Behavioral Ecology Advance Access originally published online on June 14, 2007
Behavioral Ecology 2007 18(4):689-695; doi:10.1093/beheco/arm031
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Stealing behavior and the maintenance of a visual display in the satin bowerbird
School of Integrative Biology, University of Queensland, St Lucia, Queensland 4072, Australia
Address correspondence to J.M. Wojcieszek, who is now at the Centre for Evolutionary Biology, School of Animal Biology, University of Western Australia, 35 Stirling Highway, Crawley 6009, Australia. E-mail: wojcij01{at}student.uwa.edu.au. J.A. Nicholls is now at the Institute of Evolutionary Biology, University of Edinburgh, Edinburgh EH9 3JT, UK.
Received 29 November 2006; revised 21 February 2007; accepted 19 March 2007.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Honest signals that indicate male quality have been observed in many species and are thought to have evolved to allow males to assess rivals accurately and respond to "cheaters." Females could potentially also use the same honest signals as reliable indicators of male quality. In bowerbirds, the numbers of specific bower decorations may serve as an honest signal of male quality: this study investigates whether decoration stealing among male satin bowerbirds at the Bunya Mountains, Australia, may also involve honest signals. In this study, we aimed to determine 1) predictors for the degree to which individual male satin bowerbirds steal, and are stolen from, and 2) predictors for why some male pairs interact by stealing, whereas other pairs do not. We also assessed how experimentally standardizing the number of decorations on bowers would affect the 1) frequency of stealing, 2) specific interactions among males, and 3) distribution of decorations across bowers. Bower decorations were labeled and tracked through one breeding season. Males that were successful stealers, stole from other successful stealers, had many feathers and bottle tops on their bowers and painted their bower walls often. Male pairs were more likely to interact by stealing if their bowers were in close proximity. Most of the stealing observed was of a reciprocal nature. After we standardized the numbers and types of decorations on a small group of males' bowers, the mean number of daily stealing gains and the total number of males interacting by stealing did not change. In addition, no significant novel stealing interactions were initiated after the manipulation. The average number of all bower decorations and the average number of rosella feathers on a given male's bower prior to the manipulation were proportional to the average numbers for the period after the manipulation. Furthermore, males that originally had better collections of decorations tended to suffer fewer losses due to stealing after the manipulation. Our results suggest that the total number of decorations, the total number of rosella feathers on a male's bower, and possibly stealing behavior, may form part of the basis of an honest signal indicating male quality and therefore might be correlated with mating success.
Key words: bower, bowerbirds, honest signal, Ptilonorhynchidae, Ptilonorhynchus violaceus, sexual display, sexual selection.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Honest signals that indicate male quality have been observed in many species and are thought to have evolved to allow males to assess rivals accurately and respond to "cheaters" (Rohwer 1975
; Johnstone 1995; Berglund 1996). Females could potentially also use the same honest signals as reliable indicators of male quality (Johnstone 1995; Berglund 1996; Borgia and Coleman 2000; Madden 2002). In a study on spotted bowerbirds (Chlamydera maculata), Madden (2002) supplied males with "free" bower decorations and surprisingly not all males took advantage of them. In fact, many males with experimentally enhanced levels of bower decoration suffered from increased instances of bower destruction. This showed that for some males these free bower decorations incurred a substantial cost. Madden therefore proposed that the numbers of specific bower decorations may serve as an honest signal of male quality in this species (Madden 2002). The current study investigates whether decoration stealing among male satin bowerbirds may also involve honest signals.
During the breeding season, male satin bowerbirds (Ptilonorhynchus violaceus) attract females by building elaborate stick structures, known as bowers. These bowers are the focal point of a male's display as it is here that males exhibit the various natural and artificial colorful decorations they have collected, perform courtship dances, and eventually mate with females (Frith CB and Frith DW 2004). Satin bowerbirds are polygynous and males do not care for young (Frith CB and Frith DW 2004). Because a male's only contribution to females is his sperm, females may choose mates based on traits that are indicative of "good genes" (Borgia 1986b; Borgia and Collis 1990). In fact, many aspects of a male's display are correlated with mating success, including bower quality and the numbers and types of decorations (Borgia 1985b, 1986a; Borgia and Gore 1986; Patricelli et al. 2003; Coleman et al. 2004). However, in isolation, such correlations are not necessarily indicative of male quality. Male–male competition is also intense among male satin bowerbirds, as males often steal bower decorations from, and destroy, rival males' bowers during the breeding season (Marshall 1954; Vellenga 1970; Borgia 1985a). Because both male–male competition and female choice focus on the same suite of traits, investigation into stealing interactions among males may clarify whether females are in fact selecting for good genes.
Satin bowerbirds steal bower decorations regularly (Borgia and Mueller 1992; Madden 2002; Frith CB and Frith DW 2004), but few studies have specifically investigated patterns of stealing behavior in this species. Borgia and Gore (1986) tracked the movement of rosella (Platycercus elegans and Platycercus eximius) tail feathers among males in the Wallaby Creek population in New South Wales. Stealing was frequent between males whose bowers were in close proximity and many pairs of males were involved in specific "stealing relationships," most of which involved reciprocal stealing between pairs of males (i.e., stealing back and forth between a pair of males as opposed to a thief stealing from a nonreciprocating victim). In contrast, it appeared that some males did not form stealing relationships at all and it was suggested that such behavior was linked to male dominance levels. Rosella tail feathers and blue plastic bottle tops were the decorations most frequently stolen by satin bowerbird males in the Bunya Mountains (Wojcieszek et al. 2006). It is likely that males preferred these decorations because they reflected higher amounts of ultraviolet light, offered a high degree of contrast to other aspects of the bowerbirds' visual display and therefore provided a "better" display (Wojcieszek et al. 2006). Given that possession of these decorations—feathers in particular—correlates with mating success (Borgia 1985b), and plays a role in competition among males, they may be honest signals of a male's quality.
In this study, we observed one group of males to obtain baseline data on stealing to determine 1) predictors for the degree to which individual male satin bowerbirds steal, and are stolen from, and 2) predictors for why some male pairs interact by stealing, whereas other pairs do not. We used a second group of males in the same population to assess how experimentally standardizing the number of decorations on bowers would affect the 1) frequency of stealing, 2) specific interactions among males, and 3) distribution of decorations across bowers. Given the correlation between female mate choice and the number of some types of bower decorations (Borgia 1985b), determining the contribution made by stealing to the distribution of decorations will help in understanding whether females are assessing an aspect of a male's display that is indicative of a potential mate's quality.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Study site and birds observed
This study was conducted at the Bunya Mountains, approximately 160 km northwest of Brisbane, Australia (26°51'S, 152°50'E). The site was a mixture of cleared paddocks and araucarian simple notophyll vine rain forest (Webb 1968
). Thirty-one male satin bowerbirds were observed in this study. Twenty-four of these males were already individually color banded from previous studies (Robson et al. 2005; Bravery et al. 2006). Two birds were banded during this study using the methods described in Bravery et al. (2006). Although some males were not banded, we can be confident that the same birds were observed throughout this study as bower ownership usually continues for many years (Vellenga 1970). Banding was done under license from the Australian Bird and Bat Banding Scheme. Stealing of bower decorations among males was monitored in 2 separate clusters of bowers (see Figure 1). These clusters were approximately 1.5 km apart, and no stealing was observed between males from different clusters during the study. The 21 bowers in cluster one remained in their natural state throughout the study, whereas the 10 bowers in cluster two were used in an experimental manipulation. This study was approved by the Queensland Parks and Wildlife Service (Scientific Purposes Permit: WISP02093904) and the University of Queensland Animal Experimentation Ethics Committee (reference: ZOO/ENT/355/04/UQ).
|
Marking of decorations and monitoring of stealing
Decorations on all bowers were individually marked, and the stealing of bower decorations was monitored for the bowers in cluster one between 11 September and 1 December 2004. Bowers were visited each day, at the same time and in the same sequence. During this time period, there were 44 pairs of consecutive days when stealing could be inferred. At each visit, it was noted if a bird had gained a decoration that was present on another bower the previous day; this was defined as a stealing gain. Similarly, if a male had lost a decoration to another monitored bower this was defined as a stealing loss. For further details regarding the decoration labeling and subsequent inferring of stealing see Wojcieszek et al. (2006)
. At the end of the season, we determined which male pairs were involved in stealing relationships. Stealing relationships were further categorized into reciprocal and thief–victim relationships. Reciprocal relationships existed where each male of a pair stole at least once from the other male, whereas thief–victim relationships consisted of at least one theft by one male in a pair with no reciprocal stealing by the other male in the pair.
Measuring characteristics of males and their bowers
We measured various characteristics of the 21 males from cluster one and their bowers to determine whether any of these characteristics predicted 1) a male's stealing activity or 2) why some pairs of males interacted by stealing. Most of the males observed had been banded as birds in mature satin blue plumage, thus they could only be assigned a minimum age. Because the exact age could only be determined for 8 out of the 21 males, age was excluded from all further analyses. Morphological data were used in our analyses as they were available for 17 of the 21 males (variables included body weight, head-bill size, wing length, tarsus length, and tail length).
At each of the 21 bowers, 3 sessions of continuous behavioral observation, each of 90 min, were performed over the season, providing 270 min of behavioral data per male. Observation periods commenced 1 h after sunrise and observers sat approximately 10 m from bowers. Start and finish times for the following categories of behaviors were recorded: (A) owner out of sight, (B) owner manipulating sticks, (C) owner manipulating decorations, (D) owner removing objects from bower, (E) owner masticating or painting (bowerbirds regularly masticate various types of leaves and sticks and subsequently "paint" this mixture onto their bower walls; see Bravery et al. 2006), (F) owner perching quietly on the ground or in a tree within observer's sight, (G) owner vocalizing, (H) marauding male present at bower and/or chased by owner, (I) owner displaying with no other birds observing, (J) owner displaying with birds observing (juvenile males and females were distinguished, if possible, based on bill and plumage differences; Vellenga 1970), and (K) owner on bower platform. We then determined the total time spent in each of the above categories of behavior. The total times spent "on bower" (total of B + C + D + E + I + J + K) and spent on "bower maintenance" (total of B + C + D + E) were also determined. Tallies were made of 1) the number of copulations received during the observation period, 2) the numbers of each type of vocalization performed, and instances of visitation by 3) juvenile males and 4) females.
Each bower was mapped using a "Garmin eTrex" Global Positioning System. The locations of all bowers were then mapped (see Figure 1) and the distances between all pairs of bowers calculated. Instances of complete bower destruction were noted during the daily decoration surveys. Complete bower destruction was defined as little or no evidence of the bower avenue remaining (modified from Hunter and Dwyer 1997). Once each month the following 4 variables were subjectively ranked at each bower: 1) symmetry of walls, 2) stick size, 3) stick density, and 4) overall quality of construction (Borgia 1985b). Each of these 4 aspects was ranked on a scale of 1 to 5, with scores of 5 given to bowers with symmetrical walls, sticks of similar size, high stick density, and of an overall "neat" appearance (modified from Borgia 1985b). Bower quality scores were then determined for each bower by obtaining the means of these 4 variables. Digital photographs of each bower were also taken and 2 naive observers, with no knowledge of the males studied, were asked to assign bower quality to ensure against bias (in accordance with Bravery et al. 2006). These values were correlated with those from the field (P < 0.005). The field scores were therefore used to generate a mean bower quality score for each male.
Statistical methods and model selection
After monitoring stealing for 44 days, each male's total numbers of stealing gains and losses were determined. A correlation matrix was made using all the explanatory variables and if 2 variables were correlated at r 
0.70, one of these variables was removed from all further analyses. Only 2 sets of variables were intercorrelated; the proportion of time males spent performing bower maintenance was positively correlated with the time spent manipulating sticks and the proportion of time males were on their bowers was negatively correlated with the time they were out of sight. The proportion of time that males were out of sight was not used in further analyses; the time spent on the bower platform was selected as it provided a more reliable measure of males' attendance at their bowers. This is because there would have been times when a male was within the vicinity of his bower, but obscured from the view of observers. The time males spent manipulating sticks was removed from further analyses in favor of the bower maintenance measure, as this was a more complete measure of a male's activity while on his bower.
All data were analyzed using the statistical program "R" (version 1.9.1). To determine predictors for the degree to which individual male satin bowerbirds steal, and are stolen from, correlation tests were performed between all the uncorrelated explanatory variables and 2 response variables—the total numbers of stealing gains and stealing losses of each male. If an explanatory variable was not correlated with either of the response variables, it was not considered for any further analysis. All possible combinations of the variables that were significantly correlated with the stealing success or stealing losses of males were then examined using generalized linear models. Optimal models were selected using Akaike's Information Criterion corrected for small sample sizes (AICC). The AIC "method" allows one to select the model with the best predictive power from a list of alternative models and the degree to which a given model "fits" the data can be reported using the "Akaike weight." The closer the Akaike weight is to 1.0 the better the predictive power of the model (see Burnham and Anderson (2002) for more details regarding the AIC method). Stealing events were count data, thus Poisson error distributions were specified when modeling both stealing gains and stealing losses. The "offset" function in R was utilized in these models to account for the fact that individual males started building and attending bowers at different times during the mating season, thus were not stealing and being stolen from for the same lengths of time.
The same explanatory variables were used to determine why some male pairs interacted by stealing, whereas other pairs did not. Each possible male pair was assigned a single value for each of the explanatory variables. This variable was the difference between the measures determined for each bird in a pair for that variable. We included as another explanatory variable the distances between the bowers of each pair of males. Generalized linear models with binomial error distributions were then created (the binary response being scored "yes" if a male pair interacted by stealing and "no" if they did not). In these models, the number of days in which both males of a pair were simultaneously attending bowers was used as the offset.
Experimental manipulation
The decorations on the 10 bowers in cluster two were labeled and their movements recorded as described previously. After monitoring stealing for 21 days (decorations surveyed 11 September–16 October 2004), the decorations on all 10 bowers were experimentally manipulated. This involved removing each male's collection of blue decorations and replacing these with identical collections of decorations for all 10 bowers. All the replaced items were "new" to the birds; that is, they were not sourced from bowers, although they were identical to decorations that had been seen on bowers in the Bunya Mountains satin bowerbird population. Each new collection included 1 blue crimson rosella tail feather, 2 blue straws, 2 blue plastic bottle tops, 2 blue bottle rings, and 4 blue pegs. These 11 objects were placed at the northern end of each bower platform—where satin bowerbirds' objects are concentrated (Borgia 1995b). A total of 604 decorations were removed from the 10 bowers during this manipulation and these were replaced by a total of 110 new decorations. Because each of the 10 males now held identical collections, they each held 10% of the total decoration pool available to the group. This manipulation took place on one evening (16 October 2004) between 1900 and 2100 h. Thus, each male owner arrived at his bower the following morning to a new "standardized" set of bower decorations. Stealing events were quantified for 21 days after the manipulation (decorations surveyed 17 October–18 November 2004) and at the end of the experiment all the original decorations were returned to their source bowers.
We compared the distribution of blue objects across bowers for 21 days preceding, and 21 days after, the manipulation. We examined whether the stealing interactions among males resulted in blue objects, in particular the favored feathers and bottle tops, being distributed among the bowers in the same relative proportions as before the manipulation. Rates of stealing losses and gains and stealing relationships were also compared for specific males before and after the manipulation. We determined the proportion of the total decoration pool that each male held on the day prior to the manipulation. At the manipulation, males were given decoration collections that equated to possessing 10% of the total pool of decorations. If a male's new decoration collection meant he then possessed more than the proportion of decorations he held prior to the manipulation he was termed a "cheater." For example, one male possessed only 4% of the total decoration pool prior to the manipulation; he was therefore labeled a cheater once his decoration collection was experimentally enhanced to 10%. Three birds were classified as noncheaters and 7 as cheaters.
The decorations of nonmanipulated males around the periphery of the 10 manipulated males' bowers were monitored during the study and no labeled decorations were ever found on any of these peripheral bowers. It was therefore assumed that the main stealing interactions were confined to the 10 males we included in our experimental manipulation.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
What factors were correlated with individual males' stealing gains and losses?
Males that were successful stealers were also frequently stolen from (Figure 2), spent small proportions of their time on their bower platforms and maintaining their bowers, painted their bowers frequently (Figure 3), and had high numbers of rosella feathers and blue plastic bottle tops on their bowers (statistics reported in Table 1). The best of the generalized linear models included all 6 of the above variables (Akaike weight = 0.99). The same explanatory variables were the only variables found to be significantly correlated with males' stealing losses (Table 1) and, again, the best generalized linear model explaining stealing losses also included all the above variables (Akaike weight = 0.84). There was no correlation between stealing success and average bower quality (Pearson's product-moment correlation test, r = 0.19, n = 21, degrees of freedom [df] = 19, P = 0.39). Thus, although the more active stealers spent less time on their bowers they did not have bowers of poorer quality. None of the other variables (age, morphological variables, and other behavioral variables) were correlated with either response variable.
|
|
|
What predicted whether male pairs interacted by stealing?
The most likely generalized linear model of the factors that determined whether birds interacted contained only the distance between bowers of males in pairs (Akaike weight = 0.99) (Figure 4). Only 2 of the 21 birds did not steal from their nearest neighboring bower and one of these birds did not steal all season. Most of the stealing observed was between pairs of males that were involved in reciprocal stealing relationships, whereas thief–victim relationships accounted for only a small proportion of the total number of stealing events (mean reciprocal stealing events ± standard error [SE] = 14.53 ± 0.21; mean thief–victim stealing events ± SE = 2.4 ± 0.07; t-test: t = 4.76, df = 14.71, P < 0.001).
|
How did the manipulation affect the redistribution of bower decorations?
The average total number of decorations that males had for the 21 days preceding the manipulation was related to the average total number of decorations they had during the 21 days that followed the manipulation (Pearson's product-moment correlation test, r = 0.72, n = 10, df = 8, P = 0.019) (Figure 5). Similarly, the average number of feathers that males had for the 21 days preceding the manipulation was related to the average number they possessed during the 21 days that followed the manipulation (Pearson's product-moment correlation test, r = 0.80, n = 10, df = 8, P = 0.006) (Figure 6). The average number of bottle tops males had prior to the manipulation was not significantly related to the average number they had after the manipulation (Pearson's product-moment correlation test, r = 0.55, n = 10, df = 8, P = 0.10).
|
|
Did "cheaters" with experimentally enhanced levels of bower decorations have more decorations stolen after the manipulation than before?
We compared the total numbers of losses due to stealing experienced by cheaters during a 21-day period prior to the manipulation with the total numbers experienced during a 21-day period after the manipulation. On average, the noncheater males experienced fewer losses to stealing postmanipulation and this change approached significance (paired t-test: t = 3.71, df = 2, P = 0.07), whereas the levels of stealing experienced by cheater males did not change postmanipulation (paired t-test: t = –0.61, df = 6, P = 0.56). In addition, the difference in the levels of stealing experienced by cheaters and noncheaters approached significance (t-test: t = 1.90, df = 7.41, P = 0.097).
How were instances of stealing affected by the manipulation?
The mean number of daily stealing events among the 10 males in cluster two declined postmanipulation (premanipulation: X ± SE = 3.95 ± 0.63 stealing events, n = 10; postmanipulation: X ± SE = 2.73 ± 0.27 stealing events, n = 10), although this drop was not statistically significant (Wilcoxon signed ranks test: V = 109.5, n = 21, P = 0.12). There was no difference in the change in stealing rate between cheater males (change of –0.57 ± 1.02 stealing gains per day) and noncheater males (change of 0.33 ± 0.67 stealing gains per day; t-test: t = 0.74, df = 7.9, P = 0.48). After the manipulation, many of the birds stopped interacting with previous "stealing partners." Where male pairs were still interacting postmanipulation it was with the same partners as before the manipulation. In fact, only 2 novel stealing interactions were observed after the bowers were manipulated and both of these involved only one stealing event that was never reciprocated.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Rosella tail feathers and blue plastic bottle tops are the most frequently stolen decorations in this population (Wojcieszek et al. 2006
). Based on the current study, we propose that feathers, along with the total number of decorations a male possesses, are agents of an honest signaling system in satin bowerbirds mediated by stealing interactions.
The males that were the most active stealers were those most frequently stolen from (a phenomenon also observed by Borgia and Gore 1986 and Madden et al. 2004), had high numbers of crimson rosella feathers and blue plastic bottle tops, and spent less time maintaining their bowers than other males, but painted their bower walls more frequently. The fact that the most active stealers painted their bowers more often than less active stealers is interesting in the light of a recent study (Robson et al. 2005) involving many of the males examined in this present study, in which painting rate was found to be the best predictor of mating success among males with high-quality bowers. Because painting effort has been shown to be highly correlated with both mating success and stealing behavior in our study population of satin bowerbirds, it is possible that the most active stealers were the males that gained the most matings. Hence, stealing and painting rate may both be reliable indicators of male quality. The signal redundancy hypothesis would suggest that some aspects of a male's multicomponent display provide redundant information to females (Candolin 2003), as has been found for some other males' male traits in satin bowerbirds (Coleman et al. 2004, Robson et al. 2005, Bravery and Goldizen 2007). Thus, possession of frequently stolen decorations and regular painting effort may provide similar information regarding male quality to females.
Our study showed that the most active stealers spent less time on their bowers and, as would be expected, rival males took advantage of these unattended bowers by raiding them. A recent modeling approach has also made the same conclusions; males cannot both leave their bowers frequently to steal decorations and constantly repel thieves (see Morrell and Kokko 2004). In addition, actively stealing males almost certainly displayed more feathers and bottle tops because of their extra efforts in stealing, and displaying these sought-after items may have then made these bowers stealing targets. Bower quality was expected to increase with the time spent on bowers (Morrell and Kokko 2004). However, although the less active stealers spent more time on their bowers, their bowers were not of higher quality. This could, of course, be a consequence of the subjective nature of bower quality assignment, but this is unlikely as naive observers agreed with our field scores of bower quality. Coupled with the fact that the less active stealers displayed few of the popular decorations over the season and painted their bowers infrequently, it seems more likely that they were inexperienced and/or low-ranking males with displays of overall lower quality.
Whether a pair of males interacted by stealing depended exclusively on the proximity of their bowers. In fact, all but 2 of the males observed stole at least once from their nearest neighbors. Although the home ranges of adult male satin bowerbirds are yet to be established, those of juvenile males range from 3.59 to 28.67 ha (Maxwell et al. 2004), thus it is conceivable that adult males would have been able to steal from any other adult males in the observed pool of bowers, rather than being limited to stealing from neighboring bowers. Studies have consistently shown, however, that distances between bowers are important in determining male–male interactions in bowerbird species (Borgia and Gore 1986; Diamond 1986; Borgia and Mueller 1992; Borgia 1995a). Most of the observed stealing occurred between reciprocally stealing pairs and although most stealing occurred between near neighbors, reciprocal relationships were not always between nearest neighbors (Wojcieszek J, unpublished data), as was also found by Borgia and Gore (1986). When males acquire bower decorations via theft, they not only improve their own decoration collection but also diminish that of a rival (Morrel and Kokko 2004). By targeting near neighbors, males need not leave their bowers unattended for extended periods of time and they have the potential to outcompete their most geographically immediate rivals. It is also possible that the stealing relationships observed are operating under a "retaliatory" or tit-for-tat strategy (as has been suggested previously by Borgia and Gore 1986; Pruett-Jones MA and Pruett-Jones M 1994). Tit-for-tat behavior involves a pair of "players" interacting via some activity (stealing in this instance) and the tit-for-tat pattern emerges when each member of the pair reacts by imitating the previous actions of the opponent (Axelrod 1984). Future work could investigate whether stealing between reciprocal males operates under a tit-for-tat strategy via manipulating the bower decorations (particularly the most popular decorations—crimson rosella tail feathers and blue plastic bottle tops) of individual males to determine whether this instigates thievery.
The present study provides some experimental evidence for honest signaling among male satin bowerbirds in which male quality is communicated, at least in part, via the numbers of bower decorations. The total numbers of decorations and the numbers of feathers on bowers before the manipulation were correlated with their respective numbers after the manipulation, suggesting the distribution of objects across bowers is a true reflection of the respective males' abilities to steal and then maintain a collection of objects. The experimental manipulation suggested that noncheater males may have been males of higher quality as they could better defend their collection of objects, although this result only approached statistical significance, perhaps due to the small sample size used. These findings suggest that the numbers of blue objects on bowers, and primarily the number of feathers, could be an indirect way of inferring a male's relative quality among a group (also suggested Borgia 1985b), with the males having the greatest numbers of feathers and other objects being those of highest quality. The fact that feather possession is one of the best predictors of mating success (Borgia 1985b) supports the idea that these decorations could form part of the basis of an "honest signaling" system that females may also be monitoring in order to choose a high-quality mate (Coleman et al. 2004). It should also be noted that these decorations may impose a cost to owners, not necessarily because these objects are significantly rare in the surrounding environment and therefore hard to acquire (Madden and Balmford 2004), but rather because these items are in high demand and must be heavily guarded (see Wojcieszek et al. (2006) for discussion of why certain bower decorations are in high demand).
Hunter and Dwyer (1997) suggested that when bower decorations were scarce within a population, individual items became more valuable and overall stealing effort increased. Given these findings, we expected that the marked decrease in the total number of decorations available across bowers as a result of the manipulation would cause an increase in the amount of stealing and initiate a number of new interactions between male pairs. In contrast, we found no significant change in the number of daily stealing events, no new stealing interactions, and in fact observed some reduction in the number of stealing partners. However, although the difference only approached statistical significance, our results are consistent with the idea that males who initially had better displays suffered reduced rates of stealing postmanipulation. If we assume the decrease in the total number of decorations available did increase the value of individual decorations, and hence their attractiveness for being stolen, then this response is consistent with these males being of higher quality, as only they were capable of defending their now more valuable collection of decorations.
In summary, we have shown that there are predictors for the degree to which a male steals as the most active stealers displayed high numbers of feathers and bottle tops on their bowers, painted their bower walls frequently and spent low proportions of their time on their bower platforms and maintaining their bowers. Stealing was more common between males with bowers in close proximity and thieving males mostly stole from other males of high quality. Stealing interactions returned the numbers of decorations to their original proportional distribution over a set of experimentally manipulated bowers, with males that originally had better collections of decorations being the only ones capable of maintaining their manipulated collection of decorations. This study provides evidence for an honest signaling system in satin bowerbirds, similar to that described for spotted bowerbirds. Because females appear to select mates based in part on the same traits that are altered by stealing among males, the stealing interactions among males probably provide females with an honest indication of the relative qualities of males in a population.
|
ACKNOWLEDGEMENTS |
|---|
We thank the Stirling and Foxton families for allowing us to work on their properties, the Bjelke-Peterson family for allowing us to rent "Possum Lodge," and all the volunteer field assistants for their time and assistance. We thank David Pavlacky for statistical help and the members of the Behavioural Ecology Research Group at the University of Queensland for all their advice. Funding for this study was provided by the University of Queensland.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Axelrod R. The evolution of cooperation. (1984) New York: Basic Books Inc.
Berglund A. Armaments and ornaments: an evolutionary explanation of traits of dual utility. Biol J Linnean Soc (1996) 58:385–399.[CrossRef]
Borgia G. Bower destruction and sexual competition in the satin bowerbird (Ptilonorhynchus violaceus). Behav Ecol Sociobiol (1985a) 18:91–100.[CrossRef][ISI]
Borgia G. Bower quality, number of decorations and mating success of male satin bowerbirds (Ptilonorhynchus violaceus): an experimental analysis. Anim Behav (1985b) 33:266–271.[CrossRef][ISI]
Borgia G. Satin bowerbird parasites: a test of the bright male hypothesis. Behav Ecol Sociobiol (1986a) 19:355–358.[CrossRef][ISI]
Borgia G. Sexual selection in bowerbirds. Sci Am (1986b) 254:70–79.
Borgia G. Threat reduction as a cause of differences in bower architecture, bower decoration and male display in two closely related bowerbirds Chlamydera nuchalis andC. maculata. In: Emu (1995a) 95:1–12.[ISI]
Borgia G. Why do bowerbirds build bowers? Am Sci (1995b) 83:542–547.[ISI]
Borgia G, Coleman SW. Co-option of male courtship signals from aggressive display in bowerbirds. Proc R Soc Lond B (2000) 267:1735–1740.[Medline]
Borgia G, Collis K. Parasites and bright male plumage in the satin bowerbird (Ptilonorhynchus violaceus). Am Zool (1990) 30:279–285.[ISI]
Borgia G, Gore MA. Feather stealing in the satin bowerbird (Ptilonorhynchus violaceus): male competition and the quality of display. Anim Behav (1986) 34:727–738.[CrossRef][ISI]
Borgia G, Mueller U. Bower destruction, decoration stealing and female choice in the spotted bowerbird Chlamydera maculata. Emu (1992) 92:11–18.[ISI]
Bravery BD, Goldizen AW. Male satin bowerbirds (Ptilonorhynchus violaceus) compensate for sexual signal loss by enhancing multiple display features. Naturwissenschaften (2007) 94:473–476.[CrossRef][Medline]
Bravery BD, Nicholls JA, Goldizen AW. Patterns of painting in satin bowerbirds Ptilonorhynchus violaceus and males' responses to changes in their paint. J Avian Biol (2006) 37:77–83.[CrossRef]
Burnham KP, Anderson DR. Model selection and multimodel inference: a practical information-theoretic approach. (2002) New York: Springer Verlag.
Candolin U. The use of multiple cues in mate choice. Biol Rev (2003) 78:575–595.[Medline]
Coleman SW, Patricelli GL, Borgia G. Variable female preferences drive complex male displays. Nature (2004) 428:742–746.[CrossRef][Medline]
Diamond JM. Biology of birds of paradise and bowerbirds. Annu Rev Ecol Syst (1986) 17:17–37.[CrossRef][ISI]
Frith CB, Frith DW. The Bowerbirds—Ptilonorhynchidae. (2004) Oxford: Oxford University Press.
Hunter CP, Dwyer PD. The value of objects to satin bowerbirds Ptilonorhynchus violaceus. Emu (1997) 97:200–206.[ISI]
Johnstone RA. Sexual selection, honest advertisement and the handicap principle: reviewing the evidence. Biol Rev (1995) 70:1–65.[Medline]
Madden JR. Bower decorations attract females but provoke other male spotted bowerbirds: bower owners resolve this trade-off. Proc R Soc Lond B (2002) 269:1347–1351.[Medline]
Madden JR, Balmford A. Spotted bowerbirds Chlamydera maculata do not prefer rare or costly bower decorations. Behav Ecol Sociobiol (2004) 55:589–595.[CrossRef][ISI]
Madden JR, Lowe TJ, Fuller HV, Coe RL, Dasmahaptara KK, Amos W, Jury F. Neighbouring male spotted bowerbirds are not related, but do maraud each other. Anim Behav (2004) 68:751–758.[CrossRef][ISI]
Marshall AJ. Bower-birds. (1954) Oxford: Clarendon Press.
Maxwell PS, Robson TE, Goldizen AW. Home-range sizes and bower visitation patterns of immature male Satin Bowerbirds (Ptilonorhynchus violaceus). Emu (2004) 104:75–77.[CrossRef][ISI]
Morrell LJ, Kokko H. Can too strong female choice deteriorate male ornamentation? Proc R Soc Lond B (2004) 271:1597–1604.[Medline]
Patricelli G, Uy JAC, Borgia G. Multiple male traits interact: attractive bower decorations facilitate attractive behavioural displays in satin bowerbirds. Proc R Soc Lond B (2003) 270:2389–2395.[Medline]
Pruett-Jones MA, Pruett-Jones M. Sexual competition and courtship disruptions: why do male bowerbirds destroy each others bowers? Anim Behav (1994) 47:607–620.[CrossRef][ISI]
Robson TE, Goldizen AW, Green DJ. The multiple signals assessed by female satin bowerbirds: could they be used to narrow down females' choice of mates? Biol Lett (2005) 1:1264–1267.
Rohwer SA. The social significance of avian winter plumage variability. Evolution (1975) 29:593–610.[CrossRef][ISI]
Vellenga R. Behaviour of the male satin bowerbird at the bower. Aust Bird Bander (1970) 8:3–11.
Webb LJ. Environmental relationships of the structural types of Australian rain forest vegetation. Ecology (1968) 49:296–311.[CrossRef][ISI]
Wojcieszek JM, Nicholls JA, Marshall NJ, Goldizen AW. The stealing of bower decorations among male satin bowerbirds (Ptilonorhynchus violaceus): why are some decorations more popular than others? Emu (2006) 106:175–180.[CrossRef][ISI]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||