Darwinian argument and Neanderthal discoveries persuaded later Victorian science of humanity's animal origins. By the early 1900s, Eugene Dubois's recovery of Java man - Pithecanthropus erectus , the "missing link" - had further secured that connection. While Neanderthals were viewed as primitive humans, Pithecanthropus was, as his ape-man name indicated, definitely subhuman and our link with the animals. Dubois's discovery also established for many the primacy of Asia in human evolution. That position seemed confirmed when, from the 1920s onwards, human fossils were recovered from the Longgushan (Dragon Bone Hill) cave system at Zhoukoudian, near Beijing. Until excavation was disrupted by the Sino-Japanese War, the increasing record of Sinanthropus , or Peking man, provided the main evidence for pre-Neanderthal human evolution.

Despite undoubted similarities between Pithecanthropus and Sinanthropus (reflected in the later incorporation of both within Homo erectus ), the primitive nature of the former continued to be emphasised, whereas the humanity of the Peking material was stressed from the outset - with hunting, fire, cooking and family relationships all attributed to the Longgushan hominids. Their discovery close to the heart of an ancient civilisation, compared with Java man's domicile in a lesser colonial backwater, just might have had something to do with that.

Noel Boaz and Russell Ciochon review the history of the Longgushan finds and assess their significance for current human evolutionary studies. The earlier part of the story and the loss of the original fossils in late 1941 is well known, having been summarised by Ruth Moore half a century ago in her admirable Man, Time and Fossils . Boaz and Ciochon take the story on, describing the postwar excavations that yielded only a few more hominid fossils but added greatly to our knowledge of the cave system and of its non-hominid contents and context. This, combined with palaeoanthropology's broader perspectives and more cautious approach compared with 70 years ago, results in a markedly different evaluation of the material.

Longgushan was a hyena den, not a hominid living site, and H. erectus was prey, not predator. Evidence for fire is equivocal - "hearths" have been shown to be water-laid deposits, and other supposed indicators of human activity more likely the results of natural taphonomic processes.

ÁñÁ«ÊÓƵ

Nonetheless, burnt bone indicates the presence of fire, while undoubted stone tools and cut marks on animal bones attest to hominid agency. But the cut marks often overlie hyena tooth marks, pointing to secondary scavenging not primary hunting, and there is no evidence of cut marks underlying tooth marks. Some erectus crania display cut marks as well as hyena damage, pointing to cannibalism.

Hominids appear to have occasionally ventured into the Zhoukoudian cave to exploit the carcasses of hyena victims. Such behaviour may have been the mark of desperation, for the hyenas were of a lion-sized and particularly nasty variety, and it is unsurprising that H. erectus involuntarily contributed to the prey accumulation.

ÁñÁ«ÊÓƵ

Hominids were in the vicinity from about 670,000 years ago down to about 400,000 years ago or possibly later. Their presence was not continuous, being confined to warm interglacial phases and even then highly episodic.

Boaz and Ciochon provide a plausible "clinal replacement" model of Pleistocene human evolution involving partial substitution of local populations, so contrasting with "out of Africa's" total replacement and multiregionalism's emphasis on local continuity.

They view H. erectus as a basically tropically adapted species, driven by climatic forces, especially greater aridity, beyond Africa, and by Middle Pleistocene times with a fragile, tenuous presence at Longgushan, which must have been close to its geographical limit, even during optimal climes.

Before 500,000 years ago, conditions were relatively balmy with warm summer monsoons, but after that colder conditions intruded, peaking before 400,000 years in cold, dry, glacial climates and the mainland disappearance of H.erectus , replaced by the more advanced H. heidelbergensis , which differentiated from erectus populations further west.

ÁñÁ«ÊÓƵ

The book's breezy, informal style makes it highly accessible. The illustrations include informative three-dimensional reconstructions of the cave system and the location of finds. Other illustrations are adequate but hardly more than that; the half-tones in particular could be of better quality. The account effectively integrates the Longgushan material into the broader evolutionary picture with, for example, a well balanced and thoughtful section on the relative roles of fossil and molecular data in unravelling human phylogeny.

All in all, as well as updating readers on Zhoukoudian and its significance, Dragon Bone Hill provides a fair and painless overview of much current palaeoanthropology.

Alan Bilsborough is professor of anthropology and pro vice-chancellor, Durham University.