We consider envelopes of DB white dwarfs that are not composed of pure He-4
but rather a mixture of He-3 and He-4. Given this assumption, the same dif
fusive processes that produce a relatively pure H layer overlying an He lay
er in the DA's should work to produce an He-3 layer overlying an He-4 layer
in the DB's. We examine the relevant timescales for diffusion in these obj
ects and compare them with the relevant evolutionary timescales in the cont
ext of the DBV white dwarfs. We then explore the consequences that He-3 sep
aration has on the pulsational spectra of DBV models. Since GD 358 is the b
est-studied member of this class of variables, we examine fits to its obser
ved pulsation spectrum. We find that the inclusion of an He-3 layer results
in a modest improvement in a direct fit to the periods, while a fit to the
period spacings is significantly improved.