1 Introduction

This communication is a reply to Burgess et al. (BCT, 1997) who question the accuracy of the effective collision strength computed within the IRON Project (IP) by Galavís et al. (IP-X, 1995) for the electron impact excitation of the ${\rm 3s}\sp2{\rm 3p}\sp4\ \sp1$D $-\sp1$S transition in Ar III. The IP (Hummer et al. 1993; Butler 1996) is an international on-going collaboration whose goal is to compute collisional and radiative data for ions of astrophysical interest. Since the initial stages of the IP targeted isoelectronic sequences rather than single ionic systems, general practical approximations had to be adopted. The issue of accuracy consequently became a main concern because tricky individual features perhaps did not get the attention that was required. In their assessment of the IP data for the quadrupole transition in Ar III, BCT made use of a helpful technique developed by Burgess & Tully (BT, 1992) whereby the effective collision strength is scaled and mapped onto the finite reduced electron-temperature interval (0,1). In this particular case, they computed the limiting point at infinite temperature (i.e. reduced temperature $T_{\rm r}=1$), and they adjusted the scaling parameter in such a way that an electron temperature interval of 10⁵ K occupied 90% of the unitary scale. In their analysis, BCT suggested that the IP data could be in error by a factor of 2 due to the neglect of the contribution of high partial waves. Since the accuracy rating quoted in IP-X for the slightly ionised species is 20%, the magnitude of the alleged discrepancy indeed deserves some consideration.

We have rerun the IP-X calculation examining in detail possible sources of error. Also, the original energy range has been extended so as to study the high-temperature behaviour of the effective collision strength with respect to the limit estimated by BCT. The present report is organised as follows. We provide short summaries of the original IP calculation and the BT formalism in Sects. 2 and 3, respectively, followed in Sect. 4 by a presentation and analysis of the recalculated collisional data. Some conclusions are drawn in Sect. 5.