delays occurred below Little Goose Dam. Radio- 

 tagged salmon travelled mean distances of 26.5 

 km in 1976 and 42.0 km in 1977 before crossing 

 Little Goose Dam, despite its location only 6.4 km 

 above the Texas Rapids release site. Dropbacks 

 after arrival of fish in the Little Goose Dam spill 

 were common, averaging 1 or 2/fish. Delay times 

 and distances ranged up to 100 h and 40 km/ 

 dropback episode. Radio-tagged (1976-77) salmon 

 exhibited three movement patterns after release 

 at Texas Rapids until arrival at Little Goose Dam: 

 31% (12/39) moved to the dam within 4 to 12 h; 

 31% ( 12/39) remained within ±5 km of the release 

 point overnight and moved to the dam the next 

 day; and 38% (15/39) moved downstream as far as 

 32 km, and then upstream to the dam 1-6 days 

 later. 



Once an individual salmon began moving up- 

 stream, it traveled 2-5 km/h and, generally, did 

 not stop until reaching Little Goose Dam. After 

 entry into the dam spill, three behavior patterns 

 were observed: 34% (12/35) crossed the dam 2-5 

 days after release without dropping back; 40% 

 (14/35) crossed the dam after averaging 1.6 

 dropbacks/fish; and 26% (9/35) were not observed 

 or recorded crossing Little Goose Dam. However, 

 at least four salmon in the latter group were ob- 

 served passing Lower Granite Dam or were recov- 

 ered upstream by anglers or at hatcheries, and 

 properly belong in the second group. 



At Little Goose Dam, especially with continuous 

 spring 1976 spilling, salmon appeared "confused." 

 Movements to and from the spill area were com- 

 mon. Substantial milling, previously reported at 

 Lower Granite Dam (Liscom and Monan see foot- 

 note 8), occurred in a large back eddy on the north 

 side of Little Goose Dam in 1976 and in front of 

 turbine outflows in 1977. Salmon may use 

 rheotactic, olfactory and/or acoustical cues to 

 navigate upstream (Harden Jones 1968). These 

 cues may be distorted by continuous and heavy 

 spilling near dams. Milling may be an attempt to 

 relocate orientation cues (Hasler et al. 1978). In 

 both study years movements into the fish ladder 

 were frequent, but salmon frequently fell back and 

 returned to the spilling basin, especially upon 

 nearing the trapping facility in midladder. 



Periodically, fish trapping operations were 

 halted to allow large groups of salmon, which were 

 suspected of accumulating in the spill to pass Lit- 

 tle Goose Dam (Slatick"). From 26 April to 30 

 May 1977, the trap was inoperative 17% of the 

 time (6 of 35 days). However, 30% of all salmon 



counted (7,382 of 24,238) at the fish viewing win- 

 dow by NMFS personnel and 59% (16/27) of our 

 tagged salmon crossed Little Goose Dam during 

 nontrapping periods. Daily viewing window 

 counts of salmon passage averaged 562±497 fish 

 during trapping and 1,230± 1,040 fish during non- 

 trapping periods. A ^-test showed these differences 

 were significant (P<0.05) and indicated trapping 

 operations at Little Goose Dam impeded chinook 

 salmon passage. 



Other aspects of dam operations affected salmon 

 passage at Little Goose Dam. One morning in May 

 1976, spillways were closed for several hours. The 

 five radio-tagged salmon present left the spill and 

 moved downstream as far as 16 km. During the 24 

 h after spilling was restored, fish returned to the 

 dam. In the absence of spilling in spring 1977, 

 radio-tagged salmon moved into turbulent areas 

 created by water leaving power generating tur- 

 bines at Little Goose and Lower Granite Dams. 

 When turbine operations were altered, in response 

 to power generation demands, salmon generally 

 exited the area and returned after water-flow con- 

 ditions stabilized. Salmon passage through the 

 fish ladder was also impeded by turbine operations 

 (Slatick see footnote 11). Finally, the fish ladder at 

 Little Goose Dam uses pumped river water rather 

 than a gravity flow. Of all salmonid species 

 studied, chinook salmon may be most sensitive to 

 and least likely to swim through pumped water 

 (Slatick see footnote 11). 



Gray and Haynes (1977) showed that mean 

 swimming depths of adult chinook salmon in the 

 Snake River were significantly greater (P<0.05) 

 in spring 1976 than in spring 1977. However, in 

 both years, mean swimming depths in the Little 

 Goose Dam spill were significantly greater 

 (P<0.05) than in all other sections of the study 

 area (Haynes 1978). In contrast, swimming depths 

 in the Lower Granite Dam spill were similar to 

 those in the open river between dams. As delays 

 increased in the Little Goose Dam spill, salmon 

 swimming depths increased. Since fish ladder en- 

 trances are near the surface, this decreased oppor- 

 tunities for passage. 



Although some delays may have resulted from 

 tagging, we believe other factors caused the exten- 

 sive delays observed at Little Goose Dam. First, 

 our radio-tagged and control salmon had similar 

 passage times at Little Goose and Lower Granite 



•'E. Slatick, National Marine Fisheries Service, Little Goose 

 Dam, Starbuck, Wash., pers. commun. 1977. 



188 



