Laserfiche WebLink
sampling design may not have been suitable to test for flow effects on <br />channel catfish reproduction. <br />Larval and post-larval mosquitofish were negatively correlated with peak <br />flow in all reaches, though not significantly so. Their numbers were <br />highest in 1989, the year of lowest flow. Correlations for catch rates in <br />fall (juveniles and adults) were relatively high and might become signifi- <br />cant with a greater sample size. Meffe (1984) documented the susceptibili- <br />ty of mosquitofish to mortality from flash-flooding in desert streams. <br />Effects from flushing flows during spring runoff in the upper Colorado <br />River may be less acute due to the gradual rather than sudden increase in <br />flow, and perhaps due to a greater availability of off-channel, refugia <br />habitats. <br />Numbers of young centrarchids were not correlated with peak flows. Corre- <br />lations for both larval and YOY green sunfish and YOY largemouth bass, <br />though negative, were generally not significant, regardless of the reach. <br />This, coupled with the apparent rarity of adults in riverine habitats <br />(Nesler 1990), indicates that reproduction occurs in protected off-channel <br />habitats, removed from but connected to the main channel. No larvae of <br />largemouth bass were ever detected in river backwaters during our `four <br />years of study. Green sunfish and largemouth bass inhabit ponds throughout <br />the Grand Valley (Osmundson 1987), and green sunfish are abundant in <br />irrigation return canals (USFWS unpublished data): Thus, chronic immigra- <br />tion of young to the river from these ponds and canals probably accounts <br />for their continued presence in backwaters rather than from significant <br />riverine reproduction (Osmundson 1987, Nesler 1990). The abundance of <br />young would therefore not be strongly influenced by high spring flow <br />25