Annelida and Arthropoda Are Not Sister Taxa: A Phylogenetic Analysis of Spiralian Metazoan Morphology

Sysl. Biol.
41(3):305-330, 1992
ANNELIDA AND ARTHROPODA ARE NOT SISTER TAXA:

A PHYLOGENETIC ANALYSIS OF SPIRALIAN
METAZOAN MORPHOLOGY
DOUGLAS J. EERNISSE, JAMES S. ALBERT, AND FRANK E. ANDERSON 1
Museum of Zoology and Department of Biology, University of Michigan,
Ann Arbor, Michigan 48109-1079, USA
Abstract.—Annelids and arthropods have long been considered to be each other's closest
relatives, as evidenced by similarities in their segmented body plans. In the first cladistic analysis
of metazoan morphology accompanied by an explicit data matrix, Schram (Meglitsch and Schram,
1991, Invertebrate zoology, 3rd edition, Oxford Univ. Press, New York) suggested tentative
Downloaded from http://sysbio.oxfordjournals.org/ at Russian Archive on November 13, 2013

support for this conventional "Articulata" hypothesis. Our reanalysis of the Schram data matrix
yielded weak support for an alternative "Eutrochozoa" grouping of annelids, molluscs, and
certain other spiralian phyla, exclusive of arthropods. Likewise, recent 18S ribosomal RNA se-
quence comparisons have favored the Eutrochozoa hypothesis. This study presents a new analysis
of 141 independently assembled characters, purported to represent the current state of knowledge
of metazoan morphology and embryology. This maximum parsimony analysis resulted in robust
support of Eutrochozoa. For this data compilation and method of analysis, the Articulata hy-
pothesis could only be supported by adding multiple ad hoc proposals of evolutionary events.
Instead, the more parsimonious Eutrochozoa hypothesis is favored as the best-supported current
reconstruction of higher level animal genealogy. [Phylogeny; Metazoa; animal; Arthropoda;
Annelida; Mollusca; morphology; embryology; RNA.]
An enormous literature of descriptive gram (Hennig, 1966; Patterson, 1982), and

and experimental work on the ontogeny only shared derived characters (synapo-
and morphology of animals has accumu- morphies) are considered candidates for
lated over the last century and a half. Much homology. Other similarities that result
is now known about the life history, anat- from independent or parallel evolution
omy, and genetic organization of species (homoplasies) or are retained from more
in most of the more than 30 recognized ancient evolutionary transformation (ple-
animal phyla. Our modern interpretation siomorphies) are not considered relevant
of character homology rests on the results even when they are phenotypically iden-
of these studies and on the understanding tical.
of animal evolution they make possible. Studies on the early evolution of major
The recent emphasis of cladistic methods animal lineages and on the origins of im-
in systematic biology, however, has fo- portant features in their structural design
cused new attention on the evidential basis have only recently begun to incorporate
for asserting hypotheses of homology the character congruence approach of eval-
among characters (Wiley, 1981; Patterson, uating phenotypic similarities. One ex-
1982). The criteria for homology formal- ample is the investigation of genealogical
ized by Remane (1956) involve evaluation relationships among spiralian metazoans,
of similarities of ontogeny, composition, including Arthropoda, Mollusca, Anneli-
and anatomical position on a case-by-case da, and several other less speciose phyla.
basis. These similarities are now regarded Spiralians comprise >90% of all living
by many systematists to be preliminary to metazoan (multicellular animal) species
the test of character congruence on a clado- (Barnes, 1987; Brusca and Brusca, 1990), yet
the genealogical relationships among many
1
of them remain undocumented by phylo-
Present address: Institute of Marine Sciences and genetic criteria. Members of the Spiralia
Biology Department, University of California, Santa
Cruz, California 95064, USA. have also been grouped as Protostomia by
305
306 SYSTEMATIC BIOLOGY VOL. 4 1
Grobben (1909) on the basis of certain em- tebrate zoology textbooks (Lutz, 1986;
bryological similarities, especially spiral Barnes, 1987; Pearse et al., 1987; Brusca and
cleavage of the blastomeres, derivation of Brusca, 1990; Kozloff, 1990; Meglitsch and
the mesoderm from a single (4d mesen- Schram, 1991) revealed that the Articulata
toblast) cell, and protostomous mouth for- hypothesis continues to dominate discus-
mation. sions of animal relationships. Although
The overt segmental arrangement of these texts differ in their treatment of hy-
parts in the adult body plan of Annelida potheses on the origin of character systems
and Arthropoda, but not Mollusca, led Cu- such as the coelom and mesoderm, there
vier (1817:508) to erect "Les articules" as is an apparent unanimity regarding the or-
"le troisieme grande division du regne an- igin of spiralian segmentation. The union
imal." Later, Haeckel (1866) popularized of Annelida and Arthropoda, exclusive of
the use of "articulates" as an evolutionary Mollusca, is one of the few consistent el-

lineage, which in his opinion rivaled the ements in the diversity of their summary
vertebrates as among the greatest of his 12 phylogenetic diagrams. Some of these au-
stems of the animal kingdom. Although its thors (e.g., Barnes, 1987) even echo Haeck-
exact composition has varied in the eyes el's (1866) depiction of Arthropoda emerg-
of subsequent authors, the "Articulata" su- ing directly from Annelida, although few
perphylum-level grouping is still widely researchers today would posit the deriva-
used. Some nomenclatural problems have tion of one higher taxon from another
arisen from the proposal of an alternative (Ghiselin, 1974; Wiley, 1981).
name (Hadzi, 1963) as well as the use of The hypothesis of a monophyletic Ar-
"Articulata" in brachiopod taxonomy ticulata (Figs, la, lb) has yet to be docu-
(Huxley, 1869). More recently, some zool- mented by an analysis of data using the
ogists have elevated Articulata to include character congruence approach. For ex-
most spiralian taxa, even including Mol- ample, Brusca and Brusca (1990:882) pre-
lusca (Nielsen, 1985; Ghiselin, 1988). Al- sented one of the first character-labeled
though we do not dispute the potential branching diagrams for metazoans. None-
value of this suggestion, Articulata is used theless, it is not clear whether the phylo-
here in the conventional sense as a hy- genetic distribution of characters was used
pothesis that axial mesodermal segmenta- to build the tree. The reader is left to con-
tion arose but once in the shared common clude with the authors (1990:682) that
ancestor of Annelida and Arthropoda, per- "[t]here is little argument that annelids and
haps also including Pogonophora, Tardi- arthropods are closely related" and that
grada, and Onychophora, but not Mollus- "[t]he body plans of these two phyla are
ca. more similar to one another than to any
The near universal use of the Articulata other major protostome group." This con-
hypothesis in textbooks and classrooms of clusion is reinforced elsewhere (1990:765):
this century has left a deep impression on "molluscs probably arose early in the pro-
the interpretation of patterns and process- tostome clade, soon after the origin of the
es of animal evolution by zoologists. From coelom but before the origin of annelid-
Haeckel's perspective as an advocate of the arthropod metamerism" (emphasis in orig-
biogenetic law, the observation that an- inal). In the final analysis, however, a sin-
nelids and arthropods pass through an un- gle character diagnosing a clade including
segmented early ontogenetic stage is evi- Annelida and Arthropoda (with Pogo-
dence that segmentation is derived with nophora) was proposed, "true segmenta-
respect to the molluscan condition. This, tion arising by teloblastic growth and re-
of course, reinforced his opinion that seg- sulting in serial repetition of body parts,"
mentation is homologous between the two although conflicting data were not report-
groups. ed.
The theory of recapitulation is no longer An older alternative view (Pelseneer,
emphasized, yet a survey of recent inver- 1899, 1906; Naef, 1913, 1924), more re-
1992 PHYLOGENETIC ANALYSIS OF METAZOAN MORPHOLOGY 307
cently advocated by investigators who have

examined 18S ribosomal RNA (rRNA) data
(Field et al., 1988,1989; Ghiselin, 1988; Pat-
terson, 1989; Raff et al., 1989; Lake, 1990;
Eernisse, unpubl. manuscript), posits that
annelids, molluscs, and certain other less
speciose phyla share a more recent com-
mon ancestor with one another than any
do with arthropods. This assemblage (Figs, (c)
lc, Id) was referred to as "Eutrochozoa"
by Ghiselin (1988) and approximately co-
incides with a long-standing grouping of
those taxa with a trochophore larva in at
least some marine representatives of each

group (Hatschek, 1891; De Beer, 1930; Hy-
man, 1951; Gruner, 1982; Nielsen, 1987;
Strathmann, 1987; cf. Salvini-Plawen, (e) j v * y y
1980b). Although ciliary bands have been p JP
reported from widely divergent metazoan
taxa, no species of arthropod is known to
possess any element of the trochophore
larva.
Because of the incongruent character FIGURE 1. Diagrammatic cladograms of the three
distribution of trochophore larvae and seg- possible resolved hypotheses of branching relation-
mented body plans among spiralian taxa, ships among Arthropoda, Annelida, and Mollusca,
assuming each is monophyletic with respect to the
Articulata and Eutrochozoa are mutually others, with the phylogenetic distributions of two
exclusive hypotheses of relationships. The characters plotted; tr = trochophore larva, sg = me-
Articulata hypothesis requires that the lack sodermal segmentation. A solid bar indicates derived
of a trochophore larva in arthropods be presence, an outlined bar indicates derived loss, (a),
explained either by its independent deri- (b) Articulata. (c), (d) Eutrochozoa. (e), (f) Two of the
four most-parsimonious distributions of these char-
vation in molluscs and annelids (Fig. la) acters required by an Arthropoda-Mollusca grouping
or by its derived loss in arthropods (Fig. exclusive of Annelida.
lb). If a trochophore larva and segmenta-
tion are each viewed as unitary characters,
these two hypotheses are equally parsi- hypothesis requires at least four transfor-
monious, each requiring three steps. mations (only two of the four possible com-
The hypothesis of a monophyletic Eu- binations of losses or gains of trochophore
trochozoa also requires at least three trans- larvae and segmentation are shown in Figs,
formations (Figs, lc, Id). Either mesoder- le, If). Although less parsimonious for ex-
mal segmentation is an independent plaining these particular attributes, this
derivation in annelids and arthropods or hypothesis is more consistent with still
Eutrochozoa is primitively segmented, with other character evidence, such as the dis-
molluscs exhibiting a derived loss of this tribution of hemocyanin respiratory pig-
phenotype. Members of some extant mol- ments, which is known only from Mollus-
luscan taxa do possess elements of serially ca and Arthropoda (Mangum, 1985;
arranged (or possibly metameric) struc- Ghiselin, 1989). This third hypothesis was
tures, especially Polyplacophora, Mono- developed shortly after the discovery of
placophora, and Nautiloidea (Cephalopo- living monoplacophorans and arose from
da) (Naef, 1924; Wingstrand, 1985). observations of serial repetition in several
The third logical alternative is that mol- of their organ systems (Lemche, 1959a,
luscs and arthropods are more closely re- 1959b; Fretter and Graham, 1962). This view
lated than either is to annelids. This considers molluscs to be derived from
short-bodied segmented animals with very from the text of Willmer (1990) was un-
small coelomic sacs and an open circula- dertaken by Wheeler (1990), who failed to
tory system. find unambiguous support for the Articu-
Objections to viewing molluscs as prim- lata depicted in graphical form.
itively metameric have come from advo- Our goal in this paper was to abstract
cates for an unsegmented, possibly acoe- from primary reports a representative ac-
lomate, flatwormlike ancestor (Clark, 1964, counting of embryological and morpho-
1979; Stasek, 1972; Salvini-Plawen, 1980a, logical evidence bearing information on
1985), perhaps similar to the wormlike spiralian metazoan phylogeny. It was not
aplacophoran molluscs. Recent treatments our intent to present an exhaustive survey
of molluscan evolution (Runnegar and Po- of this literature but rather a fair approx-
jetta, 1985; Salvini-Plawen, 1985; Barnes, imation of the current state of knowledge
1987; Scheltema, 1988) have tended to rein the zoological community. The main
purpose of this work is to recover from the

place an older archetypical gastropodlike
hypothetical ancestral mollusc (HAM) with available information the best hypothesis
a flatwormlike HAM. Although at least two or set of hypotheses of relationships con-
of these authors have since abandoned in gruent with the phylogenetic distribution
part their earlier published views (Run- of variation.
negar, pers. comm.; Scheltema, pers. METHODS
comm.) and new studies reconstructing
plesiomorphic molluscan characters are Characters used in the present analysis
being published (Wingstrand, 1985; Eer- were extracted from the primary and sec-
nisse and Kerth, 1988), many (e.g., Will- ondary literature on metazoan morpholo-
mer, 1990) continue to believe a consensus gy and embryology and were screened for
has been reached in favor of the flatworm phenotypic similarity so that patterns in
hypothesis. their phylogenetic distribution across taxa
Perhaps the publication of only a single could be examined using maximum-par-
phylogenetic analysis of metazoans (Meg- simony algorithms. A complete accounting
litsch and Schram, 1991) accompanied by of published information pertaining to
an explicit matrix of morphological char- metazoan phylogeny was complicated by
acter hypotheses, including all spiralian the tendency of authors to omit evidence
phyla and appropriate outgroups, is the re- inconsistent with their narrative interpre-
sult of the emphasis on taxonomic spe- tations of character evolution. Studies
cialization in the field of zoology. It is of- whose data are best represented in the
ten asserted that the tremendous present character matrix were those with
divergence of these taxa prohibits studies explicit, if perhaps a priori, proposals of
of morphological variation from recover- homology (e.g., Beklemishev, 1969; Brusca
ing phylogenetically relevant informa- and Brusca, 1990).
tion. Patterson (1990:199), for example, Character data were extracted from mul-
suggested that higher level metazoan phy- tiple sources, seeking to incorporate all en-
logeny "has been something of a back- codable discrete characters that appeared
water for decades, largely because all the to show variation at an appropriate level.
morphological clues had been pushed be- Character descriptions were given binary
yond their limits, and mutually contradic- alternative states, usually present or ab-
tory speculations led only to dead ends." sent, accompanied by reasonably precise
In our view, reproducible testing of hy- character descriptions (Appendix 1). The
potheses has only just begun. A reanalysis 141 binary characters were grouped into
of Schram's data set discussed herein sug- 16 morphological categories for descrip-
gests that the connection between evi- tive purposes only. For the sake of brevity,
dence and hypotheses deserves closer scru- the character list in Appendix 1 includes
tiny. A reanalysis of 55 characters extracted only the major citations used. To avoid rep-
etition, a few of the more general refer- not related to metamerism. We suspect that
ences utilized for specific taxonomic groups such asymmetries, whether or not meta-
are listed at the start of Appendix 1. meric in origin, could instead be due to the
Two assumptions regarding character nondeterministic consequences of induc-
independence and weighting are central tive cues during development (Hall, 1992).
to this analysis. The independence of char- If characters are hypotheses of phylo-
acters was assessed by spatial or topological genetic transformation, then they are also
segregation within an organism, temporal ontologically equivalent as heritable
discoupling in ontogeny, or hierarchical changes that occur in the evolution of a
distinction (e.g., organs, cells, molecules). lineage. Accordingly, we weighted all
We accepted at face value the indepen- characters equally and avoided using prob-
dence of characters with unique distribu- abilistic models of character evolution
tions across taxa. Complex morphological (sensu Ghiselin, 1991) to guide us in
systems such as "trochophore larvae" and weighting characters a priori. Some, no

"segmentation" were considered compos- doubt, will not accept our explicit sugges-
ite suites of many phylogenetically inde- tion that a seemingly minor morphological
pendent characters. feature (e.g., character 34: cilia with one
For example, we distinguished between basal body) should have a weight equal to
serially repeated characters ("meres" of that of manifestly more important evolu-
Bateson, 1894) and segmented body plans. tionary transformations (e.g., character 96:
This distinction can be difficult because anus with proctodeum). However, a priori
there is no consensus regarding the mech- weighting is an epistemological issue; no
anisms leading to the production of either claim is made that nature produces varia-
the segmented body plans of spiralians or tion in all characters with equal frequency,
serially repeated characters. Here we refer and equal weighting is used in the absence
to segmentation or metamerism as the se- of compelling reasons to weight otherwise.
rial arrangement of similar body parts along Also, equal weighting is not intended to
the bilateral axis of the animal, containing be a null hypothesis. We have simply de-
congruent and repeated patterns of tissue cided to break down the known variation
organization. A segment (metamere) is into as many encodable characters as the
therefore a block or unit of a segmented data will permit, and this seems as good a
organism. The pattern of serially repeated reason as any to weight them equally. Some
characters, on the other hand, does not characters turn out to be more useful at*
necessarily correlate with those of other certain levels than others, i.e., useful be-
repeated patterns in an organism. Al- cause they help predict the set of relation-
though repeated characters are often re- ships that carry a library of phylogenetic
ferred to as "metameric," we prefer to re- information.
strict this term to cases of complete body Variation in all characters was atomized
segmentation, a condition referred to as into derived and plesiomorphic states to
"true" or "mesodermal" segmentation by reflect specific hypotheses of evolutionary
Hyman (1951:28). In many cases, the phy- transformation. As hypotheses of unique
logenetic distribution of repeated charac- historical events, characters are ontologi-
ters in one organ system is different from cally equivalent and comparable, making
that of other organ systems. We took this it possible to combine them in the study
variation as evidence that these characters of patterns of their phylogenetic distri-
do not share an identical history. Russell- bution in a parsimony analysis. Another
Hunter and Brown (1965), Russell-Hunter methodological consequence of character
(1988), and others have argued that bilat- equivalence involves the use of multistate
eral asymmetries in numbers of repeated characters. Additive binary coding, com-
characters are a clear indication of "pseu- putationally equivalent to ordered multi-
dometamerism," or duplication of systems state characters (Farris et al., 1970), was em-
ployed to make explicit each hypothesis of ation for the question of this study must
transformation (e.g., characters 5-8). not have been reported within its repre-
A rather broad selection of terminal taxa sentatives unless that variation could be
was chosen for the analysis, rather than polarized unambiguously. No attempt was
restrict the analysis to only major spiralian made to restrict or otherwise balance the
phyla, because the relationships of spirali- selection of terminal taxa according to Lin-
an phyla to other metazoans remain high- naean rank, numbers of species, or ecolog-
ly controversial (e.g., Nielsen, 1987; Lake, ical diversity. For example, Conchifera
1990). Inclusion of multiple outgroup taxa (sensu Wingstrand, 1985) includes the di-
increases the likelihood of resolving par- vergent monoplacophorans, gastropods,
ticular ingroup relationships of interest cephalopods, bivalves, and scaphopods;
(Swofford and Olsen, 1990). We did not Pogonophora includes the hot-vent inhab-
attempt to address the putative monophyly iting vestimentiferans; and Clitellata in-
cludes oligochaetes and leeches. Conven-

of Metazoa itself, which has been brought
into question by recent authors (Field et tional taxa that were split include
al., 1988; Lipscomb, 1989; Christen et al., Arthropoda (Crustacea, Chelicerata, and
1991). In investigating the evidence for Uniramia), Platyhelminthes (Acoelomor-
metazoan monophyly, we assembled sev- pha and Rhabditophora, sensu Ax, 1985),
eral potential synapomorphies (characters Aplacophora (Caudofoveata and Soleno-
24-26, 73, 100, 102, 135). We intentionally gastres), and Annelida (Polychaeta and
left these in our matrix, although they are Clitellata).
uninformative for this restricted metazoan Characters for which both states are re-
comparison, with the hope that future au- ported in different representatives of ter-
thors will provide more details about their minal taxa were scored in one of two ways.
condition in potential protistan outgroups. In certain cases, the results of previous
Likewise, our analysis does not speak di- studies provided compelling evidence for
rectly to the issue of the monophyly of establishing character polarity within a
"Bilateria" (Fig. 4: node 1) because only taxon, permitting the assignment of the
one taxon, Cnidaria, was included that is more plesiomorphic (i.e., primitive) con-
outside this grouping. Thus, transforma- dition to that taxon. Without such
evidence, the taxon was coded as "poly-
tions postulated between Cnidaria and the
morphic." The states "missing" or "inap-
node uniting all other taxa could either be plicable" were distinguished in Figure 3
synapomorphic for bilaterians or autapo- but were treated identically (as "missing")
morphic for cnidarians. We did not con- in all analyses.
sider some metazoan phyla, notably En-
toprocta, Bryozoa, Porifera, Placozoa, We used PAUP version 3.0r (Swofford,
Ctenophora, Rotifera, Gastrotricha, Hemi- 1990) for all reported phylogenetic anal-
chordata, Nematomorpha, Acanthocepha- yses. Table 1 summarizes select analyses
la, Dicyemida, Orthonectida, Chaetogna- performed on the data in Figure 3, accord-
tha, or Loricifera. None of these phyla, ing to the differing options used and the
except perhaps Entoprocta, are thought to resulting minimum-length tree statistics.
be particularly closely related to the spira- This PAUP file is available in electronic
lian phyla considered here. Still, their in- form upon request. We used the random-
clusion may have influenced our results. addition-sequence option of PAUP for
As in any phylogenetic survey, our se- stepwise addition of taxa, with 100 repli-
lection of terminal taxa was constrained by cates per search, and the MULPARS option
the need to sample the appropriate and to save all minimum-length trees. The de-
informative combinations of primitive and fault accelerated transformation (ACCT-
derived characters. The decision to unify RAN) character optimization was specified
or split particular lineages was based on during searches, but the inferred ancestral
two criteria: (1) the lineage must be thought conditions at internal nodes reported in
to be monophyletic and (2) relevant vari- Appendix 2 include only those that are
consistent with all three character opti- TABLE 1. Summary of constraint analyses of some
mization methods available in PAUP metazoan phyla.
(ACCTRAN, DELTRAN, and MIN F). The Length Consis-
analyses were duplicated in entirety, with Analy- (no. of tency
polymorphic terminal taxa first treated as sis Constraint8 trees) index
polymorphic (Tables 1, 2) and then as un- 1 None 384(6) 0.458
certainties, with nearly identical results ex- Articulata clade
cept for a proportionate increase of length 2 (6, 17) 400 (13) 0.440
estimates in the polymorphic analysis. The 3 (6, 16) 399(4) 0.441
heuristic search algorithm of PAUP was 4 (6, 15) 401(12) 0.439
used for all analyses because the large 5 (5, 16) 394(2) 0.447
6 (5 + 14) 393(2) 0.448
number of taxa precluded practical use of
the branch-and-bound algorithm. Using Arthropods as eutrochozoans
software by Eernisse (1992), we processed 7 (7 + 14 + 18) 400(3) 0.440

8 (6 + 14 + 18) 400 (3) 0.440
the PAUP output of 39 different heuristic 9 (10 + Nemertea + (5, 14, 18)) 397 (3) 0.443
constrained or unconstrained searches, 10 (5 + 14 + 18) 394 0.447
each with 100 random-addition-sequence 11 (5 + 14 + Nemertea + 18) 390 (3) 0.451
replicates. On average, searches found 12 (5, (14, (Nemertea, 18))) 390(2) 0.451
13 (5 + 9) 387 (4) 0.454
>91% of all minimum-length trees found 14 (5,9) 387(4) 0.454
in the first of 100 replicate searches, >97%
by replicate six, and 100% by replicate 16, Arthropod-mollusc clade
15 (8, 18) 407(3) 0.432
suggesting that the heuristic algorithm was 16 (7, 18) 402(32) 0.438
relatively effective for our matrix but also 17 (6, 18) 401(16) 0.439
that the replicate searches were necessary 18 (5, 18) 397(13) 0.443
to be reasonably confident that all mini- Flatworm-mollusc clade
mum-length trees were found. Figure 4 is 19 (Rhabditophora, 18) 396 (14) 0.444
presented in the form of a PAUP "phylo- 20 (11 + 18) 390(4) 0.451
gram," which provides information re- 21 (10 + 18) 390 (4) 0.451
garding the relative proportion of apo- 22 ((10, (Nemertea, 18)), 14) 387(4) 0.454
morphic characters supporting each Not Eutrochozoa
internal node or terminal taxon. Branch 23 not (12) 387(4) 0.455
24 not (13) 385(4) 0.457
lengths are best interpreted as a measure
a
of data available at particular levels of gen- Numbers represent nodes in Figure 4 of clades that were
erality rather than as a measure of anage- forced into sister-group relationships, as indicated by one or
more sets of enclosing parentheses. Numbered clades sepa-
netic rates. rated by commas were each constrained to maintain their
Although commonly used, bootstrap- own monophyly, whereas those separated by plus signs were
not. Constrained searches found minimum-length trees that
ping and similar techniques have un- satisfied (analyses 1-22) or did not satisfy (analyses 23, 24)
known statistical properties, given likely the indicated constraint.
deviations from the assumptions required
to use them to estimate confidence inter-
vals on phylogenies (Felsenstein, 1988). In-
stead, we used the "Constraints" option of a clade (analyses 2-6). Alternatively, ar-
PAUP to search for the minimum number thropods (with or without Onycophora,
of additional evolutionary events (steps) Tardigrada, or Kinorhyncha) or flatworms
required for various competing hypothe- (either Acoelomorpha or Rhabditophora,
ses, relative to the minimum-length to- or both and with or without Gnathosto-
pologies, given the character matrix and mulida) were forced into a sister-taxon re-
method of analysis (Table 1: analyses 2- lationship with molluscs or more inclusive
24). We tested various combinations of clades (analyses 7-22). Searches were con-
constraints that, given our data matrix, strained (analyses 23-24) to find the min-
forced annelids and arthropods, along with imum-length tree(s) that did not support
various or no additional taxa, together as the more or less inclusive "eutrochozoan"
312 SYSTEMATIC BIOLOGY VOL. 41
TABLE 2. . Comparison of previous hypotheses of metazoan relationships.
Tree as shown Tree resolved

Consistency Consistency
Analysis Lengtha index Lengthb index
Strict consensus0 390 0.444 378 (3) 0.458
Hyman, 1940-1967 497 0.354 452 0.389
Hadzi, 1953, 1963 543 0.324 534 0.330
Marcus, 1958 490 0.359 452 0.389
Salvini-Plawen, 1982, 1985 469 0.375 441 (3) 0.399
Nielsen, 1985, 1987 459 0.383 424 0.415
Barnes, 1987 465 0.378 417 0.422
Pearse et al., 1987 440 0.400 419 0.420
Brusca and Brusca, 1990 437 0.403 410 (5) 0.429
Kozloff, 1990 453 0.389 419 (2) 0.420

Willmer, 1990 649 0.271 449 (2) 0.392
Meglitsch and Schram, 1991 448 0.393 443 0.397
Meglitsch and Schramd 481 0.366 403 0.437
a
See Figures 4 and 5. Reported value is the length of a figured tree when optimized on the data matrix in Figure 3.
b
Length of minimum-length tree(s) found during searches constrained to preserve all clades contained in the figured tree
but allowing any polytomies to be further resolved to the one or more hypotheses requiring the fewest postulated steps
when optimized on the data matrix in Figure 3. Number in parentheses is the number of equally parsimonious minimum-
length trees found, if greater than one.
c
See Figure 4.
d
See Figures 2a and 51.
groupings that resulted from uncon- strict consensus tree. A special case was our
strained searches (analysis 1). replacement of Solenogastres, Caudofove-
To compare the results of this study with ata for Brusca and Brusca's (1990) Aplaco-
those of previous efforts, we present cer- phora, Caudofoveata, which we attributed
tain published hypotheses in the form of to typographical error. Once the topologies
cladograms, with taxa limited to those in were created, we calculated tree statistics
our analyses. Because the original trees as optimized on our data matrix using Cni-
were drawn in diverse styles, we did our daria as an outgroup. Figure 5 presents the
best to recreate the trees as cladogram in- assembled topologies, with all but one
terpretations of the original drawings, with rooted with Cnidaria; Hadzi (1953, 1963)
the following caveats. We entered unre- (Fig. 5b) advocated Acoelomorpha as the
solved groupings as polytomies and used root, but because our characters were all
the accompanying text discussions to clar- binary (albeit in some cases additive bi-
ify uncertainties or to confirm the place- nary) and undirected, the tree length was
ment of unresolved basal taxa. Some au- independent of rooting in this case.
thors clearly postulated paraphyletic Table 2 presents tree lengths that were
concepts for certain taxa, which often led computed for Figures 5a-k and Figure 2a
to difficulties in translating their diagrams as depicted when optimized on our data
into cladograms. Depending on the case, matrix. These values are sometimes mis-
we either reduced resolution to a polytomy leading because various authors often (but
or treated them as if the authors had in- not always) intended polytomies to rep-
tended them to be monophyletic concepts. resent their uncertainties of relationships
In other cases, we found no reference to rather than explicit hypotheses of inde-
certain taxa (Hyman, 1951: Tardigrada, On- pendent character evolution. To make
ychophora; Salvini-Plawen, 1982, 1985: comparison of these cases with our mini-
Priapula, Tardigrada, Gnathostomulida, mum-length results more meaningful, we
Pogonophora, Nematoda). To make tree- also computed tree lengths of the most re-
length comparisons possible, we added solved hypothesis or hypotheses resulting
these taxa in a position consistent with our from an analysis in which only the clades
they depicted were used to constrain the (and many more than 10,000 alternative
search for shorter length trees. trees of length 140). As expected from this
During completion of our study, large number of minimum-length trees, a
Schram's (Meglitsch and Schram, 1991; strict consensus of all trees of length 139
Schram, 1991) analysis of metazoan phy- (Fig. 2a) expressed little resolution within
logeny was published. Concurrent with our the spiralian phyla, leaving no support for
own analysis, we entered Schram's matrix either the Articulata or Eutrochozoa hy-
and performed an independent analysis of pothesis or for the proposed placements of
his data matrix. Because the reanalysis of Onychophora and Pogonophora. A major-
Schram's matrix resulted in multiple min- ity-rule consensus tree (Fig. 2b) showed
imum-length trees, we calculated strict and support for the Eutrochozoa hypothesis
50% majority-rule consensus trees (Swof- (including Echiura and Sipuncula) in 96%
ford and Olsen, 1990) as result summaries of the 1,422 trees.
(Fig. 2).

Morphological Analysis
For the analysis of our independently
RESULTS
assembled matrix of morphological and
Reanalysis of Schram's Morphology Matrix embryological data, 141 characters (Ap-
Schram (Meglitsch and Schram, 1991) pendix 1, Fig. 3) were compiled, empha-
presented results of two cladistic analyses sizing variation among spiralian phyla and
of the same matrix of 77 binary morpho- effectively employing nonspiralian meta-
logical characters. The first assumed all zoans as multiple outgroups. The parsi-
characters were undirected, for which mony analysis resulted in six minimum-
Schram presented a single resulting clado- length trees of length 384 steps (or 342 if
gram of length 140. The second analysis, taxa coded as polymorphic were treated as
resulting in a cladogram of length 158, as- having uncertain states), the strict consen-
sumed complete irreversibility of all char- sus of which is presented in Figure 4. To-
acters. One curious result for both clado- pological variation observed among the six
grams was that Onychophora was nested trees is less than that implied by the three
within Arthropoda as the sister taxon to unresolved polytomies represented in the
Uniramia, with Crustacea basal. Another strict consensus. These differences are re-
was that Pogonophora was the sister taxon stricted to the branching order within a
of Arthropoda. The first cladogram indi- clade composed of Chordata, Echinoder-
cated general support for Articulata, with mata, Brachiopoda, and Phoronida (Fig. 4:
Annelida the sister taxon of an Arthropo- node 2), the placement of Priapula (Fig. 4:
da-Pogonophora clade. In contrast, the node 3), and the branching order of the
second cladogram placed Annelida with three extant arthropod subphyla (Fig. 4:
Mollusca, along with Sipuncula and node 8). Trees 1 and 2 of the six minimum-
Echiura, as members of a clade. Although length trees supported a "lophophorate"
the second analysis favors the alternative clade as the sister taxon to Echinodermata,
Eutrochozoa hypothesis, it is subject to the (Chordata (Echinodermata (Brachiopoda,
criticism that (as Schram himself points out) Phoronida))), whereas tree 3 supported the
the assumption of character irreversibility reverse branching order, (Brachiopoda
is a burdensome one, difficult to justify and (Phoronida (Echinodermata, Chordata))),
seldom used because it yields solutions that in which lophophorates are paraphyletic.
are far from the most-parsimonious ones. Tree 1 placed Priapula as the sister taxon
Our reanalysis confirmed that with all to clade 4 (Fig. 4), whereas trees 2 and 3
characters treated as unordered as reported placed Priapula as the sister taxon to a more
Schram's cladogram did indeed have a inclusive clade, itself composed of clades
length of 140 as analyzed with PAUP 3.0 4 and 9. Trees 4-6 mirrored topological
(Swofford, 1990). In addition, we found variation among the first three trees but
1,422 alternative topologies of length 139 differed in supporting a "mandibulate" ar-
(a) (b)
Ancestor Ancestor
Mesozoa 541 Mesozoa
Placozoa Placozoa
Porifera Porifera
Cnidaria Cnidaria
Ctenophora — Ctenophora
Gnathostomulida Gnathostomulida
Platyhelminthes C Platyhelminthes
Gastrotricha Mollusca
Rotifera Sipuncula
Acanthocephala Echiura

Loricifera Annelida
Priapulida Pogonophora
Kinorhyncha Pentastomida
Nematomorpha Tardigrada
Nematoda Onychophora
Chaetognatha Uniramia
Mollusca Cheliceriformes
Sipuncula Crustacea
Echiura Nemertinea
Annelida Entoprocta
Pogonophora Phoronida
Pentastomida Brachiopoda
— Tardigrada Echinodermata
— Onychophora Enteropneusta
Uniramia Urochordata
Cheliceriformes Cephalochordata
L
— Crustacea Pterobranchia
Nemertinea Ectoprocta
— Phoronida Gastrotricha
— Ectoprocta Rotifera
— Brachiopoda Acanthocephala
— Echinodermata Loricifera
— Enteropneusta Priapulida
— Pterobranchia Kinorhyncha
— Urochordata Nematomorpha
— Cephalochordata Nematoda
Entoprocta Chaetognatha
FIGURE 2. Consensus cladograms for our reanalysis of the Schram (Meglitsch and Schram, 1991) data
matrix. All 77 characters for 38 taxa were binary and treated as unordered, with a hypothetical ancestor
declared as an outgroup to root each otherwise unrooted network topology, (a) Strict consensus diagram for
the 1,422 equal and minimum-length trees found, each with a length of 139 and a consistency index of 0.554.
(b) The 50% majority-rule consensus diagram, depicting only those nodes supported in at least 50% of the
minimum-length topologies. Only nodes supported in less than 100% of the 1,422 trees are labeled with the
percentage value.
• 1 • 2 * 3 • 4 * 5 * 6 • 7
Cnidaria 00100000000?OmJNNNNNNNN11100110000010000000000000000000000000000000000
Chordata 11101000000 1 P 1-1 1-1-1
Echlnodermata 11110100100 1 1--P—1 1
Phoronlda ?—10110100100 1-0 PI—1—11 1 1 1
Brachiopoda ?—1111PP00100 1-0 PI 1-1—11-? P 1 1-
Nematoda -101 7-10000000001 1100 0? P—1
Acoelomorpha -101 110-0000000000 10-11-11-? ?
Rhabditophora 11011—P-110-0000000000 110-11-11 1-11P-P-P-P-P
Gnathostomulida ?10?????—7-10000000000 1 1-?—1 P-?
Nemertea 11011-1 7-10001000000 l-0-ll-?l-?-l-l-l-l 1-1-1-1-1 PI
Prlapula •>•>—10000000100 1?0 7-1-1-7 1-
Sipuncula 11011111—1010001000100 1-0-1 7—11-1111—1 1-
Caudofoveata
Solanogattrea 1101 7010001001710 1-0-111-1-1 111—1-1111-1111-1 1 11-?
Polyplacophora 11011111—1010001011010 1-0-111-1-l l l l - l — l l l l l P l l l - l - l - l 1-11-1
Conchlfera 11011111—1010001011010 1-0-111-1-P l l P P l l - l l l l l P l l l - l l l - l — 1 - 1 - 1 1 - 1

Echlura 110-1711—1110001000100 1-0-1-1-1-1 111111? 1-1 1
Pogonophora -10777771 1P101000100 1-0-1-1 7 171—1?—111-1-P 1—11
Polychaata 110111111-1111101000100 1-0-1-1-l-l lllllPlllllll-lllllll-llllll—l-
Clitallata 11011111177-11101000100 1-0-1-1-1-1 1 11-1-1-1111111-111111
Onychophora ???????71-?-1010??11000—1-0 7 111-1-11111—1 11-1—?
Cruatacaa 110111-11 10700011000 1-00 0? 1 11111-1111-1-1 11-1—1
Unlramla —011 l-?-ll?0001?000—1-00 0? 11—P-llll 1 11-1—1
Chalicarata ??7?ll-ll-?-10?0001?100 1-00 0? 11111-1111 1 11-1—1
Tardigrada ????ll-7???-10100011101 1100 0? 11—P-llll—1 1 ?
Klnorhyncha ??7??1?????-10000000001 110 1—1 111-1-1111 1 ?
* 8 • 9 • 0 * 1 * 2 * 3 * 4.
Cnidaria 70111110000000000000000000000111000000000000000000000000000000001000000
Chordata 0000 1 P 11-1 1 1 PI 1 1
Echlnodermata 0000 1 1 11-1 1 1 1 1
Phoronlda 00 1 1-1—7—1 1 1 11 1
Brachiopoda 0 11 1 11-1—7—71 1 1 1 1
Nematoda o-ll—l ?-i 0-1-1111—7 1—1 1 1—1 1
Acoelomorpha 0000 1 0-1 1-1 111 PP 1-1
Rhabditophora 0000 1 1-0-1PP-P—P-l 111 1-1
Gnathoatomulida 000 1??-?-? 1-? 1111 1-1 — 1
Nemertea 000 1 0-1—IP-?-? IP 11 1
Priapula 0—1 1-1-11-7-1-1 ? 1 1 11 1
Sipuncula 00 11 1-1-1—0-1-11 ?-l 1 1 11 1
Caudofoveata -1-00?? 1 1 1— 1 0-17111-17-1 1 1 11 1
Solenogaatrea -1-000? 1 1-1-1—1 0-l?lll-l?-l 1 1 11 1-1-1-
Polyplacophora 11-000? 11--7-111 — 1 0-11111-11-1 1—11 1 1-1-1-
Conchifera 11-000 11 111—1 0-11111-11-1 P-P1--11 P-l 1-1-1-
Echlura 0—1 7—1-1-1—7-1-1 ?-l 1 11 1
Pogonophora ?1 o P—1-? 1 ?-l 1 — 1 11 1
Polychaeta 0—1—1-1 1-1-1—0—111—1P1—11—1 1 11—11
Clltellata 0—1 7-1-1—0—11 1 11-1 11 1
Onychophora 00—1—17 1-111—7—11 11—1-11 1-11-1-11
Crustacea 0—00—11-1111-111 1-111—0—111 1-1111—1111 1111—11-11-1-1
Uniramla 0—00—11-1111-111 1-111—0—111 1-1111—1-11 1111—11-11-1 —
Chelicerata 0—00—11-1111-111 1-111—0—111 1-11—1111 11—11-11-1-1
Tardigrada 00—11-71? 1-1 7—111 7—71—1 1 1-171-11
Kinorhyncha 0—1-7-17? 1 ?-l 7—11 7 1—1 1 71111—1
FIGURE 3. Data matrix of 141 morphological characters (Appendix 1) for 26 taxa selected for their relevance
to our investigation of higher spiralian relationships. All characters are binary, assigned states of 0, 1, P
(polymorphic; 0 and 1), N (inapplicable), or ? (unknown). Inapplicable states were analyzed as if they were
unknown. The 141 columns correspond to the character numbers used in Appendix 1; dashes denote a match
to the observed state in the first taxon, Cnidaria.
— Cnidaria larger clade (node 4), which (listed in order

Chordata of decreasing proximity to arthropods) in-
— Echinodermata cludes Onychophora, Tardigrada, Kino-
— Phoronida rhyncha, and Nematoda, with priapulans
— Brachiopoda as the next outgroup in two of the six most-
— Nematoda parsimonious trees. Opposing this clade is
Onychophora a large sister clade that includes all other
Crustacea protostomous phyla, including a flat-
- Uniramia worm-gnathostomulid clade (node 10),
81 Nemertea, and a remaining clade referred
Chelicerata to hereafter as Eutrochozoa (node 13).
!— Tardigrada Within this clade, molluscs are the sister
I— Kinorhyncha taxon to a group that includes annelids and
Acoelomorpha other allied worms. Nemertea is the sister

io L
L Rn Rhabditophora taxon to the mollusc-annelid clade, to-
Gnathostomulida gether forming a clade that is the sister
r- Nemertea taxon to a flatworm-gnathostomulid clade.
Sipuncula Lists of synapomorphies that unite Eu-
r Echiura trochozoa (Fig. 4: clade 13) are presented
15 I— Pogonophora in Appendix 2. Three unambiguous syn-
13
Polychaeta apomorphies (characters 40, 44, and 49)
171— Clitellata unite taxa in this node. Strong support for
Caudofoveata the Eutrochozoa hypothesis was most ev-
Lrs'
18 Solenogastres ident from searches constrained to keep
Annelida and Arthropoda together as sis-
is] r Polyplacophora
2 0 l IConchifera ter taxa (Table 1: analysis 1 vs. 2-5). These
I Priapula Articulata-constrained trees required 9-16
FIGURE 4. Strict consensus phylogram of the six additional steps, depending on how inclu-
minimum-length trees resulting from maximum par- sive we made "annelid" and "arthropod"
simony analysis of the matrix in Figure 3. All 141 clades. A 5+14 constraint (enforcing a
characters (Appendix 1) for 26 taxa were binary and
treated as undirected, with Cnidaria declared as an
clade including all taxa encompassed by
outgroup to root each otherwise unrooted network clades 5 and 14 of Fig. 4 but not requiring
topology. The six trees (Fig. 4) summarized in this clades 5 and 14 to remain monophyletic)
consensus diagram each had a length of 384, a con- required only nine additional steps (Table
sistency index of 0.458, a homoplasy index of 0.651, 1: analysis 6). The shortest of these trees,
and a retention index of 0.684. The ingroup (1) and
all interior (2-20) nodes are labeled, and the apo- however, would require reversal(s) of seg-
morphy hypotheses supporting each interior node mentation in Sipuncula and Echiura to
are presented in Appendix 2. Branch lengths corre- support a common origin for "articulate"
spond to the relative number of apomorphic hypoth- segmentation. Searches constrained to in-
eses supporting each interior node or terminal taxon. clude arthropods (as clades 5, 6, 7, or 8) as
derived eutrochozoans (i.e., within clade
13) minimally required 10 additional steps
thropod clade (Crustacea, Uniramia) in- (Table 1: analysis 10), and even then clades
stead of a "biramous" (Crustacea, Chelicer- 14 and 18 (i.e., molluscs and annelids, etc.)
ata) clade. were joined as sister taxa. Somewhat short-
The results reported by the strict con- er trees were found when Nemertea (Table
sensus diagram (Fig. 4) include support for 1: analyses 11, 12) or Nemertea and clade
a clade (node 3) approximately consisting 10 of Figure 4 (Table 1: analyses 13, 14)
of a conventional grouping of protostome were included in the constraint. Con-
phyla, exclusive of deuterostome-lopho- straints enforced to keep molluscs and ar-
phorate phyla (node 2). The monophyly of thropods together as sister taxa required
arthropods is supported (node 8) within a 13-23 additional steps (Table 1: analyses
15-18), depending again on the composi- posed polytomies (Figs. 5a-k) were re-
tion of the arthropod clade. Enforcing con- solved by branch swapping to minimize
straints to join molluscs and flatworms as required steps on our matrix, these trees
a clade (Table 1: analyses 19-22) required remained substantially longer than the
3-12 extra steps, depending on the inclu- best-fitting hypotheses, requiring no fewer
siveness of "flatworms." The shortest of than 26 additional steps (Table 2: Brusca
these trees (Table 1: analysis 22) included and Brusca, 1990).
Nemertea as the sister taxon to molluscs, a
result that was topologically equivalent to DISCUSSION
that of analysis 23 (Table 1), which was In certain respects, our results were an-
generated by imposing a converse con- ticipated by Ghiselin (1988), whose paper
straint, searching for only those trees that should be consulted for historical aspects
did not satisfy the monophyly of clade 12 of the Articulata versus Eutrochozoa con-
(Fig. 4). Shorter trees, requiring only a sin- troversy. Ghiselin and other authors (Field

gle step more than the minimum-length et al., 1988; Lake, 1990) have argued for the
trees (Table 1: analysis 24), were found in consistency of morphology with their
searching for trees that did not maintain rRNA results supporting the Eutrochozoa
clade 13. All four resulting trees were char- hypothesis. The results reported here con-
acterized by the placement of Nemertea as firm these suggestions by supporting the
the sister taxon to molluscs within an oth- monophyly of Eutrochozoa (Mollusca plus
erwise unmodified clade 13. Annelida plus several less speciose spirali-
As in the constraint analyses, characters an phyla). We base our conclusion on a
from our matrix optimized onto the tree maximum-parsimony analysis of as many
topologies of previously published hy- relevant, putatively independent, mor-
potheses (Table 2, Figs. 5a-k) required phological and embryological characters as
multiple additional steps. The closest al- we could compile. It would take at least
ternative proposals were those of Brusca nine additional ad hoc hypotheses of char-
and Brusca (1990; length = 437) and Pearse acter evolution (Table 1) to claim support
et al. (1987; length = 440), requiring 53 and for Articulata. This analysis is the first to
56 additional steps, respectively, over our recover evidence from congruence in the
six minimum-length hypotheses (length = phylogenetic distribution of morphologi-
384). Part of these length differences can cal and embryological characters, corrob-
be attributed to these authors' support of orating the results of the 18S rRNA se-
Articulata, which requires at least nine ad- quence data (Field et al., 1988, 1989;
ditional steps. Even Nielsen's diagram, Ghiselin, 1988; Patterson, 1989; Raff et al.,
which, based on his text statement "[t]here 1989; Lake, 1990; Eernisse, unpubl. manu-
are so many similarities between sipun- script). The resulting placement of Nemer-
culans, echiurans, annelids and molluscs, tea in close proximity to eutrochozoans is
that these phyla must be closely related" also in accordance with recent 18S rRNA-
(1985:256), was interpreted as support for based comparisons by Turbeville et al.
the Eutrochozoa clade, required an addi- (1992).
tional 75 steps. Another factor contribut- Previous hypotheses of metazoan
ing to these length differences was the ten- branching patterns are all substantially less
dency of various authors to express parsimonious than our own, when opti-
uncertainties as polytomies, which as hy- mized on our data matrix. We acknowledge
potheses of relationship require multiple that comparisons of previous hypotheses
acquisition of traits. This well-known are difficult to interpret for several reasons.
shortcoming of consensus summaries (Mi- For instance, it was often necessary to in-
yamoto, 1985) is illustrated by variation clude certain terminal taxa neglected by
among our six most-parsimonious trees particular authors. The translation of con-
relative to the strict consensus summary ventional evolutionary trees into clado-
(Fig. 4). Yet, even when these authors' pro- grams is further complicated by the dif-
(a) Hyman, 1940-1967 (b)Hadzl, 1953,1963

Cnidaria i — Cnidaria
Chordata ' — Rhabditophora
Echinodermata Chordata
Pogonophora Pogonophora
Phoronida Echinodermata
Brachiopoda Brachiopoda
Sipuncula Phoronida
Caudofoveata Uniramia
Solenogastres Tardigrada
Polyplacophora Onychophora
Conchifera Chelicerata
Echiura Crustacea
Polychaeta Sipuncula
Clitellata Echiura
Onychophora Polychaeta
Crustacea Clitellata
Uniramia Caudofoveata
Chelicerata Solenogastres
Tardigrada Polyplacophora
Nematoda Conchifera
Gnathostomulida Nematoda
Priapula Nemertea

Kinorhyncha Priapula
Rhabditophora Gnathostomulida
Nemertea Kinorhyncha
Acoelomorpha Acoelomorpha
(c) Marcus, 1958 (d) Salvinl-Plawen, 1982,1985
Cnidaria Cnidaria
Chordata Chordata
Echinodermata Echinodermata
Pogonophora Phoronida
Phoronida Brachiopoda
Nematoda Sipuncula
Gnathostomulida Echiura
Priapula Pogonophora
Kinorhyncha Polychaeta
Ir-F
Caudofoveata Onychophora
Solenogastres Crustacea
Polyplacophora Uniramia
Conchifera Chelicerata
Echiura Tardigrada
Polychaeta Clitellata
Clitellata Nematoda
Onychophora Gnathostomulida
Crustacea Priapula
Uniramia Kinorhyncha
Chelicerata Rhabditophora
Tardigrada Nemertea
Acoelomorpha Caudofoveata
Rhabditophora Solenogastres
Nemertea
Sipuncula "-HE Polyplacophora
Conchifera
Brachiopoda Acoelomorpha
(e) Nielsen, 1985,1987 (f) Barnes, 1987
Cnidaria Cnidaria
Chordata Chordata
Phoronida Phoronida
Nematoda Nematoda
Priapula Priapula
Kinorhyncha Kinorhyncha
Rhabditophora Rhabditophora
Nemertea Gnathostomulida
Gnathostomulida Nemertea
Sipuncula
Sipuncula Echiura
Echiura Pogonophora
Caudofoveata Polychaeta
Solenogastres Clitellata
Polyplacophora Onychophora
Conchifera
Pogonophora Crustacea
Polychaeta Uniramia
Clitellata Chelicerata
Onychophora Caudofoveata
Crustacea Solenogastres
Uniramia Polyplacophora
Tardigrada Conchifera
Chelicerata Tardigrada
FIGURE 5. Summary of topologies translated from published figures into cladogram hypotheses. Taxon
composition of each cladogram has been adjusted to compare to Figure 4. Tree statistics, optimized on the
(g) Pearse at al., 1987 (h) Brusca and Brusca, 1990

Cnidaria Cnidaria
Chordata Chordata
Echinodermata Phoronida
Brachiopoda Echinodermata
Nematoda Nematoda
Rhabditophora Rhabditophora
Gnathostomulida Gnathostomulida
Nemertea Nemertea
Sipuncula Sipuncula
Caudofoveata Caudofoveata
Solenogastres Solenogastres
Polyplacophora Polyplacophora
Conchifera Conchifera
I Echiura Echiura
I l _ _ i — Polychaeta Pogonophora
I — ' — Clitellata Polychaeta
Onychophora Clitellata
Crustacea Onychophora
Uniramia Crustacea
Uniramia

Chelicerata
Pogonophora Chelicerata
Priapula Tardigrada
Tardigrada Priapula
Kinorhyncha Kinorhyncha
(')Kozloff,1990 Cnidaria (l)Wlllmer, 1990

Cnidaria
Chordata Chordata
Phoronida Phoronida
Nematoda Nematoda
Priapula Acoelomorpha
Kinorhyncha Rhabditophora
Sipuncula Gnathostomulida
Caudofoveata Nemertea
Solenogastres Priapula
Polyplacophora Sipuncula
Conchifera Caudofoveata
Echiura Solenogastres
Pogonophora Polyplacophora
Polychaeta Conchifera
Clitellata Echiura
Onychophora Pogonophora
Crustacea Polychaeta
Uniramia Clitellata
Chelicerata Uniramia
Tardigrada Onychophora
Acoelomorpha Crustacea
Rhabditophora Chelicerata
Gnathostomulida Tardigrada
Nemertea Kinorhyncha
(k) Meglltsch and Schram, 1991: their Fig. 38.2 (I) Meglltsch and Schram, 1991: herein
Cnidaria Cnidaria
Chordata Chordata
Brachiopoda Phoronida
Nemertea Nematoda
Sipuncula Priapula
Echiura Kinorhyncha
Caudofoveata Acoelomorpha
Solenogastres Rhabditophora
Polyplacophora Gnathostomulida
Conchifera Nemertea
Pogonophora Sipuncula
Onychophora Caudofoveata
Uniramia Solenogastres
Chelicerata Polyplacophora
Tardigrada Conchifera
Crustacea Echiura
Polychaeta Pogonophora
Clitellata Polychaeta
Acoelomorpha Clitellata
Rhabditophora Onychophora
Gnathostomulida Crustacea
Nematoda Uniramia
Priapula Chelicerata
Kinorhyncha Tardigrada
matrix in Figure 3, are presented in Table 2. Citations are indicated above each figure (a-k). Figure 51
corresponds to Figure 2a.
ferent concepts of evolutionary entities age mere comparison of our results with
held by each author. Taxa considered an- those inferred from molecules. Like Kluge
cestral to other higher taxa were treated as (1989) and Eernisse (unpubl. manuscript),
terminal taxa to make comparison with our we advocate that ultimately, the best esti-
results meaningful. Because of these con- mates of branching order for metazoans
cerns, Figures 5a-k cannot be used to judge will come from combined analyses of all
the relative merit or worth of these works. relevant data, whether classical in nature
Rather, they are presented to make eval- or a site-by-site comparison of nucleic acid
uation of the relative support for different or peptide sequences. Ghiselin (1988) ob-
tree topologies possible. These disclaimers served that molecules have morphology
aside, we believe we have established a and can be analyzed as such. In this spirit,
more rigorous precedent for reporting sup- we have coded variation in the structure
port for alternative hypotheses of meta- of the hemocyanin molecule as characters

zoan phylogeny. 71 and 72.
Because we were interested in the ques- One consequence of character congru-
tion of spiralian interrelationships, little ence analyses is that specific homoplasies
effort was made to collect characters lack- required to support alternative views are
ing variation among these taxa. Moreover, highlighted. To further judge the confi-
beyond affirming the monophyly of Mol- dence of the Eutrochozoa hypothesis, the
lusca and Arthropoda, relatively little of structure of the present data could be an-
the enormous literature bearing on these alyzed to discern whether the homoplastic
groups was included in this analysis. Our characters are randomly and indepen-
analysis was unable to resolve some im- dently distributed or produced by conflict-
portant aspects of metazoan phylogeny. In ing suites of characters. This issue is only
particular, our lack of resolution of deu- beginning to be addressed in cladistic
terostome-lophophorate relationships re- studies (Farris, 1991) but would be a nat-
flects the largely unsettled nature of the ural extension of the present analysis.
corresponding literature (cf. Jefferies, 1986; Because the character congruence ap-
Nielsen, 1987; Ax, 1989; Brusca and Brusca, proach assumes character independence,
1990; Kozloff, 1990), as is the case for re- this analysis is subject to the criticism that
lationships among arthropod subphyla (cf. a suite of ciliary characters have been lost
Brusca and Brusca, 1990; Emerson and as a unit in the arthropods in association
Schram, 1990). This lack of resolution may with yolk-rich development. For example,
largely be due to a combination of sam- at least two of the three characters (40, 44)
pling biases, especially the emphasis of unambiguously optimized to diagnose Eu-
variation among basal spiralian lineages trochozoa (Fig. 4: node 13) are features as-
and the inclusion of the particular phyla sociated with early larval development and
used in this study. A similar condition may the presence of cilia. Absence of these
apply to our results with regards to priapu- characters could be interpreted as the re-
lans, kinorhynchs, and nematodes. In ad- sult of a single evolutionary reversal in
dition, these and other relationships may Arthropoda. This hypothesis is appealing
reflect the biases of particular sources used; because many of the differences between
for example, the characters used to diag- arthropods and other metameric spiralians
nose clade 10 (Fig. 4) are mainly those em- could be explained by relatively few and
phasized by Ax (1985). Further research simple evolutionary changes. The pres-
should be directed to the careful addition ence of flagella in crustacean and pycno-
of more relevant character evidence, in- gonid sperm (Franzen, 1987) and of cili-
cluding fossil and molecular sequence data ated gonoducts and nephridial segmented
not included in the present study. organs in onycophorans (Boudreaux, 1979)
We anticipate such a combination of data indicates that cilia were not entirely lost.
recovered from morphological and molec- A similar argument could be made for the
ular studies. We do not, however, encour- position of chordates as the sister taxon to
other deuterostomes in four of the six min- luscs, sipunculans, and echiurans descend-
imum-length trees of Figure 4 as the result ed from a segmented ancestor, our under-
of the difficulty of comparing larval char- standing of the biology of these groups
acters in these taxa. Although considera- will be deepened. We hope that the data
tion of all the data available for this anal- matrix compiled for this study will serve
ysis does not support these interpretations, as a nucleus from which further discus-
close scrutiny of the characters in question sions of animal relationships can grow.
could lead to a reappraisal of the present
notions of phenotypic similarity.
The support for Eutrochozoa recognized ACKNOWLEDGMENTS
here does not settle the controversy sur- The compilation of a character matrix for spiralian
rounding the extent to which molluscs are metazoans and select nonspiralian outgroups was ini-
tiated in fall 1989 during a graduate seminar led by
primitively metameric. Two important un- one of us (D.J.E.). We thank members of that seminar
certainties remain regarding the identity

group for stimulating discussion and suggestions. Ar-
of the immediate sister taxa to Mollusca nold G. Kluge and James N. Cather provided helpful
and the optimization of overt metamerism, suggestions at various stages of this project. We give
special thanks to Vicki B. and John S. Pearse for their
which may be absent in this group due to many useful observations in review of the data ma-
derived loss or retained plesiomorphy (Fig. trix. Rich Mooi, Malcolm Telford, and an anonymous
1). The lack of apparent metamerism in reviewer also provided many valuable suggestions.
Solenogastres, Caudofoveata, Sipuncula, We thank William R. Dawson, Director of the Mu-
and Echiura presents difficulties for the ar- seum of Zoology, University of Michigan, for use of
facilities.
gument of a primitively metameric eutro-
chozoan. Furthermore, Nemertea may be
more closely aligned to Mollusca than has REFERENCES
been generally appreciated. Of all the con-
straint analyses previously reported, place- ANDERSON, D. T. 1973. Embryology and phylogeny
in annelids and arthropods. Pergamon Press, Ox-
ment of Nemertea as the sister taxon to ford, England.
Mollusca required only one additional step Ax, P. 1985. The position of the Gnathostomulida
(Table 1: analysis 24), yet we see no reason and Platyhelminthes in the phylogenetic system of
to prefer this less-parsimonious arrange- the Bilateria. Pages 168-180 in The origins and re-
ment to our best-fitting hypotheses. lationships of lower invertebrates (S. Conway Mor-
ris, J. D. George, R. Gibson, and H. M. Platt, eds.).
If annelids and arthropods are not sister Oxford Univ. Press, Oxford, England.
taxa, then similarities of their segmentation Ax, P. 1989. Basic phylogenetic systematization of
must have been much more ancient, or in the Metazoa. Pages 229-245 in The hierarchy of life
many respects independently derived, or (B. Fernholm, K. Bremer, and H. Jornvall, eds.).
Elsevier Science B.V., Amsterdam.
both. There is now a rapidly advancing AZARIAH, J. 1973. Studies on the cephalochordates
knowledge of the genetic and develop- of the Madras coast. 15. The nature of the structural
mental bases of a basic level of metamerism polysaccharide in amphioxus, Branchiostoma lanceo-
in metazoans as diverse as insects, verte- latum. Acta Histochem. 46:10-17.
brates, sea urchins, brachiopods, molluscs, BAKKE, T. 1980. Embryonic and post-embryonic de-
velopment in the Pogonophora. Zool. Jahrb. Anat.
and leeches, especially involving homeo- 103:276-284.
domain-encoding ("homeobox") gene BARNES, R. D. 1987. Invertebrate zoology, 5th edi-
clusters (Holland and Hogan, 1986; Gould, tion. W. B. Saunders, Philadelphia.
1991; Wedeen et al., 1991). This burst of BATESON, W. 1894. Materials for study of variation.
activity is leading to a reevaluation of the Macmillan, London.
BEKLEMISHEV, W. N. 1969. Principles of comparative
basic underlying mechanisms of segmen- anatomy of invertebrates, Volumes 1, 2 (Z. Kabata,
tation and a search for likely homologies. ed.). Univ. Chicago Press, Chicago.
Homeotic homologies, extending to the se- BEREITER-HAHN, J., A. G. MATOLTSY, AND K. S. RICH-
quence and functional level, will probably ARDS (eds.). 1984. Biology of the integument, Vol-
be hypothesized for molluscs and other ume 1. Invertebrates. Springer-Verlag, Berlin.
BOUDREAUX, H. B. 1979. Arthropod phylogeny, with
phyla. If further analyses and new data special reference to insects. Wiley, New York.
continue to support the notion that mol- BOYER, B. C. 1971. Regulative development in a spi-
ralian embryo as shown by cell deletion experi- FARRIS, J. S. 1991. Excess homoplasy ratios. Cladistics
ments on the acoel, Childia. J. Exp. Zool. 176:97-106. 7:81-91.
BOYER, B. C. 1989. The role of the first quartet mi- FARRIS, J. S., A. G. KLUGE, AND M. J. ECKARDT. 1970.
cromeres in the development of the polyclad Ho- A numerical approach to phylogenetic systematics.
ploplana inquilina. Biol. Bull. 177:338-343. Syst. Zool. 19:172-189.
BROWN, C. H. 1975. Structural materials in animals. FAUTIN, D. G., AND R. N. MARISCAL. 1991. Cnidaria:
Wiley, New York. Anthozoa. Pages 267-358 in Microscopic anatomy
BRUSCA, R. C, AND G. J. BRUSCA. 1990. Invertebrates. of invertebrates, Volume 2. Placozoa, Porifera, Cni-
Sinauer, Sunderland, Massachusetts. daria, and Ctenophora (F. W. Harrison and J. A.
CARTER, J. G. (ed.). 1990. Skeletal biomineralization: Westfall, eds.). Wiley-Liss, New York.
Patterns, processes, and evolutionary trends, Vol- FELSENSTEIN, J. 1988. Phylogenies from molecular
umes 1, 2. Van Nostrand Reinhold, New York. sequences: Inference and reliability. Annu. Rev.
CATHER, J. N. 1971. Cellular interactions in the reg- Genet. 22:521-565.
ulation of development in annelids and molluscs. FIELD, K. G., G. J. OLSEN, S. J. GIOVANNONI, E. C. RAFF,
Adv. Morphog. 9:67-125. N. R. PACE, AND R. A. RAFF. 1989. Phylogeny and
CHRISTEN, R., A. RATTO, A. BAROIN, R. PERASSO, K. G. molecular data. Science 243:550-551.
FIELD, K. G., G. J. OLSEN, D. J. LANE, S. J. GIOVANNONI,

GRELL, AND A. ADOUTTE. 1991. An analysis of the
origin of metazoans, using comparisons of pal se- M. T. GHISELIN, E. C. RAFF, N. R. PACE, AND R. A.
quences of the 28S RNA, reveals an early emer- RAFF. 1988. Molecular phylogeny of the animal
gence of triploblasts. EMBO J. 10:499-503. kingdom. Science 239:748-753.
CLARK, R. B. 1964. Dynamics in metazoan evolution: FRANZEN, A. 1987. Spermatogenesis. Pages 1-45 in
The origin of the coelom and segments. Clarendon Reproduction of marine invertebrates, Volume 9.
Press, Oxford, England. General aspects: Seeking unity in diversity (A. C.
CLARK, R. B. 1979. Radiation of the Metazoa. Pages Giese, J. S. Pearse, and V. B. Pearse, eds.). Blackwell
55-102 in The origin of major invertebrate groups Scientific/Boxwood Press, Pacific Grove, California.
(M. R. House, ed.). Systematics Association Special FREEMAN, G., AND J. W. LUNDELIUS. 1992. Evolution-
Volume 12. Clarendon Press, Oxford, England. ary implications of the mode of D quadrant speci-
COSTELLO, D. P., AND C. HENLEY. 1976. Spiralian de- fication in coelomates with spiral cleavage. J. Evol.
velopment: A perspective. Am. Zool. 16:277-291. Biol. 5:205-247.
CUVIER, G. 1817. Le regne animal distribue d'apres FRETTER, V., AND M. A. GRAHAM. 1962. British proso-
son organisation, pour servir de base a l'histoire branch molluscs. Ray Society, London.
naturelle des animaux et d'introduction a l'anato- GANS, C, AND R. G. NORTHCUTT. 1983. Neural crest
mie comparee, Volume 2. Deterville, Paris. and the origin of vertebrates: A new head. Science
DALY, H. V., J. T. DOYEN, AND P. R. EHRLICH. 1978. 220:268-274.
Introduction to insect biology and diversity. Mc- GHISELIN, M. T. 1974. A radical solution to the spe-
Graw-Hill, New York. cies problem. Syst. Zool. 23:536-544.
DE BEER, G. R. 1930. Embryology and evolution. GHISELIN, M. T. 1988. The origin of molluscs in the
Clarendon Press, Oxford, England. light of molecular evidence. Ox. Surv. Evol. Biol.
DEMUYNCK, S., P. SAUTIERE, J. VAN BEEUMEN, AND N. 5:66-95.
DHAINAUT-COURTOIS. 1991. Homologies entre les GHISELIN, M. T. 1989. Summary of our present
hemerythrines des sipunculiens et une metallopro- knowledge of metazoan phylogeny. Pages 261-272
teine complexant le cadmium (MP II) d'une anne- in The hierarchy of life (B. Fernholm, K. Bremer,
lide polychete, Nereis diversicolor. C.R. Acad. Sci. Ser. and H. Jornvall, eds.). Elsevier Science B.V., Am-
Ill Sci. Vie 312:317-322. sterdam.
EERNISSE, D. J. 1992. DNA Translator and Aligner: GHISELIN, M. T. 1991. Classical and molecular phy-
HyperCard utilities to aid phylogenetic analysis of logenetics. Boll. Zool. 58:289-294.
molecules. Comput. Appl. Biosci. 8:177-184. GIESE, A. C, AND J. S. PEARSE (eds.). 1975. Repro-
EERNISSE, D. J., AND K. KERTH. 1988. The initial stages duction of marine invertebrates, Volume 3. An-
of radular development in chitons (Mollusca: Poly- nelids and echiurans. Academic Press, New York.
placophora). Malacologia 28:95-103. GOLDBERG, W. M. 1976. Comparative study of the
EHLERS, U. 1985. Phylogenetic relationships within chemistry and structure of gorgonian and antipa-
the Platyhelminthes. Pages 143-158 in The origins tharian coral skeletons. Mar. Biol. 35:253-267.
and relationships of lower invertebrates (S. Con- GOULD, S.J. 1991. Of mice and mosquitoes. Nat. Hist.
way Morris, J. D. George, R. Gibson, and H. M. Platt, July:12-20.
eds.). Systematics Association Special Volume 28. GROBBEN, K. 1909. Die systematische Einteilung des
Clarendon Press, Oxford, England. Tierreichs. Verh. Zool.-Bot. Ges. Wien 58:491-511.
EHLERS, U. 1986. Comments on a phylogenetic sys- GROSBERG, R. K. 1990. Out on a limb: Arthropod
tem of the Platyhelminthes. Hydrobiologia 132:1- origins. Science 250:632-633.
12. GRUNER, H. E. 1982. Stamm Sipunculida. Pages 246-
EMERSON, M. J., AND F. R. SCHRAM. 1990. The origin 260 in Lehrbuch der Speziellen Zoologie, 4th edi-
of crustacean biramous appendages and the evo- tion, 1 Band, 3. Teil (A. Kaestner, ed.). Gustav Fi-
lution of Arthropoda. Science 250:667-669. scher, Stuttgart.
HADFIELD, M. G. 1979. Aplacophora. Pages 1-25 in LAKE, J. A. 1990. Origin of the Metazoa. Proc. Natl.
Reproduction of marine invertebrates, Volume 5. Acad. Sci. USA 87:763-766.
Molluscs: Pelecypods and lesser classes (A. C. Giese LANG, W. H. 1988. cDNA cloning of the Octopus
and J. S. Pearse, eds.). Academic Press, New York. dofleini hemocyanin: Sequence of the carboxyl-ter-
HADZI, J. 1953. An attempt to reconstruct the system minal domain. Biochemistry 27:7276-7282.
of animal classification. Syst. Zool. 2:145-154. LANG, W. H., AND K. E. VAN HOLDE. 1991. Cloning
HADZI, J. 1963. The evolution of the Metazoa. Mac- and sequencing of Octopus dofleini hemocyanin
millan, New York. cDNA: Derived sequences of functional units Ode
HAECKEL, E. 1866. Generelle morphologie der or- and Odf. Proc. Natl. Acad. Sci. USA 88:241-248.
ganismen, 2 volumes. Georg Reimer, Berlin. LEMCHE, H. 1959a. Molluscan phylogeny in the light
HALL, B. K. 1992. Evolutionary developmental bi- of Neopilina. Proc. XV Int. Congr. Zool., pages 380-
ology. Chapman and Hall, London. 381.
HANSON, E. D. 1977. The origin and early evolution LEMCHE, H. 1959b. Protostomian interrelationships
of animals. Wesleyan Univ. Press, Middletown, in the light of Neopilina. Proc. XV Int. Congr. Zool.,
Connecticut. pages 381-389.
HASZPRUNAR, G. 1985a. The fine morphology of the LlNZEN, B., N. M. SOETER, A. F. RlGGS, H. J. SCHNEIDER,

osphradial sense organs of the Mollusca. I. Gas- W. SCHARTAU, M. D. MOORE, E. YOKOTA, P. Q. BEH-
tropoda—Prosobranchia. Philos. Trans. R. Soc. Lond. RENS, H. NAKASHIMA, T. TAKAGI, T. NEMOTO, J. M.
B 307:457-496. VEREIJKEN, H. J. BAK, J. J. BEINTEMA, A. VOLBEDA, W.
HASZPRUNAR, G. 1985b. The fine morphology of the P. J. GAYKEMA, AND W. G. J. HOL. 1985. The struc-
osphradial sense organs of the Mollusca. II. Allo- ture of arthropod hemocyanins. Science 229:519-
gastropoda (Architectonicidae and Pyramidelli- 524.
dae). Philos. Trans. R. Soc. Lond. B 307:497-505. LIPSCOMB, D. L. 1989. Relationships among the eu-
HASZPRUNAR, G. 1987. The fine morphology of the karyotes. Pages 161-178 in The hierarchy of life (B.
osphradial sense organs of the Mollusca. III. Pla- Ferholm, K. Bremer, and H. Jornvall, eds.). Elsevier
cophora and Bivalvia. Philos. Trans. R. Soc. Lond. Science B.V., Amsterdam.
B 315:37-61. LUTZ, P. E. 1986. Invertebrate zoology. Benjamin/
HASZPRUNAR, G. 1988. On the origin and evolution Cummings, Menlo Park, California.
of major gastropod groups, with special reference MANGUM, C. P. 1985. Oxygen transport in inverte-
to the Streptoneura. J. Molluscan Stud. 54:367-441. brates. Am. J. Physiol. 248:505-514.
HATSCHEK, B. 1891. Lehrbuch der Zoologie, 3 Lie- MANGUM, C. P., K. I. MILLER, J. L. SCOTT, K. E. VAN
ferung. Wilhelm Engelmann, Leipzig. HOLDE, AND M. P. MORSE. 1987. Bivalve hemo-
HENNIG, W. 1966. Phylogenetic systematics. Univ. cyanin: Structural, functional, and phylogenetic re-
Illinois Press, Urbana. lationships. Biol. Bull. 173:205-221.
HOLLAND, P. W. H., AND B. L. M. HOGAN. 1986. Phy- MANGUM, C. P., J. L. SCOTT, R. E. BLACK, K. I. MILLER,
logenetic distribution of Antennapedia-like homeo AND K. E. VAN HOLDE. 1985. Centipedal hemo-
boxes. Nature 321:251-253. cyanin: Its structure and its implications for ar-
HUXLEY, T. H. 1869. An introduction to the classi- thropod phylogeny. Proc. Natl. Acad. Sci. USA 82:
fication of animals. Churchill, London. 3721-3725.
HYMAN, L. 1940-1967. The invertebrates, Volumes MANTON, S. M., AND D. T. ANDERSON. 1979. Poly-
1-6. McGraw-Hill, New York. phyly and the evolution of the arthropods. Pages
HYMAN, L. 1951. The invertebrates, Volume 2. Platy- 269-322 in The origin of major invertebrate groups
helminthes and Rhynchocoela. McGraw-Hill, New (M. R. House, eds.). Systematics Association Special
York. Volume 12. Clarendon Press, Oxford, England.
HYMAN, L. 1958. The occurrence of chitin in the MARCUS, E. 1958. On the evolution of the animal
lophophorate phyla. Biol. Bull. 114:106-112. phyla. Q. Rev. Biol. 33:24-58.
IVANOV, A. V. 1963. Pogonophora. Academic Press, MEGLITSCH, P. A. 1972. Invertebrate zoology, 2nd
London. edition. Oxford Univ. Press, New York.
IVANOV, A. V. 1988. Analysis of the embryonic de- MEGLITSCH, P. A., AND F. R. SCHRAM. 1991. Inver-
velopment of Pogonophora in connection with the tebrate zoology, 3rd edition. Oxford Univ. Press,
problems of phylogenetics. Z. Zool. Syst. Evolu- New York.
tionsforsch. 26:161-185. MIYAMOTO, M. M. 1985. Consensus cladograms and
JAMEISON, B. G. M. 1988. On the phylogeny and general classifications. Cladistics 1:186-189.
higher classification of the Oligochaeta. Cladistics NAEF, A. 1913. Studien zur generellen Morphologie
4:367-410. der Mollusken, 2. Teil. Das Colomsytem in seinen
JEFFERIES, R. P. S. 1986. The ancestry of the verte- topographischen Beziehungen. Ergeb. Zool. 3:329-
brates. British Museum (Natural History), London. 462.
KLUGE, A. G. 1989. A concern for evidence and a NAEF, A. 1924. Studien zur generellen Morphologie
phylogenetic hypothesis of relationships among der Mollusken, 3. Teil. Die typischen Beziehungen
Epicrates (Boidae, Serpentes). Syst. Zool. 38:7-25. der Weichtierklassen untereinander und das Ver-
KOZLOFF, E. N. 1990. Invertebrates. W. B. Saunders, haltnis ihrer Urformen zu anderen Colomaten. Er-
Philadelphia. geb. Zool. 6:28-124.
NELSON, B. H., AND D. A. WEISBLAT. 1991. Conver- of metazoan circulatory systems. Zoomorphology
sion of ectoderm to mesoderm by cytoplasmic ex- 103:193-208.
trusion in leech embryos. Science 253:435-438. RUSSELL-HUNTER, W. D. 1988. The gills of chitons
NELSON, D. R. (ed.). 1982. Proceedings of the third (Polyplacophora) and their significance in mollus-
international symposium on Tardigrada. East Ten- can phylogeny. Am. Malacol. Bull. 6:69-78.
nessee State Univ. Press, Johnson City. RUSSELL-HUNTER, W. D., AND S. C. BROWN. 1965.
NIELSEN, C. 1985. Animal phylogeny in the light of Ctenidial number in relation to size in certain chi-
the trochaea theory. Biol. J. Linn. Soc. 25:243-299. tons, with a discussion of its phyletic significance.
NIELSEN, C. 1987. Structure and function of meta- Biol. Bull. 128:508-521.
zoan ciliary bands and their phylogenetic signifi- SALVINI-PLAWEN, L. V* 1978. On the origin and evo-
cance. Acta Zool. 68:205-262. lution of the lower Metazoa. Z. Zool. Syst. Evolu-
PATTERSON, C. (ed.). 1982. Methods of phylogenetic tionsforsch. 16:40-88.
reconstruction. Zool. J. Linn. Soc. 74:197-344. SALVINI-PLAWEN, L. V. 1980a. A reconsideration of
PATTERSON, C. 1989. Phylogenetic relations of major systematics in the Mollusca (phylogeny and higher
groups: Conclusions and prospects. Pages. 471-488 classification). Malacologia 19:249-278.
in The hierarchy of life (B. Fernholm, K. Bremer, SALVINI-PLAWEN, L. V. 1980b. Was ist eine Trocho-
and H. Jornvall, eds.). Elsevier Science B.V., Am- phora?—Eine analyse der larventypen mariner pro-

sterdam. tostomier. Zool. Jahrb. Abt. Anat. Ontog. Tiere 103:
PATTERSON, C. 1990. Reassessing relationships. Na- 389-423.
ture 344:199-200. SALVINI-PLAWEN, L. V. 1982. A paedomorphic origin
PAULUS, H. F. 1979. Eye structure and the mono- of the oligomerous animals? Zool. Scr. 11:77-81.
phyly of Arthropoda. Pages 299-377 in Arthropod SALVINI-PLAWEN, L. V. 1985. Early evolution and the
phylogeny (A. P. Gupta, ed.). Van Nostrand Rein- primitive groups. Mollusca 10:59-150.
hold, New York. SALVINI-PLAWEN, L. V. 1988. Annelida and Mollus-
PEARSE, V., J. PEARSE, M. BUCHSBAUM, AND R. ca—A prospectus. Pages 383-396 in The ultrastruc-
BUCHSBAUM. 1987. Living invertebrates. Boxwood ture of Polychaeta (W. Westheide and C. O. Her-
Press, Pacific Grove, California. mans, eds.). Gustav Fischer Verlag, Stuttgart.
PELSENEER, P. 1899. Recherches morphologiques et SANNASI, A., AND H. R. HERMANN. 1970. Chitin in
phylogenetiques sur les mollusques archaiques. the Cephalochordata, Branchiostoma floridae. Experi-
Mem. Sav. Etr. Cour. Acad. R. Belg. 57(3):1-112. entia 26:351-352.
PELSENEER, P. 1906. Mollusca. Pages 1-354 in A trea- SCHELTEMA, A. H. 1988. Ancestors and descendents:
tise on zoology, Volume 5 (E. R. Lankester, ed.). Relationships of the Aplacophora and Polypla-
Adam and Charles Black, London. cophora. Am. Malacol. Bull. 6:57-68.
RAFF, R. A., K. G. FIELD, G. J. OLSEN, S. J. GIOVANNONI, SCHLEIP, W. 1929. Die determination der primitiv-
D. J. LANE, M. T. GHISELIN, N. R. PACE, AND E. C. entwicklung. Akademische Verlagsgesellschaft,
RAFF. 1989. Metazoan phylogeny based on anal- Leipzig.
ysis of 18S ribosomal RNA. Pages 247-260 in The SCHRAM, F. R. 1986. Crustacea. Oxford Univ. Press,
hierarchy of life (B. Fernholm, K. Bremer, and H. New York.
Jornvall, eds.). Elsevier Science B.V., Amsterdam. SCHRAM, F. R. 1991. Cladistic analysis of metazoan
REMANE, A. 1956. Die Grundlagen des naturlichen phyla and the placement of fossil problematica.
Systems der vergleichenden Anatomie und Phy- Pages 35-46 in The early evolution of Metazoa and
logenetik, 2nd edition. Geest und Portig K. G., Leip- the significance of problematic taxa (A. M. Simo-
zig. nette and S. Conway Morris, eds.). Cambridge Univ.
RICE, M. E., AND M. TODOROVIC (eds.). 1975-1976. Press, Cambridge, England.
Proceedings of the international symposium on the SMITH, A. B. 1984. Classification of the Echinoder-
biology of the Sipuncula and Echiura, Volumes 1 mata. Paleontology 27:431-459.
(1975), 2 (1976). Nauchno Delo Press, Belgrade, Yu- SMITH, J. P. S., Ill, S. TYLER, AND R. M. RIEGER. 1986.
goslavia, and the Smithsonian Institution, Wash- Is the Turbellaria polyphyletic? Hydrobiologia 132:
ington, D.C. 13-21.
RICHARDSON, D. E., M. EMAD, R. C. REEM, AND E. I.
STASEK, C. 1972. The molluscan framework. Pages
SOLOMON. 1987. Allosteric interactions in sipun-
1-44 in Chemical zoology, Volume 7 (M. Florkin
culid and brachiopod hemerythrins. Biochemistry
and B. T. Scheer, eds.). Academic Press, London.
26:1003-1013.
RiEDL, R. J. 1969. Gnathostomulida from America. STRATHMANN, M. F. 1987. Reproduction and devel-
Science 163:445-452. opment of marine invertebrates of the northern
ROSEN, M. D., C. R. STASEK, AND C. O. HERMANS. 1979.
Pacific coast. Univ. Washington Press, Seattle.
The ultrastructure and evolutionary significance of STRATHMANN, R. R. 1978. The evolution and loss of
the ocelli in the larva of Katharina tunicata (Mol- feeding larval stages of marine invertebrates. Evo-
lusca: Polyplacophora). Veliger 22:171-176. lution 32:894-906.
RUNNEGAR, B. 1983. Molluscan phylogeny revisited. SWOFFORD, D. L. 1990. PAUP: Phylogenetic analysis
Mem. Assoc. Aust. Palaeontol. 1:121-144. using parsimony, Version 3.0. Illinois Natural His-
RUNNEGAR, B., AND J. POJETA. 1985. Origin and di- tory Survey, Champaign.
versification of the Mollusca. Mollusca 10:1-57. SWOFFORD, D. L., AND G. J. OLSEN. 1990. Phylogeny
RUPPERT, E. E., AND K. J. CARLE. 1983. Morphology reconstruction. Pages 411-501 in Molecular system-
atics (D. M. Hillis and C. Moritz, eds.)- Sinauer, erythrin alpha subunit from Lingula unguis. J. Bio-
Sunderland, Massachusetts. chem. (Tokyo) 110:376-380.
TAKAGI, T., AND J. A. COX. 1991. Primary structure
of myohemerythrin from the annelid Nereis diver- Received 29 October 1991; accepted 19 January 1992
sicolor. FEBS Lett. 286:25-27.
THOMAS, M. B., AND N. C. EDWARDS. 1991. Cnidaria:
APPENDIX 1
Hydrozoa. Pages 91-183 in Microscopic anatomy of
invertebrates, Volume 2. Placozoa, Porifera, Cni- Sources used are listed by character with the ex-
daria, and Ctenophora (F. W. Harrison and J. A. ception of certain taxon-specific references, including
Westfall, eds.). Wiley-Liss, New York. those pertaining to Cnidaria (Fautin and Mariscal,
TIDBALL, J. G. 1984. Cnidaria: Secreted surface. Pages 1991; Thomas and Edwards, 1991), Chordata (Gans
69-78 in Biology of the integument, Volume 1. In- and Northcutt, 1983), Echinodermata (Smith, 1984),
vertebrates (J. Bereiter-Hahn, A. G. Matoltsy, and Platyhelminthomorpha (Ax, 1985; Ehlers, 1985), Mol-
K. S. Richards, eds.). Springer-Verlag, Berlin. lusca (Runnegar and Pojeta, 1985; Wingstrand, 1985;
TURBEVILLE, J. M. 1991. Nemertinea. Pages 285-328 Haszprunar, 1988), Annelida (Giese and Pearse, 1975;
in Microscopic anatomy of invertebrates, Volume Jameison, 1988), Echiura and Sipuncula (Giese and
Pearse, 1975; Rice and Todorovic, 1975-1976; Strath-

3. Platyhelminthes and Nemertinea (F. W. Harrison
and B. J. Bogitsh, eds.). Wiley-Liss, New York. mann, 1978), Crustacea (Schram, 1986; Brusca and
Brusca, 1990), Uniramia (Daly et al., 1978; Boudreaux,
TURBEVILLE, J. M., K. G. FIELD, AND R. A. RAFF. 1992.
1979), and Chelicerata (Meglitsch and Schram, 1991).
Phylogenetic position of phylum Nemertini, in-
ferred from 18S rRNA sequences: Molecular data
Characters by System"
as a test of morphological character homology. Mol.
Biol. Evol. 9:235-249. A. Cleavage
TURBEVILLE, J. M., AND E. E. RUPPERT. 1985. Com- 1. Spiral quartet cleavage; cells B and D with
parative ultrastructure and the evolution of nem- ventral transverse contact, cells A and C with
ertines. Am. Zool. 25:53-71. dorsal saggital contact (Wilson, 1898; Cather,
TYLER, S. 1979. Distinctive features of cilia in meta- 1971; Anderson, 1973; Freeman and Lunde-
zoans and their significance for systematics. Tissue lius, 1992) [a/p]
& Cell 11:385-400. 2. Spiral cleavage with nuclear migration15
(Schleip, 1929; Costello and Henley, 1976)
VERDONK, N. H., AND J. A. M. VAN DEN BIGGELAAR.
1983. Early develoment and the formation of the
3. Radial holoblastic cleavage (Wilson, 1898;
germ layers. Pages 91-122 in The Mollusca, Volume
Brusca and Brusca, 1990) [p/a]b
3. Development (N. H. Verdonk, J. A. M. van den
4. Cell fates of primary germ layers fixed by
Biggelaar, and A. S. Tompa, eds.). Academic Press,
end of fifth cleavage (Anderson, 1973; Brusca
New York.
and Brusca, 1990; Meglitsch and Schram, 1991)
VOIT, B., AND G. FELDMAIER-FUCHS. 1990. Arthropod
[a/p]
hemocyanins. J. Biol. Chem. 265:19447-19452.
5. Entomesoblast cell (blastomere 4d) (Beklemi-
VOLBEDA, A., AND W. G. J. HOL. 1989. Pseudo 2-fold shev, 1969; Boudreux, 1979, Verdonk and
symmetry in the copper-binding domain of arthro- Biggelaar, 1983) [a/p]
podan haemocyanins. J. Mol. Biol. 206:531-546.
6. Entomesoblast proliferation into paired an-
WEDEEN, C. J., D. J. PRICE, AND D. A. WEISBLAT. 1991. terior coelomic sacs (Brusca and Brusca, 1990)
Cloning and sequencing of a leech homolog to the [a (noncavitated mesenchyme)/p]
Drosophila engrailed gene. FEBS Lett. 279:300-302. 7. Entomesoblast proliferation contributing to
WHEELER, W. 1990. Review of Invertebrate Relation- mesoderm (Wilson, 1898; Anderson, 1973)
ships, by P. Willmer. J. N.Y. Entomol. Soc. 98:502- [mesoderm absent or forming from ectoderm
504. or archenteron/p]
WILEY, E. O. 1981. Phylogenetics: The theory and 8. Entomesoblast proliferation into paired dor-
practice of phylogenetic systematics. Wiley-Inter- soposterior mesodermal tissue bands (Salvi-
science, New York. ni-Plawen, 1985, 1988; Brusca and Brusca,
WILLMER, P. 1990. Invertebrate relationships. Cam- 1990)[a/p]
bridge Univ. Press, Cambridge, England. 9. Teloblastic segmentation of mesodermal sacs
WILSON, E. B. 1898. Considerations in cell-lineage with pygidial growth zone (Salvini-Plawen,
and ancestral reminiscence. Ann. N.Y. Acad. Sci. 1985, 1988)[a/p]
11:1-27. 10. Epidermal mitosis by parenchymal kineto-
WINGSTRAND, K. G. 1985. On the anatomy and re-
lationships of Recent Monoplacophora. Galathea a
Character numbers 1-141 correspond to columns
Rep. 16:7-94. in Figure 3. Alternative states, in square brackets fol-
WIRTH, U. 1984. Die struktur der Metazoen—Sper- lowing the character description, are listed in order
mien und ihre bedeutung fur die phylogenetik. of states 0 and 1 with no implied polarity (a = absent;
Verh. Naturwiss. Ver. Hambg. 27:295-362. p = present)
b
YANO, H., K. SATAKE, Y. UENO, K. KONDO, AND A. See supplemental notes following listing of char-
TSUGITA. 1991. Amino acid sequence of the hem- acters.
some-containing cells (Ehlers, 1986; Smith et sory centriole (Ehlers, 1986; Nielsen, 1987)
al., 1986) [a/p] [a/p]
11. Apical and intermediate micromere quartet 35. Coordinated cilia with ciliated necklace
form cross (Meglitsch, 1972) [a/p] (Nielsen, 1987) [a/p]
12. Cross pattern (Meglitsch, 1972; Brusca and 36. Motile somatic cilia or flagella (Nielsen, 1987)
Brusca, 1990) [radiate/interradiate] [a/p]
13. Triploblastic tissue organization (Hanson, 37. Chemoreceptor cells with paddle-shaped
1977; Brusca and Brusca, 1990; Meglitsch and discociliab (Haszprunar, 1985a) [a/p]
Schram, 1991; cf. Nelson and Weisblat, 1991)
[a/p] E. Larval
38. Upstream collecting bands of cilia in larvae
B. Coelom with separate cilia on monociliate cellsb
14. Bilaterally paired coelomic anlagenb (Brusca (Nielsen, 1987) [a/p]
and Brusca, 1990; Meglitsch and Schram, 1991) 39. Swimming/feeding band(s) of cilia in larvae
[a/p] with compound ciliab (Nielsen, 1987) [a/p]
15. Longitudinally metameric coelomic cavities 40. Prototroch; locomotory equatorial ciliary
with mesodermal contribution to mesenteric band(s) with two or four rows of broad cilia

partitions (Brusca and Brusca, 1990; Meg- formed before gastrulation (Strathmann, 1987;
litsch and Schram, 1991) [a/p] Brusca and Brusca, 1990) [a/p]
16. Tripartite coelom (Brusca and Brusca, 1990) 41. Pelagic larvae with apical ciliary tuft and plate
[a/p] (Nielsen, 1987) [a/p]
17. Schizocoelous formation of body cavity lined 42. Nutritive metatroch with opposed bands;
with mesodermal peritoneumb (Brusca and postoral (segmentally added) paired ciliary
Brusca, 1990; Meglitsch and Schram, 1991) bands beating in opposite directions and
[a/p] serving in food capture1" (Strathmann, 1978;
18. Enterocoelous formation of body cavity lined Salvini-Plawen, 1988) [a/p]
with mesodermal peritoneumb (Brusca and 43. Cerebral rhabdomeric larval ocelli or integ-
Brusca, 1990; Meglitsch and Schram, 1991) umentary pigment cupsb (Rosen et al., 1979;
[a/p] Brusca and Brusca, 1990; Meglitsch and
19. Hemocoel; main body cavity unlined with Schram, 1991) [a/p]
lymph-filled vacuities (Brusca and Brusca, 44. Telotroch; pelagic larvae with para- or cir-
1990; Meglitsch and Schram, 1991) [a/p] cumanal ciliary tuft (Nielsen, 1987; Brusca
20. Gonocoel; coelom reduced to perigonadal re- and Brusca, 1990; Meglitsch and Schram, 1991)
gion (Brusca and Brusca, 1990; Meglitsch and [a/p]
Schram, 1991) [a/p] 45. Preoral fold covering larval hyposphere
21. Hydrostatic skeleton; coelomic compartment (Brusca and Brusca, 1990; Meglitsch and
under relatively high pressures (Brusca and Schram, 1991) [a/p]
Brusca, 1990; Meglitsch and Schram, 1991) 46. Pericalymna; unpaired, mineralized, epi-
[a/p] spheral test enveloping larvaeb (Salvini-Pla-
22. Pericardial excretory complex of coelomo- wen, 1988; Brusca and Brusca, 1990) [a/p]
ducts connected to cloaca (Salvini-Plawen, F. Bilateral symmetry
1985)[a/p] 47. Paired ventral nerve bundles'5 (Beklemishev,
23. Pseudocoelom; blastocoel persisting as un- 1969; Brusca and Brusca, 1990; Kozloff, 1990)
lined cavity(ies) between endoderm and me-
[a/p]
soderm (Meglitsch and Schram, 1991) [a/p]
C. Cellular 48. Paired ventral lateral pedal retractor muscle
24. Multicellular with population of gamete cells bundlesb (Wingstrand, 1985) [a/p]
(Hanson, 1977; Nielsen, 1985) [a/p] 49. Paired excretory organs and ducts open ex-
25. Special sense cells (Hanson, 1977; Nielsen, ternally (nephridiopores) (Brusca and Brus-
1985)[a/p] ca, 1990)[a/p]
26. Gap junctions (Nielsen, 1985) [a/p] 50. Paired gills; ectodermal filamentous or la-
27. Basal lamina (Nielsen, 1985) [a/p] mellar respiratory surfaces (Brusca and Brus-
28. Mitosis lacking in epidermal cells (Ehlers, ca, 1990) [a/p]
1986; Smith et al., 1986) [a/p] 51. Paired gonads and gonoducts (or nephridio-
pores used as gonoducts)b [a/p]
D. Ciliary 52. Paired endothelium-lined pericardial diver-
29. Collared ciliary units (Willmer, 1990) [a/p] ticulae (auricle) (Wingstrand, 1985) [a/p]
30. Ciliated somatic cellsb (Nielsen, 1987) [a/p]
31. Multiciliary epidermis with ciliated rootlets G. Serial repetition
(Smith et al., 1986; Nielsen, 1987) [a/p] 53. Serially repeated nerve collaterals; ladder-
32. Ciliated ventral surface in adult (Nielsen, like nervous system with ventrolateral nerve
1987)[a/p] cords and lateral connectives'1 (Beklemishev,
33. Monociliated cells with accessory centriole 1969; Wingstrand, 1985) [a/p]
(Nielsen, 1987) [p/multiciliated without cor- 54. Serially repeated transverse discrete muscle
responding accessory centrioles] bundles (Beklemishev, 1969; Wingstrand,
34. Cilium with one basal body; without acces- 1985) [a/p]
55. Serially repeated nerve gangliab (Beklemi- domains with one oxygen-binding dinuclear
shev, 1969; Wingstrand, 1985) [a/p] copper site per domain]
56. Serially repeated transverse discrete muscle 72. Mesodermal origin of pericardioducts (Sal-
bundles (Beklemishev, 1969; Wingstrand, vini-Plawen, 1985, 1988) [a/p]
1985) [a/p] I. Integumentary
57. Serially arranged series of excretory ducts; 73. Cellular production of collagenous proteinsb
nephridiopores (Brusca and Brusca, 1990) (Brusca and Brusca, 1990; Meglitsch and
[a/p] Schram, 1991) [a/p]
58. Serially arranged ectodermal filamentous or 74. Cuticle; continuously secreted, nonliving ex-
lamellar respiratory surfaces'* (Brusca and ternal layer(s) containing protein11 (Brusca and
Brusca, 1990) [a/p] Brusca, 1990; Meglitsch and Schram, 1991)
59. Serially arranged series of gonads (Brusca and [a/p]
Brusca, 1990) [a/p] 75. Collagenous proteins sequestered in cuticleb
60. Atria; serially arranged muscularized regions (Brown, 1975; Bereiter-Hahn et al., 1984;
of a dorsal blood vessel (Salvini-Plawen, 1985; Brusca and Brusca, 1990; Meglitsch and
Wingstrand, 1985) [a/p] Schram, 1991) [a/p]
76. Chitinous proteins invested in cuticle b

61. Schizocoelous metamerism between preoral
prostomium and nonmetameric pygidium (Brown, 1975; Bereiter-Hahn et al., 1984;
Brusca and Brusca, 1990; Meglitsch and
62. One or more transverse coelomic septa (Brus- Schram, 1991) [a/p]
ca and Brusca, 1990) [a/p] 77. Cellular secretion of chitinous proteins'5
(Brown, 1975; Bereiter-Hahn et al., 1984;
63. Serially repeated ventricles; branchioauricu-
Brusca and Brusca, 1990; Willmer, 1990; Meg-
lar sinuses with ctenedial pores (Ruppert and
litsch and Schram, 1991) [a/p]
Carle, 1983; Wingstrand, 1985; Brusca and
78. Cuticular covering of entire external body
Brusca, 1990) [a/p] surface (Brusca and Brusca, 1990; Meglitsch
64. Metamerism in associated cuticular, muscu- and Schram, 1991) [a/p]
lar, and nervous tissues (Brusca and Brusca, 79. Sclerotinization of cuticle with tannin pro-
1990; Meglitsch and Schram, 1991) [a/p] teins (Brusca and Brusca, 1990; Meglitsch and
H. Circulatory Schram, 1991) [a/p]
65. Heart(s); dorsal blood vessel with contractile 80. Protrusible and retractable chitinous setae
epithelium formed around a vascularized (Brusca and Brusca, 1990; Meglitsch and
longitudinal lumen by fusion of coelomic Schram, 1991) [a/p]
walls and lined by a basal laminab (Ruppert 81. Alpha and/or beta ecdysone (Willmer, 1990)
and Carle, 1983; Brusca and Brusca, 1990) [a/ [a/p]
P] 82. Periodic ecdysis under hormonal control
66. Closed posterior circulation (Ruppert and (Brusca and Brusca, 1990; Meglitsch and
Carle, 1983; Salvini-Plawen, 1985; Brusca and Schram, 1991) [a/p]
Brusca, 1990) [a/p] 83. Anterior ecdysome-producing gland (Brusca
67. Atrial ostia; muscularized opening(s) in dor- and Brusca, 1990; Meglitsch and Schram, 1991)
sal blood vessel (Ruppert and Carle, 1983; [a/p]
Salvini-Plawen, 1985, 1988) [a/p] 84. Aciculae; tonofibrillae penetrating epi-
dermis with muscle attachment (Brusca and
68. Atrial ultrafiltration (Salvini-Plawen, 1985,
Brusca, 1990; Meglitsch and Schram, 1991)
1988)[a/p]
[a/p]
69. Hemerythrin or myohemerythrin as a respi-
85. Mantle; thick epidermal cuticular sheet with
ratory pigment molecule (Mangum et al.,
band(s) of glands capable of secreting a hard
1985; Richardson et al., 1987; Volbeda and
calcareous skeleton (Brusca and Brusca, 1990;
Hoi, 1989; Demuynck et al., 1991; Takagi and
Meglitsch and Schram, 1991) [a/p]
Cox, 1991; Yano et al., 1991) [a/p]
86. Calcified skeletal covering secreted by epi-
70. Hemocyanin as a respiratory pigment mol- dermis" (Brown, 1975; Tidball, 1984; Salvini-
ecule hypothesized to be homologous based Plawen, 1988; Brusca and Brusca, 1990; Car-
on spectroscopic and sequence similarities ter, 1990) [a/p]
(Mangum et al., 1985, 1987; Lang, 1988; Vol- 87. Lateral tergal folds or paranotal lobes (Ghi-
beda and Hoi, 1989; Voit and Feldmaier- selin, 1988; Salvini-Plawen, 1988) [a/p]
Fuchs, 1990; Lang and Holde, 1991) [a/p] 88. Anterior cephalic tagma formed from meta-
71. Hemocyanin structure (Linzen et al., 1985; mere(s) and the primary sensory acron (Brus-
Mangum et al., 1987) [hexamer or multihexa- ca and Brusca, 1990; Meglitsch and Schram,
mer "boxcar" molecules with subunits of 1991)[a/p]
about 75,000 molecular weight, together 89. Serially arranged mineralized ectodermal
combining to up to 3-5 million, each con- plates (Brusca and Brusca, 1990; Meglitsch
taining one dinuclear copper site/cylindrical and Schram, 1991) [a/p]
molecules made up of about 10-20 "stacked J. Alimentary
petri-plate" subunits, each of about 350,000 90. U-shaped alimentary canal (Brusca and Brus-
molecular weight, containing seven or eight ca, 1990)[a/p]
91. Crystalline stylus and associated ciliated 110. Endon; median cerebral ganglion and adja-
midgut digestive organs (Brusca and Brusca, cent aboral statocyst organ (Beklemishev,
1990)[a/p] 1969)[a/p]
92. Pharyngeal diverticulae (Salvini-Plawen, 111. Three pairs of cerebral ganglia; an anterior
1988) [a/p] one receiving ocular input, a second receiv-
93. Esophageal pouches (Salvini-Plawen, 1988) ing palpar or antennal input, and a third con-
[a/p] tributing to circumenteric connectives (Bek-
94. Terminal alimentary zones of cuticle (Boud- lemishev, 1969; Brusca and Brusca, 1990) [a/
reaux, 1979)[a/p] P]
95. Secondary mouth formation (Meglitsch and 112. Paired olfactory fossae of preoral lobes (Bek-
Schram, 1991) [a/p] lemishev, 1969) [a/p]
96. Anus with proctodeum; complete unidirec- 113. Compound eyes with ommatidia (Paulus,
tional alimentary canal b (Meglitsch and 1979; Brusca and Brusca, 1990; Meglitsch and
Schram, 1991) [a/p] Schram, 1991) [a/p]
K. Excretory 114. Ommatidium consisting of a cornea with two
97. Antennal gland excretory ducts; mandibular corneagen cells, a tetrapartite eucone crys-

(first) pair metanephridia [a/p] talline cone, and a retinula of eight cells
98. Metanephridia; paired mesodermal excreto- (Brusca and Brusca, 1990) [a/p]
ry ducts with ciliated funnel draining coe- 115. Prostomial sensory antennae with basal ocel-
lomic cavity(ies) (Brusca and Brusca, 1990) li with or without lens structure (Brusca and
[a/Pi Brusca, 1990) [a/p]
99. Protonephridia; ampullary (blind) vessels 116. Unicellular sensory (tactile) setae in epi-
bearing multiciliated cells serving excretory / dermis (Brusca and Brusca, 1990) [a/p]
osmoregulatory function (Brusca and Brusca, M. Reproductive
1990; Meglitsch and Schram, 1991) [a/p] 117. Hermaphroditic sexual system (Brusca and
L. Nervous Brusca, 1990) [gonochoric/present in more
100. Acetocholine" (Willmer, 1990) [a/p] than isolated species]
101. Creatine phosphotase (Willmer, 1990) [a/p] 118. Filiform morphology of sperm (Wirth, 1984;
102. Population of specialized polar neurons with Ax, 1985; Smith et al., 1986; Franzen, 1987)
neurites and synaptic terminals (Nielsen 1985; [a/p]
Brusca and Brusca, 1990) [a/p] 119. Direct internal fertilization (Ax, 1985) [ex-
103. Orthogon; dense diffuse neural plexus with ternal/internal]
short peripheral connections and very long N. Respiratory
interganglionic connections (Beklemishev, 120. Gill with counter-current O2 exchange11
1969; Brusca and Brusca, 1990; Meglitsch and (Brusca and Brusca, 1990) [a/p]
Schram, 1991) [a/p] 121. Cuticle-lined tracheal tubes (Brusca and
104. Subcutaneous neural plexus; subepithelial
Brusca, 1990) [a/p]
location of (at least some) epidermally de-
rived neurons (Beklemishev, 1969; Brusca and 122. Semi-internal lateroventral respiratory
Brusca, 1990; Meglitsch and Schram, 1991) chamber (Brusca and Brusca, 1990) [a/p]
[a/p] O. Oral
105. Circumpharyngeal chain of ganglia (buccal, 123. Subradular organ (Wingstrand, 1985) [a/p]
pharyngeal, or subenteric) attached to lon- 124. Radula; ribbon or plates or recurved chitin-
gitudinal ventral nerve cord(s) (Beklemish- ous teeth stretched over a supportive carti-
ev, 1969; Kozloff, 1990) [a/p] laginous (or hemocoelic) basal expansion of
106. Adnate ventral nerve cords (Kozloff, 1990) the foregut epithelium (Wingstrand, 1985;
Eernisse and Kerth, 1988) [a/p]
[a/p] 125. Introvert (proboscis) at anterior end of di-
107. Lateral nerve cords; paired longitudinal cu- gestive tract with barbs and hooks (Nielsen,
taneous or subepidermal axon bundles de- 1987; Brusca and Brusca, 1990; Meglitsch and
scending from an anterior commissure or Schram, 1991) [a/p]
ganglion (Beklemishev, 1969; Brusca and 126. Lophophore; anterior ring of hollow ciliated
Brusca, 1990; Meglitsch and Schram, 1991) tentacles formed by coelomic evaginations
[a/p] (Brusca and Brusca, 1990) [a/pi
108. Dorsal nerve cord; median longitudinal cu- 127. Mandibles; appendages of the third post-
taneous or subepidermal axon bundle de- acronal head somite (Brusca and Brusca, 1990)
scending from an anterior commisure or gan- [a/p]
glion (Beklemishev, 1969; Brusca and Brusca, 128. Two pairs of maxillae; appendages of post-
1990; Meglitsch and Schram, 1991) [a/p] acronal head somites four and five (Brusca
109. Osphradia; chemosensory epithelial surfaces and Brusca, 1990) [a/p]
located on or near the gill(s); mechanorecep-
tive collar cells with eight or nine stereo mi- P. Muscular
crovilli (Haszprunar, 1985a, 1985b, 1987; 129. Oblique striated muscle fibers (Brusca and
Brusca and Brusca, 1990) [a/p] Brusca, 1990) [a/p]
130. Cross striated muscle fibersb (Brusca and 38. Characters of larval morphology: taxa with ex-
Brusca, 1990) [a/p] clusive direct development coded 0 if absent
131. Dermal circular (or "external transverse") 39. Lack of compound cilia in oweniids (Polychaeta)
muscular fibers (Salvini-Plawen, 1978, 1985; considered a derived exceptional case (Nielsen,
Brusca and Brusca, 1990) [a/p] 1987)
132. Longitudinal muscle sheet(s) or band(s) (Sal- 42. Sipunculans and nemerteans without opposed
vini-Plawen, 1978, 1985; Brusca and Brusca, band mechanism (Strathmann, 1978)
1990) [a/p] 43. Uniramia coded 1, although myriopods possess
133. Intersegmental tendon system (Brusca and direct development (Brusca and Brusca, 1990)
Brusca, 1990) [a/p] 46. Sipunculan "serosa larva"; polychaete "endo-
134. Locomotor coxae with extrinsic and intrinsic larva" (Salvini-Plawen, 1988)
muscles (Brusca and Bursca, 1990) [a/p] 47. Characters of paired lateral structures sharing
135. Myofilaments (Hanson, 1977; Nielsen, 1985) similar relative topology and ultrastructure
[a/p] 48. Solenogastres (Salvini-Plawen, 1985)
136. Smooth muscle fibers (Hanson, 1977; Niel- 51. Uniramia coded present; Diplopoda present,
sen, 1985; Brusca and Brusca, 1990) [a/p] Chilopoda variable, Pauropoda females have
single ovary, all with paired gonoducts

137. Scleratized terminal structure on coxae
(Brusca and Brusca, 1990) [a/p] 53. See Methods for our distinction between serial
138. Broad creeping sole or narrow hydrostatic repetition and metamerism
foot in ventral furrow ("pedal groove") 55. Pogonophora coded polymorphic (Meglitsch and
(Wingstrand, 1985) [a/p] Schram, 1991)
139. Segmented serially arranged locomotor ap- 58. Chordata (C. Gans, pers. comm.)
pendages with basal coxite and distal telo- 65. Characters of the blood vascular system (Rup-
pide (Brusca and Brusca, 1990) [a/p] pert and Carle, 1983); Crania (Brachiopoda) with
140. Pedal glands; ventral and large with mucous several, lkeda (Echiura) with one, neither coded
secretion (Brusca and Brusca, 1990) [a/p] as plesiomorphic (Brusca and Brusca, 1990)
141. Biramous appendages (Emerson and Schram, 73. Characters of the ectodermal integument: ar-
1990; Grosberg, 1990; Meglitsch and Schram, thropods (Manton and Anderson, 1979), mol-
1991)[a/p] luscs (Runnegar, 1983; Salvini-Plawen, 1985,
1988; Wingstrand, 1985)
Supplementary Notes by Character Number 74. Arthropod codes largely from Manton and An-
derson (1979); see notes for character 75 for Cni-
2. Pogonophoran ontogeny poorly understood daria
(Ivanov, 1963, 1988; Bakke, 1980), pogonophor- 75. Collagenous exoskeletal perisarc of gorgonians
ans represented by Perviata (Meglitsch and (Cnidaria) considered similar to the cuticle of
Schram, 1991) for all characters except no. 14; insects (Goldberg, 1976)
Gnathostomulida (Riedl, 1969); "acoel" and "po- 76. Chitin found in tube but not cuticle of Phoron-
lyclad" flatworm cleavage (Boyer, 1971,1989, re- ida (Hyman, 1958)
spectively) 77. Report of chitin in Chordata (Sannasi and Her-
3. Solenogastres cleavage and ontogeny (reviewed mann, 1970) rejected by Azariah (1973)
by Hadfield, 1979; Salvini-Plawen, 1985), Cau- 86. The calcification mechanisms of anthozoan and
dofoveata ontogeny undescribed; "Turbellaria" hydrocoral (Cnidaria) skeletons, which occurs
interpreted as polyclads (Salvini-Plawen, 1988) within specialized region of epidermis, not com-
14. Characters of coelom and mesoderm ontogeny: pletely understood (Fautin and Mariscal, 1991)
mollusc references reviewed in Salvini-Plawen 96. Vestimentifera have transitory alimentary canal
(1985); monoplacophoran "dorsal coeloms" not in early ontogeny; character coded as polymor-
coded as present (Wingstrand, 1985); uniramian phic for Pogonophora (Meglitsch and Schram,
coelomic metamerism from Brusca and Brusca 1991)
(1990); pogonophorans represented by Vesti- 100. Characters of the nervous system (Beklemishev,
mentifera (Brusca and Brusca, 1990) 1969)
17. Nemertean rhynchocoel (Turbeville and Rup- 120. Characters of the respiratory system: arthropod
pert, 1985; Turbeville, 1991) terminals coded from sources providing ingroup
18. Tardigrades (in Nelson, 1982); priapulids (Lang, polarizations as described in Methods.
1848, in Meglitsch, 1972) 130. Characters of the somatic musculature and lo-
30. Ciliary and flagellar character 30-39 (Tyler, 1979; comotor apparatus: arthropods (Manton and An-
Nielsen, 1985, 1987) derson, 1979; Emerson and Schram, 1990), mol-
37. The discocilia's paddle shape itself may be a luscs (Runnegar, 1983; Salvini-Plawen, 1985,
preservation artifact (Nielsen, 1987) 1988; Wingstrand, 1985)
330 SYSTEMATIC BIOLOGY VOL. 41
APPENDIX 2. Apomorphy hypotheses for internal

nodes of metazoan tree in Figure 4.°
Node Character no.

2 38, 45, 49, 62, - 7 6 , 86, 90, 95, 98, 120, 126
3 51, - 1 0 1 , 103, 116
4 23, 28, 55, 82, 119
5 6,47,48,54,56,64,-103,104,130
6 15, 19, 20, - 7 4 , 134, 137
7 - 2 3 , - 2 8 , 60, 65, 67, 97, 98, 121, 122, -125
8 81,84,86,87,88,113,-132,139
9 11, 32, - 7 4 , - 7 6 , 110, 112
10 - 9 6 , 117, 118, 119
11 10, - 1 3 , 28, -77, 92, -116, -132, 138
12 7, 17, 39, 41, 43, 131
13 40, 44, 49

14 21, - 3 2 , 74, 75, 94, 98
15 12, 66, 80
16 9, 15, 62, 65, -103
17 47, 54, 56, 57, 58, 59, 61, 63, 64, 104
18 20, 22, 47, 52, 67, 68, 72, 85, 93, 109, -116,
120, 124
19 48, 54, 56, 91, 138, 140
20 19,58,60,65,86,92,123,-131
a
This listing of apomorphies by labeled interior node (2-
20) of the strict consensus morphology-based cladogram (Fig.
4) gives characters hypothesized to have changed prior to
each node according to their numbering in Appendix 1. Those
preceded by a minus sign are hypothesized to have changed
from 1 to 0, whereas all other changes are from 0 to 1. Only
those changes consistent with all different optimization
methods available in PAUP (ACCTRAN, DELTRAN, MIN F)
are included in this listing. The strict concensus cladogram
differs from the six individual minimum-length hypotheses
as explained in the text.

Annelida and Arthropoda Are Not Sister Taxa: A Phylogenetic Analysis of Spiralian Metazoan Morphology

Uploaded by

Copyright:

Available Formats

Annelida and Arthropoda Are Not Sister Taxa: A Phylogenetic Analysis of Spiralian Metazoan Morphology

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Annelida and Arthropoda Are Not Sister Taxa: A Phylogenetic Analysis of Spiralian Metazoan Morphology

Uploaded by

Copyright:

Available Formats

Sysl. Biol.

ANNELIDA AND ARTHROPODA ARE NOT SISTER TAXA:

Downloaded from http://sysbio.oxfordjournals.org/ at Russian Archive on November 13, 2013

An enormous literature of descriptive gram (Hennig, 1966; Patterson, 1982), and

Downloaded from http://sysbio.oxfordjournals.org/ at Russian Archive on November 13, 2013

cently advocated by investigators who have

Downloaded from http://sysbio.oxfordjournals.org/ at Russian Archive on November 13, 2013

Downloaded from http://sysbio.oxfordjournals.org/ at Russian Archive on November 13, 2013

Downloaded from http://sysbio.oxfordjournals.org/ at Russian Archive on November 13, 2013

Downloaded from http://sysbio.oxfordjournals.org/ at Russian Archive on November 13, 2013

Downloaded from http://sysbio.oxfordjournals.org/ at Russian Archive on November 13, 2013

TABLE 2. . Comparison of previous hypotheses of metazoan relationships.

Tree as shown Tree resolved

Downloaded from http://sysbio.oxfordjournals.org/ at Russian Archive on November 13, 2013

Downloaded from http://sysbio.oxfordjournals.org/ at Russian Archive on November 13, 2013

Downloaded from http://sysbio.oxfordjournals.org/ at Russian Archive on November 13, 2013

Downloaded from http://sysbio.oxfordjournals.org/ at Russian Archive on November 13, 2013

— Cnidaria larger clade (node 4), which (listed in order

Downloaded from http://sysbio.oxfordjournals.org/ at Russian Archive on November 13, 2013

Downloaded from http://sysbio.oxfordjournals.org/ at Russian Archive on November 13, 2013

(a) Hyman, 1940-1967 (b)Hadzl, 1953,1963

Downloaded from http://sysbio.oxfordjournals.org/ at Russian Archive on November 13, 2013

(g) Pearse at al., 1987 (h) Brusca and Brusca, 1990

Downloaded from http://sysbio.oxfordjournals.org/ at Russian Archive on November 13, 2013

(')Kozloff,1990 Cnidaria (l)Wlllmer, 1990

Downloaded from http://sysbio.oxfordjournals.org/ at Russian Archive on November 13, 2013

Downloaded from http://sysbio.oxfordjournals.org/ at Russian Archive on November 13, 2013

Downloaded from http://sysbio.oxfordjournals.org/ at Russian Archive on November 13, 2013

Downloaded from http://sysbio.oxfordjournals.org/ at Russian Archive on November 13, 2013

Downloaded from http://sysbio.oxfordjournals.org/ at Russian Archive on November 13, 2013

Downloaded from http://sysbio.oxfordjournals.org/ at Russian Archive on November 13, 2013

Downloaded from http://sysbio.oxfordjournals.org/ at Russian Archive on November 13, 2013

Downloaded from http://sysbio.oxfordjournals.org/ at Russian Archive on November 13, 2013

Downloaded from http://sysbio.oxfordjournals.org/ at Russian Archive on November 13, 2013

Downloaded from http://sysbio.oxfordjournals.org/ at Russian Archive on November 13, 2013

APPENDIX 2. Apomorphy hypotheses for internal

Node Character no.

Downloaded from http://sysbio.oxfordjournals.org/ at Russian Archive on November 13, 2013

You might also like