Taxacom: replacing un-sequenceable types (was Re: Minimalist revision of Mesochorus)

[Richard Pyle]

> To be honest, I'm not quite following the suggestion to establish a
parallel
> "DNA-type" system. It basically boils down to "DNA-type" (well, not the
DNA,
> but the entire specimen, including its DNA) serving the role of the
neotype, and
> setting aside the real type

Not sure if this was a reference to what I was suggesting and asking about,
but it doesn't describe what I was talking about.  Rather, what I was trying
to characterize is a system where the name-bearing type remains the
name-bearing type, but if it's unsequenceable, then there would be a
mechanism in the code to formally link another specimen that is sequenceable
to the name explicitly for the purposes of attaching a nomenclaturally
relevant DNA sequence.  This would be more like an epitype than a neotype.
If a neotype is needed, then the Code already has provisions for that, and
presumably one would select a neotype that can be sequenced, in which case
the sequence would directly be from the name-bearing type.

> The way to do it, is to genome-sequence (not
> "barcode") the real type, and if it is "unsequenceable" (that implies
> unidentifiable, not fit to serve as the name-bearer for taxa that are
currenly
> eligible for the neotype designation), set it aside, and replace it with
the real
> neotype. From our experience, there aren't that many "unsequenceable"
types.

Most fishes have been fixed in formalin and are for all practical purposes
unsequenceable (at least via existing technologies).  I'm not sure how
common that is among other organisms.

I'm curious whether you mean that, without a sequenceable name-bearing type,
that automatically makes it unidentifiable (and thus a neotype is required)?
Or if you mean this only in the context of the subset of species for which
there is no known morphological characters that are diagnostic.

This potentially opens up another contentious debate (that has graced
Taxacom many times before) along the lines of:  If the only way to diagnose
a species is via it's DNA, then should it really be thought of as a distinct
"species" in any useful sense of that word?  Surely there are some examples
where even the most diehard traditionalist would acknowledge the recognition
that some sets of organisms ought to be considered as distinct "species"
even when there is no consistent morphology to use for diagnostic purposes.
But I would hope that the meaning of "species" never gets so diluted that
such examples are anything more than a very rare exception.

Note:  This is a very different debate than the one of "triage taxonomy"
where only molecular characters are used for diagnostic purposes when
proposing a series of new names, with the idea that the cost of creating a
few junior synonyms is exceeded by the benefit of assigning standard labels
(scientific names) to likely legitimate taxa in order to accelerate the
biodiversity documentation process.  In such cases, one would presume that
the diagnostic DNA sequences were taken directly from the actual
name-bearing type; otherwise how would one ever know if the sequence was
relevant to the name-bearing type, if there's no morphology to associate the
specimen from which the DNA was taken with the specimen fixed as
name-bearing type?

Aloha,
Rich