Attention is drawn to the fact that the approximation formula of Bédard, Chang, and Mandel for photoelectric counting yields different probabilities for different spectral distributions, and not just one function as implied in a recent paper of Lachs.