Theories of life-history evolution and parental investment,

Behavioral Ecology Vol. 13 No. 6: 742–749 Phenotypic selection and function of reproductive behavior in the subsocial bug Elasmucha putoni (Heteropte...
Author: Alexis Sherman
1 downloads 2 Views 112KB Size
Behavioral Ecology Vol. 13 No. 6: 742–749

Phenotypic selection and function of reproductive behavior in the subsocial bug Elasmucha putoni (Heteroptera: Acanthosomatidae)

To investigate the function of maternal care and determinants of reproductive success in the subsocial bug Elasmucha putoni (Heteroptera: Acanthosomatidae), I used two different approaches, the measurement of phenotypic selection and femaleremoval experiments under conditions differing in biotic-environmental pressure. For two field populations, unattended eggs and younger nymphs consistently suffered severe predation pressure and attendance by parent females greatly enhanced their survival. In contrast, under enemy-excluded conditions, offspring performance was not reduced in broods without parent females, indicating that maternal care functions as a physical defense against predators. However, the determinant of female reproductive success in E. putoni in the field was not the care behavior alone. Selection gradient analysis showed that early season oviposition and larger clutch size, as well as a longer duration of care by a female, was favored during the breeding episode. This study is the first to evaluate phenotypic selection on parental care and other reproductive traits in arthropods. Key words: Elasmucha, Heteroptera, parental care, phenotypic selection, reproductive success, subsociality. [Behav Ecol 13:742– 749 (2002)]

T

heories of life-history evolution and parental investment, including post-ovipositional parental care, generally assume that increased investment in current offspring provides fitness benefits to the offspring but also imposes costs in future survival and/or reproduction of the parent (e.g., CluttonBrock, 1991; Roff, 1992; Stearns, 1992). Optimal parental care strategies have been considered to be designed to maximize fitness gain of the parent under such reproductive trade-offs (Grafen and Sibly, 1978; Lazarus, 1990; Maynard Smith, 1977; Winkler, 1987; Yamamura and Tsuji, 1993). Parental care of eggs or young is not common among arthropods but has evolved repeatedly in different lineages and resulted in current diverse forms (Choe and Crespi, 1997; Zeh and Smith, 1985). For a variety of arthropod taxa, there has been an increasing number of empirical studies that evaluate costs and benefits of parental care by different research methods (reviewed in Tallamy and Wood, 1986; Trumbo, 1996). To quantify costs and benefits of parental behavior in arthropods, experiments in which treatment groups with different investment in care behavior are established by removing the parent from the offspring have usually been employed (for benefits, see, e.g., Diesel, 1989; Kudo et al., 1995; FilippiTsukamoto et al., 1995; Tallamy and Denno, 1981a; for costs, see, e.g., Fink, 1986; Nalepa, 1988; Tallamy and Denno, 1982). Such an experimental approach provides quantitative evidence for the contribution of care behavior to reproductive success of the parent, but it is not enough to understand the determinants of reproductive success because reproductive

Address correspondence to S. Kudo, Naruto University of Education, Naruto, Tokushima, 772-8502 Japan. E-mail: skudo@naruto-u. ac.jp. Received 25 July 2001; revised 25 January 2002; accepted 12 February 2002.  2002 International Society for Behavioral Ecology

traits other than care behavior, such as clutch size, oviposition site selection, and oviposition schedule, can also affect reproductive success. Moreover, reproductive traits, including care behavior, are often correlated with each other, and natural selection acts on individual phenotype through the net effects of these traits. There have been only a few studies that clarify how parental care behavior varies with other traits and affects reproductive success of the parent in the field (e.g., Scott and Traniello, 1990). Selection gradient analysis is a powerful tool to quantify phenotypic selection on such correlated traits (Arnold and Wade, 1984; Lande and Arnold, 1983; reviewed in Brodie et al., 1995). However, to the best of my knowledge, there has been no attempt to apply this method to understanding natural selection on reproductive traits of subsocial arthropods. Such a regression analysis approach and an experimental approach complement each other in the study of phenotypic evolution (Mitchell-Olds and Shaw, 1987; Wade and Kalisz, 1990): The former will identify target traits and modes of selection in the study population, and the latter will clarify the agent and the mechanism of the detected selection on a particular trait. Elasmucha bugs are well-known subsocial insects, and their reproductive history and maternal care vary considerably among species (e.g., Kaitala and Mappes, 1997). Elasmucha putoni Scott has two generations per year, each of which depends on different host plants, and overwintered females breed on fruit-bearing wild mulberry, Morus bombycis Koizumi (Kudo, 1994). E. putoni is semelparous in the field. Females lay single egg masses on the underside of host leaves usually in early June. First-instar nymphs remain in tight aggregations on natal leaves, and second- or later instar nymphs move among the shoots to feed on fruit, maintaining their aggregations. The nymphal feeding aggregations become smaller as the instars progress, and final (fifth) instar nymphs often feed solitarily. Females attend their eggs and nymphs until as late as the third instar on wild mulberry in the field (Honbo

Downloaded from http://beheco.oxfordjournals.org/ at Pennsylvania State University on March 6, 2014

Shin-ichi Kudo Laboratory of Applied Zoology, Faculty of Agriculture, Hokkaido University, Sapporo, 060-8589 and Department of Biology, Naruto University of Education, Naruto, Tokushima, 772-8502 Japan

Kudo • Phenotypic selection in subsocial bugs

and Nakamura, 1985; Kudo, 1996). Honbo and Nakamura (1985) have shown that in an E. putoni population of central Japan, eggs suffer high predation when the parent female is removed. However, the function of maternal behavior has not been well studied. Moreover, potential environmental pressure against offspring can vary temporally and spatially (e.g., Jeanne, 1979). Costs and benefits of maternal care may also vary according to variation in selection pressure among different populations. In this study, I sought to clarify ecological and behavioral factors determining reproductive success of females on wild mulberry in E. putoni populations of northern Japan, using both experimental and regression analysis approaches.

Measurement of phenotypic selection Data collection I carried out field investigations from mid-May to July in 1991 using an E. putoni population of Misumai located in Sapporo, Hokkaido, northern Japan (42⬚57⬘ N, 141⬚15⬘ E). I censused branches of seven wild mulberry trees, which had been marked with colored tapes, and when first found, I marked individual females guarding eggs with quick-drying paint for later discrimination. At the first census on 21 May, no E. putoni adults were found on the wild mulberry. Copulating pairs were found first on 25 May, and the first oviposition occurred on 30 May. In total, 57 females with offspring were found and monitored every day or two until the collection of their families (see below). I measured and recorded the following traits for each of the 57 females: clutch size, duration of care (see below), the date when oviposition occurred (30 May ⫽ 1), length of the oviposition leaf, relative position of the oviposition leaf on a shoot (order of the leaf from the apex/no. leaves per shoot; 1 ⫽ most basal position), number of leaves on the shoot with the clutch, number of fruits on the shoot with the clutch, and the minimum distance from the ground to the oviposition leaf along the tree surface. The latter five traits are characters of oviposition-site selection by females. The last trait, distance from the ground to the oviposition leaf, might correlate significantly with ant predation because ants have often been reported to be important predators of the offspring of Elasmucha bugs (Kudo et al., 1989; Mappes and Kaitala, 1995; Melber et al., 1980). I examined clutch size by removing the female temporarily from the clutch. This female removal did not affect subsequent female behavior. Two days after the time when broods molted into the second instar, I collected all of the broods and parent females, which were usually found on or near the natal shoot, and counted the number of surviving nymphs. It was difficult or impossible to identify the broods after that time: The later instar broods move away from natal sites and often split into different groups (or join again in other groups) through development. Thus, data on care duration did not cover its full range in the field, which sometimes extends to the third instar. The number of surviving nymphs in a brood was regarded as reproductive success for each parent female and was used in the selection analyses as a measure of fitness (see below). I measured the prothorax width of the collected females under a stereomicroscope using an ocular micrometer. Some females had left their broods before the collection and were subsequently caught when found solitary on wild mulberry trees. However, most of them had disappeared from the trees and were not discovered later. Their body size data are missing. Therefore, body size was not included in the following selection analyses.

Estimation of selection coefficients I used multivariate statistical techniques for quantifying natural selection acting on reproductive traits in a current population of E. putoni (Lande and Arnold, 1983; Mitchell-Olds and Shaw, 1987; reviewed in Brodie et al., 1995). Individual measures of fitness (reproductive success) were converted to relative fitness with a mean value of 1 within a population. Some trait variables were log- or square-root–transformed, and then all of the variables were standardized to a mean zero and unit variance. Thus, selection coefficients presented here are standardized selection coefficients. The opportunity for selection, the upper limit of the strength of selection, was calculated as the variance in relative fitness. I calculated linear selection differentials, which describe changes in trait means, as the covariance of trait values with relative fitness (Lande and Arnold, 1983). Significance tests of the differentials were made with Spearman’s rank correlation (Lande and Arnold, 1983). I estimated linear selection gradients, which describe direct effects of a given trait on fitness adjusting for indirect effects of other correlated traits, as the partial regression coefficients in a linear multiple regression model of relative fitness on all traits (Lande and Arnold, 1983). In multivariate regression, multicollinearity can lead to inflation of the standard errors of the regression coefficients, thus reducing the power of the analysis (Mitchell-Olds and Shaw, 1987). To assess the impact of multicollinearity on the results, I examined the variance inflation factors for trait variables in the regression model (Chatterjee and Price, 1977). The variance inflation factors for the linear selection gradients were not high (1.2–1.6), indicating that the multicollinearity was not problematic in the analysis. Residuals of relative fitness in the model were approximately normally distributed (KolmogorovSmirnov one-sample test, z ⫽ 1.33, p ⫽ .06). Microenvironmental heterogeneity among host trees might affect reproductive behavior of E. putoni, as suggested in E. grisea (Mappes and Kaitala, 1995). There was no significant difference in female reproductive success among seven wild mulberry trees on which oviposition occurred (F6,50 ⫽ 0.67, p ⫽ .67). Trait variables, however, differed among the trees (MANOVA: Pillai’s trace ⫽ 1.71, F42,294 ⫽ 2.78, p ⬍ .0001). To examine effects of the trees on female reproductive success, I included those as dummy variables in the regression model (Chatterjee and Price, 1977). The effects of trees were not significant and thus were excluded from the final model. To detect nonlinear and correlational selection acting on trait variances and covariances, large sample size is ordinarily needed. In the preliminary analysis, neither quadratic nor correlational (cross-product) terms were significant in the full quadratic regression model (quadratic, F8,12 ⫽ 2.15, p ⫽ .11; cross-product, F28,12 ⫽ 1.26, p ⫽ .35). Moreover, as mentioned above, the present data do not completely cover the breeding episode of females, and this may cause underestimates of the nonlinear selection (see also the Discussion). Therefore, I omitted the detailed analysis. Female-removal experiments Under field conditions I carried out field experiments at Nopporo (43⬚02⬘ N, 141⬚30⬘ E) in 1987 and at Misumai in 1990 and 1991. In early June, I removed mothers from clutches randomly chosen on wild mulberry trees (the female-removal group). Other females guarding clutches were kept intact as the control group. I examined clutch size by temporarily removing the female from the clutch. There were no significant differences in initial clutch size between the female-removal and control groups in any of the experiments conducted in the Nopporo (female-removal, n ⫽ 15, mean ⫾ SD ⫽ 48.38 ⫾ 4.91, control,

Downloaded from http://beheco.oxfordjournals.org/ at Pennsylvania State University on March 6, 2014

MATERIALS AND METHODS

743

Behavioral Ecology Vol. 13 No. 6

744

Table 1 Summary of reproductive traits of E. putoni females on wild mulberry in the Misumai population in 1991 Trait

n

Mean ⫾ SD

Body size (mm) Clutch size Care duration (days) Oviposition date (30 May ⫽ 1) Leaf size (mm) Leaf position No. of leaves No. of fruit Distance (cm)

45 57 57 57 57 57 57 57 57

4.54 48.47 19.16 2.32 44.23 0.91 3.14 3.18 306.6

⫾ ⫾ ⫾ ⫾ ⫾ ⫾ ⫾ ⫾ ⫾

0.20 5.47 4.68 2.03 14.49 0.16 0.93 1.12 89.4

n ⫽ 23, 46.33 ⫾ 3.32; t36 ⫽ ⫺0.524, p ⫽ .60) and Misumai populations (in 1990, n ⫽ 16, 48.38 ⫾ 4.91, n ⫽ 25, 46.33 ⫾ 3.32; t39 ⫽ ⫺0.294, p ⫽ .77; in 1991, n ⫽ 18, 48.38 ⫾ 4.91, n ⫽ 25, 46.33 ⫾ 3.32; t41 ⫽ ⫺0.439, p ⫽ .66). I monitored the clutches in the two experimental groups at 1–3 day intervals. Enemies attacking clutches (or females) were identified. When the oldest brood had molted into the second instar, I collected all of the experimental broods in 1987 in Nopporo and 1991 in Misumai. In 1990 in Misumai, broods were collected just before hatching. I compared the survival of offspring between the female-removal and the control groups. Under enemy-excluded conditions I conducted experiments using wild mulberry trees in the sericultural farm of the Hokkaido University in 1990. Nineteen branches with many shoots bearing fruit were selected and enclosed by cylindrical nylon nets (ca. 100 cm long, 40 cm diam). These enclosures contained a large amount of fruit to allow bugs to complete development. I introduced a copulating pair which had been collected from the Misumai popu-

Statistics Statistical analyses were performed with StatView 5.0 (SAS Institute, 1998), JMP 3.1 (SAS Institute, 1995), and SAS program software (SAS Institute, 1990). I used a sequential Bonferroni correction (Rice, 1989) to evaluate the table-wide significance. RESULTS Variation of reproductive traits and phenotypic selection A summary of the traits measured and phenotypic correlations among them (female body size excluded) are shown in Tables 1 and 2. As previously reported for this population in a different year and for another population (Kudo, 2001), there was a positive correlation between female body size and clutch size (rs ⫽ .505, n ⫽ 45, p ⫽ .0008) in the sample collected. No other significant correlations with body size were detected (see also Materials and Methods). The mean reproductive success for 57 females, as measured by the number of surviving nymphs, was 36.88 ⫾ 14.62 (SD),

Table 2 Phenotypic correlations among reproductive traits of E. putoni females on wild mulberry

Trait Clutch size Care duration Oviposition date Leaf size Leaf position No. of leaves No. of fruit

Care duration

Oviposi- Leaf tion date size

Leaf position

No. of leaves

No. of fruit

.199 (.137)

⫺.053 (.691) ⫺.005 (.967)

⫺.080 (.549) ⫺.013 (.923) .229 (.087) ⫺.413 (.002)

.119 (.373) ⫺.240 (.072) ⫺.081 (.546) .036 (.788) ⫺.175 (.191)

⫺.298 (.026) ⫺.513* (⬍.001) ⫺.157 (.241) ⫺.348 (.009) .135 (.312) .261 (.051)

.114 (.392) .314 (.019) .092 (.493)

Distance from ground ⫺.060 (.651) ⫺.282 (.035) ⫺.282 (.035) ⫺.277 (.038) ⫺.039 (.772) .233 (.081) .183 (.172)

See Table 1 for further explanations of traits. Spearman’s rank correlation coefficients are shown; values in parentheses are probabilities at the independent test. * p ⬍ .05 after the sequential Bonferroni correction.

Downloaded from http://beheco.oxfordjournals.org/ at Pennsylvania State University on March 6, 2014

Body size: prothorax width. Leaf size: length of the oviposition leaf. Leaf position: relative position of the oviposition leaf on a shoot (order of the leaf from the apex/no. leaves per shoot, 1 ⫽ the most basal position). No. of leaves: number of leaves on the shoot with the clutch. No. of fruit: number of fruits on the shoot with the clutch. Distance: the minimum distance from the ground to the oviposition leaf along the tree surface.

lation to each of the experimental branches. Other arthropods were removed from the experimental branches before the introduction of E. putoni pairs. I checked each branch daily to confirm ovipositions. When an oviposition occurred, I measured clutch size by temporarily removing the female from the clutch. Males were also removed from the branch at this time. The day after oviposition I removed the females from six randomly chosen clutches (the female-removal group). The other 13 females guarding clutches were kept intact as the control group. There was no significant difference in initial clutch size between the two groups (control, mean ⫾ SD ⫽ 48.38 ⫾ 4.91, female-removal, 46.33 ⫾ 3.32; t17 ⫽ 0.69, p ⫽ .50). Conducting frequent censuses that involve opening the nylon-net enclosures could disturb females attending nymphs on shoots. Thus, I checked experimental branches only once or twice a week. I compared the survival and the size of offspring (prothorax width) between the female-removal and the control groups when they had grown into adults. The development rate of offspring was not compared because oviposition did not occur simultaneously among females.

Kudo • Phenotypic selection in subsocial bugs

745

Table 3 Standardized selection differentials (Sⴕ) and standardized selection gradients (␤ⴕ) of linear (directional) selection on reproductive traits of E. putoni females on wild mulberry Trait Clutch size Care duration Oviposition date Leaf size Leaf position No. of leaves No. of fruit Distance

S⬘ 0.170 0.253 ⫺0.230 0.035 ⫺0.024 ⫺0.042 ⫺0.027 0.011

␤⬘ ⫾ SE

p ⬍.001* .002* .082 .622 .738 .549 .383 .897

0.147 0.186 ⫺0.134 ⫺0.017 0.017 ⫺0.059 0.064 0.045

⫾ ⫾ ⫾ ⫾ ⫾ ⫾ ⫾ ⫾

p 0.038 0.041 0.039 0.042 0.039 0.040 0.041 0.037

⬍.001* ⬍.001* .001* .680 .653 .147 .127 .236 Downloaded from http://beheco.oxfordjournals.org/ at Pennsylvania State University on March 6, 2014

See Table 1 for further explanations of traits. Significance test of selection differentials was made with Spearman’s rank correlation. The total model for selection gradients: r2 ⫽ .648, F8,48 ⫽ 11.02, p ⬍ .0001. *p ⬍ .05 after the sequential Bonferroni correction.

and the opportunity for selection was 0.157. Selection differentials for clutch size and care duration showed significant positive values (Table 3). Overall, females producing larger clutches and those taking care of offspring longer were favored during the breeding episode. The regression model was highly significant, and trait variables explained about 65% of the variance in female reproductive success (Table 3). Significant directional selection for clutch size, care duration, and oviposition date (positive values for the former two and a negative value for the latter) were detected. I conclude that earlier and larger clutch production with longer care by the female was selected directly during the breeding episode in the Misumai population. Experiments under field conditions The survival of offspring decreased rapidly in the female-removal group in the Nopporo population (Figure 1C). This was also the case for the Misumai population in different years (Figure 1A, B). In all of the three experiments, there were significant differences in offspring survival between the female-removal group and the control group (Mann-Whitney U test: Nopporo, z corrected for ties ⫽ ⫺4.01, p ⬍ .0001; Misumai in 1990, z ⫽ ⫺4.21, p ⬍ .0001; in 1991, z ⫽ ⫺5.13, p ⬍ .0001). Eggs and young nymphs in the female-removal group were attacked by a variety of arthropod enemies (Table 4). Some females in the control group (4 of 23 females in Nopporo, 2 of 25 in Misumai in 1990, 6 of 25 in 1991) had disappeared before the brood molted into the second instar. The reason for these disappearances was usually unknown, but four females were confirmed to be killed by spiders (by direct observations), indicating that even brooding females were not always free from predation. Excluding those orphaned broods from analysis, mean survival of offspring during experiments was high in all of the three control groups (Nopporo, mean ⫾ SD ⫽ 87.16 ⫾ 13.79%; Misumai in 1990, 98.55 ⫾ 1.80%; in 1991, 86.12 ⫾ 12.35%). To investigate the effect of clutch-size on predation rate, I performed regressions of offspring survivorship on clutch size (data were arcsine square-root transformed before analysis). No significant relationships were detected in any of the experimental groups in the two populations (female-removal: Nopporo, r2 ⫽ .023, F1,13 ⫽ 0.30, p ⫽ .59; Misumai in 1990, r2 ⫽ .018, F1,14 ⫽ 0.26, p ⫽ .62; in 1991, r2 ⫽ .043, F1,16 ⫽ 0.71, p ⫽ .41; control: Nopporo, r2 ⫽ .001, F1,21 ⫽ 0.03, p ⫽ .87; Misumai in 1990, r2 ⫽ .022, F1,23 ⫽ 0.52, p ⫽ .48; in 1991, r2 ⫽ .035, F1,23 ⫽ 0.84, p ⫽ .37).

Figure 1 Results of female-removal experiments in Nopporo and Misumai populations of E. putoni. Filled circles: female-removal groups in which parent females were removed from clutches. Open circles: control groups in which females were not removed. Means with SDs of brood survival are shown.

Behavioral Ecology Vol. 13 No. 6

746

Table 4 List of arthropod predators of E. putoni immatures on wild mulberry Stages of immatures attacked Predators Hemiptera Coleoptera Hymenoptera

Miridae Lygocoris (Nepolygus) sp. Nabidae gen. sp. Anthocoridae gen. sp. Coccinellidae gen. sp. Formicidae Formica yessensis Formica japonica Myrmica kotokui Myrmica jessensis Leptothorax sp. Anystidae gen. sp. Araneidae Araneus cornutus Theridiidae gen. sp. Thomisidae Oxytate striatipes Tmarus rimosus Xysticus sp. Salticidae gen. sp. Linyphiidae gen. sp.

Experiments under enemy-excluded conditions All of the 13 females in the control group continued to attend their nymphs when they left natal leaves to feed on fruit. There were no significant differences in bug density in enclosures between the female-removal and the control groups (female removal, mean ⫾ SD ⫽ 40.17 ⫾ 9.62; control, 38.15 ⫾ 9.10; t test, t17 ⫽ ⫺0.44, p ⫽ .67), and thus crowding effects can be ignored. No significant difference in offspring survival, measured as the proportion of adults eclosing in each clutch, was detected between the female-removal and the control groups (data were arcsine-root transformed, t test, t17 ⫽ ⫺1.18, p ⫽ .25; Figure 2A). There were also no significant differences in the body size of emerging adults in both sexes between the two groups (nested ANOVA with clutch [treatment] as a random variable: for males, treatment, F1,17.17 ⫽ 1.78, p ⫽ .20, clutch, F17,347 ⫽ 69.67, p ⬍ .0001; for females, treatment, F1,17.02 ⫽ 0.40, p ⫽ .54, clutch, F17,350 ⫽ 74.04, p ⬍ .0001; Figure 2B). DISCUSSION Since the comprehensive review of insect sociality by Wilson (1971), many studies have revealed that, in a variety of subsocial taxa, ecological factors such as stress from the physical environment, food availability, competition, and natural enemies are associated with the evolution of parental care (reviewed in Eickwort, 1981; Tallamy and Wood, 1986; Trumbo, 1996). In hervivorous insects, the most important factor leading to parental care is predation by other arthropods (e.g., Kudo et al., 1992; Nafus and Schreiner, 1988; Tallamy and Denno, 1981b; Windsor, 1987; Wood, 1976). In subsocial acanthosomatid bugs, offspring are potentially exposed to severe predation pressure, and parent females provide highly effective defense for their offspring (Kudo et al., 1989; Kudo and Nakahira, 1993; Mappes and Kaitala, 1994; Mappes et al., 1997; Melber and Schmidt, 1975; Melber et al., 1980). There are only a few subsocial insects (e.g., Elasmucha grisea; Mappes and Kaitala, 1994; Melber and Schmidt, 1975; Melber et al., 1980) in which effects of the parental care on offspring survival have been evaluated in different populations and/or different years, among which environmental pressures against the offspring might vary. Furthermore, the previous studies on insect parental care, with a few exceptions

X X X X X X X X

Nymphs X X X X

X X X X

X X X X X X

(e.g., Filippi et al., 2000; Kudo et al., 1995; Tallamy and Denno, 1981b), have not used experiments conducted under enemy-excluded conditions to evaluate the possible roles of the parent. Such experiments may provide evidence of functions of parental attendance other than simple protection from enemies if offspring survival and/or development is reduced in the experimental group. The present study and that by Honbo and Nakamura (1985) demonstrate that in different populations of E. putoni, arthropod predators are consistently important mortality agents of the offspring and that physical defense by the parent female is essential to offspring survival. The results of phenotypic selection analysis suggest that natural selection is acting on maternal care behavior in the study population: Females that leave their broods earlier had lower reproductive success. Although there is no quantitative measurement of potential predation against late instar nymphs, the defensive effectiveness of females against predators may be high. However, it should be noted that the present analysis does not completely cover the breeding episode of females; potential reproductive success of females with longer care was not considered. E. putoni females usually leave their broods during the period when the broods are in the second instar on wild mulberry in the field, and only some of the females attend the third instar broods (Honbo and Nakamura, 1985; Kudo, 1994, 1996, unpublished data). E. putoni nymphs depend on host fruits (probably developing seeds) for food (Kudo, 1994, unpublished data). They usually maintain aggregations in early instars. Forming an aggregation may provide benefits such as facilitating feeding for nymphs, but it may also induce competition between nymphs, particularly in later instars, for limiting food resources (e.g., Godfray, 1987). The optimal feeding aggregation size for nymphs would change with their growth and resource conditions. Wild mulberry has shoots bearing a small amount of fruits (three or four fruits per shoot), and maturing fruits fall off during the period when E. putoni nymphs are developing on them (Kudo, 1994, unpublished data). The fact that nymphal feeding aggregations on wild mulberry become smaller in later instars and that nymphs cease aggregating if fruits are experimentally removed from branches on which the nymphs are enclosed (Kudo, 1994, unpublished data) suggests growth-dependent competition for fruits among nymphs in aggregations. But nymphs in aggregations can also have

Downloaded from http://beheco.oxfordjournals.org/ at Pennsylvania State University on March 6, 2014

Acarina Araneida

Eggs

Kudo • Phenotypic selection in subsocial bugs

higher survival rates in the presence of maternal defense. E. putoni females usually settle on the stems of shoots which harbor nymphs feeding on fruit and face toward the base of the shoot (Kudo, 1996). This position and orientation of females attending nymphs is probably effective for detecting predators approaching the nymphs by walking along the stem. Females may keep their nymphs in aggregations on single shoots to provide better physical defense. As in some other Elasmucha species (Kaitala and Mappes, 1997; Kudo et al., 1989), E. putoni is semelparous in the field, which is mediated by host-plant phenology. Some females produce second clutches under laboratory conditions (Kudo, 1994, unpublished data). However, the experimental reduc-

tion of maternal care does not promote second clutch production in terms of clutch size and recovery period before oviposition (Kudo, 1994, unpublished data). There is also no correlation between the first clutch size and the recovery period. Even when females are removed from eggs and then reared, they lay second clutches more than 6 weeks after the removal. Nymphs derived from a second clutch cannot survive in the field because host fruits become unavailable far before they complete the development (Kudo, 1994, unpublished data). In fact, I have found no females that produce the second clutch in the field. Thus, E. putoni is considered a functionally semelparous species (sensu Tallamy and Brown, 1999). In such species, females would be free from the cost of parental care on their future survival and/or reproduction (e.g., Kudo et al., 1989, 1995; see also Tallamy and Brown, 1999). In E. putoni on wild mulberry, the duration of maternal care is probably determined by the balance between fitness benefits and costs to current offspring; the former derived from enhanced survival by the defense against predators and the latter from sib competition in maintained aggregations. The self-defense capabilities of nymphs will increase with development, while competition for host fruit will also increase. It is possible that taking care of nymphs longer imposes costs on females through the reduction of performance of the nymphs, and consequently a stabilizing selection is acting on the duration of maternal care. The detected directional selection on care duration reported here may be a part of such a stabilizing selection. Further studies will be needed to confirm this possibility. The determinant of female reproductive success in E. putoni was not parental care behavior alone: Larger initial clutch-size and earlier oviposition were also favored. If larger investment in egg production in a single breeding opportunity imposes energetic costs on females (e.g., Monaghan et al., 1998), females producing larger clutches might show shorter periods of care behavior n. In the data reported here (Table 2), however, such a trade-off was not detected. Even when females were monitored until they completed care, no correlation was detected between clutch size and care duration (Kudo, 1994, unpublished data). The offspring situated on the periphery of the clutch are more vulnerable than those on the center in Elasmucha bugs as well as in other subsocial insects in which parent females straddle their clutches to guard them (e.g., Eberhard, 1975; Mappes and Kaitala, 1994; Mappes et al., 1997). Thus, larger clutch size relative to female body size probably reduces defense efficiency. Clutch size is positively correlated with female body size in E. putoni (Kudo, 2001). Although effects of body size of females on their reproductive success were not evaluated directly in the present analysis, body size, as well as clutch size, may also be under positive directional selection. The mechanisms and agents responsible for selective advantages from early oviposition are unclear. In phytophagous insects, the timing of oviposition is often an important determinant for offspring performance through phenological interactions of the insects and host plants (e.g., Keese and Wood, 1991). This appears to be the case in E. putoni (see above). Wild mulberry trees leaf in May in northern Japan, and E. putoni females begin oviposition soon after completion of leaf expansion and flowering (Kudo, 1994). When the second instar broods were collected (20–28 June), many of the mulberry fruits were mature and still remained on the tree. The total amount of mulberry fruit was almost at the maximum. First-instar nymphs remain on natal leaves and take no food (Honbo and Nakamura, 1985). Thus, at the time of the collection, the nymphs had fed on fruit for only 2 days. It is unlikely that, within the observed variation in the oviposition

Downloaded from http://beheco.oxfordjournals.org/ at Pennsylvania State University on March 6, 2014

Figure 2 Comparison of (A) offspring survival and the (B) body size (prothorax width) of eclosing adults between the female-removal and control groups under enemy-excluded conditions. In the female-removal group, parent females were removed from clutches (n ⫽ 6). In the control group, females were not removed (n ⫽ 13). Total number of individuals measured: 366 males and 369 females. See the text for statistics. Means with SDs are shown.

747

748

I thank S. Makino, M. Ohara, Y. Sakamaki, M. Sato, and K. Sayama for assistance in the field, and F. Ito, J. Ogura, Y. Saito, M. Suwa, and T. Yasunaga for identifying predators. T. Iizuka permitted me to use the sericultural farm. I also thank S. Akimoto, E. Kasuya, and H. Ueno for their suggestions for the analysis and M. E. Carter, A. F. G. Bourke, and two anonymous referees for their helpful comments on the manuscript. This study was supported in part by Grants-in-Aid (nos. 09740581, 11740430, 13440227) for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology, Japan.

REFERENCES Arnold SJ, Wade MJ, 1984. On the measurement of natural and sexual selection: theory. Evolution 38:709–719. Brodie ED III, Moore AJ, Janzen FJ, 1995. Visualizing and quantifying natural selection. Trends Ecol Evol 10:313–318. Chatterjee S, Price B, 1977. Regression analysis by example. New York: John Wiley & Sons. Choe JC, Crespi BJ, 1997. The evolution of social behavior in insects and arachnids. Cambridge: Cambridge University Press. Clutton-Brock TH, 1991. The evolution of parental care. Princeton, New Jersey: Princeton University Press. Diesel R, 1989. Parental care in an unusual environment: Metopaulias depressus (Decapoda: Grapsidae), a crab that lives in epiphytic bromeliads. Anim Behav 38:561–575. Eberhard WG, 1975. The ecology and behavior of a subsocial pentatomid bug and two scelionid wasps: strategy and counterstrategy in a host and its parasites. Smithson Contrib Zool 205:1–39. Eickwort GC, 1981. Presocial insects. In: Social insects, vol. 2 (Hermann HR, ed). New York: Academic Press; 199–280. Filippi L, Hironaka M, Nomakuchi S, Tojo S, 2000. Provisioned Parastrachia japonensis (Hemiptera: Cydnidae) nymphs gain access to food and protection from predators. Anim Behav 60:757–763. Filippi-Tsukamoto L, Nomakuchi S, Kuki K, Tojo S, 1995. Adaptiveness parental care in Parastrachia japonensis (Hemiptera: Cydnidae). Ann Entomol Soc Am 88:374–383. Fink LS, 1986. Costs and benefits of maternal behaviour in the green lynx spider (Oxyopidae, Peucetia viridans). Anim Behav 34:1051– 1060. Godfray HCJ, 1987. The evolution of clutch size in invertebrates. In: Oxford surveys in evolutionary biology, vol. 4 (Harvey PH, Partridge L, eds). Oxford: Oxford University Press; 117–154. Grafen A, Sibly R, 1978. A model of mate desertion. Anim Behav 26: 645–652. Gross P, 1993. Insect behavioral and morphological defenses against parasitoids. Annu Rev Entomol 38:251–273. Honbo Y, Nakamura K, 1985. Effectiveness of parental care in the bug Elasmucha putoni Scott (Hemiptera: Acanthosomatidae) [in

Japanese with English summary]. Jpn J Appl Entomol Zool 29:223– 229. Jeanne RL, 1979. A latitudinal gradient in rates of ant predation. Ecology 60:1211–1224. Kaitala A, Mappes J, 1997. Parental care and reproductive investment in shield bugs (Acanthosomatidae, Heteroptera). Oikos 80:3–7. Keese MC, Wood TK, 1991. Host-plant mediated geographic variation in the life history of Platycotis vittata (Homoptera: Membracidae). Ecol Entomol 16:63–72. Kudo S, 1994. Comparative ecology of parental care in acanthosomatid bugs (PhD dissertation) [in Japanese]. Sapporo, Japan: Hokkaido University. Kudo S, 1996. Ineffective maternal care of a subsocial bug against a nymphal parasitoid: a possible consequence of specialization to predators. Ethology 102:227–235. Kudo S, 2001. Intraclutch egg-size variation in acanthosomatid bugs: adaptive allocation of maternal investment? Oikos 92:208–214. Kudo S, Ishibashi E, Makino S, 1995. Reproductive and subsocial behaviour in the ovoviviparous leaf beetle Gonioctena sibirica (Weise) (Coleoptera: Chrysomelidae). Ecol Entomol 20:367–373. Kudo S, Maeto K, Ozaki K, 1992. Maternal care in the red-headed web-spinning sawfly, Cephalcia isshikii (Hymenoptera: Pamphiliidae). J Insect Behav 5:783–795. Kudo S, Nakahira T, 1993. Brooding behavior in the bug Elasmucha signoreti (Heteroptera: Acanthosomatidae). Psyche 100:121–126. Kudo S, Sato M, Ohara M, 1989. Prolonged maternal care in Elasmucha dorsalis (Heteroptera, Acanthosomatidae). J Ethol 7:75–81. Lande R, Arnold SJ, 1983. The measurement of selection on correlated characters. Evolution 37:1210–1226. Lazarus J, 1990. The logic of mate desertion. Anim Behav 39:672–684. Maynard Smith J, 1977. Parental investment: a prospective analysis. Anim Behav 25:1–9. Mappes J, Kaitala A, 1994. Experiments with Elasmucha grisea L. (Heteroptera: Acanthosomatidae): does a female parent bug lay as many eggs as she can defend? Behav Ecol 5:314–317. Mappes J, Kaitala A, 1995. Host-plant selection and predation risk for offspring of the parent bug. Ecology 76:2668–2670. Mappes J, Mappes T, Lappalainen T, 1997. Unequal maternal investment in offspring quality in relation to predation risk. Evol Ecol 11:237–243. Melber A, Ho¨scher L, Schmidt GH. 1980. Further studies on the social behaviour and its ecological significance in Elasmucha grisea L. (Hem.-Het.: Acanthosomatidae). Zool Anz 205:27–38. ¨ ologische Bedeutung des SozialverMelber A, Schmidt GH, 1975. O haltens zweier Elasmucha-Arten (Heteroptera: Insekta). Oecologia 18:121–128. Mitchell-Olds T, Shaw RG, 1987. Regression analysis of natural selection: statistical inference and biological interpretation. Evolution 41:1149–1161. Monaghan P, Nager RG, Houston DG, 1998. The price of eggs: increased investment in egg production reduces the offspring rearing capacity of parents. Proc R Soc Lond B 265:1731–1735. Nafus DM, Schreiner IH, 1988. Parental care in a tropical nymphalid butterfly Hypolimnas anomala. Anim Behav 36:1425–1431. Nalepa CA, 1988. Cost of parental care in the woodroach Cryptocercus punctulatus Scudder (Dictyoptera: Cryptocercidae). Behav Ecol Sciobiol 23:135–140. Rice WR, 1989. Analyzing tables of statistical tests. Evolution 43:223– 225. Roff DA, 1992. The evolution of life histories: theory and analysis. London: Chapman and Hall. SAS Institute, 1990. SAS/STAT user’s guide, version 6. Cary, North Carolina: SAS Institute. SAS Institute, 1995. JMP statistics and graphics guide, version 3.1. Cary, North Carolina: SAS Institute. SAS Institute, 1998. StatView 5.0. Cary, North Carolina: SAS Institute. Scott MP, Traniello JFA, 1990. Behavioural and ecological correlates of male and female parental care and reproductive success in burying beetles (Nicrophorus spp.). Anim Behav 39:274–283. Stearns SC, 1992. The evolution of life histories. Oxford: Oxford University Press. Tallamy DW, Brown WP, 1999. Semelparity and the evolution of maternal care in insects. Anim Behav 57:727–730. Tallamy DW, Denno RF, 1981a. Alternative life history patterns in risky environments: an example from lace bugs. In: Insect life his-

Downloaded from http://beheco.oxfordjournals.org/ at Pennsylvania State University on March 6, 2014

date, later oviposition had induced keen competition and high mortality among the second-instar nymphs during the 2 days. Although females that laid clutches later did not leave offspring earlier (Table 2), clutches deposited later suffered higher mortality; there was a significant negative correlation between survival rates and oviposition date (rs ⫽ ⫺.322, p ⫽ .016). In E. putoni, maternal defense is highly effective but not perfect: Under field conditions up to 20% of offspring were lost in female-guarded broods at the time of collection. If predation pressure increases rapidly with the breeding season of E. putoni, this could explain the advantage of earlier oviposition. In insects, there are a variety anti-enemy tactics other than parental defense (e.g., Gross, 1993; Mappes and Kaitala, 1995). Heteropterans living in environments in which the offspring suffer high mortality may adopt alternative reproductive strategies (e.g., Tallamy and Denno, 1981a). Such alternatives may explain the current distribution of parental care in the Heteroptera (Tallamy and Schaefer, 1996). The quantification of selection on multiple reproductive traits in related species will promote our understanding of the evolution and maintenance of the diversified reproductive strategies.

Behavioral Ecology Vol. 13 No. 6

Kudo • Phenotypic selection in subsocial bugs

tory patterns: habitat and geographic variation (Denno RF, Dingle H, eds). New York: Springer Verlag; 129–147. Tallamy DW, Denno RF, 1981b. Maternal care in Gargaphia solani (Hemiptera; Tingidae). Anim Behav 29:771–778. Tallamy DW, Denno RF, 1982. Life history trade-offs in Gargaphia solani (Hemiptera: Tingidae): the cost of reproduction. Ecology 63: 616–620. Tallamy DW, Schaefer CW, 1997. Maternal care in the Hemiptera: ancestry, alternatives, and current adaptive value. In: The evolution of social behavior in insects and arachnids (Choe JC, Crespi BJ, eds). Cambridge: Cambridge University Press; 94–115. Tallamy DW, Wood TK, 1986. Convergence patterns in subsocial insects. Annu Rev Entomol 31:369–390. Trumbo ST, 1996. Parental care in invertebrates. Adv Stud Behav 25: 3–51.

749

Wade MJ, Kalisz S, 1990. The causes of natural selection. Evolution 44:1947–1955. Wilson EO, 1971. The insect societies. Cambridge: Belknap Press. Windsor DM, 1987. Natural history of a subsocial tortoise beetle, Acromis sparsa Boheman (Chrysomelidae, Cassidinae) in Panama. Psyche 94:127–150. Winkler DW, 1987. A general model for parental care. Am Nat 130: 526–543. Wood TK, 1976. Alarm behavior of brooding female Umbonia crassicornis (Homoptera: Membracidae). Ann Entomol Soc Am 69:340– 344. Yamamura N, Tsuji N, 1993. Parental care as a game. J Evol Biol 6: 103–127. Zeh DW, Smith RL, 1985. Paternal investment by terrestrial arthropods. Am Zool 25:785–805.

Downloaded from http://beheco.oxfordjournals.org/ at Pennsylvania State University on March 6, 2014

Suggest Documents