WEINSTEIN: SHALLOW MARSH HABITATS AS PRIMARY NURSERIES 



Table 9. — Partial correlations i given tempera turei of species abundance with salinity in 

 with < 10 individuals of a given species in any month were omitted from the calculations. 



Aug 



the Cape Fear River estuary, N.C. Collections 

 N.C. = >10 individuals collected. **P<.01. 



Species 



Feb, 



Mar 



Apr. 



May 



June 



July 



Sept. 



Oct 



Dec Pooled 



'v^ values are based on lasts of equality among correlations ot the 1 1 monthly collections; none were slgnl^ca^l, therefore, all individual correlation values were 

 pooled 



a lack of correlation of distribution with salinity. 

 This effect occurred in other months for other 

 species and is consistent with the patterns of dis- 

 tribution of estuarine organisms and their ability 

 to withstand wide ranges of salinity, at least over 

 the short term. What is important, however, is 

 that during the course of residency in the marshes, 

 the presence of several species was significantly 

 correlated with salinity. Gunter (1961) draws a 

 similar conclusion by stating that correlation is 

 not necessarily with a given salinity but rather 

 with the gradient as a whole. 



Of the species tested (Table 9), striped mullet; 

 blackcheek tonguefish; mummichogs; silver 

 perch, Bairdiella chrysura; pinfish, Lagodon ' 

 rhomboides; bay anchovy; and striped anchovy, A. 

 hepsetus, were distributed independently of salin- 

 ity. In several instances, a considerable portion of 

 the variance associated with abundance data was 

 explained by salinity alone; this was true for the 

 Atlantic and tidewater silversides and for the 

 striped killifish and Atlantic croaker. Although 

 other/- values were significant (P<0.01) it is obvi- 

 ous that factors other than salinity ~were con- 

 tributing to patterns of distribution. 



Substrate characteristics have been shown to 

 influence invertebrate populations and the struc- 

 ture of fish communities (Mills 1975). In this 

 study, the distribution of several species also ap- 

 peared influenced by properties of the sediment 

 (Table 10). The abundance of Menidia menidia and 

 M. heryllina was negatively correlated with per- 

 cent organics. and the former species displayed a 

 similar relationship with sorting coefficient. This 

 is not at all sm-prising in light of their mode of 

 feeding and the presence of currents which proba- 

 bly act to carry food items through the area. On the 



Table 10.— Partial correlations (given salinity) of species 

 abundances with several sediment parameters in the Cape Fear 

 River estuary. N.C. Collections with < 10 individuals of a given 

 species in any month were omitted from the calculations. Values 

 in parentheses do not include Bamards Creek. ** =P<.01. 



Other hand, young striped mullet which relies 

 heavily on detritus in its diet (Odum 1968) was 

 expected to display a positive association with per- 

 cent organics, but this did not occur with respect to 

 all creeks concerned. If Barnards Creek was omit- 

 ted from the calculations, however, the relation- 

 ship became highly significant. The extremely 

 high organic content of Barnards Creek sediments 

 probably is indicative of highly reducing condi- 

 tions, and may, in fact, have contributed to the low 

 total fish catch in this creek. 



Two other species exhibited a positive relation- 

 ship with percent organics, the blackcheek 

 tonguefish and the Atlantic croaker. The 

 blackcheek tonguefish commonly is associated 

 with muddy bottoms and high salinity (Gunter 

 1945; Kilby 1955) although salinity did not seem 

 to play a role in governing its distribution in the 

 Cape Fear region (Table 7). Darnell (1958) has 

 described the feeding preference of young croaker 

 for organic matter and since this species tends to 

 accumulate toward the headwaters of many sys- 



353 



