FISHERY BULLETIN: VOL. 70, NO. 2 



experiments were examined. The analysis of 

 variance from an experiment on November 8, 

 1967 (Table 1, IV) showed that the catch rates 

 of larvae differed significantly with location and 

 with the interaction of depth and tide. The un- 

 transformed catch rates showed that concentra- 

 tions of larvae were shallow during the ebb tide 

 and deep during the flood tide. Also, more lar- 

 vae occurred at the landward end of the channel 

 than at the seaward end. These data and those 

 from other experiments having significantly dif- 

 ferent distributions in catch were summarized 

 in Table 2. For example, the interaction on No- 

 vember 8 was illustrated by first summing all of 

 the cubic meters of water strained by the shallow 

 nets and then all of the larvae captured by them. 

 These totals were used to obtain an average 

 catch per 100 m^ of water strained; the process 

 was repeated for the deep nets. Beginning 

 anew, the average catch rates were determined 

 for the two locations. The numbers in paren- 

 theses in Table 2 locate the concentrations of 

 larvae within the channel diagrammed in Fig- 

 ure 3. Those concentrations of larvae captured 

 in the shallow nets had a net transport seaward 

 during an experiment and those in the deep nets 



offshore 



inshore  



® 



® 



® 



o 



® 



'///,, 



Bottom 



Figure 3. — Diagram of a system of larval movements 

 thought to retain larval herring in the Sheepscot estuary. 

 The numbered distributional components of the system 

 coincide with the numbered catches in Table 2. 



were transported landward. A transition in 

 such transport occurred for those larvae concen- 

 trated throughout the water column at one lo- 

 cation or the other. The arrows in Figure 3 

 infer a system of movements from the location 

 and transport of the larvae that appears analo- 

 gous to those of other larval fishes (Pearcy and 

 Richards, 1962) and planktonic organisms 

 (Bousfield, 1955) which are retained in estuar- 

 ies by tidal currents. The suggested movements 



Table 2. — Catch rates from 11 experiments in the autumn and spring. The distributions of rates for 

 single factors and for two factors (interaction X) are given. Numbers in parentheses are from Figure 

 3; unnumbered catch rates are discussed in the text. 



302 



