From calculations of the third harmonic generation (THG) spectra of finite polyene chains McWilliams and Soos [J. Chem. Phys. 95, 2127 (1991)] have attempted to explain the experimental THG spectrum of trans-polyacetylene, the infinite polyene chain. According to the authors, the frequency of the two-photon resonance, as well as its intensity, are too large, and substantial interchain Coulomb interactions are necessary to explain experimental data. We point out here that these results are finite size effects. Determination of correlation effects requires comparison of correlated and uncorrelated chains of same lengths. Comparison of short correlated chains with long uncorrelated chains is not the ideal approach, as results assigned to correlation effects can in reality be finite size effects.