Behavioral Ecology Vol. 15 No. 1: 109-119
© 2004 International Society for Behavioral Ecology
Phylogenetic analysis of life-history adaptations in parasitic cowbirds
Departamento de Ecología y Comportamiento Animal, Instituto de Ecología A.C., Km 2.5 antigua carretera a Coatepec No 351, Congregación "El Haya", Xalapa, Veracruz 91070, México
Address correspondence to M. E. Mermoz, who is now at the Laboratorio de Ecología y Comportamiento, Departamento de Ecología, Genética y Evolución, Facultad de Ciencias Exactas y Naturales, Universidad de Buenos Aires, Pabellón II 4° Piso, (1428) Ciudad Universitaria, Buenos Aires, Argentina. E-mail: mermoz{at}bg.fcen.uba.ar.
Received 21 March 2002; revised 27 January 2003; accepted 27 February 2003.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Parasitic cowbirds lay eggs in the nests of other species and dupe them into caring for their young. Unlike other brood parasites, cowbirds have not developed egg mimicry or bizarre chick morphology. However, most of them parasitize a large number of hosts. Several features of cowbirds have been proposed as more general adaptations to brood parasitism. In this study, we used a recent molecular phylogeny as a historical framework to test the possible adaptations of the parasitic cowbird, including egg size, eggshell thickness and energy content of the eggs, length of the incubation period, and growth pattern of cowbird nestlings. We used a recently developed extension of independent contrasts to test whether the five cowbird species deviate from general allometric equations. We generated prediction intervals for a nonparasite that evolved in the place of the cowbirds. By using these prediction intervals, we found that parasitic cowbirds had not reduced weight or energy content of their eggs, nor their incubation period over evolutionary time. Cowbird chicks and those of nonparasitic relatives had similar growth pattern. The only characteristic that separated parasitic cowbirds from their nonparasitic relatives was an increase in eggshell thickness. All these findings were robust and resisted the use of three models of character evolution. The fact that most traits exhibited by cowbirds were inherited from a nonparasitic ancestor does not rule out that they are advantageous for parasitism. Future research should focus on such traits of cowbird relatives and on how these traits preadapted a particular lineage to become parasites.
Key words: adaptation, cowbirds, eggshell, egg size, Icteridae, incubation period, independent contrasts, nestling development, phylogeny.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Avian brood parasitism is a breeding strategy used by almost 1% of the extant species. Parasitic females lay eggs in nest of other species, the hosts, which incubate the eggs and perform all parental care needed for the normal development of parasitic chicks. In this relationship, the brood parasite depends entirely on the host for successful reproduction, while the host's reproductive output is lowered by the parasite (Payne, 1977
; Rothstein, 1990). Therefore, we expect that natural selection will favor the evolution of adaptations in the brood parasites and the evolution of defense mechanisms among their host species (Payne, 1997a; Rothstein, 1990). Avian brood parasitism has evolved seven times independently, including three times among the cuckoos (Cuculinae and Neomorphininae) and once in the honeyguides (Indicatoridae), the old-world finches (Ploceidae), the parasitic cowbirds (Icteridae), and the black-headed duck (Heteronetta atricapilla) (Aragón et al., 1999; Lanyon, 1992; Sorenson and Payne, 1998).
The brood parasite–host system is ideal for the study of adaptation (Redondo, 1993; Rothstein, 1990). Common cuckoo females are highly territorial, defending access to host nests from other females (Davies, 2000). Egg mimicry of parasitic eggs has been described in many Cuculinae–host systems whose parasitic females select a host with better egg matching (Brooker and Brooker, 1990; Davies and Brooke, 1989). In some brood parasites (Indicatoridae and Neomorphininae), the hatchlings have a beak with a raptorial hook that they use to stab host nestlings (Friedmann, 1955; Morton and Farabaugh, 1979). In the majority of the Cuculinae species, the chick at hatching evicts all host eggs and nestlings by toppling them over onto their backs and pushing them out (Davies, 2000; Payne, 1997b). Therefore, parasite chicks are reared alone, monopolizing all parental care from their host parents. Cuckoo eggs have shorter incubation periods because of their relative small size and rapid embryo development in the female's oviduct, so they usually hatch first (Payne, 1973; Vernon, 1970). Among the parasites that do not harm host eggs or nestlings, most Vidua chicks have special mouth markers that resemble host's chicks, increasing the parent's acceptance (Friedmann, 1960; Payne et al., 2000).
The five species of parasitic cowbird lack most of those specialized features found in other brood parasites. Cowbird females do not defend territories to avoiding intraspecific competition (Dufty, 1982; Fleischer, 1985; Fleischer and Smith, 1992; Fraga, 1998; Mermoz and Reboreda, 1999; but see Alderson et al., 1999). Similarly, cowbird females do not avoid hosts that reject their eggs (Mermoz and Reboreda, 1999; Scott, 1977) or those that rarely could rear their chicks successfully (Elliot, 1978; Middleton, 1991; Scott and Lemon, 1996). Egg or chick mimicry is absent in at least four of the five species (Fraga, 1998; Rothstein, 1990; Smith, 1968). Moreover, except for accidental reports (Dearborn, 1996), cowbird chicks lack any aggressive behavior toward host eggs or chicks. Consequently, cowbird chicks usually have to compete with host chicks for parental care. Furthermore, three species, Molothrus aeneus, M. bonariensis, and M. ater, are extreme generalists in host use, ranging from 80 to 200 species (Friedmann and Kiff, 1985). In addition, it has been suggested that the extremely fecund cowbird females (Davis, 1942; Kattan, 1993; Scott and Ankney, 1983), lay "cheap" eggs (i.e., low energy content and relative small size; Kattan, 1995; Strausberger, 1998a; but see Rahn et al., 1988). Cowbird eggs have thicker eggshells than nonparasitic Icterids (Rahn et al., 1988; Spaw and Rohwer, 1987). The low energy content and/or small size of the eggs allows cowbird chicks to hatch before host chicks do (Kattan, 1995; Strausberger, 1998a). Finally, cowbird chicks seem to have special growth traits that enable them to avoid competitive exclusion by host chicks (Kattan, 1995, 1996; Ortega and Cruz, 1992).
The evolutionary definition of adaptation states that a true adaptation is a feature that was built by natural selection for its current role. Likewise, a feature that today is advantageous but was originally built for a different role is known as an "exaptation" (Gould and Vrba, 1982; Reeve and Sherman, 1993). Therefore, putative adaptations such as those mentioned above must be tested taking into account their evolutionary history—that is, not only the species of interest but also its sister taxa (Brooks and McLennan, 1991; Coddington, 1988; Harvey and Pagel, 1991). If proposed life-history characteristics of parasitic cowbirds are real adaptations, they must be present only in parasitic species and must have evolved in the context of brood parasitism. However, all life-history adaptations suggested for the parasitic cowbirds were proposed before a resolved phylogeny for the group was available (Briskie and Sealy, 1990; Kattan, 1995, 1996; Ortega and Cruz, 1992; Rahn et al., 1988; Scott and Ankney, 1983; Spaw and Rohwer, 1987; Strausberger, 1998a; Weatherhead, 1991). Furthermore, in some instances authors compared one of the extremely generalist cowbirds with a particular host (Kattan, 1995, 1996; Ortega and Cruz, 1992; Strausberger, 1998a). Performing two-species comparisons to test adaptation has multiple logical and statistical problems (Garland and Adolph, 1994). Those problems increase when hosts turn out to be a very distantly related species (Kattan, 1995, 1996; Ortega and Cruz, 1992; see Lanyon and Omland, 1999).
In this study, we used a recent molecular phylogeny (Johnson and Lanyon, 1999) as a historical framework to test the possible adaptations of the parasitic cowbird, including egg size, eggshell thickness and energy content of the eggs, length of the incubation period, and growth pattern of cowbird nestlings. We used a recently developed extension of a well-known comparative method, the phylogenetic independent contrasts (Felsenstein, 1985). This extension is used to test whether a particular species deviates from a previously established allometric equation (Garland and Ives, 2000; Garland et al., 1999). We extended the analysis to the five cowbird species that conform a monophyletic group in the phylogeny (Johnson and Lanyon, 1999). Only if values of cowbirds differ from those predicted by the allometric regression generated with data for nonparasitic relatives do the proposed adaptations to brood parasitism hold (Garland and Ives, 2000).
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Characters of the species and allometric functions
All information on characters of species was obtained from the literature and unpublished sources (Appendix). Egg weight can be described as a positive allometric function of female weight (Ar et al., 1975
; Rahn and Ar, 1975). Similarly, eggshell thickness is a positive allometic function of egg weight (Ar et al., 1974). Energy content of the egg also is a positive allometric function of egg weight (Vleck and Vleck, 1987). However, data on energy content of the egg is only available for one nonparasitic species and two parasitic species of cowbirds (Kattan, 1995; Vleck and Vleck, 1987). Kattan (1995) suggested that the low energy content of parasitic cowbird eggs implies that hatchlings are low weight. Therefore, we regressed the weight of nestling on the hatching day (as an estimate of energy content of the egg) to egg weight.
Incubation period is also an allometric function of egg weight (Rahn and Ar, 1974; Vleck and Vleck, 1987), and it has been proposed that parasitic cowbirds should have shorter incubation periods than those predicted under the general equation for passerines (Kattan, 1995; Strausberger, 1998a). Other authors have stated that incubation in cowbirds only differs from other passerines in the minimum reported (Briskie and Sealy, 1990). Therefore, we used data on the average and the minimum incubation period recorded for each species. Data for parasitic species in some instances was obtained via artificial incubation (Kattan, 1995; Strausberger, 1998a), but this information is not available for most nonparasites. To avoid a possible bias, we performed the analysis using incubation length estimated in the field for all species. Relating to nestling development, Ricklefs (1968) found an inverse relationship of the average growth rate (K) and the asymptotic weight reached in the nest (A). It has been suggested that parasitic cowbirds should have higher growth rates than those predicted under a general equation for altricial birds (Kattan, 1996; Ortega and Cruz, 1992). Many species of the group have marked sexual dimorphism in size that can be detected in the nestlings (Teather and Weatherhead, 1994). Therefore, we used data of females when possible. To minimize errors, we used data on the average growth rate and the asymptotic weight reached in the nest estimated using the methodology of Ricklefs (1967). If an author did not use this approximation, we used the original data to estimate both parameters. Kattan (1996) suggested that the parasitic Molothrus bonariensis has accelerated its development by leaving the nest with a lower proportion of the adult weight. Therefore, we regressed the asymptotic weight reached in the nest to female body weight. We increased the number of species included in that analysis using asymptotic weight data as estimated by Ricklefs' (1967) methodology and those reported as weight of the nestlings before leaving the nest (Appendix).
Authors that suggested adaptations to brood parasitism worked with different subspecies of cowbirds. Specifically, in Molothrus bonariensis, seven subspecies have been described whose adult female weights vary from 31.9 g in M. b. minimus (Wiley, 1988) to 54.6 g in M. b. cabaniisi (Kattan, 1995). Therefore, we tried to include data of more than one subspecies of M. bonariensis.
We applied the logarithmic transformation to all variables before the analysis.
Phylogenetic hypothesis
Parasitic cowbirds were included in one of the five natural groups within the Icteridae: the "grackles and allies" (Johnson and Lanyon, 1999; Lanyon and Omland, 1999). The grackles and allies' phylogeny was based on the sequences of cytochrome-b and ND2 mitochondrial genes and was completely resolved with rather good support (Johnson and Lanyon, 1999). As an estimation of the evolutionary divergence between taxa (tip or node species), we used the number of substitutions in both sequences with the character weights given by Johnson and Lanyon. This information was not available in Johnson and Lanyon (1999). Therefore, to obtain branch length data, we had to access the sequences of both genes for all species in the GenBank, and then we repeated the author's analysis with PAUP 3.1 (Swofford, 1993). When performing our comparative analysis, we only replaced Dives warsaweski by the most known congener, D. dives (Figure 1).
|
Comparative analysis: regression lines with prediction intervals
Species cannot be considered independent evolutionary units because they are interconnected to each other by common ancestry. Consequently, comparative data are not statistically independent and therefore violate one of the most basic assumptions of statistical procedures (Felsenstein, 1985
; Martins and Hansen, 1996). Phylogenetic independent contrasts (PIC) are calculated as differences in trait values between adjacent pairs of nodes or terminal taxa in a phylogenetic tree. Because contrasts do not share the same branches of a tree, they are statistically independent samples of evolutionary change within a lineage. We did not use the PIC as originally stated (Felsenstein, 1985), but used an extension of that method proposed by Garland and co-workers. In this extension, PIC are used to generate an allometric regression line with its prediction intervals including the Y intercept of the original data (Garland and Ives, 2000). We generated the prediction interval for a nonparasitic grackle and allies that evolved in the place of the cowbirds. If observed values of parasitic cowbirds differ from those predicted by the regression, the adaptations to brood parasitism may hold (Garland and Ives, 2000). Otherwise, the existence of exaptations can be claimed. The assumptions of PIC are (1) the topology of the phylogenetic tree is dichotomous, fully resolved, and accurate and (2) branch length is a good estimator of the expected variance in character evolution. Consequently, contrast values can be standardized to a mean 0 and a variance 1 through dividing them by the square root of the sum of the branches involved in the comparison (Felsenstein, 1985; Garland et al., 1992). If properly standardized, contrasts can be used to perform basic statistics such as correlation and regression (Felsenstein, 1985; Garland et al., 1992; Garland and Ives, 2000).
We performed the analysis assuming two evolutionary models: Brownian motion (BM) and punctuated equilibrium or speciation (PE; Díaz-Uriarte and Garland, 1996; Felsenstein, 1985). Under the BM model of evolution, changes in character states follow a random process with mean 0 and branch length as expected variance. On the other hand, under the PE evolutionary model, changes in character states occur only associated with speciation events (Felsenstein, 1985; Díaz-Uriarte and Garland, 1996). PE can be implemented with the same methodology of PIC under a BM model but considering all branches of equal length. To guarantee that the PE analysis is not biased, the data set must include all species of the clade. Otherwise, additional opportunities of character evolution should be allowed one for each additional speciation event that has occurred along each branch segment (Díaz-Uriarte and Garland, 1996).
To perform the PIC, we used the program PDTREE (Garland and Ives, 2000; Garland et al., 1999) with the tree of Johnson and Lanyon (1999; Figure 1). We tested whether parasitic cowbirds differ from nonparasitic species in some of the proposed life-history characters. We had to prune off the clade that includes all parasitic species from the tree before the analysis (Figure 1; Garland and Ives, 2000). Under the BM evolutionary model, we used the original tree topology with branch lengths as shown in Figure 1. If information on character states of a tip species was not available, we pruned it off again. Under the PE evolutionary model, we had to avoid the bias of not considering all speciation events, so we first set up all branch length of the original tree to 1. Then we pruned off all the taxa without character information. As we transformed branch length to 1 before pruning, distances between nodes and remaining tip species in the original tree were kept in the pruned tree. Remaining branch lengths in the shortened tree thus still reflect the speciation events that have taken place on them (Björklund, 1997).
With both BM and PE evolutionary models, we tested if the generated contrasts were correctly standardized. We verify the lack of a significant correlation between contrasts and its SD as diagnostic of a correct standardization (Díaz-Uriarte and Garland, 1996; Garland et al., 1992). With the regression line built with the values of the standardized contrasts, we generated the 95% prediction intervals with PDTREE. Similar to a simple linear regression, prediction intervals of PIC are generated considering which value of the dependent variable is expected under a particular value of the independent one (Garland and Ives, 2000; Sokal and Rohlf, 1994). Moreover, PIC prediction intervals take into account the location of the species of interest and the length of the branch that attaches it to the tree (Garland and Ives, 2000). It is expected that species would be more similar to closely than to distantly related species. Moreover, the length of the branch leading to hypothetical species would be an estimator of its independent evolution (Garland and Ives, 2000). To generate the prediction interval the original tree must be re-rooted in an internal node. With this re-rooting, the species in question must be adjacent to its sister taxa and to the remaining tree (Garland and Ives, 2000; Garland et al., 1999). If the sister taxa of parasitic cowbirds were reduced to one species, we choose to re-root the tree in the midpoint of that branch (Garland et al., 1999; Garland and Ives, 2000). Generation of the prediction intervals was originally stated for only one taxon, and there are five species parasitic cowbirds. Therefore, we generated the prediction intervals using an average of the branch lengths that attached each cowbird species to the tree. Under a BM evolutionary model, the branch length we used was 150.2 (Figure 1), and under the PE model, it was 4.8. Results obtained with averages did not differ from other alternatives such as using the minimum or the maximum branch length of the cowbird's clade. When we had applied a transformation to branch lengths, the branch of the cowbirds was transformed in the same way.
When evolution is very rapid, taxa are not expected to show any similarity due to shared phylogeny (Björklund, 1997; Ricklefs and Starck, 1996). Under an evolutionary model, this approach is equivalent to all species being connected to each other in a "star phylogeny." That is, all evolved from a single common ancestor at the same time. The phylogenetic tree will look like it has a unique, hard (i.e., real) polytomy, and all branches will be the same length (Felsenstein, 1985; Martins, 2000). It has been suggested that if a range of evolutionary models all lead to the same conclusions, those conclusions are strengthened (Garland and Adolph, 1994; Garland et al., 1999). Therefore, we repeated all analysis considering the star phylogeny as an alternative. In practice, we performed simple linear regression with all variables using data of nonparasitic species to build the associated prediction intervals. In all cases, both variables were continuous and subjected to error. However, as we performed the analysis mainly for a predictive purpose, we applied the model I of the regression (Sokal and Rohlf, 1994). Again, if observed values of parasitic cowbirds were not included in the 95% prediction interval of the linear regression for a new observation, it may be considered that they differed from nonparasitic species (Sokal and Rohlf, 1994).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Adjustment and significance of the three evolutionary models
The correlations generated under BM and PE models, as well as under the simple regression of the star phylogeny, showed similar patterns of significance and adjustment (Table 1). Only the relationship of the average growth rate (K) and the asymptotic weight reached in the nest (A) under both the BM and the PE models were marginally significant (.1 > p >.05). In a similar way, the adjustments with simple linear regression were significant except for the relationship of the average and minimum incubation period and egg weight, which were nonsignificant and marginally significant, respectively (Table 1).
|
Contrasts and prediction intervals under a Brownian model
Most contrasts were adequately standardized. The only exceptions we found were the relationship of K with A and the minimum incubation period with egg size. Diagnosing plots showed a positive trend (Garland et al., 1992
). Contrasts achieved the correct standardization when we applied the fifth and sixth root to branch lengths. Egg weights of all parasitic cowbirds were included in the 95% prediction interval of the allometric function based on the weight of females (Figure 2a). Moreover, contrary to our expectation, the egg of the parasitic Scaphidura oryzivora remained on the limit of the upper prediction interval. The hypothesis of the lower energy content of the egg was not supported for any species of brood parasite. The weights of the nestling at hatching of all parasitic cowbird were included in the 95% prediction interval based on egg weight (Figure 2b). We did not we find any evidence that brood parasites have shortened the incubation periods of their eggs. Both average and minimum incubation recorded were included in the prediction interval based on egg weight (Figure 2c,d). Finally, we did not detect any special pattern in weight development in the nestling of brood parasites. Average growth rates of all parasites were included into the interval generated by the asymptotic weight reached in the nest (Figure 3a). Likewise, the weights attained before leaving the nest did not differ from expected by their adult weight (Figure 3b). However, values of many parasitic species appeared in the upper limit of the interval. The only characteristic that clearly separated brood parasites from nonparasites was eggshell thickness. All parasitic cowbirds had thicker eggshells than predicted by their egg weight. However, three nonparasitic species, Pseudoleistes virescens, P. guirahuro, and Molothrus badius, also presented significantly thicker eggshells (Figure 4).
|
|
|
Prediction intervals under a punctuated equilibrium
All contrasts were adequately standardized. Prediction intervals generated using a speciation model of evolution showed the same pattern. Resulting graphs were undistinguished from those obtained under a BM evolutionary model. When estimating 95% prediction intervals for parasitic cowbirds, brood parasites differed from nonparasites only in their eggshell thickness (Table 2). Again, Pseudoleistes virescens, P. guirahuro, and Molothrus badius had thicker eggshells (data not shown).
|
Prediction intervals under a star phylogeny
Most results obtained under the assumption of a star phylogeny did not differ from those obtained assuming a shared phylogeny (Table 2). Brood parasites did not differ from nonparasites in respect to egg weight, energy content of the egg (measured as weight of the nestling at hatching), incubation period, or average growth rate. We only detected minor differences in respect to the weight attained before fledgling and eggshell thickness. Parasitic cowbirds whose asymptotic weights before fledgling were included in the upper limit of the interval generated by the regression now were significantly larger (M. b. bonariensis, M. rufoaxillaris, and M. ater). Concerning eggshell thickness, brood parasites again differed from nonparasites except M. ater, which was included in the upper limit of the prediction interval (Table 2).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Phylogenetic analysis of adaptations in cowbirds
Most of the life-history adaptations proposed for parasitic cowbirds were not supported by an exhaustive analysis taking into account the phylogeny of the group. Parasitic cowbirds had not reduced the weight or the energy content of their eggs. Similarly, they had not reduced their incubation periods in regard to the average or the minimum recorded. Finally, we did not detect any difference in the growth pattern of the cowbird nestlings, either in respect to the average growth rate or the asymptotic weight reached before fledgling. The only tested characteristic that clearly separated parasitic cowbirds from nonparasitic relatives was the increase in eggshell thickness. All these findings were robust and, except for slight differences in the marginal results, resisted the use of Brownian motion or punctuated equilibrium models. Our results did not change in a pronounced way when we considered that species did not show similarities due to shared ancestry in most variables of interest (star phylogeny). However, eggshell thickness in M. ater was included in the prediction interval generated by nonparasites (SF; Table 2). Consequently, correcting for the non-independence of the data, we conclude that only eggshell thickness has evolved in the context of brood parasitism.
Methodological considerations: evolutionary scenario
All conclusions based on a particular phylogeny rely heavily on the accuracy of the evolutionary scenario used to test them. If that scenario is wrong, so are the conclusions (Coddington, 1992). The use of taxonomy with traditional statistics did not resolve the question, as this method also relies on taxonomy accurately reflecting the membership of species while assuming a star phylogeny model of evolution (Felsenstein, 1985; Martins, 2000). Other phylogeny that includes all cowbird species is based on presence of restriction sites of the whole mtDNA (Freeman and Zink, 1995). Based on this phylogeny, brood parasitism would have evolved three times (Freeman and Zink, 1995). However, Freeman and Zink (1995) discuss that when they combined their data with that of Lanyon (1992), they recovered the same tree as Lanyon's (i.e., brood parasitism evolving only once). This result would indicate that phylogenetic signal of restriction sites of the whole mtDNA was weaker than that of sequences of cytochrome b (Eernisse and Kluge, 1993; Freeman and Zink, 1995). Therefore, we did not consider the phylogeny of Freeman and Zink (1995) as an alternative evolutionary hypothesis.
Methodological considerations: variables and species considered
For some variables, such as the incubation periods and the average growth rate of the nestlings, prediction intervals were very broad (Figures 2c,d and 3a and Table 2). Such broad prediction intervals make it difficult to detect any evolutionary change in the brood parasites. The number of species used for generating those prediction intervals was low (9 and 15 or 16, respectively). This low number of species could be responsible for the low resolution of the associated prediction intervals. However, when analyzing the relationship of asymptotic weight reached in the nest and weight of the adult, we have data for only 11 species, but prediction intervals were narrower (Figure 3b and Table 2). Moreover, in the analysis of average growth rate, the nine species were distributed in seven different clades representing most of the original tree's topology (Figure 1; Appendix). The broad prediction intervals may be related to the high variation of incubation periods and growth rate of the nestling. Finally, it may be considered that growth rates developed by Ricklefs (1967, 1968) would be comparable only when all species fledge at the same stage of development. This may not be true if some species leave the nest when reaching near-adult full weight and others leave the nest earlier. However, our analysis of the relationship of the asymptotic weight reached in the nest and adult weight failed to detect any special pattern among parasitic cowbirds.
Why these adaptations were proposed?
Most of the proposed adaptations to brood parasitism arise from comparative analysis of only one species of cowbird with data predicted from general allometric equations. The comparison of the egg weight of M. ater and M. b. cabanisii with the general equation for passerines by Rahn and Ar (1975), egg weight = 0.34 x FW0.677 (see Appendix for nomenclature), led to the proposal that brood parasites have reduced the size of their eggs (Kattan, 1995; Strausberger, 1998a). At the same time, a comparative study revealed that the egg size of brood parasites does not differ from other Icterids (Rahn et al., 1988). When we applied the equation of Rahn and Ar (1975) to the grackles and allies, we found that it overestimates egg weight in 20/23 of the nonparasites (Appendix; paired t test, t = 4.717, p <.001). This overestimation was also observed when we analyzed incubation periods. Shorter incubation periods in brood parasites were suggested from the comparison with the general equation of Vleck and Vleck (1987): log(incubation period) = 0.97 + 0.29 log(egg weight) (Kattan, 1995; Strausberger, 1998a). Again, when we applied this equation to nonparasitic grackles and allies, it overestimated the mean incubation period of 14/15 species (Appendix; paired t test, t = 3.455, p <.005). Finally, the larger mean growth rate of brood parasites was proposed when comparing them with the general equation of Ricklefs (1968) for passerines and raptors: K = 1.11A-0.278 (Kattan, 1996). The Ricklefs (1968) equation underestimated the value of K in all of the nine nonparasitic species (Appendix; paired t test, t = 71.8, p <.001). This pattern had been reported, but with a small number of species (Teather and Weatherhead, 1994). If this lack of adjustment to general equations has biological meaning, it may imply that most of the proposed adaptations to brood parasitism were inherited from a nonparasitic ancestor. The nonbiological alternative would be the lack of adjustment is the consequence of using general equations that do not provide associated prediction intervals.
Possible function of the increase in eggshell thickness
Our analysis shows that cowbirds have developed thicker eggshells, as previously suggested by many authors (Rahn et al., 1988; Schönwetter, 1984; Spaw and Rohwer, 1987). The adaptive significance of eggshell thickness is commonly associated with host rejection of parasite eggs: some hosts do not reject cowbird eggs because costs of rejection could be higher than costs of acceptance (Lotem and Nakamura, 1998; Rohwer and Spaw, 1988). This may be true especially for small-billed hosts that must peck parasite eggs before rejection. Eggshell thickness in cowbirds increased rejection costs, forcing small-billed hosts to accept parasitic eggs (Picman, 1989; Rohwer et al., 1989; Rohwer and Spaw, 1988; Spaw and Rohwer, 1987). However, only three cowbird hosts, all of M. ater, are peck ejectors (Rohwer and Spaw, 1988; Rohwer et al., 1989; Sealy, 1996). The existence of thicker eggshells would not be advantageous when hosts manipulate parasite eggs without damage (i.e., all other rejecting hosts; Fraga, 1998; Mason, 1986; Mermoz, 1996; Rohwer et al., 1989; Smith, 1968). Another alternative is that all parasitic Molothrus females peck eggs in parasitized nests (Carter, 1986; Fraga, 1998; Friedmann, 1929; Hudson, 1874; Sealy, 1992). Consequently, thicker eggshells may protect them from pecks of other females in multiply parasitized nests (Brooker and Brooker, 1991; Hudson, 1874; Mermoz and Reboreda, 1999). Finally, the last explanation is that the thicker eggshell evolved to protect eggs from accidental breakage in the nest as a result of the larger combined clutch size (host plus parasite) of parasitized nests (Weatherhead, 1991).
No information on frequency of brood parasitism is available for Pseudoleistes guirahuro. Molothrus badius and P. virescens are the only hosts regularly parasitized by Molothrus bonariensis and M. rufoaxillaris, with frequencies of parasitism of 67–100% (Hudson, 1874; Fraga, 1998; Mason, 1980; Mermoz and Fernández, 2003; Mermoz and Reboreda, 1999). An increase in eggshell thickness among nonparasites would be expected only under the evolutionary pressure of the parasites, as neither host pecks parasite eggs before rejection (Fraga, 1998; Mermoz, 1996). In summary, there is little evidence favoring the evolution of egg thickness in cowbird eggs as a defense against rejecting hosts. Moreover, Scaphidura oryzivora apparently does not peck eggs (Smith, 1968; but see Fleischer and Smith, 1992). Consequently, both egg pecking by the parasites as well as accidental breakage would play an important role in the increased thickness of cowbird eggs. In a similar way, both characteristics might also explain the independent evolution of eggshell thickness observed in two of the most parasitized grackle and allies' hosts (Figure 1). To rule out that this is not a coincidence, one must search for independent evolution of thicker eggshells in highly parasitized hosts from other bird lineages.
The "no adaptations" of parasitic cowbirds
Based on the evolutionary definition of adaptation, the only tested characteristic of parasitic cowbirds that apparently was built by natural selection for their proposed current role (Gould and Vrba, 1982; Reeve and Sherman, 1993) was the increase in eggshell thickness. All other features tested, such as cheaper eggs as well as nestlings with shorter incubation periods and particular growth strategies, would be exaptations (i.e., inherited from a nonparasitic ancestor). Nevertheless, generalist cowbirds still have shorter incubation periods and relatively faster growth rate than most of their hosts (Rothstein and Robinson, 1998). Other proposed adaptations like rapid egg laying (Sealy et al., 1995) as well as high fecundity of females (Kattan, 1993; Scott and Ankney, 1983; but see Alderson et al., 1999) could not be tested in our study due to the lack of comparable data. The lack of specific characteristics that distinguished parasitic cowbirds from their nonparasitic relatives apparently seemed not to affect their success. During the last 50–150 years, at least all Molothrus species that are brood parasites have markedly expanded their area of geographic distribution and/or the number of hosts they use (Carter, 1986; Fraga, 1996; Friedmann, 1929; Friedmann and Kiff, 1985; Hudson, 1874; Mayfield, 1965; Post et al., 1993). Human habitat modifications such as forest clearing and livestock rearing clearly have favored them (Mayfield, 1965; Post et al., 1990). However, parasitic cowbirds could have not taken advantage of these favorable changes if they could not cope with all other factors that guarantee successful reproduction. The apparent high fecundity of parasitic cowbirds (Davis, 1942; Kattan, 1993; Scott and Ankney, 1983) may be accompanied with changes in females' behavior. If cowbird females can select the most appropriate hosts (Wiley, 1985, 1988) and monitor their nests to achieve good synchronization of the parasitic event (Massoni and Reboreda, 1998; Mermoz and Reboreda, 1999; Strausberger, 1998b; Strausberger and Ashley, 1997), they probably do not need further changes in egg or chick morphologies to be successful. Furthermore, if the grackle and allies' ancestor developed traits such as short incubation period or rapid growth rate, they would have increased the success of a proto-parasite. Finally, data on cytochrome-b nucleotide substitution rates indicate that cowbirds have been parasites for as much as half the time than the older parasitic cuckoos (Rothstein et al., 2002). The more recent origin of cowbirds may then explain their small number of specific adaptations to brood parasitism.
|
|
|
ACKNOWLEDGEMENTS |
|---|
We specially thank Ted Garland Jr. for providing us with a copy of the PDTREE program and Alejandro Espinosa de los Monteros for assisting us with getting branch lengths of Johnson and Lanyon's tree. We also thank Victoria Sosa and the technicians at the computer lab of the Instituto de Ecología, A. C. for providing us with PC computational facilities. Gustavo J. Fernández, Pablo Tubaro, Ron Ydenberg, and two anonymous reviewers made valuable comments on early versions of the manuscript. M.E.M. was supported by Postdoctoral Fellowships from the CONICET (Argentina) and from the "Programa Cuauhtémoc" of the Secretaría de Relaciones Exteriores and CONACyT (México).
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Alderson GW, Gibbs HL, Sealy SG, 1999. Determining the reproductive behaviour of individual brown-headed cowbird using microsatellite DNA markers. Anim Behav 58:895-905.[CrossRef][ISI][Medline]
Ar A, Paganelli CV, Reeves RB, Greene DG, Rahn H, 1974. The avian egg: water vapor conductance, shell thickness and functional pore area. Condor 76:153-158.
Ar A, Rahn H, Paganelli CV, 1975. The avian egg: mass and strength. Condor 81:331-337.
Aragón S, Møller AP, Soler JJ, Soler M, 1999. Molecular phylogeny of cuckoos supports a polyphyletic origin of brood parasitism. J Evol Biol 12:495-506.[CrossRef]
Bancroft GT, 1984. Growth and sexual dimorphism of the boat-tailed grackle. Condor 86:432-432.
Björklund M, 1997. Are "comparative methods" always necessary? Oikos 80:607-612.[CrossRef]
Briskie JV, Sealy SG, 1990. Evolution of short incubation periods in the parasitic cowbirds, Molothrus spp. Auk 107:789-794.
Brooks RB, McLennan DA, 1991. Phylogeny, ecology and behavior. A research program in comparative biology. Chicago: University of Chicago Press.
Brooker LC, Brooker MG, 1990. Why are cuckoo host specific? Oikos 57:301-309.[CrossRef]
Brooker MG, Brooker LC, 1991. Eggshell strength in cuckoos and cowbirds. Ibis 133:406-413.
Carter MD, 1986. The parasitic behavior of the bronzed cowbird in south Texas. Condor 88:11-25.
Coddington JA, 1988. Cladistic tests of adaptational hypotheses. Cladistics 4:3-22.
Coddington JA, 1992. The comparative method in evolutionary biology. Trends Ecol Evol 7:68-69.[CrossRef]
Davies NB, 2000. Cuckoo, cowbirds and other cheats. London: Academic Press.
Davies NB, Brooke MdL, 1989. An experimental study of co-evolution between the cuckoo, Cuculus canorus and its hosts. II. Host egg markings, chick discrimination and general discussion. J Anim Ecol 58:225-236.[CrossRef]
Davis DE, 1942. The number of eggs laid by cowbirds. Condor 44:10-12.
Díaz-Uriarte R, Garland T, Jr, 1996. Testing hypotheses of correlated evolution using phylogenetically independent contrasts: sensitivity to deviations from Brownian motion. Syst Biol 45:27-47.[CrossRef]
Dearborn DC, 1996. Video documentation of a brown-headed cowbird nestling ejecting an indigo bunting nestling from the nest. Condor 98:645-649.
Dufty AM, Jr, 1982. Movements and activities of radio-tracked brown-headed cowbirds. Auk 99:316-327.
Dunning JB, 1993. CRC handbook of avian body masses. Boca Raton, Florida: CRC Press.
Eernisse DJ, Kluge G, 1993. Taxonomic congruence versus total evidence, and amniote phylogeny inferred from fossils, molecules, and morphology. Mol Biol Evol 10:1170-1195.[Abstract]
Elliot PF, 1978. Cowbird parasitism in on the Kansas tallgrass prairie. Auk 95:161-167.
Felsenstein J, 1985. Phylogenies and the compartive data. Am Nat 125:1-15.
Fiala KL, Congdon JD, 1983. Energetic consequences of sexual size dimorphism in nestling red-winged blackbird. Ecology 64:642-647.[CrossRef]
Fleischer RC, 1985. A new technique to identify and asses the dispersion of eggs of individual brood parasites. Behav Ecol Sociobiol 17:91-99.
Fleischer RC, Smith NG, 1992. Giant cowbird egg in the nests of two icterid hosts: the use of morphology and electrophoretic variants to identify individuals and species. Condor 94:574-578.
Fraga RM, 1996. Further evidence of parasitism of chopi blackbirds (Gnorimopsar chopi) by specialized screaming cowbird (Molothrus rufoaxillaris). Condor 98:866-867.
Fraga RM, 1998. Interactions of the parasitic screaming and shiny cowbirds (Molothrus rufoaxillaris and M. bonariensis) with a shared host, the bay-winged cowbird (M. badius). In: Parasitic birds and their hosts: studies in coevolution (Rothstein SI, Robinson SK, eds). Oxford: Oxford University Press; 172–193.
Fraga RM, Casañas H, Pugnali G, 1998. Natural history and conservation of the endangered saffron-cowled blackbird Xanthopsar flavus in Argentina. Bird Conserv Intl 8:255-267.
Freeman S, Zink RM, 1995. A phylogenetic study of the blackbirds based on variation in mitochondrial DNA restriction sites. Syst Biol 44:409-420.[CrossRef]
Friedmann H, 1929. The cowbirds, a study in the biology of the social parasitism. Springfield, Illinois: C.C. Thomas.
Friedmann H, 1955. The honeyguides. US Natl Mus Bull 223:1-196.
Friedmann H, 1960. The parasitic weaverbirds. US Natl Mus Bull 233:1-273.
Friedmann H, Kiff LF, 1985. The parasitic cowbirds and their hosts. Proc West Found Vert Zool 2:226-304.
Garland T, Jr, Adolph SC, 1994. Why not to do two species comparative studies: limitations on inferring adaptations. Physiol Zool 67:797-828.
Garland T, Jr, Harvey PH, Ives AR, 1992. Procedures for the analysis of comparative data using phylogentically independent contrasts. Syst Biol 41:18-32.[CrossRef]
Garland T, Jr, Ives AR, 2000. Using the past to predict the present: confidence intervals for regression equations in phylogenetic comparative methods. Am Nat 155:346-364.[CrossRef][Medline]
Garland T, Jr, Midford PE, Ives AR, 1999. An introduction to phylogenetically based statistical methods, with a new method for confidence intervals on ancestral states. Am Zool 39:374-388.
Gould SJ, Vrba, ES., 1982. Exaptation: a missing term in the science of form. Paleobiology 8:4-15.[Abstract]
Harvey PH, Pagel M, 1991. The comparative method in evolutionary biology. Oxford: Oxford University Press.
Hudson WH, 1874. Notes on the procreant instincts of the three species of Molothrus found in Buenos Ayres. Proc Zool Soc Lond 9:153-174.
Johnson KP, Lanyon SM, 1999. Molecular systematics of the grackles and allies, and the effect of additional sequence (Cyt b and ND2). Auk 116:759-768.
Kattan GH, 1993. Reproductive strategy of a generalist brood parasite, the shiny cowbird, in the Cauca Valley, Colombia (PhD dissertation). Gainesville: University of Florida.
Kattan GH, 1995. Mechanism of short incubation period in brood-parasitic cowbirds. Auk 112:335-342.
Kattan GH, 1996. Growth and provisioning of shiny cowbird and house wren host nestlings. J Field Ornithol 67:434-441.
Lanyon SM, 1992. Interspecific brood parasitism in blackbirds (Icterinae): a phylogenetic perspective. Science 225:77-79.
Lanyon SM, Omland KE, 1999. A molecular phylogeny of the blackbirds (Icterinae): five lineages revealed by cytochrome-b sequence data. Auk 116:629-639.
Lotem A, Nakamura N, 1998. Evolutionary equilibria in avian brood parasitism: an alternative to the "arms race-evolutionary lag" concept. In: Parasitic birds and their hosts: studies in coevolution (Rothstein SI, Robinson SK, eds). Oxford: Oxford University Press; 223–235.
Martins EP, 2000. Adaptation and the comparative method. Trends Ecol Evol 15:296-299.[CrossRef][Medline]
Martins EP, Hansen TF, 1996. The statistical analysis of interespecific data: a review and evaluation of phylogenetic comparative methods. In: Phylogenies and the comparative methods in animal behaviour (Martins EP, ed). Oxford: Oxford University Press; 22–75.
Mason P, 1980. Ecological and evolutionary aspects of host selection in cowbirds (PhD dissertation). Austin: University of Texas.
Mason P, 1986. Brood parasitism in a host generalist, the shiny cowbird: I. The quality of different species as hosts. Auk 106:52-60.
Massoni V, Reboreda JC, 1998. Costs of brood parasitism and the lack of defenses on the yellow-winged blackbird-shiny cowbird system. Behav Ecol Sociobiol 42:273-280.[CrossRef]
Mayfield H, 1965. The brown-headed cowbird, with old and new hosts. Living Birds 4:13-28.
Mermoz ME, 1996. Interacciones entre el tordo renegrido Molothrus bonariensis y el pecho amarillo Pseudoleistes virescens: estrategias del parásito de cría y mecanismos de defensa del hospedador (PhD dissertation). Buenos Aires: Universidad de Buenos Aires.
Mermoz ME, Fernández GJ, 2003. Breeding success of a specialist brood parasite, the screaming cowbird, parasitizing an alternative host. Condor 105:63-72.[CrossRef]
Mermoz ME, Reboreda JC, 1999. Egg laying behaviour by shiny cowbirds parasitizing brown-and-yellow marshbirds. Anim Behav 58:873-882.[CrossRef][ISI][Medline]
Middleton LA, 1991. Failure of brown-headed cowbird parasitism on nests of the American goldfinch. J Field Ornithol 62:200-203.
Morton E, Farabaugh SM, 1979. Infanticide and other adaptations of the nestling striped cuckoo Tapera naevia. Ibis 121:212-213.
Nolan V, Jr, Thompson CF, 1978. Egg volume as a predictor of hatchling weight in the brown-headed cowbird. Wilson Bull 90:353-358.
Ortega CP, Cruz A, 1992. Differential growth patterns of nestling brown-headed cowbirds and yellow-headed blackbirds. Auk 109:368-376.
Payne RB, 1973. Individual laying histories and the clutch size and number of eggs of parasitic cuckoos. Condor 75:414-438.
Payne RB, 1977. The ecology of brood parasitism in birds. Annu Rev Ecol Syst 8:1-28.[CrossRef][ISI]
Payne RB, 1997a. Avian brood parasitism. In: Host-parasite evolution: general principles and avian model (Clayton DH, Moore J, eds). Oxford: Oxford University Press; 338–369.
Payne RB, 1997b. Family Cuculidae (Cuckoos). In: Handbook of the birds of the word, vol. 4. Sandgrouse to cuckoos (del Hoyo J, Elliot A, Sargatal J, eds). Barcelona: Lynx Ediciones; 508–607.
Payne RB, Payne LL, Woods JL, Sorenson MD, 2000. Imprinting and the origin of parasite-host association in brood parasites indigobirds, Vidua chalybeata. Anim Behav 59:69-81.[CrossRef][ISI][Medline]
Peer BD, Bollinger EK, 1997. Explanations for the infrequent cowbird parasitism on common grackles. Condor 99:151-161.
Picman J, 1989. Mechanism of increased puncture resistance of eggs of brown-headed cowbirds. Auk 106:577-583.[ISI]
Post W, Cruz A, McNair DB, 1993. The North-American invasion pattern of the shiny cowbird. J Field Ornithol 64:32-41.
Post W, Nakamura TK, Cruz, A, 1990. Patterns of shiny cowbird parasitism in St. Lucia and southwestern Puerto Rico. Condor 92:461-469.
Rahn H, Ar A, 1974. The avian egg: incubation time and water loss. Condor 76:147-152.
Rahn H, Ar A, 1975. Relation of avian egg weight to female body weight. Auk 92:750-756.
Rahn H, Curran-Everett L, Booth DT, 1988. Eggshell differences between parasitic and nonparasitic Icteridae. Condor 90:962-964.
Ralph MH, 1975. Development of young Brewer's blackbird. Wilson Bull 87:207-230.
Redondo T, 1993. Exploitation of host mechanisms for parental care by avian brood parasites. Etologia 3:235-297.
Reeve HK, Sherman, PW, 1993. Adaptation and the goals of evolutionary research. Q Rev Biol 68:1-31.[CrossRef]
Ricklefs RE, 1967. A graphical method of fitting equations to growth curves. Ecology 48:978-983.[CrossRef]
Ricklefs RE, 1968. Pattern of growth in birds. Ibis 110:419-451.