Ross Mayhew asks how someone determines that specimens belong to a new
species, and then names them. That is a tall order. General works on
taxonomy tend to skip over this part in favor of telling the details of how
to name a species properly. But there is a good description of the "Shell
Name Game" in Raup and Stanley's "Principles of Paleontology", first or
second edition.

In brief, the process of discovery goes something like this:

1. Try to identify the specimens with your existing library and reference
collection. Fail to find a valid name. Allow yourself a raised eyebrow.
2. Gather ALL of the comparative literature and specimens that might be
relevant and try to identify it again. Fail again. A small measure of
excitement is permissible, even expected.
3. Ask experts for advice. They fail. Get excited, but don't uncork the
bottle until the reviewers approve the manuscript for publication.

At this point, it is likely that your specimens belong to a new species,
and the negative process of failure has been converted into the positive
process of discovery. If step 2 is not completed, then you may end up by
naming a new, unwanted junior synonym of an already named species. Because
it is difficult or impossible to amass and comprehend all of the possibly
relevant literature and specimens, even the experts name new junior
synonyms now and then.

Of course, if your failure to identify the specimens is due to your making
a sloppy, overly fine, or even nonexistent distinction from already named
species, nothing will help you from gaining the reputation of a splitter.
About a third of all the taxa that have ever been named are invalid
synonyms or homonyms. As shown in monographic revisions of genera and
families, this ratio remains remarkably constant through the years despite
changes in taxonomic philosophy, which is a philosophic concept I leave the
reader to ponder.

Andrew K. Rindsberg
Geological Survey of Alabama