I have shown [Phys. Rev. B 38, 6901 (1988)] that weak dilution of n≥2 Ising antiferromagnets, with no applied external fields, generates random anisotropies. Let ε=φ-(γ-2β), where φ is the corresponding crossover exponent, and γ and β are the usual exponents. I have claimed, wrongly, that ε=0 in general. Aharony points this out in the preceding Comment. ε is evaluated by Monte Carlo simulations here for some values of ${\mathit{J}}_1$/${\mathit{J}}_2$. ε≠0 is indeed obtained, but ε turns out to be too small to make any significant numerical difference in the particular case that I have treated.

• Received 26 December 1990

DOI:https://doi.org/10.1103/PhysRevB.44.425