Geology and Discovery Record of the Trinil Pithecanthropus erectus Site, Java

The scientific utility of Eugène Dubois’ Pithecanthropus erectus (P.e.) specimens and other paleontological finds from Trinil has been impaired for over a century by questions about provenience. Firsthand accounts and contemporaneous field photographs, presented here, extensively document the site geology and discovery record, and enhance Trinil’s paleoanthropological value. The P.e. and numerous-other fossils originated from small excavations dug near the seasonal low-water level of the Solo River along its left bank. The discovery bed was thin bioclast-rich, conglomeratic, volcaniclastic sandstone of fluvial origin (main bonebed). It was overlain by about nine meters of indurated, generally poorly fossiliferous strata which held up the river’s incised embankment. The strata were approximately horizontal, and provenience was tracked by elevation with respect to river level. Dubois’ on-site supervisors specified that the P.e. Skullcap (Trinil 1) and Femur I (Trinil 3) came from a ~0.2m-thick subunit of the bonebed that was traced for ~12m from the 1891 Skullcap pit (~30m2) to the 1892 Femur-discovery excavation and then across an enlarged 1892-1893 trench (~170m2). The bonebed contained fossils referable to Trinil fauna species such as Axis lydekkeri, Duboisia santeng and Stegodon trigonocephalus. Trinil fauna fossils continued to be unearthed from near seasonal low water during 1895-1908, as excavations were expanded southward over about two-thousand-square meters. Field studies by the Selenka Expedition (1907-1908) and others (1930s to 1970s) confirmed central aspects of the site geology and paleontology. The bonebed has features indicative of an unusual origin. The biotic content is uniquely broad for a Homo erectusdeposit in Java. The bioclasts range from proboscidean craniums and logs to rat teeth and leaves. Terrestrial-vertebrate skeletal elements are overwhelmingly disarticulated, frequently broken, and little-abraded by fluvial transport. Their fossilization is uniform and surfaces finely preserved. Ungulate materials indicate a Minimum Number of Individuals >100. No evidence has been found of hominin- or terrestrial-carnivore involvement in death or accumulation. The bioclasts were matrix supported. Place-to-place and vertical changes in bioclast density occurred, but no substantial internal depositional hiatus was reported. Considering the paleogeographic context of the bonebed, these features implicate a catastrophic mortality of ungulates in a population aggregation along the Trinil paleo-river floodplain, followed by lahar-flood transport and deposition of gravel-size lithic- and biotic-materials. Trinil continues to play pivotal roles in the paleoanthropological record of southern Sundaland. The Trinil fauna is a lynch-pin in a long-lasting paleobiogeographic linkage between Homo erectus and certain lineages of large bovids, cervids, proboscideans rhinoceros, suids and tiger. The paleogeographic setting of the Trinil main bonebed exemplifies stratovolcanic watersheds that H. erectus occupied in eastern Java for ?0.8my, based on discoveries at Kedungbrubus, Ngandong, Mojokerto, Sangiran Dome and Trinil. Large-mammal faunas occupied a Pleistocene volcanic island (Patiayam) ~100km north of these sites, and occurred 400km to their west. Seismic data in the Java Sea north of Java image immense river and coastal systems of the Pleistocene Sunda Shelf where H. erectus- and large-mammal populations like those evident at Trinil could have lived.