FEEDING BY LARVAL COD IN DIFFERENT WATER-MASSES ON WESTERN BANK, SCOTIAN SHELF

Gut contents were obtained from 1406 cod larvae from 94 stations in se ven water masses related to a gyre around Western Bank, Scotian Shelf, 22 November-16 December 1992. Initial samples were from: well-mixed w ater over the bank crest (CW); a surrounding convergent FRONT; relativ ely cold, fresh water (CFW) largely east of CW; warmer, salty water (W SW) west of CW anal FRONT. After a storm on 3-6 December, samples were from CW and CFW displaced south-east on the bank and, after further w inds 11-12 December, from CW displaced north-west off the bank. Zoopla nkton biomass (300-333 mu m mesh, mostly Calanus copepodids) did not d iffer among water masses, but larval concentrations were significantly higher in FRONT than elsewhere. The small-copepod diets of larvae var ied among water masses, partly attributable to larval growth during th e sampling period. Numbers of prey in guts, and indices of fullness an d digestion, varied among water masses. More reliably, after ANCOVAs s ignificant independent variables were: overwhelmingly time of day (max imum prey numbers and fullness at similar to 19:00) and larval size; w ater mass; weaker interactions of the above among themselves and with sample depth and date; a very weak negative turbulence-index effect on gut prey numbers in depth-stratified samples. After ANCOVAs, larvae f rom prestorm CW had significantly higher prey numbers and fullness tha n did those from FRONT, WSW, and CFW. Larvae in CFW were significantly fuller when sampled closer to sites of former CW after the storm. Alt hough numbers of prey in larvae advected off the bank in CW decreased significantly, prey averaged larger, so that gut fullness did not decr ease. We conclude that lan ae were best fed in the 'centre' of the Wes tern Bank gyre, but nut greatly affected by subsequent displacement of f the bank.