The high-temperature susceptibility of the q-state Potts model behaves as G
amma\T- T-c\(-gamma) as T --> T-c + while for T --> T-c - one may define bo
th longitudinal and transverse susceptibilities, with the same power law bu
t different amplitudes Gamma(L) and Gamma(T). We extend a previous analytic
calculation of the universal ratio Gamma/Gamma(L) in two dimensions to the
low-temperature ratio Gamma(T)/Gamma(L), and test both predictions with Mo
nte Carlo simulations for q = 3 and 4. The data for q = 4 are inconclusive
owing to large corrections to scaling, while for q = 3 they appear consiste
nt with the prediction for Gamma(T)/Gamma(L), but not with that for Gamma/G
amma(L). A simple extrapolation of our analytic results to q --> 1 indicate
s a similar discrepancy with the corresponding measured quantities in perco
lation. We point out that stronger assumptions were made in the derivation
of the ratio Gamma/Gamma(L), and our work suggests that these may be unjust
ified. (C) 2000 Elsevier Science B.V. All rights reserved.