SANGIRAN 483
tektite from a higher Trinil faunal level at Sangiran, and
with the fairly generally held notion that deposits close
to the Grenzbank date from around 1 Ma, even if this
does stretch the Pucangan time scale somewhat. How-
ever, it has also been argued that this part of Java may
not yet have emerged from the sea in the very early
Pleistocene (H. Widianto, personal communication).
Furthermore, Larick et al. (2001) recently reviewed the
evidence for the Swisher et al. dating and concluded
that the hornblende pumice samples tested may have
been of dubious provenance and/or history. Their own
Ar/Ar datings in the less-deformed southeast quadrant
of the Sangiran dome ranged from 1.5 to 1.0 Ma,
bracketing almost all the hominids within this time
frame. Total agreement on the age of the Sangiran de-
posits thus still is lacking. For the time being, however,
the Larick et al. framework seems set to become widely
accepted, and Carl Swisher now apparently has con-
firming dates for the age of the Bapang (Bower, 2001).
ARCHAEOLOGICAL CONTEXT
Most Sangiran hominid finds are without archaeolog-
ical association, although some crude flake artifacts
have been reported from some Kabuh localities (von
Koenigswald and Ghosh, 1973).PREVIOUS DESCRIPTIONS AND ANALYSES
The study of the Sangiran hominids has been almost as
chaotic as that of the geology. Even the Trinil-like San-
giran 2 calvaria, reasonably assigned to Pitbecantbropus
by von Koenigswald and Weidenreich (1939), was sep-
arated into Homo wadjakensis by Dubois (1940). Wei-
denreich (1945) assigned the Sangiran 4 partial calvaria
and maxilla to the new species Pitbecanthopus robustus,
and the Sangiran 5 hemimandible to P dubius, while
retaining the partial mandible Sangiran lb in P erectus.
Later, von Koenigswald (1950) transferred Sangiran 4
to H. erectus modjokertensis; but, following the lead of
Mayr (1950), it rapidly became conventional to refer
almost all of the Sangiran hominids to Homo erectus,
usually without benefit of subspecies-although
Sartono still favored the appellation of Pitbecantbropus
erectus in his initial description of Sangiran 17 (Sartono,
1971), and as late as 1982 Jacob argued for grouping
the Java hominids into three species of Pitbecanthopus.
The major exception to this lumping trend has
been a few “hyperrobust” specimens epitomized by the
Sangiran 6 mandibular fragment. In 1945 Weidenreich,
using a name informally proposed by von Koenigswald,described this fossil as the holotype of the new genus
and species Megantbropus palaeojavanicus, a taxon into
which a few other pieces (the Sangiran 8 partial
mandible and the Sangiran 27 and 31 crushed partial
crania) eventually found their way. Robinson (1954)
later changed the species allocation of Sangiran 6 to
Parantbropus, while Tobias and von Koenigswald
(1964) compared this specimen to the Olduvai 7 Homo
babilis holotype. Lovejoy (1970) was unable to make a
clear distinction between the Megantbrops mandibular
specimens and Sangiran Homo erectus; but von Koe-
nigswald (1973) continued to insist that the former was
a separate, and terminal, form. Wolpoff (1992) has reit-
erated the view that all Sangiran hominids lie within a
regional lineage that ultimately gave rise to modern
Australasians, and would refer all to Homo sapiens. Tat-
tersall (1997) articulated an increasing awareness that
classical Homo erectus is a terminal east Asian develop-
ment, and Schwartz and Tattersall (2000) have adduced
additional detailed resemblances of the classic Sangiran
crania to the type calotte from Trinil. Holloway (2000)
gives the following cranial capacities (using a different
specimen numbering system): Sangiran 2: 815 ml;
Sangiran 4: 900 ml; Sangiran 10: 855m1; Sangiran 12:
1004 ml; Sangiran 17: 1059 ml.MORPHOLOGY
Large numbers of mostly fragmentary hominid fossils
have been recovered from sites in the Sangiran
Dome. We describe below the major specimens that
we have been able to access, grouped according to
element, and with references to other specimens as
appropriate.Crania
Sangiran 17 was the most complete cranial specimen
available to us for study and is described separately, as is
the crushed cranium Sangiran 27. In comparable parts,
Sangiran 2 and 4 represent extremes of gracility (S2)
and robusticity (S4) in the sample. These important
specimens are also described individually, as are many
other of the accessible fossils from the large Sangiran
collection. Some other specimens that appear to fall be-
tween S2 and S4 in degree of gracility or robusticity
are noted comparatively in the descriptions of other fos-
sils. In general, there is, within the cranial sample, a
good deal of variability in the degree of cranial-vault
bone thickness, extension of the mastoid process, exca-
vation of the mastoid notch, swelling of the arcuate