Behav Ecol Sociobiol(1991) 29:121-126

Behavioral Ecology and Sociobiology © Springer-Verlag1991

Components of variance in male lifetime copulatory and reproductive success in a seed bug Denson Kelly McLain Biology Department, Landrum Box 8042, Georgia Southern College, Statesboro, GA 30460-8042, USA ReceivedJuly 27, 1990 / Accepted March 10, 1991

Summary. Males and females of the seed bug, Neacoryphus bicrucis Say, were individually numbered in the field in southeastern Georgia (USA) and census taken daily for 6 weeks. Variation in male mating efficiency (ME = no. copulations/no, sightings) exceeded that in females and was significantly greater than that generated by a null model. Lifetime copulatory success, estimated as the product of ME and longevity, ranged from 0 to 41, with ME accounting for over 88% of the variation. Lifetime reproductive success (LRS), estimated as the product of ME, longevity, and clutch size, ranged from 0 to 898. Among males copulating at least once, ME accounted for over 40% of the variation in LRS, while longevity alone or in combination with clutch size accounted for 21% and 46%, respectively, of the variation in LRS. Failure of some males (26%) to copulate contributed 38% of the total variation in LRS. Thus, among males surviving to adulthood, sexual selection pressure arising from variation in ME is approximately as strong a "force" for phenotypic change as is natural selection pressure arising from variation in longevity.

Introduction Sexual selection arises from variation in the number of mates and promotes the evolution of traits that (1) enhance attractiveness to members of the opposite sex or (2) enhance ability to win contests with members of the same sex for access to members of the opposite sex (Darwin 1871). Typically, the opportunity for sexual selection is greater in males than females since the relative inexpense of male gametes does not limit the number of copulations that can be achieved (Trivers 1972; Borgia 1979). In insects, males frequently possess behavioral and morphological traits that appear to have evolved in the context of sexual selection (Thornhill and Alcock 1983). The opportunity for sexual selection to effect further phenotypic evolution is proportional to the variance in

mating success (Wade and Arnold 1980), while the intensity of sexual selection will depend on the relative contribution of variation in mating success to the total variation in lifetime reproductive success (LRS) (Arnold and Wade 1984; Wade 1987). Thus, genetic correlations between traits experiencing natural and sexual selection within the same and different life history episodes will constrain the evolution of phenotype (Lande 1980, 1982; Lande and Arnold 1983; Arnold and Wade 1984). An estimate of the relative opportunities for sexual and natural selection requires that variance in LRS [V(LRS)] be partitioned among those components that give rise to sexual versus natural selection (McCauley 1983; Arnold and Wade 1984; Brown 1988). Few studies have assessed the relative contributions of fitness components to total variation in either lifetime mating success (LMS) or LRS (Clutton-Brock 1988). In some insect species, V(LRS) is higher in males than females and may be a consequence of greater variation among males in copulatory success (e.g., Fincke 1988; McVey 1988). However, in other species, variation in mating efficiency (ME; no. matings/unit time) does not differ between the sexes (e.g., McCauley 1983; McLain and Boromisa 1987). In the species examined herein, individual variation in longevity was at least twice as important as variation in ME in determining relative LMS or LRS. Thus, natural and not sexual selection provided the greater opportunity for selection. Greater variation in LMS or LRS among males than among females may not reflect the disparity in gamete size and the consequent potential for one or a few successful males to monopolize matings. If females are slower than males to return to receptiveness after copulation, then variance in LMS [V(LMS)] will necessarily be greater in males (Sutherland 1985). Chance variation in male longevity and rate of encounter of receptive females will introduce further variation in male LMS (Hubbell and Johnson 1987). Consequently, V(LMS) may not reflect competition among males as much as chance environmental variation coupled with betweensex differences in physiology (tied to differential invest-

122 m e n t in offspring). Thus, the o p p o r t u n i t y for sexual selection to effect p h e n o t y p i c c h a n g e m a y be o v e r e s t i m a t ed in field studies ( S u t h e r l a n d 1987). The present study estimates LMS and LRS and partitions the v a r i a n c e in each p r o d u c t for males o f the seed bug, Neacoryphus bicrucis (Say) ( H e m i p t e r a : L y g a e i d a e ) . T h e s t u d y was c o n d u c t e d o n a small, d i s j u n c t p o p u l a t i o n o c c u p y i n g a single p a t c h o f h o s t p l a n t , Senecio anonymus, m e a s u r i n g o n l y 2 × 12 m. T h e small size o f the h a b i tat p a t c h a n d the h o m o g e n e o u s d i s t r i b u t i o n o f p l a n t s within it s h o u l d p r e c l u d e the i m p o r t a n c e o f s o m e p o t e n tial e n v i r o n m e n t a l c o m p o n e n t s o f V ( L M S ) such as o p e r a t i o n a l sex ratio, e x p o s u r e to m o r t a l i t y agents, a n d density o f c o m p e t i t o r s a n d p o t e n t i a l mates. T h e seed bug, N. bicrucis, is a h o s t p l a n t specialist on r a g w o r t , Senecio spp., in G e o r g i a ( s o u t h e a s t e r n U S A ) ( M c L a i n 1984). O v e r w i n t e r i n g a d u l t s b r e a k diapause in early to m i d - s p r i n g , s y n c h r o n i z i n g their activity with the initial flowering o f r a g w o r t . A d u l t s feed a n d m a t e exclusively o n r a g w o r t for 4~6 weeks, after which few survive; n y m p h s b e c o m e a b u n d a n t as h o s t p l a n t s d r y a n d release seeds, 3 5 weeks after the initial a p p e a r ance o f a d u l t s ( M c L a i n a n d Shure 1990). T h e seed b u g is a c o n v e n i e n t a n i m a l for the s t u d y o f L M S since m a t i n g activity p e a k s in early a f t e r n o o n a n d the d u r a t i o n o f c o p u l a t i o n in the field is l o n g (generally > 10 h; M c L a i n 1989). A l s o , the bugs feed a n d m a t e on o r n e a r flower h e a d s (i.e., high u p on plants) w h e r e they are easily sighted a n d o b s e r v e d .

Methods Field site and marking. The field site was a pasture located 11 km west of Statesboro, Georgia (USA), on SR 25, that contained a single 2 × 12 m patch of ragwort, Senecio anonymus. At this site, seed bugs, Neacoryphus bicrucis, fed only upon ragwort. There were no other patches of ragwort within 2 km of the field site. Ragwort began flowering April 5 (1990). The density of flower heads within the patch ranged from 250-650 per 0.25 m 2 (mean 385.00, SD 145.40, n=20). Seed bugs colonized the patch beginning with initial ragwort flowering on April 5 but were relatively rare until 1 week later. Seed bugs were marked April 12-April 20. In all, 206 males and 159 females were marked. This represented approximately 1/3 of all seed bugs present, based on estimates of population size. Population size on d a y / w a s estimated as Ni= (ai-lniri.i z)/(ri-1.i-2r~,~ 1) where al i is the number marked on day i - 1 while r~ x,i y is the number marked on day i - y that were recaptured on day i - x (Southwood 1966). The following population sizes were estimated: 350 (April 17), 300 (April 18), 1000 (April 19), 1000 (April 20), 750 (April 21). Population size appeared steady between 750 and 1000 until May 1, at which point it began to decline. Seed bugs were numbered on the hemilytra with black Pigma 01 pens (Sakura Color Products, Japan). The ink is waterproof and did not fade or chip during a l-month preliminary laboratory study. Marked and unmarked bugs showed identical survivorship and mating efficiency. Insects were marked in the field and released at the base of the plant from which they were collected. Typically, marked insects immediately climbed back up the plant and either began to feed or flew approximately 1 m to another cluster of flower heads. Insects were sampled daily, at mid-afternoon (1500-1800 h), when mating activity peaked. Thus, each marked individual could be sighted only once per day.

Statistics. Data analysis is restricted to those individuals sited at least twice. For all such individuals the ME is calculated as the no. matings/no, sitings (McCauley 1983). LMS is estimated as longevity × ME, where longevity is the number of days an individual survived after the initiation of mating activity on April 12. The date of last recapture is assumed to be the last day of life. LRS for males is the product : ME × longevity × average clutch size. Because females oviposit and remate almost daily (Solbreck /978; McLain 1989) and because sperm precedence strongly favors and last male to copulate (McLain 1989), it is assumed that males fertilize a single clutch per copulation. Also, the long duration of copulation, typically 8-24 h, generally limits the number of copulations to only one per day. Average clutch size is a function of male longevity because clutch size increases if females survive to relatively old age (> 28 days). By week, clutch size averaged: (1) 4.12, (2) 19.80, (3) 12.68, (4) 15.17, (5) 35.27, and (6) 56.49 (n=58 field-collected females, unpublished data). The average number of eggs fertilized per copulation for a male surviving to age x is [~ kCj]/x where Cj is the average clutch size in week j, and k is the number of days in weekj through which the male survived. The opportunity for sexual selection is I~E=V(ME)/ME 2, where V(ME) is the variance in ME and ME 2 is the square of mean ME (Wade and Arnold 1980; McCauley 1983). The opportunity for selection on fitness component j, lj, is proportional to the rate at which average fitness at component j can evolve; i.e., Wi+l =IjWi+W~, if fitness is completely heritable, where Wi and Wi+ 1 are absolute average fitness of component j in generations i and i + 1 (Arnold and Wade 1984). Ij represents the maximum opportunity for selection to effect adaptive change since fitness at component j will not be completely heritable and may negatively covary with other components of fitness (Arnold and Wade 1984). To determine whether observed values of IME are significantly larger than expected on the basis of random mating, IME is also calculated for a null model in which matings are assigned at random with the restrictions that (1) distribution of the number of sightings in the null data set is the same as in the real data set, and (2) the number of matings assigned to the null data set equals the number actually observed (McLain 1986). The mean ME will not vary between real and null data sets, but the variance about the respective means may differ. Significance of this difference is tested with Levene's test (McLain 1986). I used the method of Brown (1988) to partition the variance in mating success into 2 components, longevity and ME, and to partition the variance in LRS into 3 components, longevity, ME, and clutch size. The partitioning of variance in mating success gives rise to a complex covariance term which is then partitioned into components representing simultaneous independent variation and covariation. The partitioning of LRS entails the assessment of contributions from joint variables which are themselves further partitioned into components representing, again, simultaneous independent variation and covariation. To partition variance in LRS, a separate calculation is made to assess the contribution to variation from failure to mate, the upward partition into components being restricted to males having mated at least once (Brown 1988). The percentage of variation, C~, attributable to each component, i, of the upward partition has the following interpretation: C~ is the percent of total variation observed that would still be present if all other components were held constant at their means. Negative C~ (observed with some joint variables or joint covariation terms) indicates that the observed variation would have been even larger if not for negative covariation (Brown 1988). Statistical tests were conducted on a microcomputer using the Systat system for statistics (L. Wilkinson 1987).

Results Demographics In all, 46 m a r k e d males a n d 28 m a r k e d females were n o t resighted. F o r t h o s e resighted, the n u m b e r o f sight-

123

o>, 70

• Males I [] Females

0.8 ~ 0.6

0.4 '~

0.2 0.0

1

2

3

5

6

Week

Fig. 1. Average population mating efficiency for males or females as a function of the number of weeks since the initiation of mating activity

u)

60 50

[m Males ] [] Females

40

Lifetime mating success

> 55 .E 6 Z

The opportunity for sexual selection, IME, was 2.71 times greater in males (IME=0.60, mean ME 0.44, SD 0.34) than females (IME=0.22, mean ME 0.66, SD 0.31). The difference between sexes in mean ME was significant (F=20.84, df=1,289, P<0.001), as was the difference between sexes in variance in ME (Levene's test t = l . 9 1 , P<0.05). Observed variance in ME was significantly greater than that generated by a null model for males (Levene's test t=3.57, P<0.001) but not for females (Levene's test t = 1.04, P>0.05). There was a positive correlation between the ME of individuals early and late in their lives. For ME in weeks 1 3 versus weeks 4-6, Pearson's linear correlation coefficient r = 0 . 5 5 (t=4.38, P<0.001, n = 4 7 ) for males, while r = 0 . 6 8 (t=5.01, P<0.001, n = 3 2 ) for females. Similarly, for ME in weeks 1-2 versus weeks 5-6, r = 0 . 8 9 ( t = 6.93, P<0.001, n = 1 5 ) for males, while r = 0 . 6 8 ( t = 3.98, P<0.001, n = 2 0 ) for females.

30 20

1: 0.125

0.375

0.625

0.875

Mating efficiency Fig. 2. Distribution of lifetime mating efficiency in males and females

ings ranged from 2 to 11 for males (mean 3.56) and from 2 to 9 for females (mean 3.40). Longevity of males (mean 25.74, SD 7.68, n = 160) and females (mean 26.52, SD 9.05, n = 1 3 1 ) was similar (t=0.78, P > 0 . 0 5 for difference in mean; Wilcoxon test, Z = 1 . 0 2 , P > 0 . 0 5 for difference in distribution). Mean population ME varied over time for both males and females, rising in weeks 1-2, peaking in week 3, then declining in weeks 4-6 (Fig. 1). Significant weekly variation in mean population ME is explained by models of the form M E = c o n s t a n t + m l w e e k i - m z w e e k i 2 (males: F = 1 0 . 4 1 , df=2,405, P < 0 . 0 0 1 ; females: F = 65.87, df=2,310, P < 0 . 0 0 1 ) but not of the form M E = constant+mlweek~ (males: F = 0 . 0 6 , df=1,406, P > 0.05; females: F = 2.02, d f = 1,311, P > 0.05). ME did not vary as a function of the number of sightings for males sighted 2, 3, 4-5, or > 5 times (F = 1.31, df=3,156, P>0.05). There was no significant difference in the variance in mating efficiency between groups of males sighted 2, 3, 4-5, or > 5 times ( F = 1.58, d f = 3,156, P>0.05).

The estimated mean number of matings was 11.41 (SD 9.66, range 0~41) for males and 16.32 (SD 9.64, range 0 35) for females. The opportunity for selection on number of matings was twice as great in males (ILMs= 0.72) as in females (ILMs= 0.35). Most of the variation in LMS, in both males (88.3%) and females (76.7%), was attributable to variation in ME not longevity (Table 1). V(LMS) would have been greater in both sexes if not for the negative covariation between longevity and ME (Table 1).

Lifetime reproductive success Estimated LRS among males ranged from 0 to 898 (mean 162.46, SD 154.71 for all males; mean 218.44, SD 141.13 for males mating at least once). Failure to mate accounted for 38.0% of V(LRS) among adult males. Among males mating at least once, ME alone accounted for 43.1% of V(LRS), while ME alone and in combination with other components of LRS accounted for 50.1% of V(LRS) (Table 2). Furthermore, longevity alone accounted for 20.5% of V(LRS), while longevity alone and in combination with other variables

Table 1. Partitioning of the variance in lifetime mating success among males or females Percentage of variance due to Mating efficiency

Longevity

88.3 76.7

14.8 34.6

Q - 3.1 -- 11.3

I

J

8.9 9.0

- 12.0 - 20.3

Mating efficiency

Males Females

ME ranged from 0 to 1.0 in both males and females, although the proportion of individuals with ME = 0 was higher for males (0.26) than females (0.11) (Fig. 2).

Q = covariance term for longevity and mating efficiencyand is composed of I, representing simultaneous independent variation (between longevity and mating efficiency),and J, representing covariation. For further details, see Methods, Brown (1988)

124 Table 2. Partitioningof the variancein lifetimereproductivesuccess

(LRS) among males having mated at least once [failure to mate accounted for 38.0% of V(LRS)] LT LT 20.5 ME 3.8 CS 0.9 LT*ME*CS: 0.2\8.5

ME

CS

-8.3 43.1 1.8

24.5 -4.6 9.6

Components of LRS: longevity (LT), mating efficiency (ME), clutch size (CS), and joint components(LT*ME, LT*CS, ME*CS, LT*ME*CS). Joint variables are themselves decomposed into a component representing simultaneous independent variation (below diagonal) and a component representing covariation (above diagonal). Numbers represent percentage of observed variance remaining if all other variables are held constant at their means (see Brown 1988)

accounted for 50.1% of V(LRS) (Table 2). Mating efficiency covaried negatively with clutch size and longevity while clutch size and longevity covaried positively (Table 2). The total opportunity for selection on LRS of adult males, ILRS, is 0.91 [=V(LRS)/meanLRs2]. This can be partitioned into opportunities arising from variation in female fecundity (clutch size, here) and V(LMS) according to ILRs=RIclutch+ILMs, where R is the number of female mates per male (= mean ME, here) (Wade 1987). This gives 0.89=(0.44[=R])(0.38)+0.72 which agrees well with ILRS(= 0.91) calculated directly from estimates of LRS. This partitioning reveals that V(LMS) contributes about 4 times as much as variance in female fecundity to the total opportunity for selection on males.

Discussion

Variation in lifetime mating success Among male N. bicrucis, variation in ME contributes more to variation in LMS than does longevity. This result contrasts with other studies of insects (e.g., McCauleT 1983; McVey 1988). The relative unimportance of longevity apparently stems primarily from the reduction in mating activity in the population as a whole in the latter half of the breeding season. This results in part because as females age, they spend more of their time on the ground, away from males, where they oviposit in seed mats and other litter. In addition, ME and longevity covary negatively. Thus, mating may cause males to lose vigor, perhaps as a consequence of nutritional contributions to females (see McLain et al. 1990) or as a cost of territorial defence (McLain 1984; McLain and Shure 1987). Also, courtship and mating may expose males to increased risk of predation or parasitization. Among females, longevity and ME were even more strongly negatively correlated. Assassin bugs and spiders, both of which were seen to prey upon seed bugs, were abundant in the ragwort patch (density=l-2/ 0.25 m2).

While male mating success was negatively correlated with longevity, there was a strong correlation between ME early and late in the season. This suggests that loss of vigor does not account for the negative covariation between longevity and ME. Frequently, females exercise mate choice by resisting courtship, mating only males capable of subduing them (McLin and Shure 1987; McLain et al. 1990). Thus, vigorous males have a mating advantage. Variation among males in ME gives rise to sexual selection (McCauley 1983). In the N. bicrucis population, the variance in ME was significantly larger than expected on the basis of random mating. Thus, in the absence of countervailing natural selection (e.g., GS Wilkinson 1987), sexual selection may effect phenotypic change if there is additive genetic variation for sexually selected traits. Variance in ME was greater in males than in females. If females are clumped and males acquire groups of females at random, variation in male ME may be much greater than that in females (Sutherland 1987). If the postmating latency is higher in females than males, variance in ME will necessarily be greater in males (Sutherland 1985; Hubbell and Johnson 1987). Furthermore, chance variation in longevity and mate encounter rate can increase variation in ME or LMS (Hubbell and Johnson 1987). In the present study, some sources of environmental variation are minimized due to the small size of the habitat patch and its isolation from other patches of host plant. Thus, variation in mate encounter rate, risk of predation, and operational sex ratio varied across a spatial scale that could easily be traversed, with repeated sampling (e.g., searching for mates), by a single individual within a day. For this reason, variation among males in ME should correlate with relevant phenotypic traits. Still, environmental variation in mating success provides an opportunity for sexual selection. For instance, if clumping of females causes high variance in LMS among males, sexual selection may favor traits that permit males to locate groups of females better. Similarly, if variation in longevity contributes much to variation in LMS, then natural selection may favor traits that remove chance from mortality, such as the ability to detect predators and parasites, flee from dangerous conditions, detect ample food and water supplies, etc. Sex difference in postmating latency is probably a consequence of differential parental investment (PI) (see Trivers 1972). The difference in PI may reflect the joint effects of sexual and natural selection in promoting and maintaining disruptive selection on gamete size (Parker 1978). Even if differential PI necessarily causes greater variance in male ME or LMS, the low investment by males still affords the opportunity for some males to enjoy great reproductive success at the expense of other males. Sex difference in postmating latency does not reduce the impact of sexual selection, rather it provides the opportunity for greater sexual selection in males. Cheap male gametes and a male bias among individuals receptive to mating enhance the opportunity for sexual selection to act through the monopolization of matings by a small number of males.

125

Variance in ME among females is not greater than expected from random assignment of matings. Thus, there is little opportunity for sexual selection on females. Unless variation in clutch size contributes much to variation in LRS of males, sexual selection may be a weak force on females. Males will have little to gain from choosing a more fecund female over a less fecund one if the probability of finding and competing for the more fecund female is less than 1 (see Hammerstein and Parker 1987). The male-biased operational sex ratio (56.4% males) also contributes to low variance in female ME (Sutherland 1987; Hubbell and Johnson 1987).

Lifetime reproductive success Variation in longevity contributes more to variation in LRS than to variation in LMS. This results in large part from the greater clutch size of females mated later in the season. Only one-third of the males survived to encounter females of high fecundity. Even if all of the variation in LRS resulting from failure to mate (38.0%) is attributed to low ME (= 0), longevity would still account for 30.9% of V(LRS), or twice its contribution to V(LMS). If we assume that failure to mate is a result of ME = 0, then ME accounts for 69.1% of the variation in LRS among adults. This suggests that sexual selection could be a strong force for phenotypic change if, for example, environmental change relaxed countervailing natural selection. Since the rate of mortality on immature stages is high, countervailing natural selection may overwhelm sexual selection (McLain 1991). If population size is constant, males must average two progeny surviving to adulthood. The estimated average number of progeny is 162, giving a survival rate of 1.2%. Variance in LRS arising from failure to breed, assuming the same population size and same V(LRS) among breeders in parental and progeny generations, is now 70.3%, much of which is attributable to variation in juvenile survivorship. Thus, opportunity for sexual selection as a component of the total opportunity for selection on males may be overestimated when juvenile survivorship and genetic correlations between juvenile and adult traits are ignored. Genetic correlation between the sexes may further reduce the effectiveness of sexual selection if males and females have different phenotypic optima (Lande 1980). Variation in longevity, not ME, probably accounts for most of the variation in LRS among females (e.g., Fincke 1988; McVey 1988). Thus, natural selection favoring female longevity may be an important source of countervailing natural selection on male sexually selected traits.

Conclusions There is a significant opportunity for sexual selection on males of the seed bug. V(ME), which gives rise to sexual selection, accounts for most of V(LMS) and at least half of V(LRS). Since the opportunity for sexual selection is a large fraction of the total opportunity of

selection on adult males, there is a potential for the evolution of sexually selected traits, given genetic variation. However, this potential is constrained to the extent that sexually selected traits are negatively genetically correlated with traits experiencing natural selection in juveniles or females. A phenotypic response to sexual selection may occur if the environment changes and countervailing natural selection is relaxed.

References Arnold S J, Wade MJ (1984) On the measurement of natural and sexual selection : theory. Evolution 38 : 709-719 Borgia G (1979) Sexual selection and the evolution of mating systems. In: Blum MS, Blum NA (eds) Sexual selection and reproductive competition in insects. Academic, New York, pp 19-80 Brown D (1988) Components of lifetime reproductive success. In: Clutton-Brock TH (ed) Reproductive success. University of Chicago Press, Chicago, pp 439-453 Clutton-Brock TH (1988) Introduction. In: Clutton-Brock TH (ed) Reproductive success. University of Chicago Press, Chicago, p p l 10 Darwin C (1871) The descent of man and selection in relation to sex. John Murray, London Fincke O (1988) Sources of variation in lifetime reproductive success in a nonterritorial damselfly (Odonata: Coenagrionidae). In: Clutton-Brock TH (ed) Reproductive success. University of Chicago Press, Chicago, pp 24~43 Hammerstein P, Parker GA (1987) Sexual selection: games between the sexes. In: Bradbury JW, Andersson MB (eds) Sexual selection: testing the alternatives. John Wiley, New York, pp 119142 Hubbell SP, Johnson LK (1987) Environmental variance in lifetime mating success, mate choice, and sexual selection. Am Nat 130:91 112 Lande R (1980) Sexual dimorphism, sexual selection, and adaptation in polygenic characters. Evolution 34:292-307 Lande R (1982) Rapid origin of sexual isolation and character divergence in a cline. Evolution 36:213-223 Lande R, Arnold SJ (1983) The measurement of selection on correlated characters. Evolution 37:1210-1226 McCauley DE (1983) An estimate of the relative opportunities for natural and sexual selection in a population of milkweed beetles. Evolution 37:701-707 McLain DK (1984) Host plant density and territorial behavior of the seed bug, Neacoryphus bicrucis (Hemiptera: Lygaeidae). Behav Ecol Sociobiol 14 : 181-I 87 McLain DK (1986) Null models and the intensity of sexual selection. Evol Theory 8:49 52 McLain DK (1989) Prolonged copulation as a post-insemination guarding tactic in a natural population of the ragwort seed bug. Anita Behav 38:65%664 McLain DK (1991) The r--k continuum and the relative effectiveness of sexual selection. Oikos 59:263-265 McLain DK, Boromisa RD (1987) Male choice, fighting ability, assortative mating and the intensity of sexual selection in the milkweed longhorn beetle, Tetraopes tetraophthalmus (Coleoptera: Cerambycidae). Behav Ecol Sociobiol 20:239 246 McLain DK, Shure DJ (1987) Pseudocompetition: interspecific displacement of insect species through misdirected courtship. Oikos 49:291-298 McLain DK, Shure DJ (1990) Spatial and temporal density dependence of host plant patch use by the ragwort seed bug, Neacoryphus bicrucis (Hemiptera: Lygaeidae). Oikos 58:306-312 McLain DK, Hancock MB, Hope MA (1990) Fitness effects of nonrandom mating in the ragwort seed bug, Neacoryphus bicrucis (Hemiptera: Lygaeidae). Ethol Ecol Evol 2 : 253-262 McVey ME (1988) The opportunity for sexual selection in a territorial dragonfly, Erythemis simplicicolis. In: Clutton-Brock TH

126 (ed) Reproductive success. University of Chicago Press, Chicago, pp 59-75 Parker GA (1978) Selection on non-random fusion of gametes during the evolution of anisogamy. J Theor Biol 73 : 1-28 Solbreck C (1978) Migration, diapause, and direct development as alternative life histories in the seed bug, Neacoryphus bicrucis. In: Dingle H (ed) Evolution of insect migration and diapause. Springer, Berlin Heidelberg New York, pp 196-217 Sutherland WJ (1985) Can chance produce a sex difference in variance in mating success and account for Bateman's data? Anim Behav 33:1349-1352 Sutherland WJ (1987) Random and deterministic components of variance in mating success. In: Bradbury JW, Andersson MB (eds) Sexual selection: testing the alternatives. John Wiley, New York, pp 209-220

Thornhill R, Alcock J (1983) Evolution of insect mating systems. Harvard University Press, Cambridge Trivers RL (1972) Parental investment and sexual selection. In: Campbell B (ed) Sexual selection and the descent of man. AIdine, Chicago, pp 136-179 Wade MJ (1987) Measuring sexual selection. In: Bradbury JW, Andersson MB (eds) Sexual selection: testing the alternatives. John Wiley, New York, pp 197-208 Wade Mj, Arnold SJ (1980) The intensity of sexual selection in relation to male sexual behaviour, female choice, and sperm precedence. Anim Behav 28:446-461 Wilkinson GS (1987) Equilibrium analysis of sexual selection in Drosophila melanogaster. Evolution 41 : 11 21 Wilkinson L (1987) Systat: the system for statistics. Systat Inc, Evanston