Testing the Validity of Traditional Morphological Characters by
Seeing Whether They Are Consistent With Molecular Trees
To be consistent with a given phylogenetic tree, a character must
map onto the tree with few changes in character state.
example, bilateral symmetry is consistent with the traditional morphology-based
cladogram for the Big Nine phyla (those with more than 5,000 named
species), since it can be mapped onto the cladogram with only one
change in character state.
- The basic
phylogenetic outline shown in Figure 14 has been standard since
the publication of The Invertebrates by Hyman. Hyman's
original tree had a "primitive acoel flatworm" as the base of
the Bilateria, with Platyhelminthes and Nematoda in the Protostomia.
Many authors now make the latter two phyla separate branches basal
to the other Bilateria.
14. The traditional morphology-based phylogeny of the major animal
phyla, based on the "hypothetical diagram" by Hyman (1940, vol.
1, p. 38). Bilateral symmetry is consistent with this phylogeny
because it requires only one change in character state to account
for its distribution among the phyla.
however, is less consistent with traditional morphology based phylogenetic
trees, because it requires at least two changes in character state:
one for annelids and arthropods and one for chordates (Fig. 15).
15. Traditional phylogeny of the Big Nine showing that segmentation
is less consistent, because it requires at least two changes in
character state to account for its distribution.
Lack of consistency
implies that either the character is not synapomorphic (homologous),
or the phylogenetic tree is incorrect.
characters can evolve convergently more than once, as segmentation
apparently did in Arthropoda and Chordata. Convergent evolution
results in analogous characters that are not indicative of phylogeny.
In the language of cladistics, such an evolutionary convergent
character is a homoplasy. A homoplasy is not a synapomorphy (shared,
derived, homologous character), which is the only kind of character
that is useful in cladistics.
- If analysis
indicates that an incharacter is nevertheless a synapomorphy,
the tree is less likely to be correct, because it does not provide
a parsimonious explanation for the evolution of the character.
A different tree should be tried.
- These same
principles apply to molecular characters as well as to morphological
that is consistent with both a morphological and a molecular phylogeny
is more likely to be phylogenetically informative.
- A morphological
character is not likely to be useful if it is not consistent with
either a tree based on other morphological characters or one based
characters have not led to a consensus phylogeny.
- The use
of different characters and methods of analysis have also resulted
in numerous dif-ferent proposed phylogenies (Jenner and Schram
1999). For example, Willmer (1990), like Hyman, used a noncladistic
approach in which one or a few striking characters, such as the
fate of the blastopore, were heavily weighted. Nielsen (1995)
and other cladists give equal weight to a larger number of characters.
Pechenik (2000, pp. 18-20) provides a sampling of different trees,
and Eernisse et al. (1992) show a dozen variations gleaned from
(1940, vol. 1, pp. 37-38) phylogenetic tree of Animalia has a
trunk from which Pro-tozoa, Mesozoa, Porifera, and Radiata sprout
before giving rise to the Bilateria, as outlined in Figures 14
and 15. She recognized 22 phyla (listed below in bold face with
some spellings changed) and arranged them in order of complexity:
Branch A. Mesozoa
Branch B. Parazoa
Branch C. Eumetazoa
Aschelminthes (Rotifera, Gastrotricha, Kinorhyncha, Nematoda,
(1990, p. 361) presented a morphology-based tree that she characterized
as "undoubtedly wrong" but a useful summary of her conclusions
from the morphological evidence. Her summary tree could be shown
in a cladogram of the Big Nine phyla that differs from Figures
14 and 15 mainly in that each of four lineages for Nematoda, Mollusca,
Annelida, and Echinodermata plus Chordata diverges from a flat-worm-like
ancestor. In addition, she divided Arthropoda into three phyla,
with only Uniramia allied to Annelida. Her tree for the 36 phyla
that she recognized can be summarized as follows:
Lines of descent from "acoelomate Platyhelminthes"
Gastrotricha, Nematoda, Nematomorpha
Echiura, Pogonophora, Onychophora,
Ectoprocta, Phoronida, Brachiopoda
Hemichordata, Echinodermata, Chordata
- A morphology-based
cladogram for the Big Nine phyla derived from Nielsen's (1995,
p. 6) summary cladogram is similar to Figures 14 and 15 except
that Nematoda is in a separate line from the one in which Platyhelminthes
then Mollusca, Annelida, and Arthropoda branch. His cladogram
showing the 31 phyla he recognized (in boldface) is summarized
in the following slightly simplified indented list:
morphological characters traditionally used in phylogenetics are
also consistent with the widely accepted molecular phylogenetic
tree of the Big Nine Phyla: bilateral symmetry and triploblasty,
deuterostomy, and spiral cleavage pattern.
16 summarizes the currently accepted molecular phylogeny for the
Big Nine phyla. It differs from the traditional phylogeny in Figures
14 and 15 mainly in dividing the protosomes into two distinct
clades, with one including Platyhelminthes, Mollusca, and Annelida,
and the other including Nematoda and Arthropoda.
16. Bilateral symmetry and triploblasty, deuterostomy, and spiral
cleavage pattern are consistent with the molecular phylogenetic
tree of the Big Nine Phyla.
and triploblasty are also consistent with a molecular phylogenetic
tree that includes all animal phyla.
In the first volume of The Invertebrates (1940, pp. 32-39),
Hyman placed the Bilateria in a grade above the Radiata (Cnidaria
and Ctenophora), which were in turn above the Porifera. She eschewed
the term "diploblast," noting that sponges do not develop from
two germ layers and that all except Hydrozoa have a middle layer
of cells that could be called tissue.
- RECENT MORPHOLOGICAL
STUDIES: Willmer (1990) rejected not only the dip-loblast/triploblast
dichotomy, but also the Radiata/Bilateria distinction, noting
that most sponges and placozoans are asymmetric, and many cnidarians
and especially ctenophorans tend toward bilateral symmetry. Nevertheless,
she (1990, p. 361) placed Porifera, Cnidaria, and other traditional
"Radiata" or "Diploblasts" below a flatworm-like ancestor of the
bilaterally symmetric animals. Nielsen (1995, p. 64) also concluded
that bilateral symmetry is "a highly questionable synapomorphy
of the bilaterians," but for different reasons he (p. 72) considered
it "reasonable to regard the Bilateria as a monophyletic group
and as the sister group of the Cnidaria." His Bilateria includes
Ctenophora as the sister group of Deuterostomia (p. 307).
- As shown
in Figure 17 (next page), a tree based mainly on 18S rDNA supports
the traditional view that "diploblasts" or "radiata" are basal
to the bilateria.
- This molecular
phylogenetic tree is a composite of many separate studies, some
of which will be discussed in Part III. It recognizes 30 phyla.
Mesozoa are included within Platyhelminthes, and Echiurida and
Pogonophora are included within Annelida. See Figure 1.B of Adoutte
et al. (2000) and Figure 2 of Zrzavý et al. (1998) for somewhat
different molecular phylogenetic trees of all phyla. It may prove
convenient to print a copy of Figure 17 for reference in later
- The Bilateria
are also recovered as a monophyletic clade in molecular phylogenies
based on 5S rDNA (Hori and Osawa 1987) and 28S rDNA (Christen
et al. 1991).
in Echinodermata, Hemichordata, and Chordata is also consistent
in a molecular phylogenetic tree of all animal phyla.
Hyman (1951, vol. 2, p. 5) accepted the then-prevailing view that
deuterostomes (Chaetognatha, Echinodermata, Hemichordata, and
Chordata) were a heterologous assemblage of groups that were not
closely related. She considered the lophophorates (Ectoprocta,
Phoronida, and Brachiopoda) to be intermediate between protostomes
- RECENT MORPHOLOGICAL
STUDIES: Willmer (1990, p. 349) rejected the protostome/deuterostome
dichotomy, but she agreed that Hemichordata, Echinodermata, and
Chordata (branching in that order) represented a monophyletic
lineage that did not include the lophophorates. Nielsen (1995,
pp. 76-77) regarded the fate of the blastopore as an unreliable
character, but using other characters, he (p. 62) divided the
Bilateria into the two clades Protostomia and Protornaeozoa (Ctenophora
plus Deuterostomia). His Deuterostomia clade included hemichordates,
echinoderms, chordates, Phoronida, and Brachiopoda, but not Ectoprocta
- If the problematic
Chaetognatha are excluded, the molecular tree based on 18S rDNA
(Fig. 17) suggests that the traditional deuterostomates are in
fact monophyletic. Chaetognaths, as well as lophophorates, will
be discussed in more detail later in Part III.
17. Summary tree from molecular phylogenetic studies of animal
phyla. (There are no molecular data for Loricifera.)
pattern is somewhat consistent with a molecular phylogenetic tree
that includes all animal phyla.
Hyman recognized the importance of the cleavage pattern, but she
gave more weight to the schizocoelous versus enterocoelous origin
of the coelom. She appears not to have regarded "Spiralia" as
a distinct group.
- RECENT MORPHOLOGICAL
STUDIES: Willmer (1990, pp. 128-129) cautiously recognized a "core
group of spiralians… including certain polyclads, nemerteans,
annelids, uniramian arthropods, pogonophorans, echiurans, sipunculans
and molluscs." However, she (p. 268) regarded the molluscs as
pseudocoelomates with little in common with other phyla. Nielsen
(1995, p. 96) considered it best to treat the Spiralia as the
sister clade of Aschelminthes within Protostomia. His Spiralia
comprised Sipuncula, Mollusca, Annelida, Onychophora, Arthropoda,
Tardigrada, Entoprocta, Platyhelminthes, and Nemertea.
phylogenetics (Fig. 17) supports the existence of a clade in which
the spiral cleavage pattern may be plesiomorphic (primitive).
However, this clade also includes lophophorates, which generally
have radial cleavage. It also includes flatworms, which have spiral
cleavage but are often not included with coelomate spiralians.
It does not include onychophorans, tardigrades, or arthropods.
These groups will be discussed in more detail in Part III.
by itself is not consistent with the molecular phylogenetic tree,
and it may not be a homology.
Hyman (1959, vol. 5, p. 229) defined the lophophore as "a tentaculated
extension of the mesosome that embraces the mouth but not the
anus and has a coelomic lumen." She considered it a homologous
character uniting Ectoprocta, Phoronida, and Brachiopoda as lophophorates.
- RECENT MORPHOLOGICAL
STUDIES: Willmer (1990, p. 349) considered the lophophor-ates
to be a "rather close-knit assemblage." Nielsen (1995, p. 183)
found no synapomorphy uniting the ectoprocts with the phoronids
and brachiopods. He dismissed the lophophore as nonsynapomorphic,
because numerous other characters linked Ectoprocta to Spiralia
and Phoronida and Brachiopoda to Deuterostomia.
studies (Fig. 17) place the "lophophorates" in the clade of protostomates
that includes Platyhelminthes, Mollusca, and Annelida. The lophophorates
do not appear to be monophyletic within that clade, however, because
Phoronida and Brachiopoda may be more closely related to each
other than to Ectoprocta.
and type of body cavity (whether the animal is acoelomate, pseudocoelomate,
or coelomate) is not consistent with the molecular phylogenetic
tree and is not homologous.
Hyman (1940, vol. 1, p. 35), following Schimkevitch (1891), was
primarily responsible for the distinction among acoelomates, pseudocoelomates,
and coelomates. "Such a division," she wrote, "stands firmly on
a realistic anatomical basis and eschews all theoretical vaporizings…."
She dismissed the once-common view that acoelomates and pseudocoelomates
originated from coelomates. This "Archecoelomate Theory" has been
revived in recent times, but it is not widely accepted or even
familiar among American zoologists. (See Willmer 1990, pp. 33-37
for discussion.) Support for it comes from the fact that the pseudocoel
can form by loss of the peritoneum or part of the mesoderm enclosing
a coelom, among other ways (Maggenti 1976). In addition, some
nematodes, all leeches, and some other presump-tive coelomates
and pseudocoelomates have secondarily become acoelomate, showing
that a coelomate-to-pseudocoelomate or coelomate-to-acoelomate
evolutionary sequence is at least conceivable. All of this casts
doubt on the homology of the acoelomate and pseudocoelomate conditions.
- RECENT MORPHOLOGICAL
STUDIES: Willmer's (1990, p. 22-38) review of body cavities led
her to the conclusion that "it may well be that--contrary to most
of the simple invertebrate textbooks--the body cavities of animals
are amongst the most misleading of all possible characters." She
(p. 246) concluded that the pseudocoelomates (including molluscs;
p. 268) were polyphyletic derivatives of several acoelomate lines.
Nielsen (1995) saw "nothing to indicate that the acoelomate condition
is ancestral" or "that the various coeloms are homologous" (p.
65), and he opined that the pseudocoelomate versus coelomate distinction
had been "strongly overemphasized" (p. 235). He rejected, however,
the idea that the acoelomate and pseudocoelomate conditions were
derived from coelomate ancestors (p. 236). A cladistic analysis
by Wallace et al. (1996) also indicated that pseudocoelomates
were polyphyletic, with one clade comprising Rotifera and Acanthocephala
and another comprising two lesser clades: (Nematoda + Nematomorpha)
and (Kinorhyncha + Loricifera + Priapulida).
phylogenetic studies (Fig. 17) support the nontraditional view
that the pseudocoel is apomorphic (derived) with respect to the
coelom. The traditional pseudocoelomates (including Nematoda,
Nematomorpha, Priapulida, Kinorhyncha, Rotifera, and Entoprocta)
are scattered in two distinct clades, each of which also includes
coelomates. In addition, the out-group of these two clades, Deuterostomia,
is coelomate. Thus the most parsimonious hy-pothesis is that the
coelom is plesiomorphic in all Bilateria, and the pseudocoel is
derived from it.
phylogenetic studies also suggest that the acoelomates are derived
from coelomates. "Acoelomates" (including Platyhelminthes and
Gnathostomulida) occur within a clade that mostly comprises coelomates.
The most parsimonious hypothesis is that the acoelomate condition
evolved from a coelomate, not that coeloms evolved from acoelomates
many times in these other phyla.
may, however, be monophyletic within this clade (Giribet et al.