Armstrong et al.: Food habits of Leptocottus armatus 



459 



calculation of a modified form (Stevens et al., 1982) 

 of the Index of Relative Importance (IRI) (Pinkas et 

 al., 1971; Hyslop, 1980) based on estimated food 

 weight rather than food volume. For a particular prey 

 category, an IRI value was calculated as 



IRI = (JVC + GOFO, 



where NC (numerical composition) is the number of 

 a particular prey item divided by the total number 

 of all prey items in that sample multiplied by 100, 

 GC (gravimetric composition) is the combined weight 

 of a particular prey item divided by the total weight 

 of all stomach contents in the sample multiplied by 

 100, and FO (frequency of occurrence) is the number 

 of stomachs from a sample containing a given prey 

 item divided by the total number offish sampled mul- 

 tiplied by 100. IRI values from all prey categories 

 were summed to derive a grand total IRI value. The 

 relative importance of each prey category was then 

 expressed as a percentage of this total IRI (hereaf- 

 ter referred to as %IRI). 



Data analysis 



Sculpin stomach content data were analyzed from 

 four perspectives. First, an overall summer diet for 

 staghorn sculpin was derived by combining all fish 

 from both sites (E and S), inter- and subtidal, across 

 months to determine dominant prey taxa. Second, a 

 temporal comparison of sculpin diets from late spring 

 through summer was made by examination of sculpin 

 stomachs grouped by sampling trip. Third, the ef- 

 fect of sculpin size (and thus age) on diet was inves- 

 tigated for two size groupings derived from length- 

 frequency histograms of sculpin caught during the 

 six trips. Size ranges and age equivalents were set 

 from the literature (Anaheim Bay, CA, [Tasto, 1975]; 

 Yaquina Bay, OR [Bayer, 1985]) to group fish by age 

 class as either 0+ (n=210) or >1+ (n=256). IRI analy- 

 ses were run separately on 0+ and >1+ sculpin to 

 determine whether there were changes in diet com- 

 position based on predator size. Fourth, prey compo- 

 sition of sculpins feeding at different epibenthic habi- 

 tats was compared by examining diets of fish col- 

 lected from the intertidal eelgrass site (n=55) and diet 

 of sculpin caught at the intertidal shell site (rc=44). 



To compare dietary overlap of sculpin of the two 

 size/age categories or of sculpin collected at the two 

 different habitats (eelgrass vs. shell), Schoener's 

 (1970) index of overlap (S ;o ) was calculated as 



yi 



S IO = 1-0.5 



\J= 



where P • = the proportion of prey i (%IRI) in the diet 

 of sculpin of size x (or from habitat x); 

 = the proportion of prey i (%IRI) in the diet 

 of sculpin of size y (or from habitat y), 

 and j = the number of prey types. An overlap 

 value of S l0 > 0.6 (Schoener, 1970) was 

 considered significant in this study. 



Results 



Overall summer diet 



A total of 482 staghorn sculpin stomachs were col- 

 lected from April to August 1989, of which 458 had 

 some stomach contents. Sculpins averaged 118 mm 

 TL (range: 61-215 mm TL) during the five-month 

 sampling period (Table 1). Mouth gape width ranged 

 from 13 to 17% of standard length (gape width = 0. 1818 

 SL-2.6487, r2=0.909; Fig. 2). Average fish length and 

 weight varied by sampling month and were gener- 

 ally greater in April and May when samples were 

 composed primarily of 1+ fish, decreased in June as 

 0+ sculpin moved into the sampling sites, and in- 

 creased in July and August as individuals grew 

 through summer (Fig. 3; Table 1). Size-frequency 

 data indicated the presence of two modes equivalent 

 to 0+ and 1+ age classes. Young of the year ranged 

 from 60 to 94 mm TL in April and had grown to 75- 

 119 mm by August (Fig. 3). For purposes of analyses 

 of diet based on size-age composition of sculpin, 118 

 mm TL was used as the boundary between 0+ and 

 1+ groups. These size-age categories were corrobo- 

 rated by otolith annuli determinations of a similar 

 size range of staghorn sculpins from Vancouver Island, 

 British Columbia, estuaries by Mace (1983, p. 369). 



The mean number of individual prey items per 

 sculpin was 3.4, and gut fullness was generally high; 

 75% of all stomachs examined were rated 50% full 

 or greater whereas fewer than 5% were empty (Fig. 

 4). Sculpin stomach contents included 23 prey taxa 

 (Armstrong, 1991) which were grouped into 14 prey 

 categories for IRI analyses (Table 2). Principal cat- 

 egories of sculpins' overall summer diet (% IRI) were 

 the polychaete Nereis (Neanthes) brandti (30% IRI), 

 the ghost shrimp Neotrypaea californiensis (27% IRI), 

 sand shrimps Crangon spp. (13% IRI, including C 

 franciscorum, C. nigracauda, and C. stylirostris), 

 juvenile Cancer magister (9% IRI), and the mud 

 shrimp Upogebia pugettensis (6% IRI). 



Temporal changes in diet 



Stomach contents were analyzed on a per trip basis 

 to note change in sculpin diet over the course of the 



