Table 3. — Splittail reproductive success, outflows of the Sacramento River at Chipps Island, Calif., and the correlations between them (* = 

 P< 0.05,** = P< 0.01). Abundance index is based on the strength of the year class, by percentage of sample (1 = weak, 3 = strong). The index 

 is derived from the collection data of this study, from California Department of Fish and Game unpublished records, and Caywood (1974). 



relatively constant, representing 30-50 f / ( of the 

 catch. The 1980 year class did not begin to dominate 

 monthly samples until the disappearance of the 1978 

 year class in November 1980. The 1980 year class 

 remained dominant from that date to the end of the 

 study. The importance of the 1978 year class was in- 

 dicated by the significant negative correlation 

 (P < 0.01) between catch per minute and number of 

 months that had elapsed since the beginning of our 

 study. 



Data for 1969 through 1974 from Caywood (1974) 

 and for 1972 through 1978 from the California 

 Department of Fish and Game (CDFG) (D. Kohlhorst 

 unpubl. data) also show strong and weak year classes 

 in splittail. These data sets are not strictly compa- 

 rable with ours because different methods were used 

 to capture the fish and because Caywood's samples 

 were not from the Suisun Marsh. However, the con- 

 sistency of the patterns in the data from two sources 

 on the same years (1972-74, 1977-78) allowed us to 

 rank year class strength from weak to strong, on a 1 to 

 3 scale (Table 3). Although the use of such a scale is 

 likely to bias an analysis against finding significant 

 correlations (Thorndike 1978), strong positive cor- 

 relations (P < 0.01) were nevertheless found be- 

 tween year class strength and outflows in February, 

 March, April, and May, as well as with total outflows. 

 Splittail may spawn on flooded vegetation in March 

 and April. Thus, wet years, such as 1974 and 1978, 

 show extremely large year classes that remained 

 abundant for several successive years. During 

 drought years, such as 1976 and 1977, nearly com- 

 plete year class failures occurred. CDFG collected 

 only oneyoung-of-the-yearin 1976 and none in 1977. 

 Our samples in January-March of 1978 yielded only 

 large adults, none of which was from the 1976 or 1977 

 year classes. 



Feeding 



Splittail prey upon a variety of organisms but de- 

 tritus dominates gut contents (Table 4). A small part 

 of the gut content was unidentifiable animal matter. 

 Neomysis mercedis was the dominant prey item, and 

 crustaceans accounted for almost 85% by volume of 

 the animal portion of the diet. Neomysis mercedis 

 dominated the diet seasonally and through a diel cy- 

 cle. Annelids, molluscs, insects, and fish completed 

 the prey list; all were relatively unimportant com- 

 ponents. These results are in contrast to those of 

 Caywood (1974) who analyzed three collections of 

 splittail from the upper Delta. He found the domi- 

 nant animal matter in the stomachs to be clams (Cor- 

 bicula manilensis), amphipods (Corophium spp.), 

 copepods, and dipteran larvae and pupae. In a sam- 



TaBLE 4.— Prey items taken by 635 splittail collected from the 

 Suisun Marsh, Calif., 1979-80. 



651 



