Acta Zoologica Taiwanica 13(2): 53 -62 (2002) Comparative Analysis of the Diets of Pygmy Sperm Whales and Dwarf Sperm Whales in Taiwanese Waters Ming-Chih Wang1, William A. Walker2, Kwang-Tsao Shao3, Lien-Siang Chou1* 1.Department of Zoology, National Taiwan University, No. 1 Roosevelt Rd., Sec. 4, Taipei, Taiwan 106, R.O.C. 2.National Marine Mammal Laboratory, Alaska Fisheries Science Center, National Marine Fisheries Service, NOAA, 7600 Sand Point Way N.E., Seattle, WA , USA 3.Institute of Zoology, Academia Sinica, No. 128, Academy Rd., Sec. 2, Nankang, Taipei, Taiwan115, R.O.C. ABSTRACT Stomach contents were analyzed of six pygmy sperm whales (Kogia breviceps) and five dwarf sperm whales (Kogia sima) which were taken as by-catch or were stranded specimens in coast al Taiwan from 1998 through 2000. Twenty-t wo species in 12 families of oceanic cephalopods were identified. In pygmy sp erm whales, Enoploteuthis chunii, Sthenoteuthis oualaniensis, and Taonius pavo were the primary prey in the diet, while E. chunii, Histioteuthis miranda, and T. pavo were the most important prey items ingested by dwarf sperm whales. Although the primary p rey items these two species ingested were very similar, each item comprised a different proportion for each whale species. A similarity test demonstrated a significant difference in prey composition, and SIMPER analysis showed that E. chunii was ranked first and contributed 37.1% to the average dissimilarity between pygmy and dwarf sperm whales. Pygmy sperm whales fed on much larger T. pavo compared to those ingested by dwarf sperm whales, while dwarf sperm whales ingested more H. miranda than did pygmy sperm whales. These results support the view that pygmy sperm whales live seaward of the continental shelf and that dwarf sperm whales live more in coastal waters. Key words:Kogia, Enoploteuthis chunii, cephalopod, stomack contents INTRODUCTION slope in eastern Taiwan (Yeh, 2000). Several studies have indicated that pygmy sperm whales Pygmy (Kogia breviceps) and dwarf sperm live mostly beyond the edge of the continental whales (Kogia sima) occur in tropical and tem- shelf, while dwarf sperm whales inhabit the shelf- perate latitudes worldwide. They are generally edge and slope waters (e.g., Rice, 1998). sighted in waters over the continental shelf and Several studies have reported the s tomach *Corresponding author: Department of Z oology, National Taiwan University , No. 1 Roosevelt Rd., Sec. 4, Taipei, Taiwan 106, R.O.C., Tel: +886-2-23661331, Fax: +886-2-23639902, E-mail: [email protected] 53 M. C. Wang, et. al. contents of stranded Kogia species, i.e., from from by-caught and stranded specimens of South Africa (Sekiguchi et al., 1992) and south- pygmy and dwarf sperm whales, res pectively, eas tern and s outhern Brazil (Santos and from the west and east coasts of Taiwan, during Haimovici, 2001). These reports s howed that 1998 through 2000. Stomachs were tied off at most of the stomach contents of Kogia species the esophageal and pyloric ends prior to being are oceanic cephalopods, with Histiotuethis spp. removed intact from the animal. Each stomach dominating the diets, with only a few remains of was tagged with an individual specimen number fishes and crustaceans being found (Sekiguchi and stored in a freezer. Data of each cetacean et al., 1992; Santos and Haimovici, 2001). individual, including its length and s ex, were Stranded animals are often used for diet analyses, collected in port. In the laboratory, stomachs were but they are usually unhealthy and have fewer thawed overnight at room temperature prior to stomach contents (Jones, 1981) than do non- preliminary sorting and specimen preservation. stranded animals, and may have different prey The stomach of each specimen was weighed full compositions (Leatherwood et al., 1978) in their and empty to the nearest 0.1 g with an A&D GF- stomachs; therefore, they might not provide data 6000 balance. representative of normal diets. In the laboratory, the contents of each stom- The beaks of cephalopod remains in stomach ach were carefully removed, separated into iden- contents are usually in a condition where identi- tifiable components, and drained of excess fluid. fication to family, genus , and often species is The stomach lining was thoroughly rinsed into a possible, and estimations from beak size can be shallow tray and then through a series of sieves made of their relative contribution to the diet by with mesh sizes of 1.4 and 0.5 mm, and 500 ìm weight and number of various taxa (Clarke, in order to recover all isolated cephalopod beaks. 1986). This is not only important for dietary Dorsal mantle length and weight of the cephalo- analysis, but also informative with respect to the pod prey that were whole or nearly whole were distribution and relative abundance of the vari- recorded, and the beaks were extracted from the ous cephalopod taxa in the sea, because many of intact prey and saved as reference material for the cephalopods eaten by oceanic cetaceans are species identification. The cephalopod beaks rarely caught in nets (Clarke, 1986). were sorted and preserved in 70% ethanol. Little is known about the stomach contents of Kogia species from Taiwanese waters . The Prey identification and enumeration present collections of the stomachs of two Kogia Cephalopod beaks were identified using the s pecies were from stranded and by-caught private reference collection of W. A. Walker and specimens. The aim of this paper was to provide from illus trations and keys presented in Clarke the first dietary information for pygmy and dwarf (1986). The maximum number of upper or lower sperm whales from Taiwanese waters. This study beaks was used to estimate the minimum num- also presents a qualitative description of the diet, ber of cephalopods ingested. The relative impor- and the relative importance of prey species as- tance of prey species was evaluated by means of sessed through frequency of occurrence, as well the frequency of occurrence and the percentage as numerical and biomass indices. by number (Hyslop, 1980). MATERIALS AND METHODS Estimation of prey size Sample collection Estimations of the original dors al mantle Six and five stomach s amples were collected length (DML) and weight of the commonly oc- 54 Tthe Diets of Pygmy Sperm Whales and Dwarf Sperm Whales curring prey were based on meas urements of each taxon to the average dissimilarity between lower beak rostral length (LBRL). The LBRL was the two Kogia species. The mean contribution of measured to the neares t 0.01 mm with vernier each species to the dissimilarity of two clusters calipers or an optical micrometer. Damaged or is defined as an average overall cross-group pair eroded specimens were not measured. The DML of samples. This yields an assessment of which and weight of Enoploteuthis chunii were es ti- prey species are diagnostic species between the mated using equations developed from intact two Kogia species. specimens in stomach contents. Regression equa- tions for other cephalopod prey were from Clarke RESULTS (1986). When identification to species level was not possible, cephalopod weight and length were Both pygmy sperm whales and dwarf sperm estimated by using regressions available for the whales fed on oceanic squid. Muscular squid genus or family. comprised more than 60% of the diets of these two Kogia species, both numerically and on a Comparative analysis weight basis (Table 1). Cephalopods inges ted Non-parametric multivariate techniques were we re n eutr ally buo yant s qu id s uch as used to compare differences between the diets of Histioteuthis miranda and Taonius pavo which the pygmy and dwarf sperm whales. The com- amounted to 30% of the specimens eaten and puter s oftware package PRIMER (Plymouth muscular squids like Enoploteuthis chunii and Routines In Multivariate Ecological Research) those of the Ommastrephidae which comprised was used in the analysis. The data were put into 70%. For both species of whales, E. chunii and triangular matrices bas ed on Bray-Curtis T. pavo made up about 70% of the total number similarities. An ANOSIM permutation test was of prey items ingested. They were als o the im- performed to test the null hypothesis that there portant prey on a weight basis. were no differences between species. It is based on a non-parametric permutation procedure and Kogia breviceps (Pygmy sperm whale) is applied to the ranked similarity matrix under- Eighteen species in nine families of squids lying the classification or ordination of samples. were identified by beaks or undiges ted remains. Randomization tes ts for significance were used The enoploteuthid squid, E. chunii, was the pri- for a priori selection by comparing the ranked mary prey, representing 48.6% of the total prey s imilarities of s amples between and within counts, with an overall occurrence of 50.0%. The species. ANOSIM derives a test statistic, R, from cranchiid squid, Taonius pavo, ranked second and a matrix of ranked similarities between all repli- made up 18.9% of the total number of prey cate samples, which represents the degree of dif- ingested, with an overall occurrence of 33.3%. ference between seasons or species. In theory, R The ommas trephid s quid, Sthenoteuthis can take any value between 1 and –1, and R = 1 oualaniensis, was the third most abundant prey if all replicate samples within a species are more item and represented 5.4% of the total number similar to each other than to any samples from of prey items, with an occurrence of 66.7% (Table other species, while an R value of 0 indicates that 2). The squids eaten were small to large sized, within group similarities are, on average, equal the estimated mantle length of the prey ranged to between-s pecies similarities.