2101q2003 The Society for the Study of Evolution. All rights reserved.Evolution, 57(9), 2003, pp. 2101–2118LINEAR HABITATS AND THE NESTED CLADE ANALYSIS: AN EMPIRICALEVALUATION OF GEOGRAPHIC VERSUS RIVER DISTANCES USING AN OZARKCRAYFISH (DECAPODA: CAMBARIDAE)JAMESW. FETZNER,JR.1,2ANDKEITHA. CRANDALL1,31Department of Integrative Biology, Brigham Young University, 675 Widtsoe Building, Provo, Utah 846023Monte L. Bean Life Science Museum, Brigham Young University, Provo, Utah 84602E-mail: [email protected]. The nested clade analysis can be extremely useful in testing for an association between genetic variationand geography and in explaining these observed patterns in terms of historical or contemporary population processes.The strength of this method lies in its ability to test a variety of processes simultaneously under a rigorous statisticalframework. Indeed, many recent studies have used the nested analysis in a wide range of terrestrial and aquatic taxa.However, it has been suggested that riverine, riparian, or coastal species may be better examined using river (orcoastal) distances rather than the standard geographic (great circle) distances among populations. It is thought thatthe standard geographic distances may not adequately describe the actual distances involved between populations ofspecies inhabiting these one-dimensional (riverine) habitats. Therefore, we analyzed population data from an Ozarkcrayfish, Orconectes luteus, to examine the effects on the results of a nested clade analysis using river distances. Inmost cases, the haplotypes detected in this crayfish were unique to a particular drainage or a group of neighboringdrainages, indicating very little movement of individuals among drainages. Five major population groups were detected,corresponding to many of the major river drainages sampled in this study. The two types of distance analyses obtainsimilar results for higher-level (older) clades, but differ in many of the inferences made for lower-level (younger)clades. However, we suggest that the comparison of both types of analyses for riverine species may enhance theprocess of elucidating historical and contemporary population processes, especially in cases where the transfer ofindividuals among different drainages are involved.Key words.Cambaridae, crayfish, genetic variation, linear habitats, nested clade analysis, phylogeography.Received August 23, 2002. Accepted March 17, 2003.The basic goal of population genetic studies is to describehow genetic variation is distributed within and among pop-ulations. However, additional information from gene gene-alogies has added a temporal perspective to phylogeographicanalyses, which allows such studies to examine the geo-graphic distribution of genetic variation in a historical frame-work. This combination of spatial and temporal informationhas enlightened many of our views of underlying populationprocesses. Many phylogeographic methods fall short, how-ever, because the detection of an association between hap-lotype distributions and geography does not necessarily leadto an explanation of this observed pattern (Templeton et al.1995). One method developed to specifically address thisproblem is the nested clade analysis (NCA) of Templeton etal. (1987) and later extended by Templeton et al. (1995) andTempleton (1998, 2001).The NCA is a powerful method for examining the geo-graphic associations of haplotypes under a rigorous statisticalframework. The method also goes further than other phylo-geographic methods by explaining these associations in termsof contributions from either historical (e.g., fragmentation,colonization, or range expansion) or present-day (e.g., re-stricted gene flow) processes that have played a role in de-fining the currently observed patterns of population structure(Templeton et al. 1995; Templeton 1998).Several other methods are also currently available that usetemporal information contained in haplotype data. These in-2Present address: Section of Invertebrate Zoology, Carnegie Mu-seum of Natural History, 4400 Forbes Avenue, Pittsburgh, Penn-sylvania 15213; E-mail: [email protected] the analysis of molecular variance (AMOVA) of Ex-coffier et al. (1992) and the variance in ordered allelesmethod(NST) of Dumolin-Lapegue et al. (1997). However, the NCAallows for the analysis of more complex population processesthan the NSTmethod and provides a more objective assess-ment of geographic partitioning of haplotypes (and clades)than does the AMOVA method (Cruzan and Templeton 2000;Emerson et al. 2001; but see also Knowles and Maddison2002).In addition, a few alternative methods have been used toinfer population histories and include the mismatch pair dis-tribution method of Rogers and Harpending (1992) and theskyline plot method of Pybus et al. (2000). However, bothof these methods are limited in that they only deal with asingle historical event or deal strictly with range expansions(or contractions).While the NCA has been used in numerous analyses (e.g.,Turner et al. 2000; Bernatchez 2001; Carbone and Kohn2001; Sivasundar et al. 2001a), most such studies use geo-graphic coordinates (latitude and longitude) to calculate thegreat circle distance between sampling sites or populations.For many riverine, riparian, or coastal species, however,thesegeographic distances may not adequately represent the actualdistances separating such populations (Fig. 1). In such cases,user defined distances (hereafter referred to as ‘‘river dis-tances’’) calculated by following river courses or coastlinesmay better reflect the actual distances that must be traversedif individuals were to migrate between populations.Linear river distances are often much larger than the geo-graphic distances calculated among populations (see Fig. 2,Appendix 1). However, no empirical studies have been con-2102J. W. FETZNER, JR. AND K. A. CRANDALLFIG. 1. An illustration of the difference in estimating distances among populations of a riverine or riparian species using either geographicor river distances. (A) Demonstration of the different distances calculated among three hypothetical populations (n) from the same riverdrainage. (B) An illustration of the difference in distances calculated using the different methods. For geographic distances, the distancebetween populations 1 and 2 is the smallest, whereas for the river distances, populations 1 and 2 have the largest distance. (C) Howthese distances are
View Full Document
Unlocking...