Cranial morphology of Javanese Homo erectus: New evidence for continuous evolution, specialization, and terminal extinction Yousuke Kaifu a, b, * , Fachroel Aziz c , Etty Indriati d , Teuku Jacob d , Iwan Kurniawan c , Hisao Baba a a Department of Anthropology, National Museum of Nature and Science, 3-23-1 Hyakunincho, Shinjuku-ku, Tokyo 169-0073, Japan b Department of Biological Science, The University of Tokyo, 6-3-1 Hongo, Bunkyo-ku, Tokyo 113-0033, Japan c Geological Museum, Geological Survey Institute, Jl. Diponegoro No. 57, Bandung 40122, Indonesia d Laboratory of Bio- and Paleoanthropology, Gadjah Mada University, Yogyakarta 55281, Indonesia article info Article history: Received 30 October 2006 Accepted 8 January 2008 Keywords: Sangiran Sambungmacan Ngandong Autapomorphy abstract Our current knowledge of the evolution of Homo during the early to middle Pleistocene is far from complete. This is not only because of the small number of fossil samples available, but also due to the scarcity of standardized datasets which are reliable in terms of landmark identification, interobserver error, and other distorting factors. This study aims to accurately describe the cranial morphological changes of H. erectus in Java using a standardized set of measurements taken by the authors from 18 adult crania from Sangiran, Trinil, Sambungmacan, and Ngandong. The identification of some obscure landmarks was aided by the use of micro-CT imaging. While recent studies tend to emphasize evolu- tionary conservatism in Javanese H. erectus, our results reinforce the theory that chronologically later groups experienced distinct morphological changes in a number of cranial traits. Some of these changes, particularly those related to brain size expansion, are similar to those observed for the genus Homo as a whole, whereas others are apparently unique specializations restricted to Javanese H. erectus. Such morphological specializations in Java include previously undescribed anteroposterior lengthening of the midcranial base and an anterior shift of the posterior temporal muscle, which might have influenced the morphology of the angular torus and supramastoid sulcus. Analyses of morphological variation indicate that the three crania from Sambungmacan variously fill the morphological gap between the chrono- logically earlier (Bapang-AG, Bapang Formation above the Grenzbank zone in Sangiran) and later (Ngandong) morphotypes of Java. At least one of the Bapang-AG crania, Sangiran 17, also exhibits a few characteristics which potentially indicate evolution toward the Ngandong condition. These strongly suggest the continuous, gradual morphological evolution of Javanese H. erectus from the Bapang-AG to Ngandong periods. The development of some unique features in later Javanese H. erectus supports the hypothesis that this Javanese lineage went extinct without making significant contributions to the ancestry of modern humans. Ó 2008 Elsevier Ltd. All rights reserved. Introduction In the past two decades, debates over the variation of Homo erectus have been centered on the question of whether or not morphological differences between African and Asian forms of this species (H. erectus sensu lato) are significant enough to recognize two distinct species; namely, H. ergaster and H. erectus (Andrews, 1984; Stringer, 1984; Wood, 1984, 1994; Groves, 1989; Rightmire, 1990, 1998; Bra ¨uer and Mbua, 1992; Kramer, 1993; Bra ¨uer, 1994; Clarke, 1994; Asfaw et al., 2002; Manzi, 2004; Villmoare, 2005; Terhune et al., 2007). The ‘‘Asian’’ vs. ‘‘African’’ morphological dichotomy has been investigated intensively in order to approach this question. In this context, Anto ´ n (2002a) correctly emphasized the need to look at regional and temporal variations within Asian H. erectus. This is reasonable because we can understand overarching themes in hominid evolution only by accumulating and analyzing valid, detailed information from each local area. In this study, we focus on H. erectus from Java and investigate chronological changes in its cranial morphology. Fossil remains of Javanese H. erectus are sampled from several different sites with varying dates. While the Trinil and Sangiran specimens belong to the Early Pleistocene, Ngandong possibly belongs to the late Middle or Late Pleistocene. Dates of the Sambungmacan specimens are unclear, but are generally believed to be contemporaneous with, or older than, Ngandong. In spite of * Corresponding author. E-mail addresses: kaifu@kahaku.go.jp (Y. Kaifu), azizf@bdg.centrin.net.id (F. Aziz), ettykurtz@idola.net.id (E. Indriati), baba@kahaku.go.jp (H. Baba). Contents lists available at ScienceDirect Journal of Human Evolution journal homepage: www.elsevier.com/locate/jhevol 0047-2484/$ – see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.jhevol.2008.05.002 Journal of Human Evolution 55 (2008) 551–580