116 FISHERY BULLETIN OF THE FISH AND WILDLIFE SERVICE 



were made 1 week later than that indicated by the marginal date, the Bonneville 

 count and the estimated catch above Bonneville were made 2 weeks later than that 

 indicated by the marginal date, and the Rock Island count 4 weeks later. For con- 

 venience we shall refer below to the assumed position of the fish during their upward 

 migration as in Zones 1 and 2 the first week, in Zones 3 to 5 the second week, at Bonne- 

 ville and in Zone 6 the third week, and at Rock Island the fifth week of then- fresh- 

 water migration. The same system was followed in preparing the similar tables for 

 the other species. 



Thus, reading across any one line, say the line for May 7 in table 12, the first col- 

 umn gives the estimated catch made in Zones 1 and 2 during the week ending May 7, 

 the second column the estimated catch made in Zones 3 to 5 during the week ending 

 May 14, the fourth column the count at Bonneville during the week ending May 21, 

 the fifth column the estimated catch above Bonneville during the week ending May 21, 

 and the seventh column the count at Rock Island during the week ending June 4. 

 Columns 3 and 6 are derived by summing across the rows in the appropriate columns 

 and therefore show totals for the run as a whole — all referred back to the week that 

 the fish were presumably in the extreme lower part of the river and, therefore, approxi- 

 mately to the time that they entered the river. 



Individual fish undoubtedly vary greatly in respect of their rate of travel up- 

 stream, but the obvious similarity in the trends of all the columns in this table is 

 evidence that, on the average, these assumptions are well founded. 



NATURE OF THE ANALYSIS OF RUNS 



From the tables of this structure it is possible, for those species that largely spawn 

 above the site of the Bonneville Dam, to estimate the number of fish of each species 

 that escaped the intensive fishery below Celilo Falls (the upper limit of commercial 

 fishing) in 1938 and were available for reproduction above Bonneville Dam. This is 

 readily done for any desired portion of the season by subtracting the catch above 

 Bonneville from the Bonneville count. Such an estimate of the escapement is subject 

 to error from several causes, of which the following may be mentioned: (1) Error in 

 the count of fish of the different species at Bonneville, (2) error in the catch figures 

 due to the fact that a considerable catch that does uot appear in the record is made by 

 Indians, and to some extent by Whites for home use, and (3) error in converting 

 pounds to number of fish. While these sources of error are present, it is believed that 

 their total effect is relatively small and will not affect the general conclusions that 

 may logically be drawn. Furthermore, in making these estimates no attempt has 

 been made to correct for the spawning that takes place in the tributaries below 

 Bonneville Dam. In the case of the silver and chum salmon such a large percentage 

 of the spawning takes place below Bonneville that a similar analysis has not been made. 

 Also, as mentioned above, there is a considerable part of the faU run of chinooks 

 that spawns below Bonneville so that our study of the fall run is probably Jess reliable 

 than that for the spring season. Since our estimate of the escapement is based 

 primarily upon the count at Bonneville (from which is subtracted only the estimate 

 of the number of fish in the recorded commercial catch above Bonneville) the spawning 

 in the tributaries between Bonneville and the upper end of the commercial fishing 

 district at Celilo Falls will not affect the results. If any considerable portion of the 

 run that is actually derived from the tributaries below Bonneville be ascribed to the 



