In 2004 a newly discovered hominin fossil from the Indonesian island of Flores made headlines worldwide. Although an adult, it was tiny – about a metre tall, had a commensurately small brain (the size of a grapefruit), had made tools and hunted small elephants and giant rats. Dates from the cave floor sediments that had entombed it gave ages as young as 13 to 11 thousand years and as far back as 850 ka. So H. floresiensis was regarded as being the last human to share the Earth with us; that is, if it was a different species rather than a product of evolutionary shrinkage of anatomically modern humans stranded and isolated on the island for a very long time. Then there was talk among locals of the legendary Ebo Go-Go, with whom their ancestors had shared the island – they had arrived between 35 to 55 thousand years ago.
Unsurprisingly, a major controversy raged in palaeoanthropology circles, between those who demanded either island dwarfism or congenital deformity of modern humans, and the other camp focused on many anatomical differences that pointed to a bona fide companion to later immigrants who perhaps survived into modern times. The ‘Hobbit’ became a cause celebre, but many of the original protagonists are now left with the proverbial egg on their faces. The cave sediments turn out to have a much more complex stratigraphy than previously thought, following further excavations led by the original discoverer Thomas Sutikna of the Pusat Penelitian Arkeologi Nasional in Jakarta Indonesia (Sutikna, T. and 19 others 2016. Revised stratigraphy and chronology for Homo floresiensis at Liang Bua in Indonesia. Nature, v. 532, p. 366-369.
The delayed appearance of the revision is hardly surprising, given the lengthy political squabbles surrounding access to the site. And neither are the outcomes, for cave sediments are notoriously tricky because of their episodic reworking by cave floods and roof falls, together with the difficulty in finding materials suited to dating in tropical settings. The original charcoal used in radiocarbon dating and sand grains subject to the thermoluminescence method were in fact from a unit that lies unconformably against the stratum that hosted the fossils. More sophisticated luminescence dating of the actual fossil-hosting sediments yield ages between 100 to 60 ka, tool-bearing units range from 190 to 50 ka. The origins of H. floresiensis are thus pushed back beyond the date of supposed colonisation by H. sapiens, and remain an open question.