FISHERY BULLETIN: VOL. 70, NO. 2 



had a small net upstream flow at the surface. 

 Apparently, the winds reversed the tidal flow 

 to some extent and thus the larval distribution 

 as well. 



Although I cannot account with certainty for 

 the system of larval movements in the sampling 

 area, the most reasonable explanation is that the 

 larvae responded to changes in the character of 

 the tidal flow. Larvae are transported up the 

 estuary at about 1.8 knots per semidiurnal tidal 

 cycle (Graham and Davis, 1971) which ap- 

 proaches the average length of tidal excursions 

 recorded at 15 m, below the level of no-net-mo- 

 tion. This level of no-net-motion is obliterated 

 in the upper estuary where the maximum depth 

 is 10 m and the channel widens and extensive 

 mud flats are present. The transition zone be- 

 tween the lower and upper estuaries is near the 

 town of Wiscasset (Figure 2) and coincides with 

 the maximum shoreward penetration of an 

 abundance of larvae. Stickney (1959) states 

 that the upper estuary compared to the lower 

 has a greater exchange ratio, lower and more 

 variable salinity and a wider range in temper- 

 ature. A reversing tidal falls 3.5 km above the 

 transition zone completely mixes the water. The 

 most likely agency to initiate the ascending re- 

 sponse (Figure 3, No. 2) by the larvae would 

 be the change in the character of the tidal flows 

 caused by shoaling of the bottom rather than 

 agencies peculiar to the estuary. Herring are 

 retained throughout their larval life over the 

 banks and ledges of the open waters of the Gulf 

 of Maine where the water is shallow and the tidal 

 flows are well developed but conditions are not 

 estuarine. After a brief transport downstream 

 near the surface the larvae would resume their 

 usual transport by descending towards the 

 bottom. 



The results suggested that ithe arrangement 

 of buoyed and anchored nets in the estuarine 

 channel was appropriate to sampling larval her- 

 ring transported within the tidal currents. Lar- 

 vae were sampled efficiently on both the flood 

 and ebb tides, since no important main eflfect 

 was obtained for tidal phase during the exper- 

 iments. To obtain such an eff"ect would have 

 required a larval concentration to pass beyond 

 the positions of the nets ; then the larvae would 



either have to miss the nets on the return tidal 

 current or drift to some area in the channel 

 where they were not subject to the currents. 

 Because the larvae were transported by the tidal 

 currents, tide was always one of the interacting 

 factors and an interaction between depth and 

 location was not obtained. 



One explanation for the appropriateness of 

 the sampling design was that concentrations of 

 the larvae were of sufficient length to overlap 

 a pair of nets when transported into a given lo- 

 cation. When this occurred, the diflference in 

 catch rates between the two lines of nets within 

 the location would be small, and the diflference 

 between the two lines of nets in the other lo- 

 cation where the concentration did not occur 

 would also be small. These diff"erences within 

 the locations provided a measure of experimental 

 error in the sampling design and if small they 

 would yield a statistically significant i^-ratio 

 when compared to diff"erences in catch rates be- 

 tween location and depth and the interaction of 

 these two factors with tidal phase (e.g., F — 

 mean square between locations /mean square 

 within locations) . Statistical significance would 

 not occur when a concentration overlapped only 

 one line of nets in a given location; the diff"er- 

 ences in catch rates between the two lines of 

 nets would be relatively large, as indicated for 

 the exterior and interior nets of the seaward lo- 

 cation in Table 3. 



"Although net tidal flows in the channel aflfect- 

 ed the transport of the larvae, temporal varia- 

 tions in tidal currents made it difficult to deter- 

 mine the exact nature of that transport. The 

 catch rates and corresponding tidal excursions 

 obtained from the nets were accumulative and 



Table 3. — Larval catches per 100 m^ of water strained 

 from an experiment during March 28, 1968. When 

 treated as in Table 1 the distributions of the catch rates 

 were not statistically significant. 



304 



