Atholl Anderson's comment (above) on the recent finds from Jerimalai draws attention to the dangers of over-interpreting the wider significance of marine resources present in Pleistocene coastal sites without careful evaluation of at least three variables: (1) the accurate identification of the species represented and hence of their behaviour and accessibility to capture; (2) the actual quantities and rates of accumulation of the marine food remains; and (3) local ecological and oceanographic conditions. To these I would add a fourth variable: the bathymetry and submerged topography of the marine environment adjacent to the sites in question—the physical structure of what one might call the 'offshore catchment'—and changes resulting from relative sea level variation (including eustatic and isostatic/tectonic effects).

In an important paper, O'Connor et al. (2011) described evidence for marine fishing from around 42 000-year-old (cal BP) deposits at Jerimalai Shelter on Timor-Leste. The paper's title referred to evidence for pelagic fishing and the maritime skills of anatomically modern humans (AMH). Considering that not long ago human seafaring and marine fishing were considered to be limited to the terminal Pleistocene or early Holocene (see Erlandson 2001), the paper's broader significance lies in the further evidence for Pleistocene voyaging required to colonise Timor-Leste and the quantities of fish bone that represent a substantial marine fishing effort at a relatively early date.

I thank the contributors for their comments. We all agree that the Jerimalai data provide an intriguing glimpse of maritime activity in the Wallacean Pleistocene by showing that Scombridae were unusually prominent amongst 15 families of fish that had been caught around 40 000 BP. In various ways, we also agree that the hypothesis interpreting these data in O'Connor et al. (2011) is far from robust. Its fundamental proposition was that the scombrids were oceanic tuna, but O'Connor and Ono (above) now concede that, 'as we did not speciate [sic] the scombrids at Jerimalai we cannot be certain that oceanic species were represented'. Their description of samples and methods, which confirms the inadequacy of the comparative collection and notes that scombrid bone from Jerimalai Square A also may not be from tuna, simply underlines the thrust of my critique. Some of the scombrids might be oceanic tuna but since no tuna were identified, either as a class or species, the marine zones involved in scombrid fishing cannot be inferred, and the empirical argument for offshore tuna fishing collapses.

In 1959, at a meeting reviewing the 'archaeological achievements of the past 10 years' in celebration of the tenth anniversary of the 'New China (1949-)', the leading archaeologist Yin Da (1906-1983), then director of the Institute of Archaeology of the Chinese Academy of Sciences (CAS), urged all the archaeologists in China "to cooperate fully, so that in the next three or five years, in the entire nation, we can build up a scientific and holistic system out of all cultural remains of all periods; that is to say, to build up a Marxist Chinese archaeological system" (Yin 1959: 123).

This call had two keywords in it.One was 'Chinese'. Ever since the early twentieth century, growing nationalism had drum-beaten Chinese archaeologists to search for Chinese cultural origins (Liu & Chen 2001: 317). A particularly urgent matter for archaeologists of the 1950s was to dispel the notion of 'the western origin of Chinese culture' that was current among foreign and native archaeologists during theNationalist Era (1911-1949). To achieve this goal, it was imperative to undertake archaeological investigation systematically so as to prove the autochthonous origins and undisrupted development of Chinese civilisations. The second word, 'Marxist', reflects a process of cutting the umbilical cord of the reborn archaeology of the 'New China' from the 'bourgeois archaeology' of the 'Old China' and swaddling the discipline with the mantle of Marxist theories and models