Voucher specimens

[Kenelm Philip]

	Neil Jones stated:

>The deposition of voucher specimens in collections has traditionally
>been seen as they way that proof is provided of its existence at a
>site. Of course this proves nothing.

Interesting reasoning, indeed. I guess we should throw out the 20 million
insect specimens at the Smithsonian, and the 3 million butterflies at the
Allyn Museum, since they prove nothing...  You could equally well claim
that, since textbooks have errors, no texts should ever be used.  :-)

	In a slightly more serious vein, I have friends who are, or were,
curators at major lepidoptera collections, and have examined material in
such collections myself purporting to be from areas I know something about.
Neil is perfectly correct--there are errors on locality labels in many
collections. However, the curators are usually aware of which collectors
(or collections, when an entire collection is acquired) are unreliable.
On the other hand, there are collectors known to be highly reliable, and
when their labels are backed up by field notes there is no reason to sus-
pect locality errors. 

	Voucher specimens can provide information no photograph or sight
record can. Maybe the British fauna (including all micros??) is totally
known by now with _no_ present or potential taxonomic problems--but the
same is not true in North America, even for the butterflies (let alone
moths!). Only a specimen, capable (if need be) of being dissected, can
provide good evidence of what taxon was really present at locality X
at time Y. In the one area I know a little about (the western North
American arctic/subarctic), it is absolutely _essential_ to have actual
specimens of many species to document their occurrence. At least when
_specimens_ are misidentified the error can be corrected later, but
mis-ID'd sight records are pretty hopeless, and in a number of arctic
species even photographs aren't very helpful. Finally, with all due
respect to Neil's views about collecting, I know of no reason in the
western NA arctic why the levels of collecting that have been occurring
during the last 30 years should not be continued as needed, unless
someone objects on principle to deliberately killing individual insects
for any reason (in which case I beg to disagree).

	I would agree with Neil to the following extent: when a species
is well known and easy to ID, and adequate series exist from the locality
in question, it is not necessary to provide a specimen for every chrono-
logical record of its occurrence at that locality. I have some interesting
records for temporal dissociation among _Boloria_ (s.l.) in Alaska based
on sight records--but adequate series exist of all the taxa involved from
all the localities involved.

							Ken Philip
fnkwp at aurora.alaska.edu