Elephas Celebensis (Hooijer) from the Pleistocene of Java

In a recent monograph, Maglio (1973) states that Stegodon hypsilophus Hooijer (1954a) must be considered a synonym of Elephas celebensis (Hooijer, 1949). Though Elephas celebensis is smaller-toothed and more hypsodont than the "hypsilophous stegodont" of Java (width of M3 42-52 mm against 62-64 mm; height-width index 106-116 against 87-89: Hooijer, 1954a, b; Maglio, 1973: 46, table 21), Maglio (1973: 47) considers the differences not great when the degree of variation of known samples (in other species) is allowed for. Over the years, I studied molars of pygmy proboscideans from various Indonesian islands, and it never occurred to me that the Java pygmy would be same as that from Celebes until this suggestion was put forth by Dr. Maglio. I had described the Java form as a progressive Stegodon although I did notice the elephantid build of the molar ridges: "The structure of the ridges, with two large conelets at either end and a number of smaller conelets in between that form a transversely convex edge, is essentially elephantine" (Hooijer, 1954a: 100). However, a molar was cut in mid-sagittal plane, and in that section, which I illustrated, the valleys are closed in at the bottom, Yshaped, as in a stegodont. Maglio (1973: 47) observed that this is because of intermediate columns in the middle of the crown, confluent with the ridges. A cut laterally of the median line would have produced a valley section open to the bottom, V-shaped, as in an elephantid. Re-examination of the specimen (in the Geological Museum at Leiden) showed that the type of valley reproduced was indeed an artifact of sectioning. The valleys are Y-shaped only in the middle. In a stegodont the molar valleys are Y-shaped in longitudinal section all over the width of the crown, producing enamel