Reply to Evans: Use of poison remains the most parsimonious explanation for Border Cave castor bean extract

Evans (1) challenges our interpretation (2) that the presence of ricinoleic and ricinelaidic acids on a notched stick from Border Cave reflects the use of poison 24,000 B.P. We find his alternative hypotheses unlikely and reject his critiques of our methodology. He acknowledges the similarity of the notched stick to San poison applicators but ignores this similarity in his later arguments. Had the castor bean extract been found on another object, a stronger case may have been made for a function other than poison; the presence of it on a stick, virtually identical to those used by San for applying poison, cannot be dismissed.

Use of substances by recent production-based societies, such as the Zulus, should not be used to infer how the same substances were used by hunter-gatherers 24,000 y ago. Historical San (3, 4) obtained poison from numerous species, indicating that past choices may have depended on local traditions and availability.

Pounding castor beans produces a pulp containing the fatty acids present in castor oil, including ricinoleic and ricinelaidic acids, as well as the highly poisonous ricin protein. Without techniques to separate ricin from castor oil, which were developed in late prehistory (5), a certain amount of ricin will remain in substances derived from castor bean pulp. This fact challenges Evans’ interpretation (1) that the extract may have been used for food or medicinal purposes.

Evans (1) criticizes the absence of spectra for the two sampled residues but confuses spectra and chromatograms. Spectra of ricinoleic and ricinelaidic acids are documented in the literature (6), so they need not be reproduced. They do not provide information by which to evaluate relative concentrations. Mass spectra identified on the stick as ricinoleic and ricinelaidic acids are shown in Fig. 1. The chromatogram obtained from the residue from the object’s tip is shown in figure 2 (no. 27) of ref. 2. The undiagnostic nature and small amount of alcohols and other identified fatty acids listed in table S7 of ref. 2 do not allow the more in-depth interpretation for which Evans (1) argues. The chromatogram of the second sample was not shown because the peak intensities were just above the detection limit for fatty acids. Proteins degrade fairly quickly and become virtually impossible to identify when, as in our case, sample size prevents characterization by proteinomics.

Fig. 1.

Mass spectra of ricinoleic acid (A) and ricinelaidic acid (B) as their trimethylsilyl derivatives, identified in the gas chromatogram of the sample from the end of the notched wooden stick from Border Cave. O, oxigen; OTMS, trimethylsilylether.

Contrary to Evans’ claim (1), evidence that the compound was heated is not in contradiction to the interpretation that it was a poison. The temperature needed to obtain an isomerization of unsaturated fatty acids is much lower than that necessary for the denaturation of proteins; the compound may have been submitted to moderate heating while still remaining toxic. Heating of the compound most probably occurred after discard of the object, as a consequence of close proximity to hearths during later site occupations. In fact, many remains from the site show evidence of burning.

We maintain that our interpretation of the castor bean extract identified on the notched stick from Border Cave as the oldest evidence for the use of poison is the most parsimonious.