During 1891 and 1892 Eugène Dubois, searching for the missing link, found at Trinil (Java, Indonesia) a molar, which was ape-like, a skullcap, which had a cranial capacity between the apes and Homo sapiens, and a femur which was human-like. In 1894 he combined those fossils as Pithecanthropus erectus (now Homo erectus). Dubois excavated at Trinil for about 10 years, and besides these hominin fossils found a few thousand mammal and reptile fossils. In 1907 and 1908, Eleonora Selenka also excavated at Trinil and added a few more thousand fossils to the fauna.
At another Javan site, Kedung Brubus, Dubois found a mandible of his Pithecanthropus erectus. Here he excavated only a few hundred fossils. The whole collection from all the sites where Dubois excavated was sent to the Rijksmuseum van Natuurlijke Historie (now Naturalis Biodiversity Center), Leiden, The Netherlands, where it has been stored and curated since then. Sites with hominins are considered more special and important than sites without. For this reason, Trinil and Kedung Brubus have played a crucial role in the biostratigraphy of Java during the last century.
Dubois considered the faunas of the sites Trinil and Kedung Brubus as coeval, although he noticed a difference in faunal composition. This idea was adopted by researchers until the 1930s. Then the Dutch East Indies Geological Survey started a project to develop a biostratigraphy system for Java. One of their employers, the well-known Ralph von Koenigswald, came up with a scheme. Three faunas were distinguished by him for the Pleistocene of Java: Jetis (Early), Trinil (Middle), and Ngandong (Late Pleistocene) named after sites on Java.
When I started in the early 1980s as curator of the Dubois collection, I tried to understand the biostratigraphy of von Koenigswald. But his faunal list for Trinil was very long and I could not find the species in the Dubois or the Selenka collection. It turned out that von Koenigswald used locality names for his faunal successions, but the faunal assemblages differed from the fauna actually collected at those localities. According to von Koenigswald the ‘Jetis’ fauna was an assemblage of fossils from several localities, whereas the ‘Trinil’ fauna was an assemblage of fossils from Trinil, Kedung Brubus, and other sites. In this interpretation, the Kedung Brubus fauna was still coeval with Trinil.
However, checking the fauna from the type localities Trinil and Kedung Brubus in the Dubois collection showed that the Trinil fauna was poor in species, while the Kedung Brubus fauna had many more. In Trinil there was only one proboscidean, a Stegodon, while in Kedung Brubus there were two — a Stegodon and Elephas. Furthermore Kedung Brubus had a few more advanced species such as Rhinoceros kendennindicus, Epileptobos, and Tapirus, which were missing in Trinil. This bias was not caused by sampling error, because the Trinil collection contains a few thousand fossils and Kedung Brubus only a few hundred.
Based on the collections, the differences in the faunal composition of Trinil and Kedung Brubus are clear. As the sites are only a few kilometres from each other, an environmental explanation between the differences in fauna is not likely. So there must be a difference in age. Because of the absence of advanced species in Kedung Brubus, we concluded that Trinil is older.
Von Koenigswald used guide fossils for his units. In Jetis it was Leptobos (a bovid). However, this species was also present in Kedung Brubus. Based on this and other species which had Jetis and Kedung Brubus fauna in common, we concluded that Jetis was coeval with Kedung Brubus. This conclusion caused a debate about the absolute age (as a precise number of years) of the faunas and the Homo erectus specimens discussed here. Th ere is still a debate about the absolute ages of these sites, while the biostratigraphic conclusion of the sequence of these events — the Trinil specimens being older than Kedung Brubus — is still considered to be correct.
What can we learn from this? Always go back to the collection from the type locality, and do not believe faunal lists.
Vos, J de, S Sartono, S Hardja-Sasmita, and P Sondaar (1982). The fauna from Trinil, type locality of Homo erectus; a reinterpretation. Geologie en Mijnbouw 61, 207–211.