per collection averaged over both years are (number of 

 collections in parentheses): 



Surf 



Tidal pool 

 Overall 



Fall 

 4.25(12) 

 2.70(10) 

 3.55(22) 



Winter 

 3.31(13) 

 1.00(7) 

 2.50(20) 



Spring 

 4.92(13) 

 2.40(10) 

 3.82(23) 



Summer Overall 



6.00(12) 4.60(50) 



2.12(8) 2.14(35) 



4.45(20) 3.59(85) 



The surf and tidal pool collections differ significantly 

 with respect to number of species per collection when 

 averaged over both years and all seasons (Table 7). The 

 results of the analysis of variance also indicate a signifi- 

 cant seasonal effect. Student-Newman-Keuls' test was 

 used to determine which of the overall seasonal means 

 were significantly different (Sokal and Rohlf 1969). The 

 results indicate that only winter and summer differ 

 significantly in number of species per collection; all other 

 comparisons were nonsignificant. 



The results of the analysis of variance for the number 

 of specimens per collection are presented in Table 8, in- 

 dicating that neither the year effect nor the main effect of 

 season is significant. On the other hand, both the main 



Table 8. — Analysis of variance for X = log (number of specimens of 

 fishes per collection + 1). (** Highly significant, P< 0.01; ns not sig- 

 nificant, P> 0.05.) 



Source 



SS 



df 



MS 



F 



Blocks (years) 



1.4906 



1 



1.4906 



3.565 ns 



Location 



3.5807 



1 



3.5807 



8.563 ** 



Season 



0.6739 



3 



0.2246 



0.537 ns 



Location by season 



6.4340 



3 



2.1447 



5.129 ** 



Error 



31.7783 



76 



0.4181 





Total 



43.9575 



84 







effect of location and the location-by-season interaction 

 are highly significant. The back -transformed location- 

 by-season means (means expressed in the original units 

 of measurement) for both years combined are (number of 

 collections in parentheses): 





Fall 



Winter 



Spring 



Summer 



Surf 



20.66(12) 



56.00(13) 



18.88(13) 



38.69(12) 



Tidal pool 



23.02(10) 



2.50(7) 



18.19(10) 



8.84(8) 



In view of the highly significant interaction of location 

 and season, an examination of this interaction was un- 

 dertaken to determine location differences within season; 

 results of t-tests for surf versus tidal pool are: fall 0.16, 

 P>0.05; winter 4.00, P<0.01; spring 0.06, P>0.05; 

 summer 2.05, P< 0.05. These single degree-of-freedom 

 comparisons are not independent, but are warranted 

 because the interaction is highly significant. They in- 

 dicate that there is a highly significant difference 

 between the average number of specimens per collection 

 in the surf and the average number of specimens per 

 collection in the tidal pool for those collections made in 

 winter, a significant difference for summer, and a non- 

 significant difference for both fall and spring. 



Table 9 gives the results of the analysis of variance for 

 total weight per collection which indicate a highly 



Table 9. — Analysis of variance for X = log (total weight of fishes per 

 collection + 1). ('Significant, P< 0.05; "highly significant, P<0.01; 

 ns not significant, P> 0.05.) 



Source 



SS 



df 



MS 



Blocks (years) 

 Location 

 Season 



Location by season 

 Error 

 Total 



4.4421 



12.5504 



1.2729 



5.8823 



40.7355 



64.8832 



1 

 1 

 3 

 3 



76 

 84 



4.4421 

 12.5504 

 0.4243 

 1.9608 

 0.5360 



8.288" 



23.416 " 



0.792 ns 



3.658* 



significant difference between years for the total weight 

 of fishes per collection. This is due in part to the capture 

 during the first year of 20 Mugil cephalus with a total 

 weight in excess of 1.5 kg and a single specimen of 

 Pogonias cromis which weighed more than 1.2 kg. A 

 variation in seining technique (see section on Materials 

 and Methods) may have contributed to the capture of 

 these large specimens because two short tows may have 

 been more effective in collecting large individuals than a 

 single long tow. The main effect of season is not signifi- 

 cant. Both the main effect of location and the location- 

 by-season interaction were found to be significant. The 

 back-transformed location-by-season means (in grams) 

 for both years combined are (number of collections in 

 parentheses): 





Fall 



Winter 



Spring 



Summer 



Surf 



75.87(12) 



184.60(13) 



101.24(13) 



146.66(12) 



Tidal pool 



59.56(10) 



5.84(7) 



22.37(10) 



11.34(8) 



Single degree-of-freedom comparisons of location 

 differences within season show that the surf does not 

 differ significantly from the tidal pool for those collec- 

 tions made in the fall, but surf and tidal pool collections 

 differ significantly for each of the other seasons with 

 respect to the average total weight of fishes per collection 

 (results of t-tests are: fall 0.33, P>0.05; winter 4.18, 

 P<0.01; spring 2.08, P<0.05; summer 3.23, P<0.01). 



Ranking of Species 



The more common species (represented by at least 100 

 specimens) were Arenaeus cribrarius, Anchoa mitchilli, 

 Fundulus majalis, Menidia menidia, Trachinotus 

 carolinus, Menticirrhus littoralis, and Mugil curema. 

 The six species of fishes accounted for 94% of the 

 specimens of fishes collected (4,767 of 5,095) and 72% of 

 the ichthyomass (12,406 of 17,318 g). One species, 

 Areneaus cribrarius, made up 82% of the swimming in- 

 vertebrate specimens (422 of 512) and 75% of the mass of 

 swimming invertebrates (3,218 of 4,300 g). These species 

 were ranked to determine their relative importance. Of 

 the seven most important species, one (F. majalis) was 

 not a significant part of the catch in the surf, and three 

 (A. cribrarius, A. mitchilli, and M. littoralis) were of lit- 

 tle consequence in the tidal pool. 



We developed a seasonal index to allow ordering of 

 species by seasonal occurrence. It is defined as: 



11 



