The authors wish to acknowledge that the values of the rate of spread for the grass fires in Fig. 2 (blue circles) were extracted the following reference: Cheney NP, Gould JS, Catchpole WR (1998) Prediction of fire spread in grasslands. International Journal of Wildland Fire 8, 1–13. doi:10.1071/WF9980001 Additionally, the authors wish to advise of errors present in the text and in Figs 2 and 3 of their paper. An error has recently been detected in Table 2 of Bilgili and Saglam (2003) (shrub fires represented with blue Nabla in Fig 2 and Fig 3). The value of the open wind speed were indicated in ‘km/h’ and not in ‘m/min’ as written in the text. As a consequence of this error, the following paragraph in the text is now invalid: ‘Beside the main cluster of points (Fig. 2), we also noticed that a set of experimental results (shrub, with blue triangles) (Bilgili and Saglam 2003) did not fit the general trend compared with the others. These experimental fires were conducted in maquis fuel, characterised by a quite high fuel load (the data reported in the paper indicate a fuel consumption ranging between 1.3 and 4.4 kgm2, between 3 and 10 times larger than the valuesmeasured for grass fires in Australia; Sullivan 2007).Moreover, these fires were carried out under very low wind conditions (ranging between 0.05 and 0.24ms
1); the consequence of these two factors (high fuel density and low wind speed) was a quite high value of Byram’s convective number (larger than 104) in comparison with the other experimental data. Therefore, under these conditions, it was not really surprising that the behavior of these fires differed from the other ones (Fig. 3).’ Amendments to Fig. 2 and 3 are as follows (in text):