BZN Volume
58, Part 2, 29 June 2001
General
Articles & Nomenclatural Notes
General
Articles and Nomenclatural Notes with the
following titles were published on 29 June
2001 in Volume 58, Part 2 of the Bulletin
of Zoological Nomenclature
Copies
of these General articles and Nomenclatural
Notes can be obtained free of charge from
the Executive Secretary, The International
Commission on Zoological Nomenclature, c/o
The Natural History Museum, Cromwell Road,
London SW7 5BD, U.K. (e-mail: iczn@nhm.ac.uk).
General Article
The
PhyloCode: description and commentary
Peter L. Forey
Department of Palaeontology, The Natural
History Museum, Cromwell Road, London SW7
SBD, UK
Introduction
This essay is prompted by a seminar which I gave to the Department of
Palaeontology at The Natural History Museum, London. The size of the audience
indicated strong interest in the subject, and it was suggested that it would
make a suitable subject for an article in the Bulletin of Zoological Nomenclature.
Since I was critical of the proposals of the PhyloCode I was happy to agree
to this on the understanding that I could describe the aims and mechanics of
the PhyloCode in as neutral a way as possible while allowing myself the opportunity
of personal commentary. Therefore, this essay is divided into two distinct
parts. Readers may wish to cease reading at the end of the first part and form
their own opinions. The PhyloCode is published in preliminary form on the web
at www.ohio.edu/PhyloCode, and where possible
I take direct quotes (designated in italics - page numbers are irrelevant since
different web download programs will paginate differently) so as to avoid any
personal filters beyond the selection from the continuous text, which I encourage
reading in total.
Part
1. The PhyloCode
The PhyloCode is a new system of Biological Nomenclature which is designed
to provide rules to govern the naming of clades across all of biology. The
PhyloCode is the formalisation of the ideas of Phylogenetic Nomenclature (also
known as phylogenetic taxonomy, see below) which has been discussed in a series
of papers beginning with De Queiroz & Gauthier (1990), although many of
the issues raised by advocates of Phylogenetic Nomenclature had been discussed
long before. A near comprehensive bibliography of Phylogenetic Nomenclature
is given following the Preface at the PhyloCode website. It has been discussed,
refined and argued over in three symposia, with the formal proposals being
set out as a result of a meeting in 1998 at Harvard. The names of 26 people
are attached to the PhyloCode as an advisory group but it is unclear as to
whether all of these are signatories to all of the aims of the PhyloCode.
I should perhaps make it clear that terms such as Phylogenetic Taxonomy
and Phylogenetic Nomenclature were freely interchanged in the earlier papers
on Phylogenetic Nomenclature. The two are not the same. Phylogenetic Taxonomy
is effectively phylogenetic systematics. We can of course have Phylogenetic
Taxonomy without Phylogenetic Nomenclature.
Phylogenetic Nomenclature starts from the premise that there should be
congruence between phylogenetic hypotheses and nomenclature. At the moment
it is only in draft form, which may be perused at the web site cited above,
and the authors welcome comments as to its utility, practicality and the particulars.
At present the PhyloCode governs the naming of clades which may be previously
unnamed or correspond to taxa above the species level in other biological Codes.
Rules governing species names will be added in the future. `The PhyloCode
is designed so that it can be used concurrently with the preexisting Codes [International
Code of Zoological Nomenclature, International Code of Botanical Nomenclature
and International Code of Bacteriological Nomenclature] or (after rules
governing species names are added) as the sole Code governing the names of
taxa, if the scientific community ultimately decides that it should. The intent
is not to replace existing names but to provide an alternative system for governing
the application of both existing and newly proposed names.' (Preface paragraph
3).
Fundamentally the PhyloCode is designed to name the various parts of
the tree of life-clades (ultimately of species) - and it does this by explicit
and sole reference to phylogeny. It runs counter to what we are all familiar
with by giving no significance to ranks (Genus, Family, Order, etc.); it ignores
familiar endings such as (in the Zoological Code) -idae for family, -inae for
subfamily, -ini for tribe, etc. Such endings may be retained but they have
no hierarchical significance, so that -ini may come to prescribe a more inclusive
group than -idae.
The aims of the PhyloCode are directed toward reflecting phylogenetic
hypotheses through a system of names and it emphasizes that the usage of those
names should be explicit, unambiguous and stable: that is, they should not
change their meaning through time. The PhyloCode defines names by reference
to a hypothesised phylogeny but once a name is defined it may well be applicable
in the context of other phylogenetic hypotheses.
The principles of the PhyloCode are stated under six headings (PhyloCode
Division I. Principles):
`1. Reference. The primary purpose of taxon
names is to provide a means of referring to
taxa, as opposed to indicating their characters,
relationships, or membership.
2. Clarity. Taxon names should be unambiguous
in their designation of particular taxa.
Nomenclatural clarity is achieved through
explicit definitions.
3. Uniqueness. To promote clarity, each taxon
should have only one accepted name, and each
accepted name should refer to only one taxon.
4. Stability. The names of taxa should not
change over time. As a corollary, it must
be possible to name newly discovered taxa
without changing the names of previously
discovered taxa.
S. Phylogenetic context. The PhyloCode is
concerned with the naming of taxa and the
application of taxon names within a phylogenetic
context.
6. The PhyloCode permits freedom of taxonomic
opinion with regard to hypotheses about relationships;
it only concerns how names are to be applied
within the context of a given phylogenetic
hypothesis.'
It needs to be pointed out here that `taxon' refers to a clade or species.
If a clade it does not matter how many species are included. Thus, a clade
taxon may be what is referred to as a Genus or an Order or a Phylum under current
Linnaean Taxonomy.
The PhyloCode recognises that there are three ways of naming a clade
within a phylogenetic context and these lead to the explicit definitions referred
to in the Principles. These are illustrated in Figure 1.
Consider
a phylogeny as shown here in Figure 1 a which
shows a phylogeny leading to modern birds which
are traditionally called Aves. This lineage
may be considered as a series of cladogenic
events, each split being marked as a node which
was occupied by an ancestor `A'. The intervening
sections of the evolutionary history can be
thought of as a series of stems. During the
evolutionary history of the lineage changes
between successive nodes may be characterised
by the appearance of new characters (apomorphies)
such as, in this case, feathers. This entire
lineage will have a sister group, in this case
designated as crocodiles. It needs to be pointed
out to those readers more familiar with the
crown, total and stem group concept of Hennig
(see Jefferies, 1979) that there is partial
overlap between phylogenetic systematics and
Phylogenetic Nomenclature usage. In phylogenetic
systematics there are the concepts of crown,
total and stem groups. The crown group is the
latest common ancestor plus all its descendants
of a Recent group. The total group consists
of all species more closely related to the
crown group than to the Recent sister group
and the stem group is the extinct paraphyletic
assemblage leading up to the origin of the
crown group.
In Phylogenetic Nomenclature there is no requirement that the node specify
a crown group. In other words all crown groups are node-based groups but the
converse is not true. Under Phylogenetic Nomenclature it is perfectly possible
to recognise an entirely extinct node-based group. Similarly, all total groups
are stem-based but not all stem-based clades are total groups. With this clarified
we will continue within the terminology of Phylogenetic Nomenclature.
Under a node-based definition (Fig. 1 b) the name `Aves' is the name
given to a clade stemming from the most recent common ancestor of (say) Struthio
camelus and Corvus corax. Or, if we wish to strip out direct reference to ancestors,
it may be expressed as on the right here as the least inclusive clade containing Struthio
camelus and Corvus corax. The notation in parentheses below the
tree is a suggestion for abbreviating the definition (Aves must have Struthio
camelus and Corvus corax). Struthio camelus (the ostrich) and Corvus corax
(the raven) are called specifiers. They serve exactly the same function as
Linnaean types except their characters do not define the clade.
We could actually name as many birds to serve as specifiers as we wanted
but two is the minimum. No matter what other birds such as sparrows, gannets
or vultures are included, in this example the word Aves is constructed around
the ostrich and the raven. Clade membership may expand or contract to include
or exclude these extra taxa - and this depends upon the phylogeny - but the
ostrich and raven must always be included. So in Phylogenetic Nomenclature
we now have two types (specifiers).
In the stem-based definition (Fig. lc) Aves is named as the clade consisting
of Struthio camelus and all organisms sharing a more recent common
ancestor with Struthio camelus than with Crocodylus niloticus.
Or, again without specific reference to ancestors, as the most inclusive clade
containing Struthio camelus but not Crocodylus niloticus.
Here there is an included taxon-Struthio camelus-and an excluded taxon-Crocodylus
niloticus. So again we have a reference to specifiers but this time one
is included and one is specifically excluded. The stem-based definition states
that a taxon is more closely related to one specifier or type than another.
The shorthand notation is given beneath the diagram. So again we have two types
(specifiers).
In the apomorphy-based definition (Fig. l d) - with its abbreviation
shown below the diagram - the definition is a clade stemming from the first
species to possess the character feathers synapomorphic with that
in Struthio camelus. Or a clade diagnosed by feathers homologous
with those in Struthio camelus. Here there is one specifier taxon and one specifier
character. Two types (specifiers) but one is conceptually quite different from
the other.
In order to name a clade there must be some phylogenetic hypothesis before
us. Names are then applied in the context of that hypothesis. Should the hypothesis
change then the taxonomic content implied by a name may change but the important
point made by advocates of the PhyloCode is that the name is clear since it
based on an explicit definition (stem-, node- or apomorphy-based), it is unique
and stable since the taxon name is fixed to specifiers (taxa or characters).
To explain this, consider Figure 2 and the names Sarcopteryii and Choanata,
and take the phylogeny of the left-hand column as the phylogeny current when
the names Sarcopterygii and Choanata were coined. Under the PhyloCode the original
author of the name would have had three choices of definition (node-, stem-
or apomorphy¬based) and choices of reference taxa. In this example let
us say that the coelacanth and a frog (to represent a tetrapod) were used as
specifiers for the node-based definition (Fig. 2A I) of Sarcopterygii and the
lungfish and frog were used as specifiers for Choanata (the use of one anchor
taxon - in this case the frog - for different definitions has been advocated
by Lee, 1999a). Alternative phylogenies shown to the right (Figs. 2A2 and 2A3)
would result in different taxon membership.
Let us say that the name had been introduced under the stem-based definition:
that is, Sarcopterygii is the name given to the clade that includes the coelacanth
but not the perch (an actinopterygian) and that Choanata is the name given
to the clade including the frog but not the coelacanth (Fig. 2BI). The consequences
of subsequent phylogenetic revisions are shown to the right (Figs. 2132 and
2B3).
Lastly the same exercise (Figs. 2C,- 2C3) can be applied to apomorphy-based
naming, with the exception that there would be ambiguity about the homology
of fleshy fins under the second (Fig. 2C2) phylogenetic hypothesis so that
the name Sarcopterygii could not be unambiguously applied in this case.
There are a number of features of this exercise to notice as properties
of Phylogenetic Taxonomy. Firstly, a shift in taxon membership with changing
ideas of phylogeny is perfectly acceptable to the PhyloCode since principle
1 states that: The primary purpose of taxon names is to provide a means
of referring to taxa, as opposed to indicating their characters, relationships,
or membership [my emphasis].
Second, ideas of relationships can vary substantially (e.g. the three
theories given here) but, with one exception (Fig. 3C2, involving the apomorphy-based
definition) there will always be some position at a node or along an internode
on a phylogeny where the name Sarcopterygii will apply. That is also acceptable
since principle 1 states: The primary purpose of taxon names is to provide
a means of referring to taxa, as opposed to indicating their characters, relationships,
or membership.
The third feature is that the name is applied to a phylogeny without
reference to why that phylogeny should have been chosen. Again this is perfectly
consistent with the aims of the PhyloCode: The primary purpose of taxon
names is to provide a means of referring to taxa, as opposed to indicating
their characters, relationships, or membership.
Figure 2.
Stability and synonymy. In this diagram three
hypotheses of the relationships of coelacanths
to other vertebrates are given. That shown
in the left column after Forey (1980), centre
column after Lagios (1979), and that in the
right column after Schultze (1987). The effects
of the different taxonomic content of the names
Choanata and Sarcopterygii can be mapped across
each of the three definitions A - node based,
B - stem based (arrows on trees are a convention
used by Sereno, 1999), C - apomorphy based.
Note that in CZ and C3 the names Sarcopterygii
and Choanata could not be used because of problems
of interpreting the specifier character as
an homology (see text and Fig. 5 for discussion).
In B, `Choanata' is co-extensive with the frog
and presumably would not be named.
It is
important to notice that a change in the phylogenetic
hypothesis will cause a different change in
the taxon membership and its hierarchical relationship
to names of other clades depending on whether
the name Sarcopterygii is node- or stem-based.
Therefore, the PhyloCode makes it mandatory
that the intended definition is stated when
a name is proposed (see below).
Another phenomenon can be noted as a result of changing hypotheses when
using the node-based name. This is the fact that the hierarchical relationships
of names can reverse. Thus in Fig. 2A, Choanata is more exclusive than Sarcopterygii
whereas in Fig. 2A3 the reverse is true. Again, this is not a particular problem
for the PhyloCode since it is not concerned with rank. However, the PhyloCode
does suggest ways in which this and situations like it can be avoided: this
is done by adding exclusion clauses or qualifiers to the definition. Thus,
in this case we could say that the name Choanata is a name given to a clade
including the lungfish and the frog but excluding the coelacanth. This would
mean that in the phylogeny represented by Fig. 2A3 the name Choanata could
not be used.
The final point to be outlined concerns synonymy and homonymy. To some
extent these terms mean the same in the PhyloCode as in Linnaean Taxonomy.
Thus homonymy is an instance where the same name is used for different taxa
and synonymy is an instance where different names are used for the same taxon.
But the meaning of homonymy has an additional dimension in the PhyloCode because
of the different potential ways of defining a group (stem-, node- or apomorphy-based-see
PhyloCode Note 13.2.3). With regard to synonymy there is the possibility of
two names specifying the same taxon but since they may be defined in different
ways (e.g. stem- and node-based) they may both be valid (PhyloCode, Note 14.1.2).
Figure 3.
Truncated hierarchies and redundancy. A Linnaean
hierarchy is symmetrical where all terminal
taxa can be assigned inclusive rank. The hierarchy
on the right approximates to many phylogenetic
hypotheses where extinction (dotted lines)
or reality means that some ranks for some taxa
will be redundant.
Practicalities
and Governance
The PhyloCode will be part of
the activities of `The Society for Phylogenetic
Nomenclature (SPN), an international, non-profit
organization with no membership restrictions.
Two committees of the SPN have responsibilities
that pertain to this Code: the International
Committee on Phylogenetic Nomenclature (ICPN)
and the Regis¬tration Committee. [Note:
These organizations do not yet exist. They
will be established before the PhyloCode is
implemented].'
Thus the PhyloCode proposes a registration system whereby clade names
are submitted electronically. In order to register a name certain pieces of
information need to be provided (those marked with an asterisk being mandatory
and others optional):
`Definition type* (node based, stem based, apomorphy based, other. ) Phylogenetic
definition*
List of specifiers*, at least two being mandatory
Qualifying clause
Reference phylogeny (bibliographic reference, URL, or Accession number in
public repository) '
These then, are the aims and basic workings of the PhyloCode. Of course,
there are many other provisions in the Code designed to streamline the naming
process (orthography and authorship) and to deal with particular situations
(e.g. hybrids). I encourage all to visit the PhyloCode website to read the
full text.
Part
2. Commentary
Phylogenetic Nomenclature already
has a history, with the main arguments and
suggestions for its implementation having been
put forward in a series of papers (e.g. De
Queiroz & Gauthier, 1990, 1992, 1994; Rowe & Gauthier,
1992; Lee, 1998, 1999a,b; Sereno, 1999; Cantino,
2000). Counterviews have been expressed in
others (Liden & Oxelman, 1996; Dominguez & Wheeler,
1997; Moore, 1998; Benton, 2000; Nixon & Carpenter,
2000).
Supporters of Phylogenetic Nomenclature argue that because Linnaean Taxonomy
is based on the concept of rank it is ill-suited to expressing our changing
ideas of phylogenetic relationships between species. Rank is problematic because
the appli¬cation of a rigid rank system leads to redundancy and instability.
Redundancy is introduced because the Linnaean hierarchy is equidistant: that
is to say, every taxon is included in a continuity of ranks from Genus to Kingdom
(although this is not stated as mandatory in the Zoological Code). This may
be perfectly satisfactory should the phylogeny be perfectly symmetrical (Fig.
3 left). But reconstructed phylogenies are not like this, either because history
is genuinely asymmetrical or because of extinctions; they can appear to us
as truncated hierarchies (Griffiths, 1973). This means that there are empty
ranks (Fig. 3 right). Or to express this in another way: in some parts of the
phylogenetic tree the Family rank is equivalent in scope and content to the
Order rank elsewhere. Thus, confining oneself to the Recent world, ranks can
become redundant in monospecific groups; for example there is nothing more
implied by the Family Hominidae than by the Genus Homo or the Species sapiens.
Rank has been used to imply some level of morphological divergence either
in amount or kind. The boundaries of ranks (Genus, Family, Order, etc.) are
traditionally, and still usually, judged on morphological divergence. We expect
the morphological gaps between Orders to be larger than those between Families,
and in turn the latter to be larger than the gaps between Genera. At the same
time we expect the variations within Orders than to be greater than those within
Families, and these to be greater than the variations within Genera. How much
variation and how large a gap is appropriate for families, genera, etc. is
usually unstated and is indeed undefinable.
Rank has also been used to signify the kind or quality of divergence
to ascribe rank. The action of separating birds as Class Aves equivalent to
Class Reptilia is only because of the kinds of characters by which
birds differ from reptiles. Birds have characters such as wings, feathers and
air sacs that enable them to exist in a different adaptive zone. These are
deemed by mutual consent to be Class characters and because of this the paraphyletic
rubble left behind - the reptiles - also has to have Class status.
So, I have some sympathy with Phylogenetic Nomenclature in the desire
to seek a rank-free classification. But there are ways around the problem which
do not involve the adoption of a PhyloCode (e.g. Crane & Kenrick, 1997).
This is the annotated Linnaean system which by the use of a few conventions
(Nelson, 1974; Patterson
& Rosen, 1977; Wiley, 1979) can absorb
the problems caused by rank yet allow those
who wish to retain rank to do so for their
own purposes.
It must be remembered that the abolition of ranks can have some rather
unfortunate consequences for many people who compile `diversity indices' based
on generic counts or family counts. Here, abolition of rank would immediately
affect some palaeontologists and many people studying biodiversity. It is apparently
common practice (Dr Sandy Knapp, pers. comm.) in biodiversity inventories to
simply note the existence of a representative of a family or genus, because
the organism may be new and can only be recognised initially on family characters.
With rank abandoned counts are abandoned.
TypeslSpecifiers
It is difficult to see why Phylogenetic Nomenclature has adopted the
new term `specifier(s)' when, in reality the `type concept' is still with us,
only in a more complicated fashion. The types in Phylogenetic Nomenclature
are the specifiers (species, specimens or synapomorphies) coupled with a
phylogenetic hypothesis. In practice there is no difference between specifiers
and types, except that in Phylogenetic Nomenclature it is necessary to cite
at least two for every name: node-based (A + B), stem-based (A F- B), apomorphy-based
(synapomorphy a in A). Under Phylogenetic Nomenclature rules we have the additional
complication of the phylogeny, because the name is only to be used within the
context of a phylogeny (PhyloCode, Division 1, Principles, number 6).
As De Queiroz & Gauthier (1992) pointed out, it is always possible to make
a mistake about the contents of a clade (the taxa included) and the diagnostics
(the characters by which it is recognised) - but it is not possible to make
a mistake about the phylogenetic definition. Because of the way names are constructed
under phylogenetic taxonomy this must be true. However, it needs to be pointed
out that it is not the phylogeny that is important but only the part of
the phylogeny that is relevant to the name (that portion which includes the
specifiers). Other taxa which may have been part of the original phylogeny
when the name was erected are free to wander in and out of the named clade.
The specifiers and the part of the phylogeny used in erecting the name suffer
from the same problem as Linnaean types - they are acting as focal points.
How widely or narrowly their naming influence spreads is entirely at the whim
of systematists erecting new phylogenies in precisely the same way as in Linnaean
taxonomy. Therefore the substituion of types by specifiers (PhyloCode,
Preface) seems completely unnecessary.
Of course in Linnaean taxonomy the type concept is ultimately tied to
characters, attributes of specimens which we can see.
When to
name
One of the objections raised against Linnaean Taxonomy is that it is
often difficult to name clades without causing a cascade of name changes through
rank-ending changes. This, phylogenetic taxonomy claims, will result in clades
for which there is much evidence being un-named, and taxonomy becoming out
of step with phylo¬genetic knowledge. Therefore, the unrestricted ability
to name clades is seen as an advantage (PhyloCode, Preface). However,
phylogenetic taxonomy also acknowledges that not all clades need to be named.
At first sight this commonsense view may seem odd, considering that the paramount
objective of phylogenetic nomenclature is to name clades. Some reasons for
naming a clade are given as recommendations in the PhyloCode: `Criteria
that influence the decision whether to name a clade include level of support,
phenotypic distinctiveness, economic importance, etc.' (Preface, paragraph
6). I am not sure what 'etc.' covers, but taking the three that are given I
can make some comment.
Level of support. This means that the PhyloCode
recommends that we only name clades that are judged to be soundly
based with good support. What might this mean? Numbers of synapomorphies,
Bremer support, bootstrap support, jackknife support (first order
jackknife or second order jackknife), consistency index, retention
index, rescaled consistency index, resistance to successive weighting,
heavy implied weighting scores, and so on.
Phenotypic distinctiveness. This seems to me to be a curious
criterion to use, since much of the PhyloCode's objection against Linnaean
Taxonomy is based on the fact that the classical type system does not specify
how far from the type the name applies, i.e. how distinctive taxa have to be
from a name-bearing type before they become a new genus, family or whatever.
Yet, here the PhyloCode seems to be saying the same thing-only in relation
to clades. If we are only going to name clades according to phenotypic distinctiveness
then this seems to advocate an apomorphy-based definition. We name clades with
reference to one or more apomorphies which are judged to be `significant'.
However, apomorphy-based naming is less favoured than the other two definitions
because of the subjective assessments of characters; this has been emphasised
by Rowe & Gauthier (1992) using the naming of Mammalia as an example.
Economic importance. Well, there are more than enough measures
here (e.g. contribution to Gross National Products; financial impact on social
conditions, health and welfare; cost-benefit for international aid) but how
these are going to be evaluated is difficult.
The point is that the advocates of phylogenetic taxonomy really do not
have any more precise reasons for naming a group than do followers of Linnaean
Taxonomy and to include advice in the PhyloCode registers a precision which
is both unneccesary and undesirable.
How to
name
The kind of definition which should
be applied in any given clade has been discussed
in the context of phylogenetic taxonomy on
many occasions (e.g. Lee, 1999b; Sereno, 1999).
Here we meet a curiously illogical rationale,
since the reason for choosing one kind of definition
over another (node-based, stem-based, apomorphy-based)
is apparently in order to `stabilize the taxonomic
content of a taxon more than another in the
face of local changes in relationships' (Sereno,
1999, p. 329). However, taxonomic content is
not the primary purpose of Phylogenetic Nomenclature
(PhyloCode, Division 1. Principles)
and it is therefore unclear why this should
be an issue. However, anyone practicing Phylogenetic
Nomenclature must specify which definition
is to be used (see Practicalities and Governance above)
and therefore some decision has to be made.
Several suggestions have been put forward and,
for me, the most thorough discussion of this
subject is that by Sereno (1999) who advises
in which circumstances it may be best to use
node-, stem- or apomorphy-based names as well
as offering advice on selecting specifiers
(types). Despite all the discussion around
this subject, the final decision must rest
on some estimate as to the resolution, the
strength of phylogenetic signal and the potential
durability of the phylogeny (crudely put: will
those taxa stay in place with the introduction
of new data?). In other words some evaluation
of the quality of the phylogeny is required.
Not surprisingly, Phylogenetic Nomenclature
is mute in offering guidelines since there
are no agreed criteria amongst the systematic
community at large'. Therefore while the name
of a taxon may well remain stable the applicability
of that name within classifications may be
decidedly unstable.
There are instances where names can be considered unstable. PhyloCode
(Article 15. Conservation) allows that, under certain circumstances involving
synonymy and homonomy, authors may apply to the International Committee on
Phylogenetic Nomenclature to have names conserved and suppressed. Thus, suppose
that with reference to Figure 2A2 Sarcopterygii had been defined as node-based-Sarcopterygii
(coelacanth and frog)-and Gnathostomata had been similarly defined (coelacanth
and frog). These are clearly synonyms. Let us further imagine that even though
date precedence favoured Gnathostomata common usage suggested Sarcopterygii
as a more appropriate name (in principle this is similar to the `prevailing
usage' rules of the Zoological Code). As I understand the PhyloCode, Sarcopterygii
could be conserved and Gnathostomata suppressed. However, a later author might
resurrect Gnathostomata by using a different definition (e.g. Gnathostomata
[frog <--- lamprey]). This is hardly stability.
Linking a name with a particular phylogeny also leads us into theories
of homology, since it is precisely such theories which enable us to recognise
the phylogeny in the first place. This is not without difficulty for phylogenetic
taxonomy, as may best be explained with reference to the apomorphy-based definition.
Historically, in Linnaean Taxonomy apomorphy-based names are those which have
caused most confusion, as Rowe & Gauthier (1992) point out in the context
of the naming of Mammalia. However, of all of the definitions advocated by
phylogenetic Nomenclature, apomorphy-based
naming is the only one which makes specific reference to characters
observable in the objects of study (i.e. organisms). But even here there are
problems because characters are homologies and homologies are theories. This
aspect may not be fully apparent to those taxonomists unfamiliar with phylogenetic
systematics. Consider an apomorphy-based definition which may be proposed as
`Tetrapoda is the name given to the clade consisting of all those animals with
fingers and toes homologous with those in Rana esculenta'. The problem
arises over the word homologous. In phylogenetic systematics an homology is
a theory and is equivalent to synapomorphy (shared derived character). Let
us say that we had arrived at the phylogeny of organisms shown in Figure 4
where lungfishes and caecilians lack fingers and toes whereas lacertilians
(mostly), urodeles and Rana esculenta have them. There are two ways
in which we may imagine the characters `fingers and toes' to have evolved given
this phylogeny. Or, in cladistic terminology, there are two ways of optimising
this character on this tree. We could suggest that `fingers and toes' was gained
in the common ancestor of the group lacertilians + Rana esculenta and
subsequently lost in caecilians. This involves two evolutionary steps (or transformations):
one gain and one loss (Fig. 4a). In this case `fingers and toes' is an homology
(shared derived character or synapomorphy) which has been subsequently lost
in some members (caecilians) of this group. This type of optimisation is called
accelerated transformation (ACCTRAN) because it places the first transformation
- no fingers and toes --> fingers and toes - at the most inclusive hierarchical
level on the tree. One alternative is shown in Figure 4b. Here, it is assumed
that `fingers and toes' is a character that was gained twice - once in lacertilians
and again in the common ancestor of the group urodeles + Rana esculenta.
This optimisation is called delayed transformation (DELTRAN) because it delays
the transformation to the most exclusive positions in the hierarchy. In this
case `fingers and toes' is not regarded as a synapomorphy because it has arisen
twice and therefore cannot be considered an homology, and presumably would
not be used as an apomorphy-based specifier. However, these two theories of
character evolution are equally parsimonious and we would need additional information
to choose one alternative as more likely than the other. In order for there
to be no ambiguity we need a qualifying phrase to be added to our apomorphy-based
definition of Tetrapoda as `all those animals sharing fingers and toes homologous
with those of Rana esculenta under the optimising procedure of accelerated
transformation'. This is not a particularly utilitarian usage. I admit the
example may be contrived: but if fingers and toes were substituted by a particular
gene sequence then it may not be easy to argue for or against homology. Perhaps
it will be necessary to restate the apomorphy-based definition of Tetrapoda
as `all those animals sharing fingers and toes homologous with those of Rana
esculenta under any optimising procedure'.
------------------------------------------------------------------------------------------------
'There have
been many indices devised to try to assess
the support of cladograms/phylogenetic trees
such as Bremer support, Bootstrap, Jacknife,
consistency index, retention index, rescaled
consistency index, permutation tail probability
tests (and derivatives). All these have their
fields of applicability but they are really
designed to test the strength of the hierarchial
signal, not the stability of the phylogenetic
hypothesis which may only be done a posteriori with
more data. And the problem is compounded if
the analysis is carried out under Maximum Likelihood
methods - as is often the case with molecular
phylogenies - because here the tests applied
are undertaken in the context of a particular
model of character evolution.
--------------------------------------------------------------------
Pain
- no gain
The PhyloCode proposes that biologists will gain clarity, efficiency
and stability when accepting its premises and adopting its methodology. Advocates
of the PhyloCode also claim that these gains are likely to be appreciated by
those not interested in phylogeny or nomenclature (Cantino et al., 1999); it
is clear that the intention is for the PhyloCode to be understood and used
by non-systematists. We need, therefore, to assess what that gain is and at
what cost it is to be achieved within the context of biology in general. Cost
can only be measured against some standard and therefore some comparison with
Linnaean Taxonomy is essential. With respect to clarity and stability there
may be no difference between Phylogenetic Nomenclature and Linnaean taxonomy.
Within Phylogenetic Nomenclature a name is stable within the context of its
specifiers. But so are Linnaean names based on types. The specimen BMNH 1853.11.12.111
is and will remain the name-bearing type (lectotype) of Clupea harengus,
just as Struthio camelus and Corvus corax could be regarded
as specifiers of the node-based Aves under the PhyloCode. Under Linnaean taxonomy
suprafamilial names do not have formal name-bearing types but they may be said
to have specifiers. In 1861 T.H. Huxley erected the name Crossopterygii for
an assemblage of fossil fishes including Polypterus, Gyroptychius,
Holoptychius, Osteolepis, Dipterus, Phaneropleuron and Macropoma.
Two or more of these fishes are the specifiers, if you like, of the name Crossopterygii
Huxley, 1861. In the years immediately following Huxley's work the content
or membership of the Crossopterygii changed dramatically. This was not surprising
because some or all of these fishes were implicated in the ancestry of tetrapods
and therefore authors were struggling with a paraphyletic group. However, any
scientist foolish enough to struggle with such a group (and I count myself
amongst them) is forced to go back to Huxley (1861) to learn the membership
of the group and the observations which were used in its recognition. Under
phylogenetic taxonomy exactly the same would happen. We cannot gather any relevant
details directly from the name Aves (Struthio camelus and Corvus
corax) Joe Doe. We are forced to examine the contents of the clade to
understand its membership and presumably we would also be interested in how
it was recognised.
When Gunther (1871) examined Crossopterygii Huxley, 1861 he decided that
the relationships of the included taxa were not as Huxley opined. The phylogeny
had changed and so had the membership of Crossopterygii. But this is precisely
what happens under Phylogenetic Nomenclature also (e.g. Sarcopterygii in Fig.
2). If we want to understand the systematic history of a particular taxon we
still have to examine all of the phylogenies under which that name has been
used because the name itself may be compatible with more than one phylogenetic
hypothesis. Thus the claim by phylogenetic taxonomy for clarity and stability
within the context of why systematists need the name in the first place is
at best illusionary and at worse misleading. There is nothing to be gained.
The pain is administered in several ways. First, for the sake of clarity
new names may have to be coined for very familiar groups. The PhyloCode is
very clear to point out that this need not be so and suggests that existing
names can be redefined under Phylocode conventions by appending a suffix `[P]',
meaning that this name is to be used in the sense of phylogenetic taxonomy
(Cantino, 2000, p. 87). While this is perfectly feasible, we may ask - will
the redefinition be understandable to the many non-systematists who use classifications
as their comparative framework? The PhyloCode (Article 11.8) does insist that
`when a clade name is converted from a preexisting genus name or is a new or
converted name derived from the stem of a genus name, the definition of the
clade name must use the type species of that genus as an internal specifier.'.
However, it makes no recommendations as to suprageneric names. Things can go
awry. For instance, Laurin (1998) redefined the taxon Anthracosauria under
Phylogenetic Nomenclature such that it no longer included its Linnaean type
genus Anthracosaurus. To use the same name in two completely different
contexts will surely lead to confusion, and it puts the onus on the non-systematist
to find out the difference or overlap in the meaning of the names. As taxonomists
we are hardly serving the wider biological community by this duality and potential
confusion.
Second, the PhyloCode is agnostic about characters, relationships, or
membership. However, this is precisely the important information which may
be of importance to comparative biologists. Thus the retrieval of information
may not be as easy as the PhyloCode suggests.
Third, changing hypotheses of relationship will mean that names are used
and disused according to the phylogeny in fashion at that time (in Linnaean
taxonomy the name will remain the same but the membership may change). This
is hardly stability.
Fourth, the PhyloCode names clades, each of which is defined as `a monophyletic
group of species' (PhyloCode, Preface). This means that only monophyletic
groups be named (there is no other kind of clade). While this is a desirable
endpoint we are very far from achieving that phylogenetic resolution. There
remain vast branches in the tree of life where monophyly has yet to be demonstrated.
Phylogenetic Nomenclature will leave these assemblages of taxa un-named. I
find myself in the rather uncomfortable position of being one who agrees strongly
that monophyletic groups are the only real biological entities worth consideration
and I would never argue for the retention of paraphyletic taxa. But I am also
mindful of the fact that for many biologists potentially non-monophyletic groups
(e.g. Bryophyta) still serve a useful purpose for their own reasons of communication
(say, in ecological studies). Thus we will still have to live with Linnaean
names alongside PhyloCode names. The annotated Linnaean system (Wiley, 1979)
can cope with phylogenetic uncertainty to satisfy the systematists without
denying names that may be useful elsewhere.
Fifth, adoption of the PhyloCode can and probably would lead to a rapid
inflation of names because, quite naturally, individual workers will wish to
name the hard-won results of their own phylogenetic investigations. I see this
most likely to happen in two areas; molecular systematics and with newly discovered
fossil taxa. With respect to the latter de Queiroz & Gauthier (1992, p.
457) recognised this but claimed that since it is palaeontologists who are
most concerned with phylogenies they should live with this problem, which they
dismissed as minor since `there are already more taxon names than anyone can
remember - then naming clades seems preferable to leaving them unnamed . .
.'. Thus, in one sense, phylogenetic systematists get what they deserve. But
in another sense, phylogenetic systematists are not serving the wider biological
community by introducing a plethora of names, each with their own definitions
which need to be understood before they can be used by others.
Conclusion
The intention of the PhyloCode is to name clades and it is therefore
free of empirical content (with the possible exception of the apomorphy-based
definition). In trying to name hypotheses the PhyloCode puts the onus on the
users of the names to assess the confidence we may have in one particular clade
or another before selecting a name that matches that choice. Users of Linnaean
taxonomy are, of course, forced to do the same, but no name changes need be
required. The alleged clarity, efficiency and stability claimed by the PhyloCode
do not stand critical examination and it needs to be asked what exactly has
been gained. More importantly the biological community will have to judge whether
the alleged gains are worth the undoubted pain.
Acknowledgements
I wish to thank Philip Cantino and Kevin de Queiroz, both of whom were
kind enough to read the first part of this essay and correct some of my errors
of fact about the PhyloCode. Philip Cantino's further comments on my commentary
were also very useful to me and are acknowledged here-of course, disagreements
still remain.
References
Benton, M.J. 2000. Stems,
nodes, crown clades, and rank-free lists:
is Linnaeus dead? Biological Reviews, 75:
633-648.
Cantino, P.D. 2000. Phylogenetic nomenclature:
addressing some concerns. Taxon, 49:
85-93.
Cantino, P.D., Byrant, H.N., De Queiroz, K., Donoghue,
M.J., Eriksson, T., Hillis, D.M. & Lee, M.S.Y. 1999.
Species names in phylogenetic nomenclature. Systematic Biology, 48:
790-807.
Crane, P.R. & Kenrick, P. 1997. Problems
in cladistic classification: higher-level relationships in land
plants. Aliso 15: 87-104.
De Queiroz, K. & Gauthier, J. 1990. Phylogeny
as a central principle in taxonomy: phylogenetic definitions
of taxon names. Systematic Zoology, 39:
307-322.
De Queiroz, K. & Gauthier, J. 1992. Phylogenetic
taxonomy. Annual Review of Ecology and Systematics, 23:
449-480.
De Queiroz, K. & Gauthier, J. 1994. Towards
a phylogenetic system of biological nomen¬clature. Trends
in Ecology and Evolution, 9: 27-31.
Dominguez, E. & Wheeler, Q.D. 1997. Taxonomic
stability is ignorance. Cladistics, 13:
367-372.
Forey, P.L. 1980. Latimeria: a paradoxical
fish. Proceedings of the Royal Society of London, B, 208:
369-384.
Griffiths, G.C.D. 1973. Some fundamental problems
in biological classification. Systematic Zoology, 22:
338-343.
Gunther, A.C.L.G. 1871. Description of Ceratodus,
a genus of ganoid fishes, recently discovered in the rivers of
Queensland, Australia. Philosophical Transactions of the
Royal Society of London, 161: 511-571
Huxley, T.H. 1861. Preliminary essay upon the
systematic arrangement of the fishes of the Devonian epoch. Memoirs
of the Geological Survey of the United Kingdom, Decade
10: 1-110.
Jefferies, R.P.S. 1979. The origin of chordates
- a methodological essay. Pp. 443-477 in House, M.R. (Ed.), The
origin of major invertebrate groups. Systematics Association
Special Volume Series, no. 12.
Lagios, M.D. 1979. The coelacanth and the Chondrichthyes
as sister groups: a review of shared apomorph characters and
a cladistic analysis and reinterpretation. Occasional papers
of the California Academy of Sciences, 134:
25--44.
Laurin, M. 1998. The importance of global parsimony
and historical bias in understanding tetrapod evolution. Part
1. Systematics, middle ear evolution and jaw suspension. Annales
des Sciences Naturelles, Zoologie et Biologic Animale 19:
1-42.
Lee, M.S. 1998. Phylogenetic uncertainty, molecular
sequences, and the definition of taxon names. Systematic
Biology, 47: 719-726.
Lee, M.S.Y. 1999a. Reference taxa and phylogenetic
nomenclature. Taxon, 48: 31-34.
Lee, M.S.Y. 1999b. Stability of higher taxa
in phylogenetic nomenclature - some comments on Moore (1998). Zoologica
Scripta, 28: 361-366.
Liden, M. & Oxelman, B. 1996. Do we need
`phylogenetic taxonomy'? Zoologica Scripta, 25:
183-185.
Moore, G. 1998. A comparison of traditional
and phylogenetic nomenclature. Taxon, 47:
561-579.
Nelson, G. 1974. Classification as an expression
of phyogenetic relationships. Systematic Zoology, 22:
344-349.
Nixon, K.C. & Carpenter, J.M. 2000. On the
other `Phylogenetic Systematics'. Cladistics, 16:
298-318.
Patterson, C. & Rosen, D.E. 1977. Review
of ichthyodectiform and other Mesozoic teleost fishes and the
theory and practice of classifying fossils. Bulletin of the
American Museum of Natural History, 158:
81-172.
Rowe, T. & Gauthier, J. 1992. Ancestry,
paleontology, and definition of the name Mammalia. Systematic
Biology, 41: 372-378.
Schultze, H.-P. 1987. Dipnoans as sarcopterygians.
Pp. 39-74 in Bemis, W.E., Burggren, W.W. & Kemp,
N. (Eds.), The biology and evolution of lungfishes.
Alan R. Liss, New York.
Sereno, P.C. 1999. Definitions in
phylogenetic taxonomy: critique and rationale. Systematic
Biology, 48: 329-351.
Wiley, E.O. 1979. An annotated Linnaean hierarchy,
with comments on natural taxa and competing systems. Systematic
Zoology, 28: 308-337.
