Fortier and Villeneuve Cannibalism and predation by Scomber scombrus larvae 



269 



1991. Cannibalism and predation on other fish larvae 

 were studied in relation to larval fish density and the 

 availability and consumption of alternative inverte- 

 brate prey. In particular, we searched for evidence that 

 cannibalism and piscivory by Atlantic mackerel larvae 

 were density dependent. Finally, we evaluated the po- 

 tential impact of predation by Atlantic mackerel lar- 

 vae on the early survival offish larvae. 



Methods 



In spring, the northern component of the Northwest 

 Atlantic stock of Atlantic mackerel migrates from the 

 New England Continental Shelf and the Scotian Shelf 

 towards the southern Gulf of St. Lawrence where the 

 bulk of spawning takes place from early June to mid- 

 August (Sette, 1943, 1950). Some spawning also takes 

 place along the coast of Nova Scotia (Sette, 1943; 

 Berrien et al., 1981). During a 27-month series of 

 monthly ichthyoplankton surveys on Emerald, West- 

 ern, and Sable Island banks (Fig. D.Atlantic mack- 

 erel larvae were captured offshore, primarily over 

 Sable Island Bank, in June, July, and August of 1991 

 and 1992. The larvae ranged from 2.7 to 13.7 mm 

 standard length (SLXunpreserved), corresponding to 

 an approximate age span of 4 to 16 days (D'Amours 

 et al., 1990). These lengths and ages indicate that 

 the shallow waters west of Sable Island constitute a 

 spawning ground for the species. The present study 

 focuses on the trophodynamics of Atlantic mackerel 

 larvae sampled in July 1991. 



• Western Bank 



Emerald Bank 



64W 

 J |_ 



62°W 



60°W 

 J 1_ 



Figure 1 



Location of the 32 rectangular midwater trawl (RMT) stations sampled 

 for Atlantic mackerel. Scomber scombrus, larvae and their prey from 

 18 to 23 July 1991 on Emerald, Western, and Sable Island banks (Scotian 

 Shelf). The vertical distribution of fish larvae was studied with the 

 multi-net sampler on 24 and 25 July at the station marked with a circled 

 black dot. Isobaths are in meters. 



Sampling 



Atlantic mackerel larvae and their prey were sampled 

 from 18 to 23 July 1991 over a grid of 32 stations 

 covering part of Emerald, Western, and Sable Island 

 banks ( Fig. 1 ). Three different plankton nets mounted 

 on a rectangular midwater trawl (RMT) were used 

 to collect young fish and zooplankton simultaneously. 

 A main net (8-m 2 sweeping section, 1,600-jUm mesh) 

 captured juvenile fish; an intermediate net (2-m 2 

 sweeping section, 333-/im mesh) collected fish lar- 

 vae and mesozooplankton; and a small net (0.0075- 

 m 2 mouth, 64-jum mesh) sampled the microzoo- 

 plankton prey of fish larvae. The RMT was towed 

 once at each station in a multiple oblique trajectory 

 from to 75 m (or 5 m from the bottom at stations 

 shallower than 75 m) for an average (± standard de- 

 viation [SD]) of 17 (±5) min at a speed of approxi- 

 mately 2 knots (1 m/s). A CTD (conductivity, tem- 

 perature, and depth) probe and flowmeters linked to 

 a computer allowed real-time monitoring of depth, 

 water column temperature and salinity, and the vol- 

 ume of water filtered by each net. 



On 24 July, a multi-net sampler (Eastern Marine 

 Services E-Z-Net ) mounted with 10 identical nets ( 1- 

 m~ mouth, 333-^im mesh) was used to determine the 

 vertical distribution offish larvae at the grid station 

 (52-m depth) where the greatest density offish lar- 

 vae was found with the RMT ( Fig. 1 ). Six casts were 

 made at intervals of approximately four hours from 

 2115 h on 24 July to 2000 h on 25 July 1991. For 

 each cast, 4 depth intervals of 10 m were sampled 

 twice in the following order: 40-30, 30- 

 20, 20-10, 10-0, 0-10, 10-20, 20-30, and 

 30—40 m. Nets were also opened on the 

 downward (0-40 m) and upward (40-0 m) 

 trajectories of the cast, providing two 

 depth-integrated samples. The multi-net 

 system was equipped with the same real- 

 time monitoring system as that used on 

 the RMT. 



Most fish larvae (709r) were sorted im- 

 mediately at sea and preserved in 95 ( "c 

 ethanol. The remainder of the plankton 

 sample, including undetected fish larvae, 

 was preserved in 4% buffered formalin. A 

 subset of 50 Atlantic mackerel larvae were 

 videotaped on the ship for determination 

 of morphometries before preservation. 

 Twenty-six of these (3.3 to 13.7 mm fresh 

 SL) were preserved in ethanol and 24 (3.3 

 to 13.3 mm fresh SL) in formalin. 



In the laboratory, all fish larvae remain- 

 ing in the preserved samples were sorted 

 and identified. All undamaged mackerel 



