432 



FISHERY BULLETIN OF THE FISH AND WILDLIFE SERVICE 



nearest hundredth gram. Eggs were weighed and 

 enumerated from cross-sectional samples from 

 each ovary combined to approxinuite 10 percent 

 of the total egg weight. Total egg count was 

 then computed on a proportional weight-count 

 basis, with tlie assumption tliat the sample was 

 representative of the entire ovary. 

 Spawning Ground Surveys 



Spawning grounds of the Okanogan River have 

 been surveyed each year since 1947, except 1950. 

 Gangmark and Fulton (1952) describe the method 

 used to estimate the size of spawning populations. 



Estimates of age and sex composition, com- 

 puted for the years 1953 and 1956 through 1959, 

 are based on standarcUzed collections of fresh 

 dead fish, which were picked up in proportion to 

 their availability at intervals throughout the 

 entire period when spawners are dying. 



The sampling methods described herein are 

 not rigorous from a statistical viewpoint. Yet we 

 believe them to be adequate for purposes of this 

 report, which represents the difficult transition 

 from the qualitative to quantitative stages of 

 investigation. Under appropriate headings of the 

 next section are discussed certain preliminarj' 

 tests which are pertinent to the sampling and to 

 the data tliereby derived. 



RESULTS 

 Age Group Fecundity 



In the previous section we indicate that because 

 the catch contains 42's from both the Wenatchee 

 and Okanogan populations, we can merely assume 

 that mean fecundity estimates derived from this 

 source are representative of the Okanogan popula- 

 tion. The assumption is supported by the fol- 

 lowing test: 



We hypothesize that no fecundity difference 

 exists between the 42's of the two river systems. 

 Comparison of the 1957 sample-means of 16 

 Okanogan and 17 Wenatchee fish by t-test reveals 

 no significant difference at the 5 percent level 

 (t=0.15 with 15 d.f.). On the basis of this 

 evidence the hypothesis is not rejected, and, with 

 reasonable assurance that it holds, we assume tliat 

 the fecundity difference is not large and that 

 estimates derived from the catch will accurately 

 represent either area. 



To determine whether a bias exists in oui- 

 method of estimating the mean egg content of 

 individual fisli, we used a t-test to compare the 

 actual and estimated mean egg content of a 

 randondy selected subsample of 21 fish (table 3). 

 The average difference between the actual and 

 estimated means is 40 eggs, the standard error of 

 the mean, 27 eggs. Although it appears that we 

 are overestimating, tlie dift'erence is not significant 

 at the 5 percent level (t=1.49 with 20 d.f.). 



The results of the fecundity study are shown in 

 table 4 and figure 2. Of the statistics shown, the 

 mean and range figures provide the basis for age 

 group comparison, and the standard deviation 

 and tlie standard error of the mean are presented 

 merely as points of additional interest. Differ- 

 ences between the numbers of fish reported in 

 table 4 or figure 2 and those reported in the original 

 sample are due to the omission of aberrant speci- 

 mens (2 females, 1 each in 1957 and 1959, had but 

 a single ovary) and tliose fish whose ages were 

 indeterminable because of scale erosion. 



Examination of table 4 or figure 2 discloses 

 that the egg content of individual fish witiiin an 

 age group sample varied by as much as 2,749 

 eggs. Little between-year variation is evident for 

 the means of a single group. Tlie variation be- 

 tween the means of the 42's and the 53's, different 

 age groups having similar ocean age, is also small. 



Table 3. — Estimated and actual egg counts of a randomly 

 selected subsample, 1957-59 



