I have also been guilty of describing species from one specimens, in the
bivalvia Euciro queenslandica and in the scaphopoda two. The criterea I
use is to find at least three distinctive variations which set the
species apart from its nearest "lookalike". So far the species remain
valid.
There are many early species named from one specimen, Linnaeus, Lamarck
and many other early authors did this and in most cases we now know the
species as quite common dur to much more recent collecting.
Kev