Examining two standard assumptions of ancestral reconstructions: repeated loss of dichromatism in dabbling ducks (Anatini).
While these techniques have been embraced by systematists and evolutionary biologists, the assumptions underlying these methods, and the inherent limitations and pitfalls that might follow, require continued examination. A quote from two authors of the software packages commonly used for character reconstruction (PAUP and MacClade) emphasizes the caution that should be used: "Character state reconstructions can provide a powerful mechanism for studying many facets of the evolutionary process. However, the zeal with which these techniques are sometimes advocated belies the complexity of the problem." (Swofford and Maddison 1992, p. 220).
Relatively few studies have assessed the reliability of ancestral reconstructions (Hillis et al. 1992; Collins et al. 1994; Frumhoff and Reeve 1994; Pagel 1994; Maddison 1995; Donoghue and Ackerly 1996; Schultz et al. 1996; also see Maddison 1994). In contrast there has been a profusion of work designed to assess the reliability of phylogenies themselves (e.g., Felsenstein 1988; Graybeal 1994; Omland 1994; Hillis et al. 1994). Several of the general approaches used in assessing tree reliability can be applied to character state reconstructions, especially congruence analysis. Congruence analysis has been applied to tree reliability in two ways: congruence between any two independent data sources (e.g., Bousquet et al. 1992), and between a new phylogenetic hypothesis and a "known" phylogeny (e.g., Graybeal 1994). The logic of both of these approaches can be applied to character reconstruction. The current paper uses a combination of these two methods to assess the reliability of ancestral plumage reconstructions in dabbling ducks.
Waterfowl plumage patterns have several features that make them a good model system for understanding macro-evolutionary patterns at the species level. There is every reason to believe that all extant taxa are described and that their distributions are well known. Dabbling ducks have long been used as models for studies of behavior and phylogeny (e.g., Delacour and Mayr 1945; Lorenz 1971; Kessler and Avise 1984). Furthermore, plumage characters can be scored easily and reliably for all taxa in the field or from museum specimens (unlike behavioral or internal features). Duck plumage and speciation patterns are so well studied that there has emerged a widely accepted scenario of speciation and plumage evolution.
Most widespread Northern Hemisphere ducks such as the mallard (Anas platyrhynchos) are sexually dichromatic (i.e., sexually dimorphic in plumage coloration with bright males and cryptic females). In contrast many of their insular and more southerly relatives are monochromatic (usually with both sexes cryptic) (e.g., Delacour and Mayr 1945; Sibley 1957; Livezey 1991). This geographic pattern of monochromatism and dichromatism occurs repeatedly in each of the six major clades within the dabbling ducks (Anas): wigeons, mallards, blue-wings, austral teal, pintails, and the green wing clade (see [ILLUSTRATION FOR FIGURE 1A OMITTED] below) (Livezey 1991). Thus, clearly, there were frequent transformations; the question is, what was the direction of those transformations.
Delacour and Mayr (1945) were the first to write about the pattern of monochromatism and dichromatism in ducks. based on several lines of evidence that I will review below, these and many subsequent authors agree that dull plumage has been derived secondarily within the mallards and other Anas lineages (e.g., Delacour 1956; Ripley 1957; Johnsgard 1961, 1965; Kear 1970; Kear and Murton 1976; Weller 1980; Madge and Burn 1988; Avise et al. 1990; Livezey 1991, 1993; Peterson 1996; but see Palmer 1976). Thus I will refer to this scenario as the "widely accepted scenario," which posits that multiple monochromatic species evolved from the widespread dichromatic species in most of the clades (the wigeons, mallards, blue-wings, austral teal, and pintails). (Note that within the order Anseriformes, sexual dichromatism is recently derived and is limited to tribes in the sister subfamilies Tadorninae and Anatinae, Livezey 1986). Similar patterns of dichromatism and monochromatism occur in other lineages of birds (e.g., Australian flycatcher, Petroica multicolor) and invertebrates. Monochromatism seems to be derived recently in these lineages as well (Mayr 1942; Peterson 1996).
METHODS AND RESULTS
I used Livezey's (1991) morphological phylogeny as a basis for reconstructing the history of plumage evolution in dabbling ducks. Livezey's phylogenetic hypothesis includes all Recent species of Anatini. Furthermore, his data are highly congruent with Kessler and Avise's (1984) molecular data on nine North American species (Omland 1994). This result suggests that Livezey's tree may be a reliable basis for reconstructing ancestral plumage states, at least with regard to the relationships among major clades.
The 59 phylogenetic species Livezey (1991) identified in the tribe Anatini were used (although a few authors have questioned whether some monochromatic ducks should be considered full species, e.g., Sibley 1957; Johnsgard 1965). Livezey's (1991) classification was followed, except that for the wigeons the genus name Anas was retained. I used a bifurcating topology that resolved the four polytomies in Livezey's (1991) tree. These resolutions facilitated character state reconstruction: two polytomies were resolved based on reanalysis of Livezey's matrix and the remaining two polytomies were resolved arbitrarily. Other resolutions of Livezey's polytomies do not strongly affect plumage mapping because they usually involve species or clades with identical character states. For example, one of the unresolved polytomies involves the clade consisting of A. platyrhynchos, A. wyvilliana, A. laysanensis, and A. oustaleti; since all are coded as dichromatic, their ancestor is reconstructed as dichromatic regardless of the arrangement of these taxa. I coded species as monochromatic or dichromatic following Scott and Clutton-Brock (1989). This coding assumes that the mechanism(s) that cause differential expression of plumage ornaments in the sexes is homologous across species (see Discussion).
Reconstructing Plumage Dichromatism
I used "simple parsimony" (gains and losses given equal probabilities) in MacClade (Maddison and Maddison 1992) to produce a reconstruction based on the present-day character states. This assumption, also referred to as "unordered parsimony," is the default setting in MacClade (Maddison and Maddison 1992) and PAUP (Swofford 1993). The reconstruction based on equal transformational weighting suggests that the common ancestor of all Anas and the immediate ancestors of most clades were monochromatic [ILLUSTRATION FOR FIGURE 1A OMITTED]; this reconstruction disagrees with the accepted scenario. However, the standard assumption of equal probability of gains and losses of character states probably does not apply here. Later I will detail four lines of evidence supporting this assertion.
There are few good data to assist in determining more realistic estimates of the transformation probabilities between monochromatism and dichromatism. There is no specific information on the genetic or developmental bases for these transformations (but see Peterson 1996; Price and Birch 1996). Therefore, I applied a range of transformation probabilities to determine: (1) the sensitivity of the reconstruction to changes in probabilities (see Donoghue and Ackerly 1996); and (2) the magnitude of the transformation probability necessary to recover the widely accepted scenario using the existing phylogeny.
Alternative Transformation Weightings
The use of different transformation matrices in MacClade dramatically altered the ancestral reconstructions. This is not surprising given such evolutionarily dynamic character states. When I made it twice as likely to lose dichromatic plumage, the blue-wings and green-wings were connected with a dichromatic ancestor. When it was three times more likely the austral teal ancestor was shown as dichromatic [ILLUSTRATION FOR FIGURE 1B OMITTED]. When losses were four times more likely all clades had dichromatic ancestors except the mallard, which was equivocal. Only with a step-matrix that made it five times more likely to lose dichromatism were all the ancestors including the mallard reconstructed as dichromatic [ILLUSTRATION FOR FIGURE 1C OMITTED]. This reconstruction agrees with the accepted scenario, especially in that monochromatism independently evolved repeatedly in each of the six clades. Dollo transformation, which makes independent gains extremely unlikely (Swofford et al. 1996), produced this same reconstruction.
Altering transformation probabilities in the other direction, making losses of dichromatism five times less likely than gains, only slightly alters the reconstruction in Figure 1A. Within Anas the wigeon ancestor is shown as monochromatic, and three basal outgroup nodes are also reconstructed as monochromatic.
When I used Scott and Clutton-Brock's (1989) alternative continuous plumage scale, in which they scored brightness on an ordered scale of one to four, similar results were found (unpubl. data). However, using the ordered scale further biases the ancestral reconstructions away from strong dichromatism; several taxa with plumage that only differs slightly between the sexes are treated as dichromatic in the two-state coding (A. wyvilliana, A. laysanensis, and A. oustaleti in the mallard group; A. smithii, A. platalea, and A. rhynchotis in the shoveler/blue-wing group [Scott and Clutton-Brock 1989] [ILLUSTRATION FOR FIGURE 1 OMITTED]). Thus the results of the ordered scale differ even more strongly from the accepted scenario.
The Impact of Paraphyletic Species
It might seem possible that minor rearrangements of taxa in Livezey's (1991) tree would change the simple reconstruction in Figure 1A. But when using his species as terminal taxa, the predominance of monochromatic species almost invariably results in the reconstruction of monochromatic ancestors. The other crucial issue that should be considered is the assumption that these duck species are monophyletic and are the appropriate terminal taxa (see Graybeal 1995). Most of the Anas species in Livezey's (1991) tree lack autapomorphies that would support their monophyly, including two of the Holarctic dichromatic species (mallard [A. platyrhynchos] and northern shoveler [A. clypeata]). Livezey (1991) documents autapomorphies for only 18 of the 46 Anas, including three of the Holarctic dichromatic species (gadwall [A. strepera], blue-winged teal [A. discors], and northern pintail [A. acuta]) (Livezey 1991). Further examination of monophyly is warranted even for these species. Any study of dabbling ducks that treated species as the operational taxonomic unit would likely be hampered by this situation and would produce similar reconstructions. For example any molecular study that used one individual as an exemplar would not recover evidence of the paraphyletic relationships that result from peripatric speciation (Avise et al. 1990; Peterson 1992; Hoelzer and Melnick 1994; Patton and Smith 1994; Cooper et al. 1996). Because peripatric (founder effect) speciation may be prevalent in these ducks, population-based molecular sampling of each species is needed to reconstruct their evolutionary history.
Since the reconstruction based on standard unordered parsimony differs from the widely accepted scenario, is it possible to decide which is better supported? I will review seven lines of evidence consistent with ancestral dichromatism in Anas. The first three provide evidence that the monochromatic species should be nested within dichromatic northern species, which is inconsistent with the assumption of species monophyly. The last four suggest evolution may be biased toward the loss of dichromatism, which would challenge the assumption of equal gain-loss probabilities. Thus, these lines of evidence point out two sources of error that likely affect parsimony reconstruction in ducks and many other taxa. Some of the lines of evidence may not be entirely independent of others. However, each represents an aspect of speciation or plumage evolution that should be considered explicitly when reconstructing the history of dichromatism in ducks.
Seven Lines of Evidence Consistent with Ancestral Dichromatism in Anas
1. Population-Level Phylogenies of Mallards
The first line of evidence of ancestral dichromatism comes from phylogenetic studies that included multiple samples of mallards and monochromatic relatives. Cooper et al. (1996) used control region mtDNA sequences and showed that the mallard (A. platyrhynchos) is paraphyletic with respect to the Hawaiian duck (A. wyvilliana) [ILLUSTRATION FOR FIGURE 2 OMITTED]. Male Hawaiian ducks have plumage quite similar to female mallards, although some individuals show hints of features typical of adult male mallards (see below) (Madge and Burn 1988). These data (Cooper et al. 1996) suggest that the Hawaiian duck may be a direct descendant of the holarctic mallard. Their data do not place the other dull species from Hawaii, the Laysan duck (A. laysanensis), within the holarctic mallard clade; however the bootstrap support for this arrangement is not strong (Cooper et al. 1996) [ILLUSTRATION FOR FIGURE 2 OMITTED].
A second study (Avise et al. 1990) showed that the monochromatic American black duck (A. rubripes) may also be a direct descendant of the dichromatic mallard. Restriction fragment length polymorphisms revealed two mitochondrial DNA lineages in North America: the A lineage contains both mallards and black ducks, while the B lineage contains just mallards. A third study, based on mitochondrial sequences, shows that the mallard is paraphyletic with respect to several other monochromatic mallard relatives from North America and Asia (K. Johnson and M. Sorenson, pers. comm. 1997). The results of these three studies suggest that the Hawaiian duck, the American black duck, and multiple other monochromatic species may have evolved from the mallard by peripatric speciation. These studies show that the mallard is paraphyletic, and the assumption of species monophyly may not be justified. Other data are needed to determine whether similar scenarios apply to the remaining dull mallard relatives, as well as dull isolated species in other Anas lineages.
A second line of evidence that dichromatic species may be ancestral (and paraphyletic) is the geographic distribution of dichromatic and monochromatic species. Most of the dichromatic Anas are widespread migratory continental species that breed in northern temperate zones, while the monochromatic species tend to be sedentary and have restricted isolated ranges, often on oceanic islands and southern continents. For example, the mallard breeds throughout the Northern Hemisphere (Madge and Burn 1988). The other thirteen species in the mallard clade (Livezey 1991) are predominantly monochromatic and generally restricted to islands and small continental ranges in the tropics or subtropics [ILLUSTRATION FOR FIGURE 3 OMITTED]. It seems likely that wayward dichromatic mallards became isolated, thus forming the predominantly sedentary monochromatic species. The genetic data just discussed support this scenario for the mallard clade. Insular monochromatic species in other lineages seem especially likely to have formed by peripatric speciation in this way (e.g., Kerguelen Island pintail, A. eatoni; Coues' gadwall, A. strepera cousei [recently extinct]) (e.g., Delacour 1956; Johnsgard 1965). The opposite scenario involving insular species colonizing continental areas seems much less likely.
3. Vestigial Bright Plumage
Monochromatic species in five of the six major clades of Anas (all except the green-wing clade) show evidence of vestigial features of the bright dichromatic plumage of their Northern relatives. For example, some males in several dull species in the mallard clade have hints of a green head, maroon breast, curled tail feathers, and black rump (especially Hawaiian ducks [discussed above] and to a lesser extent Mexican ducks) (Murton and Westwood 1977; Madge and Burn 1988; Livezey 1991). A particularly striking case involves the Kerguelen Island pintail in which one of several hundred males shows many of the markings typical of the Northern pintail (A. acuta) (Delacour 1956; Madge and Burn 1988; Livezey 1991). Although it is difficult to say with certainty that these hints of bright features are not incipient ornaments, several factors suggest that they are indeed vestigial. In particular, such characteristics are often expressed by a few males to minor and variable degrees, which is typical of characteristics that have lost their function. Vestigial plumage brightness provides further evidence that the widespread species are ancestral, and may be paraphyletic. Consequently, this evidence supports the accepted scenario that ancestors of several monochromatic species in most of the clades may have had dichromatic plumage similar, if not identical to, that of today's dichromatic species.
The four lines of reasoning to be discussed below suggest that bright male plumage could be lost easily, and caution that the equal gain-loss probability assumption may not hold. However, these lines of reasoning provide no specific evidence that the ancestors were indeed dichromatic.
4. Weak Female Preferences for Plumage Traits
A fourth line of evidence, which is consistent with ancestral dichromatism, comes from my work on mallard mating preferences. Female mallards showed strong significant preferences for naturally bright yellow-green bills (explaining 24% of the variance in male mating success). However, I found no significant preferences for any of the eight individual plumage traits (Omland 1996). In addition, Sorenson and Derrickson (1994) did not find preferences for most plumage ornaments in Northern pintails. Weak female preferences may lead to the evolution and maintenance of bright plumage in large panmictic populations (Omland 1996). However, genetic drift, inbreeding, and few choices among males may overcome this weak selection pressure in founder populations undergoing peripatric speciation. Mayr (1942) thought that drift played some role in the loss of dichromatism on islands, and Peterson (1996) concluded that drift probably caused the repeated loss of dichromatism that he detected across all birds. But while plumage dichromatism is repeatedly lost, bill patterns and colors are evolutionarily conserved in several lineages (Madge and Burn 1988). The strongly preferred yellow-green bill of the male mallard is retained apparently in males of its three North American relatives, the black, mottled and Mexican ducks (A. rubripes, A. fulvigula, and A. diazi, respectively; Omland 1996).
5. Dichromatic Characteristics May Be Lost Easily
A fifth line of evidence that is consistent with ancestral dichromatism is that elaborate plumage in ducks may be lost more easily than gained because of the genetic basis of sex-limited traits (also see Peterson 1996). For example, every male mallard carries the genes to produce cryptic plumage, since females of that species produce dull plumage. This is the case for any sex-limited trait. The sex-determining mechanism triggers some developmental switch that leads to differential expression of ornaments in the two sexes. In ducks this switch seems to be the presence or absence of female hormones that prevent or allow the expression of elaborate plumage (Owens and Short 1995). That male ducks in dichromatic species produce female-like plumage as juveniles, and during the late summer as adults, makes it even more apparent that adult males have the genetic and physiological capability to produce cryptic plumage (Johnsgard 1961). In such species the elaborate sex can evolve crypsis by retaining the juvenile phenotype (i.e., neoteny). Such a mechanism has been suggested for the loss of bright plumage in several waterfowl lineages (Livezey 1990, 1991). This type of neoteny would fit a common pattern, in which the final part of a developmental sequence is eliminated (i.e., terminal deletion [e.g., see Gould 1977]). Developmental or genetic factors may often cause biases or constraints on the direction of evolution that are not related to selection (Wake 1991; Cunningham et al. 1997). Such biases should not be ignored when reconstructing character evolution.
6. Complex Characters May Be Lost Easily
Another reason that elaborate male duck plumage may be lost more easily than gained is that it is a complex character. Easy loss may apply especially to complexity that involves characteristics composed of several functionally related elements. A mutation affecting any of the critical elements will remove the character. Maddison (1994) suggested such an argument for the loss of flight, which has occurred independently many times in birds, apparently without subsequent regain (Lande 1978; Livezey 1989, 1990). Other complex characters for which there is compelling evidence for multiple independent losses are: limbs in several groups of vertebrates (e.g., Lande 1978), the feeding apparatus of direct developing larvae in several groups of invertebrates (Strathmann 1978; Wray and Raft 1991), and medusae in several hydroids (e.g., Cunningham and Buss 1993).
Similar arguments may extend to two or more features that are mechanistically related, although perhaps not functionally related. In the mallards and other ducks the genetic basis of the multiple discrete visual traits is not known, but it seems that a number of different loci must be involved. However, shared hormonal switches unite these ornaments (see discussion above and Owens and Short 1995). De novo evolution of each of the independent ornaments is less likely than all being turned off simultaneously, such as by a single hormonal switch. Once turned off, the likelihood that these could be re-expressed may decrease rapidly. Maddison (1994) discusses these issues in more detail.
7. Repeated Loss of Dichromatism in Other Birds
Two other concurrent studies of birds revealed repeated losses of dichromatism. Price and Birch (1996) found that dichromatism had been lost three times more frequently than gained across all passerines. Peterson (1996) surveyed geographic variation in dichromatism in all birds, and found dichromatism was lost about five times more frequently than gained. These two studies suggest that there may be biases toward the easy loss of dichromatism in all birds, which are likely acting in ducks as well (see detailed discussion below).
Repeated Loss of Dichromatism in Dabbling Ducks
These seven lines of evidence (population-level studies, biogeography, vestigial plumage, weak female preferences, easy loss of complex characters, easy loss of dichromatic characters, and repeated loss of dichromatism in other birds) together provide compelling support for the widely accepted scenario. This evidence suggests that dichromatism has been lost repeatedly in multiple lineages of dabbling ducks, probably during peripatric speciation. Ducks seemed an ideal group for employing character state reconstruction and the comparative method; there was a well-resolved comprehensive phylogenetic hypothesis for a group with interesting character states and repeated transformations among species. Two sources of error likely led to problems with the reconstruction based on simple parsimony: paraphyletic species and unequal gain-loss probabilities. Failure to account for either of these situations would likely be sufficient to result in unreliable reconstructions.
The Pitfalls and Powers of Paraphyly
Detailed population-based sampling and molecular analyses are required to resolve the evolutionary histories of each of the duck species groups (see Melnick et al. 1993; Cooper et al. 1996). If these clades indeed diversified by founder effect speciation, then such sampling should recover paraphyletic patterns in each of the species groups. This would confirm that the widespread dichromatic species violated the assumption of monophyletic terminal taxa. Graybeal (1995) points out that such ancestral paraphyletic groups, which she refers to as ferespecies, should not be used as terminal taxa in cladistic studies (see also Donoghue 1985; Olmstead 1995). In ducks and other groups in which founder effect speciation is suspected, detailed population-level sampling of multiple individuals from the putative parental ranges is needed (e.g., see Funk et al. 1995; Olmstead 1995).
While paraphyletic species are problematic on one hand, evidence that a taxon is paraphyletic provides a strong basis for inferring that its character states represent the ancestral condition. Paraphyletic taxa thus present an under-appreciated opportunity for reconstructing ancestral states with greater certainty. Molecular data have revealed paraphyletic species in many groups (e.g., birds, Peterson 1992; primates, Melnick et al. 1993; Hoelzer and Melnick 1994; rodents, Patton and Smith 1994; insects, Funk et al. 1995; Shaw 1996). Several papers have argued that peripatric speciation, which leads to the simple sorts of paraphyly discussed here, may be common (e.g., see Chesser and Zink's  response to Lynch ; Zink and McKitrick 1995). Moreover, others have suggested that paraphyletic species may be the rule rather than the exception (e.g., Larson et al. 1981; Rieseberg and Brouillet 1994; see also Sosef 1997). Although some authors have pointed out the difficulty in inferring that one species is ancestral to another (e.g., Engelmann and Wiley 1977), several cases of persistent ancestors are well documented (e.g., Theriot 1992).
If intensive population-level molecular phylogenies of ducks continue to support the widely accepted speciation scenario, then these new trees will be less likely to reconstruct monochromatic ancestors, even with equally weighted transformation probabilities. For example, if each mitochondrial haplotype found by Cooper et al. (1996) [ILLUSTRATION FOR FIGURE 2 OMITTED] is used as a terminal taxon, then the basal node for the mallard-Hawaiian duck clade is ambiguous (using equally weighted parsimony; mallard coded as dichromatic, other three species as monochromatic). Making losses only 1.1 times more likely than gains is sufficient to reconstruct "mallard-dichromatism" as ancestral for the mallard-Hawaiian duck clade. Since mallard dichromatism seems to involve so many independent elements (that show no gross differences in expression throughout the world), the other possible reconstruction positing multiple evolution of "mallard dichromatism" seems unlikely. A population-based phylogeny for the whole genus or tribe might also help provide a much better estimate of the actual transformation probabilities between monochromatism and dichromatism.
Altering the Equal Gain-Loss Probability Assumption in Parsimony Reconstruction
Four lines of evidence suggest that dichromatic plumage may be lost more easily than gained in ducks. Unfortunately many investigators interested in mapping ancestral states may work on similar character types: complex and/or sex-limited characters that may be subject to relatively easy loss. Research may be biased toward features that catch our attention, and these are likely to be complex characters, from elaborate ornaments to interesting physiological traits and life history strategies. For example, character state reconstruction has been used in many other studies of sexual dimorphism (e.g., Basolo 1990; Ryan et al. 1990; Bjorklund 1991). (In the case of the former study on swordtails, there is some evidence of vestigial expression of the sword, and this may parallel the duck example [Winquist et al. 1991; but see Basolo 1991]). A substantial portion of the new comparative method research program may be directed toward characteristics for which the equal transformation-probability assumption does not hold. Therefore, at the very least, the equal gain-loss assumption should be explicitly addressed and carefully evaluated whenever ancestral reconstruction is attempted.
Whenever there are strong reasons to predict that characteristics are prone to easy loss, alternative transformation matrices can be used. Schultz et al. (1996) showed that failing to account for transformation biases increased the probability of error in ancestral reconstructions. Accounting for such biases is accomplished readily using transformation matrices in MacClade (Maddison and Maddison 1992). (Donoghue and Ackerly  mention two cases in which unequal gainloss weighting did not alter results significantly.) Ideally, a weighting scheme should be adopted prior to phylogeny construction and character mapping to avoid subjective post-hoc justification. Such weighting should be accomplished by generalizing transformation probabilities from similar characters in related taxa, or on a model of the genetic or developmental basis of the character. For example Price and Birch's (1996) estimate that losses of dichromatism are three times more likely than gains would be a good value to apply to closely related nonpasserines. Peterson (1996) also found losses of dichromatism much more likely than gains across all birds (see below). These findings on avian dichromatism also may be generalizable to dimorphism in nonavian groups. Although the gain:loss transformation weightings of 1:3 for passerines and 1:5 for ducks may not be exactly correct, either may be more reasonable than a one-to-one weighting, and sensitivity analysis can be conducted with a range of values in this neighborhood.
Frequent Loss of Dichromatism in Birds
Three different studies of dichromatism in birds (all birds, Peterson 1996; passerines, Price and Birch 1996; dabbling ducks, present study) all found that dichromatism has been lost more frequently than it has been gained. These three studies were conducted independently during the same years, and although the taxa included only overlap slightly, all three reached strikingly similar conclusions. Price and Birch (1996) estimated that dichromatism has been lost three times more frequently than it has been gained in passerine birds. The present study on ducks suggests that it might be five times easier to lose dichromatic plumage than gain it. Across all birds, Peterson (1996) found that males were five times more likely to lose bright plumage than gain it, and females were five times more likely to gain bright plumage than lose it.
Peterson (1996) points out that his results reflected the fact that most of the species that show geographic variation are dichromatic initially. Peterson's study emphasizes that either males can change, leading to monochromatic dull species, or females can change, leading to monochromatic bright species. A flycatcher (Petroica multicolor) that is dichromatic on mainland Australia provides a particularly illustrative example of this. Nine of the surrounding islands have dichromatic races, but interspersed among these islands are two islands with monochromatic bright races, and two islands with monochromatic dull races (Mayr 1942, p. 49; also see Futuyma 1986, p. 239). Dabbling ducks also provide at least one example of a monochromatic bright species; in the Chiloe wigeon (A. sibilatrix) both sexes look similar to male American wigeons (A. americana; Madge and Burn 1988). The fact that loss of dichromatism occurs in both directions makes it less likely that selection can explain the majority of cases. In particular, the lack of a need for species recognition mechanisms (e.g., Mayr 1942; Sibley 1957) cannot explain why isolated species would become monochromatic bright, or why some continental species with overlapping ranges seem to become monochromatic dull (e.g., A. zonorhyncha, A. rubo ripes, [ILLUSTRATION FOR FIGURE 3 OMITTED]; also see Peterson 1996). Genetic drift and other factors affecting founder populations, combined with biases toward the loss of sex-limited and complex characters, likely caused repeated losses of dichromatism in ducks and other birds.
This study examined two assumptions that can lead to errors in ancestral state reconstruction: paraphyletic species and unequal probabilities for the gain and loss of dichromatism, a bias that seems widespread in birds. The possible occurrence of paraphyletic species needs to be considered whenever species-level phylogenies are estimated, or are used to reconstruct character evolution. Population-level sampling aimed at detecting and accounting for paraphyletic species is especially warranted in situations in which peripatric speciation is suspected. This study emphasized the situation in which widespread continental species likely have given rise to species in insular and other peripheral ranges, but analogous scenarios may occur in other contexts (e.g., host shifts in insects, Funk et al. 1995). Species monophyly should be critically assessed whenever possible; monophyly at the species level is crucial to reliable tree-building and character state reconstruction, just as it is at higher levels (e.g., Graybeal 1995).
In this study I advocate using step matrices to account for known or suspected biases in transformation probabilities. Some might suggest that by abandoning the assumption of equal transformation costs, one is starting along a slippery slope toward subjectivity. However, when assuming equal weights researchers are already on that slope. Swofford and Maddison write:
In general, we accept the use of "simple" assumptions-unordered character states and equal costs for all transformations as a suitable starting point. . . . However, uncertainty as to whether [transformation] weighting should be 1.5:1, 3:1 or 10:1 does not imply that a 1:1 weighting is more objective . . . (Swofford and Maddison 1992, p. 216).
Easy loss of complex and/or dichromatic characters may cause a general evolutionary trend toward the repeated loss of dichromatism in birds that is not caused by selection. Other researchers have also suggested that nonadaptive factors may cause loss of dichromatism in birds (Grant 1965; Livezey 1991; Peterson 1996). Repeated evolutionary patterns are usually interpreted as strong evidence of adaptation (i.e., convergence due to selection). However, increasingly evolutionary biologists are pointing out that biases due to genetic and developmental constraints can also cause homoplasy (e.g., Wake 1991). Such biases should be accounted for when reconstructing evolutionary history.
Further evaluation of the reliability of character state reconstruction methods should prove useful for assessing the effects of the pitfalls explored in this paper, as well as other issues. More studies of congruence between phylogenetic reconstructions and other sources of information on ancestral states would be a logical next step. Such evaluations could be based on studies of characters with detailed fossil records (e.g., Theriot 1992), or studies on groups with known ancestors (e.g., Atchley and Fitch 1991; Hillis et al. 1992; Cunningham et al. 1997). New maximum likelihood methods (Pagel 1994; Schluter 1995) show promise for reconstructing character evolution, but the challenge will be similar to that facing generalized parsimony. In either case more realistic models of character evolution are needed to reconstruct reliably the evolutionary history of morphological, behavioral, and physiological characters.
K. Able, R. Benson, J. Brown, A. Jacklet, S. Lanyon, B. Livezey, M. McKitrick, M. Murphy, J. Podos, S. Scheffer, R. Sikes, C.-B. Stewart, H. Wiley, and P. Wilson all provided helpful suggestions. C. Cunningham and the Duke University Zoology Department provided support and encouragement during the final writing. While doing this work I was supported in part by grants from the Delta Waterfowl Foundation and the E. N. Huyck Preserve (NY). T Price and A. T Peterson kindly gave me a copies of then unpublished manuscripts. Finally, I am indebted to B. Livezey whose detailed phylogenetic hypothesis of the Anatini made this paper possible.
ATCHLEY, W. R., AND W. M. FITCH. 1991. Gene trees and the origins of inbred strains of mice. Science 254:554-558.
AVISE, J. C., C. D. ANKNEY, AND W. S. NELSON. 1990. Mitochondrial gene trees and the evolutionary relationship between mallard and black ducks. Evolution 44:1109-1119.
BASOLO, A. L. 1990. Female preference predates the evolution of the sword in swordtail fish. Science 250:808-810.
-----. 1991. Male swords and female preferences: response to Winquist et al. Science 253:1426-1427.
BJORKLUND, M. 1991. Coming of age in fringillid birds: heterochrony in the ontogeny of secondary sexual characteristics. J. Evol. Biol. 4:83-92.
BOUSQUET, J., S. H. STRAUSS, AND P. LI. 1992. Complete congruence between morphological and rbcL-based molecular phylogenies in birches and related species (Betulaceae). Mol. Biol. Evol. 9:1076-1088.
BROOKS, D. R., AND D. A. MCLENNAN. 1991. Phylogeny, ecology, and behavior: a research program in comparative biology. Univ. of Chicago Press, Chicago.
CHESSER, R. T., AND R. M. ZINK. 1994. Modes of speciation in birds. Evolution 48:490-497.
COLLINS, T. M., P. H. WIMBERGER, AND G. J. P. NAYLOR. 1994. Compositional bias, character-state bias, and character-state reconstruction using parsimony. Syst. Biol. 43:482-496.
COOPER, A., J. RHYMER, H. F. JAMES, S. L. OLSON, C. E. MCINTOSH, M. D. SORENSON, AND R. C. FLEISCHER. 1996. Ancient DNA and island endemics. Nature 381:484.
CUNNINGHAM, C. W., AND L. W. BUSS. 1993. Molecular evidence for multiple episodes of paedomorphosis in the family Hydractiniidae. Biochem. Syst. Ecol. 21:57-69.
CUNNINGHAM, C. W., K. JENG, J. HUSTI, M. BADGETT, I. J. MOLINEUX, D. m. HILLIS, AND J. J. BULL. 1997. Parallel molecular evolution of deletions and nonsense mutations in bacteriophage T7. Mol. Biol. Evol. 14:113-116.
DELACOUR, J. 1956. The waterfowl of the world. Vol. 2. The dabbling ducks. Country Life, London.
DELACOUR, J., AND E. MAYR. 1945. The family Anatidae. Wilson Bull. 57:2-55.
DONOGHUE, M. J. 1985. A critique of the biological species concept and recommendations for a phylogenetic alternative. Bryologist 88:172-181.
-----. 1989. Phylogenies and the analysis of evolutionary sequences with examples from seed plants. Evolution 43:1137-1156.
DONOGHUE, M. J., AND D. D. ACKERLY. 1996. Phylogenetic uncertainties and sensitivity analysis in comparative biology. Phil. Trans. R. Soc. Lond. B 351:1241-1249.
ENGELMANN, G. E, AND E. O. WILEY. 1977. The place of ancestor-descendant relationships in phylogeny reconstruction. Syst. Zool. 26:1-11.
FELSENSTEIN, J. 1988. Phylogenies from molecular sequences: inference and reliability. Annu. Rev. Genet. 22:521-565.
FRUMHOFF, P. C., AND H. K. REEVE. 1994. Using phylogenies to test hypotheses of adaptation: a critique of some current proposals. Evolution 48:172-180.
FUNK, D. J., D. J. FUTUYMA, G. ORTI, AND A. MEYER. 1995 A history of host associations and evolutionary diversification for Ophraella (Coleoptera: Chrysomelidae): new evidence from mitochondrial DNA. Evolution 49:1008-1017.
FUTUYMA, D. J. 1986. Evolutionary biology. Sinauer, Sunderland, MA.
GOULD, S. J. 1977. Ontogeny and phylogeny. Belknap Press, Cambridge, MA.
GRANT, P. R. 1965. Plumage and evolution of birds on islands. Syst. Zool. 14:47-52.
GRAYBEAL, A. 1994. Evaluating the phylogenetic utility of genes: a search for genes informative about deep divergences among vertebrates. Syst. Biol. 43:174-193.
-----. 1995. Naming species. Syst. Biol. 44:237-250.
HARVEY, P. H., AND M.D. PAGEL. 1991. The comparative method in evolutionary biology. Oxford Univ. Press, Oxford.
HILLIS, D. M., J. M. BULL, M. E. WHITE, M. R. BADGETT, AND I. J. MOLINEUX. 1992. Experimental phylogenetics: generation of a known phylogeny. Science 255:589-592.
HILLIS, D. M., J.P. HUELSENBECK AND C. W. CUNNINGHAM. 1994. Application and accuracy of molecular phylogenies. Science 264:671-677.
HOELZER, G. A., AND D. J. MELNICK. 1994. Patterns of speciation and limits to phylogenetic resolution. Trends Ecol. Evol. 9:104-107.
JOHNSGARD, P. A. 1961. Evolutionary relationships among the North American mallards. Auk 78:3-43.
-----. 1965. A handbook of waterfowl behavior. Cornell Univ. Pres's, Ithaca, NY.
KEAR, J. 1970. Adaptive radiation of parental care in waterfowl. Pp. 357-392 in J. H. Crook, ed. Social behavior in birds and mammals: essays on the social ethology of animals and man. Academic Press, New York.
KEAR, J., AND R. K. MURTON. 1976. The origins of Australian waterfowl as indicated by their photoresponses. Pp. 83-87 in H. J. Frith and J. H. Calaby, eds. Proceedings of the 16th international ornithological congress. Australian Academy of Science, Canberra, Australian Capital Territory.
KESSLER, L. G., AND J. C. AVISE. 1984. Systematic relationships among waterfowl (Anatini) inferred from restriction endonuclease analysis of mitochondrial DNA. Syst. Zool. 33:370-380.
LANDE, R. 1978. Evolutionary mechanisms of limb loss in tetrapods. Evolution 32:73-92.
LANYON, S. M. 1992. Interspecific brood parasitism in blackbirds (Icterinae): a phylogenetic perspective. Science 255:77-79.
LARSON, A., D. B. WAKE, L. R. MAXSON, AND R. HIGHTON 1981. A molecular phylogenetic perspective on the origins of morphological novelties in the salamanders of the tribe Plethodontini (Amphibia, Plethodontidae). Evolution 35:405-442.
LIVEZEY, B.C. 1986. A phylogenetic analysis of Recent anseriform genera using morphological characters. Auk 103:737-754.
-----. 1989. Flightlessness in grebes (Aves, Podicipedidae)' its independent evolution in three genera. Evolution 43:29-54.
-----. 1990. Evolutionary morphology of flightlessness in the Auckland Islands Teal. Condor 92:639-673.
-----. 1991. A phylogenetic analysis and classification of recent ducks (Tribe Anatini) based on comparative morphology. Auk 108:471-508.
-----. 1993. Comparative morphometrics of Anas ducks, with particular reference to the Hawaiian duck Anas wyvilliana, Laysan duck A. laysanensis, and Eaton's pintail A. eatoni. Wildfowl 44:75-100.
LORENZ, K. 1971. Comparative studies of the motor patterns of Anatinae. Pp. 14-113. in R. Martin, trans. Studies in animal and human behavior. Harvard Univ. Press, Cambridge, MA.
LYNCH, J. D. 1989. The gauge of speciation: on the frequencies of modes of speciation. Pp. 527-553. in D. Otte and J. A. Endler, eds. Speciation and its consequences. Sinauer, Sunderland, MA.
MADDISON, D. R. 1994. Phylogenetic methods for inferring the evolutionary history and processes of change in discretely valued characters. Annu. Rev. Entomol. 39:267-292.
MADDISON, W. P. 1995. Calculating the probability distributions of ancestral states reconstructed by parsimony on phylogenetic trees. Syst. Biol. 44:474-481.
MADDISON, W. P., AND D. R. MADDISON. 1992. MacClade: analysis of phylogeny and character evolution. 3.04. Sinauer, Sunderland, MA.
MADGE, S., AND H. BURN 1988. Waterfowl. Houghton Mifflin Co., Boston, MA.
MAYR, E. 1942. Systematics and the origin of species. Dover, New York.
-----. 1963. Animal species and evolution. Belknap Press, Cambridge, MA.
MCKITRICK, M. C. 1993. Phylogenetic analysis of avian parental care. Auk 109:828-846.
MELNICK, D. J., G. A. HOELZER, R. ABSHER, AND M. V. ASHLEY. 1993. mtDNA diversity in rhesus monkeys reveals overestimates of divergence time and paraphyly with neighboring species. Mol. Biol. Evol. 10:282-295.
MURTON, R. K., AND N.J. WESTWOOD. 1977. Avian breeding cycles. Clarendon Press, Oxford.
OLMSTEAD, R. G. 1995. Species concepts and plesiomorphic species. Syst. Bot. 20:623-630.
OMLAND, K. E. 1994. Character congruence between a molecular and a morphological phylogeny for dabbling ducks (Anas). Syst. Biol. 43:369-386.
-----. 1996. Female mallard preferences for multiple male ornaments. I. natural variation. Behav. Ecol. Sociobiol. 39:353-360.
OWENS, I. P. E, AND R. V. SHORT 1995. Hormonal basis of sexual dimorphism in birds: implications for new theories of sexual selection. Trends Ecol. Evol. 10:44-47.
PAGEL, M. 1994. Detecting correlated evolution on phylogenies: a general method for the comparative analysis of discrete characters. Proc. R. Soc. Lond. B Biol. Sci. 255:37-45.
PALMER, R. S. 1976. Handbook of North American birds. Vol. 2. Waterfowl (Part 1). Yale Univ. Press, New Haven, CT.
PATTON, J. L., AND M. F. SMITH. 1994. Paraphyly, polyphyly and the nature of species boundaries in pocket gophers (genus Thomomys). Syst. Biol. 43:11-26.
PETERSON, A. T. 1992. Phylogeny and rates of molecular evolution in the Aphelocoma jays (Corvidae). Auk 109:133-147.
-----. 1996. Geographic variation in sexual dichromatism in birds. Bull. Brit. Ornithol. Club 116:156-172.
PRICE, T. D., AND G. k. BIRCH 1996. Repeated evolution of sexual color dimorphism in the Passerine birds. Auk 113:842-848.
RIESEBERG, L. H., AND L. BROUILLET. 1994. Are many plant species paraphyletic? Taxon 43:21-32.
RIPLEY, S. D. 1957. Book review - the waterfowl of the world. Jean Delacour. Auk 74:269-272.
RYAN, M. J., J. H. Fox, W. WILCZYNSKI AND A. S. RAND. 1990. Sexual selection for sensory exploitation in the frog, Physalaemus pustulosus. Nature 343:66-67.
SCHLUTER, D. 1995. Uncertainty in ancient phylogenies. Nature 377:108-109.
SCHULTZ, Z. R., R. B. COCROFT, AND G. A. CHURCHILL. 1996. The reconstruction of ancestral character states. Evolution 50:504-511.
SCOTT, D. K. AND T. H. CLUTTON-BROCK. 1989. Mating systems, parasites and plumage dimorphism in waterfowl. Behav. Ecol. Sociobiol. 26:261-273.
SHAW, K. L. 1996. Sequential radiations and patterns of speciation in the Hawaiian cricket genus Laupala inferred from DNA sequences. Evolution 50:237-255.
SIBLEY, C. G. 1957. The evolutionary and taxonomic significance of sexual dimorphism and hybridization in birds. Condor 59: 166-187.
SILLEN-TULLBERG, B., AND A. P. MOLLER. 1993. The relationship between concealed ovulation and mating systems in anthropoid primates: a phylogenetic analysis. Am. Nat. 141:1-25.
SORENSON, L. G., AND S. R. DERRICKSON. 1994. Sexual selection in the northern pintail (Arias acura): the importance of female choice versus male-male competition in the evolution of sexually selected traits. Behav. Ecol. Sociobiol. 35:389-400.
SOSEF, M. S. M. 1997. Hierarchical models, reticulate evolution, and the inevitability of paraphyletic supraspecific taxa. Taxon 46:75-85.
STEWART, C.-B., J. W. SCHILLING AND A. C. WILSON. 1987. Adaptive evolution in the stomach enzymes of foregut fermenters. Nature 330:401-404.
STRATMANN, R. R. 1978. The evolution and loss of feeding larval stages of marine invertebrates. Evolution 32:894-906.
SWOFFORD, D. L. 1993. PAUP: phylogenetic analysis using parsimony. 3.1.1. Smithsonian Institution, Washington, DC.
SWOFFORD, D. L., AND W. P. MADDISON. 1992. Parsimony, character-state reconstructions, and evolutionary inferences. Pp. 186-223 in R. C. Mayden, ed. Systematics, historical ecology, and North American freshwater fishes. Stanford Univ. Press, Palo Alto, CA.
SWOFFORD, D. L., G. J. OLSON, E J. WADDELL, AND D. M. HILLIS. 1996. Phylogenetic inference. Pp. 407-514 in D. M. Hillis, C. Moritz, and B. K. Mable, eds. Molecular systematics. 2d ed. Sinauer, Sunderland, MA.
THERIOT, E. 1992. Clusters, species concepts, and morphological evolution of diatoms. Syst. Biol. 41:141-157.
WAKE, D. B. 1991. Homoplasy: the result of natural selection, or evidence of design limitations? Am. Nat. 138:543-567.
WELLER, M. W. 1980. The island waterfowl. Iowa Univ. Press, Ames.
WINQUIST, S. T., D. G. WEARY, A. J. INMAN, D. J. MOUNTJOY, AND E. A. KREUS. 1991. Male swords and female preferences. Science 253:1426.
WRAY, G. A., AND R. A. RAFF. 1991. The evolution of developmental strategy in marine invertebrates. Trends Ecol. Evol. 6:45-50.
ZINK R. M., AND M. C. McKITRICK. 1995. Debate over species concepts and its implications for ornithology. Auk 112:701-719.
|Printer friendly Cite/link Email Feedback|
|Author:||Omland, Kevin E.|
|Date:||Oct 1, 1997|
|Previous Article:||Testing for unequal amounts of evolution in a continuous character on different branches of a phylogenetic tree using linear and squared-change...|
|Next Article:||North American black bear mtDNA phylogeography: implications for morphology and the Haida Gwaii glacial refugium controversy.|