ABTI's

[John Noyes]

Donald has an interesting thought, but being old-fashioned I obviously
cannot concur, at least not completely. Using molecular methodology to
inventory species might work, but I do not think so, at least not for some
time. For a start we would have to find the correct gene sequence(s) for
each group in order to recognise the species. Work on chalcidoids has shown
this to be difficult in many cases, even in closely related taxa. For
instance in Trichogramma ITS2 works very well for some species, whilst for
others (even closely related species) it does not. Also, we would still
have similar problems in quantifying variation and deciding where to draw
the line between species. In some groups a small difference in sequence may
be enough to characterise a species, whilst in others a much larger
difference may be necessary. I would also suggest that we have very little
idea how species defined by gene-sequence would correlate to biological
species so the same applies here as to morpho-species. We also have the
problem with singleton specimens or specimens that cannot sequenced for one
reason or another. God forbid that we have to give a particular sequence,
as suggested by Donald, when we are mentioning in a publication or during a
lecture a species which is the subject of some research. I can remember
names quite easily but a list of seemingly random letters (albeit it only
four different ones to consider) - no way. Even given a molecular approach
which works I do not think it would be any cheaper or easier than a
traditional morphological approach because it would still be necessary to
train, and employ, specialists who could assess the differences between
sequences (although I expect that sometime in the future it will be
possible to "train" computers to do this). It will also be necessary to
train people who can sort specimens into groups (RTU's) where selected
specimens can be sequenced otherwise it would be necessary to sequence all
15400m specimens that would be necessary to sort out the world's
chalcidoids. 15400m specimens could be mounted, labelled and sorted much
more quickly and cheaply than sequenced. No, Donald I definitely cannot
agree with you.

I do not totally throw out the molecular approach for defining species. I
am convinced that the inital approach should be one using traditional
morphological methodology, followed by molecular methodology for the
difficult groups if necssary. I see the role of the morphotaxonomist as one
who does the initial ground work, followed by other methods to resolve
difficult or complex problems (such as Anisopteromalus mentioned by Donald)
where there is a need.

Donald may be right in stating that it may eventually be easier to identify
species using molecular techniques (it already is with some groups - but is
it cheaper?) but it will certainly be easier to convey information about
species by whatever medium you care to name if the species have names
rather than an amino acid base sequence.

Finally, I think that the chances of either chalcidoid or ichneumonoids
being completely inventoried on a world-wide basis using either methodology
or a combination of both is about as likely as me walking on the moon.

John

At 03:37 PM 3/12/2001 +0000, you wrote:
>hi all
>
>I too have been pondering on doing a large group - in my case the
>Ichneumonoidea. probably not as speciose as the Chalcidoidea, but certainly
>35,000 described and roughly 3 to 5 times as many awaiting description
>(i.e. not quite so tropicocentric as chalcidoids).
>
>John considers the preoblem purely from a morpho-taxonomic point of view
>and I would propose taht in the 21st century we may be obliged to combine
>this or even to substitute this for a molecular approach.
>
>My reasons are as follows -
>
>Firstly, we have very little idea (at least among parasitic wasps) about
>how our traditional morpho-species correlate to actual biological species
>(don't let's argue about definitions - things which are interbreeding ...).
>Recent studies, e.g. by Gokhman et al., have for example, shown taht two
>parasitic wasps (Anisopteromalus in his case) can look to all intent and
>purpose identical, yet be clearcut species with totally different
>biologies. Such cryptic species seem to occur very frequently in any
>applied (i.e. biocontrol) situations, and i personally would guess that it
>could be anything between 5 and 50%. It would seem to me impotant to know
>the magnitude of this cryptic species problem first - before assuming that
>we must follow a morphology-based path - when in fact we have so far done
>rather less than half the work. Molecular approaches will be the easiest
>way of solving this , and can be applied to a random sample of putative
>species based on traditional morphological defnition.
>
>Secondly, it takes a long time to train morphologists to any competance.
>Some would say that they are born rather than trained - i would concur. It
>is far easier to train people to sequence things, and a lot can be
>automated.
>
>Thirdly, John points out that there could well be a duplication of effort
>in even a well-run scenario as he describes. It is hard to check what
>others are describing before they have described them. True too of
>sequences, but you can get several sequences done in a few days, and place
>them on a readily searchable data base. Certainly, taxonomic expertise will
>be needed to give some sort of identification to taxa - but it may not be
>necessary or advisable to describe and name each morpho-species.
>
>I am not saying do away with morphology, or don't bother about voucher
>specimens. Rather that molecular approaches will have to be brought to bear
>too. Obtaining the material, setting up large international networks of
>traps, arranging one or more dedicated sequencing centres and housing the
>collections are of course necessary costs here. But I envisage a time when
>it will be easier and cheaper to do research on biology/ecology/behaviour
>of organisms such as insects in the tropics by saying that you are working
>on the species of X whose DNA sequence is "ATGGTCGAT...." rather than on
>Aus bus.
>
>With an appropriate choice of two or three gene fragments it should be
>possible to provide a full species-specific identification. And once that
>sequence is archived, it won't matter if the world expert on the group has
>long-since retired or succumbed to drink - the species will always be
>recognisable
>
>
>Donald


John S. Noyes, Entomology Department, The Natural History Museum, Cromwell
Road, South Kensington, London, SW7 5BD, UK

Tel. +44 (0)207-942-5594  Fax: +44 (0)207-942-5229

INTERNET: jsn at NHM.ac.uk

Interactive catalogue and biological database of World Chalcidoidea:

http://www.springer-ny.com/detail.tpl?cart=97108527710714749&ISBN=354014675X
or search for Noyes at:
http://www.springer-ny.com/

Encyrtidae of Costa Rica:
http://www.nhm.ac.uk/science/intro/entom/project2/index.html

Course on taxonomy and biology of parasitic Hymenoptera:
http://www.nhm.ac.uk/entomology/hymcours