Creation of a Captive Broodstock Program for Southern Coho Salmon (Oncorhynchus kisutch): Results from the Initial Rearing and Spawning of the First Brood Year

摘要

Coho salmon (Oncorhynchus kisutch) from Scott Creek, Santa Cruz County, CA, USA, were captured as adults in the winter of 2001/02 to establish a captive broodstock program designed to reduce the risk of extinction of coho salmon at the southern end of their range following years of precipitous declines. These fish were spawned with a random mating design, and 154 progeny from these crosses were selected as founding captive broodstock fish using an ad hoc method based upon molecular genetic variation. The captive broodstock fish were divided into 3 groups of approximately 50 fish and were reared in seawater and freshwater at two different locations until just before reproductive maturity, when they were all consolidated at the freshwater facility for spawning. Total mortality (21%) was substantially less than the 50% expected from previous coho salmon captive broodstock programs. However, captive broodstock fish were 30-50% shorter and lighter than wild coho salmon from Scott Creek upon maturity in January 2005. The captive broodstock fish also had inferior measures of all reproductive traits, except gonosomatic index, relative to wild fish. In addition, significant differences in both size and reproductive traits were noted between fish reared in seawater and those reared in freshwater, with the freshwater-reared fish growing much more quickly during the last six months prior to spawning. Females from both rearing environments did not mature on the same schedule as the wild coho salmon and the majority of the captive broodstock fish, and they were treated with an experimental gonadotropin releasing hormone analogue. This quickly induced maturity, but the majority of eggs from these females did not survive. To create the next generation of captive broodstock fish, all mature wild and captive broodstock fish were mated together using a genetically-defined breeding matrix that relied upon the pair-wise coefficients of relatedness between all potential mating pairs, as estimated with molecular genotype data from 18 microsatellite loci. This genetic broodstock management used a rapid genotyping effort for all returning adults and the iterative construction of breeding matrices to include all potentially available mates for each weeks spawning event. The genetic matrices were not able to be fully utilized to create optimal crosses, due to asynchronous maturity and other logistical issues, but the mean coefficient of relatedness between mated pairs and the number of highly inbred matings were both dramatically reduced, relative to random mating.