Rouse & Fauchald Phylogeny Paper

Kirk Fitzhugh fitzhugh at
Mon May 13 14:44:49 EST 1996

To Pro-annelido-moleculoid types:

Since Bernie Cohen, Jim Blake, and Michael Satterwhite interpreted my
statement about moleculoid research as meaning that I hold no hope for
sequence data in phylogenetic research, let me clarify some matters.

I did not intend to condem sequences as data unworthy of inclusion in
cladistic analyses, much less for inferring patterns of relationship at any
level. I've no problem with the notion that science should strive to collect
observations. Sequences (qua observation) are no different from what one
might gleen from looking at gross morphology. Sequences suffer the same
(maybe more) inherent problems with respect to homology as morphology. Data
are data no matter what, and in the context of cladistic analysis, the goal
is to explain why a given set of taxa share particular states to the
exclusion of other taxa. Thus, my statement that the interest in phylum
relationships is a "growing trend with the advent of the misguided and
grossly naive notion that molecular sequence data will tell us how phyla are
related" was not a general remark on the data themselves, but on the
percieved greater importance of those data relative to other observations as
well as the generally poor degree to which phyla are represented in these
studies. To sequence, say, a couple of flatworms, a sipunculan, a
vestimentiferan, a polychaete, an oligochaete, and a fruit fly, and produce
a cladogram in the name of gaining substanative understanding of phylum
relationships is laughable.  Similarly, I can't think of anyone so
completely fluent in the systematics of all invertebrates that they are
truely qualified to discern plesiomorphic taxa representative of each
phylum. Again, this presumes a solid knowledge of within-phylum
relationships, the lack of which was the concern expressed in my earlier
posting. I certainly would agree that sequences have profound utility when
there are inherent problems characterizing homologous conditions of
morphological characters between phyla. What I vehemently disagree with is
the position that one suite of characters, say sequences, is "better" or
"more informative" than another suite for the sake of simply generating trees.

Once again, my problem is not with the data per se, but the manner in which
it has been used. As I tried to indicate earlier, working at the level of
phylum to the exclusion of having a better understanding of lower level
relationships is a very fundamental problem. To work at the level of phylum,
or even class or order, using molecular data from a limited set of taxa
representing these groups presumes the monophyly of each group. If evidence
of monophyly is lacking, then the patterns generated are not particularly
illuminating, and potentially very misleading. For instance, if one is
interested in relationships of polychaete families, the more basic issue of
polychaete monophyly must be addressed as a part of this research, which
necessitates consideration of polychaete/clitellate relationships, not to
mention how vestimentiferans/pogonophorans, echiurans, and arthropods fit
into all this. That polychaete monophyly is tenuous at best, certainly calls
into question the efficacy of a phylum-level research program, especially
one as grand as looking at relationships among the metazoa, regardless of
using sequences or morphology. 

I hope this has soothed the concerns of Cohen, Blake, and Satterwhite.  The
bottom line is that I have always been an advocate for seeking evidence of
all sorts for phylogenetic research, especially given that I support
analyses of _ALL_ data combined as opposed to the use of consensus
techniques. Having data at hand is one thing, using it responsibly and
within the purview of accepted scientific practice is another.


Kirk Fitzhugh
Associate Curator of Polychaetes
Research & Collections Branch
Los Angeles County Museum of Natural History
900 Exposition Blvd
Los Angeles CA 90007

Phone:  (213) 744-3233
FAX:    (213) 746-2999
e-mail: fitzhugh at

More information about the Annelida mailing list