REPLYING TO J. V. Ferraro et al. Nature 554, https://doi.org/10.1038/nature25165 (2018)

Contrary to our hypothesis1 that the Cerutti Mastodon (CM) site represents a 130,000-year-old archaeological site, in the accompanying Comment2 Ferraro et al. argue that the site formed through ‘common’ geological and taphonomic processes. As a source for the cobbles that we interpreted as hammerstones and anvils, they postulate a previously unrecognized alluvial fan, later removed by fluvial winnowing that somehow left our five cobbles, refitting flakes, and fragments of stone, mastodon bone and teeth in place. There is no sedimentological or geomorphic evidence of an alluvial fan, and their scenario leaves unexplained a number of taphonomic features, including the two discrete concentrations in which were found cobbles, refit stones and bones, impact-fractured bones, side-by-side femoral heads and a tusk oriented vertically.

Ferraro et al.2 also speculate that the stone and bone fractures that we analysed can be explained by post-burial processes such as sediment compaction or interaction with excavation equipment, whereas we contend that these features are part of the CM biostratinomic (pre-burial) record. Support for our view is provided by the fact that most CM bones and stones were enclosed within crusts of pedogenic carbonate that establish a ‘chain of evidence’ showing that breakage and positioning of objects occurred many thousands of years ago, and, as we contend, before burial3. The only pre-burial cause of bone breakage Ferraro et al.2 consider is trampling, which we have argued is incompatible with other site data1.

Ferraro et al.2 draw comparisons to the Inglewood Mammoth Site (IMS)4,5 and the Waco Mammoth National Monument (WMNM)6. For the IMS, they cite an observationally based study4 that proposes that excavating equipment caused the spiral fractures on many of the bones. However, this claim is compellingly refuted by an experimentally based study5 that shows that the IMS spiral fractures are ancient after all, and probably occurred before burial.

WMNM bones illustrated by Ferraro et al. (figure 1 of ref. 2) lack clear evidence of true spiral fractures or normal impact notches7, instead representing classic examples of dry bone fracture, with rough texture on fracture surfaces and contrasting coloration of broken versus cortical surfaces (figure 1b, d, g of ref. 2). The closest approach to a notch (shown in figure 1c, d of Ferraro et al.2) is a shallow, irregularly arcuate break—described as a pseudo-notch or micro-notch—that does not extend to the medullary portion of the bone, unlike the ‘normal notch’7,8 we illustrated1, which was defined by two clear inflection points, a negative flake scar, an attached cone flake and smoothly curved fracture surfaces that extend completely through the cortical portion of the bone. Only ‘normal notches’ are used to determine human agency7,8.

By overlooking the most important bone evidence, which includes impact features such as cone flakes, bulbs of percussion and a large impact notch with associated negative flake scar, as well as bone distribution patterns, bone refits and missing femoral diaphysis pieces, Ferraro et al.2 did not consider precisely those features that are individually and collectively most likely to have been caused by cultural processes. They have not offered a cogent alternative site formation hypothesis that accounts for all evidence presented.