The analytical resolution or response function (ARE) suggested by Dowsett et al. for describing measured secondary ion mass spectrometry (SIMS) depth profiles of delta doping distributions in solids was analysed with the aim of identifying the relevance and the physical meaning of the upslope length lambda(u) and the Gaussian broadening parameter sigma. It was found that it is difficult to determine the upslope length safely as Iona as lambda(u)/sigma < 0.3. For an accurate determination of lambda(u), it will usually be necessary to measure the profile of (ideal) delta markers over four orders of magnitude or more. Measured delta profiles with very sharp leading edges as well as delta profiles calculated on the basis of the diffusion approximation of atomic mixing were compared with the ARF. Irrespective of the true value of lambda(u), the peak height of the ARF was found to be too high by up to 12% and the width too small. The results suggest directions for improving the ARF. (C) 2002 Elsevier Science B.V. All rights reserved.