28 



were read and only those fish with Lyoni Terry Hatchery codee were 

 allowed to enter the Lyons ferry brood stock from 1990 on. Eggs and 

 sperm taken from non-Lyons Ferry tagged fish were transferred to a 

 downstream "mongrel" hatchery. Further, the mixed ancestry fish from 

 the Lyons Ferry Hatchery brood stock spawned in 1989 were all marked 

 with coded wire tags and returns from this tagged lot of fish have not 

 been allowed into future generations of the brood stock. Also, the fish 

 released from Lyons Ferry Hatchery and the fish released in the Umatilla 

 River (the predominant population of Columbia River strays) since that 

 time were all marked with coded wire tags to better enable an assessment 

 of future straying. 



An effort has been made to trap as many fish as possible at Lower 

 Granite Dam and remove those fish with coded wire tags in an attempt to 

 control straying of hatchery fish on the spawning grounds. This program 

 has only been partially successful because the Lower Granite trap is not 

 100% efficient; and further, strays frora other Columbia River hatcheries 

 are not 100% coded wire tagged. As a result, the straying rate for 

 Snake River hatchery fish past Lower Granite Dam reached its lowest 

 level in 1993 (4%) while the straying rate of Columbia River hatchery 

 fish has only been partially abated (see Mundy 1994 for a detailed 

 description of the results of this effort). In 1993, more Columbia 

 River hatchery strays (167 fish; 18% of the total escapement) migrated 

 past Lower Granite Daa than did Snake River hatchery strays (43 fish; 4% 

 of the total escapement). Further, because of the success of this 

 program in stemming the Snake River hatchery strays, the potential 

 effect of Columbia River hatchery strays altering the gene pool has 

 increased (i.e. the relative contribution of Columbia River strays on 

 the population of fall chinook spawning in the wild has increased) . 



Although the escapement numbers used to represent "natural" spawners in 

 the ESU have been adjusted to account for both Columbia River and Snake 

 River hatchery strays, the potential effect of this straying on the 

 "gene pool" has not been adequately evaluated. Use of the escapement 

 estimates of "natural" fish only in evaluations of escapement trends 

 past Lower Granite Dam only partially addresses the potential problem. 

 These estimates of the "natural" escapement are only realistic if 

 hatchery strays are entirely unsuccessful in reproducing themselves 

 (i.e., their fitness is zero). If fitness is anything other than zero, 



the effect of straying is cumulative. In other words, the proportion of 



"stray" genes in the population at any one time is a function of both 



fitness and the additive level of straying that continues to occur each 

 year. 



A simple dilution model was developed to better evaluate the effect of 

 straying on the "natural" escapement gene pool. The starting assumption 

 was that no straying of Snake River hatchery fish occurred prior to 1983 

 and no straying of Columbia River hatchery fish occurred prior to 1984. 

 Although these are the first years when straying was documented, it must 

 be remembered that prior to this time, coded wire tag technology was not 

 available to detect strays. It seems likely that strays, both from the 

 Snake River hatchery program (Hagerman Hatchery) and from the Columbia 

 River hatchery program (various hatcheries) likely entered the Snake 

 River spawning grounds prior to 1983. Thus the dilution model will 

 likely underestimate true effects. A second assumption used in the 

 dilution model was that random mating occurred ^unong all fish in the 

 escapement. Annual stray rates as documented in Table 4 were used along 

 with an assumed age composition for returns of 30% age 3, 56% age 4, and 

 14% age 5. Fitness values of 0.0, 0.5, and 1.0 for all hatchery strays 

 and for just Columbia River hatchery strays were used through this time 

 series dilution model to predict the composition of the current gene 

 pool . 



The additive effects of continued straying on the gene pool if fitness 

 of strays is other than rero is readily apparent from this simple 

 dilution model (Table 14 and Figure 14). If both Snake River and 

 Columbia River hatchery strays are a concern, fitness values of 0.5 and 

 1.0 result in the 1994 gene pool being composed of only 24% and 10% 

 "natural" genes; respectively. If only Columbia River hatchery strays 

 are a concern, fitness values of 0.5 and 1.0 result in the 1993 gene 

 pool being composed of only 78% and 62% "natural" genes; respectively. 

 It seems unlikely that fitness of strays is zero and it seems unlikely 



