2008 NORTHEASTERN NATURALIST 15(1):111–122 Patterns of Spatial Variability in the Morphology of Sympatric Fucoids Melinda A. Coleman1,2,* and Jessica F. Muhlin1 Abstract - Morphology is one of the most common methods used to identify among congeneric species, but its use can be problematic when spatial patterns of phenotypic variability are unknown. We examine spatial variability in the morphology of two sympatric species of algae (Fucus vesiculosus and F. spiralis) on large (>100 km) and small (<10 km) scales and determine the extent to which they can be identified based on morphology. These species could generally be distinguished based on gross morphology on large spatial scales, but on small scales some individuals exhibited much overlap in the gross morphology. This finding is not surprising given that hybridization and introgression in these species is common. Although there were some consistent patterns in morphology between species, many were species-specific. Similarly, there were few common spatial patterns when individual measures of morphology were analyzed, suggesting that selective pressures may act on each species independently. These results have implications for the use of morphology in identifying congeneric species. Given the existence of individuals of intermediate morphology, it is likely that species are often misidentified when spatial variability in morphological distinctness is not considered. In particular, stipe width may prove to be a valuable predictor of species identification for non-reproductive and young individuals, as it was the only morphological variable to vary consistently between the species. Introduction Morphology is one of the most common methods used to identify and classify species. Congeneric species, however, are often difficult to distinguish based on morphology because it can vary greatly in space. Moreover, in closely related or congeneric species, hybridization can result in individuals of “intermediate” morphology. Using morphology to identify and distinguish among species is often problematic when spatial patterns of morphological variability are unknown (Mathieson et al. 1981). Morphological variability within species of marine macroalgae can occur on small and large spatial scales. Field experiments have demonstrated that the processes that cause this variability are sometimes genetic (e.g., Chapman 1974, Roberson and Coyer 2004) but often environmental (e.g., Fowler- Walker et al. 2006). Before we can understand the processes that structure morphology and whether such processes are general or specific to particular places, morphological variability must be quantified on a number of spatial scales. Although there exists a long history of such sampling for many species (see Mathieson et al. 1981 and references therein), rarely are patterns of 1321 Hitchner Hall, School of Marine Sciences, University of Maine, Orono, ME 04469. 2Current address - 501B Biological Sciences Building, Center for Marine Biofouling and Bioinnovation, University of New South Wales, NSW 2052, Australia. *Corresponding author - melinda.coleman@unsw.edu.au. 112 Northeastern Naturalist Vol. 15, No. 1 morphological variability simultaneously assessed across more than a single species over multiple spatial scales. Understanding the extent to which the morphology of closely related or morphologically similar species co-varies in space is particularly important where morphological characteristics are used to determine species identity. Fucus vesiculosus L. and F. spiralis L. are sympatric macroalgae that are common on the intertidal rocky shores of the northeastern USA and Canada. Fucus vesiculosus occurs in mid-intertidal areas, with F. spiralis occupying the adjacent substratum higher on the shore. Common morphological traits used to distinguish between these species in the field include the presence of a receptacle ridge (F. spiralis) and vesicles (F. vesiculosus); however, these traits are not always present (e.g., in non-reproductive and young individuals, respectively) making species identification problematic (Burrows and Lodge 1951). Moreover, there is much morphological variability in these and other traits within each species (Jordan and Vadas 1972, Kalvas and Kautsky 1998, Knight and Parke 1950, Scott et al. 2001), which has often been hypothesized to be a result of hybridization (Burrows and Lodge 1951, Kniep 1925, Scott and Hardy 1994), different selective regimes (Chapman 1995), or genotypic differences (Anderson and Scott 1998). Despite great spatial variability in morphology of each individual species and subsequent difficulties in species identification, we have little knowledge of the extent to which the morphology of these congeneric species co-vary in space. Indeed, this is hampered by the fact that few studies have quantified morphological variability in each species across more than one spatial scale. Here, we examine spatial variability in the morphology of these two congeneric species of algae and determine the extent to which these species can be identified and distinguished based on morphology. We tested the hypothesis that on large spatial scales (10s to 100s of km) F. vesiculosus and F. spiralis can generally be separated based on gross morphology and individual measures of morphology, but at small spatial scales (i.e., within sites), these differences break down, potentially due to site-specific levels of hybridization and introgression. Further, we determined whether spatial differences in the morphology of each species are common to both species or are species-specific. Due to the occurrence of apparent widespread dispersal (Coleman and Brawley 2005; Muhlin et al., in press) and hybridization and introgression (Engel et al. 2005, Wallace et al. 2004) between these species, we thought it most likely that selection or differing levels of hybridization were the driving forces behind any differences in gross morphology among sites therefore we also tested the hypothesis, that for each species variability in morphology among sites is not related to spatial distance. That is, variability in morphology between sites that are close together will be similar to, or greater than, morphological variability between sites that are far apart. Materials and Methods Fucus vesiculosus and F. spiralis were collected from each of 4 sites (separated by km) at each of two coastal points, Pemaquid Point (43o83.702'N, 2008 M.A. Coleman and J.F. Muhlin 113 69o50.64'W) and Schoodic Point (44o21.928'N, 68o04.613'W), >100 km of coastal distance apart on the coast of Maine, from August to November 2002 (see Coleman and Brawley 2005 for map). F. spiralis were also collected from 3 sites (separated by 10s of meters) at Avery Point in Connecticut (CT, 41o18.95'N, 72o3.99'W), approximately 565 km from sites in Maine (linear distance). These Maine sites were Ledges (L, 44o20.585'N, 68o02.632'W), Marks (M, 44o21.928'N, 68o04.613'W), Great Pool (GP, 44o20.032'N, 68o03.527'W), and Navy (N, 44o20.311'N, 68o04.021'W) at Schoodic Point and Krezgy (K, 43o50.108'N, 69 o30.833'W), Pemaquid Lighthouse (PL, 43o50.220'N, 69o30.428'W), Yellow Head (YH, 43o51.481'N, 69o30.40'W) and West Side (WS, 43o51.481'N, 69o31.130'W) at Pemaquid Point. Distances between all sites within a point were approximately equidistant and were chosen so that two were located on either side of each coastal point. Different sites on these coastal points may represent different hydrodynamic regimes (i.e., more exposed sites versus more sheltered sites) that have the potential to influence the morphology of fucoids (Bäck 1993, Kalvas and Kautsky 1993). Individuals of each species (n = 20 to 30 adults) were randomly collected along a 60-m transect line, which was laid down in the middle of their respective “zones” in the intertidal. This avoided any potential “hybrid zone” near where the species distributions overlap. Only individuals that could be reliably identified by the presence of vesicles (F. vesiculosus) or receptacle margin (F. spiralis) were collected (see Fig. 114 in Fritsch 1945 for illustrations). Specimens were returned to the laboratory on ice for measurement. We measured and compared 5 morphological variables that have previously been shown to be highly variable in fucoids (Bäck 1993). We measured thallus length (cm) and width (cm), stipe length (cm) and width (mm), and midrib width (mm) (see Bäck 1993 for descriptions of how each variable was measured). For F. vesiculosus, we also counted the number of vesicles and determined whether each individual was male or female. Samples were pressed and stored in the University of Maine Herbarium (MAINE). Plaster of Paris clod cards (as described by Thompson and Glenn 1994) were used to measure relative time-integrated water motion at 4 sites at Schoodic Point over 7 days of low to moderate water motion from September to November 2003 (n = 3 replicate clod cards per site per day) to test whether morphology was correlated with water motion. Clod cards were cast from plaster of paris (calcium sulfate, Mallinckrodt Baker, Inc., Paris, KY) in ice cube trays, air dried for 72 hours, pre-weighed (ca. 22 g), then securely glued onto perspex backing plates. Cards were fastened by cable ties to polypropylene lines on bolts in the mid-Fucus vesiculosus zone. Time of immersion and emersion of each clod card was noted. After the fucoid bed was exposed after high tide, clod cards were collected, allowed to dry for 72 h, and reweighed to establish weight lost by dissolution. Water motion was expressed as weight loss (grams) per hour that each clod card was submerged over a single tidal cycle. Spatial variability in gross morphology was visualized using non-metric multidimensional scaling (nMDS) plots generated from Bray-Curtis dissimilarity matrices. The significance of apparent groupings was tested using 114 Northeastern Naturalist Vol. 15, No. 1 a 3-factor PERMANOVA (Anderson 2001, McArdle and Anderson 2001) on Bray-Curtis dissimilarity values calculated using raw data. Permutations (n = 4999) were of residuals under the reduced model (Anderson and Legendre 1999, Anderson and ter Braak 2003). The factors were species (fixed), point (fixed), and site (nested in point, fixed). Point and site were treated as fixed factors because they were specific distances apart, on specific sides of each point, and constituted a more comprehensive genetic sampling program (e.g., see Coleman and Brawley 2005). Pairwise post hoc tests were done for significant factors, and the Bonferroni correction was used (P < 0.008). Spatial relationships for each individual measure of morphology were analyzed using 3-factor analysis of variance (ANOVA). The factors were as above, and n = 20 replicates were used. Data were not transformed in the few cases where variances were heterogeneous because ANOVA is robust to departures from this assumption when sample sizes are large (i.e., n = 20) or where residual degrees of freedom are greater than 30 (see Underwood 1997). Student-Newmann-Keuls (SNK) tests were done for terms where ANOVA showed significant differences to reveal the exact nature of these differences. For all SNK tests, P < 0.01 was used. For sites with 30 replicates (K, M, and GP for F. spiralis), we randomly selected 20 replicates to balance our design for PERMANOVA and ANOVA. For PERMANOVA analyses that included “sex” as a factor, we randomly selected n = 6 replicates of each sex from each site because this was the lowest number of replicates of any one sex at a site. All replicates were used, however, to generate nMDS plots. For each species, correlations between each pair of variables were assessed using all data (n = 244 for F. spiralis and n = 160 for F. vesiculosus). We used the Bon-Ferroni correction, and P < 0.005 was used for correlations. Relative levels of water motion among sites were analyzed using a one-way ANOVA (n = 21 replicate measures per site over all days), and differences determined using SNK tests. Data were square root (x + 1) transformed to conform to homogeneity of variances. Results Separating F. vesiculosus and F. spiralis based on morphology As predicted, nMDS plots revealed that F. vesiculosus and F. spiralis clearly separated based on gross morphology (Fig. 1). SIMPER showed that thallus length contributed most to this difference. The significance of these groupings was confirmed using PERMANOVA (Table 1). nMDS plots also revealed varying levels of separation between species within each site, with some overlap in gross morphology (Fig. 2). At all sites, there was significant separation between species (P < 0.01); however, this pattern was weaker at YH (P < 0.05). Moreover, from nMDS plots it is clear that some individuals have morphologies that are more similar to individuals belonging to the other species than they are to other individuals of their own species (e.g., many individuals at M, L, YH, and K). At all sites, thallus length contributed between 63% and 83% of dissimilarity between species. 2008 M.A. Coleman and J.F. Muhlin 115 We tested for differences in each morphological variable between species at each site (Table 2). F. vesiculosus had a greater stipe length than F. spiralis, and this pattern was consistent at all points and sites (Table 2). Similarly, stipe width was greater in F. vesiculosus than in F. spiralis, but only at approximately half the sites (M, GP, L, and PL; Table 2). Thallus length was greater in F. vesiculosus compared to F. spiralis at all sites, except YH, where there were no differences between species (Table 2). Interestingly, midrib width was greater in F. spiralis than in F. vesiculosus at all but 3 sites (M, GP, and WS; Table 2), and thallus width was greater in F. spiralis than F. vesiculosus at all sites except GP, PL, and WS (Table 2). Figure 1. nMDS plot showing relationships in gross morphology among sites and between F. vesiculosus and F. spiralis. Distances between points on the plot represent how similar (points close together) or different (points far apart) sites are from one another. Each point represents a site centroid or average. Site abbreviations are as specified in Materials and Methods section. Pemaquid sites were PL, K, YH, and WS; Schoodic sites were L, M, GP, and N; and Connecticut sites were CT1, CT2, and CT3. The letters “V” and “S” before site abbreviations represent F. vesiculosus and F. spiralis, respectively. All data were used to generate plots, and CT sites are included for F. spiralis. Table 1. Results of PERMANOVA for all 5 morphological variables for both species. Factors are as specified in Material and Methods section. n = 20 replicate individuals per site. ** = P < 0.01, *** = P < 0.001. Results of SNK tests are given in Results section. Source d.f. SS MS F P Species (Sp) 1 61,656.97 61,653.97 259.66 *** Point (Po) 1 2282.49 2282.49 9.61 *** Site (Po) 6 8152.24 1358.71 5.72 *** Sp x Po 1 1365.04 1365.04 5.75 ** Sp x site (Po) 6 16,335.53 2722.59 11.47 *** Residual 304 72,186.55 11.47 116 Northeastern Naturalist Vol. 15, No. 1 Spatial patterns in morphology of each species For F. vesiculosus, there was no difference in gross morphology between points (Fig. 1, Table 1). This pattern can be explained by large differences among sites at Pemaquid but not at Schoodic Point. At Schoodic Point, gross morphology did not differ among sites after the Bonferroni correction (Fig. 1, Table 1). At Pemaquid, patterns were difficult to resolve, but in general, gross morphology at K and YH differed from PL and WS. Thallus length contributed Figure 2. nMDS plot showing relationships between Fucus vesiculosus (v) and F. spiralis (s) at each site at Schoodic and Pemaquid Points. Distances between points on the plot represent how similar (points close to each other) or different (points far from each other) species are from one another. Site abbreviations are as specified in Materials and Methods section. Pemaquid sites were PL, K, YH, and WS; and Schoodic sites were L, M, GP, and N. 2008 M.A. Coleman and J.F. Muhlin 117 to at least 68% of these differences. When each morphological variable was analyzed separately, there were few clear patterns between points, among sites within points, or even among different variables. First, stipe length and width did not vary at any point or site (Fig. 3, Tables 2 and 3). Some variables (e.g., thallus length) differed among sites at one point (Pemaquid) but not the other (Schoodic; Fig. 3, Tables 2 and 3). Other variables varied among sites at both points (e.g., thallus width and midrib width) (Fig. 3, Tables 2 and 3). Where there were significant differences among sites at both points, rank orders of differences among sites were similar. Midrib width was positively correlated with thallus width, and thallus length was positively correlated with Table 2. Analyses of variance (ANOVA) for each morphological variable. Factors are as specified in Materials and Methods section. n = 20 replicate individuals per site. * = P < 0.05, ** = P < 0.01, *** = P < 0.001. See Table 3 for Sp x Si (Po) SNK results. Pemaquid sites were PL, K, YH, and WS. Schoodic sites were L, M, GP, and N. Source d.f. SS MS F P SNK (a) Thallus width Species (Sp) 1 2.11 2.11 10.82 * F. spiralis > F. vesiculosus Point (Po) 1 0.13 0.13 1.76 Site (Po) 6 0.45 0.07 2.40 * PL > K = WS = YH, PL = YH Sp x Po 1 0.11 0.11 0.58 Sp x site (Po) 6 1.17 0.20 6.27 *** See Table 3 Residual 304 9.48 0.03 (b) Thallus length Species (Sp) 1 18,107.10 18,107.11 22.46 ** F. vesiculosus > F. spiralis Point (Po) 1 145.44 145.44 2.17 Site (Po) 6 402.78 67.13 1.06 Sp x Po 1 121.61 121.61 0.15 Sp x site (Po) 6 4836.56 806.09 12.72 *** See Table 3 Residual 304 19,267.76 63.38 (c) Stipe width Species (Sp) 1 9.77 9.76 12.16 * F. vesiculosus > F. spiralis Point (Po) 1 0.00 0.00 0.00 Site (Po) 6 2.93 0.49 3.90 *** WS > K, other pairs ns Sp x Po 1 0.80 0.80 1.00 Sp x site (Po) 6 4.82 0.80 6.42 *** See Table 3 Residual 304 38.01 0.13 (d) Stipe length Species (Sp) 1 472.64 472.64 64.23 *** F. vesiculosus > F. spiralis Point (Po) 1 26.85 26.85 2.93 Site (Po) 6 54.92 9.15 2.04 Sp x Po 1 0.27 0.27 0.04 Sp x site (Po) 6 44.15 7.36 1.64 Residual 304 1360.95 4.48 (e) Midrib width Species (Sp) 1 13.61 13.61 15.56 ** F. spiralis > F. vesiculosus Point (Po) 1 1.13 1.13 0.49 Site (Po) 6 13.80 2.30 7.31 *** GP > M = L = N, PL > K = WS = YH Sp x Po 1 2.42 2.42 2.76 Sp x site (Po) 6 5.25 0.87 2.78 * See Table 3 Residual 304 95.63 0.32 118 Northeastern Naturalist Vol. 15, No. 1 stipe width (Table 4). For F. vesiculosus, we did further multivariate analyses including an extra variable (number of vesicles). We also stratified our data according to sex (male versus female, n = 6 of each per site) to determine whether gross morphology varied between male and females at some or all sites. PERMANOVA found no differences in gross morphology between sexes at any point or site (Table 5). Table 3. SNK results for significant Sp x Si (Po) terms. All terms are significant at P < 0.01. ns = not significant. Site abbreviations are as specified in Materials and Methods section. Fucus vesiculosus Fucus spiralis Schoodic Pemaquid Schoodic Pemaquid Thallus width GP > L, PL > K, ns ns other pairs ns other pairs ns Thallus length ns K = YH < PL = WS, ns K = YH > WS = PL but K = PL Stipe width ns ns M = GP < N = L, YH > WS, but N = GP other pairs ns Midrib width GP > M = N = L PL > K = YH = WS ns K = YH = PL > WS, but WS = K Table 4. Matrix of correlations among all pairs of variables. The upper right of the matrix are Pearson correlation estimates for F. vesiculosus, and the lower left are values for F. spiralis. Significant terms are identified with an “*.” P < 0.005 was used. L = length, W = width. Thallus L Thallus W Stipe L Stipe W Midrib W Thallus L 0.009 0.001 0.278* 0.055 Thallus W 0.169 0.129 0.112 0.505* Stipe L 0.359* -0.021 0.049 0.174 Stipe W 0.497* 0.151 0.183 0.089 Midrib W 0.250 0.458* 0.178 0.124 Figure 3. Variability in stipe width among sites for (a) Fucus vesiculosus and (b) F. spiralis. Site a b b r e v i a - tions are as specified in M a t e r i a l s and Methods section. CT sites are included for F. spiralis. Pemaquid sites were PL, K, YH, and WS; Schoodic sites were L, M, GP, and N; and Connecticut sites were CT1, CT2, and CT3. 2008 M.A. Coleman and J.F. Muhlin 119 Gross morphology of F. spiralis varied between Schoodic and Pemaquid Points (Fig. 1, Table 1) and thallus length contributed most to this difference (66% of dissimilarity between points). Within each point, there were differences in morphology among sites. The morphology of F. spiralis at Schoodic Point showed little variability among most sites (Fig. 1), but GP was different from all other sites. There was greater variability in F. spiralis morphology among sites at Pemaquid; morphology at K and YH differed from PL and WS. Again, thallus length contributed most to these differences (at least 63%). As with F. vesiculosus, when each morphological variable was analyzed separately for F. spiralis, there were no clear patterns between points, among sites within points, or among variables. There was no difference in stipe length or thallus width between points or among sites (Fig. 3, Tables 2 and 3). Stipe width varied among sites at both points, and midrib width and thallus length varied among sites at Pemaquid but not at Schoodic (Fig. 3, Tables 2 and 3). For F. spiralis, midrib width was positively correlated with thallus width, and thallus length was positively correlated with stipe width (Table 4). In addition, stipe and thallus length were positively correlated. At Schoodic Point, water motion was generally greater at GP than at any other site (ANOVA: 3, 83 d.f.; F = 9.26; P < 0.00001). Discussion Fucus vesiculosus and F. spiralis can be distinguished based on gross morphology because all morphological variables that were measured varied between species at all spatial scales. Despite this difference, at smaller spatial scales (km) individuals exhibited a gradient of morphological variability with overlap in the gross morphology of some individuals. That is, there were always a few individuals that were morphologically more similar to the opposite species than they were to members of their own species. This overlap in morphology is not surprising. F. vesiculosus and F. spiralis occur sympatrically on intertidal shores in Maine and are known to hybridize in the laboratory (Burrows and Lodge 1953, McLachlan et al. 1971) and in the field (Engel et al. 2005, Wallace et al. 2004). Even if individuals of intermediate morphology are not recent (F1) hybrids, it is likely that some amount of introgression has occurred between F. vesiculosus and F. spiralis (Engel et al. 2005, Wallace et al. 2004). Indeed, data from microsatellite markers Table 5. PERMANOVA results for F. vesiculosus data that were separated by sex. n = 6 individuals per sex per site. Although the term Site (point) was significant, it is not relevant to this particular analysis (see Table 1). Source d.f. SS MS F Point 1 1246.28 1246.28 2.61 Site (point) 6 13,043.51 2173.92 4.55 Sex 1 1082.87 1082.87 2.27 Point x sex 1 347.66 347.66 0.73 Site (point) x sex 6 1037.62 172.94 0.36 Residual 80 38,193.24 477.42 120 Northeastern Naturalist Vol. 15, No. 1 has revealed that there is much overlap in allele sizes and no unique alleles to distinguish between F. vesiculosus and F. spiralis at the sites studied here (M.A. Coleman and J.F. Muhlin, pers. Observ.), indicating that hybridization and introgression are likely to occur. The amount of variability in gross morphology at the scale of less than 10 kilometers was not consistent on larger scales. That is, at Schoodic Point, variability in gross morphology among sites was small (F. spiralis) or nonexistent (F. vesiculosus), but at Pemaquid, the morphology of both species varied significantly among sites. This suggests that either dispersal among sites at Schoodic is great preventing the formation of genetic differences (due to drift, mutation, etc) in morphology, or that regardless of the scales of dispersal, selective pressures on morphology are similar across all sites at Schoodic and result in similar morphologies. We have some evidence to support the first of these models. Microsatellite data on gene flow in F. spiralis demonstrated that dispersal is likely to occur across large spatial scales with few genetic differences among individuals from sites at Schoodic, or between pairs of sites at Schoodic versus Pemaquid Points (Coleman and Brawley 2005, cf. Table 3). Large-scale dispersal cannot explain, however, the greater variability in morphology at Pemaquid Point, because there was also little genetic differentiation among sites at this point (Coleman and Brawley 2005, cf. Table 3). Although it appears that dispersal occurs across large spatial scales, it is possible that selective pressures differ among sites at Pemaquid and not at Schoodic Point. Consistency in patterns between species Although most patterns in morphology were often species-specific, some were consistent between species. For example, the few correlations that did exist between variables were mostly common to both species, for both species’ gross morphology was generally more variable among sites at Pemaquid than among sites at Schoodic, and spatial patterns in gross morphology at Pemaquid Point were similar for both species. Regardless of why these specific patterns occurred, the factors that influence some morphological traits of fucoids appear to be common. Interestingly, however, measures of thallus length for F. vesiculosus and F. spiralis at some places (K and YH) showed opposite patterns. That is, thallus length was shorter at K and YH than at other Pemaquid sites for F. vesiculosus, but longer at the same sites for F. spiralis (see Table 3). In contrast, this suggests that selective pressures may act on each species independently because patterns in morphology among sites were species- specific. The differing tidal heights that these species occupy may present different selective pressures on morphology and, combined with site-to-site differences in biotic and abiotic conditions, may account for varying patterns in morphology between species. The lack of correlation between morphology and distances among sites is further evidence to suggest that selection drives morphological variability in these species. For F. vesiculosus, differing levels of water motion among sites may be the driving force behind any such selection since the site with significantly greater water motion (GP) also 2008 M.A. Coleman and J.F. Muhlin 121 exhibited morphological differences where they occurred (midrib width and thallus width). Carefully designed reciprocal transplant experiments would be required to test hypotheses about selection. The results of this study have implications for the use of morphology in identifying these congeneric species. Given the existence of many individuals of intermediate morphology at some places, it is likely that at least some individuals are regularly misidentified, which may result in an increase in levels of variability in experiments, decreasing the power to detect real differences at some places. The presence of a ridge around the receptacle of F. spiralis when an individual is fertile is the most commonly used predictor of field species identification to date, but cannot be used when individuals are non-reproductive. Similarly, the presence of vesicles is used to identify F. vesiculosus in the field, but vesicles are often absent, particularly on young individuals. Stipe width was the only morphological variable to vary consistently between species at all sites and may prove to be a valuable predictor of species identification if spatial variability in this variable is taken into account. Consideration of spatial variability in patterns of morphological distinctness among congeneric species will provide a greater understanding of the genetic and environmental processes that determine species morphology and identity. Acknowledgments We are grateful to the National Science Foundation for support of this work (OCE- 99043). We also thank Susan Brawley for discussion, Dave Abelson for help with collecting and measuring F. spiralis, and 3 anonymous reviewers for thoughtful comments on the manuscript. Literature Cited Anderson, C.I.H., and G.W. Scott. 1998. The occurrence of distinct morphotypes within a population of Fucus spiralis. Journal of the Marine Biological Association of the United Kingdom 78:1003–1006. Anderson, M.J. 2001. A new method for non-parametric multivariate analysis of variance. Austral Ecology 26:32–46. Anderson, M.J., and P. Legendre. 1999. An empirical comparison of permutation methods for tests of partial regression coefficients in a linear model. Journal of Statistical Computation and Simulation 62:271–303. Anderson, M.J., and C.J.F. Ter Braak. 2003. Permutation tests for multi-factorial analysis of variance. Journal of Statistical Computation and Simulation 73: 85–113. Bäck, S. 1993. Morphological variation of northern Baltic Fucus vesiculosus along the exposure gradient. Annales Botanica Fennici 30:275–283. Burrows, E.M., and S. Lodge. 1951. Autecology and the species problem in Fucus. Journal of the Marine Biological Association of the United Kingdom 30: 161–176. Burrows, E.M., and S. Lodge. 1953. Culture of Fucus hybrids. Nature 172:1009–1010. Chapman, A.R.O. 1974. The genetic basis of morphological variation in some Laminaria populations. Marine Biology 24:85–91. Chapman, A.R.O. 1995. Functional ecology of fucoid algae: 23 years of progress. Phycologia 34:1–32. 122 Northeastern Naturalist Vol. 15, No. 1 Coleman, M.A., and S.H. Brawley. 2005. Are life-history characteristics good predictors of genetic diversity and structure? A case study of the intertidal alga Fucus spiralis (Heterokontophyta: Phaeophyceae). Journal of Phycology 41: 753–762. Engel, C.R., C. Daguin, and E.A. Serrao. 2005. Genetic entities and mating system in hermaphroditic Fucus spiralis and its close dioecious relative F. vesiculosus (Fucaceae, Phaeophyceae). Molecular Ecology 14:2033–2046. Fowler-Walker, M.J., T. Wernberg, and S.D. Connell. 2006. Differences in kelp morphology between wave-sheltered and exposed localities: Morphologically plastic or fixed traits? Marine Biology 148:755–767. Fritsch, F.E. 1945. The Structure and Reproduction of the Algae, Vol. 2. Cambridge University Press, Cambridge. UK. 939 pp. Jordan, A.J., and R.L. Vadas. 1972. Influence of environmental parameters on intraspecific variation in Fucus vesiculosus. Marine Biology 14:248–252. Kalvas, A., and L. Kautsky. 1993. Geographical variation in Fucus vesiculosus morphology in the Baltic and North Seas. European Journal of Phycology 28: 85–91. Kalvas, A., and L. Kautsky. 1998. Morphological variation in Fucus vesiculosus populations along temperature and salinity gradients in Iceland. Journal of the Marine Biological Association of the United Kingdom 78:985–1001. Kniep, H. 1925. Ueber Fucus-bastarde. Flora 118–9:331–38. Knight, M., and M. Parke. 1950. A biological study of Fucus vesiculosus L. and F. serratus L. Journal of the Marine Biological Association of the United Kingdom 29:439–514. Mathieson, A.C., T.A. Norton, and M. Neushul. 1981. The taxonomic implications of genetic and environmental induced variations in seaweed morphology. Botanical Review 47:313–347. McArdle, B.H., and M.J. Anderson. 2001. Fitting multivariate models to community data: A comment on distance-based redundancy analysis. Ecology 82:290–297. Mclachlan, J., L.C.M. Chen, and T. Edelstein 1971. Culture of 4 species of Fucus under laboratory conditions. Canadian Journal of Botany 49:1463–1469. Muhlin, J.F., C.R. Engel, and S.H. Brawley. In press. Barriers to gene flow: The influence of coastal topography and circulation patterns in structuring populations of an intertidal alga. Limnology and Oceanography. Roberson, L.M., and J.A. Coyer. 2004. Variation in blade morphology of the kelp Eisenia arborea: Incipient speciation due to local water motion? Marine Ecology Progress Series 282:115–128. Scott, G.W., and F.G. Hardy. 1994. Observations of the occurrence of hybrids between 2 sympatric species of fucoid algae. Cryptogamie Algologie 15:297–305. Scott, G.W., S.L. Hull, S.E. Hornby, F.G. Hardy, and N.J.P. Owens. 2001. Phenotypic variation in Fucus spiralis (Phaeophyceae): Morphology, chemical phenotype, and their relationship to the environment. European Journal of Phycology 36: 43–50. Thompson, T.L., and Glenn E.P. 1994. Plaster standards to measure water motion. Limnology and Oceanography 39:1768–1779. Underwood, A.J. 1997. Experiments in ecology: Their Logical Design and Interpretation Using Analysis of Variance. Cambridge University Press, Cambridge, UK. 504 pp. Wallace, A.L., A.S. Klein, and A.C. Mathieson. 2004. Determining the affinities of salt marsh fucoids using microsatellite markers: Evidence of hybridization and introgression between two species of Fucus (Phaeophyta) in a Maine estuary. Journal of Phycology 40:1013–1027.