Empirically derived models of amino acid replacement are employed to s
tudy the association between various physical features of proteins and
evolution. The strengths of these associations are statistically eval
uated by applying the models of protein evolution to 11 diverse sets o
f protein sequences. Parametric bootstrap tests indicate that the solv
ent accessibility status of a site has a particularly strong associati
on with the process of amino acid replacement that it experiences. Sig
nificant association between secondary structure environment and the a
mino acid replacement process is also observed. Careful description of
the length distribution of secondary structure elements and of the or
ganization of secondary structure and solvent accessibility along a pr
otein did not always significantly improve the fit of the evolutionary
models to the data sets that were analyzed. As indicated by the stren
gth of the association of both solvent accessibility and secondary str
ucture with amino acid replacement, the process of protein evolution-b
oth above and below the species level-will not be well understood unti
l the physical constraints that affect protein evolution are identifie
d and characterized.