larger trout produce more eggs and larger eggs 

 than do the smaller trout, we perceive differ- 

 ences in eggs that are independent of the 

 differences in fish size between streams . 



Table 2 is based on Yellowstone hatchery 

 egg-taking records for the years 1941 through 

 1952 . We do not have length data for all 

 streams for comparison with the numbers in 

 table 2, but available measurements indicate 

 that Chipmunk Creek fish average among the 

 largest, Pelican and Arnica Creek fish are 

 small, and Grouse Creek fish are intermediate 

 in size. If fish of these streams were all of 

 the same race, we would then expect ChipmunK 

 Creek females to have the largest eggs and the 

 greatest number per female, for Grouse Creek 

 to be intermediate in these relationships, and 

 for Pelican Creek and Arnica Creek females to 

 have small eggs and small numbers per female. 

 The table shows that ChipmunK Creek does, in- 

 deed, have the largest eggs (262 per ounce) and 

 Pelican Creek the smallest (307 per ounce), and 

 that Grouse CreeK occupies an intermediate 

 position (286 per ounce). Arnica Creek, how- 

 ever, with 268 eggs per ounce, has the second 

 largest eggs on the list, rather than having the 

 small eggs that the fish size would lead us to 

 expect. Chipmunk Creek has a relatively small 

 number of eggs per female (981) instead of the 

 large number we would expect . Pelican and 

 Arnica Creeks, with 979 and 899, respectively, 

 have the low numbers we would expect (but 

 with rather large discrepancy between them), 

 and Grouse Creek has the expected intermedi- 

 ate number (1,007) . We are faced with the fact 

 that we do not have length measurements in 

 these streams for all the years from 1941 

 through 1952, but measurements are available 

 for several of these years, and the relative 

 sizes of fish in these streams have been very 

 consistent through these years. 



Figure 4 plots regressions of total num- 

 bers of eggs on ovary weight for three streams . 

 These relationships are based on measurements 

 made by Dr. Stillman Wright in Pelican, Chip- 

 munk, and Grouse Creeks in 1945. The trends 

 show that Pelican Creek fish have more eggs 

 per ovary weight than those of ChipmunK Creek, 

 and Chipmunk Creek fish more than those of 

 Grouse Creek. This is evidently not related to 

 size of fish, since Chipmunk Creek fish are 



larger than are Grouse Creek fish. 



Coefficients of regression were calcu- 

 lated for each of the regressions in figure 4, 

 and then compared according to the method 

 described by Simpson and Roe (1939), page279. 

 These tests, for significance of differences be- 

 tween regression coefficients, resulted in the 

 following t values: Grouse Creek vs. Pelican 

 Creek - 17.207; Grouse Creek vs. ChipmunK 

 Creek - 8.774; Chipmunk Creek vs. Pelican 

 Creek - 8.166. These values indicate that the 

 relationship for each stream is very significant- 

 ly different from that of each of the other two 

 streams. 



These relationships involving eggs ap- 

 pear to show that the fish of certain streams 

 have specific characteristics that are different 

 from those of adjacent streams as well as re- 

 mote streams. 



SUMMARY 



Yellowstone Lake and the Upper Yellow- 

 stone Paver, and possibly part of the Upper 

 Snake River drainage, form a closed system . 

 The Yellowstone cutthroat is held within these 

 waters, and no introduction has been made within 

 the history of fishery work . These fish are thus 

 suited to studies on races. Several spawning 

 tributaries to Yellowstone Lake and the river 

 have been studied to determine the existence of 

 distinct races in this drainage. 



The association of groups of fish with 

 particular spawning streams has been established 

 through tagging. Homing to streams occurs in 

 97 percent of the spawners, suggesting that each 

 stream has its own race of trout . Migrational 

 patterns in the lake after spawning are quite con- 

 stant from year to year, races from certain 

 streams often moving great distances. Mixing 

 of races takes place in the lake, and each part 

 of the laKe appears to contain about the same 

 mixture each year. 



Times of migration into five streams 

 were examined for a five-year period, and 

 five different patterns were perceived. The 

 patterns were very constant from year to year, 

 some streams supporting early runs, some late 

 runs, some having bimodal distributions, and 



82 



