794 



Fishery Bulletin 91(4), 1993 



there at a density greater than 10/100 m :) . The other 

 anomalous tributary was the C&D Canal, which is not 

 a true tributary, but a man-made, free-flowing canal 

 constructed in the early 1900's to link the Delaware 

 and Chesapeake estuaries for navigation. The C&D 

 Canal contains virtually no freshwater drainage, and 

 the species composition there was most similar to the 

 lower region of the mainstem. No clupeid larvae and 

 very few M. americana larvae were collected there, 

 but the C&D Canal did contain densities of M. saxatilis 

 eggs considerably higher (P<0.01) than the mainstem 

 or other tributaries. 



Potential bias in the recent survey could exist be- 

 cause sampling in the upper river region was prima- 

 rily accomplished in 1987, whereas the downstream 

 areas were sampled in 1988. To determine if this was 

 a problem, we compared ichthyoplankton density and 

 species composition in the area sampled in common 

 between the two years. We found that the difference in 

 total ichthyoplankton density between the two years 

 was less than 10% (Table 4) and not significantly dif- 

 ferent. In addition, none of the dominant taxa had 

 more than a two-fold difference in density between 

 years. Temporal patterns for the dominant taxa were 

 generally consistent across both years of the study (Fig. 

 4). Seasonal peaks in larval abundance for Perca 

 flavescens (yellow perch) preceded that of the Alosa 

 spp., followed by M. americana and M. saxatilis. A. 

 sapidissima and P. flavescens larval abundance peaked 

 about a week later in 1988 than in 1987. This differ- 

 ence may be attributable to inter-annual variability in 

 spawning peaks; however, since sampling in 1988 was 

 concentrated farther downriver, the difference in tem- 

 poral patterns between years may partly reflect the 

 time required for downstream transport of larvae from 

 upstream spawning grounds. 



Comparison with historical data 



When we compared our data to the historic data, we 

 found that average density of all eggs and larvae did 

 not differ significantly in the two lower regions and 

 that historical density was higher (P<0.01) in the most 

 upstream region (Fig. 3). The larger difference among 

 periods was in composition and dominance; the num- 

 ber of taxa caught in each region in the present study 

 was almost double that in the historical data (Fig. 2), 

 despite the greater number of sampling years encom- 

 passed by the previous studies. Whereas river herring 

 accounted for more than half of the catch in every 

 region during the 1970's, no taxon accounted for more 

 than 40% of total catch in any region in the present 

 study (Table 1 ). 



Density of nine taxa increased significantly in at 

 least one region since the previous decade (Table 1). 



Of these, the increase was most dramatic for A. 

 sapidissima. In the earlier studies, A. sapidissima eggs 

 were rarely found, and larvae were not encountered at 

 all. In contrast, the density of shad larvae in the up- 

 per river region exceeded 100/100 m ( during several 

 weeks of our study (Table 3). Five taxa decreased sig- 

 nificantly in number in at least one region, and cyp- 

 rinids (minnows) and Anguilla rostrata (elvers) declined 

 significantly in all three regions. Cyprinids were only 

 about one-tenth as abundant in our study. Although 

 we did not identify all of the cyprinids in our study, 

 the cyprinids collected consisted primarily of juvenile 

 Notropis spp., whereas most cyprinids in previous stud- 

 ies were identified as juvenile C. carpio. 



Discussion 



The dramatic decline of A. sapadissima productivity 

 in the Delaware River since the last century 

 (Chittenden, 1974) has been attributed to reductions 

 in the amount of usable spawning habitat (Chittenden, 

 1976). Shad have historically used all of the Delaware 

 River as spawning and nursery grounds, but spawn- 

 ing during most of this century was limited to areas 

 200-km upstream from the tidal freshwater areas we 

 studied. Size of the spawning ground was reduced for 

 two reasons (Chittenden, 1976). First, only the earli- 

 est arriving shad, which typically spawn farthest up- 



