing

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Populations, species description and methods of Rearing
  6. Experimental design
  7. General setup
  8. Dorso-longitudinal muscle dissection
  9. Results
  10. Do populations differ in the proportion of macropterous males?
  11. Do populations differ in the call duration?
  12. Do populations differ in the rate of histolysis of dorso-longitudinal muscle in macropterous males?
  13. Is female choice dependent on both relative call duration and population?
  14. Does the trade-off between call duration and dorso-longitudinal muscle mass vary among populations?
  15. Discussion
  16. Acknowledgment
  17. References

Gryllus firmus is a large, (live weight, approximately 0.75 g) ground dwelling cricket found along the coastal region of the American southeast from Florida to Connecticut (Alexander, 1968; Harrison, 1985). It is usually found on dunes and sandy areas, which being impermanent, favour the evolution of wing dimorphism (Roff, 1990, 1994). To test the predictions outlined above we compared four populations of G. firmus: the laboratory stock on which all former experiments were conducted and three widely separated populations recently brought into the laboratory. The laboratory population is descended from approximately 40 adults (equal sex ratio) collected in northern Florida in 1981 and had passed through approximately 40–50 generations in the laboratory prior to the experiment. This stock has been maintained with a standing adult population of approximately 100–500 adults under diapause averting temperatures (>25 °C). The three recent populations came from Florida (Gainsville, designated F), Bermuda (Annies Bay, designated B) and South Carolina (Charleston, designated SC). From each location we collected approximately 100 adults and nymphs. Crickets from each population were maintained in the laboratory at approximately 100–300 breeding adults for five generations at a temperature of 25–30 °C before being used for the present experiments. The selection regime within the laboratory undoubtedly differs from that experienced by the wild populations. In particular, there can be expected to be no selection favouring dispersal traits and selection favouring early histolysis of flight muscles. Thus, the most likely population to show changes in both intercept and slope is the laboratory population.

Crickets used for the experiments were reared in four-litre buckets (diameter, 21 cm; height, 15 cm) at a density of 30 nymphs per bucket under a photoperiod of 15 h light, 9 h dark and at a temperature of 28 °C with ad libitum food and water. For the Lab, F and SC populations 10 buckets per population were set up and for the B population seven buckets (by a technical accident) were used. Food consisted of crushed Purina rabbit chow provided ad libitum. Water was provided by water soaked cotton placed in test tubes. On the day of their final ecdysis (i.e. the day nymphs became adults), the newly-eclosed adults were placed in individual containers with food and water until they were 6 days old at which time they were used in the experiment.

General setup

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Populations, species description and methods of Rearing
  6. Experimental design
  7. General setup
  8. Dorso-longitudinal muscle dissection
  9. Results
  10. Do populations differ in the proportion of macropterous males?
  11. Do populations differ in the call duration?
  12. Do populations differ in the rate of histolysis of dorso-longitudinal muscle in macropterous males?
  13. Is female choice dependent on both relative call duration and population?
  14. Does the trade-off between call duration and dorso-longitudinal muscle mass vary among populations?
  15. Discussion
  16. Acknowledgment
  17. References

Each male was housed in an individual glass jar and monitored for call duration and whether or not he attracted a female. Each glass jar was placed in a bucket which was connected to two other buckets by plastic 2.5 cm diameter tubing, thereby forming a T-maze (stem of ‘T’ = 10 cm and head of ‘T’ = 27 cm). Eight T-mazes were used, all of which were placed in incubators set at a 15L/9D photoperiod and 28 °C. For each T-maze, two males (one micropterous, one macropterous) and a single female (either macropterous or micropterous, chosen at random from the same population as the males) were placed in separate buckets. Males were placed in the ends of the head of the T and the female at the base of the stem of the T. The 2.5 cm tubing that interconnected all three buckets allowed the females free access to the males’ quarters. Cones placed on the ends of the tubes leading to the buckets containing males prevented the female from exiting these buckets. There is no data indicating visual communication between crickets but for ensure no such effect, even if it exists, the jars housing the males were painted black to prevent the male and female from seeing each other. Wire mesh covered the top of each jar allowing females to hear the calling males. A continuous playback of cricket calling was used to provide a constant background (Cade & Wyatt 1984; Crnokrak & Roff, 1995).

Male calling was monitored by Realistic tie-clip 33–105 microphone (frequency response, 50–15 000 Hz) placed in each male's jar. Microphones were connected to an analog-to-digital converter relay system monitored by a computer that recorded the time of each incoming signal. The computer scanned the relay system once every second. Chirp length averages 154 ms (Webb & Roff, 1992) and chirps are repeated in bout lengths generally exceeding one second (average bout length = 1.47 s; P. Crnokrak, personal observation). Each microphone gain was set at a level that would trigger the relay system only when the occupant of the jar called and would not be triggered by the background call or the call of neighbouring crickets. A simple binary code was used by the computer to record when a male called (1-calling, 0-not calling). Every male was monitored for 23.5 h, on day 6 as adults. Day 6 was picked since calling increases up to day 6 and then changes little afterwards (Crnokrak & Roff, 1998b). Recording began at 13.00 hours each day and ended at 12.30 hours the next day (30 min was spent on maintaining the food and water levels and cleaning the T-maze). After daily maintenance was completed, the computer was reset and monitoring recommenced. We paired males that were from the same population and switched the position of macropterous and micropterous males each day. Fifty males (25 micropterous and 25 macropterous) from each population were monitored for call duration and female attraction.

Dorso-longitudinal muscle dissection

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Populations, species description and methods of Rearing
  6. Experimental design
  7. General setup
  8. Dorso-longitudinal muscle dissection
  9. Results
  10. Do populations differ in the proportion of macropterous males?
  11. Do populations differ in the call duration?
  12. Do populations differ in the rate of histolysis of dorso-longitudinal muscle in macropterous males?
  13. Is female choice dependent on both relative call duration and population?
  14. Does the trade-off between call duration and dorso-longitudinal muscle mass vary among populations?
  15. Discussion
  16. Acknowledgment
  17. References

All crickets were preserved by freezing immediately at the end of the calling measurement period. Once thawed, crickets were dissected to remove the dorso-longitudinal flight muscles that are immediately below the dorsal mesothoracic covering and are the major muscles for flight (Du Porte, 1920; Srihari et al., 1975; Pfau & Koch, 1994). We assessed the amount of histolysis of the dorso-longitudinal flight muscle by its colour and dry weight. Dorso-longitudinal muscle colour was coded initially on a three point scale: 0-white, 1-pink and 2-brick red. A red colour is associated with the presence of mitochondria, indicating fully functional muscles (Ready & Josephson, 1982; Mole & Zera, 1993; Zera et al., 1997). Pink coloured muscles are in the process of being histolysed and white are histolysed. For the statistical analysis we combined the white and pink categories into a single category indicative of histolysis. All colour tests were carried out by one person (P.C.) to avoid any discrepancies between experimenters. Once the dorso-longitudinal muscles were dissected, they were placed on a numbered, pre-weighed, microscope coverslip placed in an oven set at 60 °C for at least 3 days. Once dried, the dorso-longitudinal muscles were measured on a Mettler digital scale to 0.0001 g. Although all crickets were dissected, only macropterous crickets had measurable dorso-longitudinal muscle; all micropterous individuals were given dorso-longitudinal muscle weights and condition scores of 0.

Do populations differ in the proportion of macropterous males?

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Populations, species description and methods of Rearing
  6. Experimental design
  7. General setup
  8. Dorso-longitudinal muscle dissection
  9. Results
  10. Do populations differ in the proportion of macropterous males?
  11. Do populations differ in the call duration?
  12. Do populations differ in the rate of histolysis of dorso-longitudinal muscle in macropterous males?
  13. Is female choice dependent on both relative call duration and population?
  14. Does the trade-off between call duration and dorso-longitudinal muscle mass vary among populations?
  15. Discussion
  16. Acknowledgment
  17. References

We calculated mean proportion macropterous for each population based on proportion macropterous per cage. The two populations collected from Florida had very similar percentages of macropterous males as did the populations from Bermuda and South Carolina (Table 1). A one-way anova using the arcsine-squareroot transformed values indicates a highly significant variation among the four populations (F3,33 = 12.65, P < 0.0005). A post hoc Tukey test shows that the two Florida populations do not differ from each other (P = 0.95), nor do the South Carolina and Bermuda populations (P = 1.00), but both Florida populations differ significantly from the Bermuda (Lab vs. B, P = 0.001: F vs. B, P = 0.003) and South Carolina (Lab vs. SC, P < 0.005: F vs. SC, P = 0.001) populations.

Table 1.  Mean values and standard errors (in parentheses) of characteristics of the four populations of Gryllus firmus.
 FloridaLabBermudaSouth Carolina
  1. *Mean of per cent macropterous per cage (10 cages for Florida, Lab, South Carolina and seven cages for Bermuda).

  2. †LW = macropterous, SW = micropterous. Mean values based on 25 macropterous males per population.

  3. ‡Percentage based on total sample (25) per wing morph per population.

  4. §Percentage of females attracted to macropterous males in 25 male pairs (one LW, one SW) per population.

  5. ¶Data from Roff et al. (2002).

Percentage macropterous*28.8 (4.5)33.3 (3.4)72.0 (13.5)79.8 (5.5)
LW† call duration (h)0.68 (0.06)0.55 (0.04)0.31 (0.02)0.32 (0.04)
SW† call duration (h)0.97 (0.07)0.84 (0.06)0.79 (0.04)0.78 (0.05)
LW dorso-longitudinal muscle (g × 103)3.0 (0.6)4.9 (0.6)7.0 (0.5)6.1 (0.7)
Percentage red dorso-longitudinal muscle‡4.032.052.056.0
Percentage attracted to LW§41.739.116.712.0
LW§ head width (mm)¶6.75 (0.04)6.65 (0.06)6.58 (0.03)6.68 (0.05)

Do populations differ in the call duration?

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Populations, species description and methods of Rearing
  6. Experimental design
  7. General setup
  8. Dorso-longitudinal muscle dissection
  9. Results
  10. Do populations differ in the proportion of macropterous males?
  11. Do populations differ in the call duration?
  12. Do populations differ in the rate of histolysis of dorso-longitudinal muscle in macropterous males?
  13. Is female choice dependent on both relative call duration and population?
  14. Does the trade-off between call duration and dorso-longitudinal muscle mass vary among populations?
  15. Discussion
  16. Acknowledgment
  17. References

Because it has been established from previous experiments that macropterous and micropterous males differ in their call duration (Crnokrak & Roff, 1995, 1998a, b), we analysed the morphs separately. As expected from previous work, micropterous males called longer than macropterous males (Table 1). There is a non-significant difference among populations in the call duration of micropterous males (F3,96 = 2.38, P = 0.0744), although this difference is significant by the Kruskal–Wallis test (χ2 = 12.4, d.f. = 3, P = 0.006). By both tests there is a highly significant difference in call duration the macropterous males (F3,96 = 19.67, P < 0.00005). A Tukey test indicates that the two Florida populations do not differ (P = 0.15), nor do Bermuda and South Carolina (P = 1.00), but the Florida populations differ from both the Bermuda and South Carolina populations (P < 0.005 in all four pairwise comparisons). The ranking of call duration in the micropterous males is the same as in the macropterous males (Table 1). The differences in call duration of the macropterous males are consistent with the prediction that call duration should decline with proportion macropterous. The lesser differences in the micropterous males is also consistent with the prediction that this morph will show relatively little response (see introduction).

Do populations differ in the rate of histolysis of dorso-longitudinal muscle in macropterous males?

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Populations, species description and methods of Rearing
  6. Experimental design
  7. General setup
  8. Dorso-longitudinal muscle dissection
  9. Results
  10. Do populations differ in the proportion of macropterous males?
  11. Do populations differ in the call duration?
  12. Do populations differ in the rate of histolysis of dorso-longitudinal muscle in macropterous males?
  13. Is female choice dependent on both relative call duration and population?
  14. Does the trade-off between call duration and dorso-longitudinal muscle mass vary among populations?
  15. Discussion
  16. Acknowledgment
  17. References

We have two measures of flight muscle histolysis; dorso-longitudinal muscle mass and histolysis category (red dorso-longitudinal muscle = not histolysed, pink–white = partial–full histolysis). For both measures, one-way anova shows a significant difference among populations (dorso-longitudinal muscle mass, F3,96 = 8.24, P < 0.0001; histolysis category, F3,96 = 7.22, P < 0.0001) A post hoc Tukey test showed that for both dorso-longitudinal muscle mass and histolysis category the Florida population differed from the Bermuda population and the South Carolina population (P < 0.002 in all comparisons), but no other pairwise comparison was significant.

One possible source of variation in dorso-longitudinal muscle mass is overall body size, larger males possibly having larger muscles. Unfortunately, body size of individuals in the present experiment was not recorded. Head widths of females grown under similar conditions (28 °C, 12L : 12D) were recorded in a previous experiment (Roff et al., 2002). Overall, there is a significant difference in head width (F3,183 = 3.89, P = 0.010), but a Tukey test shows only the pairwise comparison between Florida and Bermuda to be significant (P = 0.004). Bermuda females had the smallest head width and Florida the largest, with the Lab and South Carolina females falling between, with very similar mean head widths (Table 1). Thus, the differences among populations in dorso-longitudinal muscle mass are unlikely to be a consequence of differences in body size.

Is female choice dependent on both relative call duration and population?

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Populations, species description and methods of Rearing
  6. Experimental design
  7. General setup
  8. Dorso-longitudinal muscle dissection
  9. Results
  10. Do populations differ in the proportion of macropterous males?
  11. Do populations differ in the call duration?
  12. Do populations differ in the rate of histolysis of dorso-longitudinal muscle in macropterous males?
  13. Is female choice dependent on both relative call duration and population?
  14. Does the trade-off between call duration and dorso-longitudinal muscle mass vary among populations?
  15. Discussion
  16. Acknowledgment
  17. References

Previous experiments have shown that the probability of a female choosing a male in a pair of calling males is dependent on the relative call duration (=time spent calling by focal male divided by total time spent calling by the pair). We examined whether this probability was influenced by the population of origin by use of the proportionality model of Crnokrak & Roff (1998a). If call duration is the only variable affecting the probability of attraction, the probability of a female selecting the macropterous male is equal to the relative call duration of the macropterous male (i.e. dLW/[dLW + dSW], where dLW, dSW are the call durations of the macropterous and micropterous males, respectively). Crnokrak & Roff (1998a) tested this hypothesis by fitting the model PLW♂ = aX, where PLW♂ is the probability of a female selecting a macropterous male, a is a constant, and X is the relative call duration of the macropterous male. Under the null hypothesis of direct proportionality the coefficient of proportionality, a, is equal to 1. To include the possible effect of population we fitted the model PLW♂ = (a + b * p)X, where a, b are fitted constants and p is the proportion of macropterous males in a population. This model is predicated on the hypothesis that any differences in probability of attraction will be related to the proportion of macropterous males in the population. Model testing was performed using the maximum likelihood method with a binomial loss function, beginning by testing the constants-only model (PLW♂ = constant) against the single-parameter model (PLW♂ = aX). The single-parameter model gives a significantly better fit than the constants-only model (χ2 = 29.48, df = 1, P < 0.0001). The two-parameter model (PLW♂ = [a + b * p]X) gives a significantly better fit than the one-parameter model (χ2 = 5.09, d.f. = 1, P = 0.024). The two-parameter model predicts a probability of selecting a macropterous male that is proportional to (a + b*p), which we shall refer to as C. For the four populations the values of C are: Florida = 1.21, Lab = 1.14, Bermuda = 0.56 and South Carolina = 0.45. Values of C estimated separately for each population are: Florida = 1.20 (SE = 0.29), Lab = 1.19 (0.29), Bermuda = 0.65 (0.25), South Carolina = 0.45 (0.20). In the Florida and Lab populations the probability of being selected by a female is not significantly different from the null hypothesis of direct proportionality, whereas in the Bermuda and South Carolina populations females show a preference for micropterous males. This is also evident from the overall mean values: the Lab and Florida macropterous males call 40% and 41%, respectively, of the total call duration and attract 39% and 42% of the females, respectively (Table 1). In contrast, the Bermuda and South Carolina macropterous males call 28% and 29%, respectively, of the total call duration but attract only 17% and 12% of the females (Table 1).

Does the trade-off between call duration and dorso-longitudinal muscle mass vary among populations?

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Populations, species description and methods of Rearing
  6. Experimental design
  7. General setup
  8. Dorso-longitudinal muscle dissection
  9. Results
  10. Do populations differ in the proportion of macropterous males?
  11. Do populations differ in the call duration?
  12. Do populations differ in the rate of histolysis of dorso-longitudinal muscle in macropterous males?
  13. Is female choice dependent on both relative call duration and population?
  14. Does the trade-off between call duration and dorso-longitudinal muscle mass vary among populations?
  15. Discussion
  16. Acknowledgment
  17. References

Analysis of covariance shows that there is significant variation in call duration because of dorso-longitudinal muscle mass (F1,92 = 30.41, P < 0.0001) and population (F3,92 = 8.82, P < 0.0001) but not the interaction (F3,92 = 2.07, P = 0.109). This indicates that only the intercept of the trade-off function between call duration and dorso-longitudinal muscle mass differs statistically among the four populations. The intercept values are: Florida = 0.78, Lab = 0.72, Bermuda = 0.55 and South Carolina = 0.64. To analyse the difference among the intercepts we created a new variable using the common slope (−33.62), y = call duration −33.62 (dorso-longitudinal muscle mass), and then tested for variation with one-way anova followed by a pairwise Tukey test. There is no difference between the intercepts of the Lab and Florida populations (P = 0.073), nor between the Bermuda and South Carolina populations (P = 0.960), but all other comparisons are highly significant (P < 0.0001). Thus, the four regression lines separate into two groups; the Lab and Florida populations, and the Bermuda and South Carolina populations. The significant variation among populations is in line with the quantitative perspective of the evolution of trade-off functions. There is no indication that the slope of the Lab population differs from the recently captured populations.

Dorso-longitudinal muscle mass is a measure of how much histolysis has taken place and hence the extent to which call duration and muscle maintenance are in ‘competition’. Another measure is the histolysis (=colour) category of the muscle. To examine the effect of both measures simultaneously on call duration we used two approaches: multiple regression and principal components analysis. For the first approach we used stepwise multiple regression with population and histolysis category coded as dummy variables. Both forward and backward stepwise give the same result, namely that the best fitting model comprises dorso-longitudinal muscle mass, population, the interaction dorso-longitudinal muscle mass × population, and the interaction histolysis category × population (in all cases P < 0.006). The additive model including dorso-longitudinal muscle mass and population accounts for 54% of the variance, whereas the final multiple regression model explains 63% of the variance.

In the second approach we first reduced the two variables to a single variable using the first principal component from a principal components analysis of dorso-longitudinal muscle mass and histolysis category. Covariance analysis shows highly significant effects because of factor score (F1,92 = 61.15, P < 0.0001) and population (F3,92 = 6.22, P = 0.0007) and a marginally non-significant interaction term (F3,92 = 2.45, P = 0.0682). This model accounts for 63% of the variance, the same as the multiple regression model (see above). The analysis using the first principal component supports the conclusion reached using only dorso-longitudinal muscle mass that there is significant variation in the trade-off function among populations. It hints at differences in both intercepts and slopes: however, the most deviant slope is not that from the Lab population but the recently collected Florida population (slopes from independent regressions: Florida = −0.25, Lab = −0.11, B = −0.10, SC = −0.15).

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Populations, species description and methods of Rearing
  6. Experimental design
  7. General setup
  8. Dorso-longitudinal muscle dissection
  9. Results
  10. Do populations differ in the proportion of macropterous males?
  11. Do populations differ in the call duration?
  12. Do populations differ in the rate of histolysis of dorso-longitudinal muscle in macropterous males?
  13. Is female choice dependent on both relative call duration and population?
  14. Does the trade-off between call duration and dorso-longitudinal muscle mass vary among populations?
  15. Discussion
  16. Acknowledgment
  17. References

Proportion macropterous males, mean call duration, mean dorso-longitudinal muscle mass and histolysis category on the sixth day of adult life all vary geographically in G. firmus. The common garden design of the present experiment demonstrates that this variation is genetically based. As predicted, the Lab population shows the longest call duration and highest rate of muscle histolysis (as assessed through both dorso-longitudinal muscle mass and dorso-longitudinal muscle colour). The Lab population also has the lowest proportion macropterous, but is not significantly different from the more recently collected Florida population (Table 1). From theory and previous selection experiments we predicted that call duration of the macropterous males should decline with proportion macropterous in the population. Because they do not invest in flight muscles, the call duration of micropterous males should show much less variation. Both predictions are consistent with the present data (Table 1). In the macropterous males the time spent calling drops from 0.7 h day−1 in the Florida population (28.8% macropterous), to 0.5 h day−1 in the Lab population (33.3% macropterous), and to 0.3 h in the Bermuda and South Carolina populations (72% and 80% macropterous, respectively). There is also a decline in the call duration of the micropterous males but it is only from 1.0 h (Florida) to 0.8 h (Lab, B and SC).

Because their longer call duration attracts more females, micropterous males have an advantage over macropterous males. The success of a micropterous male appears to depend upon time spent calling relative to that of a macropterous male and the proportion of macropterous males in the population. In the low proportion populations (Lab and Florida) the success rate of macropterous males is equal to the relative time spent calling, whereas in the two high proportion populations the success of macropterous males is less than the relative time spent calling (i.e. females show a preference for micropterous males). Both of the low proportion populations originated in Florida and hence more populations varying in proportion macroptery are required to better test the relationship between success and the proportion macroptery in the population. In each trial females from the same population as the calling males were used: this protocol allows us to draw conclusions concerning the relative success of males within their own population but confounds success rates of males with differences because of both sexes. The present results do demonstrate that there is significant geographic variation in the relationship between relative call duration and male success: further research is required to disentangle effects due to differences among males vs. differences among females. The present experimental design did not permit the analysis of possible differences in call behaviour as might be exhibited by satellite males. Such behaviour has not been recorded for G. firmus but is certainly worthy of investigation. In particular, we might expect that if such behaviour existed in the population that macropterous males would be most likely to adopt it, thereby alleviating to a degree the cost of calling while maintaining flight capability.

As found in previous studies (Crnokrak & Roff, 1998a, 2000), call duration of macropterous males declines as dorso-longitudinal muscle mass increases. The size of the dorso-longitudinal muscle is an index of the degree to which it has been histolysed and hence the negative relationship between call duration and dorso-longitudinal muscle mass supports the hypothesis that calling and flight muscle maintenance ‘compete’ for resources, thereby inducing a trade-off between flight and calling capabilities. Colour of the dorso-longitudinal muscle is also an index of muscle functionality and a multiple regression analysis indicates that both indexes separately account for explained variance. We combined these two indexes using principal components analysis and obtained results qualitatively the same as the regression analysis. Because the principal components analysis produces an index that is a linear combination of the two variables, this consistency is not surprising.

We predicted that the trade-off between call duration and dorso-longitudinal muscle mass (or the pca score) would differ among populations. Because it is a function of the population mean values, the intercept of the trade-off function should show the greatest variation. Whichever measure of flight muscle competence is used, there is a highly significant difference in the intercept, as predicted. Geographic variation in the trade-off function between a reproductive trait (calling duration, fecundity) and flight capability has now been observed in both males and females (Roff et al., 2002, this study), demonstrating that analyses of evolutionary changes in trait mean values need to take into account the bivariate normal character of trade-offs.

Geographic variation in both morphology and life history is a common observation, clinal variation in body size, for example, being so common as to be given a particular name, Bergmann's rule. In sockeye salmon (Oncorhynchus nerka) this variation in body size is accompanied by variation in both the slope and intercept of the allometric relationship of fecundity and body length, whereas in kokanee (a non-anadromous form of sockeye) and chinook salmon (O. tshawytscha) only variation in the intercept was statistically significant (McGurk, 2000; Kinnison et al., 2001). Among the sockeye salmon populations the intercept had a coefficient of variation (CV) of 101%, whereas the CV of the slope was only 26% (data from Table 2 of McGurk, 2000; sample size = 51 populations). These population data are from field-collected individuals and hence it is not clear if the differences are genetic or environmental in origin. Nevertheless these data reinforce the message that trade-offs should not be considered as simple functional relationships but rather as multinomial distributions. While the assumption of fixed relationships may be adequate for some theoretical analyses, a failure to take into account the variability, both genetic and phenotypic, in life history, morphological and behavioural traits could lead to erroneous predictions and the assumption is certainly not a correct reflection of reality. An important avenue of further investigation is an empirical study of the variation in trade-off functions, most particularly with respect to the relative variation in slope and intercept. Cases in which the slope shows significant variation are of particular interest because our quantitative genetic model predicts that such variation is indicative of selection acting on the genetic or environmental variance in addition to the mean phenotype.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Populations, species description and methods of Rearing
  6. Experimental design
  7. General setup
  8. Dorso-longitudinal muscle dissection
  9. Results
  10. Do populations differ in the proportion of macropterous males?
  11. Do populations differ in the call duration?
  12. Do populations differ in the rate of histolysis of dorso-longitudinal muscle in macropterous males?
  13. Is female choice dependent on both relative call duration and population?
  14. Does the trade-off between call duration and dorso-longitudinal muscle mass vary among populations?
  15. Discussion
  16. Acknowledgment
  17. References
  • Alexander, R.D. 1968. Life cycle origins, specialization and related phenomena in crickets. Q. Rev. Biol. 43: 141.
  • Bailey, W.J., Withers, P.C., Endersby, M. & Gaull, K. 1993. The energetic costs of calling in the bushcricket, Requena verticalis (Orthoptera: Tettigoniidae: Listrosceliinae). J. Exp. Biol. 178: 2137.
  • Cade, W.H. & Wyatt, D.R. 1984. Factors affecting calling behaviour in field crickets, Teleogryllus and Gryllus (age, weight and parasites). Behavior 88: 6175.
  • Crnokrak, P. & Roff, D.A. 1995. Fitness differences associated with calling behaviour in the two wing morphs of male sand crickets, Gryllus firmus. Anim. Behav. 50: 14751481.
  • Crnokrak, P. & Roff, D.A. 1998a. The genetic basis of the trade-off between calling and wing morph in males of the cricket, Gryllus firmus. Evolution 52: 11111118.
  • Crnokrak, P. & Roff, D.A. 1998b. The contingency of fitness: an analysis of food restriction on the macroptery-reproduction trade-off in Gryllus firmus. Anim. Behav. 56: 433441.
  • Crnokrak, P. & Roff, D.A. 2000. The trade-off to macroptery in the cricket Gryllus firmus: a path analysis in males. J. Evol. Biol. 13: 396408.
  • Crnokrak, P. & Roff, D.A. 2002. Trade-offs to flight capability in Gryllus firmus: the influence of whole-organism respiration rate on fitness. J. Evol. Biol. 15: 388398.
  • Dingle, H. 1996. Migration. Oxford University Press, New York.
  • Du Porte, E.M. 1920. The muscular system of Gryllus assimilis Fabr. (=pennsylvanicus Burm.). Ann. Entom. Soc. Am. 23: 1652.
  • Dudley, R. 2000. The Biomechanics of Insect Flight. Princeton University Press, Princeton, New Jersey.
  • Fairbairn, D.J. & Roff, D.A. 1990. Genetic correlations among traits determining migratory tendency in the sand cricket, Gryllus firmus. Evolution 44: 17871795.
  • Futuyma, D.J. 1998. Evolutionary Biology. Sinauer Associates, Inc., Sunderland, Massachusetts.
  • Harrison, R.G. 1980. Dispersal polymorphism in insects. Ann. Rev. Ecol. Syst. 11: 95118.
  • Harrison, R.G. 1985. Barriers to gene exchange between closely related cricket species. II. Life cycle variation and temporal isolation. Evolution 39: 244259.
  • Hissmann, K. 1990. Strategies of mate finding in the European field cricket (Gryllus campestris) at different population densities: a field study. Ecol. Ent. 15: 281291.
  • Hocking, B. 1953. The intrinsic range and speed of flight of insects. Trans. R. Ent. Soc. Lond. 104: 223345.
  • Johnson, C.G. 1969. Migration and Dispersal of Insects by Flight. Methuen, London.
  • Johnson, C.G. 1976. Lability of the flight system: a context for functional adaptation. In: Insect Flight (R. C.Rainey, ed.), Blackwell Science Publisher, Oxford.
  • Kinnison, M.T., Unwin, M.J., Hendry, A.P. & Quinn, T.P. 2001. Migratory costs and the evolution of egg size and number in introduced and indigenous salmon populations. Evolution 55: 16561667.
  • Langellotto, G.A., Denno, R.F. & Ott, J.R. 2000. A trade-off between flight capability and reproduction in males of a wing-dimorphic insect. Ecology 81: 865875.
  • McGurk, M.D. 2000. Comparison of fecundity-length-latitude relationships between nonanadromous (kokanee) and anadromous sockeye salmon (Oncorhynchus nerka). Can. J. Zool. 78: 17911805.
  • Mole, S. & Zera, A.J. 1993. Differential allocation of resources underlies the dispersal-reproduction trade-off in the wing-dimorphic cricket, Gryllus rubens. Oecologia 93: 121127.
  • Mountford, M.D. 1968. The significance of litter size. J. Anim. Ecol. 37: 363367.
  • Pfau, H.K. & Koch, U.T. 1994. The functional morphology of singing in the cricket. J. Exp. Biol. 195: 147167.
  • Ready, N.E. & Josephson, R.K. 1982. Flight muscle development in a hemimetabolous insect. J. Expt. Zool. 220: 4956.
  • Roff, D.A. 1984. The evolution of life history parameters in teleosts. Can. J. Fish. Aquat. Sci. 41: 9841000.
  • Roff, D.A. 1986. The evolution of wing dimorphism in insects. Evolution 40: 10091020.
  • Roff, D.A. 1989. Exaptation and the evolution of dealation in insects. J. Evol. Biol. 2: 109123.
  • Roff, D.A. 1990. The evolution of flightlessness in insects. Ecol. Monograph. 60: 389421.
  • Roff, D.A. 1992. The Evolution of Life Histories: Theory and Analysis. Chapman and Hall, New York.
  • Roff, D.A. 1994. Habitat persistence and the evolution of wing dimorphism in insects. Amer. Nat. 144: 772798.
  • Roff, D.A. 1996. The evolution of threshold traits in animals. Quart. Rev. Biol. 71: 335.
  • Roff, D.A. 1997. Evolutionary Quantitative Genetics. Chapman and Hall, New York.
  • Roff, D.A. 2002. Life History Evolution. Sinauer Associates, Sunderland, MA.
  • Roff, D.A. & Fairbairn, D.J. 1999. Predicting correlated responses in natural populations: changes in JHE activity in the Bermuda population of the sand cricket, Gryllus firmus. Heredity 83: 440450.
  • Roff, D.A. & Fairbairn, D.J. 2001. The genetic basis of dispersal and migration and its consequences for the evolution of correlated traits. In: Causes, Consequences and Mechanisms of Dispersal at the Individual, Population and Community Level (C.Clobert, J.Nichols, J. D.Danchin & A.Dhondt, eds), Oxford University Press, Oxford.
  • Roff, D.A., Tucker, J., Stirling, G. & Fairbairn, D.J. 1999. The evolution of threshold traits: effects of selection on fecundity and correlated response in wing dimorphism in the sand cricket. J. Evol. Biol. 12: 535546.
  • Roff, D.A., Mostowy, S. & Fairbairn, D.J. 2002. The evolution of trade-offs: testing predictions on response to selection and environmental variation. Evolution 56: 8495.
  • Rowell, G.A. & Cade, W.H. 1993. Simulation of alternative male reproductive behavior: calling and satellite behavior in field crickets. Ecol. Modell. 65: 265280.
  • Sahai, H. & Ageel, M.I. 2000. The Analysis of Variance. Birkhauser, Boston.
  • Schaffer, W.M. 1974. Selection for optimal life histories: the effects of age structure. Ecology 55: 291303.
  • Smith, C.C. & Fretwell, S.D. 1974. The optimal balance between size and number of offspring. Am. Nat. 108: 499506.
  • Sokal, R.R. & Rohlf, F.J. 1995. Biometry. W. H. Freeman and Co., San Francisco.
  • Srihari, T., Gutmann, E. & Novak, V.J.A. 1975. Effect of ecdysome and juvenoid on the development involution of flight muscles in Acheta domestica. J. Insect Physiol. 21: 18.
  • Stirling, G., Fairbairn, D.J.F., Jensen, S. & Roff, D.A. 2001. Does a negative genetic correlation between wing morph and early fecundity imply a functional constraint in Gryllus firmus. J. Evol. Biol. 3: 157177.
  • Webb, K.L. & Roff, D.A. 1992. The quantitative genetics of sound production in Gryllus firmus. Anim. Behav. 44: 823832.
  • Young, S.S.Y. & Weiler, H. 1960. Selection for two correlated traits by independent culling levels. J. Genet. 57: 329338.
  • Zera, A.J. & Denno, R.F. 1997. Physiology and ecology of dispersal polymorphism in insects. Ann. Rev. Ecol. 42: 117.
  • Zera, A.J., Sall, J. & Grudzinski, K. 1997. Flight-muscle polymorphism in the cricket Gryllus firmus: muscle characteristics and their influence on the evolution of flightlessness. Physiol. Zool. 70: 519529.
  • Zuk, M. & Simmons, L.W. 1997. Reproductive Strategies of the Crickets (Orthoptera: Gryllidae). In: Mating Systems in Insects and Arachnids (J. C.Choe & B. J.Crespi, eds), pp. 89109. Cambridge University Press, Cambridge.