Sainte Mane and Chabot Natural diet of Homarus americanus off the Magdalen Islands 



107 



1988; Barshaw, 1989; Lavalli and Barshaw, 1989; Lawton 

 and Lavalli, 1995 ). Wahle ( 1992 ) offered a conceptual mod- 

 el suggesting that lobsters shift from a cryptic to a wide- 

 roaming behavior as predation risk becomes offset by 

 the need for a high-energ\' diet that cannot be satisfied 

 through shelter-restricted feeding. 



Our study was conducted at the Magdalen Islands, east- 

 ern Canada, to resolve the natural diet of SRJ lobsters and 

 to compare it with that of larger lobsters by using stomach 

 content analysis. We found a gradual ontogenetic shift in 

 lobster diet over the size range of 4 to 112 mm CL. SRJs 

 were carnivorous and probably derived their meals main- 

 ly through predation and scavenging. We also determined 

 the predator-prey size relationship for one of the lobster's 

 preferred and most important prey, i.e. Atlantic rock crab. 

 Cancer irroratus (Reddin, 1973; Evans and Mann, 1977; 

 Carter and Steele, 1982a I. 



Materials and methods 



The study site was a narrow 2-km rocky section (47°14.5'N. 

 6r50..5' to erSl.S'W) of the south shore of Baie de Plai- 

 sance, Magdalen Islands, eastern Canada. This site corre- 

 sponds to the Butte-a-la-Croix location that Hudon (1987) 

 determined to be a settlement ground for lobster Divers 

 collected lobsters by hand or by suction-sampling at depths 

 of 1 to 7 m. Lobsters were processed live usually within 

 minutes and at most two hours after collection. The sex 

 of collected specimens was determined and their CL was 

 measured to the nearest 0.1 mm with a vernier caliper. 

 Lobsters that were not berried and that were judged to be 

 intermolt, based on criteria of shell hardness, coloration, 

 and fouling in Aiken (1980), were dissected to remove the 

 stomach which was preserved in buffered formalin diluted 

 to 4'7(- in seawater Stomachs with calcified gastroliths were 

 subsequently disregarded, thereby effectively eliminating 

 from the present study all premolt lobsters from stage 

 D'-5 (=Dq) on (Aiken, 1980). The resulting sample con- 

 sisted of 471 stomachs from lobsters of 7-112 mm cepha- 

 lothorax length (CL) collected from 24 July to 31 October 

 1996, and of 35 stomachs from lobsters of 4—12 mm CL col- 

 lected between 4 August and 13 September 1997. The 1997 

 lobsters were added to improve coverage of stomach con- 

 tents of the early juveniles because very little settlement 

 occurred in 1996 iSainte-Marie et al., 2001). There was no 

 commercial fishery during the sampling periods; therefore 

 items in lobster stomachs were not discards or bait. 



In the laboratory, stomachs were opened and their con- 

 tent was emptied into dishes for examination under a Wild 

 M8 compound microscope (10-50x). Identity of food items 

 was determined to the lowest taxonomic level possible, 

 based on comparisons with illustrations in literature and 

 samples of benthic and pelagic fauna from our study site. 

 Particular care was taken when examining the stomach 

 contents of lobsters <12 mm CL; for these stomach con- 

 tents we often resorted to higher magnification (>100x) 

 with a Leitz Dialux 20 microscope. 



The contribution of each food item, exclusive of miner- 

 als and nylon debris, to the volume of stomach contents 



of each lobster was visually scored from to 10, by 10% 

 increments (0=0% of volume, 1=1-10%, 2=11-20%, etc.). 

 The total for all food items could exceed 10, for example, 

 if more than two minor food items each were scored 1 in 

 addition to one predominant food item that was scored 8. 

 In such cases, the corrected contribution of each food item 

 was obtained by dividing its score by the sum of scores for 

 all organic food items in a given stomach. Corrected volu- 

 metric contribution of each food item was expressed as a 

 proportion of stomach content volume. 



To obtain information on the size spectrum of rock crab 

 consumed by lobsters, we established predictive (least 

 squares) linear regressions (Sokal and Rohlf 1995) be- 

 tween 30 measurements of distinctive hard body parts 

 and cephalothorax width (CW) of 26 crabs ranging from 

 7 to 62 mm CW (following the approach in Lovrich and 

 Sainte-Marie, 1997). All the predictive regressions were 

 highly significant (/■-=0.970-0.999. P<0.001). When rock 

 crab remains were encountered in lobster stomachs, dis- 

 tinctive hard body parts were measured with an eyepiece 

 micrometer to estimate crab CW from predictive regres- 

 sions. When more than one body part could be measured, 

 the crab's CW was determined as the mean of the various 

 estimates unless it was obvious that multiple crabs had 

 been ingested. Such was considered to be the case when 

 more than two similar fragments of a paired structure 

 (e.g. eyes or claws) were found in one lobster stomach or 

 when there was considerable divergence among crab CW 

 estimates based on different body parts. The functional re- 

 lationship between the CW of rock crab prey and the CL of 

 lobster predators was established with a model II regi'es- 

 sion (Laws and Archie, 1981; Sokal and Rohlf 1995). 



The stomach contents, once identified and scored for 

 volume, were transferred separately to preweighed trays, 

 dried to constant mass at 60°C, and weighed to the near- 

 est mg. Dry mass was not obtained for eight stomach con- 

 tents because of manipulation errors. The allometric rela- 

 tionship between the dry mass of stomach contents and 

 lobster CL was established by least squares linear regres- 

 sion, after logarithmic transformation of both variables. 



Diet was described by occurrence, volumetric contribu- 

 tion, and the specific abundance of food items in the stom- 

 achs of lobsters grouped into 5-mm CL size classes (2.5 to 

 <7.5 mm, 7.5 to <12.5 mm, etc.). Percent occurrence (PO) 

 was the percentage of stomachs in one size class that con- 

 tained a given food item. Volumetric contribution ( VC ) was 

 the average of corrected contributions of each food item 

 to the stomachs of all lobsters in a given size class. Spe- 

 cific abundance (SA) was the average volumetric contribu- 

 tion of a food item determined only for lobsters that had 

 this food item in their stomach. This index is useful for 

 food items with a low average volumetric contribution be- 

 cause it allows the distinction between the case when few 

 animals consume large quantities of a given food item or 

 when many animals consume small quantities of the same 

 food item (Amundsen et al., 1996). The mathematical rela- 

 tionship of the three indices is SA = VC x 100/PO. 



To assess how the overall diet varied with lobster size, 

 and thus whether or not there were size-related shifts 

 in diet supporting the ontogenetic phases of lobster, we 



