FISHERY BULLETIN: VOL. 74, NO. 4 



device separated a larger subsample from the total 

 numbers for marking. 



We calibrated the cone device as a sampler with 

 which to estimate total numbers of emerging fry 

 from the incubators. Inspection of the relationship 

 of total fry emigrating from an incubator (y) 

 plotted against fry retained by the sampler (x) on 

 24 occasions indicated a constant ratio (straight 

 line through the origin) with increasing variation 

 at higher subsampler counts (Figure 3). Con- 

 sequently, the average of the 24 ratios (y/x) 

 available from the calibration study is taken as the 

 slope estimate (Snedecor 1956: 153-156) and was 

 calculated as 24.537 (SE: 1.072). The major portion 

 of the fry passed the cone sampler and were then 

 routed through the parallel troughs, one of which 

 emptied into a holding tank and four of which 

 emptied into the hatchery drain and then into 

 Auke Creek. With these two devices we captured 

 about one-fourth of the gravel incubator fry each 

 day without impeding the seward migration of the 

 other three-fourths. 



Twice weekly, samples of 50 fry from each 

 gravel incubator and the fyke nets were preserved 

 in 5% Formalin. The preserved fry were allowed to 

 stand for 6 wk before lengths were measured to 

 the nearest millimeter and wet weights to the 

 nearest milligram. An inde.x to stage of develop- 

 ment (Bams 1970) of the fry was computed from 

 the formula 



^ 10 ^ weight in milligrams 



^D — . 



length in millimeters 



This index is used only on unfed fry to indicate the 

 relative yolk content. It is not a condition factor. 



Weighted means and variances of pooled data 

 were computed on the basis of the fraction of the 

 migrant fry represented by each sample. Statis- 

 tical comparisons were made of lengths, wet 

 weights, and developmental index as follows: 



where Y,,. 

 W. 



= weighted mean 



= observed mean measurement in ith 



period 

 = proportionofrunleavingin/th period 



from index sampling, and 



V{Y,y)= J^ WfViY^) 



100 200 300 400 



NUMBER OF FRY IN SUBSAMPLER (x) 



Figure 3.- Relation of number of fry migrating to number of fry 

 in subsampler. Each point represents one sample. 



where F(r„) 

 V{Y,) 



n = 



variance estimate of weighted 



mean 



sample variance of estimated 



mean in ('th period 



number of periods sampled. 



i = 1 



Technicians marked hatchery fry by clipping 

 the adipose and left ventral fins and creek fry by 

 clipping the adipose and right ventral fins. Before 

 marking, the fry were anesthetized in a solution of 

 1:7,500 MS-222 (Tricaine methanesulfonate) 

 buffered with sodium bicarbonate to pH 6.1-6.4; 

 the solution was kept cool in a water table and 

 recirculated to keep the oxygen content high. 

 Surgical iris scissors were used to excise fins under 

 a 3x magnifying lens. Technicians marked an 

 average of about 200 fry/h on this study, whereas 

 technicians marked about 400 fry/h on a similar 

 study in Canada (R. A. Bams pers. commun.). 

 Samples of fry from each technician were ex- 

 amined several times daily to ensure that the 

 correct fins were excised as close to the body as 

 possible. All marked fry were released at 1130 h 

 the same day they were marked; most of the 

 unmarked fry that had left the incubator or the 

 stream at the same time had migrated seaward 24 

 h earlier. Dead fry remaining in the release tank 

 were counted each morning. The immediate mor- 

 tality from marking was less than 0.1% for both 

 hatchery and creek fry. Totals of marked fry 

 released were 84,000 from the hatchery and 5,500 



964 



