Jim Blake is correct in pointing out the increasing detailing of our
descriptions. I would like to make two minor points: First, earlier
scientists were just as bright as we are. They had a different
knowledge-base to work with, and as a consequence made decisions
that we would not make at this time (I hope). We all describe the
organisms as much as we believe it necessary, so that we can separate
them from all other known organisms: Most descriptions could be
characterized as extended differential diagnoses, rather than as true
descriptions. Secondly, we have increased the numbers of features
that must be described tremendously over the last several years: We
are now discussing detail we would not even consider 50 years ago.
This means that collections are becoming increasingly important: The
only way we can correct our previous identifications is to go back to the
material and describe those features which we previously thought were
unimportant. As a moral to this little story: Deposit voucher material in
permanent collections whenever you can and especially if you are
introducing a "new" taxon to your fauna, be sure to do so, so that it can
be carefully checked on a later date by somebody. In this business
nobody ever says the final word: We must all be prepared to be
corrected by future development; this is what science is all about.