The Sambungmacan 3 specimen (Sm 3) was first excavated in 1977 near the village of Poloyo in Central Java, Indonesia, and later resurfaced in New York City before being returned to Indonesian authorities (Boedhihartono, 1998; Márquez et al., 2001). The discovery of a new hominid fossil from Java is always a boon for scientists, but this specimen is particularly important for the elucidation of later human evolution in Southeast Asia. The interest in this specimen stems from its overall similarity to the Ngandong fossils (Mowbray et al., 2000; Márquez et al., 2001, Delson et al., 2001), a sample many scientists have proposed to be intermediate between earlier Javan Homo erectus and later Homo sapiens from Australia (Thorne and Wolpoff, 1981; Wolpoff et al., 1984; Wolpoff, 1989, 1992, 1999). This notion has been questioned by other scientists, who instead perceive the Asian record as one of stasis, with no evidence of a morphological transformation in the direction of modern humans (Rightmire 1981, 1986, 1990, 1996; Stringer, 1989, 1992, 1994).
In their examination of the Sm 3 specimen, Delson and colleagues (2001) note that the Ngandong fossils, including those from Sambungmacan, possess a very distinctive morphology of the mandibular fossa. In the Ngandong specimens the squamotympanic fissure (called the “Glaserian fissure” by Weidenreich (1951)) lies entirely along the roof of the fossa, and the fossa itself is deep but anteroposteriorly narrow. In their comparison of the Ngandong and Sambungmacan morphology to a series of other Asian and African H. erectus specimens, Delson and colleagues (2001) claimed that several other specimens from Indonesia and Africa reflect a similar pattern in the mandibular fossa. A closer look at the specimens involved in this comparison would be helpful in determining the degree to which these specimens share this unique morphology. The purpose of this brief review is to examine their findings.
Most of the fossils in question, including Ngandong 1, 6, 7, 9, 10, 11, and 12, Sambungmacan 1, and Sangiran 17, were examined by the author during a visit to Indonesia in July, 1999. The remaining specimens were viewed as casts in the labs of either the University of Tennessee or Northern Illinois University, except for Ngawi, which was studied at Balai Arkeologi in Bandung, West Java, Indonesia.
Sangiran 4 was the first specimen noted as having a fossa similar to that of the Ngandong specimens (Delson et al., 2001). Indeed, Weidenreich (1951: 274) states that “[i]t is of greatest interest that the mandibular fossa of Pithecanthropus robustus [Sangiran 4] generally has exactly the same form as that of Solo man” (italics mine). He later states that “[t]here are, however, some differences. The tubercle is shorter and higher in Pithecanthropus. A postglenoid process separates the most lateral part of the fossa from the vestibulum of the ear entrance, but the tympanic plate is almost perpendicular and forms the main part of the posterior wall of the fossa” (Weidenreich, 1951: 274). Delson et al. (2001: 390) concur with this diagnosis, stating that “in Sangiran 4 the squamotympanic fissure courses mediolaterally and coincides with the deepest portion of the mandibular fossa.” However, the mandibular fossa in Sangiran 4 does not correspond to the Ngandong hominids in the course of this fissure along its entire length. The squamotympanic fissure is the separation between the mandibular fossa, formed by the temporal squama, and the tympanic portion of the temporal bone. According to Weidenreich (1951), the postglenoid tubercle of Sangiran 4 separates the mandibular fossa of that specimen from the vestibulum of the ear at the lateral end of the fossa. Sangiran 4 is unlike the Ngandong specimens in that it has a large postglenoid tubercle (Weidenreich, 1951; Durband, 1997). The squamotympanic fissure in the Sangiran 4 fossil curves posteriorly to this process, and is thereby removed from the deepest portion of the mandibular fossa at its lateral aspect. While the more medial course of the fissure is indeed in the roof of the fossa, as the fissure moves laterally it is no longer coincident with the deepest portion of the fossa. In fact, in the lateral view of the fossa the fissure can be clearly seen posterior to the postglenoid tubercle and clearly posterior to the roof of the mandibular fossa. This is illustrated very well in Fig. 1, which is borrowed and modified from Picq (1983; as cited in Wood, 1991). In this figure it is clear that while the deepest portion of the articular area of the mandibular fossa always lies anterior to the squamotympanic fissure, only in Ngandong does the roof of the fossa continue to slope superiorly to the fissure, thus containing the entire course of the fissure along its deepest extent. In short, Sangiran 4 does not exhibit the Ngandong morphology in this area.
The same is true for the Sangiran 17 specimen, also noted by Delson and coauthors (2001) as showing a morphological similarity to the Ngandong and Sambungmacan specimens. In this fossil, as in Sangiran 4, an eroded but clearly present postglenoid tubercle lies at the posterior end of the mandibular joint surface and anterior to the squamotympanic fissure. While once again the course of this fissure corresponds to the deepest portion of the fossa at its most medial aspect, the postglenoid tubercle lies anterior to this fissure and diverts it from the roof of the fossa at it moves laterally (Durband, 1997; see also Fig. 1). This is particularly important for discussions concerning the morphological affinities of Indonesian H. erectus. Jacob (1975, 1981) in particular has persisted in grouping Sangiran 17 with the Ngandong and Sambungmacan 1 cranium in Pithecanthropus soloensis due to these perceived similarities in form. In fact, Sangiran 17 is excluded from any grouping with the Ngandong hominids on the basis of several features, including the course of the squamotympanic fissure currently under consideration (Durband, 1997, 1998; Durband and Kidder, 2000).
The WT 15000 specimen provides another interesting case for comparison with the Ngandong fossils. Delson and colleagues (2001: 380) note that this specimen has “a flat and long fossa, but the squamotympanic fissure courses in its deepest portion.” Admittedly, this specimen's status as a juvenile does somewhat confound interpretation of its morphology. Even in this fossil, however, the course of the squamotympanic fissure does not run in the roof of the fossa for its entire length. In the very complete right mandibular fossa of this individual the course of this fissure is clearly posterior to the joint surface and does not traverse the roof of the fossa (see Fig. 6.31 in Walker and Leakey, 1993). It lies posterior to the postglenoid tubercle, and is again visible in a lateral view of this bone as lying quite posterior to the roof of the mandibular fossa.
Finally, OH 9 had also been considered as being very similar to the Ngandong hominids in the area of the mandibular fossa (Delson et al., 2001). The morphology does indeed appear broadly similar to the latter group, but once again differs in the course of the fissure. Like Sangiran 4, OH 9 possesses a very large postglenoid tubercle that clearly shunts the course of the squamotympanic fissure to the posterior of the curved temporomandibular joint surface. While the fissure does indeed run along the deepest portion of the fossa medially, as it does in many other specimens, the fissure moves posteriorly as it runs further laterally. Once again, this morphology is visible in the lateral view of the fossa, with the squamotympanic fissure clearly posterior to the postglenoid tubercle and not in the deepest portion of the mandibular fossa. As with Sangiran 4 and 17, this is also clear in Fig. 1.
DISCUSSION AND CONCLUSIONS
As Delson and colleagues (2001) have shown, the Sm 3 specimen shares the unique morphology of the mandibular fossa, among other traits, with the Ngandong specimens. The striking overall similarity of the Sm 3 specimen to Sambungmacan 1 and the rest of the Ngandong sample provides an opportunity to examine the presence or absence of some of these proposed Ngandong autapomorphies in a specimen outside of the Ngandong sample itself. However, the mandibular fossae in the Sm 3 calvarium, as well as the other Ngandong crania, were claimed to be similar to several other crania included in that particular study. As this brief examination has shown, this is clearly not the case. The Ngandong mandibular fossa is characterized by a squamotympanic fissure that lies along the very roof of the fossa for its entire length. In addition, we see little evidence of true postglenoid tubercles in the Ngandong specimens. Any postglenoid structures in these hominids are vestigial at best and appear as low ridges, while true tubercles are present in virtually every other specimen of H. erectus found to date. Each of the specimens noted by Delson et al. (2001) as being similar to the Ngandong fossils in this area possesses a large postglenoid tubercle, and the course of the squamotympanic fissure moves posterior to that structure and out of the deepest portion of the fossa. While the fissure may run along the deepest portion of the fossa at its most medial aspect in these other specimens, this course does not continue to the lateral edge of the fossa as it does in the Ngandong people.
So why is one seemingly insignificant fissure in the mandibular fossa worth this much trouble? The illustration of these differences serves to clarify the patterns present in the H. erectus sample from Africa and Asia. While most characteristics identified as potentially separating African and Asian members of H. erectus have been found to be variable in both populations (Kennedy, 1991; Bräuer and Mbua, 1992; Durband and Kidder, 2000), there are features present in the Ngandong sample (including the crania from Sambungmacan and Ngawi) that appear to be autapomorphic for this group (Weidenreich, 1951; Durband 1997, 1998; Durband and Kidder, 2000). This would have profound implications for the course of later human evolution in Southeast Asia, as well as in Australia (Durband, 1997). Therefore, it is important to clarify the morphologies present in these fossil samples. If, after continued study, these proposed autapomorphies are found to be restricted to the group represented by Ngandong, Ngawi, and the Sambungmacan specimens, the role of these fossils in the evolution of more modern humans in the region would be called into question. Therefore, a debate on these seemingly esoteric but potentially unique features found in the Javan fossils is a welcome and necessary one.
I thank Fachroel Aziz and Teuku Jacob for allowing me to study original fossil specimens in their care. Tony Djubiantono also allowed me to examine casts in his laboratory, and he and Dr. Aziz made my stay in Indonesia a very enjoyable one. I also thank Andrew Kramer for his comments on a draft of this manuscript. Any errors or omissions that remain are, of course, my own.