Open Access
Translator Disclaimer
29 March 2022 Utility of Shell-Valve Outlines for Distinguishing among Four Lampsiline Mussel Species (Bivalvia: Unionidae) in the Great Lakes Region
Madison R. Layer, Russell L. Minton, Todd J. Morris, David T. Zanatta
Author Affiliations +
Abstract

Four freshwater mussel species from the tribe Lampsilini found in the Laurentian Great Lakes region—Lampsilis fasciola (Wavy-rayed Lampmussel), Lampsilis cardium (Plain Pocketbook), Ortmanniana ligamentina (Mucket), and Lampsilis siliquoidea (Fatmucket)—have similar and variable shell morphologies that make some specimens difficult to identify in the field. Identification is further confounded by sexual dimorphism in three of the four species. We used landmark-based morphometric analyses of shell shape in conjunction with DNA barcoding to quantify shell-shape differences between the species. We collected specimens (N = 388) from Great Lakes tributaries in Michigan, USA, and Ontario, Canada. We photographed each specimen and made an initial identification in the field. We then took a tissue biopsy or swab from 248 of the specimens, sequenced a fragment of the mitochondrial cytochrome c oxidase subunit 1 (COI) gene, and confirmed identifications by comparing our sequences with sequences for all four species accessioned in GenBank. On the photographs, we digitized 21 two-dimensional landmarks along the shell margin and used multivariate methods to evaluate the correspondence of shell shape to our COI-confirmed species identifications and sex determinations. Principal-components analysis and linear-discriminant analysis of shell shape correctly identified only 77.8% of specimens to species and 72.2% to species and sex. Sex determination was particularly confounded by the similar shapes of female L. fasciola and female L. cardium specimens. This study demonstrates the limitations of using only two-dimensional valve shape in differentiating among some mussel species.

INTRODUCTION

Early classifications of freshwater mussel species in North America were often based almost solely on descriptions of shell morphology (Haag 2012). Even today, species are usually identified by shell characteristics. However, these identifications can be inaccurate due to wide intraspecific variation in shell characters. Genetic and morphometric techniques can improve the ability to differentiate among mussel species with similar and overlapping shell characteristics (Beauchamp et al. 2020; Beyett et al. 2020; Willsie et al. 2020).

In the Laurentian Great Lakes region, four lampsiline mussel species can be difficult to differentiate based on external shell features: Lampsilis fasciola (Rafinesque 1820), Wavy-rayed Lampmussel; Lampsilis cardium (Rafinesque 1820), Plain Pocketbook; Ortmanniana ligamentina (Lamarck 1819) (= Actinonaias ligamentina), Mucket; and Lampsilis siliquoidea (Barnes 1823), Fatmucket. At sites where these species co-occur, identification can be challenging even for experts (Cummings and Mayer 1992). In the three Lampsilis species, identification is further confounded by sexual dimorphism (Watters et al. 2009; Mulcrone and Rathbun 2018). Sex determination based on shell characters also can have a high degree of error (Hess et al. 2018).

Table 1.

Site locations of Lampsilis fasciola, Lampsilis cardium, Ortmanniana ligamentina, and Lampsilis siliquoidea and the number of field-identified and cytochrome c oxidase subunit 1 (COI)–confirmed specimens. Numbers represent the total number collected from the site, and, in parentheses, the number of specimens that had COI sequences generated.

img-z2-2_37.gif

Accurate species and sex determination is important for many reasons. For example, Lampsilis fasciola is a species of special concern in Canada (COSEWIC 2010) and is threatened in Ontario and Michigan (OMNRF 2021; MNFI 2020). Confusion between L. fasciola and more common lampsiline species could result in an inaccurate assessment of its status. If the more common L. cardium, O. ligamentina, and L. siliquoidea are misidentified as L. fasciola, the latter species' distribution and abundance may be overestimated, resulting in a potential loss of protection needed to ensure its persistence. Lampsiline species are often used for laboratory studies including studies on the impacts of invasive species and toxicological studies (e.g., Wang et al. 2011; Gilroy et al. 2014; Larson et al. 2016; Waller and Bartsch 2018; Gillis et al. 2021). Improper identification of test organisms may lead to misinterpretations of laboratory results and can lead to improper management recommendations (Shea et al. 2011).

DNA barcoding (Hebert et al. 2003) has become an important tool for species identification. Partial mitochondrial cytochrome c oxidase subunit 1 (COI) gene sequences are frequently used as diagnostic barcode markers for many unionid species (e.g., Inoue et al. 2013, 2014; Beauchamp et al. 2020; Beyett et al. 2020; Willsie et al. 2020). The large and growing number of unionid COI sequences accessioned in GenBank serve as references to improve identifications.

Geometric morphometric analysis also can be a useful tool for species identification. Landmark-based analyses allow for quantification of mollusk shell shape while removing the effects of size, position, and rotation. The resulting shape data can be analyzed using traditional multivariate statistics to detect differences among individuals or a priori groups (Webster and Sheets 2010). Recent studies combining DNA barcoding and geometric morphometric analysis have been used to distinguish between morphologically similar species (Beauchamp et al. 2020; Beyett et al. 2020; Willsie et al. 2020).

We tested the utility of geometric morphometric analyses of shell shape in conjunction with DNA barcoding to differentiate between L. fasciola, L. cardium, O. ligamentina, and L. siliqouidea. Our specific objectives were (1) to assess whether two-dimensional geometric morphometric techniques can differentiate accurately among species and sexes, and, if so, (2) to establish diagnostic and quantifiable morphological characters for distinguishing among species and sexes.

METHODS

Field Collections

We collected 388 specimens of the four target species from eight rivers in Ontario and Michigan (Table 1). As we were seeking only to differentiate among species and sexes, we did not investigate intraspecific variation within and among source populations (i.e., environmental influences of shape variation), although this could be an interesting avenue for further study. We attempted to collect a minimum of 20 individuals of each species and sex (for dimorphic species) at each site, but this was not always possible. Field identifications and sex determinations were made by the field team upon collection. Mussel identification experience of field team members ranged from novice (<1 yr of experience), to intermediate (2 to 10 yr), to advanced (>10 yr). Species identifications in the field were made based on shell morphology, beak structure, and shell coloration using a consensus approach. Sex determination was made based on the degree of shell inflation and expansion of the posterior portion of the shell; more inflated or expanded shells are characteristic of females. We photographed the left valve of each specimen. Photographs were later reviewed by the authors with advanced identification experience, and some field identifications or sex determinations were revised based on those reviews prior to analyses. We took mantle tissue biopsies (Berg et al. 1995) from a subset of individuals for each species except for L. fasciola; because of its protected status, we took less invasive swab samples from the foot and visceral mass (Henley et al. 2006). We obtained usable COI sequences from a total of 248 specimens. We preserved tissue biopsies in 95% ethanol and swabs were preserved in a lysis buffer (Sambrook et al. 1989). We measured shell length, width, height, and hinge length of every specimen using Vernier calipers (Appendix 1). After processing, all specimens were returned to the river alive.

DNA Barcoding

A Qiagen Blood and Tissue kit (Qiagen, Inc., German-town, MD, USA) was used to extract DNA from the tissue and swab samples collected in the field. Extraction success and relative quality of genomic data were assessed by electrophoresing 2-lL amplicons of the extracted DNA on a 1.5% agarose gel. Polymerase chain reaction was used to amplify a 600-bp COI fragment using primers and amplification conditions described in Campbell et al. (2008). Amplification success and relative quality were assessed by electrophoresing 2 lL of amplicons (stained with SYBR green) on a 1.5% agarose gel. Amplicons were purified using exonuclease I and shrimp alkaline phosphates (EXOSAP). The EXOSAP solution was made using 78 lL double distilled H2O, 2 lL exonuclease I, and 20 lL shrimp alkaline phosphates. To denature any remaining primers and enzymes, 1.5 lL of EXOSAP solution was added to each sample, which were then incubated at 378C for 40 min and 808C for 20 min. Once purified, amplicons were shipped to Eton Biosciences (San Diego, CA, USA) for Sanger sequencing. Generated sequences were compared to COI sequences for all four species in the GenBank database using BLAST ( http://blast.ncbi.nlm.nih.gov/Blast.cgi; accessed November 20, 2020). The BLAST result with the highest percentage of identity was chosen as the most likely species and used as the confirmed identification for the sample.

Geometric Morphometrics

We digitized shell photographs of all 388 individuals using the MakeFan application in IMP8 (Sheets 2014). We placed homologous (Type I) anchor landmarks at the peak of the umbo and the posterior edge of the hinge ligament. We established a 40-ray fan anchored at the midpoint between landmarks 1 and 2; 19 additional (Type II) landmarks were located at equidistant points where fan rays intersected the shell margin (Fig. 1). Photographs of the left valves are available on MorphoBank ( https://morphobank.org, Project Code 3918, accession nos. M738948–M739052).

Data Analysis

We obtained shape variables from our landmark configurations of COI-confirmed individuals using a generalized Procrustes analysis (Rohlf and Slice 1990). We performed two Procrustes analyses of variance (ANOVAs) (Goodall 1991) in the R package geomorph 4.0 (Adams et al. 2021): one to test for significant shape differences between the four species and the second to test for significant shape differences using species identity, sex, and the interaction between species and sex. Our sum-of-squared Procrustes distances were used as the measure of sum-of-squares (SS), with the observed SS evaluated through residual randomization permutation (Collyer and Adams 2018, 2021). Additionally, geomorph uses z-score centering and log-transformation to ensure that statistics are normally distributed. We determined significance at α = 0.05.

We performed a principal components analysis (PCA) on the Procrustes-transformed landmark dataset. A broken-stick model was used to determine the number of dimensions to retain for further analyses (Jackson 1993). We subsequently used PCA–linear discriminant analysis (LDA) and Bayesian clustering to test the utility of shell shape in identifying specimens to species and sex. In a PCA-LDA, the dimensionality of the data is reduced through an initial PCA to preserve variance, remove collinearity, and reduce overfitting in the subsequent LDA of the components (Quinn and Keough 2002). We used PAST 4 (Hammer et al. 2001) to generate principal components from our Procrustes shape variables. We then performed an LDA in PAST on the components using the COI-confirmed species identities and used the jackknifed confusion matrix to compare COI identifications with those predicted by shape. We repeated the LDA using the COI identities by sex as groups and used the jackknifed confusion matrix to assess successful discrimination.

For Bayesian model-based clustering independent of a priori classification, we used the R package mclust 5.4.5 (Scrucca et al. 2016). We generated Bayesian information criteria (BIC) values for competing clustering models and chose the model with the highest BIC score (mclust reports BIC multiplied by –1). We created a model with four clusters (representing the four species) and a model with seven clusters (species and sex where applicable). We assessed the method by calculating classification errors as the percentage of incorrect group assignment relative to the COI species identification. We also calculated incorrect group assignment relative to the COI identities by sex.

RESULTS AND DISCUSSION

BLAST analysis identified 41 L. fasciola, 96 L. cardium, 28 O. ligamentina, and 83 L. siliquoidea (Appendix 1). We recovered 16 unique haplotypes from the 248 COI sequences generated: two L. fasciola (GenBank accession nos. MW753043–MW753044), eight L. cardium (GenBank accession nos. MW752863–MW752870, one O. ligamentina (GenBank accession no. MW752989), and five L. siliquoidea (GenBank accession nos. MW752895–MW752899). Overall, field identifications were 92.3% accurate when compared to the COI identifications. The most frequently misidentified specimens in the field were L. cardium and L. siliquoidea from the Maitland River, Ontario, with 73.6% correct identification; six L. cardium were mistaken for L. siliquoidea and nine L. siliquoidea were mistaken for L. cardium. A possible reason for the misidentification of these two species in the Maitland River is that there were instances when the shape of the shell or mantle lure morphology indicated one species (i.e., inflation and truncation of the shell and lure type typical of L. cardium), but the beak sculpture indicated another (i.e., 6–12 bars typical of L. siliquoidea, as opposed to 4–5 elevated ridges for L. cardium; Mulcrone and Rathbun 2018).

Figure 1.

Examples of a fan and 21 landmarks superimposed on the left valve using the MakeFan application in IMP8 software. Type I landmarks are represented by the green points. Type II landmarks along the edge of the shell are represented by the red points. Shell specimens are (A) female and (B) male Lampsilis fasciola, (C) female and (D) male Lampsilis cardium, (E) Ortmanniana ligamentina, and (F) male and (G) female Lampsilis siliquoidea.

img-z4-1_37.jpg

Procrustes ANOVA based on the transformed shape variables revealed significant differences in shape among the COI-confirmed species (F = 8.569, P < 0.001). ANOVA using both COI-confirmed species and field- and photo-assigned sex also showed significant differences between species and sexes (F1,2 = 1.824, P = 0.027). Pairwise post-hoc residual randomization permutation procedures (RRPP; 1,000 permutations) tests revealed significant differences between these six (of 42) pairs: male L. cardium and male L. siliquoidea (P = 0.037), male L. cardium and male L. fasciola (P = 0.016), male L. cardium and male L. siliquoidea (P = 0.001), male L. fasciola and male L. siliquoidea (P = 0.001), female and male L. siliquoidea (P = 0.001), and male L. siliquoidea and O. ligamentina (P = 0.034).

The first two principal components explained 90.5% of the total variation in valve shape (Fig. 2). However, there was considerable overlap among females of all Lampsilis species and between male L. cardium and male L. fasciola. Male L. siliquoidea and O. ligamentina had limited overlap, corresponding to the results of the ANOVA.

The PCA-LDA had 77.8% mean accuracy (73.1% to 83.1%) in assigning specimens to the correct species (Table 2) and 72.1% (57.1% to 93.1%) mean accuracy in assigning specimens to the correct species and sex (Table 3). The species with the highest accuracy in the PCA-LDA model was L. siliquoidea (83.1%), but all four species were generally similar. Groups with the highest accuracy in the PCA-LDA model were male L. siliquoidea (93.1%) and O. ligamentina (82.1%) (Table 3). Groups with the lowest accuracy were female L. cardium (57.1%) and female L. fasciola (68.4%), each of which was usually misidentified as the other species. Female L. cardium were misidentified as female L. fasciola 19.0% of the time, and female L. fasciola were misidentified as female L. cardium 21.0% of the time. Similar to the field identifications, the Maitland River samples had the highest error rates for the LDA model: 20 out of 57 (35.1%) Maitland specimens of L. cardium and L. siliquoidea were misidentified by shell morphometrics. Thirteen out of these 20 specimens were a result of misidentifying L. cardium as L. siliquoidea. Of the remaining 122 genetically confirmed L. cardium and L. siliquoidea specimens (from all rivers), only three L. siliquoidea and 14 L. cardium (13.9%) were misidentified in the morphometric model. Most of these misidentifications were L. cardium being mistaken for L. fasciola (10 of 17). A possible reason for lower accuracy in the LDA model compared to field identification accuracy is that the model only accounts for the two-dimensional shape of the specimen. Other characters, such as color, ray pattern, beak sculpture, overall size and three-dimensional attributes (e.g., shell inflation), are important characters that are also taken into consideration when making field identifications (Mulcrone and Rathbun 2018).

Figure 2.

Principal component analysis (PCA) of 21 Procrustes-transformed landmark points from female (squares) and male (circles) Lampsilis fasciola (orange/salmon), Lampsilis cardium (pink shades), Lampsilis siliquoidea (blue shades), and Ortmanniana ligamentina (green). Filled symbols represent specimens with cytochrome c oxidase subunit 1–confirmed identifications. Open symbols represent specimens that only have morphological data and were assigned to a group using the PCA–linear discriminant analysis model. Numbers in parentheses on each axis indicate the percentage of variation explained.

img-z5-1_37.jpg

Table 2.

Jackknifed confusion matrix of the four lampsiline species to the assignments based on results of the linear discriminant analysis of the principal components of 21 Procrustes-transformed landmark points. Darkened cells represent specimens that were correctly assigned by the linear discriminant analysis (LDA).

img-z5-6_37.gif

Table 3.

Jackknifed confusion matrix of the four lampsiline species and sexes to the assignments based on results of the linear discriminant analysis of the principal components of 21 Procrustes-transformed landmark points. Darkened cells represent specimens that were correctly assigned by the linear discriminant analysis (LDA).

img-z6-2_37.gif

Figure 3.

Deformation grids of two-dimensional shell shape showing difference between the combined mean shape of all specimens and the mean shape of: (A) Lampsilis fasciola, (B) Lampsilis cardium, (C) Ortmanniana ligamentina, and (D) Lampsilis siliquoidea.

img-z6-6_37.jpg

For the three species with distinct sexual dimorphism, males of each species were more accurately assigned by LDA to the correct species and sex than females (Table 3). Overall, 81.0% of males were assigned correctly in the LDA model compared to only 64.7% of females. The greater similarity of females across species could result from convergence of female shape necessary to accommodate the greatly swollen gills of gravid females (Haag 2012, Zieritz and Aldridge 2011, Hewitt et al. 2021).

Figure 4.

Deformation grids of two-dimensional shell shape showing difference between the combined mean shape of all specimens and the mean shape of: (A) female and (B) male Lampsilis fasciola, (C) female and (D) male Lampsilis cardium, (E) Ortmanniana ligamentina, and (F) male and (G) female Lampsilis siliquoidea.

img-z7-1_37.jpg

Using Bayesian clustering, a four-cluster model (model = VEI, BIC = 2,807.75, log-likelihood = 1,461.72) and a seven-cluster model (model = EII, BIC = 2,785.30, log-likelihood = 1,469.78) were created and assessed to determine how they performed in assigning specimens to groups based on their Procrustes valve shapes. The arbitrary groups created by Bayesian clustering were agnostic to the four COI-confirmed species groups and seven COI-confirmed species + sex groups, but performed similarly (79.0% for four groups, 77.8% for seven groups) to the PCA-LDA assignments. The agnostic Bayesian groupings performed similarly to the confirmed groupings, suggesting that patterns of intra- and interspecific variation in the four lampsilines are not necessarily as diagnostic as previously thought and thus require additional characters for species diagnosis (e.g., Mulcrone and Rathbun 2018 and other field identification guides).

The thin-plate splines show that the generalized mean shape across sexes of L. fasciola and L. cardium is more rounded, whereas the mean shape of L. siliquoidea and O. ligamentina is more elongate (Fig. 3). Thin-plate splines also show the truncated and rounded posterior end characteristic of females of the three species with distinct sexual dimorphism (Fig. 4A, C, and G). These shape characteristics match descriptions of the species found in field guides (e.g., Mulcrone and Rathbun 2018).

In contrast to other studies that showed the utility of landmark-based morphometric analysis for species identification (Inoue et al. 2014; Beauchamp et al. 2020; Beyett et al. 2020; Willsie et al. 2020), our results show that this method is of limited utility for these four lampsiline species. Landmark-based morphometric analysis could help improve field identifications of O. ligamentina and L. siliquoidea because it was somewhat useful for differentiating these two species from the other two species we studied. However, the high degree of overlap in shell shape among other species, particularly female L. siliquoidea and female L. cardium, limits the utility of morphometric traits for identification. Improvements to the model could be made by incorporating an assessment of shell variation among different watersheds. Local variation in water chemistry, hydrology, and other factors can influence shell shape, and two distinct shell morphologies of L. fasciola have been described (Watters et al. 2009).

Using two-dimensional landmarks to assess variation in valve shape to differentiate among four species of lampsiline mussels examined in this study has limited utility. Differentiating among more than two species and species with sexual dimorphism was problematic and had error rates between 20% and 30%. In addition to two-dimensional valve shape, we recommend exploring methods for including three-dimensional landmarks that reflect shell inflation. A DNA barcoding–calibrated morphometric key also could be used to examine differences among the closely related species L. cardium, Lampsilis ovata (Say 1817), Lampsilis cariosa (Say 1817), and Lampsilis ornata (Conrad 1835), including potential hybrids of L. cardium and L. ovata (Hewitt et al. 2019) and L. siliquoidea and Lampsilis radiata (Rafinesque 1820) (supposedly restricted to the Lake Ontario, St. Lawrence, and Atlantic Coast drainages; Krebs et al. 2013, Porto-Hannes et al. 2021). Improving the ability to correctly differentiate among species using nongenetic techniques remains important for field biologists. Misidentifications could result in inaccurate population estimates and biases in field surveys, which could in turn mislead conservation and management strategies (Shea et al. 2011).

ACKNOWLEDGMENTS

Funding for this project came from Fisheries and Oceans Canada and the Central Michigan University (CMU) Office of Research and Graduate Studies Summer Scholars Program. Specimens from Michigan were collected using scientific collection and threatened and endangered species permits issued by the Michigan Department of Natural Resources. Ontario specimens were collected under Canadian Species at Risk permit 19-PCAA-0036. We thank everyone who assisted in field collections and laboratory work including Nichelle VanTassel, Shay Keretz, Dylan Powell, Emmett Smrcka, and Julia Willsie (CMU Biology Department) as well as Dr. Patty Gillis and staff from Environment and Climate Change Canada and Fisheries and Oceans Canada. The paper is contribution 164 of the CMU Institute for Great Lakes Research.

LITERATURE CITED

1.

Adams, D. C., M. L. Collyer, A. Kaliontzopoulou, and E. K. Balken. 2021. Geomorph: Software for geometric morphometric analyses. R package version 4.0. Available at https://cran.r-project.org/packages/geomorph(accessed October 12, 2021). Google Scholar

2.

Beauchamp, K. A., T. W. Beyett, M. W. Scott, and D. T. Zanatta. 2020. Detection of hybrid Pyganodon grandis and P. lacustris (Bivalvia: Unionidae) using F- and M-lineage mtDNA sequences and geometric morphometrics. Journal of Molluscan Studies 86:233–239. Google Scholar

3.

Berg, D. J., W. R. Haag, S. I. Guttman, and J. B. Sickel. 1995. Mantle biopsy: A technique for nondestructive tissue-sampling of freshwater mussels. Journal of the North American Benthological Society 14:577–581. Google Scholar

4.

Beyett, T. W., K. McNichols-O'Rourke, T. J. Morris, and D. T. Zanatta. 2020. Use of morphometric analyses and DNA barcoding to distinguish Truncilla donaciformis and Truncilla trucata (Bivalvia: Unionidae). Freshwater Mollusk Biology and Conservation 23:99–108. Google Scholar

5.

Campbell, D. C., P. D. Johnson, J. D. Williams, A. K. Rindsberg, J. M. Serb, K. K. Small, and C. Lydeard. 2008. Identification of ‘extinct’ freshwater mussel species using DNA barcoding. Molecular Ecology Resources 8:711–724. Google Scholar

6.

Collyer, M. L., and D. C. Adams. 2018. RRPP: An R package for fitting linear models to high-dimensional data using residual randomization. Methods in Ecology and Evolution 9:1772–1779. Google Scholar

7.

Collyer, M. L., and D. C. Adams. 2021. RRPP: Linear model evaluation with randomized residuals in a permutation procedure. Available at https://cran.r-project.org/web/packages/RRPP(accessed October 12, 2021). Google Scholar

8.

COSEWIC (Committee on the Status of Endangered Wildlife in Canada). 2010. COSEWIC assessment and status update report on the Wavy-rayed Lampmussel Lampsilis fasciola in Canada. Committee on the Status of Endangered Wildlife in Canada. Ottawa. Xi+ 60 pp. Available at https://wildlife-species.canada.ca/species-risk-registry/virtual_sara/files/cosewic/sr_Wavy-rayed%20Lampmussel_0810_e.pdf(accessed February 11, 2022). Google Scholar

9.

Cummings, K. S., and C. A. Mayer. 1992. Field Guide to Freshwater Mussels of the Midwest. Illinois Natural History Survey Manual 5, Champaign. 194 pp. Google Scholar

10.

Gillis, P. L., J. Salerno, V. L. McKay, C. J. Bennett, K. L. K. Lemon, Q. J. Rochfort, and R. S. Prosser. 2021. Salt-laden winter runoff and freshwater mussels; Assessing the effect on early life stages in the laboratory and wild mussel populations in receiving waters. Archives of Environmental Contamination and Toxicology 82:239–254. Google Scholar

11.

Gilroy, E. A. M., J. S. Klinck, S. D. Campbell, R. McInnis, P. L. Gillis, and S. R. de Solla. 2014. Toxicity and bioconcentration of the pharmaceuticals moxifloxacin, rosuvastatin, and drospirenone to the unionid mussel Lampsilis siliquoidea. Science of the Total Environment 487:537–544. Google Scholar

12.

Goodall, C. 1991. Procrustes methods in the statistical analysis of shape. Journal of the Royal Statistical Society B 53:285–339. Google Scholar

13.

Haag, W. R. 2012. North American freshwater mussels: Natural history, ecology, and conservation. Cambridge University Press, New York. 505 pp. Google Scholar

14.

Hammer, Ø., D. A. T. Harper, and P. D. Ryan. 2001. PAST: Paleontological statistics software package for education and data analysis. Palaeontologia Electronica 4:9. Software available at https://www.nhm.uio.no/english/research/infrastructure/past/(accessed February 11, 2022). Google Scholar

15.

Hebert, P. D. N., A. Cywinska, S. L. Ball, and J. R. Dewaard. 2003. Biological identifications through DNA barcodes. Proceedings of the Royal Society of London Series B: Biological Sciences 270:313–321. Google Scholar

16.

Henley, W. F., P. J. Grobler, and R. J. Neves. 2006. Non-invasive method to obtain DNA from freshwater mussels (Bivalvia: Unionidae). Journal of Shellfish Research 25:975–977. Google Scholar

17.

Hess, M. C., K. Inoue, E. T. Tsakiris, M. Hart, J. Morton, J. Dudding, C. R. Robertson, and C. R. Randklev. 2018. Misidentification of sex for Lampsilis teres, Yellow Sandshell, and its implications for mussel conservation and wildlife management. PloS ONE 13:e0197107. Google Scholar

18.

Hewitt, T. L., A. E. Haponski, D. Ó Foighil. 2021. Evolution of diverse host infection mechanisms delineates an adaptive radiation of lampsiline freshwater mussels centered on their larval ecology. PeerJ 9:e12287. Google Scholar

19.

Hewitt, T. L., D. A. Woolnough, and D. T. Zanatta. 2019. Population genetic analyses of Lampsilis cardium (Bivalvia: Unionidae) reveal multiple post-glacial colonization routes into the Great Lakes drainage. American Malacological Bulletin 37:21–34. Google Scholar

20.

Inoue, K., D. M. Hayes, J. L. Harris, and A. D. Christian. 2013. Phylogenetic and morphometric analyses reveal ecophenotypic plasticity in freshwater mussels Obovaria jackoniana and Villosa arkansasensis (Bivalvia: Unionidae). Ecology and Evolution 3:2670–2683. Google Scholar

21.

Inoue, K., A. L. McQueen, J. L. Harris, and D. J. Berg. 2014. Molecular phylogenetics and morphological variation reveal recent speciation in freshwater mussels of the genera Arcidens and Arkansia (Bivalvia: Unionidae). Biological Journal of the Linnean Society 112:535–545. Google Scholar

22.

Jackson, D. A. 1993. Stopping rules in principal components analysis: a comparison of heuristical and statistical approaches. Ecology 74:2204–2214. Google Scholar

23.

Krebs, R. A., W. C. Borden, N. M. Evans, and F. P. Doerder. 2013. Differences in population structure estimated within maternally- and paternally-inherited forms of mitochondria in Lampsilis siliquoidea (Bivalvia: Unionidae). Biological Journal of the Linnean Society 109:229–240. Google Scholar

24.

Larson, J. H., W. B. Richardson, R. J. Kennedy, J. C. Nelson, M. A. Evans, and J. S. Schaeffer. 2016. Spatial variation in biofouling of a unionid mussel (Lampsilis siliquoidea) across the western basin of Lake Erie. American Midland Naturalist 176:119–129. Google Scholar

25.

MNFI (Michigan Natural Features Inventory). 2020. Michigan's rare animals. Available at https://mnfi.anr.msu.edu/species/animals(accessed February 11, 2022). Google Scholar

26.

Mulcrone, R. S., and J. E. Rathbun. 2018. Field Guide to the Freshwater Mussels of Michigan. Michigan Department of Natural Resources, Lansing. 59 pp. Google Scholar

27.

OMNRF (Ontario Ministry of Natural Resources and Forestry). 2021. Species at risk in Ontario. Available at https://www.ontario.ca/page/species-risk-ontario(accessed February 11, 2022). Google Scholar

28.

Porto-Hannes, I., L. E. Burlakova, D. T. Zanatta, and H. R. Lasker. 2021. Boundaries and hybridization in a secondary contact zone between freshwater mussel species (Family: Unionidae). Heredity 126:955–973. Google Scholar

29.

Quinn, G. P., and M. J. Keough. 2002. Experimental Design and Data Analysis for Biologists. Cambridge University Press, New York. 557 pp. Google Scholar

30.

Rohlf, J. F., and D. Slice. 1990. Extensions of the Procrustes method for the optimal superimposition of landmarks. Systematic Biology 39:40–59. Google Scholar

31.

Sambrook, J., E. F. Fritsch, and T. Maniatis. 1989. Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York. Google Scholar

32.

Scrucca, l., M. Fop, T. B. Murphy, and A. E. Raftery. 2016. Mclust 5: Clustering, classification and density estimation using Gaussian finite mixture models. The R Journal 8:289–317. Google Scholar

33.

Shea, C. P., J. T. Peterson, J. M. Wisniewski, and N. A. Johnson. 2011. Misidentification of freshwater mussel species (Bivalvia:Unionidae): Contributing factors, management implications, and potential solutions. Journal of the North American Benthological Society 30:446–458. Google Scholar

34.

Sheets, H. D. 2014. Integrated Morphometrics Package (IMP) v. 8. Canisius College, New York. Available at: https://www.animal-behaviour.de/imp/(accessed November 12, 2020). Google Scholar

35.

Waller, D. L., and M. R. Bartsch. 2018. Use of carbon dioxide in zebra mussel (Dreissena polymorpha) control and safety to a native freshwater mussel (Fatmucket, Lampsilis siliquoidea). Management of Biological Invasions 9:439–450. Google Scholar

36.

Wang, N., R. A. Consbrock, C. G. Ingersoll, and M. C. Barnhart. 2011. Evaluation of influence of sediment on the sensitivity of a unionid mussel (Lampsilis siliquoidea) to ammonia in 28-day water exposures. Environmental Toxicology and Chemistry 30:2270–2276. Google Scholar

37.

Watters, G. T., M. A. Hoggarth, and D. H. Stansbery. 2009. The Freshwater Mussels of Ohio. The Ohio State University, Columbus. 400 pp. Google Scholar

38.

Webster, M., and H. D. Sheets. 2010. A practical introduction to landmark-based geometric morphometrics. Pages 163–188 in J. Alroy and G. Hunt, editors. Quantitative Methods in Paleontology. Vol. 16. Yale University Press, New Haven, Connecticut. Google Scholar

39.

Willsie, J. A., T. J. Morris, and D. T. Zanatta. 2020. Morphometric analyses distinguish Wabash Pigtoe (Fusconaia flava) and Round Pigtoe (Pleurobema sintoxia) mussels. Diversity 12:337. Google Scholar

40.

Zieritz, A., and D. C. Aldridge. 2011. Sexual, habitat-constrained and parasite-induced dimorphism in the shell of a freshwater mussel (Anodonta anatina, Unionidae). Journal of Morphology 272:1365–1375. Google Scholar

Appendices

Appendix 1.

Length, height, width, and hinge-length measurements and field, cytochrome c oxidase subunit 1, and (jackknifed) morphometric identifications for all specimens collected.

img-z10-2_37.gif

continued.

img-z11-1_37.gif

continued.

img-z12-1_37.gif

continued.

img-z13-1_37.gif

continued.

img-z14-1_37.gif

continued.

img-z15-1_37.gif

continued.

img-z16-1_37.gif

continued.

img-z17-1_37.gif
© Freshwater Mollusk Conservation Society 2022
Madison R. Layer, Russell L. Minton, Todd J. Morris, and David T. Zanatta "Utility of Shell-Valve Outlines for Distinguishing among Four Lampsiline Mussel Species (Bivalvia: Unionidae) in the Great Lakes Region," Freshwater Mollusk Biology and Conservation 25(1), 37-53, (29 March 2022). https://doi.org/10.31931/fmbc-d-21-00007
Published: 29 March 2022
JOURNAL ARTICLE
17 PAGES


Share
SHARE
KEYWORDS
DNA barcoding
geometric morphometrics
Species at risk
RIGHTS & PERMISSIONS
Get copyright permission
Back to Top