64 



Current Her pet ol. 20(2) 2001 



Fig. 1. Representatives of large- (a) and small-headed larvae (b) of Hynobiits retardatus (redrawn 

 from photographs). 



(Holomuzki et al., 1994). With respect to 

 H. retardatus larvae, however, no studies 

 have been performed on the effects of head 

 shape on their diet selectivity. 



1 compared the stomach contents 

 between large-headed and small-headed 

 larvae of H. retardatus collected from a 

 natural pond. The purpose of this study 

 was to examine the effect of larval head 

 shape on diet selection in this species. 



Materials and Methods 



A total of 44 larvae of H. retardatus 

 were sampled on 3 August 1998 from a 

 pond (330 m- surface area, 75 cm deep) 

 in Tomakomai Experimental Forest of 

 Hokkaido University (TOEF; 42°37'N, 

 141°20'E), southwestern Hokkaido, Japan. 

 The species lays eggs in the pond from late 

 April to early May and metamorphoses in 

 September. 1 captured larvae at the middle 

 stage (i.e., digital differentiation 1-Vl; 



stage 48-53 of Iwasawa and Yamashita, 

 1991). The sampled larvae were immedi- 

 ately sacrificed with tricane methane 

 sulfonate (MS-222), and were measured for 

 snout-vent length (SVL, distance from 

 anterior tip of snout to posterior margin of 

 vent) and head width (HW, width of head 

 at the broadest portion) with a digital 

 caliper to the nearest 0.1 mm. Their stom- 

 ach contents were then examined by 

 abdominal dissection under a binocular 

 microscope. Each food item was first cate- 

 gorized as salamander larvae, detritus, and 

 terrestrial and aquatic invertebrates. Items 

 categorized as invertebrates were further 

 identified to the order level (Appendix). 

 The wet mass of items composing each 

 dietary category was weighed to the nearest 

 0.01 mg after blotting. 



The head shape of larval H. retardatus 

 shows continuous variation within a popu- 

 lation (Nishihara, 1996). To assess the 

 head shape of larvae, residuals of a simple 



