Tt,r Condor 102:52-59 0 The Cooper Ornithological Society 2000

THE PHYLOGENETIC APPROACH TO AVIAN LIFE HISTORIES: AN IMPORTANT COMPLEMENT TO WITHIN-POPULATION STUDIES ’

DAVID W. WINKLER Department of Ecology and EvolutionaryBiology, CorsonHall, Cornell UniversiQ, Ithaca, NY 14853.2701, e-mail: dww4Qcornell.edu

Abstract. In recent years, two approaches have emerged for the analysis of character evolution: the largely statistical “convergence” approach and the mainly cladistic “homol- ogy” approach. I discuss the strengths and weaknesses of these approaches as they apply to phylogenetic analyses of life-history variation in . Using examples from analyses of character variation in , I suggest that the phylogenetic approach yields distinctive insights into the selective role of the environment and other characters of the organism on the evolution of life-history traits. This view thus has the potential of bringing together micro- and macro-evolutionary views of life-history evolution. Key words: birds, character evolution,life-history traits, phylogeny,swallows

INTRODUCTION ganismal biology, and ecologists sometimes use The life histories of birds are fascinating because similar comparisons among sites to evaluate the their study requires an integration of organismal roles of various ecological processes.In the past and population biology. Here I relate some of 10 years or so, however, comparative studies my experiences in taking the phylogenetic ap- have often been thought of as having an explicit proach to studying the evolution of avian life phylogenetic component, and it is to that nar- histories. I will mention briefly many dynamic rower set of comparative evolutionary methods areas: systematicsis growing and changing rap- for analyzing adaptation and character change idly, accumulating an immense literature and a that I limit my attention. great variety of ideas; comparative studies of the ecological characteristicsof birds are increasing AN OVERVIEW OF THE METHODS rapidly as well. My intention is to discussbriefly Two principal approachesto analyzing character the ways that comparative methods can be ap- evolution have found broad application during plied to character evolution. I then give a few the last decade (Coddington 1994, Page1 1994, examples of how I have applied these methods Sheldon and Whittingham 1997). There is the to life-history traits in swallows. Finally, I con- convergence approach (also called the “func- clude with a brief detour on another life-history tional” or “homoplasy” approach) and the ho- trait, dispersal, which is suitable for analysis on mology approach (also called the “cladistic” several levels, including a comparative one. My approach). The convergence approach explores goal is to demonstrate that comparative analysis correlations among similar events across line- provides an important tool for the study of life ages. Methods have been developed by Ridley histories, as it provides the foundation for com- (1983), Felsenstein (1985), Grafen (1989), Har- bining organismal and population perspectives. vey and Page1 (1991) Maddison (1990), Gittle- Before proceeding, I must clarify the sensein man and Kot (1990) and Garland et al. (1992, which “comparative” is and has been used in 1999). These, and many other comparative biology. For decades, comparative anatomists methods that have been developed in the last 15 have used comparisons among taxa as natural years, have the common feature that they at- experiments to evaluate the mechanisms of or- tempt to draw inferences about the adaptive na- ture of charactervariation. They look across lin- 1Received 8 February 1999. Accepted14 October eages at repeated changes in charactersand cor- 1999. relate these with changes in ecological circum- stances or with changes in other characters. Because these methods require statistical as- sumptions about the evolutionary process that are difficult to evaluate, I prefer the homology Molothrus bonarlenrls (176) approach, which explores unique events within lineages. Homology methods are those of Wann- torp (1983), Padian (1985), Greene (1986) Cod- dington (1988), Carpenter (1989), Donoghue (1989), Lauder (1990), and Baum and Larson (1991)-many of whom are cladistic systema- Scaphldura oryzivora (7) tists as well. The homology approach explores unique events within lineages. Many conver- gence comparative methods have been created Molothrus rufoaxlllarls (1) to solve the problem of pseudoreplication across taxa in studies of character evolution, and one could reasonably argue that the homology ap- proach, instead of dealing with pseudoreplica- FIGURE 1. Scott Lanyon’s phylogeny of cowbirds tion, eschews any replication at all. Its greatest with his compendium of the number of hosts possessed by each in parentheses for each taxon (after disadvantage is that it deals with single events. Lanyon 1992). There is no pretense of a statistical interpretation of the variation in characters that we see. This approach does, however, have compensating based on mitochondrial DNA variation and then merits. mapped onto that tree the number of host species Before exploring these, I will present a few used by each of the cowbird species analyzed. examples and a few speculations on the history He found an obvious trend within the cowbirds of these approaches as they pertain to life-his- from basal species with few hosts to derived tory variation. I see the convergence approaches species with numerous hosts (Fig. 1). At first in as arising out of comparative an- sight, this pattern of occurrence of host species alyses used by David Lack (1968) and Oscar acrossthis phylogeny contradicts the predictions Heinroth (1922) before him, in which mean life- of the coevolutionary scenarios entertained by history traits acrossspecies were plotted relative many evolutionary ecologists. According to to some other feature of the organism or to some these scenarios, as species coexisted with their ecological circumstance. Phylogeny was usually hosts for longer periods of time, counteradapta- dealt with graphically by plotting speciesin dif- tions in the host would reduce the number of ferent taxa with different symbols and connect- hosts that the parasite could utilize successfully ing them with different lines. So, in a sense, over evolutionary time. Rothstein et al. (1997) Lack and Heinroth attempted to control for phy- criticized Lanyon’s interpretation, pointing out logeny by highlighting the relationships that po- that more primitive taxa could have been in con- tentially differ between different groups. The tact with their host species for a longer period more modern approacheshave dealt with phy- of time, thus losing host speciesas predicted by logeny either by analyzing charactersaccording the co-evolution hypothesis. Regardless of its to the taxonomic rank of the groups possessing outcome, this issue has stimulated a healthy de- them, or by using phylogenetically based statis- bate, which emphasizes the validity of the ho- tics (Felsenstein 1985, Garland et al. 1993, Mar- mology approach. It causesus to focus carefully tins and Hansen 1997). on the characters of interest and on the lineage Perhaps the earliest application of a homolo- of interest and to think logically and explicitly gy-like approach to non-morphological traits in about what kinds of forces may have produced ornithology was Lorenz’s (1941) attempt to an- the observed character distribution in that line- alyze the evolution of waterfowl courtship. A age. modern example of the homology approach Thus far I have glossed over the importance comes from Lanyon’s (1992) work on brood par- of constructing a phylogenetic tree. No evolu- asitism in cowbirds (Molothrus). Lanyon con- tionary comparative analysis can be any better structed a phylogenetic tree for the cowbirds than the phylogenetic tree upon which it is based. Unfortunately, the construction of a tree ism, it is valuable to consider the ecologies of is not easy. There are many different types of its relatives. Swallows are an excellent group in data, each of which requires different assump- which to interpret life histories using compara- tions in its analysis and different analytical tive methods. They are extremely conservative methods. Cladistic methods involve many im- in their overall morphology and foraging ecol- portant decisions about character coding and ogy, but they are quite diverse in their nesting weighting and combination of data sets, and the biology. Beginning in the mid-1980s I have methods are being refined all the time. Also, in- been working with Frederick Sheldon to study creasing numbers of non-cladistic methods of in- the phylogenetic relationships of swallows using creasing sophistication are being applied, espe- DNA-hybridization (Sheldon and Winkler cially to molecular data. In short, if you need a 1993), and we have been joined in mitochondrial tree and are not a systematist,I suggestthat you sequencing work by Linda Whittingham (Shel- get help. However, I encourage population bi- don et al. 1999). ologists and behavioral ecologists to get in- Once we started getting phylogenetic trees, volved in the process of tree construction be- the phylogenetic hypothesesallowed us to make cause it can be extremely informative. Not only inferences about ancestral character states and is it interesting in its own right, but anyone who the history of character change. There are three has worked on building a tree understands the main clades of swallows (Fig. 2): (1) the African limitations of that tree. sawwings, (2) what we call the “core martins,” One more caveat should be made clear: phy- and (3) a group consisting of and its logenetic hypothesesare historical. Each group allies. Nest construction is more variable in the of organisms has only one history, and the best swallows than it is among any other of we can do is to construct a hypothesis for its oscine birds; the nest construction of phylogeny. We can never subject phylogenetic each of the terminal taxa is indicated on Figure hypotheses to experimental tests in the sense 2. Species of swallows range from burrowers that we can test ecological hypotheses.We can through a large group of cavity-adopters to mud- add new information to an analysis to improve nesters that build either a simple cup, a closed the resolution or robustness of a phylogenetic cup, or a retort-shapednest. Figure 2 also shows tree or of a character mapping. But ultimately a reconstruction of ancestral charactersfor nest we cannot do a manipulative experimental test. construction. A sense of how ancestral charac- Equally frustrating for population biologists is ters are inferred can be gained by considering the fact that phylogenetic analysis is relatively the group of Hirundo and its allies. Notice that resistant to statistical evaluation. Clearly, we these species all make nests of mud, and within could deal with evolution statistically if we un- this lineage of “mud-nesters” we have plotted derstood the evolutionary process. However, in the reconstructionsto indicate that cup-construc- a phylogenetic tree of any given size, it is likely tion was the primitive state which gave rise to that the blend of potential causative factors for forms that made closed cups, which in turn gave character evolution has been different within ev- rise to forms that made retorts. The strategy of ery single lineage it contains. Thus, the search ancestral character-state reconstruction is to for one common statistical model of evolution minimize the number of changes in character for any group of organisms may be severely state on the tree while producing the observed handicapped from the start. We may just need to distribution of character states among the ter- accept from the outset that we will never know minal taxa. In this particular case, for instance, absolutely the history of a group or its charac- the same number of steps (i.e., two steps from ters, and we need to be cautious in our phylo- a primitive state to closed cup and retort) is ob- genetic interpretations. tained by assuming that the ancestral was a retort: that a retort-maker gave rise to a cup- LIFE-HISTORY VARIATION IN nester; that a retort-maker also gave rise to a SWALLOWS closed-cup nester; and that the retort nest in the For many years I have been interested in swal- derived species represents the retention of the lows, most specifically Tree Swallows (Tachy- primitive character state. To circumvent this am- cineta bicolor), but I have always felt that, in biguity of character state reconstruction, we ob- order to understand the ecology of one organ- served that retort-nesting swallows, when they pyrrhonota q 0 T

Petrochelidon spilodera q a “4

Cecropis semirufa Jest Types: qm-@j

0 Burrow urbica =: atzl*“p?F m Adoptedcavity 2 Hitundo rusfica qm+&@ w Open mud cup t Ptyonoprogne fuligula di3-q Covered mudcup < Phaeoprogne tapera ~rxl -q m Mud retort Lc chalybea 9Elo

social System: fascia& Eejjm 0

Solitary D Neochelidon tibia/is 3l5lo

Semi-colonial m Nofiochelidon cyanoleuca 3 m qq

Colonial Is Ste/gidoptetyxrufico//is 3 m 7

:eather Use: bicolor Am

Extensive I -----_l‘ riparia ZIBI Moderate -4 I Riparia cincta am Flare 9 , Pseudhrundo gnseopyga 3 m None Reported 0 1 Psabdoprocne holomelas 3 D 0

FIGURE 2. Sheldon and Winkler’s (1993) DNA-hybridization phylogeny of most of the putative genera of swallows, with the type of nest built by each taxon as well as their putative ancestors. Terminal character states for nest sociality and feather nest-lining, respectively, are added to the right of the terminal taxa. English names of taxa in tree, from bottom to top, are: Black Sawwing, Gray-rumped , , (Bank Swallow), , Southern Rough-winged Swallow, Blue-and-white Swallow, White-thighed Swallow, White-banded Swallow, Gray-breasted Martin, Brown-chested Martin, , , House Martin, Rufous-chested Swallow, South African , Cliff Swallow. build their nests, proceed from a cup-shapednest is an active area of research for us. Here the stage to a closed-cup stage, and then finish with homology approach becomes very interesting: a retort (Winkler and Sheldon 1993). We as- when we can examine the covariation of several sumed that nest evolution proceededin the same traits on a tree and consider whether changes in sequence of steps as does nest construction to- one character state may have been antecedent to day. Character state reconstruction can be much changes in another, we can begin to draw con- more complicated (Cunningham et al. 1998), clusions about the relative adaptive linkages be- even for nests (Sheldon and Winkler 1999, Zys- tween these different character states (Codding- kowski and Prum 1999). ton 1988, 1994). When charactersare mapped on a phylogeny, Tree Swallows line their nests with the feath- we usually see immediately that different char- ers of other species (Winkler 1993), and map- acters vary on the tree in different ways. For ping the use of feathers as nest-lining on the tree example, Figure 2 shows ancestral character across the entire family (Fig. 2) is instructive. state reconstructions for nest type, along with Like sociality, the distribution of the use of terminal statesfor sociality. The latter character feathers in the phylogenetic tree is not concor- has states ranging from territorial solitary nest- dant with the structure of the tree. Most of the ers through semi-colonial nesters to colonial mud-nesting swallows appear to use moderate nesters. However, the distribution of these types numbers of feathers, although I think even this is not so concordant with the tree as is nest con- generalization might fall apart if we look care- struction. Specifically, highly colonial forms oc- fully at more species. Within the cavity-adopt- cur among both mud-nesters and burrowers. ers, there is generally relatively little use of Elsewhere, Winkler and Sheldon (1993) have feathers, although they are sometimes used by speculated about the origins of nest coloniality Phaeoprogne and . But in No- and how it relates to nest construction, and this tiochelidon, they are used moderately, and in Tachycineta bicolor and other Tachycineta spe- few life-history traits one can measure in a col- cies, they are used quite extensively. Within the lection. And it is one of the few life-history burrowers, feathers tend not to be used much traits for which the quantification of the trait is except in the Bank Swallow (Riparia riparia), fairly unequivocal. As soon as one attempts which uses them extensively. There are a couple comparative analyses of life-history trait varia- of points that I would like to make here. First, tion, one sees the need for more quantitative character coding can be extremely subjective, as data on life-history variation among birds-such the variation of this continuous character can be as is generated by the BBIRD program (Martin of two types: whether or not the birds use feath- et al. 1997)-replicated with appropriate mea- ers at all, and if they do use feathers, how many surements for many groups of species around feathers they use. For many species we have the world. very few data, so it is hard to assign a character There has been an important debate over state for sure. Furthermore, if feather use had whether or not characters are phylogenetically been scored in Figure 2 as a 2-state character constrained (Antonovics and van Tienderen instead of a 4-state character,the pattern of char- 1991, M&&rick 1993), and the cladistic ap- acter variation on the phylogenetic tree would proach to tree construction causesone to realize be very different. that how we think about a character depends on The phylogenetic approach forces us to think our perspective within the tree. When one views hard about characters of interest and how we the tree for swallows from the root looking out define character states. It also forces us to think toward the branch tips, nest construction appears hard about homologies across the group of in- to be a flexible character,having changed many terest. Consider the case of feather use by Tree times during the diversification of this group. Swallows and Bank Swallows, species which However, viewed from a branch tip back toward likely have independent evolutionary origins. In the root, nest construction appears quite conser- Tree Swallows, the feathers are placed in the vative, having undergone few changes, and with nest lining with the feather shafts pushed down no evidence of character reversal, over the his- into the base of the nest and the vanes of the tory of a single lineage. Once a nest character feathers overlapping so that the feathers form a has changed in this phylogeny of swallows (and discrete and continuous boundary between the it has very seldom changed), it has never gone nest contents and the grass nest itself. In well- back to its former state in the evolution of a feathered nests, the rachisesof those feathersex- lineage. So, even though nest-construction is tend up and actually curve over the top of the quite variable when viewed from the root up, nest so that it forms a canopy. Based on research from the tips-down nest-construction appears done with Scott Turner, we know that this nest conservative. So where do we stand on this no- canopy is probably very important for thermo- tion of phylogenetic constraint? There is no regulation in the nest (unpubl. data). In contrast, strong evidence at any point in this tree of bind- the Bank Swallow nest is essentially a flat plat- ing organismal constraints that have prevented form at the back of the burrow on which feathers evolutionary change in nest construction. Rather, are placed rather haphazardly; there is none of at every step along the way there has been a the secondary structure that we see in Tree blend of selective factors that has subjected nest Swallow nests. The homology approach forced construction to stabilizing selection. Occasion- me to think carefully not only about whether or ally, when swallows have colonized new habi- not the birds use feathers, but about the details tats, stabilizing selection may have turned to di- of how they use feathers. It made me realize that rectional selection and new modes of nest con- the similarity in this character is likely conver- struction may have arisen. “Phylogenetic con- gent rather than homologous. This again, to me, straint” is too-strong a phrase that summarizes is a great strength of the homology approach. the history-dependency of character-statevaria- The phylogenetic approach,and especially the tion, and only by looking at a tree can we get a homology approach, highlights the inadequacies feel for the heritage of selective environments in our knowledge of groups of interest. For life- that have affected different characters. With a history traits, it is no accident that clutch size is phylogenetic approach,one need not jump to the the trait that has been subjected to most analysis conclusion that a hard genetic constraint or lack over the years because clutch size is one of the of time for change has caused a trait’s changes to follow phylogeny. The phylogeny gives us a framework for evaluating the hypothesis that \‘ trait variation is the result of selection for “mul- .- ._Heritage 1 _/ ’ tiple adaptive peaks.” Each character has an in- dependent history, and we can view organisms -I! - ’ as a suite of characters,each of which takes its Cavity Adoption own independent evolutionary path in respond- I ing to both the environment and to changes in / Earlier Timing other characters of the organism. Thus phylog- d of Breeding enies, by teaching us about the histories of char- Aggressive acters, can help us understand connections be- Social tween life-history traits and responsesto selec- , ’ \ Behavior tive pressures. ,, Chick Survival One example of this has been extremely in- 1 and Recruitment structive to me in my studies of Tree Swallows. -\ I believe that cavity adoption, which is the mode Extensive Feather Nest Lining ,,; Clutch Size of nest construction in Tree Swallows, arose \ from an evolutionary heritage in which, for good selective reasons, nest-construction was a rela- tively conservative trait (Winkler and Sheldon 1994). Once cavity adoption existed, however, it could act as a selective influence on other char- Fitness acters. For example, I believe that cavity adop- tion selected for earlier breeding in Tree Swal- lows and more aggressive social behavior (Fig. FIGURE 3. A personal view of the connections be- tween early-season life-history traits in Tree Swallows. 3). Earlier breeding has, in turn, led to selection For details see text. for extensive feather nest-lining (for insulation) which has fed back on aggressive social behav- ior through competition for feathers. Earlier tim- other life-history trait that is of great interest: ing of breeding directly influences clutch size, dispersal. Studies to date have been microevo- chick survival and recruitment, and the feather lutionary, but dispersal also can be approached lining has implications for chick growth. These from a macroevolutionary perspective. Levey four factors together are the principal contribu- and Stiles (1992) have suggestedthat long-dis- tors to fitness variation in Tree Swallows. tance seasonalmigratory movements arose from Notice that in considering interactions be- seasonal food-driven shifts in distribution: does tween characters and the influence of their var- the same sort of gradation link aseasonal no- iation on fitness, one shifts from between-spe- madic life styles with the shorter once-in-a-life- cies comparisons to between-individual compar- time dispersal movements between natal and isons, what some regard as the difference be- breeding sites that we observe in most birds? tween macroevolution and microevolution. The The fact that extreme nomadic dispersers are transition between studying variation among limited to a few phylogenetic groups (e.g., wax- speciesin charactermeans and dissecting out in- wings, Cardueline finches, and Australian hon- dividual contributions to variance in a life-his- eyeaters) suggests that such a simple selective tory trait is an exciting and valuable aspect of shift to nomadism is difficult to achieve. How- the study of life-history variation. The compar- ever, phylogenetic information is increasing for ative approach provides the phylogenetic foun- these groups, and it would be very interesting to dation that makes us aware of this transition. see what their phylogenies can tell us about the interaction of ecological and organismal forces DISPERSAL AS A LIFE HISTORY TRAIT in the evolution of patterns of dispersal. There When we think about micro- and macroevolu- may also be some interesting feedbacksbetween tionary forces acting upon trait variation at the dispersal and both speciation and extinction. Or- same time, it leads to some very interesting con- nithologists might profitably follow the lead of ceptual challenges. Consider for a moment an- Jablonski (1986) and Hansen (1978), who have shown that patterns of extinction and speciation straint terminology. Trends Ecol. Evol. 6:166- in gastropods can be affected strongly by the 168. BAUM, D. A., AND A. LARSON. 1991. Adaptation re- dispersal mode of their larvae. This is a fasci- viewed: a phylogenetic methodology for studying nating case where the evolution of a life history character macroevolution. Svst. Zool. 40: l-l 8. trait at the micro level can have important ef- CARPENTER,J. M. 1989. Testing scenarios: wasp social fects on the macroevolution of the group. behavior. Cladistics 5: 13 l-144. CODDINGTON.J. A. 1988. Cladistic tests of adautational CONCLUSIONS hypotheses. Cladistics 4:3-22. CODDINGTON, J. A. 1994. The roles of homology and The phylogenetic approach places organismal convergence in studies of adaptation, p. 53-78. In biology in a rich historical context. It helps us I? Eggleton and R. I. Vane-Wright [EDS.], Phylo- genetics and ecology. Academic Press, London. to clarify the likely roles of micro- and macro- CUNNINGHAM, C. W., K. E. OMLAND, AND T H. OAK- evolutionary forces on the evolution of any giv- LEY. 1998. Reconstructing ancestral character en character.I am often reminded of a joke that states: a critical reappraisal. Trends Ecol. Evol. 13: Fred Adler (a mathematician) once had on his 361-366. office wall. Beneath a picture of a giraffe were DONOGHLJE,M. J. 1989. Phylogenies and the analysis of evolutionary sequences, with examples from the words “This is an organism; usually denoted seed plants. Evolution 43: 1137-l 156. X.” In the context of life-history variation in FELSENSTEIN,J. 1985. Phylogenies and the comparative birds, I feel that much of our failure to make method. Am. Nat. 125:l-15. progress in the elaboration and testing of life- GARLAND, T., JR., A. W. DICKERMAN, C. M. JANIS, AND history theory has been that we have often left J. A. JONES. 1993. Phylogenetic analysis of co- variance by computer simulation. Syst. Biol. 42: organismal biology out of the mix. We have of- 265-292. ten thought we could have a life-history theory GARLAND, T., JR., I? H. HARVEY, AND A. R. IVES. 1992. that applies equally well to blue whales and fruit Procedures for the analysis of comparative data . While the basic tenets of life-history the- using phylogenetically independent contrasts. ory must pertain to any population, a Syst. Biol. 41: 18-32. GARLAND, T, JR., P E. MIDFORD, AND A. R. IVES. 1999. great deal of what we find of interest in life- An introduction to phylogenetically based statis- history variation in includes a rich blend tical methods, with a new method for confidence of the effects over history of both macro- and intervals on ancestral values. Am. 2001. 39:374- microevolutionary forces acting on character- 388. GITTLEMAN, J. L., AND M. KOT. 1990. Adaptation: sta- state variation. By encouraging us to think care- tistics and a null mode1 for estimating phyloge- fully about many life-history traits and their in- netic effects. Syst. Zool. 39:227-241. teractions, comparative approachesbring the or- GRAFEN. A. 1989. The uhvloaenetic regression. Phil. ganism back into life-history studies. Compara- Trans. R. Sot. Lond. B Byol. Sci. 3??6:119-156. tive approaches will be an important tool for GREENE, H. W. 1986. Diet and arboreality in the em- erald monitor, Varanus prusinus, with comments investigating life-history variation for a long on the study of adaptation. Fieldiana Zool. 31: l- time to come. They cannot replace detailed stud- 12 ies within single populations, but I see them as HANSEN, T A. 1978. Larval dispersal and species lon- a necessary complement. gevity in lower Tertiary gastropods. Science 199: 885-887. ACKNOWLEDGMENTS HARVEY, l? H., AND M. D. PAGEL. 1991. The compar- ative method in evolutionary biology. Oxford Many thanks to Robert Ricklefs for inviting me to par- Univ. Press, Oxford. ticipate in his symposium on life histories. My re- HEINROTH, 0. 1922. Die Beziehungen zwischen Vo- search on swallows has been funded by the NSE gelwicht, Eigewicht, Gelegegewicht und Brut- USDA, and the Whitehall Foundation. Fred Sheldon, dauer. J. Ornithol. 70:172-285. Amy McCune, and Harry Greene have provided me JABLONSKI,D. 1986. Larva1 ecology and macroevolu- many challenges and insights in thinking about char tion in marine invertebrates. Bull. Mar. Sci. 39: acter evolution, Linda Harrington was very helpful in 565-587. preparing the manuscript, and Bob, Amy, Harry, Wal- LACK, D. 1968. Ecological adaptations for breeding in ter Koenig, Arnon Lotem, and two anonymous review- birds. Methuen, London. ers provided many useful comments on the manuscript. LANYON, S. M. 1992. Interspecific brood parasitism in blackbirds (Icterinae): a phylogenetic perspective. LITERATURE CITED Science 255~77-79. LAUDER, G. V 1990. Functional morphology and sys- ANTONOVICS, J., AND P. H. VAN TIENDEREN. 1991, On- tematics: studying function patterns in an histori- toecogenophyloconstraints? The chaos of con- cal context. Annu. Rev. Ecol. Syst. 21:317-340. LEVEY, D. J., AND E G. STILES. 1992. Evolutionary 1997. Phylogeny, specialization, and brood para- precursors of long distance migration: resource site-host coevolution. Am. Ornithologists’ Union availability and m&ement patterns in Neotropical Meeting, Minneapolis-St. Paul, MN. (Abstract) land birds. Am. Nat. 140:447-476. SHELDON, E H., AND L. A. WHITTINGHAM. 1997. The LORENZ, K. 1941. Vergleichende Bewegungsstudien an use of phylogeny in studies of ecology, be- Anatinen. J. Ornithol. 89:Sonderheft 19-29. havior, and morphology. In D. Mindell [ED.], Avi- MADDISON, W. F? 1990. A method for testing the cor- an molecular evolution and systematics. Academ- related evolution of two binary characters: are ic Press, New York. gains or losses concentrated on a certain branch SHELDON,E H., L. A. WHITTINGHAM, AND D. W. WINK- of a phylogenetic tree? Evolution 44:539-557. LER. 1999. A comparison of cytochrome b and MARTIN, T E., C. R. PAINE, C. J. CONWAY, W. M. Ho- DNA hybridization data bearing on the phylogeny CHACHKA, F’. ALLEN, AND W. JENKINS. 1997. of swallows (Aves:Hirundinidae). Mol. Phylo- BBIRD field protocol. Montana Cooperative gene. Evol. 2:320-331. Wildl. Res. Unit, Univ. Montana, Missoula, MT SHELDON, E H., AND D. W. WINKLER. 1993. Interge- Available: http://pica.wru.umt.edu/BSIRD/protocol/ neric phylogenetic relationships of swallows esti- protocol.htm mated by DNA-DNA hybridization. Auk I 10: MARTINS, E. P., AND T E HANSEN. 1997. Phylogenies 798-824. and the comparative method: a general approach SHELDON, E H., AND D. W. WINKLER. 1999. Nest ar- to incorporating phylogenetic inf&mation;nto the chitecture and avian systematics. Auk 116:875- analvsis of interspecific data. Am. Nat. 149:646- 877. 667: WANNTORP, H. 1983. Historical constraints in adapta- MCKITRICK, M. C. 1993. Phylogenetic constraint in tion theory: traits and non-traits. Oikos 41:157- evolutionary theory: has it any explanatory pow- 160. er? Annu. Rev. Ecol. Syst. 24:307-330. WINKLER, D. W. 1993. Use and importance of feathers PADIAN, K. 1985. The origins and aerodynamics of as nest lining in Tree Swallows (Tachycineta bi- flight in extinct vertebrates. Paleontology 28:413- color). Auk 110:29-36. 434. WINKLER, D. W., AND E H. SHELDON. 1993. Evolution PAGEL, M. D. 1994. The adaptationist wager, p. 29- of nest construction in swallows (Hirundinidae): a 51. In I? Eggleton and R. I. Vane-Wright [EDS.], molecular phylogenetic perspeciive. Proc. Natl. Phylogenetics and ecology. Academic Press, Lon- Acad. Sci. USA 90:5705-5707. don. WINKLER, D. W., AND E H. SHELDON. 1994. Phyloge- PAGEL, M. D. 1998. Inferring evolutionary processes netic hierarchy in character variability and its from phylogenies. Zool. Ser. 216:331-348. causes: lessons from character-state distributions RIDLEY, M. 1983. The explanation of organic diversity: in swallows, Hirundinidae. J. Ornithol. I35:342. the comparative method and adaptations for mat- ZYSKOWSKI, K., AND R. 0. PRUM. 1999. Phylogenetic ing. Oxford Univ. Press. Oxford. analysis of nest architecture of the Neotropical ov- ROTHSTEIN, S. I., M. A. PATTEN, AND R. C. FL.EISCHER. enbirds (Furnariidae). Auk I 16:891-9 11.