Agreement Finally!!? (was Paraphyly)
Richard Pyle
deepreef at BISHOPMUSEUM.ORG
Thu Jan 17 15:43:31 CST 2002
O.K., as I said before, I've got other things that I need to do -- so I'm
going to resist asserting any of my own opinions.
However, I am very intrigued by the direction this thread has taken in these
past few posts, so I want to make sure I clearly understand the distinction
between the perspectives held by Ken and Tom (and others on the list as
well). Both are very eloquent writers, but just to make sure we have a
mutually identical vision of the distinction, I want to further clarify with
the aid of "ASCII Art" (make sure you're displaying this using a fixed-width
font, such as Courier).
Except for the last question below, let's ignore the issue of branches
moving around and assume that we are able to perfectly place all critters
that we know about in their "true" evolutionary context. Let's also ignore
the sticky issue of hybrid origins of taxa. So, consider the following
cladogram:
^ A B C D E F G
| \ / / \ / W / /
^ 9 / \ / \ / /
| \ / X 7 Y 8 /
^ 6 / / Z / / /
| \ / / \ / / /
T 4 / 5 / /
I \ / / / /
M 3 / / /
E \ / / /
^ \ / / /
| \ / / /
^ 2 / /
| \ / /
^ 1 /
| \ /
^ 0
| \
The top line represents present-day (with the progression of time indicated
on the left axis). So, let's call the terminal endpoints A-G "species" (by
whatever definition you want -- the point is they are extant populations of
critters with zero gene-flow between them, and good gene-flow within them).
The endpoints W-Z also represent "species" in this same context, except that
none of their descendents are alive today (i.e., they are extinct).
The numbers represent "nodes" in the cladogram -- that is, events in history
when a single population of critters with good internal gene flow somehow
divided into two populations (by whatever mechanism you wish to invoke),
each retaining adequate internal gene-flow, but permanently cutting off
gene-flow between the two populations for all history thereafter.
Note that I have *not* indicated any measure or degree of morphological
distinction along the horizontal axis. All we have is a map of "true"
evolutionary history of populations over time. While these relationships
might have been derived in part from observable morphological
synapomorphies, the point is that the "true" evolutionary history exists
independently of the morphological characters that may or may not accurately
reflect it.
Thus far, I think most of the contributors to this thread (on all sides)
will be nodding their heads in agreement, with no points of contention (If
I'm wrong about this, I'm sure I'll hear about it). So now that I've taken
the time to draw my little picture to be used as a common reference point,
I'd like to use it to get some explicit assertions out of Ken and Tom (and
others), so that we can eliminate semantic ambiguities.
We could use it to address the issue of whether the population that existed
at node 9 represents the "mother" of the extant species "A" and "B"
(equivalent sisters to each other), versus whether it would be more
appropriate to think of it in terms of extinct "W" being a short-lived
"daughter" of the otherwise stable mother "F", etc. -- but I'd rather keep
the focus on what started all of this discussion; which is, how should we
apply scientific names to critters to represent our "classification"?
I'll start with some specific questions/comments, based on the "nodes" and
"species" as indicated in the cladogram. For convenience, I've reproduced
it below to minimize the need to scroll up and down. For an example of
naming conventions, we'll assume that the monophyletic clade consisting of
A-G + W-Z, which is defined by node 0, bears the higher taxon name "Zero",
and everyone is happy with that nomenclatural assignment.
^ A B C D E F G
| \ / / \ / W / /
^ 9 / \ / \ / /
| \ / X 7 Y 8 /
^ 6 / / Z / / /
| \ / / \ / / /
T 4 / 5 / /
I \ / / / /
M 3 / / /
E \ / / /
^ \ / / /
| \ / / /
^ 2 / /
| \ / /
^ 1 /
| \ /
^ 0
| \
Questions:
1. Suppose that we decide that the differences between the species cluster
A-C warrant a sub-grouping name that nomenclaturally identifies them as
distinct from the species cluster consisting of "D" and "E". I think most
people would be able to name the D-E complex easily enough -- we just call
them "Seven". But what about the A-C complex? We could restrict the
definition of the sub-grouping name to include only those descendants of the
most recent common ancestor of "A" and "C", in which case the name would be
"Six" -- but that would leave "X" out hanging in the paraphyletic breeze.
So, do we:
a) Define the subgroup name "Four" as those descendants of the most recent
common ancestor of "A" and "X", instead of defining "Six" as stated above?
b) Recognize three subgroups of: "Seven", "Four", and "Six" (or, perhaps,
"Seven", "X", and "Six")?
c) Assign a separate name to every single node *and* every single endpoint
separately?
d) Something else?
2. To make things interesting, let's say the really "profound" morphological
divergence (e.g., feathers+flight) happened between nodes 4 & 6, such that
"X" is much more closely associated with "D" and "E" in all aspects of
morphology/ecology/behavior -- except for a tiny, trivial divergence (e.g.,
a change in eye color from turquoise to light blue) that occurred between
nodes 3 & 4. Would it be justifiable to then use the name "Six" for the
group consisting of species A-C, and the name "ThreetoFour" to refer to
species X, D, and E? (i.e., using Ken's example of defining a paraphyletic
group in terms of its "start point" and "end point")
3. What name would you apply to the organisms that represent the generations
spanning the gap between nodes 3 and 7? Would they also fall under "Seven"?
Or, would "Seven" be defined as only those critters from node 7 and forward?
4. To throw the problem of ranks into this, suppose we're all happy with a
family-level designation of "Three". To keep things strictly monophyletic
(holophyletic), we'd need to additionally define the families "Five",
"Eight", and "G". But now the name "G" applies to both a species *and* a
family. We can't have that (although the two names can certainly share the
same root, they can't be the same name). We could define the family "Zero",
but we already said we'd use that to refer to the whole set (i.e., the clade
stemming from node 1, plus "G") -- perhaps at the rank of Order. One option
is to recognize species "G" as "incertae sedis" into the Order "Zero"
without assigning it to any family-level name (even though that's an
improper use of "incertae sedis", because the relationships are known).
Another option would be to arbitrarily create the family name "StemOfG" (But
then what do we call the genus of "G"? "DistalStemOfG"?), thereby defining
the name in terms of a stem (or rather, a section of stem), instead of
defining it in terms of a node or an endpoint. Or, we can assume there are
other offshoots that went extinct, and don't know about, in which case we
hypothesize additional nodes to define the family name "Ten" (and genus name
"Eleven"). Which is the best way to assign the names?
5. Suppose we have no knowledge of any of the extinct species "W", "Y" and
"Z" (and, therefore, no knowledge of nodes 2, 5 or 8). We're all happy with
our holophyletic designations and definitions of the taxa "Three",
"StemOfF", and "StemOfG". Now we discover a new fossil bed that reveals
"W", "Y" and "Z", with enough character preservation that we can recover the
correct phylogenetic placement as illustrated above. Do we then re-define
the taxon name "Three" to include "Y" and "Z" (and all the generations of
critters that descended from the population represented by node 2), or do we
instead erect the name "Five" as distinct from "Three"? And what about
"StemOfF"? Would it be equivalent to what we would have called "Eight"? If
so, then what name applies to all the generations of critters that represent
the stem between nodes 1 & 8?
6. Suppose we not only lacked knowledge of the extinct species, but we were
unable to resolve the root affinities:
^ A B C D E F G
| \ / / \ / / /
^ 9 / \ / / /
| \ / 7 / /
^ 6 / / /
| \ / / /
T \ / / /
I \ / / /
M 3 / /
E \ / /
^ 0--
| \
In this case, I've drawn (as best I can) the basal node 0 as a trichotomy,
meaning we are completely unable to resolve whether "F" branched off near
the very base of the stem that leads to "G", or the very base of the stem
that leads to node 3. If we want to name the clade "Three", then how do we
deal with "F" and "G"? If we called them "StemOfF" and "StemOfG", then what
happens if we later discover the that "F" actually branches off of the line
connecting nodes 0 and 3 (as originally illustrated)? Do we then lump
"StemOfF" in with "Three", or do we still treat them as three separate taxa?
What I'm hoping for in all of this is for Tom to answer these questions from
the perspective of a "strict" cladist who wishes to use the existing
Linnaean-based (ranked) nomenclature, and Ken to answer them using the
Kinman system. I think I have a pretty good idea on how Phylocode would
handle it, but I'd rather see someone more intimately familiar with that
system to take a shot at it instead. Of course, all other takers are
welcome. If your preferred answer is an "it depends" sort of answer, then
please feel free to make assumptions about unstated
morphological/ecological/behavioral affinities (or other factors) in any of
your answers, provided that you state them yourselves.
What I'm really interested to find out is to what extent the three (or
more?) perspectives arrive at different answers for different reasons, the
same answers for different reasons, or the same answers for the same
reasons. Also, I hope others will pose additional questions against these
(or alternate) cladograms, that we can use to explicitly dissect what we are
really talking about and identify the real distinctions in our respective
positions.
Aloha,
Rich
Richard L. Pyle
Ichthyology, Bishop Museum
1525 Bernice St., Honolulu, HI 96817
Ph: (808)848-4115, Fax: (808)847-8252
email: deepreef at bishopmuseum.org
http://www.bishopmuseum.org/bishop/HBS/pylerichard.html
More information about the Taxacom
mailing list