Behavioral Ecology Advance Access originally published online on July 21, 2004
Behavioral Ecology 2005 16(1):1-7; doi:10.1093/beheco/arh122
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Addition of arthropod cocoons to house wren nests is correlated with delayed pairing
Behavior, Ecology, Evolution, and Systematics Section, Department of Biological Sciences, Illinois State University, Normal, IL 61790-4120, USA
Address correspondence to K.P. Eckerle, who is now at Behavior, Ecology, Evolution and Systematics Section, Department of Biological Sciences, Illinois State University, Normal, IL 61790-4120. E-mail: keckerle{at}umich.edu.
Received 8 December 2003; revised 19 March 2004; accepted 21 April 2004.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Males in the cavity-nesting house wren (Troglodytes aedon) frequently add arthropod cocoons to their nests during building, possibly as an ornamental cue for female choice. We tested this hypothesis by comparing the time to pairing for males that did and did not add cocoons to their nests and for males in whose nests we manipulated the number of cocoons prior to pairing. We also tested the hypothesis that females acquire fitness-related benefits by selecting mates based on their use of cocoons. The use of cocoons by males was not consistently related to habitat, but the number of cocoons added per nest increased during the course of the breeding season. Contrary to prediction, the time to pairing for males adding cocoons was significantly longer than that for males without cocoons in their nests at both unmanipulated and experimental nests. There was also no consistent fitness-related benefit for females related to the use of cocoons by their mates. Therefore, we conclude that females did not prefer males that added cocoons to their nests, and that the increased time to pairing for males that add cocoons likely results in fitness-related costs brought about by delayed breeding. Nonetheless, male house wrens routinely use cocoons, and why they do so remains unknown.
Key words: female choice, house wrens, nest building, sexual selection, Troglodytes aedon.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
In sexually monomorphic and monochromatic species of birds, sexual selection often leads to the evolution of conspicuous male behavioral traits (Andersson, 1994
). One trait known to influence female choice in birds is male nest building (Soler et al., 1998), with females preferring males building the largest number of dummy (cock) nests (see Savalli, 1994; Verner and Engelsen, 1970) or the biggest nests (see Grubbauer and Hoi, 1996). Females may use male nest building to assess male condition (Evans, 1997b), territory quality (Evans, 1997a), the thermal environment for their eggs (Grubbauer and Hoi, 1996), or the quality of future paternal care (Moreno et al., 1994; Palomino et al., 1998).
Male nest building could play a role in female mate choice in house wrens (Troglodytes aedon), because males generally arrive on the breeding grounds prior to females, initiate nest building in one or more potential nest sites on their territory, and usually complete a stick platform within a nest cavity before attracting a mate (Kendeigh, 1941; McCabe, 1965). Although neither the quantity of nesting material added to a single nest cavity (Alworth, 1996) nor the number of dummy nests a male builds (Eckerle, 2001; Kendeigh, 1941) appears to influence female choice, males also routinely add arthropod silk and egg sacs (cocoons) to their nests (see Methods) and such "decorations" could serve as an ornament for mate attraction (Collias and Collias, 1984; Skutch, 1976).
If a trait is to serve as a mate-choice cue, it must vary among males, be assessable by females before mating, and be related to male mating success (Kirkpatrick and Ryan, 1991; Searcy, 1982). The first two criteria are met in house wrens: males vary in their use of cocoons (Eckerle KP and Thompson CF, unpublished data; McCabe, 1965; Pacejka et al., 1996) and display their nest site to females during courtship (Johnson, 1998; Kendeigh, 1941). In the present study, we investigated whether female house wrens use cocoons as a basis for mate choice by analyzing the relationship between cocoon use and male mating success.
We tested the hypothesis that females preferred males adding cocoons to their nests by comparing the time to pairing for males (1) that did and did not add cocoons and (2) in whose nest we manipulated the number of cocoons before they acquired a mate. If females select mates based on their use of cocoons, we predicted that in unmanipulated nests time to pairing would be significantly shorter for males adding cocoons to their nests than for males not adding cocoons. Likewise, at experimental nests, we predicted that the time to pairing would be significantly different across treatments, with the time to pairing decreasing as the number of cocoons in the nest increased. Finally, if females derive direct benefits by selecting males that added cocoons, we predicted that one or more of the following should occur in unmanipulated nests: (1) clutch size and number and condition of offspring produced should be significantly greater at nests with than without cocoons, (2) male provisioning rates should be significantly higher at nests with than without cocoons, and (3) male provisioning rates should be positively correlated with the number of cocoons added.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Species, study site, and cocoon use
The house wren is a small (10–12 g), sexually monomorphic and monochromatic migratory songbird that breeds throughout much of North America (Johnson, 1998
). House wrens breeding at the Mackinaw study site in McLean County, Illinois (40°40' N, 88°53' W), overwinter along the coast of the Gulf of Mexico and arrive on the study site in late April. Since 1982 at this site, 585 identical nest-boxes, each mounted on a 1.5-m steel pole, have been aligned 30 m apart on north–south lines 60 m apart in upland and floodplain forests of oak (Quercus spp.), maple (Acer spp.), hickory (Carya spp.), hackberry (Celtis occidentalis), and ash (Fraxinus spp.; Drilling and Thompson, 1988). Males arrive first on the breeding site (Neill, 1990) and establish territories around several nest-boxes, into which they add small sticks and twigs (Kendeigh, 1941). Females are attracted to potential nest sites by male song (Johnson and Searcy, 1996) and inspect the cavity containing the stick platform (Kendeigh, 1941); once paired, females complete nest building (Alworth and Scheiber, 2000; McCabe, 1965). House wrens on the study site are double brooded, with eggs of the first brood produced in early to mid May and eggs of the second brood in late June to early July (Finke et al., 1987).
Males bring individual cocoons to the nest independently of other nesting material (i.e., not attached to sticks). We have watched banded males (more than 10 observation periods over a number of years) transport both natural and artificial cocoons to their nests. Others in eastern North America have watched males bring cocoons to their nests (Alworth T, Llambia PE, personal communication), and P.E. Llambia (personal communication) has also seen males in Argentina carry cocoons to nests in both nest-boxes and natural cavities. Kendeigh (1941: 99; 1952: 14) reported observing males inserting "spider nests" while building. Cocoons are typically placed among the sticks and twigs of the nest platform that males are building. They are also attached to vertical sticks and twigs that rise above the nest platform, stuck to the inner walls of the nest-box, and placed near the nest-box entrance, where they frequently span or hang down from the cavity entrance and are sometimes visible from outside the nest-box (Eckerle KP, personal observations). Cocoons are not used to attach or bind the nest (sensu Hansell, 2000).
Although some cocoons could be placed in nest-boxes by the arthropods themselves, the preponderance of evidence indicates that male house wrens are the source of most of these cocoons. In addition to our and others' direct observations of males carrying single cocoons to their nests, the timing of the appearance of cocoons supports this conclusion. Cocoons appear in nest-boxes before females return to the breeding grounds in spring and, later in the season, before females are attracted to an advertising male (Eckerle KP, personal observation). Furthermore, cocoons normally appear in nest-boxes only after a male has begun to build a stick platform (Eckerle KP, Thompson CF, personal observations).
Testing the time to pairing
Unmanipulated nests
We collected data in 1996–1998 on cocoon use during nest building and male time to pairing. We removed all nest material from each nest-box before male arrival in 1997 and 1998 in order to determine exactly when males began adding nest material. We did not do this in 1996, but in most cases it was clear when males began adding new nest material. We began checking nest-boxes before male arrival and continued thereafter at least twice weekly to determine male settlement and pairing dates and to count the cocoons in each newly built nest. Following the method of Johnson and Searcy (1993) we considered males to have settled when new nest material first appeared in a nest-box and paired when nest-lining material was first added. We defined male time to pairing as the time from settlement to pairing.
Experimental nests
Before male arrival and after each nesting attempt throughout the 1997 breeding season, we removed all nest material from the 310 nest-boxes north of the Mackinaw River (see Drilling and Thompson, 1988). To control for differences in male and territory quality, we randomly assigned experimental treatments just after male arrival, according to each male's use of cocoons early in nest building. Where males did not add cocoons, we left the nest as a control (without-control) or added 10 artificial cocoons (without-add); artificial cocoons were small cotton balls stained with coffee and were readily accepted by males and females. Where males added cocoons, we removed all male-added cocoons (with-remove), left the nest as a control (with-control), or added 10 artificial cocoons to the nest (with-add). We did not initiate treatments until at least 50% of the nest-box floor was covered with nest material because males may add a few sticks to several nest-boxes. We checked nests at least every other day to count the cocoons in each newly built nest, to maintain treatments by replacing missing cocoons or removing new cocoons, and to determine male pairing date (following the method of Johnson and Searcy, 1993). We calculated time to pairing for each male as the time from treatment initiation until pairing.
Reproductive success
Following pairing at unmanipulated nests in 1996 and 1998, we regularly visited each nest to determine clutch initiation date (egg-1 day), clutch size, and when the first egg hatched (brood-day 0; see Harper et al., 1992). The number of young fledged was calculated as the number of nestlings on brood-day 12 minus the number of dead nestlings remaining in the nest after fledging 2–4 days later.
Measuring adults and nestlings
We used sliding trapdoors permanently attached to each nest-box or mist-nets to catch adults during the incubation and nestling stages at unmanipulated nests in 1996 and 1998. During 1998 we also weighed nestlings on brood-days 0, 2, 4, 6, 8, 10, and 12. We weighed all birds to the nearest 0.1 g on a digital balance (Acculab Pocket Pro 250) and banded adults at first capture and nestlings on brood-day 12 with a numbered US Fish and Wildlife Service aluminium leg band. Males also were banded with a unique combination of three colored plastic leg bands (total of four bands, two per leg). In 1998, we also measured the flattened wing chord (wing) and tail length of each adult to the nearest 0.5 mm with a stopped metal ruler (Avinet, Inc.), and the right tarsometatarsus (tarsus) length for each adult and nestling to the nearest 0.1 mm with dial calipers (Avinet, Inc.; following the method of Svensson, 1992).
Male and female body mass vary with time of day and that of females also with the stage of the nesting cycle (see Cavitt and Thompson, 1997). Therefore, for adults captured in 1996, we calculated female residual mass from a multiple linear regression of body mass on time from egg-1 day to capture and hour of capture (R2 = .05), and male residual mass from a linear regression of body mass on hour of capture (R2 = .07) (PROC GLM; SAS Institute, 1990). For adults captured in 1998, we calculated female condition index as the residual from a multiple linear regression of body mass on tarsus, wing, and tail length; the period from egg-1 day to capture; and hour of capture (R2 = .48). For male condition index, we used the residual from a multiple linear regression of body mass on tarsus, wing, tail length, and hour of capture (R2 = .25). Adult residual mass and condition index were correlated in 1998 (Pearson product-moment correlations; females: r = .93, N = 97, p < .001; males: r = .89, N = 55, p < .001).
Nestling mass varies with time of day and season (Cavitt and Thompson, 1997; Finke et al., 1987); therefore, we calculated nestling condition index as the residual from a multiple linear regression of brood-day 12 nestling mass on tarsus length, hatching date, and hour of weighing (R2 = .24; PROC GLM, SAS Institute, 1990). To avoid pseudo-replication, we used brood means for mass and condition index. For broods in which nestlings were weighed at least five times, we used nonlinear least-squares estimation (PROC NLIN; SAS Institute, 1990) to calculate an exponential rate constant for nestling growth (g) by fitting mean nestling mass on successive 2-day intervals to the generalized logistic growth equation (following the method of Styrsky et al., 1999). This constant describes the rate (day–1) at which asymptotic mass is achieved.
Adult provisioning of nestlings
We measured adult provisioning rates during a single observation period after 0900 h on brood-days 8, 9, or 10 in 1996 and 1998. All but four observations were carried out between 1300–1630 h, and none were made during rain. Observation periods consisted of either 45–60-min visual watches by using a Bushnell 60-mm spotting telescope or 34–60-min video camera recording sessions. Visual observations were made from at least 30 m, and data were not recorded until 10 min after observer arrival and when there was no evidence of adult response to the observer. To acclimate adults to the video camera, 24 h before filming we placed a 3.8-l, black-painted milk jug atop a 1.5-m metal pole approximately 3 m from the nest-box, where the video camera would later be placed. We also monitored adult behavior while reviewing each videotape and discarded the session if adults still responded to the camera after 10 min (approximately 6% of sessions).
Wrens typically deliver only one prey item per trip to the nest (Johnson, 1998), but provisioning rates vary with time of day (Kendeigh, 1952), season, and brood size (Eckerle KP, unpublished data). Therefore, we obtained residuals from multiple linear regressions of female and male per capita provisioning rate on hour of observation, brood-day 0, and brood size to generate adult provisioning indices (1996: females, R2 = .03; males, R2 = .14; 1998: females, R2 = .12; males, R2 = .08; PROC GLM; SAS Institute, 1990). We also correlated the number of cocoons added to nests and adult provisioning rates by calculating a per capita provisioning index from linear regressions of male and female per capita provisioning rate on hour of observation (PROC GLM; SAS Institute, 1990).
Data analysis
Data exclusion
Males initiated 1265 nests over the 3 years, but we confined our analyses to the first nest of each male and female in each year. Therefore, we excluded data from (1) the second or third nest of an individual male or female in the same year (N = 310), (2) any partial or completed nest in which no eggs were laid (N = 199), (3) any nest at which time to pairing exceeded 21 days (N = 31), and (4) any nest at which the presence or absence of cocoons could not be determined (N = 12). We also excluded data from experimental nests at which males did not initially add cocoons, but did so after treatment initiation (N = 35). This reduced the number of unmanipulated and experimental nests to 611 and 67, respectively. We classified each of these nests as an early-season nest (egg-1 day before that season's median egg-1 day) or a late-season nest (egg-1 day on or after the median egg-1 day) and by habitat (upland or floodplain).
Statistical tests
We used Statistical Analysis System Software for PCs (SAS Institute, 1990) for all analyses. We analyzed years separately because not all individuals were identified in each year, making it likely that some birds were present, but undetected, in more than 1 year, thus violating the assumption of independent observations. At unmanipulated nests, we tested the effects of season (early versus late), habitat (floodplain versus upland), and their interaction on the proportion of unmanipulated and experimental nests with cocoons added with tests of independence using maximum likelihood (PROC CATMOD). We also used a two-factor ANOVA (PROC GLM) to test the effects of season, habitat, and their interaction on the mean number of cocoons added per unmanipulated nest; we transformed (log10) the number of cocoons added per nest to meet test assumptions. We compared the proportion of experimental nests in upland and floodplain habitats with and without cocoons added before treatment and the proportion of nests in upland and floodplain habitats in each treatment (PROC CATMOD).
We conducted independent analyses of male time to pairing at unmanipulated and experimental nests by using accelerated failure-time analysis (PROC LIFEREG; Fox, 1993), a parametric method for estimating survivorship functions. For these analyses, we selected a log-normal survival distribution and determined the distributional parameters and regression coefficients for each factor by using maximum likelihood (Fox, 1993). We tested for effects of season and habitat, treating season as a continuous variable, using the initiation date of either nest building (unmanipulated nests) or treatment (experimental nests), and habitat as a categorical variable (upland versus floodplain). We also tested for the effect of cocoon use (presence versus absence) at unmanipulated nests and of treatment at experimental nests. If there was a significant treatment effect, we conducted pair-wise comparisons using Wald chi-square statistics (following the method of Allison, 1995).
At unmanipulated nests, we used t tests (PROC TTEST) to compare adult residual mass and condition indices, the mass and condition index of nestlings, and the provisioning index of adults. We used ANCOVA (PROC GLM) to compare clutch size, the number of young fledged, and nestling growth rates, using clutch initiation date (for analyses of clutch size and number of young fledged) or brood-day 0 (for analysis of nestling growth rates) as a covariate to control for the effect of season. We used one-tailed Spearman rank correlations (PROC CORR) to compare the number of cocoons added per nest and adult provisioning rates at unmanipulated nests because we predicted a positive correlation.
Data are presented as means (±1 SE) throughout. All t tests were equal variance tests, unless otherwise specified. Differences were considered significant if p 
.05. For analyses of experimental data, we used a Bonferroni correction (Sokal and Rohlf, 1995) to calculate the comparison-wise 
-value for post-hoc comparisons. Sample sizes vary because some nests were excluded from particular analyses because of missing data.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Unmanipulated nests
Use of cocoons
Males added cocoons to 50.1% (1996: 47.5%, N = 305; 1997: 54.1%, N = 98; 1998: 51.9%, N = 208) of unmanipulated nests. The proportion of nests to which males added cocoons was independent of season (early versus late; 1996:
p = .14; 1997: 
p = .67; 1998: 
p = .27), and there was no season by habitat interaction in any year (1996: 
p = .87; 1997: 
p = .10; 1998: 
p = .65). Upland nests were more likely than were floodplain nests to have cocoons added in 1997 (
p = .05) but not in 1996 (
p = .26) or 1998 (
p = .09) (Figure 1a).
|
Males added one to 35 cocoons per unmanipulated nest (mean = 4.40 ± 0.28 cocoons, N = 301). The number of cocoons in nests varied with both season and habitat, although there was no significant season by habitat interaction in any year (all p
.31). More cocoons were added to nests in floodplain than upland habitat in 1997 (F1,49 = 9.92, p = .003), but the effect of habitat was not significant in 1996 (F1,139 = 0.19, p = .66) or 1998 (F1,101 = 0.72, p = .40). In each year, more cocoons were added to late-season than early-season nests (1996: F1,139 = 25.91, p < .0001; 1997: F1,49 = 6.50, p = .01; 1998: F1,101 = 31.10, p < .0001) (Figure 1b). Twelve of 25 males building two nests in one year added cocoons to both nests, three did not add cocoons, and 10 added cocoons to one nest only. Nine of 10 males added cocoons to each nest they built over 2 years; the remaining male did not add cocoons in one year, but did to one of two nests the following year. The majority of cocoons were those of spiders (Families Tetragnathidae, Araneidae, Clubionidae, and Theridiidae) and the white tussock moth (Orgyia leucostiga), all of which are common in deciduous forest (Sierwald P, personal communication).
Time to pairing
Contrary to predictions, time to pairing each year was significantly longer for males that added cocoons to their nests (1996: 
2 = 11.60, p = .001; 1997: 2 = 6.43, p = .01; 1998: = 29.70, p < .001) (Figure 2a). Male time to pairing also decreased significantly in each year as the breeding season progressed (1996: ß = –0.004 ± 0.002, 2 = 3.70, p = .05; 1997: ß = –0.006 ± 0.003, 2 = 3.01, p = .08; 1998: ß = –0.004 ± 0.002, 2 = 3.75, p = .05) but was not significantly related to habitat in any year (1996: 2 = 0.12, p = .73; 1997: 2 = 0.55, p = .46; 1998: 2 = 0.71, p = .40).
|
Adult mass and condition
The residual mass in 1996 and the condition index in 1998 of females mated to males that did and did not add cocoons were not significantly different (Table 1). Similarly, the residual mass in 1996 and the condition index in 1998 of males that did and did not add cocoons were not significantly different (Table 1).
|
Offspring number, mass, and condition
Clutch size and number of young fledged decreased significantly across the season in 1996 (clutch size: F1,222 = 141.96, p < .0001, ß = –0.029 ± 0.002; number of young fledged: F1,105 = 12.40, p = .0006, ß = –0.020 ± 0.006) and 1998 (clutch size: F1,182 = 62.04, p < .0001, ß = –0.027 ± 0.003; number of young fledged: F1,66 = 8.15, p = .006, ß = –0.028 ± 0.010). Neither clutch size nor number of young fledged differed significantly between nests with and without cocoons in 1996 (clutch size: F1,222 = 0.05, p = .83; number of young fledged: F1,105 = 0.22, p = .64) or 1998 (clutch size: F1,182 = 0.00, p = .99; number of young fledged: F1,66 = 0.25, p = .62). In 1998, nestling growth rates at nests with cocoons and without cocoons were not significantly related to season (F1,34 = 0.37, p = .55, ß = –0.0004 ± 0.001) or cocoon use (F1,34 = 0.13, p = .72). Likewise, mean nestling mass and mean nestling condition index were not significantly different on brood-day 12 at nests with and without cocoons (Table 1).
Provisioning of nestlings
Neither male nor female provisioning indices were significantly different at nests with and without cocoons in 1996 or 1998 (Table 1). At nests with cocoons added in 1996 (N = 14), male (rs = .60, p = .02), but not female (rs = .16, p = .59), provisioning index was significantly correlated with the number of cocoons added; at nests with cocoons added in 1998 (N = 9), neither male (rs = .15, p = .70) nor female (rs = –0.49, p = .18) provisioning index was significantly correlated with the number of cocoons added.
Experimental nests
Use of cocoons
Males added cocoons to 79.1% (N = 67) of experimental nests. Before treatment, the proportion of nests with cocoons added was marginally higher in floodplain (92.0%, N = 25) than upland habitat (71.4%, N = 42) (
p = .06). The proportion of nests with cocoons in upland and floodplain habitat did not vary with treatment (
p = .40).
Before manipulating cocoon abundance, the major difference in the number of cocoons per nest was attributable to males either adding or not adding cocoons; among males that added cocoons, the cumulative number of cocoons added to each nest was not significantly different (all p > .18) (Table 2). As intended, the experimental manipulation of cocoons resulted in a significant difference across treatments in the total number of cocoons (natural + artificial) in each nest that females encountered (F4,62 = 18.50, p < .0001, R2 = .54) (Table 2). Among males that added cocoons, the experimental manipulation of cocoons resulted in addition-treatment (with-add) males having significantly more cocoons per nest than either control (with-control) or removal-treatment (with-remove) males (p < .0001); the number of cocoons present in nests of control (with-control) and removal-treatment (with-remove) males were not significantly different (p = .17). Among males that did not add cocoons, the experimental manipulations resulted in addition-treatment males (without-add) having significantly more cocoons per nest than control-treatment (without-control) males (p = .005).
|
Time to pairing
Habitat did not significantly affect male time to pairing (
2 = 1.17, = .28), but male time to pairing tended to decrease over the breeding season (ß = –0.01 ± 0.01, 2 = 3.20, p = .07). Treatment significantly affected male time to pairing (2 = 32.75, p < .0001) (Figure 2b); however contrary to expectation, control-treatment males not adding cocoons (without-control) paired significantly faster than did control-treatment males adding cocoons (with-control) (Bonferroni corrected value = 0.01; 
p < .01). Similarly, among males that added cocoons, the time to pairing for removal-treatment males (with-remove) was significantly shorter than for control-treatment males (with-control; 
p < .01) and addition-treatment males (with-add; 
p < .01). Male time to pairing did not differ significantly among males not adding cocoons (without-control versus without-add; 
p > .10) or among control (with-control) and addition-treatment males (with-add; 
p > .10). |
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Use of cocoons
Male house wrens added cocoons to half of the first nests they built in each year. Cocoon use is not unique to this house wren population, as it occurs across much of the species' breeding range (see McCabe, 1965
). The use of cocoons did not vary consistently with habitat but increased seasonally in each year, paralleling seasonal increases in spider populations (Kendeigh, 1979; Williams et al., 1995). Variation in cocoon use was unrelated to differences in male body mass or condition index.
Mate-choice function of cocoon use
The use of cocoons by male house wrens during nest building meets two of the three criteria (Kirkpatrick and Ryan, 1991; Searcy, 1982) for serving a mate-choice function: the trait varies among males (McCabe, 1965; Pacejka et al., 1996; the present study) and the quantity of cocoons in a nest is assessable by females before mating (Kendeigh, 1941; Johnson, 1998). However, the third criterion is not met: adding cocoons did not enhance male mating success. Instead, adding cocoons was correlated with increased male time to pairing. Time to pairing did not differ among experimental males not adding cocoons (without-add versus without-control); however, in all other comparisons, males with real or artificial cocoons in their nests took longer to acquire a mate than did males without cocoons. In particular, males from whose nests we removed cocoons (with-remove) obtained mates significantly faster than did males from whose nests we did not remove cocoons (with-control) and to whose nests we added artificial cocoons (with-add). This result is not attributable to removal-treatment males (with-remove) adding more cocoons to their nests after the experimental removal of cocoons, because among males that did add cocoons, the cumulative number of natural cocoons added to each nest did not differ significantly. Likewise, this result is not attributable to temporal differences in the addition of new cocoons, as the time from the initiation of nest building to the highest daily count of cocoons in control (with-control) and addition (with-add) treatment nests (the only treatments in which the number of male-added cocoons was not manipulated) was not significantly different (Eckerle KP, unpublished data). Thus, females did not prefer males adding cocoons to their nests, and males with cocoons in their nests were apparently less attractive to females than were males with cocoons.
Fitness-related benefits of choosing males based on cocoon use
Females did not acquire a direct fitness benefit by selecting or avoiding males based on the presence of cocoons in their nests. Clutch size; number of young fledged; and the mass, condition, and growth rate of nestlings did not differ between nests with and without cocoons. Cocoon use by males also did not reliably predict paternal care, as the male provisioning index was positively correlated with the number of cocoons added per nest in 1996 but not in 1998. Further, adult provisioning rates at nests with and without cocoons did not differ significantly. It is, however, possible that annual variation in breeding conditions could lead to variation in male provisioning behavior. If so, 1996 may have been a poor year, when male provisioning ability was expressed, and 1998 a good year, when male provisioning was not needed to ensure successful reproduction. However, we have no direct data bearing on this issue. Overall, we found no evidence that females benefited by using the presence or absence of cocoons when choosing a mate.
Why do males add cocoons to their nests?
Males that add cocoons to their nests may incur significant fitness-related costs when it takes them longer than other males to acquire a mate. Clutch size, probability of nest success, and number of nestlings produced decline over the breeding season in this population (Drilling and Thompson, 1991; Finke et al., 1987), and are often accompanied by reduced nestling growth rates (Styrsky et al., 1999) and lower nestling mass before nest leaving (Finke et al., 1987; Eckerle KP, unpublished data). Further, a delay in acquiring a first mate decreases the chance that a male will mate polygynously because male house wrens do not advertise for a secondary female until after their primary mate has begun incubation (Johnson and Kermott, 1991). In addition, males breeding later than their neighbors are more likely to suffer reproductive losses by being cuckolded (Johnson et al., 2002). Thus, any delay in male pairing likely produces a fitness-related cost by lowering annual reproductive success in this short-lived songbird.
If not an ornament used for mate attraction, why do males bring cocoons to their nests? Any explanation for the persistence of cocoon use in the study population and its occurrence in widely separated populations over the species' range must identify benefits that offset the costs incurred by males that add cocoons. Thus, the possibility that spiders and other arthropods simply exploit male nest building behavior to gain some benefit (e.g., increased dispersal) seems unlikely. Pacejka et al. (1996) proposed that cocoons are a biological control agent, because in the laboratory spiderlings hatching from cocoons prey on ectoparasitic mites that inhabit house wren nests. However, in a field experiment, mite population sizes in control nests and in nests from which cocoons were removed, sterilized, and replaced did not differ significantly (Pacejka AJ, unpublished data), and experimental reduction in mite populations had no discernible effect on house wren reproductive success (Pacejka et al., 1998). Cocoons could serve as food, but they are not eaten by adults or nestlings, as they rarely disappear from active nests (Eckerle KP, personal observation), and it is unlikely that the tiny spiderlings hatching from these cocoons serve as an important food resource. One untested possibility centers on the well-documented observation that house wrens routinely destroy the eggs of conspecifics (Belles-Isles and Picman, 1986; Finch, 1990). Cocoons might distract predatory wrens when they enter the nest-cavity, reducing the chance that they will find the eggs in the nest cup.
Conclusion
Male house wrens regularly add arthropod cocoons to their nests, but females seem to prefer males that do not add cocoons to their nests. Furthermore, there is no evidence of a fitness-related benefit to females for preferring or avoiding males that use cocoons. The addition of cocoons likely reduces male reproductive success by increasing male time to pairing; therefore, why males add cocoons during nest building is an enigma.
|
ACKNOWLEDGEMENTS |
|---|
We thank the David Davis family, Mr. and Mrs. D. Sears, and the ParkLands Foundation for permission to work on their properties. Field assistance over the course of the study was provided by Robert Dobbs, Lisa Ellis, Colleen Gratton, Alexander Gubin, Alexandra Latham, Eric Sgariglia, and John Styrsky. We thank Drs. Steven Juliano and Scott Sakaluk for their assistance in experimental design and statistical analysis. We also thank Drs. Angelo Capparella, Steven Juliano, Sabine Loew, William Perry, and Scott Sakaluk and two anonymous referees for their thoughtful comments on this manuscript. Dr. Petra Sierwald from the Field Museum of Natural History, Chicago, Illinois, identified the arthropod and spider cocoons. K.P.E. was partially funded during this study by an E.L. Mockford Fellowship, an R.D. Weigel research grant from the Beta Lambda chapter of the Phi Sigma Biological Honor Society, and a GAANN Grant (no. P200A980122–00) from the US Department of Education. Additional financial assistance was provided by the Graduate School and the Department of Biological Sciences, Illinois State University.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Allison PD, 1995. Survival analysis using the SAS system: a practical guide. Cary, North Carolina: SAS Institute Inc.
Alworth T, 1996. An experimental test of the function of sticks in the nests of house wrens. Condor 98:841–844.[CrossRef][Web of Science]
Alworth T, and Scheiber IBR, 2000. Nest building in house wrens (Troglodytes aedon): a reexamination of male and females roles. J Field Ornithol 71:409–414.
Andersson M, 1994. Sexual selection. Princeton, New Jersey: Princeton University Press.
Belles-Isles JC, and Picman J, 1986. House wren nest-destroying behavior. Condor 88:190–193.[CrossRef][Web of Science]
Cavitt JF, and Thompson CF, 1997. Mass loss in breeding house wrens: effects of food supplements. Ecology 78:2512–2523.[CrossRef][Web of Science]
Collias NE, and Collias EC, 1984. Nest building and bird behavior. Princeton, New Jersey: Princeton University Press.
Drilling NE, and Thompson CF, 1988. Natal and breeding dispersal in house wrens (Troglodytes aedon). Auk 105:480–491.[Web of Science]
Drilling NE, and Thompson CF, 1991. Mate switching in multibrooded house wrens. Auk 108:60–70.[Web of Science]
Eckerle KP, 2001. An experimental analysis of the mating preferences of female house wrens (Troglodytes aedon) (PhD dissertation), Normal, Illinois: Illinois State University.
Evans MR, 1997a. The influence of habitat and male morphology on a mate-choice cue: the display nests of wrens. Anim Behav 54:485–491.[CrossRef][Web of Science][Medline]
Evans MR, 1997b. Nest building signals male condition rather than age in wrens. Anim Behav 53:749–755.[CrossRef]
Finch DM, 1990. Effects of predation and competitor interference on nesting success of house wrens and tree swallows. Condor 92:674–687.[CrossRef][Web of Science]
Finke MA, Milinkovich, DJ, and Thompson CF, 1987. Evolution of clutch size: an experimental test in the house wren (Troglodytes aedon). J Anim Ecol 56:99–114.[CrossRef]
Fox GA, 1993. Failure-time analysis: emergence, flowering, survivorship, and other waiting times. In: Design and analysis of ecological experiments (Scheiner SM, Gurevitch J, eds). New York: Chapman and Hall; 253–289.
Grubbauer P, and Hoi H, 1996. Female penduline tits (Remiz pendulinus) choosing high quality nests benefit by decreased incubation effort and increased hatching success. Ecoscience 3:274–279.[Web of Science]
Hansell MH, 2000. Bird nests and construction behavior. Cambridge: Cambridge University Press.
Harper RG, Juliano SA, and Thompson CF, 1992. Hatching asynchrony in the house wren (Troglodytes aedon): a test of the brood-reduction hypothesis. Behav Ecol 3:76–83.
Johnson LS, 1998. House wren (Troglodytes aedon). In: The birds of North America (Poole A, Gill F, eds). Philadelphia: The Birds of North America, Inc; no. 380.
Johnson LS, Hicks BG, and Masters BS, 2002. Increased cuckoldry as a cost of breeding late for male house wrens (Troglodytes aedon). Behav Ecol 13:670–675.
Johnson LS, and Kermott LH, 1991. The functions of song in male house wrens (Troglodytes aedon). Behaviour 116:190–209.[CrossRef][Web of Science]
Johnson LS, and Searcy WA, 1993. Nest site quality, female mate choice, and polygyny in the house wren Troglodytes aedon. Ethology 95:265–277.[Web of Science]
Johnson LS, and Searcy WA, 1996. Female attraction to male song in house wrens (Troglodytes aedon). Behaviour 133:357–366.[CrossRef][Web of Science]
Kendeigh SC, 1941. Territorial and nesting behavior of the house wren. Ill Biol Monogr 18:1–120.
Kendeigh SC, 1952. Parental care and its evolution in birds. Ill Biol Monogr 22:1–356.
Kendeigh SC, 1979. Invertebrate populations of the deciduous forest: fluctuations and relation to weather. Ill Biol Monogr 50:1–107.
Kirkpatrick M, and Ryan MJ, 1991. The evolution of mating preferences and the paradox of the lek. Nature 350:33–38.[CrossRef]
McCabe RA, 1965. Nest construction by house wrens. Condor 67:229–234.[CrossRef]
Moreno J, Soler M, Møller AP, and Linden M, 1994. The function of stone carrying in the black wheatear, Oenanthe leucura. Anim Behav 47:1297–1309.[CrossRef]
Neill AJ, 1990. Territorial behavior of house wrens (MS thesis). Normal, Illinois: Illinois State University.
Pacejka AJ, Gratton CM, and Thompson CF, 1998. Do potentially virulent mites affect house wren (Troglodytes aedon) reproductive success? Ecology 79:1797–1806.[CrossRef][Web of Science]
Pacejka AJ, Santana E, Harper, RG, and Thompson CF, 1996. House wrens Troglodytes aedon and nest-dwelling ectoparasites: mite population growth and feeding patterns. J Avian Biol 27:273–278.
Palomino JJ, Martín-Vivaldi M, Soler M, and Soler JJ, 1998. Functional significance of nest size variation in the rufous bush robin Cercotrichas galactotes. Ardea 86:177–185.[Web of Science]
SAS Institute, 1990. SAS/STAT user's guide, release 6, 4th ed. Cary, North Carolina: SAS Institute.
Savalli UM, 1994. Mate choice in the yellow-shouldered widowbird: correlates of male attractiveness. Behav Ecol Sociobiol 35:227–234.[CrossRef][Web of Science]
Searcy WA, 1982. The evolutionary effects of mate selection. Ann Rev Ecol Syst 13:57–85.[CrossRef][Web of Science]
Skutch AF, 1976. Parent birds and their young. Austin, Texas: University of Texas Press.
Sokal RR, and Rohlf FJ, 1995. Biometry, 3rd ed. New York: W.H. Freeman and Company.
Soler JJ, Møller AP, and Soler M, 1998. Nest building, sexual selection and parental investment. Evol Ecol 12:427–441.[CrossRef]
Styrsky JD, Eckerle KP, and Thompson CF, 1999. Fitness-related consequences of egg mass in nestling house wrens. Proc R Soc Lond B 266:1253–1258.
Svensson L, 1992. Identification guide to European passerines, 4th ed. Tetford: British Trust for Ornithology.
Verner J, and Engelsen GH, 1970. Territories, multiple nest building, and polygyny in the long-billed marsh wren. Auk 87:557–567.[Web of Science]
Williams DD, Ambrose LG, and Browning LN, 1995. Trophic dynamics of two sympatric species of riparian spider (Araneae: Tetragnathidae). Can J Zool 73:1545–1553.[CrossRef]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||