eggs in the stomach as a function of elapsed time 



(Figure 1). 



Incidence of Cannibalism 



No relationship existed between the volume 

 of the stomach contents and the time of collec- 

 tion, indicating that anchovy fed throughout the 

 night and day (Table 1). In the night samples, the 

 mean stomach volume was 23'7c of that of a 

 full stomach and was 12'7c in the 3 day samples. 

 Nearly all stomachs contained greenish to brown- 

 ish material, presumably phytoplankton remains, 

 and somewhat less frequently copepods and 

 euphausiids were mixed with this material. 



Larval fishes occurred in only 7 of the 368 

 stomachs (2%) and only 1 stomach contained 

 larvae that could be identified as northern ancho- 

 vy. This stomach contained 21 relatively large 

 northern anchovy larvae; the only measurable 

 specimen was 17 mm SL. Northern anchovy eggs 

 occurred more frequently than larvae. Forty-two 



1000 



100 



X 



o 

 <t 



o 

 t- 

 cn 

 \ 

 c/) 



<S> 



UJ 



u. 



o 



cr 



UJ 

 DD 



< 



10 I- 



Ol 



Ln X = 6.432 -OTGlT 

 r2 = 934 



4 6 8 



ELAPSED TIME (h) 



10 



percent of the northern anchovy stomachs sampled 

 at night contained northern anchovy eggs and 

 88% of those sampled in the day contained eggs. 

 Other fish eggs were rare, occurring in 4.2±2.7 

 (±2 SE) of the stomachs. The mean number of 

 anchovy eggs per stomach, including zeros, was 

 4.6±3.3 (±2 SE) for night samples, 5.6±3.5 for 

 day samples, and the mean for night and day was 

 5.1 eggs/stomach. 



The distribution of the number of eggs con- 

 sumed per fish was highly skewed; about 909^ of 

 the eggs occurred in only 38% of the stomach 

 samples containing eggs (19% of all stomachs). 

 The maximum number of eggs in a single stomach 

 was 730, which was about 32% of all the eggs 

 found in stomachs. The patchiness of northern 

 anchovy eggs in the sea may be responsible for this 

 skewed distribution. The distribution was not 

 greatly different from that of northern anchovy 

 eggs taken in plankton samples. For example, 

 Smith found that about 90% of northern anchovy 

 eggs occurred in only 20% of the positive plankton 

 net hauls (iV = 453). 



The mean number of eggs per fish in a sample 

 (night trawl samples) increased with egg density 

 in the sea (Figure 2). Although variability was 

 high (r^ = 0.47), the 99% confidence interval 

 about the regression coefficient was 1.107-2.095 

 and did not include a coefficient value of 1. This 

 indicates that the relation was exponential and 

 that the mean number of eggs in the stomach was 

 not increasing in direct proportion to the mean egg 

 abundance. Patchiness of eggs combined with 

 selectivity in filtering could explain the exponen- 

 tial nature of the relationship. Certainly, oblique 

 net tows provided only a relative measure of egg 

 density encountered by filtering northern ancho- 

 vy. Tows were begun at 70 m (below the maximum 

 depth of northern anchovy larvae (Hunter and 

 Sanchez 1977), or in shallow water near the bot- 

 tom to insure that all eggs in the water column 

 were sampled. Stomach contents, on the other 

 hand, would be expected to be more closely related 

 to the size of patches and density of eggs within 

 egg patches and not to the integrated egg density 

 for the water column. 



Measurements of the density of eggs behind and 

 in front of a single northern anchovy school were 

 probably a more realistic measurement of the 



Figure l. — Rate of gastric evacuation of northern anchovy fed 

 northern anchovy eggs. Points are log,, mean number of eggs 

 per stomach for 10-15 fish sampled at 2-h intervals after feeding. 



^P. E. Smith, Southwest Fisheries Center, National Marine 

 Fisheries Service, NOAA, PO. Box 271, La Jolla, CA 92038, 



pers. commun. December 1979. 



813 



