HOOPES: SELECTION OF SPAWNING SITES 



streambed with a shovel and placed in a cotton 

 soil sample bag. It was later washed into indi- 

 vidual 19-liter tins, drained, and allowed to air 

 dry. The dried samples were separated into 10 

 size classes by passing them through a stacked 

 series of square-meshed 20-cm U.S. Standard 

 sieves. After the sample was placed on the top 

 sieve, the series was agitated for 10 min in a 

 hand-powered portable sieve shaker. The ma- 

 terials retained on each sieve were weighed to 

 the nearest gram on a triple-beam trip scale. 



I describe the particle size distribution of the 

 streambed by relating the percent of the total 

 sample weight retained on each sieve to the 

 logarithm of its mesh size. This method of anal- 

 ysis is described more fully by Spangler (1951) 

 and by the American Society for Testing Ma- 

 terials (1958). 



Because the particle composition of streambed 

 samples from pools and riffles was so similar 

 within each section, I combined the data by sec- 

 tion. I then derived a cumulative size distribu- 

 tion for each section by summing the weights 

 retained by each sieve size in all samples for a 

 section. These data were used to construct 

 cumulative distribution curves with which I 

 could compare the composition of the stream 

 bottom in sections of high spawning density 

 (Figure 6) with that in sections of low spawning 

 density (Figure 7). 



The curves for sections of high spawning den- 

 sity (Figure 6) show that bottom composition 

 was quite similar in all three streams. The only 

 difference among the four sections was the 

 slightly higher percentage of small particles and 

 the somewhat lower percentage of large particles 

 in section 9 of Hidden Creek, as evidenced by 

 the lateral displacement toward the left of the 

 composition curve for this section. 



The cumulative curves for sections with low 

 spawning density (Figure 7) showed that par- 

 ticle size distribution was similar in three of the 

 sections (one from each stream) but quite dif- 

 ferent in the fourth section (Hidden Creek, 2). 

 It is clear that spawner density was low in Hid- 

 den Creek in areas where particle sizes were 

 either very large or very small. 



Sockeye salmon spawned most frequently in 

 areas where the particles were intermediate in 

 size. In sections of high spawning density and 

 intermediate gradient, cobbles larger than 7.6 cm 

 in diameter made up about 6^f of the bottom, 

 cobbles from 2.5 to 7.6 cm about 50%, particles 

 1.3 to 2.5 cm about 20%, and particles less than 

 1.3 cm the remainder (Figure 8A). The bot- 

 tom in sections of low spawning density and high 

 gradient was made up of almost 40% cobbles 

 exceeding 7.6 cm in diameter and contained very 

 little material less than 2.5 cm in diameter (Fig- 

 ure 8B) . The bottom in sections of low spawn- 



PARTICLE SIZE IN INCHES. LOG SCALE 



-T 

 .01 .1 



PARTICLE SIZE ININCHES.LOG SCALE 



Figure 6. — Curves of cumulative particle size distribu- 

 tion in four sections with high-spawning density. Hid- 

 den Creek sections 9 and 33, Up-a-tree Creek section 17, 

 and One Shot Creek section 4. All four sections had 

 intermediate stream gradients. 



Figure 7. — Curves of cumulative particle size distribu- 

 tion in four sections with low-spawning density. Hidden 

 Creek sections 2 and 24, Up-a-tree Creek section 15, and 

 One Shot Creek section 8. All of the sections had high 

 gradients except Hidden Creek section 2, which had low. 



453 



