404 



Fishery Bulletin 93(2), 1995 



Figure 1 



Locations of the swordhsh Xiphias gladius fishing grounds 

 in the central east Atlantic Ocean where samples were 

 obtained in 1990-91. 



and C respectively because they were used as bait. 

 Prey items were identified to the lowest taxonomic 

 category possible. Fish were identified from otoliths 

 and bones (otolith guides of Harkonen ( 1986), Hecht 

 and Hecht ( 1979), and author's collection). Crusta- 

 ceans were classified from external parts of the skel- 

 eton (Zariquiey -Alvarez, 1968). Lower beaks (LB) 

 were used as the primary means for classification of 

 cephalopods, and beak identity was established by 

 methods described by Clarke ( 1962, 1980, 1986a) and 

 supplemented by a collection of cephalopod beaks (in- 

 cluding beaks removed from locally caught cephalo- 

 pods); upper beaks, other morphological characters 

 (Hess and Toll, 1981), and distributional knowledge 

 (Nesis, 1987) were used in some cases as well. Nearly 

 all the beaks collected were fresh. All lower beaks 

 were identified to the lowest taxon possible and the 

 lower rostral length (LRL) or, in the order Sepiida, 



the hood length (LHL) (defined in Clarke, 1962, 1980) 

 were measured by digital caliper or steroscopic mi- 

 croscope and eyepiece micrometer. Mantle lengths 

 (ML in mm) and weights (W in g) of the cephalopods 

 from which lower beaks came were then estimated 

 from LRL's or LHL's by using relationships published 

 elsewere (Clarke 1962, 1980, 1986a; Perez-Gandaras, 

 1983; Wolff, 1982 (cited by Clarke, 1986a), Wolff and 

 Wormuth, 1979). 



Despite the advanced state of digestion of stom- 

 ach contents, an attempt was made to examine the 

 importance of each prey item. Two methods of stom- 

 ach content analysis were used: percent numerical 

 abundance and percent occurrence (Hyslop, 1980). A 

 coefficient of prey numerical frequency, % No. = {Nil 

 Nt) x 100, and a coefficient of prey frequency, % oc- 

 currence = (NsilNsf) x 100 were calculated; where 

 Ni is the number of prey of each group i, Nt is the 

 total number of prey, Nsi is the number of stomachs 

 containing each group i, and Nsf is the total number 

 of stomachs with food. An index of prey numerical 

 importance (Castro, 1993) was also obtained as % 

 importance = (% No. x % occurrence) 1 ' 2 x 100. 



Comparative analysis of degradation of hard 

 structures 



Free hard structures (otoliths, beaks, and eye lenses) 

 were often found in the stomach contents. To esti- 

 mate the importance of each prey item from the re- 

 fractory or hard structures, it is important to know 

 the rate of degradation by stomach acids. 



Otoliths (sagittae) from four chub mackerel, 

 Scomber japonicus (18 to 20 cm SL), similar in size 

 to those found in the swordfish stomachs, and five 

 otoliths of blue whiting, Micromesistius poutassou, 

 from head-specimens found in swordfish stomachs 

 were used to determine the rate of otolith degrada- 

 tion. Otolith lengths, taken on the longest axis, were 

 between 3.70 and 4.10 mm for Scomber and between 

 10.81 and 13.53 mm for Micromesistius. Beaks from 

 eleven Illex coindetii and two Todarodes sagittatus 

 were used, with LRL's between 3.90 and 6.19 mm 

 and 7.53 and 8.41 mm, respectively. Degradation of 

 six eye lenses, three belonging to T. sagittatus and 

 three belonging to fish species, were also analyzed. 



Hard structures were placed in a hydrochloric acid 

 (HC1) solution (pH=l.l; value lower than the range 

 reported by Jobling and Breiby [1986] for fish in 

 which the digestive process had begun), and the tem- 

 perature was maintained between 18 and 20°C. The 

 experiment was carried out for 48 hours. Old acid solu- 

 tions were replaced with fresh solutions after 24 hours. 



Cephalopod beak digestion (measured as decreas- 

 ing upper rostral length, lower rostral length, and 



