USING SCANNING ELECTRON MICROSCOPY TO IDENTIFY MUSSEL LARVAE

Jessica N. Ramirez, Macalester College
Sarah L. Boyer, Macalester College
Mark C. Hove, Macalester College
Daniel J. Hornbach, Macalester College

Freshwater mussels in North America are of special interest to conservationists because nearly 70% of all known species are listed as special concern, threatened or endangered. This project focuses on distinguishing two species, Actinonaias ligamentina and Lampsilis higginsii, found in the Mississippi and St. Croix rivers. Identifying L. higginsii is of particular importance because this species is federally endangered.

At the larval stage freshwater mussels are hard to distinguish morphologically, because of their minute size, so DNA barcoding is a useful method for species identification. Unfortunately, using mitochondrial DNA has proven to be ineffective in distinguishing between L. higginsii and A. ligamentina, as is shown by their placement in previously generated phylogenetic trees. Since DNA barcoding is not an effective tool in this case, an alternative approach is needed. Morphology of these animals can be described using scanning electron microscopy (SEM), which allows imaging and measurement of minute specimens.

Glochidia were studied from five A. ligamentina females and four L. higginsii females. Several hundred glochidia from each female were mounted onto a stub and studied using SEM. The following measurements were made from approximately ten glochidia from each stub: hinge length, height, length, hinge:length ratio, height ratio and length:height ratio. We used JMP software to identify statistically significant differences between the species in size and shape.

The results indicated that there were significant differences between A. ligamentina and L. higginsii glochidial length (p=0.02), height (p=0.0005) and hinge:length ratio (p<=0.04). In addition, discriminate analysis was performed using JMP software to demonstrate the efficacy of using shell measures to distinguish the species. A discriminate analysis attempts to predict a classification variable (species) based on known continuous responses (glochidial size and shape). According to our analysis, there is a 90% probability of correctly classifying an unknown specimen as A. ligamentina and an 86% probability of correctly classifying an unknown specimen as L. higginsii using the glochidial measurements we made.