Comment on `Two-body random ensembles: from nuclear spectra to random polynomials

In PRL 85, 3773 (2000) it was suggested to use random polynomials to analyze and understand the properties of two-body random ensembles. In this comment we point out that for the vibron model the random polynomial is not quadratic, but has a more general form. We make a comparison for the percentage of ground states with L=0 in the vibron model with random interactions.