Technical Comments

Comment on “The earliest modern humans outside Africa”

See allHide authors and affiliations

Science  26 Oct 2018:
Vol. 362, Issue 6413, eaat6598
DOI: 10.1126/science.aat6598


  • Fig. 1 Isotope evolution diagram showing U-Th activity ratios for crust samples Maxilla-1 through -8 calculated from data reported in (1).

    Straight lines are isochrons labeled in ky at their upper ends. Curved heavy lines show closed-system evolution with time for selected initial 234U/238U values. Uncorrected (green error ellipses) and detritus-corrected (red crosses) compositions of crusts are shown. Errors for detritus-corrected compositions are not shown for clarity but would be much larger than the ellipses shown for measured ratios. Thick arrow indicates expected displacement produced by loss of uranium (or gain of 230Th). The region containing Maxilla-6, to the right of and below closely spaced isochrons defining the infinite-age line, cannot be reached by evolution in closed U-Th systems. Thin solid arrows that connect uncorrected and detritus-corrected compositions for crust samples project away from the model detritus (red star). Detritus-corrected compositions are those reported in (1), except for that of Maxilla-2, which was recalculated from uncorrected ratios in (1). Dashed arrow shows direction of accurate detritus correction for crust Maxilla-6. Laser-ablation data for dentine (open brown ellipses between 50- and 100-ky isochrons) are from table S1 of (1).

  • Fig. 2 Projection of a three-dimensional 230Th-232Th-234U-238U isochron showing activity ratios for all maxilla crusts analyzed in (1) except Maxilla-6.

    The isochron regression (dash-dot line) is controlled mainly by data for Maxilla-7 and -8, the least contaminated crusts, and is consistent with the assumed model detritus composition (red star). More contaminated crusts Maxilla-2 through -5 have 232Th/238U ratios approaching those of the model detritus. MSWD = mean square weighted deviates, a measure of the observed scatter relative to that expected from analytical errors alone.

Navigate This Article