did not include measuring the body length of 

 each fish, an alternative method had to be 

 employed. Several methods for estimating 

 body length from the size of stomach were 

 tested. Body length estimates based upon 

 linear measurements of the stomach proper 

 were lejected because rather large errors 

 resulted from their use. Gravimetric meas- 

 urements were found to be fairly accurate, 

 because even though the stomach dimensions 

 varied according to the quantity of food with- 

 in It, the weight of the empty stomach re- 

 mained constant . The relationship between 

 body length and stomach weight was deter- 

 mined from measurements on 325 haddock 

 (table 3). A regression of this relationship 

 is plotted in figure 2 and the standard devia- 

 tion from regression (or standard error of 

 the estimate) is indicated by the dashed lines. 

 The regression formula is log Y = - 0.0498 

 / 0.3557 log X and the correlation coeffic- 

 ient is 0.972. It was possible to estimate 

 the body length of a haddock by simply weigh- 

 ing the empty stomach and referring to the 

 graph . Since only approximate body lengths 

 were required for this work, the method 

 worked out very satisfactorily . 



In the laboratory the stomachs were 

 opened and all contents removed for exam- 

 ination. The food mass in each stomach was 

 measured volumetrically by water displace- 

 ment. Food items were sorted to the lowest 

 category to which they could be identified. 

 They were then counted and the volume of 

 each item was estimated. The volume of 

 large items was measured in order to obvi- 

 ate large errors due to estimation. When 

 numerous amphipods and small annelids 

 were encountered, estimated totals were 

 arrived at from counting a small sample . 



RESULTS 



Georges Bank haddock were found to be 

 exceedingly omnivorous in habit. Three 

 species of fish and nearly all major groups 



of marine invertebrate animals were repre- 

 sented in their diet A large majority of 

 the organisms in the haddock's dietary were 

 sedentary or slow -moving benthic animals. 

 Crustaceans were the primary food, Moi 

 lusks, echinodeims, and annelids comprised 

 a substantial, but secondary, share of the 

 diet . Fish were a minor component in the 

 haddock's diet The major categories of 

 the stomach content components for all 

 specimens examined in this study are listed 

 in table 4 and illustrated in figure 3 . 



Items of subordinate status in the had- 

 dock's dietary were grouped together under 

 the heading Miscellaneous. These items 

 consisted of 5 food categories and 2 non-food 

 categories . The food categories were: 

 Brachiopoda, Coelenterata, Nemertea, 

 Turbellaria, Urochorda, and unidentified 

 animal flesh. Non-food categories were: 

 parasitic nematodes and sand and stone . 

 Individually these miscellaneous groups were 

 not especially abundant in the diet, but to- 

 gether they constituted a significant share . 



Mucus was designated a major classifica- 

 tion heading because of the large quantities 

 (9 . 4 percent) encountered in the stomach 

 contents. It was usually yellow in color and 

 its consistency varied from cream- to jelly- 

 like . Inasmuch as fish are known to possess 

 mucus -producing glands in the mouth and 

 pharynx, to provide a food lubricant, it 

 seems likely that most of the mucus recovered 

 from haddock stomachs originated in their 

 mouth and pharyngeal region and was swal- 

 lowed with their food. 



The unidentified material encountered in 

 the stomach contents was composed of approx- 

 imately equal parts of relatively undigested 

 fragments of invertebrate organisms and 

 rather well macerated matter of undeter- 

 mined origin . 



One outstanding feature of the food 



