Message 2004-10-0001: Re: REPOST: Crowns, Panstems, and their Correspondence to ea=

Thu, 09 Sep 2004 15:40:25 -0400

[Previous by date - RE: Ichthyornis paper]
[Next by date - Re: Panstems]
[Previous by subject - Re: RE: crown clade convention]
[Next by subject - Re: REPOST: Crowns, Panstems, and their Correspondence to ea=]

Date: Thu, 09 Sep 2004 15:40:25 -0400
From: [unknown]
Subject: Re: REPOST: Crowns, Panstems, and their Correspondence to ea=

>My responses to TMK=3D92s comments are enclosed in these symbols.<

I notice that a number of the abstracts follow de Queiroz and Gauthie=
r =3D
2002 in creating panstem/crown pairs, in each case with the name of t=
he =3D
panstem formed by adding the prefix "Pan-" to the name of the crown c=
lade. =3D
I've commented on this and discussed it on another forum, and thought=
 I'd =3D
summarize some thoughts here.

It is, of course, interesting to note that crowns and panstems have a=
 1:1 =3D
correspondence with each other, and it does seem that this would warr=
ant =3D
at least a recommendation with regards to the definition and/or nomen=
e of these clades, possibly having them always be named in pairs. I'm=
curious; does anyone think a rule would be warranted?

>We took a straw vote on this (i.e., universal use of the Pan- prefix=
 for =3D
panstems) at the Paris meeting.  Only a few people, including myself,=
 were =3D
in favor of such a rule, though many supported a recommendation.  Nam=
ing =3D
the clades in pairs is a related, but different, issue (see below).<

There is one instance where the correspondence is not really 1:1, at =
least =3D
for practical purposes, and that is, of course, if the extant outgrou=
p to =3D
a crown clade is also the immediate ancestor. In this case the panste=
m and =3D
the crown clade become heterodefinitional synonyms. Does anyone think=
 that =3D
this means certain crown clades should not have panstems named for th=
em? I =3D
personally think that it's fine for the two clades to be heterodefini=
 synonyms, as our understanding may shift. (Of course, in every insta=
nce, =3D
the crown clade should be senior to the panstem.)

>I would also treat these names as heterodefinitional (subjective) =
synonyms, given that we never actually know that the two names refer =
to =3D
the same clade.<


This has been discussed before, and is mentioned in Note 9.4.1, which=
states that the author should clarify their exact meaning for the wor=
d =3D
"extant". It seems to me it may be preferrable for the code to define=
 a =3D
consistent meaning. The most stable and easily applicable definition =
I can =3D
think of is "living and published at the time of the definition's =
publication", but I'd certainly like to see others' suggestions. Shou=
ld =3D
"...and published..." be included? It would help maintain the stabili=
ty of =3D
previously established crown and panstem clades.

>The approach we have taken is not to have a rule specifying the mean=
ing =3D
of =3D93extant=3D94 but to have a rule stating that a specified meani=
ng (the =3D
same one that TMK suggested) is to be adopted if the original author =
did =3D
not specify the meaning.  This gives the authors of definitions more =
freedom but also prevents ambiguity.  Here is the rule:
9.5. If the author of a crown clade definition (Note 9.4.1) did not =
specify the meaning of "extant" or "crown clade" or an analogous term=
 used =3D
in the definition (e.g., "living", "Recent"), then subsequent authors=
 are =3D
to interpret that definition as referring to organisms or species tha=
t =3D
were extant on its publication date (Art. 5).<


I notice two schools of thought in the abstracts for the upcoming mee=
ting =3D
as to definitions for crown and panstem clades. In Sereno's definitio=
ns, =3D
there really are no panstems, only stem-based clades that happen to h=
ave =3D
extant specifiers. The crown clades, then, are defined in terms of th=
e =3D
stem-based clades:

Stem =3D3D Clade(A <-- B) (A and B are extant)
Crown =3D3D Clade(extant Stem)

However, in other abstracts, such as the one for Gauthier et al.'s pa=
per =3D
on major amniote clades, it goes the other way: the panstem is based =
on =3D
the crown.

Crown =3D3D Clade(A + B) (A and B are extant)
Panstem =3D3D Clade(Crown <-- extant non-Crown)

In the former, the stem-based clade is the more stable one, while, in=
 the =3D
latter, the node-based (crown) clade is more stable. It seems preferr=
able =3D
to me that the crown be the more stable one, but I'd be very interest=
ed to =3D
hear other opinions.

I would allow one exception, and that is Wagner's brilliant definitio=
ns =3D
for _Panbiota_ and _Biota_ which, while not exactly following the =
stem-modified crown approach (it would be impossible to do so, as the=
re is =3D
no outgroup), are in the same spirit, with the crown clade being modi=
fied =3D
by a more inclusive clade.

>I found this comment very interesting, as I hadn=3D92t noticed this =
difference until TMK pointed it out.  I=3D92m not sure what he means =
when he =3D
says that one definition is more stable than another in this context.=
What is clear to me is that the stem concept is primary in the first =
approach in that the crown concept is based on it, while in the secon=
d =3D
approach the crown concept is primary with the (pan) stem concept bei=
ng =3D
based on it.  Although I=3D92ve always thought of things the second w=
ay (up =3D
until now), the reason was that I had never considered the alternativ=
e.  =3D
Now that the alternative has been pointed out, I like it just as well=
.  In =3D
other words, even though I might be expected to favor the primacy of =
crowns, I don=3D92t.  Both approaches seem equally reasonable.  My in=
tion:  perhaps this is a reason for naming the clades bounded by exta=
nt =3D
organisms in pairs (see TMK=3D92s first comment, above), as I have my=
self =3D
suggested (though not clearly distinguishing this issue from the rela=
ted =3D
one about how the names are to be formed).  That is, perhaps we shoul=
d =3D
have a rule stating that establishing a name for one clade of this ty=
pe =3D
(crown or total) automatically establishes a name for the other membe=
r of =3D
the pair.  Of course, this only works if the names are tied to one =
another, as in the Pan- convention.  Notice, however, that it does no=
t =3D
require the Pan- convention (and here I=3D92m generalizing this conve=
ntion =3D
to include other prefixes than Pan- that could be used for total/stem=
clades).  That is, we could have a convention that bases the name on =
the =3D
crown on that of the stem rather than vice versa.  For example, rathe=
r =3D
than using Mammalia for the crown and PanMammalia for the stem, we co=
uld =3D
use something like AcroMammalia from the crown and Mammalia for the s=
tem.  =3D
My preference, for reasons outlined in several papers, is the former =
(i.e., Mammalia and PanMammalia).  However, this does not mean that I=
prefer the primacy of crowns when it comes to definitions (as opposed=
 to =3D
names).  That is, even if I were to use Mammalia for the crown and =
PanMammalia for the stem, I could still define the names as follows:
PanMammalia =3D3D Clade (Mammalia [Homo sapiens and Ornithorhynchus =
anatinus] not Reptilia [Testudo graeca and Lacerta agilis])
Mammalia =3D3D Clade (extant PanMammalia). =3D20
Of course, this approach pretty much forces one to name the clades in=
pairs, since one wouldn=3D92t know what name to use for the stem unti=
l one =3D
had decided on a name for the crown, but as I said above, perhaps nam=
ing =3D
these clades in pairs is a good idea.<


While it seems like a good idea to delineate the correspondence betwe=
en =3D
crown and panstem clades, I see some issues ahead with using the "Pan=
-" =3D
prefix. (And I thought mandated affixes were something PhyloCode was =
trying to get away from....)

>The PhyloCode is not necessarily trying to get away from standard =
affixes=3D97only from tying particular affixes to particular ranks.<

A number of currently named taxa actually do start with "Pan-", and a=
re =3D
not panstem clades. Is someone who is not familiar with carnivoran =
taxonomy to see the clade _Panthera_ and assume it is the panstem of =
"Thera"? (When, in actuality, by the philosophy being followed, it wo=
uld =3D
be a crown clade, and "Panpanthera" the panstem!)

>To solve this problem, people have suggested conventions involving =
hyphens and/or double capitalization.  Thus, Panthera would not be a =
panstem name, but PanThera (or Pan-Thera) as well as PanPanthera (or =
Pan-Panthera) would.<

Among dinosaurs, there is a _Panoplosaurus_ which is certainly not a =
panstem clade containing _Oplosaurus_ (which is an actual unrelated =
dinosaur genus).

>See previous comment.<

Then there are some existing "Pan-" taxa which have actually been nam=
ed =3D
after other taxa, such as _Panarthropoda_ and _Pancrustacea_. These w=
ere =3D
not named as the panstem clades for _Arthropoda_ and _Crustacea_, =
respectively, but for more inclusive groups.

>As far as I can tell, the original authors of the name Pan-Crustacea=
(Walossek and M*ller, 1990) did use that name for the total clade/pan=
stem =3D
of Crustacea.  You=3D92re right about Panarthropoda  (Nielsen, 1995).=
  That =3D
name would have to be redefined so that it applied to the total clade=
 of =3D
Arthropoda (remember, the PhyloCode is not retroactive).  To achieve =
a =3D
unified system, some names, such as this one, would have to be define=
d =3D
differently than originally used.  The only way around this problem w=
ould =3D
be to use affixes that had never been used before=3D97ever, but this =
strategy isn=3D92t ideal either.<

Closer to home, if we were to have panstem clades separating our spec=
ies =3D
=66rom its closest living relatives, it seems they would be "Panhomo"=
 and =3D
"Panpan" (_Pan_ being the crown clade of chimps, not the panstem clad=
e for =3D
""). "Panhomo" is rather confusing, as it looks like a combination of=
_Pan_ and _Homo_.

>This seems like a minor drawback.<

(On another topic, while it may seem like there can be no crown clade=
called _Homo_, as there is only one extant species, I submit that the=
re is =3D
one: "the clade stemming from _Homo sapiens_". Of course, this would =
necessitate removing _neanderthalensis, _erectus_, etc. from _Homo_, =
which =3D
would be highly disruptive, to say the least....)

>I (and several others=3D97Wagner, Wolsan) have independently noticed=
 that =3D
this kind of definition is possible.  I call it ancestor-based.<

Another point to make about the proposed naming convention for crown/=
m pairs is that, while it may be obvious that _Panmammalia_ is the pa=
nstem =3D
of _Mammalia_ (even if, for example, _Panderichthys_ is not the panst=
em of =3D
"Derichthys"....), it's not obvious that _Mammalia_ is a crown clade =
=66rom =3D
its name alone. If there is a mandated prefix for panstems, why not o=
ne =3D
for crowns as well? (I'm not saying I would like one, just why one an=
d not =3D
the other?)

>A standard crown prefix is certainly possible.  As I argued in my ta=
lk at =3D
the Paris meeting, the disadvantage is that it creates discontinuity =
by =3D
replacing existing names with new ones.  Of course, the same could be=
argued for the stem clades, but if we want a relationship between the=
names of the members of a crown-stem pair, at least one of the names =
has =3D
to change.  I suppose the difference here has to do with what one =
considers the primary reason for the Pan- convention.  In my view, th=
e =3D
primary reason is not to identify a name as the name of a member of t=
he =3D
class of total clades generally but to identify it as the name of the=
total clade corresponding to a particular crown.  To identify the nam=
es =3D
associated with the members of the classes of crown and total clades,=
 I =3D
prefer to use symbols (inverted triangle and inverted triangle with a=
basal leg, respectively).<

One solution might actually be found in the traditional system, where=
 taxa =3D
may optionally be written with a title, e.g. Kingdom Animalia, Class =
Reptilia, Order Primates, etc. Why not have a similar option for Phyl=
oCode =3D
taxa, so that the names themselves can be free to be formed in any ma=
nner? =3D
Crown Clade _Amniota_, Panstem Clade _Synapsida_, Stem Clade _Saurisc=
hia_, =3D
etc. (I find it often handy to use just "Clade" as a title when contr=
 with traditional taxa, e.g., Clade _Reptilia_ vs. Class Reptilia.)

>Yes, some people are already explicitly stating the type of clade.  =
In =3D
addition, there=3D92s no reason why both conventions can=3D92t be use=
d =3D
simultaneously (e.g., Crown Clade Mammalia, Stem Clade PanMammalia).<

Of course, this has the disadvantage of not nomenclaturally linking =
panstems with their crowns. (Although it might be useful in any event=
.) =3D
Another thought would be to allow punctuation just in this case, name=
ly, a =3D
hyphen: _Pan-Mammalia_, _Pan-Amniota_, _Pan-Vertebrata_, etc. Then it=
would be clear that _Panoplosaurus_ is not a panstem clade, while =
_Pan-Sauria_ is.

>Exactly; see my earlier comment.<

9 Sep 2004

Kevin de Queiroz
Division of Amphibians & Reptiles
Smithsonian Institution
P.O. Box 37012
NHB, Room W203, MRC 162
Washington, D.C. 20013-7012
Voice:  202.633.0727
FAX:  202.357.3043


Feedback to <> is welcome!