in defense of Cavalier-Smith

Thu May 5 18:15:45 EST 1994

In the absence of Cavalier-Smith's contribution to this debate, I think
that his proposal should be defended from several of the criticisms 
levelled against it.

Firstly, I am partial to Mark Farmer's feeling that a panel of experts
would be ideal for the purposes of making a comprehensive classification
of protists.  However, there are also advantages to individuals taking on
such a task.  For instance, someone not directly involved on working with
a particular group of protists is less likely to inflate their
"differentness" and therefore their rank in a classification. But this
is not the central issue.

Cavalier-Smith's reliance on the specific branching order of his rRNA tree
to create higher ranked taxa may be problematic as Mark point's out.  For
instance, if any one of the 15 protist phya that TCS group's in the
Infrakingdom Neozoa are an outgroup to the infrakingdom Euglenozoa (which
may be the case for Entamoebids) then the classification must be emended
quite radically by either the creation of a new infrakingdom or the demotion 
of one of the existing ones.  Stability at this high level
of taxanomic rank might be useful.  On the other hand, TCS emphasizes 
the phylum rank above his new ranks (in the title, for instance) and
very few of his protozoan phyla would any protistologist disagree with
strongly, I think (with the possible exception of the Opalozoa). Perhaps
someone could correct me one this.  For popular use, his phyla reflect
current knowledge of protist relationships based on rRNA and ultrastructural
characters better than any other system I know. 

Mark's other criticisms include Tom's "controversial" statement that masking 
of sequences in rRNA trees is subjective....well isn't it?  In fact, Mark's
assumption that a "proper alignment" is known for rRNA sequences highlights 
the central problem with the field of rRNA phylogeny; there is no objective 
way of arriving at a good alignment of this dataset.  All of the "experts"
ultimately rely on non-algorithmic judgement calls.  So, it seems to me that
Tom's judgment calls need not, a priori, be any worse than, for instance, Mitch
Sogin's or Carl Woese's.  The fact that the topology of Tom's rRNA tree is at
odds with published topologies from Sogin's lab and DeWachter's group,
indicates, perhaps, the limits of the ability of anyone to arrive at "the proper
alignment" for 18S rRNAs.

In addition the criticism that Tom has ignored the suggestions of others on 
how to interpret molecular phylogenies based on their data, seems a little 
odd. Shouldn't a good interpretation of data stand up to an independent test 
by another researcher?  If the data strongly supports a given conclusion, then
any competent researcher in the field should be able to arrive at the same
conclusion given this data.  This, it seems to me, is the essence of how science

O'Kelly's first criticism of TCS's paper appears to be a difference in
opinion as to what the purpose of the classification is.  Tom erects
new high level groups as cell evolutionary hypotheses which need to be 
tested. As new data comes in, he sometimes rejects his old hypothesis for a 
new one.  A case in point is the removal of Entamoeba from the Archezoa
(where it would be primitively amitochondrial) to the Metakaryota (where
it is hypothesized to be secondarily so).  Hopefully, the collection of
more molecular data from which to construct more robust trees (analyses of
a several gene dataset perhaps) will confirm or reject his new hypothesis.
Contrary, to what O'Kelly implies, it will be data which ultimately forces
us to reject or confirm a TCS hypothesis, not personal opinion.

The question of whether the Jakobids or the Percolozoa are the outgroup
lineage to all other mito-bearing cells is an interesting one.  As I see it, the
major disagreement between TCS and O'Kelly, is the relative
stability of ultrastructural characters.  For O'Kelly, the three main 
shapes of mito cristae are indicative of the divergence of three primary mito-
bearing lineages after the initial endosymbiosis.  Each of his three
types of Jakobids which display one of the tree types of cristae is an outgroup
to each of the main lineages of metakaryotes.  Tom argues that
mito cristae shape is more evolutionarily plastic and that the change from
tubular to flat cristae (at least) has taken place several times.  He
argues that the Percolozoa, which lack golgi stacks (like archezoa) but possess
a mito (like metakaryotes) , are the first mito-bearing lineage.
Right now the rRNA data that Tom presents is consistent with this.  
Perhaps O'Kelly's case will be strengthened by rRNA sequences from the

It is rather ironic that O'Kelly's final criticism, that Tom creates
too many paraphyletic groups, applies rather aptly to his own group
the Jakobids.  According to O'Kelly, the Jakobids are thrice paraphyletic,
representatives of which have given rise to the three major mito-bearing
lineages.  This is not a fault of O'Kelly's group Jakobid, rather his
assumption that there is something more "proper" about holophyletic 
groups (monophyletic sensu Hennig).  Paraphyletic groups are 
phylogenetically coherent (they contain shared derived characters), some
members of them simply have given rise to other groups.  As long is
this is clear to people interpreting classifications, I see no reason
to purge the world of paraphyletic, yet natural, groups.

Andrew Roger     

More information about the Protista mailing list