FULL PAPER Anatomy

Morphological Adaptation of the Skull for Various Behaviors in the Tree Shrews

Hideki ENDO1), Tsutomu HIKIDA2), Masaharu MOTOKAWA3), Loke Ming CHOU4), Katsuhiro FUKUTA5) and Brian J. STAFFORD6,7)

1)Department of Zoology, National Science Museum, Tokyo, 3–23–1 Hyakunin-cho, Shinjuku-ku, Tokyo 169–0073, 2)Department of Zoology, Faculty of Science, 3)The University Museum, Kyoto University, Kyoto 606–8501, Japan, 4)Raffles Museum of Biodiversity Research, National University of Singapore, Singapore, 5)Laboratory of Morphology and Function, Graduate School of Bioagricultural Sciences, Nagoya University, Nagoya, Aichi 464–8601, Japan, 6) Division, National Museum of Natural History, Smithsonian Institution, Washington D.C. and 7)Department of Anatomy, Howard University College of Medicine, Washington D.C., U.S.A.

(Received 5 February 2003/Accepted 8 April 2003)

ABSTRACT. Skull size and shape were examined among 14 species of the tree shrews ( montana, T. picta, T. splendidula, T. mulleri, T. longipes, T. glis, T. javanica, T. minor, T. gracilis, T. dorsalis, T. tana, melanura, D. murina, and Ptilocercus lowii). The bones of face were rostro-caudally longer in T. tana and T. dorsalis, contrasting with T. minor and T. gracilis, D. melanura, D. murina and P. lowii which have smaller facial length ratios. The arbo-terrestrial species (T. longipes and T. glis) were similar to terrestrial spe- cies in length ratios of bones of face unlike the other arbo-terrestrial species (T. montana, T. picta, T. splendidula, and T. mulleri). We propose that T. longipes and T. glis have adapted to foraging for termites and ants as have T. tana and T. dorsalis. Additionally small body size in T. javanica may be the result of being isolated in Java. We separated the species into 5 groups from the measurment values of skulls: 1) Terrestrial species; T. tana and T. dorsalis, 2) Arboreal species; T. minor and T. gracilis, 3) Arbo-terrestrial species group 1: T. montana, T. splendidula, T. picta and T. mulleri, and T. javanica, 4) Arbo-terrestrial species group 2: T. glis and T. longipes, 5) Arboreal species of Dendrogale and Ptilocercus. Principal component analysis separated species into 8 clusters as follows: 1) T. tana, 2) T. dorsalis, 3) T. montana, T. splendidula, T. picta and T. mulleri, 4) T. glis and T. longipes, 5) T. javanica, 6) T. minor and T. gracilis, 7) D. melanura and D. murina, and 8) P. lowii. We suggest that these clusters correspond to behavioral strategies and peculiarities observed in foraging, feeding and locomotion in each species. KEY WORDS: adaptation, behavior, osteometry, skull, . J. Vet. Med. Sci. 65(8): 873–879, 2003

Tree shrews (Order Scandentia) consist of about 18 spe- National University of Singapore, in the Department of cies [1, 2, 6, 13, 22]. Since this order shows us the evolu- Wildlife and National Parks (Kuala Lumpur, Malaysia), and tionary process of arboreal adaptation in terrestrial , in Bogor Zoological Museum (Bogor, Indonesia). Sex the morphological variation in locomotor and feeding mech- determination was dependent on the description of biologi- anisms is noticeable among species [1, 22]. The behavior of cal data of specimens. Only specimens with fully erupted various species has been also detailed in field works [7]. molars were considered to be adults. Species composition, Previously we functional-morphologically examined 4 spe- origin, and sex are shown in Table 1. Skull measurements cies of Tupaia, and suggested that T. tana, T. javanica and were obtained with vernier calipers to the nearest 0.05 mm. T. minor have evolved different behaviors [10]. Here we Measurements are defined in Table 2, and were based on extend this study to include most members of Scandentia Driesch [5]. All measurements were then divided by the and to quantitatively examine morphological adaptations for geometric mean of PL in each species to remove the effects various behaviors (i.e. locomotion, foraging, feeding and of size. In these measurement ratios statistical differences nesting). In this study, we use 14 species of tree shrews among species were examined in the non-parametric U-test including Dendrogale and Ptilocercus to compare interspe- by the use of the software of Statistica (StatSoft, Inc., cies variations, and to clarify the adaptation strategy in skull Tokyo, Japan). The t-test was not used, since the normal morphology. distribution is not guaranteed with measurement ratios. Principal component analysis was used with all measure- MATERIALS AND METHODS ment data to examine variation among taxa. A package soft- ware for multivariate analysis (Shakai-Joho Service, Tokyo, We examined 337 skulls of 14 species of tree shrews Japan) added to Microsoft Excel 98 was used for this analy- (Tupaia montana, T. picta, T. splendidula, T. mulleri, T. lon- sis. gipes, T. glis, T. javanica, T. minor, T. gracilis, T. dorsalis, T. tana, Dedrogale melamura, D. murina, and Ptilocercus RESULTS lowii). The specimens have been stored in Muséum National d’Histoire Naturelle (Paris, France), in the Smith- The species were grouped into 5 morphological types sonian Institution, in The University Museum, Kyoto Uni- according to the data of the osteometrical ratio data in each versity, in the Raffles Museum of Biodiversity Research, species (Tables 3–5). In this grouping, the pattern of behav- 874 H. ENDO ET AL.

Table 1. Species, behavioral character, origin and sex composition of the specimens Species Behavior Origin of Specimens Male Female Tupaia montana Arbo-terrestrial Borneo 52 54 Tupaia picta Arbo-terrestrial Borneo 3 0 Tupaia splendidula Arbo-terrestrial Borneo 2 2 Tupaia mulleri Arbo-terrestrial Borneo 1 1 Tupaia longipes Arbo-terrestrial Borneo 9 3 Tupaia glis Arbo-terrestrial Malayan Peninsula 22 38 Tupaia javanica Arbo-terrestrial Java 9 6 Tupaia minor Arboreal Borneo, Sumatra 5 4 Tupaia gracilis Arboreal Borneo 14 6 Tupaia tana Terrestrial Borneo, Sumatra 24 26 Tupaia dorsalis Terrestrial Borneo 2 3 Dendrogale melanura Arboreal Borneo 10 5 Dendrogale murina Arboreal Thailand, Vietnam 2 4 Ptilocercus lowii Arboreal Malayan Peninsula 13 17 Total 168 169

Table 2. List of skull and mandibular measurements and restrial species (Tables 3–5). The skull was laterally wider their abbreviations in the two species, and the ratios GNB/PL and GMB/PL Cranium were much larger than those in the species of other locomo- Profile length PL tion. Condylobasal length CL Arbo-terrestrial species group 1, T. montana, T. picta, T. Short lateral facial length SL splendidula, T. mulleri and T. javanica: The strong elonga- Zygomatic width ZW tion was not confirmed in these species unlike T. tana and T. Least breadth between the orbits LBO dorsalis, although the bones of face was relatively longer Greatest neurocranium breadth GNB than that of arboreal species. Median palatal length MPL Length from Basion to Staphylion LBS Arbo-terrestrial species group 2, T. longipes and T. glis: Dental length DL The ratios SL/PL, MPL/PL, DL/PL and LIA/PL were large Greatest palatal breadth GPB (Tables 3–5). The bones of face in the two species were ros- Greatest mastoid breadth GMB tro-caudally longer than the arbo-terrestrial group 1 in many Height from Akrokranion to Basion HAB cases. Especially the ratios DL/PL and LIA/PL of T. lon- Mandible gipes and T. glis were not different from those of terrestrial Length from the condyle LC Length from the angle LA T. tana and T. dorsalis (Tables 3–5). Length from Infradentale LIA Arboreal species of Dendrogale and Ptilocercus: The to aboral border of the alveolus bones of face were not elongated in these species, and length Height of the vertical ramus HR ratio is extremely small in P. lowii. In contrast the width Height of the mandible at M1 HM ratios ZW/PL and GMB/PL were large in Ptilocercus. First and second principal components represent size and proportion factor, respectively. We could point out that ior and locomotion such as terrestrial, arbo-terrestrial and these components segregated specimens into 8 clusters of arboreal in each species was based upon the descriptions species that were similar for both sexes (Fig. 1, Table 6): and the original data of field works (Table 1) [1, 2, 6, 7, 13, Cluster 1, T. tana, Cluster 2, T. dorsalis, Cluster 3, T. mon- 22]. We also distinguished the two groups of arbo-terres- tana, T. splendidula, T. picta, and T. mulleri, Cluster 4, T. trial species from the morphological similarities. glis and T. longipes, Cluster 5, T. javanica, Cluster 6, T. Terrestrial species, T. tana and T. dorsalis: The facial minor and T. gracilis, Cluster 7, D. melanura and D. part was rostro-caudally elongated in the skull. The ratios murina, and Cluster 8, P. lowii. SL/PL, MPL/PL, DL/PL and LIA/PL were larger than those of the other species (Tables 3–5). The ratios SL/PL and DISCUSSION MPL/PL were especially large in both sexes of the two spe- cies. The elongation of the bones of face resulted in small The skull is morphologically divided into two major ratios of skull width in terrestrially-adapted species, and the functional units, bones of cranium and bones of face, to clar- ratios ZW/PL, LBO/PL, GNB/PL and GMB/PL were much ify adaptational strategy in each species and behavior pat- smaller (Tables 3–5). tern. Measurement ratios of the bones of face are Arboreal species, T. minor and T. gracilis: The bones of noteworthy (Tables 3 and 4). SL/PL, MPL/PL, DL/PL and face were obviously short, and the ratio SL/PL was statisti- LIA/PL represent rostro-caudal length ratio in the bones of cally different from that of terrestrial species and arbo-ter- face. These ratios are relatively large in both T. tana and T. SKULL ADAPTATION IN TREE SHREWS 875

Table 3. Mean and SD in mm for craniometric measurement in various species SEX PL CL SL ZW LBO GNB MPL LBS DL GPB GMB HAB LC LA LIA HR HM Tupaia montana M 48.41 41.62 20.99 25.30 13.80 18.61 25.27 16.49 24.86 15.55 19.28 11.43 32.64 32.74 20.03 13.28 3.49 1.10 1.10 0.81 0.84 0.55 0.47 0.89 0.57 0.84 0.53 1.00 0.41 1.09 1.15 0.72 0.76 0.26 F 48.18 41.55 20.89 24.84 13.55 18.64 25.31 16.45 24.84 15.60 19.10 11.31 32.35 32.45 20.00 13.29 3.38 1.19 1.14 0.80 0.85 0.56 0.62 0.81 0.57 0.75 0.51 1.02 0.46 0.89 0.92 0.66 0.59 0.24 Tupaia picta M 51.08 44.40 22.33 26.28 14.82 19.90 27.33 17.32 27.02 16.57 21.28 12.30 35.27 35.48 21.73 14.83 3.87 1.79 2.04 0.06 0.68 0.88 1.11 1.50 0.86 1.41 0.94 1.12 0.71 2.00 2.45 1.63 0.63 0.10 Tupaia M 47.63 41.58 20.03 25.45 13.78 18.90 24.65 17.30 24.55 15.00 19.83 11.55 32.55 32.93 19.65 14.23 3.78 splendidula 1.03 0.67 0.88 0.42 0.11 0.78 0.78 0.07 0.78 0.35 0.32 0.14 0.07 0.11 0.49 0.32 0.46 F 47.68 41.55 20.30 24.45 12.88 19.00 25.15 16.48 24.75 15.55 19.78 11.65 32.23 32.03 20.18 13.50 3.38 1.52 0.99 0.85 0.57 0.18 1.34 0.42 0.39 0.57 0.78 0.74 0.00 1.10 0.53 0.46 0.57 0.18 Tupaia mulleri* M 47.75 40.10 19.35 25.00 13.80 18.80 24.55 15.80 24.35 15.30 19.95 12.45 32.20 32.55 19.80 13.75 3.40 F 48.70 42.35 20.60 25.50 13.50 18.60 25.60 17.10 24.65 15.40 19.60 11.35 33.30 33.45 20.40 13.80 3.55 Tupaia longipes M 52.34 45.42 22.46 25.59 14.49 19.85 28.34 17.22 27.97 17.67 21.45 12.70 35.86 35.73 22.62 14.20 3.91 1.02 1.05 0.74 1.76 0.53 0.49 0.75 0.35 0.59 0.47 0.75 0.46 0.88 0.81 0.56 0.53 0.27 F 52.20 45.03 22.80 26.55 14.88 20.98 28.28 17.18 27.80 17.88 21.85 13.12 35.85 35.59 22.57 14.67 4.12 1.48 1.38 1.05 0.78 0.53 0.65 1.20 0.12 0.95 0.21 0.74 0.19 0.73 0.48 0.68 0.88 0.33 Tupaia glis M 52.57 45.59 23.02 26.34 14.85 19.67 28.18 17.72 27.64 16.39 21.01 12.51 36.02 36.14 22.61 14.33 3.91 1.07 1.06 0.96 0.87 0.67 0.30 0.81 0.54 0.62 0.53 0.58 0.42 1.03 1.12 0.68 0.84 0.39 F 51.81 44.98 22.49 25.34 14.38 19.70 27.68 17.54 27.32 16.38 20.74 12.22 35.10 35.22 22.23 13.81 3.74 1.14 1.04 0.73 0.89 0.54 0.45 0.72 0.54 0.79 0.50 0.47 0.35 0.86 0.89 0.64 0.52 0.25 Tupaia javanica M 42.31 36.89 16.92 22.14 12.65 17.53 22.04 15.02 21.40 13.37 16.10 10.30 28.35 28.34 17.08 11.85 3.11 1.11 0.79 0.41 0.63 0.50 0.54 0.52 0.49 0.47 0.32 0.89 0.41 0.97 0.72 0.33 0.57 0.25 F 42.48 36.82 16.90 21.70 12.48 17.58 22.02 14.88 21.38 13.54 15.87 10.43 27.98 28.46 17.35 11.78 3.19 0.91 0.47 0.63 1.09 0.58 0.50 0.45 0.48 0.86 0.66 0.85 0.38 0.82 0.90 0.34 0.59 0.31 Tupaia minor M 36.09 30.85 13.26 20.15 11.86 16.19 17.31 13.38 17.54 11.85 14.92 9.30 24.05 23.84 14.36 10.48 2.74 0.42 0.44 0.32 0.64 0.70 0.22 0.43 0.55 0.20 0.26 0.24 0.28 0.37 0.30 0.53 0.29 0.22 F 36.19 30.51 13.40 19.40 11.54 16.15 17.41 12.96 17.59 11.75 14.91 9.31 24.10 23.74 14.26 9.98 2.76 0.48 0.56 0.20 0.50 0.45 0.74 0.26 0.45 0.30 0.29 0.54 0.40 0.49 0.39 0.38 0.26 0.18 Tupaia gracilis M 38.86 32.87 15.00 20.53 12.17 17.16 19.01 13.85 19.08 12.79 16.83 10.26 25.31 25.24 15.33 10.75 2.72 0.63 0.63 0.36 0.56 0.39 0.48 0.71 0.24 0.40 0.45 0.54 0.55 0.49 0.67 0.34 0.46 0.20 F 39.26 32.93 15.13 20.78 12.06 17.36 19.35 13.85 19.27 12.94 17.04 10.18 25.38 25.37 15.50 10.66 2.78 1.32 1.04 1.06 1.06 0.50 0.74 0.88 0.45 0.90 0.63 0.94 0.33 0.94 0.87 0.87 0.43 0.19 Tupaia tana M 60.59 52.72 30.17 26.67 15.35 20.57 33.96 19.21 31.99 15.90 21.45 13.40 41.07 40.82 25.97 13.78 3.99 2.31 2.16 1.54 1.30 0.78 0.54 1.71 0.68 1.47 0.56 1.27 0.57 1.94 2.36 1.58 1.08 0.41 F 59.33 51.54 29.07 26.09 14.96 20.25 33.08 18.86 31.50 16.01 21.23 13.21 40.50 40.45 25.65 13.80 3.94 1.90 1.86 1.45 0.99 0.63 0.53 1.41 0.55 1.30 0.54 1.31 0.53 1.43 1.61 1.17 0.91 0.30 Tupaia dorsalis M 50.00 42.60 24.33 22.30 13.33 19.15 27.75 15.20 26.30 14.45 18.50 11.60 33.20 32.75 21.63 10.90 3.15 0.28 1.48 0.53 0.07 0.25 0.42 0.28 1.13 0.28 0.07 0.64 0.92 0.85 0.92 1.31 0.42 0.14 F 49.35 43.27 24.40 22.48 13.07 18.12 27.80 15.33 26.30 13.93 18.08 11.37 33.55 33.25 21.67 11.13 3.30 1.21 0.78 0.61 0.03 0.49 0.38 0.66 0.59 0.44 0.06 0.94 0.24 0.49 0.53 0.45 0.96 0.13 Dendrogale M 34.42 29.03 14.30 17.07 9.98 14.99 17.33 11.82 16.71 9.70 13.94 8.70 21.03 20.55 13.10 8.97 2.09 melanura 0.73 0.70 0.56 0.55 0.37 0.26 0.59 0.31 0.40 0.29 0.37 0.41 0.49 0.51 0.26 0.29 0.15 F 33.76 28.37 13.99 16.44 9.70 15.22 16.81 11.71 16.58 9.70 13.81 8.58 20.55 20.25 12.95 8.45 2.06 0.69 0.66 0.59 0.52 0.13 0.43 0.59 0.26 0.49 0.40 0.51 0.49 0.67 0.58 0.64 0.37 0.10 Dendrogale M 36.13 31.25 14.58 18.73 10.95 14.60 19.05 12.48 18.70 10.83 13.98 9.00 23.58 22.60 15.03 9.75 2.48 murina 0.18 0.07 0.18 1.38 0.35 0.07 0.21 0.04 0.14 0.18 0.46 0.21 0.18 0.64 0.04 0.35 0.18 F 33.81 29.15 13.20 17.23 10.21 14.73 17.26 11.86 17.35 10.66 13.75 8.81 21.66 21.55 13.80 8.89 2.44 0.19 0.14 0.40 0.35 0.32 0.31 0.38 0.37 0.12 0.23 0.35 0.29 0.61 0.60 0.51 0.28 0.05 Ptilocercus lowii M 37.15 33.22 12.07 21.59 8.25 14.57 17.69 15.90 17.27 11.36 16.20 8.41 24.89 25.96 14.71 11.59 2.89 0.64 0.73 0.33 0.48 0.25 0.36 0.45 0.44 0.31 0.27 0.28 0.20 0.53 0.67 0.29 0.20 0.17 F 37.36 33.39 12.18 21.44 8.11 14.54 17.47 16.06 17.29 11.37 16.23 8.35 25.11 26.18 14.69 11.50 3.05 0.89 0.94 0.59 0.77 0.26 0.26 0.45 0.40 0.45 0.30 0.32 0.30 0.74 0.74 0.42 0.54 0.16 Mean values are shown in the upper rows, and standard deviations in the lower rows. * Only one specimen was used for this species in each sex. dorsalis. Although the behavioral data have not been In contrast, these ratios are smaller in the arboreal T. reported in T. dorsalis, we suggest that this species may minor and T. gracilis. In SL/PL, statistical differences are show a terrestrial form similar to T. tana [10]. The charac- found between species adapted for different locomotion, ters of bones of face in these two species are related to dig- foraging, feeding and nesting behaviors (Tables 1 and 5). ging activity and feeding on ants and invertebrates in soil Unlike the rostro-caudal elongation in terrestrial species, [7]. specialization in length of bones of face is not shown in 876 H. ENDO ET AL.

Table 4. The measurement ratios (%) in various species of tree shrews SEX SL/PL ZW/PL LBO/PL GNB/PL MPL/PL DL/PL GMB/PL LIA/PL LBO/LBS Tupaia montana M 43.35 52.28 28.52 38.47 52.21 51.35 39.83 41.38 83.74 F 43.36 51.56 28.12 38.70 52.53 51.55 39.66 41.52 82.38 Tupaia picta M 43.76 51.48 28.99 38.94 53.49 52.87 41.66 42.51 85.85 Tupaia splendidula M 42.04 53.46 28.93 39.68 51.75 51.54 41.63 41.28 79.61 F 42.57 51.29 27.01 39.83 52.77 51.92 41.48 42.32 78.16 Tupaia mulleri M 40.52 52.36 28.90 39.37 51.41 50.99 41.78 41.47 87.34 F 42.30 52.36 27.72 38.19 52.57 50.62 40.25 41.89 78.95 Tupaia longipes M 42.89 48.88 27.69 37.94 54.15 53.44 40.98 43.22 84.19 F 43.66 50.86 28.51 40.23 54.17 53.25 41.86 43.23 86.57 Tupaia glis M 43.78 50.10 28.24 37.43 53.60 52.59 39.98 43.01 83.78 F 43.41 48.92 27.76 38.03 53.44 52.74 40.04 42.92 81.96 Tupaia javanica M 40.01 52.34 29.90 41.45 52.11 50.59 38.05 40.37 84.23 F 39.77 51.06 29.36 41.39 51.83 50.32 37.33 40.84 83.77 Tupaia minor M 36.74 55.83 32.86 44.86 47.96 48.60 41.34 39.79 86.96 F 37.03 53.62 31.89 44.64 48.12 48.60 41.22 39.41 89.04 Tupaia gracilis M 38.62 52.84 31.32 44.17 48.91 49.11 43.32 39.46 87.86 F 38.49 52.90 30.71 44.21 49.28 49.06 43.41 39.46 87.03 Tupaia tana M 49.77 44.03 25.33 33.97 56.03 52.78 35.39 42.85 79.84 F 48.97 43.99 25.22 34.15 55.74 53.08 35.78 43.23 79.33 Tupaia dorsalis M 48.65 44.60 26.65 38.30 55.50 52.60 37.00 43.24 87.68 F 49.44 45.58 26.48 36.71 56.34 53.31 36.63 43.93 85.22 Dendrogale melanura M 41.53 49.60 28.99 43.57 50.35 48.56 40.52 38.08 84.38 F 41.43 48.69 28.74 45.10 49.78 49.10 40.90 38.36 82.84 Dendrogale murina M 40.35 51.84 30.31 40.42 52.73 51.76 38.69 41.59 87.78 F 39.04 50.95 30.20 43.55 51.06 51.31 40.66 40.81 86.08 Ptilocercus lowii M 32.49 58.11 22.21 39.22 46.82 46.48 43.60 39.59 51.88 F 32.59 57.38 21.72 38.95 46.78 46.30 43.46 39.34 50.53

arbo-terrestrial species such as T. montana, T. picta, T. to T. longipes (Table 5). splendidula, T. mulleri and T. javanica. SL/PL, MPL/PL, We first attributed the small value of LBO to the estab- DL/PL and LIA/PL are also large in T. longipes and T. glis. lishment of binocular vision in the arboreal life. The LBO/ SL/PL and MPL/PL are largest in T. tana and T. dorsalis LBS (least breadth between orbits per length of bones of (Table 4). In the statistical analysis T. glis hardly indicates cranium) was also calculated in each species to remove the statistical differences from T. longipes in these two ratios effect of facial elongation in terrestrial species. T. tana is (Table 5). However, DL/PL and LIA/PL of T. longipes and not significantly larger in LBO/LBS ratio, but smaller than T. glis are as large as those of T. tana and T. dorsalis (Tables the arboreal and arbo-terrestrial species (Tables 4 and 5). 4 and 5). Although T. longipes and T. glis are arbo-terres- We could conclude that the development of the binocular trial species, the data indicate dental row extension in the vision was not confirmed in arboreal species within the two species as extensive as in completely terrestrial species. Scandentia from the data of relative size of LBO, although Since T. longipes and T. glis are dependent on termites and P. lowii exhibits the smallest LBO/PL and LBO/LBS. ants in soil [7], we suggest that the similarities in the length For many skull ratios, significant differences do not exist ratios of bones of face between tana-dorsalis and longipes- between T. longipes and T. glis, between T. minor and T. glis are consistent with the searching and foraging behavior gracilis, and between T. tana and T. dorsalis (Table 5). We on the ground in both groups. In fact, T. longipes also lives suggest that the morphological similarities in these pairs are on trees more than 1.6 m in height, although the species has consistent with behavioral partitioning, because T. minor been observed on ground and in low branches [7]. We and T. gracilis are typically arboreal and T. tana and T. dor- should examine whether the grouping of the two species as salis may be completely terrestrial [1, 2, 6, 7, 13, 22]. In T. arbo-terrestrial animal is valid in further field observations. dorsalis significant distinctions are not found in comparison Width ratio, ZW/PL, LBO/PL, GNB/PL and GMB/PL, with other species in some ratios. This may be due to small are much smaller in the terrestrial T. tana and T. dorsalis sample size for this species. (Table 4). In contrast, GNB/PL and GMB/PL are much Dendrogale and Ptilocercus are arboreal [1, 7, 8, 14, 22], larger in the arboreal T. minor and T. gracilis except for the and length ratios of bones of face are smaller in these spe- female of GMB/PL. In these three ratios, differences are not cies. Each length ratio is extremely small in P. lowii in both statistically significant between T. minor and T. gracilis. sexes. ZW/PL and GMB/PL show the largest values in These two species are significantly different from all other Ptilocercus, indicating the development of wide zygomatic species by the U-test except for the female GMB/PL related arch and jaw articulation in this species. LBO/PL is small- SKULL ADAPTATION IN TREE SHREWS 877

Table 5. Significant differences of measurement ratios among various species of tree shrews SL/PL ZW/PL LBO/PL GNB/PL MPL/PL DL/PL GMB/PL LIA/PL LBO/LBS 1. Male T. montana X T. longipes 11.27 0.10 4.63 10.82 0.00 0.00 13.77 0.00 72.96 T. montana X T. glis 26.63 0.00 29.26 0.02 0.00 0.00 80.39 0.00 69.20 T. montana X T. minor 0.00 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.54 T. montana X T. gracilis 0.00 25.88 0.00 0.00 0.00 0.00 0.00 0.00 0.01 T. montana X T. tana 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 T. montana X T. dorsalis 1.72 1.72 3.13 96.35 1.95 5.44 6.69 14.27 23.37 T. longipes X T. glis 6.13 27.66 22.30 38.41 11.72 0.70 7.44 86.18 93.06 T. longipes X T. minor 0.00 0.00 0.00 0.00 0.00 0.00 0.28 0.00 12.45 T. longipes X T. gracilis 0.01 0.04 0.01 0.01 0.01 0.01 0.07 0.01 1.98 T. longipes X T. tana 0.00 0.04 0.01 0.00 0.01 3.56 0.00 68.60 0.86 T. longipes X T. dorsalis 3.39 9.90 23.86 63.74 3.39 5.94 3.39 100.00 34.58 T. glis X T. minor 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 5.41 T. glis X T. gracilis 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.53 T. glis X T. tana 0.00 0.00 0.00 0.00 0.00 34.44 0.00 69.22 0.19 T. glis X T. dorsalis 2.16 2.16 4.72 25.06 2.83 83.45 2.83 75.40 21.01 T. minor X T. gracilis 0.00 0.48 21.85 30.21 3.54 1.24 69.14 72.09 41.58 T. minor X T. tana 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 T. minor X T. dorsalis 2.39 2.39 2.39 2.39 2.39 2.39 2.39 2.39 100.00 T. gracilis X T. tana 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 T. gracilis X T. dorsalis 2.62 2.62 2.62 2.62 2.62 2.62 2.62 3.90 100.00 T. tana X T. dorsalis 12.37 38.65 6.75 2.09 21.10 50.06 14.89 84.74 3.43 D. melanura X D. murina 8.57 51.93 13.26 3.17 3.17 3.17 13.26 3.17 39.02 D. melanura X P. Lowii 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.13 0.01 D. murina X P. Lowii 2.73 2.73 2.73 12.64 2.73 2.73 2.73 2.73 2.73

2. Female T. montana X T. longipes 59.19 17.45 43.17 13.34 1.01 1.01 4.54 0.59 5.36 T. montana X T. glis 91.79 0.00 15.58 2.10 0.00 0.00 70.05 0.00 73.61 T. montana X T. minor 0.09 2.32 0.09 0.09 0.09 0.09 23.14 0.19 1.97 T. montana X T. gracilis 0.01 2.83 0.01 0.01 0.01 0.02 0.05 0.08 0.36 T. montana X T. tana 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.07 T. montana X T. dorsalis 0.38 0.38 1.37 0.82 0.38 2.93 2.44 0.82 12.44 T. longipes X T. glis 51.52 4.01 16.10 1.63 13.31 34.15 0.80 58.19 4.01 T. longipes X T. minor 3.39 3.39 3.39 7.71 3.39 3.39 28.89 3.39 28.89 T. longipes X T. gracilis 2.01 7.07 2.01 2.01 2.01 2.01 12.13 2.01 79.63 T. longipes X T. tana 0.52 0.52 0.52 0.52 1.22 77.45 0.52 88.61 0.81 T. longipes X T. dorsalis 4.95 4.95 4.95 4.95 4.95 51.27 4.95 82.73 51.27 T. glis X T. minor 0.11 0.15 0.11 0.11 0.11 0.11 15.74 0.11 1.84 T. glis X T. gracilis 0.01 0.02 0.01 0.01 0.01 0.01 0.02 0.01 0.45 T. glis X T. tana 0.00 0.00 0.00 0.00 0.00 22.38 0.00 30.53 0.49 T. glis X T. dorsalis 0.43 0.51 3.14 1.42 0.51 84.13 0.43 36.75 5.71 T. minor X T. gracilis 20.08 52.24 13.56 83.12 8.81 28.64 8.81 100.00 20.08 T. minor X T. tana 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.41 T. minor X T. dorsalis 3.39 3.39 3.39 3.39 3.39 3.39 3.39 3.39 28.89 T. gracilis X T. tana 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.04 T. gracilis X T. dorsalis 2.01 2.01 2.01 2.01 2.01 2.01 2.01 2.01 30.17 T. tana X T. dorsalis 39.02 15.21 2.64 0.65 61.62 94.29 43.09 61.62 0.52 D. melanura X D. murina 2.75 2.75 2.75 14.17 14.17 1.43 62.42 5.01 14.17 D. melanura X P. Lowii 0.09 0.09 0.09 0.09 0.09 0.09 0.09 7.79 0.09 D. murina X P. Lowii 0.23 0.23 0.23 0.23 0.23 0.23 0.23 4.88 0.23 Each value indicates the limit percentage in which the significant differences are confirmed between the two species. est in P. lowii in contrast to Dendrogale. Although the 2 Recognition of 8 clusters (Fig. 1) is advocated from the genera have been suggested to occupy similar ecological principal component analysis, and does not contradict the 5 positions [1, 7], morphological adaptations in the skull are groups separated by the osteometrical description. Clusters obviously different between the 2 genera. Some reports 3 and 4 consist of 6 medium-sized species. Some authors have actually suggested that Dendrogale murina are more considered T. longipes as a subspecies of T. glis, and T. mul- terrestrial than Ptilocercus lowii in behavioral tendency [1, leri as a subspecies of T. splendidula [4, 19]. These 2 14]. groups include both arbo-terrestrial and functionally non- 878 H. ENDO ET AL.

Fig. 1. Principal component plots of the 1st and 2nd transformed variables obtained from 17 skull mea- surements. A) Male, B) Female. Symbols indicate the species as follows: M, Tupaia montana, P, T. picta, S, T. splendidula, U, T. mulleri, L, T. longipes, G, T. glis, J, T. javanica, m, T. minor, g, T. graci- lis, D, T. dorsalis, T, T. tana, 1, Dendrogale melamura, 2, D. murina, and 3, Ptilocercus lowii. Hori- zontal axis, the first principal component. Vertical axis, the second principal component.

Table 6. Character loading factors observed from the principal component analysis PL CL SL ZW LBO GNB MPL LBS DL Male PC1 0.998 0.997 0.973 0.892 0.870 0.928 0.990 0.915 0.995 PC2 -0.038 -0.018 -0.210 0.394 0.061 0.056 -0.114 0.207 -0.040 Female PC1 0.998 0.996 0.976 0.876 0.871 0.902 0.993 0.880 0.993 PC2 -0.019 0.004 -0.195 0.436 -0.006 0.058 -0.100 0.294 -0.058

GPB GMB HAB LC LA LIA HR HM P.V.E.* Male PC1 0.861 0.913 0.944 0.997 0.991 0.991 0.820 0.864 96.3 PC2 0.377 0.284 0.035 0.042 0.090 -0.002 0.488 0.302 1.8 Female PC1 0.867 0.902 0.945 0.997 0.990 0.992 0.815 0.848 96.1 PC2 0.329 0.292 -0.034 0.030 0.065 -0.036 0.501 0.295 1.7 Percentages of the variation explained by each factor (*). specialized species. We suggest that adaptation for feeding (Table 1), T. minor and P. lowii are widely-distributed as on termites, ants and invertebrates on the ground may distin- arboreal species, although the latter is not present in Java guish T. glis and T. longipes from the Cluster 3 (Tables 3–5) Island. T. glis is one of the common small mammalian spe- [7]. It also has been suggested that T. javanica occupies the cies with arbo-terrestrial life in this region [1–3, 7, 18, 20, same niche as the species of Cluster 3. T. javanica is 21, 25]. These species represent the fauna of tree shrews in smaller in PL and body size, and it is separated from the Malayan Peninsula and Sumatra Island. T. glis and T. lon- other arbo-terrestrial species in our plots (Fig. 1). This may gipes also occupy this region including Borneo Island as be due to the dwarfism caused by isolation in Java Island. arbo-terrestrial species. Java Island is exceptionally poor in Cluster 6 consists of T. gracilis and T. minor, species liv- tree shrew diversity, and it has only two arbo-terrestrial spe- ing mainly on trees. Clusters 7 and 8 including Dendrogale cies, T. glis and T. javanica. and Ptilocercus also represent arboreal species. The separa- Borneo Island, especially the northern district in the tion among species is dependent on the differences in the island, is of paramount importance for the tree shrew fauna second principal component. We suggest that all arboreal [15–17]. The complete terrestrial species T. tana and T. species in Scandentia, including T. minor and T. gracilis, dorsalis are distributed in this Island [20, 25]. In addition, may require a small body size in which the PL is about 35 the arbo-terrestrial species such as T. montana and T. picta mm. have been recorded in the northern district, whereas T. As for the zoogeography from the Malayan Peninsula to splendidula and T. mulleri in the southern part of this Island. Sumatra, Java, Borneo and some small islands in this region Furthermore T. gracilis, D. melanura and P. lowii share the SKULL ADAPTATION IN TREE SHREWS 879 arboreal fauna of the tree shrews with T. minor in the north- shrew (Dendrogale murina). Ann. Anat. 181: 397–402. ern district [7, 21]. 9. Endo, H., Cuisin, J., Nadee, N., Nabhitabhata, J., Suyanto, A., D. murina has not been recorded in the region from the Kawamoto, Y., Nishida, T. and Yamada, J. 1999. Geographical Malayan Peninsula to Borneo Island but recorded in the variation of the skull morphology of the common tree shrew southeastern regions of the Indochinese Peninsula. How- (Tupaia glis). J. Vet. Med. Sci. 61: 1027–1031. 10. Endo, H., Nishiumi, I., Hayashi, Y., Rerkamnuaychoke, W., ever, we used D. murina in the statistical analysis to clarify Kawamoto, Y., Hirai, H., Kimura, J., Suyanto, A., Nabhitab- the adaptation strategies of Scandentia. The common spe- hata, J. and Yamada, J. 2000. Osteometrical skull character in cies in the northern region of the Isthmus of Kra consists of the four species of tree shrew. J. Vet. Med. Sci. 62: 517–520. T. belangeri. This species is distinguishable from T. glis in 11. Endo, H., Nishiumi, I., Hayashi, Y., Rashdi, A. B. M., Nadee, the southern districts by external skin color, principal com- N., Nabhitabhata, J., Kawamoto, Y., Kimura, J., Nishida, T. ponent plots, and karyological type [9, 11, 12, 18, 23, 24], and Yamada, J. 2000. Osteometrical analysis in skull of the although we think that the similar behavioral adaptation has common tree shrew from both sides of the Isthmus of Kra. J. been established in both T. belangeri and T. glis. Vet. Med. Sci. 62: 375–378. Since the sample size was small in T. picta, T. splendid- 12. Endo, H., Hayashi, Y., Rerkamnuaychoke, W., Nadee, N., ula, T. mulleri and T. dorsalis, we should measure more Nabhitabhata, J., Kawamoto, Y., Hirai, H., Kimura, J., Nishida, T. and Yamada, J. 2000. Sympatric distribution of the two mor- specimens in the future. phological types of the common tree shrew in Hat-Yai Districts (South Thailand). J. Vet. Med. Sci. 62: 759–761. ACKNOWLEDGEMENTS. We thank Dr. J. Cuisin of 13. Hill, J.E. 1960. The Robinson collection of Malaysian mam- Muséum National d’Histoire Naturelle (Paris, France), Dr. mals. Bull. Raffles Mus. 29: 6–22. J. Mead and Ms. L. Gordon and staffs of the Smithsonian 14. Kloss, C. B. 1916. On a collection from the coast and islands of Institution (Washington D.C., U.S.A.) for their kind help in southeast Siam. Proc. Zool. Soc. Lond. 1: 27–75. our examinations of museum specimens. We are also grate- 15. Kobayashi, Y. and Hotta, M. 1978. Biological expedition to the ful to the curatorial staff of the Raffles Museum of Biodiver- rain-forest of Sabah in 1976. Contrib. Biol. Lab. Kyoto Univ. sity Research, National University of Singapore. This study 25: 255–271. was financially supported by Grant-in-Aids for Scientific 16. Kobayashi, T., Maeda, K. and Harada, M. 1980. Studies on the small mammal fauna of Sabah, East Malaysia. I. Order Chi- Research nos. 13640705, 13575027 and 14405030 from the roptera and Genus Tupaia (Primates). Contrib. Biol. Lab. Ministry of Education, Science and Culture, Japan. Kyoto Univ. 26: 67–82. 17. Kobayashi, T., Harada, M., Maeda, K. and Matsumura, S. REFERENCES 1980. Small of Sabah. Acta Phytotaxonomica et Geobotanica 31: 91–97. 1. Boonsong, L. and McNeely, J. A. 1988. Mammals of Thailand, 18. Lyon, M. W. 1913. Treeshrews: an account of the mammalian 2nd ed., Saha Karn Bhaet, Co., Bangkok. family Tupaiidae. Proc. U. S. Nat. Mus. 45: 1–183. 2. Corbet, G. B. and Hill, J. E. 1992. The Mammals of the 19. Medway, L. 1961. The status of Tupaia splendidula Gray (Pri- Indomalayan Region: A Systematic Review. Oxford Univ. mates: Tupaiidae). Treubia 25: 269–272. Press, Oxford. 20. Medway, L. 1965. Mammals of Borneo. Malaysian Branch of 3. Davis, D. D. 1962. Mammals of the lowland rain-forest of the Royal Asiatic Society, Singapore. North Borneo. Bull. Singapore Nat. Mus. 31: 43–70. 21. Medway, L. 1978. The Wild Mammals of Malaya (Peninsular 4. Dene, H., Goodman, M. and Prychodko, W. 1978. An immu- Malaysia) and Singapore. 2nd ed., Oxford Univ. Press, Oxford. nological examination of the systematics of Tupaioidea. J. 22. Nowak, R. M. 1999. pp. 244–249. Walker’s Mammals of the Mammal. 59: 697–706. World, vol. 1. 6th ed. (Nowak, R. M. ed.), Johns Hopkins Univ. 5. Driesch, A. 1976. A Guide to the Measurement of Animal Press, London. Bones from Archaeological Sites. Harvard Univ., Cambridge. 23. Steele, D. G. 1983. Within-group variation in coat-color char- 6. Ellerman, J. R. and Morrison-Scott, T. C. S. 1951. Checklist of acteristics of the common tree shrew, Tupaia glis Diard, 1820. Palaearctic Region. Oxford Univ. Press, Oxford. Int. J. Primatol. 4: 185–200. 7. Emmons, L. H. 2000. Tupai: A Field Study of Bornean Tree- 24. Toder, R., von Holst D. and Sshempp, W. 1992. Comparative shrews. University of California Press, Berkeley. cytogenetic studies in tree shrews (Tupaia). Cytogenet. Cell 8. Endo, H., Rerkamnuaychoke, W., Kimura, J., Sasaki, M., Genet. 60: 55–59. Kurohmaru, M. and Yamada, J. 1999. Functional morphology 25. Yasuma, S. 1994. An Invitation to the Mammals of East Kali- of the locomotor system in the northern smooth-tailed tree mantan. Pusrehut Special Publication No.3, Singapore.