Only recently have substantial efforts been made to understand phylogenetic relationships among freshwater mussels of the western United States and Canada. Genetic studies show the existence of two divergent clades in western Anodonta, one containing Anodonta californiensis and Anodonta nuttalliana, and another containing Anodonta oregonensis and Anodonta kennerlyi, but relationships within these two clades remain unclear. For example, some authors have placed A. californiensis in the synonymy of A. nuttalliana, but additional taxonomic information is needed to resolve these issues. We examined glochidial shell size and fine structure of these four species to assess the taxonomic utility of these characters. Glochidia of A. oregonensis and A. kennerlyi were similar in size and fine structure, which supports their proposed close relationship. Glochidia of A. californiensis and A. nuttalliana were smaller in all dimensions than A. oregonensis and A. kennerlyi, which supports the existence of two divergent clades. However, shell size and fine structure also differed between A. californiensis and A. nuttalliana, which supports the distinctiveness of these two taxa. Glochidial characters may help to clarify evolutionary relationships among western U.S. Anodonta and other problematic groups.
Until recently, six species of freshwater mussels in the genus Anodonta were recognized from western North America (Turgeon et al. 1998): Yukon Floater, Anodonta beringiana A. Middendorf, 1851; California Floater, Anodonta californiensis I. Lea, 1852; Woebegone Floater, Anodonta dejecta Lewis, 1875; Western Floater, Anodonta kennerlyi I. Lea, 1860; Winged Floater, Anodonta nuttalliana I. Lea, 1838; and Oregon Floater, Anodonta oregonensis I. Lea, 1838. A recent taxonomic revision reassigned A. beringiana to the genus Sinanodonta based on its closer genetic relationship to Asian Sinanodonta woodiana Lea, 1834, than to North American Anodonta (Chong et al. 2008; Lopes-Lima et al. 2017). Williams et al. (2017) also synonymized A. dejecta under A. californiensis based primarily on adult shell morphology (see Bequaert and Miller 1973; AZGFD 2017).
Genetic studies show that the remaining four species represent two highly divergent clades: A. oregonensis/ kennerlyi and A. californiensis/nuttalliana (Chong et al. 2008; Mock et al. 2010). Both species within each clade are genetically similar, and their distinctiveness is unclear. Blevins et al. (2017) suggested synonymizing A. californiensis under A. nuttalliana based on overlapping adult shell morphology. However, adult shell morphology in western Anodonta is highly variable, and additional characters are needed to evaluate the status of these taxa (Mock et al. 2010).
Gross glochidial shell morphology has been used to inform mussel taxonomy since the early 1900s (LeFevre and Curtis 1910; Surber 1912). More recently, glochidial fine structure as revealed by scanning electron microscopy (SEM) has been used to inform phylogenetic hypotheses (Sayenko et al. 2005; Pimpão et al. 2012; Sayenko 2016a, Sayenko 2016b). We examined glochidial shell morphology of A. californiensis, A. kennerlyi, A. nuttalliana, and A. oregonensis to assess whether these characters may be useful for better understanding relationships among these taxa.
We collected two to four gravid females of each species from the following locations (Fig. 1): A. californiensis, Wildhorse Creek, tributary of the Umatilla River, Oregon (A. californiensis is extirpated at the type locality); A. kennerlyi, Lake Chilliwack, British Columbia (type locality); A. nuttalliana, Columbia Slough, near Portland, Oregon (type locality); and A. oregonensis, Walla Walla River, Washington (A. oregonensis was unavailable at the type locality). The specimens of A. kennerlyi and A. nuttalliana we used for this project were the same specimens whose identification was described by Chong et al. (2008) based on mitochondrial cytochrome c oxidase subunit I (COI) sequencing. We verified our specimens of A. californiensis and A. oregonensis based on adult shell shape and examination of mitochondrial COI sequences that allow assignment to the clades described by Chong et al. (2008).
Upon collection, we examined the gills of each mussel to assess gravidity; gravid gills were identified as having a puffy or swollen appearance. We transported gravid mussels to the laboratory where we collected glochidia by rupturing the gill and flushing out glochidia with a wash bottle filled with water. We used only fully mature glochidia for analysis. Maturity was determined by introducing several grains of salt into a subsample of glochidia; fully mature glochidia snapped shut after exposure to salt (Zale and Neves 1982).
We preserved and examined glochidia in two ways. We preserved one subsample of glochidia from each female in 70% ethanol and measured the size of 20 glochidia from each subsample. For each glochidium, we measured the following shell dimensions under a light dissecting microscope using ImageJ image analysis software (NIH 2004): height (the widest point from the dorsal to ventral shell edge), length (the widest point from the anterior to posterior shell edge), and hinge length (Fig. 2).
We preserved a second subsample of glochidia from each female for examination of shell fine structure with SEM. We removed glochidial tissue by soaking glochidia in a 5% sodium hypochlorite solution for 2 min, followed by five rinses in tap water and preservation in 70% ethanol (Kennedy et al. 1991; O'Brien et al. 2003). Glochidial shell samples were shipped to the Interdisciplinary Center for Biotechnology Research at the University of Florida, Gainesville, for SEM, where several hundred shells of each species were mounted on double-sided carbon tape, air dried for 15 min, and coated with gold. Photos were taken of the exterior and interior valve; the flange region, a flattened area along the ventral margin of the glochidial valve; and shell sculpture.
We examined the following fine structures: the styliform hook, the projection from the ventral edge of the valve; microstylets, larger (>1.0 µm) toothlike projections located on the styliform hook; micropoints, smaller points (<1.0 µm) located along the ventral valve edge; and exterior shell sculpture, the fine surface texture on the valve (Fig. 3; see Clarke 1981; Hoggarth 1999).
We used multivariate analysis of variance (MANOVA) to examine how glochidia size varied among the four species. All linear combinations of the dependent variables were approximately normally distributed based on examination of scatter plots, and there were no departures from normality or homogeneity of variance (Shapiro-Wilks test; Bartlett's test for homogeneity of variance). Because the MANOVA overall F test was significant, we examined the three size variables separately using ANOVA and Tukey's HSD post-hoc tests. All analyses were conducted with JMP 11 (SAS Institute, Cary, NC, USA).
We made qualitative comparisons of fine structure morphology among species.
There were significant overall differences in shell size among the four species (Wilks lgr; = 0.05, F177,81 = 45.3, P < 0.001). Shell length, height, and hinge length each differed significantly among the four species (F3,78 = 94.76, 143.37, 167.03, respectively; P < 0.0001 for all tests; Table 1). Shell length was greatest in A. kennerlyi and smallest in A. californiensis, and length differed among all four species. Shell height and hinge length were not significantly different between A. kennerlyi and A. oregonensis but were significantly larger than in A. nuttalliana or A. californiensis. Shell height and hinge length differed significantly between A. nuttalliana and A. californiensis, with A. californiensis having the smallest values.
Glochidial shell measurements of four western North American Anodonta. Values are means ± standard deviation (µm) and are based on univariate ANOVA for each size variable. Values within a column with different superscripted letters are significantly different (P < 0.05, Tukey's HSD post-hoc tests).
Microstylet morphology was similar within each clade, but it differed between the two clades. However, microstylet arrangement was similar between the A. oregonensis/kennerlyi clade and A. nuttalliana, but it differed in A. californiensis. Shell sculpture differed between the A. californiensis/nuttalliana and A. oregonensis/kennerlyi clades, but this character was similar within clades (Table 2; Figs. 3–6). Shell sculpture of A. oregonensis and A. kennerlyi was intermediate between two previously described sculpture patterns, beaded and loose-looped (Hoggarth 1999). Shell shape and micropoint morphology did not provide consistent discrimination of clades or species.
Glochidial fine structure characters of four western North American Anodonta.
The similar glochidial size of A. oregonensis and A. kennerlyi supports the close genetic relationship between these species (Chong et al. 2008). The smaller size of A. californiensis and A. nuttalliana potentially supports the close relationship between these two species and their distinctiveness from the A. oregonensis/A. kennerlyi clade. However, the consistent and marked differences in size between A. californiensis and A. nuttalliana do not support placement of A. californiensis in the synonymy of A. nuttalliana (Blevins et al. 2017).
Patterns of glochidial shell fine structure among the four species were similar in most respects to patterns of size. Anodonta oregonensis and A. kennerlyi had similar patterns of shell sculpture, which supports their close genetic relationship (Chong et al. 2008). This pattern, which was intermediate between beaded and loose-looped sculpture, has been described in one other North American anodontine, Utterbackiana implicata (Hoggarth 1999), and in an Asian species, Kunashiria haconensis (Sayenko 2016a). Shell sculpture also was similar between A. californiensis and A. nuttalliana, which supports their close genetic relationship, and the difference in sculpture between these species and A. oregonensis/kennerlyi supports the existence of two divergent clades in western Anodonta as proposed by Chong et al. (2008). Sculpture similar to that of A. californiensis and A. nuttalliana also is present in North American Utterbackiana suborbiculata (Hoggarth 1999) and Asian Anodonta cygnea and Cristaria tuberculata (Sayenko 2016a, 2016b). Microstylet morphology also supported differences between the two clades, but microstylet arrangement (continuous versus broken rows) of A. californiensis differed markedly from A. nuttalliana, which does not support the placement of A. californiensis in the synonymy of A. nuttalliana as proposed by Blevins et al. (2017). However, microstylet arrangement of A. nuttalliana was more similar to the more distantly related A. oregonensis/kennerlyi than to its apparent close relative, A. californiensis. Micropoint morphology did not appear to be useful for diagnosing clades or species.
Patterns of glochidia shell size and fine structure among these four species of western Anodonta largely support proposed phylogenetic relationships based on genetic data (Chong et al. 2008; Mock et al. 2010), but they provide additional information about the potential distinctiveness of A. californiensis and A. nuttalliana. Glochidial shell features provide less ambiguous and less variable characters than notoriously vague and highly variable adult shell characters, which can be influenced to a large extent by environmental factors. Our conclusions are based on glochidia from a single population for each species. Within-population variation in glochidial size generally is low, but little is known about among-population variation (Kennedy and Haag 2005), and our data do not reflect this latter potential source of variation. Nevertheless, use of glochidial characters in conjunction with genetic data, adult anatomical characters, reproductive traits, and other data may help to clarify evolutionary relationships among western North American Anodonta and other problematic groups.
We thank Kimberly Backer-Kelley, a laboratory technician at the University of Florida Interdisciplinary Center for Biotechnology Research, for her scanning electron microscopy photos. David Wolf and Katrina Burnside provided valuable assistance with field collections. We thank lab and field volunteers for their contribution to this work, including Jonathan Arthur, Emma Stevens, Justin Pearson, and John Patterson. Gary James, Julie Burke, Gene Shippentower, and Celeste Reves of the Confederated Tribes of the Umatilla Indian Reservation (CTUIR) provided project support and guidance, as did Debbie Docherty of the Bonneville Power Administration. Colette Coiner of CTUIR produced the map highlighting the study area, and Catherine Nano provided statistical input. Funding for this research was provided by the CTUIR Freshwater Mussel Project and Bonneville Power Administration.