Mendez-Macfas et al.: Trophic ecology and ontogenetic shift in the diet of Mustelus lunulatus in the southeastern Pacific Ocean 
247 
The specimens collected during November 2003- 
October 2004 were grouped by trimester in chronological 
order: November 2003-January 2004 (Trimester I), Febru¬ 
ary 200A-April 2004 (Trimester II), May 2004—July 2004 
(Trimester III), and August 2004-October 2004 (Trimester 
IV). The information was grouped in this way in order to 
obtain a significant quantity for each trimester of the year. 
To test for ontogenetic shifts in diet, a Bray-Curtis sim¬ 
ilarity matrix was constructed by using standardized esti¬ 
mates of the contribution of each prey taxon to the diet 
based on abundance (number of samples [n]). A one-way 
analysis of similarities (ANGSIM) was then performed, by 
using individuals as samples and stages of sexual maturity 
(sexes separated) and size classes (sexes combined) as fac¬ 
tors. This test is analogous to analysis of variance and was 
used to evaluate similarity (or difference) within predefined 
groups (factors) versus similarity between groups and to 
calculate the statistic R, which varies between -1 and +1 
(Clarke and Gorley, 2006). A post-hoc multiple comparison 
test (Benfferroni test) was then performed to identify spe¬ 
cific differences between categories. 
A nonmetric multidimensional scaling analysis was 
used to determine the maturity differences by diet and to 
determine the size-class dietary changes that occurred. 
Similarity percentage (SIMPER) analysis was used to 
estimate the contribution of each prey category to ontoge¬ 
netic differences in diet. The analyses were performed by 
using the software PAST, vers. 3.14 (Hammer, 2001). 
Dietary niche breadth, overlap, and trophic position 
Dietary niche breadth was estimated by using Levin’s 
index (B ; ) (Krebs, 1999): 
( 2 ) 
where Pj is the fraction (in PSIRI) of each food item in 
the diet j (ZPj=l). The B i values were standardized to 
B a so that they ranged from 0 to 1 by using this equa¬ 
tion: B^fBj-lXIV-ir 1 , where N is the number of classes 
(Krebs, 1999). Low B A values indicate diets dominated by 
few prey items (i.e., selective predators), and values close 
to 1 indicate opportunistic, generalist diets. 
Trophic overlap was assessed by calculating the Moris- 
ita-Horn index (CA.) (Smith and Zaret, 1982) to detect pos¬ 
sible differences in diet between sexes, stages of sexual 
maturity, and size classes: 
Ck = 2- 
I'-,( p * x 4.) 
V p 2 +\ n p 2 
(3) 
where P^ is the proportion of t th prey with respect to all 
prey of predator x, P vi is the proportion of the i th prey with 
respect to predator y, and n is the total number of prey 
species. This index ranges from 0 to 1, with values close 
to 0 indicating dietary differences and values close to 1 
indicating similarity in prey consumed. 
Standardized trophic position (TP) was calculated by 
using the trophic index proposed by Cortes (1999): 
TP k =l + 
f \ 
y £P 1 xTP ] 
V i=1 
(4) 
where n is the number of prey, TP j is the trophic position 
of each prey taxon j (Cortes, 1999), and Pj is the proportion 
of each prey category j in the predator’s diet, based on IRI 
values (percentages) (Bornatowski, 2014a). The trophic 
positions of all prey were taken from Cortes (1999), Hob¬ 
son and Welch (1992), Sea Around Us (Pauly and Zeller 1 ), 
and FishBase (Froese and Pauly 2 ). 
For all analyses, the crustacean, fish, and squid 
remains and unidentified organic matter were not con¬ 
sidered because it was not possible to identify them with 
precision. 
Results 
A total of 314 stomachs were collected from specimens of 
the sicklefin smoothhound measuring 63-142 cm TL, and 
95.2% (n=299) of the stomachs contained food. Of these 
stomachs, 171 were taken from females (69-142 cm TL) 
and 143 were taken from males (63-118 cm TL), with 
96.5% (n=165) and 93.7% (rc=134) of the stomachs contain¬ 
ing food, respectively. 
Regarding the stages of sexual maturity, there were 90 
juvenile females and 81 adult females, of which 96.7% and 
96.3%, respectively, had food in their stomachs. Of the 
54 juvenile males and 90 adult males, 88.9% and 96.7%, 
respectively, had food in their stomachs. 
When examined by size class, 91.0% (/z=100) of the 
Size-I sharks (60-90 cm TL), 97.3% (n=183) of the Size-II 
sharks (90-120 cm TL), and 96.9% (n=32) of the Size-Ill 
sharks (120-150 cm TL) had food in their stomachs. 
In addition, of the specimens studied in Trimester I, Tri¬ 
mester II, Trimester III, and Trimester IV, 91.6% (n= 87), 
98% (n=58), 92% (n= 73), and 100% (n=81), respectively, 
contained food in their stomachs. 
Cumulative prey curve 
The coefficient of variation indicates that all the cumula¬ 
tive prey curves approached their asymptote. This finding 
indicates that sample sizes were sufficient to adequately 
describe the diet of sicklefin smoothhounds (Fig. 2A). 
Indeed, the minimum number of stomachs required was 
58 for females (51 for juveniles and 44 for adults; Fig. 2, 
B, D, and F) and 75 for males (44 for juveniles and 38 for 
adults; Fig. 2, C, E, and G). 
1 Pauly, D., and D. Zeller (eds.). 2015. Sea Around Us concepts, 
design, and data. [Available from website, accessed June 
2016], 
2 Froese, R., and D. Pauly. 2016. FishBase, vers. 06/2016. World 
Wide Web electron, publ. [Available from website, accessed 
June 2016], 
