Hickory Shad (Alosa mediocris) Population Identification Using Geometric Morphometrics and Otolith Shape
Loading...
Date
2019-08-05
Access
Authors
Meyer, Steve D.
Journal Title
Journal ISSN
Volume Title
Publisher
East Carolina University
Abstract
Hickory Shad Alosa mediocris is an anadromous species of fish belonging to the Family Clupeidae, first described in 1814 by Samuel L. Mitchill. Little is known about its life history or stock status. It is unknown whether Hickory Shad exhibit natal homing observed in other anadromous species. Fishery management agencies like the North Carolina Wildlife Resources Commission require more information pertaining to Hickory Shad basic life history to better manage the species. To provide information on distribution of stocks of the species, we used a suite of methods including meristics, morphometrics, otolith shape, otolith microchemistry, genetics, and body shape (geometric morphometrics). This thesis covers the body shape and otolith shape analysis. Photographs were taken of 757 Hickory Shad samples from 21 watersheds for body shape analysis. Landmarks were digitized using the software TPSDig2 and differences in body shape were determined by landmark anaylsis using the software MorphoJ. Results of Analysis of Variance (ANOVA) and Discriminant Function Analysis (DFA) determined highly significant differences (p<0.0001) in body shape between male and female samples. Principle Components Analysis (PCA) determined that body depth comrised the majority of the variation between sexes. Therefore, all analyses were run on male and female samples separately. DFA and ANOVA detected varying levels of difference from highly significant difference (p<0.0001) to no significant difference (p>0.05) between males and females of different states, watersheds, and tributaries within watersheds. Out of 153 pairwise comparisons of females by watersheds, 43 (28.1%) resulted in highly significant differences (p<0.0001), and 99 (64.7%) resulted in significant differences (p<0.05). Out of 153 pairwise comparisons of males by watersheds, 102 (66.7%) resulted in significant differences (p<0.05), 44 (28.8%) of which were highly significant differences (p<0.0001), and DFA on females correctly classified watershed of capture by pairwise discriminant funtion ranging from 88.2% to 100% and correctly classified watershed of capture by cross validation ranging from 12.5% to 100%. DFA on males correctly classified watershed of capture by pairwise discriminant funtion ranging from 83.3% to 100% and correctly classified watershed of capture by pairwise cross validation ranging from 33.3% to 100%. Type I error in watershed pairwise comparisons of body shape using DFA was controlled using Benjamini and Hochberg's false discovery rate (FDR)-controlling procedure. All 153 watershed pairwise comparison results were adjusted for both males and females. In female specimen, adjusted P-values resulted in 95 or 62.1% significant differences (p<0.05) as compared to 99 or 64.7% significant differences before adjusting. In male specimen, adjusted P-values resulted in 97 or 63.4% significant differences (p<0.05) as compared to 102 or 66.7% significant differences before adjusting. PCA determined that body depth was the majority of variation in body shape. The most heavily weighted landmark positions were PC1 = y13 and y14, PC2 = y6 and y7, and PC3 = x13 and x14. Landmarks 6 and 7 are the origin and insertion of the dorsal fin and landmarks 13 and 14 are the insertion and origin of the pelvic fin. Differences in body shape are thought to be affected by a combination of environmental and genetic factors. Inconsistency and varying results in both male and female samples to discriminate watersheds using body shape led to the conclusion that the methods used here were not dependable for differentiating spawning populations of Hickory Shad. However, more than half of the watershed pairwise comparisons of body shape were significantly different. This suggests that homing to natal tributaries is plausible. Photographs were taken of 696 right sagittal otoliths from 22 watersheds for otolith shape analysis. Shapes of the otolith outlines were generated and analyzed using the ShapeR package in RStudio. ANOVA determined there was highly significant differences in otolith shape by watershed (F21, 674 =3.4242, p=0.001) and by state (F6, 689=5.0122, p=0.001). Most of the variation among otolith shape can be attributed to the region including the antirostrum, excisura major, and dorsal side of the rostrum. Pairwise comparisons of otolith shape for every watershed using an ANOVA-like permutation test resulted in P-values ranging from highly significant (p=0.001) in 55 pairwise comparisons of watersheds, to (p=0.902) in the comparison of the Rappahannock River and the Potomac River. Out of 231 total pairwise comparisons of watersheds, 148 or 64.1% resulted in significant differences (p<0.05), 55 or 23.8% of which were highly significant differences (p=0.001). Type I error in watershed pairwise comparisons of otolith shape using ANOVA was controlled using Benjamini and Hochberg's false discovery rate (FDR)-controlling procedure. Benjamini and Hochberg adjusted P-values resulted in 131 or 56.7% significant differences (p<0.05) as compared to 148 or 64.1% significant differences before adjusting. Differences in otolith shape were hypothesized to be influenced by a combination of environmental and genetic factors. Varying results of this study attempting to discriminate sample sets captured in 22 watersheds by otolith shape led to the conclusion that the methods used here were not reliable for differentiating spawning populations of Hickory Shad; however, the fact that more than half of the otolith shape comparisons were significantly different suggests that homing to natal tributaries is plausible. More sampling and standardization of sample size, timing, and age may help in differentiating spawning populations of Hickoy Shad using otolith shape and body shape.