But the paleontological method would only reveal physical differences, not the gene pool isolation, correct?
Correct, unless these were samples recent enough to allow recovery of some gentic material, as in the case of the neanderthal mtDNA sequences. Such sequences might not completely rule out gene flow, but they could be suggestive.
According to the latter definition, if you had two groups of bears of the same type, one group in Oregon and the other in Vermont, then they would constitute two different species. That doesn't make sense to me.
Generally geographic isolation itself is not considered sufficient to classify different populations as seperate species. If the two populations are subsequently reintroduced to each other and can successfully interbreed then they would not be considered to be seperate species under the biological species concept.
The distinction is more commonly drawn between pre-zygotic, i.e. behavioural or morphological, barriers to reproduction as opposed to post-zygotic, i.e. physiological or genetic or developmental, barriers to reprodcution leading to either a lack of interfertility or a lack of viability or sterility in hybrid offspring.
TTFN,
WK