Date of Award


Level of Access Assigned by Author

Campus-Only Thesis

Degree Name

Master of Science (MS)




Michael T. Kinnison

Second Committee Member

Judith A. Rhymer

Third Committee Member

John Kocik


Conservation units for protection of endangered or threatened species under the United States Endangered Species Act of 1973 often incorporate subsets of partly isolated populations. These populations, however, may differ in important trait variation critical to the evolutionary legacy of the species and recovery of self-sustaining populations in the wild. Unfortunately, managers rarely possess sufficient knowledge of such trait variation to preserve locally adapted trait variation while also minimizing threats of inbreeding and drift in small isolated populations. Compounding this problem, managers may often not know which trait variation is most useful for assessing adaptive divergence and planning subsequent management strategies. In this study I compare patterns of heritable, trait variation in seven populations of Maine Atlantic salmon that have experienced precipitous population declines. Six of these populations have been listed as endangered within a single conservation unit, the Gulf of Maine Distinct Population Segment (GMDPS), and are currently managed at the Craig Brook National Fish Hatchery. In particular, I consider life history trade-offs as a framework for assessing important adaptive trait variation within the GMDPS. Because most of these populations were reared to maturity under similar captive conditions, I was better able to control for some of the confounding issues that may have limited other studies. I found that the salmon populations in Maine differ in their total ovarian investment, their relative investment into egg size and egg number, hatching time and larval growth. These differences suggest that the populations currently subsumed in the GMDPS merit continued conservation at the population level. However, if the populations continue to decline and it becomes necessary to mix them to some degree to reduce risks of inbreeding depression, my findings may be used to conduct such mixing in a fashion that will present the least risk to disruption of locally adaptive trait variation. This research provides an example of how assessment of adaptive trait variation can become an integral part of an endangered species management program and I provide suggestions for future research that might further enhance manager's abilities to preserve endangered salmon and other species.