Statistical Analyses of Soil Quality

the peak temperature in the sample was directly measured while the laser power was slowly increased. As soon as melting occurred in the hottest portion of the sample, the absorption increased drastically, leading to "runaway" melting and subsequent melting phenomena [figure 2B of (1)]. The melting temperatures we reported are the last temperatures of the solid before melting set in. We stated (1, p. 554) that melting temperatures were recorded at the onset and not during melting. Temperatures measured during melting were very much higher and the emission spectra were of poor quality, which was most likely due to rapid temperature fluctuations caused by convection in the molten sample. From figure 2B of our report, Heinz et al. inappropriately estimate temperature gradients in our sample based on the assumption that its outer portion did not convert to perovskite. No such statement was made in our report. We only stated that the sample surrounding the molten area was perovskite. More correctly: All of the sample surrounding the molten area had converted to perovskite. To avoid confusion, we would like to emphasize the statement in our report (1) that figure 2B represents conditions after a melting experiment, which are significantly different from those at the onset of melting. Heinz et al. discount the drastic differences in their three previous estimates of melting. Curves showing zero, positive, or negative slopes, respectively, are clearly in disagreement with each other (2, 3, 4). We fully agree with the statement of Heinz et al. that "A direct association of the peak temperature with a melting slope, without a characterization of the temperature distribution, is therefore unreliable." R. Boehler A. Zerr Max-Planck-Institut fur Chemie, Postfach 3060, 55020 Mainz, Germany