We demonstrate that the analytic calculation of the 1RSB break point parame
ter in a paper by de Oliveira and Fontanari (de Oliveira V M and Fontanari
J F 1998 J. Phys. A: Math. Gen, 32 2285) is erroneous, due to the omission
of a higher-order term in a lengthy perturbative calculation, and provide a
refinement of the accompanying numerical results.