WHAT IS THE WRITER OF A FLORA TO DO? EVOLUTIONARY TAXONOMY OR PHYLOGENETIC SYSTEMATICS?

28 Pages·2002·17.7 MB·English

by  George M Diggs

WHAT
THE WRITER OE A ELORA TO
DO?
IS
TAXONOMY
EVOLUTIONARY OR PHYLOGENETIC
SYSTEMATICS?
George L Lipscomb
M.Diggs, Barney
Jr.
Department
of Biology
Austin College
&
Botanical Research of Tex
Institute
[email protected].
ABSTRACT
ant clasbidcation and nomenclature are in a continuing
opposing schoolb-1) traditional taxon omists support!
d 2) cladists supporting "phylogenetic systematics" or "c
spanned
3veib> that has several decades relatively
little
e.amificationslorflor.stics.Thetwog
wr
plications of the controversy for the ers of floras, an
m
guments a format accessible to a dive se audience of b
some we
e of the issues have confronte inourfloristi
m
used floras (allowing paraphyletic groups but eliminating polyphyletic groups des
3e all
omenclatural instability). Further, these floras should incorporate information on newly
phylogenetic relationships (even too preliminary, tentative, or inappropriate for noir
i if
/arias decadas, se ha prestado relativamente poca atencion a las
un
letas de este articulo son 1) dar enfasis especial a las implica
de floras, y 2) ofrecer una vision general de algunos de los argun
is
Examins
diversa de botanicos interesados en
liencia la loristica.
f
como
Ds confrontado en nuestro trabajo loristico, discutimos h
f
ambos
s y debilidades de metodos, e indicamos lo que cree
Argumentamos
metodo de escribir floras en la actualidad. c
to,
INTRODUCTION
taxonomy
Plant the science that deals with the identification, nomenclature,
is
and The term plant systematics systematic botany)
classification of plants. (or
often used synonymously with plant taxonomy done but some-
(as here),
is is
mainly
times has the connotation of using recently developed techniques or
approaches such as chromosomal studies, electron microscopy, molecular biol-
From
ogy, or cladistics to answer questions about plant relationships. the defi-
taxonomy primary
nition of plant follows that the goals of the discipline are
it
and
1) identify describe all the various kinds of plants;
2) develop a system of naming plants International Code of Botanical No-
[e.g.,
menclature (Greuter et 2000) or potentially a future version of the
al.
PhyloCode
(2002)];
common
3) arrange plants with characteristics into groups that reflect their
evolutionary relationships (Lawrence 1951; Porter 1967; Radford et 1974;
al.
&
Jones Luchsinger Judd 2002).
1986; et 1999,
al.
taxonomy
In terms of nomenclature, the goal of plant has been to develop a
uniform, practical, and stable system of naming plants—one that can be used
by both plant taxonomists and others needing a way to communicate precisely
and retrieve information about plants. In the words of Stevens (2002), "The value
how
any naming
system conventions allow
of effectively establishes that
is
it
communicate and
people develop
to to their ideas.
..."
common
Regarding classification, the goal has been to arrange plants with
characteristics into groups that reflect their relationships— in other words, to
maximum
scheme
develop useful— conveys
a of classification that that in-
is
formation and has predictive value. Since the time of Darwin, a primary goal of
plant taxonomists has been to reflect phylogeny or evolutionary history in their
One
systems of classification. There are several reasons for this. that taxono-
is
want
mists their classification system to reflect the reality of the evolutionary
history of on earth. Second, a system that reflects evolution should have
life
maximum
predictive value and usefulness (since related species should share
common
similarities due to descent). While this basic evolutionary approach
agreed on by virtually botanists, in recent years there has been heated
is all
debate between two main schools of taxonomists:
(Brummitt taxonomy"
classification" 1997) or "evolutionary Traditional or evo-
lutionary taxonomists, while attempting have system based
to a classification
on amount
evolutionary relationships, also try to reflect the of evolutionary
change undergone by
groups. In addition, they try to incorporate other goals,
and Brummitt
including practicality stability into the classification system (see
The names
1997 for a detailed discussion of traditional classification). Linnaean
classification or Linnaean taxonomy (Stuessy 2000; Forey 2001, 2002; Nicolson
2002), are perhaps inappropriate since the system very different from that
is
established by Linnaeus. In more recent version perhaps better called
its it is
&
taxonomy"
"evolutionary (Sanders Judd 2000) or "evolutionary systematics"
(Grant 2001b) signifying the attempt to reflect evolutionary relationships. Grant
acronym TIES
(2001b) uses the "to include the two subschools of traditional
When
taxonomy and from
evolutionary systematics." considered the standpoint
of nomenclature, this a system that incorporates binomial nomenclature (two-
is
names genus name and and
part scientific consisting of a specific epithet) a
hierarchy of formal ranks family, genus, The nomenclatural applica-
(e.g., etc.).
viewpoint has been and
tion of this referred to as the "L-code" principles are
its
embodied Code Nomenclature
in the International of Botanical (Greuter
et
al.
2000).
whose method
2) cladists, of constructing phylogenies derived from the ideas
is
German
of the entomologist Willi Hennig, practicing "phylogenetic
systematics"or "cladistic classification" (referred to as "cladonomy" by Brummitt
1997) based explicitly and solely on phylogenetic relationships. In other words,
the overriding goal that classification should reflect the branching patterns
is
of evolution. should be noted that in a clade-based classification and result-
It
ing nomenclature system, there are no formal ranks, including family or genus,
&
and no formal binomial nomenclature Queiroz Gauthier Liden
(de 1992;
et
1997; Cantino 1998; Brummitt 2002). The nomenclatural application of this
al.
viewpoint has sometimes been and em-
referred to as the "P-code" ideas are
its
bodied PhyloCode
in the (2002).
Currently classification and nomenclatural systems are in a state of flux between
two main opposing camps— both which attempt
these of to reflect evolution-
made
ary relationships. Those practicing cladistic systematics have major con-
tributions to our understanding of plant evolution, and have brought about some
long overdue changes. In fact, some of their methodology has contributed to a
Some
well-recognized revival in taxonomy/systematics. of the most evident
examples of this are the incredible breakthroughs in knowledge of plant rela-
tionships resulting from molecular phylogenetics. should also be noted that
It
which
there significant variation in the extent to various "phylogenetic" sys-
is
approach—
tematists follow all of the implications of the cladistic most, for ex-
may
ample, use binomial nomenclature even though they disagree with
still
it
same
on theoretical grounds. At the time, the system of nomenclature (bino-
mial, and hierarchical classification that has developed over the past 250
etc.)
and and communi-
years has served continues to serve the botanical broader
ties well. There are thus positive aspects to both of these approaches.
Because of major differences in philosophy and methodology, the classifi-
cation systems produced by proponents of evolutionary taxonomy and cladis-
are often quite different. not unexpected then that the two conflicting
tics It is
&
viewpoints have produced vigorous and heated debate Nixon Carpen-
a
(e.g.,
ter 2000), which has even been referred to as a "maelstrom" (Benton 2000). The
proposed approach of phylogenetic classification has certainly not "mostly been
community"
politely accepted by the systematic as stated by Schander (1998).
and
In fact, the tone of a few of the articles discussions (on both sides of the
modern
argument) has been surprisingly impolite by the standards of scien-
discourse, with Webster (2002) referring to the arguments as "an ideologi-
tific
cacophony bombast and There voluminous on
cal of invective." a literature
is
the subject, including numerous recent articles about the different taxonomic
and nomenclatural approaches Moore Stevens Diggs 1999
1998; 1998; et
(e.g., al.
Cantmo
(Appendix Mishler Benton 2000; 2000; de Queiroz 2000;
1999;
6);
& &
McNeill 2000; Nixon Carpenter 2000; Sanders Judd 2000; Stevens 2000;
&
Cantmo
Stuessy 2000, 2001; Withgott 2000; de Queiroz 2001; Grant 2001a,
2001b; Langer 2001; Lee 2001; Pennisi 2001; Berry 2002; Brummitt 2002; Forey
&
Sym-
2002; Kress DePriest 2002; Nicolson 2002; Stevens 2002; Webster 2002).
posia and workshops have also been held (XVI International Botanical Con-
Museum
March
gress—August Smithsonian's National Natural History—
1999; of
Hunt Documentation-June and new
2001; Institute for Botanical 2002), a sys-
tem of nomenclature has been proposed (PhyloCode 2002). However, few au-
thors or discussions have specifically addressed the special problems faced by
&
writers of floras (but see Stevens 1998, Sanders Judd 2000, and Berry 2002).
many
This controversy actually multifaceted, with quite different aspects
is
names
are taxon defined?-Stuessy 2000, 2001, de Queiroz 2000, de Queiroz
(e.g.,
which more
Cantino 2001; system will ultimately be stable?-Forey 2002)
(Sa:
we
that are beyond the scope of the discussion here. In this paper are focusing
on
the implications
for lonstics.
f
IMPLICATIONS FOR FLORISTICS
As
writers of a flora (the Illustrated Flora of East Texas project— Diggs et in
al.
we
prep.; www.easttexasflora.org), and our co-authors are faced with the o^ts-
tionoiwhattype andriornendaturalcoriceptsshouldhefoUowed
classification
of
roughly North
in a large regional flora 3,300 species or 1/6 the species in
(ca.
We
America
north of Mexico). are acutely interested in this question, because
we must and
as loristicians translate synthesize a variety of types of botanical
f
research, both practical and theoretical, into a form usable by a very diverse
audience—one that ranges from professional taxonomists and other scientists
whom
many
and
to lay botanists, students, interested amateurs, of are unac-
we
quainted with taxonomic methods. In believe that addressing the needs
fact,
one most important
of diverse users of the tasks of loristicians. Further, de-
is f
veloping a "general-purpose system" that effectively addresses the needs of
and
multiple users often considered to be the "historical continuing func-
is
taxonomy whole The answer
tion" of as a (Cronquist 1987). to the question of
who
what
type of concepts should be followed varies greatly depending on one
The most would
asks. conservative voices say that the traditional system of
nomenclature, ranked system
a hierarchical of classification, traditional fami-
should be used due to both theoretical and practical considerations.
lies, etc.,
Some would
even argue that clearly polyphyletic traditional families
(e.g.,
Liliaceae in the broad sense) should continue to be used in floras since this a
is
very useful and practical approach. The most extreme voices on the other side
extreme cladists) would say that no set categories should be recognized
(i.e. (e.g.,
no families, no genera; instead, only supportable clades), only monophyletic
common
and
groups ancestor descendants; should be given
(= a all its Fig. 1)
taxonomic recognition no paraphyletic groups should be allowed— cur-
(i.e.,
many
and
rently genera families are paraphyletic; paraphyletic groups are de-
common
and
fined as those containing a ancestor some, but not of de-
all, its
scendants). Further, the extreme cladists argue that binomial nomenclature
should be replaced (since genera have no objective there can be no ge-
reality,
neric names and hence no binomials; only clade-based names should be used).
our previous (Shinners Mahler's IHustrnted Flora o/North
Interestingly, flora
(S-
Central Texas— Diggs et 1999) was criticized by individuals from both ex-
al.
tremes. Despite having more information on cladistics (lengthy appendix dis-
numerous any
cussing the issue, discussions in family synopses) than other
we know Hickman
Rhodes
regional or state flora of 1993; Yatskievych 1999;
(e.g.,
& &
Block 2000; Wunderlin Hansen 2000), we were criticized for not applying
some
the cladistic approach throughout the flora. Likewise, conservative bota-
nists were disturbed by decisions such as lumping some groups Najas into
(e.g.,
and
the Hydrocharitaceae), splitting others Senecio into Senecio Packera),
(e.g.,
and following an alphabetical rather than traditional sequence (showing sup-
a
and
posed relationships) of families genera logical impossibility since a writ-
(a
One
ten flora linear and evolution a branching process). thing to keep in
is is
mind when any
discussing this clash of viewpoints the realization that sys-
is
tem of classification, nomenclature, and written presentation will be an im-
perfect reflection of the complexity represented by the evolutionary history of
on words Benton "phylogeny
life earth. In the of (2000), is real, classification is
While we now and
not." have access to increasingly sophisticated diverse sources
of data, factors such as extinction, an incomplete fossil record, and the com-
Wagner
plexity of evolutionary processes reticulate evolution— 1954; see
(e.g.,
discussion below) will prevent us from producing completely accurate phylo-
genetic reconstructions. Despite these limitations, as floristicians attempting
we
to produce a useful flora, have to use approaches to classification and no-
menclature and complex
that best reflect a diverse set of needs.
generally agreed that the primary goal of a flora to allow identifica-
It is is
GROUPS
POLYPHYLETIC
MONOPHYLETIC GROUPS
Groups
Paraphyuetic
Diagrammatic
representatioi
I .
of the plants treated. However, there nber of secondary goals. Ac-
ar(
1
and
ding to Sanders Judd (2000), these a provide entry into the sys-
3
thumbnail
provide
latic hterature; 2) to
our knowledge (including systematic, ecological, ethnobotanical, 3) to serve
etc.);
and concepts
as a reference for other professionals; 4) to fix the of taxa, espe-
cially families and genera, in the minds of users. Generally we agree with these
secondary goals, and expended considerable thought and effort in applying
We
them in our previous f loristic effort (Diggs et al. f 999). also agree with Sand-
ers and Judd (2000) that there a critical need for the collaboration (and prob-
is
ably more importantly cross-training) of monographic, and phyloge-
loristic,
f
netic researchers. However, we disagree with Sanders and Judd (2000) in how to
They methods
accomplish the fourth of their stated goals. argue that the of
phylogenetic classification should be applied consistently in floras only
(e.g.,
We
monophyletic groups allowed, hence precluding paraphyletic families).
believe that this approach, inflexibly applied, would hinder the primary goal
if
(of a flora) of allowing effective identification. Further, all aspects of the cla-
if
distic approach are followed elimination of ranks and binomial nomen-
(e.g.,
we primary The
clature) in a flora, envision significant erosion of this goal. loss
many
and
morphologically coherent recognizable paraphyletic fami-
of easily
the discontinuity in information retrieval due to a radical change in no-
lies,
mnemonic
such widely
menclature, the lack of effective devices to replace rec-
ognized and practically important ranks as family and genus, and instability
in nomenclature (and hence identification) associated with rapidly changing
cladograms are a few of the reasons for this concern.
some unable
In cases, loristicians, for practical or historical reasons, are to
f
apply even the most important recent phylogenetic discoveries in their floras.
For example, the critically important Flora of North America Project, because
had
of the long time span necessary for such a massive multi-volume work, to
adopt standard years ago (Cronquist system of families)— hence, the Liliaceae
a
now known
broad be polyphyletic, being recog-
(in the sense), to clearly is still
nized in a forthcoming volume (with an extensive discussion of phylogeny).
we
While strongly agree with cladists that polyphyletic groups should be elimi-
we
nated (whenever disagree with the advisability of eliminating the
possible),
numerous and meaningful paraphyletic groups, particularly the
useful at lev-
We
family and genus further discussion below). would add three other
els of (see
&
goals to the four (Sanders Judd 2000) enumerated above: 5) to address spe-
work
cifically the needs of diverse users (discussed above); 6) to connect the of
monographers and other researchers to the "consumers of botanical informa-
who
tion" Barkley pers. comm.; Barkley 2000) need to use these discoveries;
(T.
and use systems of classification and nomenclature that allow meaning-
7) to
ful comparisons with other loristic works. In other words, for conservation,
f
compare
biogeographical, ecological, purposes, should be easy to data
etc.
it
number number number
such as the total of species, the of endemics, or the of
from compari-
introduced species flora to flora (with the realization that the
we somehow
sons will be far from perfect, but useful nonetheless). Ultimately,
hope to combine several important but not necessarily compatible approaches.
We
not only want to produce a useful, informative, and user-friendly flora, but
also one that accurately reflects evolutionary history be phylogenetically
(i.e.,
and
informative) incorporates recent discoveries in botany.
APPROACH TAKEN THE ILLUSTRATED FLORA OF EAST TEXAS
IN
After considerable thought, discussion with a variety of individuals, and a re-
view of the pertinent literature, we are taking what we hope an intermediate,
is
somewhat approach
albeit conservative, in the Illustrated Flora East Texas.
of
maximum
Our goal to provide information while retaining a practical and
is
framework.
utilitarian
Cladistic side
On
number
the cladistic/phylogenetic systematics side of the argument, a of our
decisions have been influenced by the desire increase information content
to
and
accuracy:
We
known
1) are attempting to provide detailed information on the evolution-
A
ary relationships of various plant groups. tremendous amount of new infor-
mation has become available recently (primarily, but not exclusively, as a re-
much
sult of the application of Hennigian principles to molecular data), and as
and
of this as possible is being included references provided. For example, in
the draft family synopsis of the Lemnaceae (duckweed family), we (Diggs
et
al.
in prep.) have included the following statement.
m
Lemnaceae are tiny and extremely reduced morphologically making difficult the past to deter-
it
mine the phylogenetic relationships of the family Kvacek (1995) suggested that the fossil genus
Limnohiophyllum is a fossil link between Araceae and Lemnaceae, and Stockey et al. (1997), using a
cladistic approach and material of Limnohiophyllum. concluded that Pistia free-floating member
(a
et al. 1993b) or more recently to Araceae subfamily Aroideae (French et al. 1995). In fact, the Lemnaceae
many
is considered by authorities to have evolved from within Araceae Oack-in-the-pulpit family) by
extreme reduction, and it has been suggested that Lemnaceae be reduced to a subgroup within a
Even where we
very preliminary information have included and
available,
is
referenced an attempt understanding
in to foster a better of evolutionary
it re-
lationships.
2) Also on the cladistic side (and on that of most other plant taxonomists), we
when
are rejecting all clearly polyphyletic groups, even these are practical and
of long-term or wide usage. The best example of this the Lihaceae (lily fam-
is
ily) sensu lato (in the broad sense). Extensive morphological and molecular data
now
clearly indicate that as broadly conceived, this family a heterogeneous
is
on Fay
mixture based superficial similarities in flower structure et 2000;
(e.g., al.
Rudall et al. 2000b). In fact, recent molecular studies (e.g., Chase et al. 2000)
shown
have that species traditionally treated in the Liliaceae should be placed
m
As we
at least four different orders. a result, are recognizing 14 separate fami-
However, from
previously treated in the Liliaceae) East Texas. the
lies (all for
we
(narrow
standpoint of usability, are incorporating a table in the Liliaceae
what
sense) family treatment that cleariy indicates in family the genera for-
now
merly included in the Liliaceae (broad sense) are placed. Furthermore, in
many
the main key to families, as as possible of the liliaceous (broad sense)
and genus
families will be clustered together clearly indicated. Likewise, the
Nolina (bear-grass) and relatives, which have often been included in the
its
now known
Agavaceae (agave family), are to not be closely related to that fam-
m
we we
ily and are excluding them. In a draft family synopsis (Diggs et al.
prep.)
say,
Liliaceae (e.g., Kartez 1999) or often in the Agavaceae (e.g., Correll &Johnston 1970; Diggs et al. 1999;
&
How-
Verhoek Hess 2002 following Cronquist 1988) based on certain morphological similarities.
&
Simpson
be recognized separately (Dahlgren et al. 1985; Eguiarte et al. 1994; Bogler 1995, 1996;
Kubitzki et al. 1998; Chase et al. 2000). Molecular evidence indicates that Nolinaceae is closely re-
lated to Convallariaceae and Ruscaceae, and some studies (e.g., Chase et al. 1995a; Chase et al. 2000;
Fay 2000) have suggested that the Nolinaceae should be included in the Convallariaceae. Fol-
et
al.
On the other hand, Rudall et al. (2000a) and Judd et al. (2002) included the Nolinaceae in a broadly
&
Tamura
and presumed Rudall 2000a; Yamashita 2000). Since the Nolinaceae
its relatives (e.g., et al.
&
appears be well-defined monophyletic group (Bogler Simpson 1995, 1996), and until the phy-
to a
logeny of this complex is clarified and the nomenclature more stable, we are recognizing it as a dis-
understanding
Hopefully, such explanations will allow users to see that the of
we
changing and improving. With such hope
plant relationships insights,
is still
non-taxonomists be needed nomenclatural changes.
that will less resistant to
when
3) Again on the cladistic side, established useful family concepts are not
we lumping small groups whose relationships have
excessively distorted, are
now
become clear For example, the monogeneric family Najadaceae (the ge-
now known
nus Najas) be derived from within the Hydrocharitaceae (R.
to
is
&
Haynes
Haynes, pers. comm.; Shaffer-Fehre 1991; Les et 1993; Les 1995;
al.
more
Haynes et 1998; Haynes 2000). Including Najas in the Hydrocharitaceae
al.
modify
accurately reflects evolutionary history, yet does not substantially the
beyond bounds
concept of the Hydrocharitaceae nor distort the of usability.
it
We
Thorne
are therefore following several recent loristic treatments 1993;
f (e.g.,
lumping
Diggs 1999) in Najas into the Hydrocharitaceae.
et al.
Our and Some
4) families, genera, species are arranged alphabetically. very
tra
want
taxonomists
ditional "related" families placed together in the linear se
quence physically required of a book. However, the complex branching patten
of evolution does not follow such a linear form and thus any linear sequence
i
An
and
highly arbitrary distorts actual evolutionary relationships. easy to
us(
alphabetical sequence, while not reflecting relationships, at least does not dis
tort them. In addition, an alphabetical arrangement allows quick and easy ac
cess to the material so arranged.
Evolutio
On
number
the evolutionary taxonomy side of the argument, a of our decisions
have been influenced by both and/or
practical theoretical considerations:
We
1) are continuing to use the system of nomenclature that has developed
250 Code
over the past years (International Botanical Nomenclature-Greuter
of
et 2000). This (particularly the use of binomials) an eminently useable
al. is
system that addresses the needs of an audience far broader than the taxonomic
community
consumers
("the of botanical information"—!. Barkley, comm.;
pers.
We
Barkley would
2000). believe that eliminating cause great confusion
it
plant taxonomy went to a specialized non-binomial, clade-based system, some
if
separate static system of "accepted plant names" would be developed by the
community
horticultural or other user groups agricultural, ecological, con-
(e.g.
move
servation). Such a would both marginalize plant taxonomy and ultimately
much
nomenclatural system with
result in a less information content than
at
may
present. This practical consideration well be one of the most important
reasons for maintaining our current system of classification and nomenclature.
In fact, even those developing the PhyloCode (2002), the nomenclatural system
being produced by phylogenetic systematists, have not yet come to grips with
what do naming
to regarding the of "species." interesting to note that Stevens
It is
(2002) has argued that binomials have been used so long and so widely (across
many many
and human
cultures in contexts) because they inherent
are in per-
ception—in other words, having such a two word nomenclature system may be
built into the organization of our nervous systems. Nixon and Carpenter (2000)
in a similar vein suggested that, "Our natural form of communication (even as
common human
evidenced by the binomial system naming
of ourselves)
is
clarified by the use of ranks and binomials." Likewise, anthropologist Brent
Berlin (1992) noted that there are widespread cross-cultural regularities in the
and naming
classification of living organisms by people in traditional, nonlit-
erate societies— these systems more closely approximate Lmnaean binomials
when
than
clade-based nomenclature.
Further, the diverse users of a flora are
considered, a radical shift in the system of nomenclature used seems particu-
when
larly ill-advised, especially at a time the public needs to be brought closer