III. Molecular Phylogenetic Trees as Alternative Hypotheses

As noted before, if a character is not consistent with one phylogenetic tree, then a molecular phylogenetic tree provides an alternative hypothesis that may better explain the evolution of the character.

Protozoa are not monophyletic.

• HISTORY: Hyman (1940, vol. 1) considered Protozoa to be a phylum of "acellular" animals. With the widespread acceptance of the Five-Kingdom System, however, Protozoa were moved from the kingdom Animalia to the kingdom Protista, together with algae. This kingdom would be paraphyletic under any hypothesis that animals evolved from a protozoan. It soon became apparent that the protozoa were so diverse that they comprised several, if not dozens, of phyla. In 1980 a committee of the Society of Protozoologists (Levine et al. 1980) proposed a tentative classification that admittedly did not reflect phylogeny. This scheme divided protozoa into seven phyla, including Apicomplexa, Ciliophora, Sarcomastigophora (sarcodines and flagellates), and Myxozoa.
• Molecular phylogenetics confirms that "protozoa" are polyphyletic, belonging to numerous separate branches within the domain Eucarya (Fig. 18, next page). The flagellates in particular are distributed widely among several clades as follows:
• Giardia and Trichomonas belong to separate clades near the base of the domain Eucarya (Baroin et al. 1988; Hasegawa et al. 1993;Sogin et al. 1986; Sogin et al. 1989; Yamamoto et al. 1997).
• Dinoflagellates appear to be closely related to ciliates and apicomplexans (Wolters 1991). The clade Alveolata that comprises these groups is supported by a variety of molecular data. (See Baldauf et al. 2000, Fig. 2 for a summary of support for this and other eukaryotic clades. See Patterson 1999 for a useful guide to eukaryotic groups.)
• Volvox is a green alga more closely related to plants than to animals (Baldauf et al. 2000; Rausch et al. 1989). (The custom of including Volvox in zoology courses is simply a relic of Haeckel's blastea theory.)
• Choanoflagellates are closer to metazoans than to other protozoans, as will be discussed shortly (Wainright et al. 1993).
• Plantae (including some algae), Fungi (excluding slime molds and some others), and Animalia (including choanoflagellates) form a monophyletic clade at the tip of the Eucarya. The name Metakaryota has been proposed for this clade.

Figure 18. Molecular phylogeny of eukaryotes showing the polyphyly of protozoans. See Baldauf et al. (2000) for a somewhat different cladogram based on protein sequences.

Metazoans are monophyletic, and choanoflagellates may be their sister group.

• HISTORY: Hyman, writing when there were only two or three kingdoms, appears never to have doubted that Animalia, including protozoa, was monophyletic. She noted that "many zoologists believe the presence of choanocytes in sponges can only be interpreted to indicate the direct descent of sponges from Choanoflagellata," and that the colonial choanoflagellate Protospongia is a link between choanoflagellates and sponges (Hyman 1940, vol. 1, pp. 358, 107).
• RECENT MORPHOLOGICAL STUDIES: Willmer (1990), after reviewing the numerous theories on the origin of metazoans (pp. 165-187), concluded (p. 196) that "the best we can do is to remain agnostic, but with suspicions of polyphyly." Nielsen (1995, p. 27), however, concurred with the traditional view that "the Animalia is a monophyletic group, and the specific characters shared with the choanoflagellates make it natural to consider the two groups as sister groups."
• The conclusion by Field et al. (1988) that Cnidaria had a separate origin from the other meta-zoa was shown by (Lake 1990) to be due to long-branch attraction. Subsequent analyses of 18S rDNA indicate that metazoans are monophyletic.
• Analysis of 18S rDNA indicates that choanoflagellates are sister to the monophyletic Metazoa (Wainright et al. 1993). Choanoflagellata are now sometimes included in the clade Animalia (Fig. 18, previous page).

Myxozoans appear to be metazoans rather than protozoans.

• HISTORY: Myxozoan parasites, such as the species responsible for "whirling disease" in salmonids, have long been considered to be protozoans. Although their infective stage is multicellular, they are extremely small and apparently without cellular differentiation, gametes, or blastula. The occurrence of nematocysts in the infective stage, however, has led to speculation for more than a century that they are related to Cnidaria. (See Siddall and Whiting 1999 for references.)
• Analyses of 18S rDNA indicate that Myxozoa is most likely derived from a metazoan near the base of the Bilateria (Smothers et al. 1994). Siddall et al. (1995) concluded from both 18S rDNA and morphology that Myxozoa are cnidarians related to the parasitic narcomedusan Polypodium. The myxozoan sequences all have high rates of evolution, however, so it has been argued that the result could have been due to long-branch attraction. Siddall and Whiting (1999) have vigorously defended their conclusion against this charge, showing that the position of Myxozoa remains the same even when Polypodium is not included.
• Cavalier-Smith et al. (1996) tentatively accepted the 18S rDNA evidence that Myxozoa were derived from Cnidaria. Unable to decide whether they evolved from a bilaterian intermediate or directly from a cnidarian, they made Myxozoa a separate subkingdom, along with Radiata, Mesozoa, and Bilateria.

Mesozoans may be flatworms.

• HISTORY: Hyman (1959, vol 5, p. 714) lamented that most zoologists persisted in considering Mesozoa to be degenerate flatworms 19 years after she had argued that they were a distinct phylum. Her view appears finally to have triumphed, however. Because of their simple construction, mesozoans are usually considered to be a distinct phylum of a grade somewhere between that of Porifera and Platyhelminthes. Some zoologists, however, point to their complex life cycles as evidence that they derive from flatworms. Many consider mesozoans to be not merely one phylum, but two: Orthonectida and Rhombozoa (=Dicyemida).
• RECENT MORPHOLOGICAL STUDIES: Willmer (1990, p. 351) acknowledged that mesozoans "might yet prove to be 'degenerate flatworms,'" but she thought it more plausible that they evolved in parallel with other bilaterians from a flatworm-like ancestor. Nielsen (1995, p. 436) merely noted the orthonectids and dicyemids as enigmatic groups and did not include them in his cladogram.
• Two studies based on 18S rDNA suggested that orthonectids are not closely related to dicyemids, and that Mesozoa is therefore polyphyletic (Hanelt et al. 1996; Pawlowski et al. 1996). Different analyses, however, suggest that Mesozoa is a monophyletic clade (Siddall and Whiting 1999; Winnepenninckx, Van de Peer, and Backeljau 1998).
• Studies based on 18S rDNA (Katayama et al. 1995; Van de Peer and De Wachter 1997) and Hox-gene sequences (Kobayashi, Furuya, and Holland 1999) support a close relationship between dicyemids and flatworms.

Anthozoa appear to be basal within Cnidaria.

• HISTORY: Hyman (1959, vol. 5, pp. 750-753) scarcely veiled her contempt for the notion that the "advanced" Anthozoa were basal to Scyphozoa and Hydrozoa.
• RECENT MORPHOLOGICAL STUDIES: It is now obvious that one cannot infer the ancestral position of a group from the perceived "grade" of its extant members. Nielsen (1995, p. 58) considered the Anthozoa to be the basal clade of Cnidaria, followed by Scyphozoa then Cubozoa and Hydrozoa.
• Bridge et al. (1992) found that in Anthozoa the mitochondrial DNA is circular, as in Ctenophora and most other organisms, but that in the other classes of Cnidaria mtDNA is linear. From this they concluded that Anthozoa is the most basal class of Cnidaria. Bridge et al. (1995) subsequently found that 18S rDNA sequences and morphological characters also support the placement of Anthozoa at the base of the Cnidaria, with Hydrozoa, Scyphozoa, and Cubozoa in an unresolved trichotomy.
• This result suggests that the polyp-only life cycle is plesiomorphic in Cnidaria, and the medusa is apomorphic.

Protostomates appear to be divided into two major clades: Ecdysozoa and Lophotrochozoa. Annelida and Arthropoda belong to Lophotrochozoa and Ecdysozoa, respectively, and are therefore not closely related.

• HISTORY: Hyman's "hypothetical diagram" (see Fig. 14) placed all the protostomes except Nemertea, Aschelminthes, and Platyhelminthes in a single lineage with Annelida closer to Arthropoda than to Mollusca. Her list of schizocoelous eucoelomates did not separate Arthropoda from Mollusca and Annelida. Hyman (vol. 1, 1940, p. 38) and many authors since have assumed a close relationship between Annelida and Arthropoda because of several shared features, including segmentation and a paired ventral nerve cord.
• RECENT MORPHOLOGICAL STUDIES: Willmer (1990) divided the protostomes into numerous lines rather than major groups, as shown in the summary table. She (p. 298) concluded that the so-called uniramian arthropods were derived from "a proto-annelid group," but other arthropods were not. A cladistic analysis led Eernisse, Albert, and Anderson (1992) to conclude that Arthropoda are not in the same major clade (Eutrochozoa) with Annelida. Nielsen (1995) grouped Arthropoda, Annelida, and Mollusca together in Teloblastica within his Spiralia, as shown in the indented list above. He placed the clade Panarthropoda (Arthopoda, Tardigrada, and Onychophora) as the sister group to Annelida.
• Among the earliest and most robust conclusions from 18S rDNA studies is that Arthropoda are in a clade separate from Annelida and Mollusca (Field et al. 1988; Lake 1990). As sequences from other protostomates were studied, they generally grouped with one or the other of the two clades (Fig. 17).
• Comparison of Hox genes supports the conclusion that protostomes divide into these two clades (De Rosa et al. 1999).

Ecdysozoa, the major protostomate clade that includes Arthropoda, also includes Nematoda and other groups with a cuticle that molts all at once.

• HISTORY: Hyman (1959, vol. 5, p. 745) did not consider Nematoda to be a separate phylum, but a class within the phylum Aschelminthes, along with other pseudocoelomates except Entoprocta. She regarded the aschelminths as being in a separate line of Protostomia from that of Arthropoda. Many authors, following Hyman's list of phyla, consider the Nematoda and other "aschelminths" to be on a branch beneath the coelomates.
• RECENT MORPHOLOGICAL STUDIES: As noted previously, there has been considerable doubt over the usefulness of the pseudocoel as a character. Nematodes have therefore wandered over the phylogenetic tree. Willmer (1990, pp. 245-246) cautiously suggested that Nematoda, Nematomorpha, and Gastrotricha are distinct phyla related to each other in a line that descended from acoelomates independently from the lines of other "aschelminths" and arthropods, as shown in the summary list above. Nielsen (1995, p. 234) included Nematoda, Nematomorpha, Priapulida, Kinorhyncha, Loricifera, Rotifera, Acanthocephala, Gastrotricha, and Chaetognatha as phyla in the clade Aschelminthes. He included the phylum Arthropoda in the Spiralia, which he made the sister of Aschelminthes in the Protostomia. (See indented list.) Another cladistic analysis by Wallace et al. (1996), however, found that Nematoda, Nematomorpha, Kinorhyncha, Priapulida, and Loricifera were in a clade separate from that of Rotifera and Acanthocephala, and that the "pseudocoelomates" were derived from one or more coelomate ancestors.
• Phylogenies based on 18S rDNA sequences (for example, Winnepenninckx et al. (1995) generally show "pseudocoelomates" divided between the two protostomate clades, with Rotifera and Acanthocephala in the clade with Platyhelminthes, Annelida, and Mollusca, while Nematomorpha and Priapulida are in the clade with Arthropoda. Nematoda often appears at the base of the other Bilateria.
• However, 18S rDNA has evolved too rapidly in most nematodes to provide a signal free of long-branch attraction. By using only the nematode sequence with the slowest rate of base substitution (from Trichinella), Aguinaldo et al. (1997) grouped Nematoda with Arthropoda, Nematomorpha, Kinorhyncha, and Priapulida. Because all these groups share the feature of molting the cuticle all at once, they named the clade Ecdysozoa (Fig. 17).
• A different study by Aleshin et al. (1998) also found the slowly evolving 18S rDNA sequence from Enoplus grouping with Arthropoda. However, there was only weak support for a clade that also includes Nematomorpha, Kinorhyncha, and Priapulida. Two subsequent analyses based on both 18S rDNA sequences and morphology support both the inclusion of Nematoda in the same clade with Arthropoda and the validity of Ecdysozoa as originally defined (Giribet et al. 2000; Zrzavý et al. 1998).
• Independent support for Ecdysozoa comes from the fact that drosophila, an onychophoran, a priapulidan, and Caenorhabditis elegans, but not flatworms, molluscs, or annelids, share the Hox genes Ubx and Abd-B (De Rosa et al. 1999).

Arthropoda is monophyletic; Uniramia is not valid.

• HISTORY: The monophyly of Arthropoda was virtually unquestioned until Sydney Manton proposed in the 1960s that arthropods comprise three different phyla: Chelicerata, Crustacea, and Uniramia (Onychophora + Myriapoda + Insecta). Her conclusion was based on the differences among these groups rather than a cladistic analysis of shared, derived homologies. Her anatomical studies led her to conclude that the mandibles of crustaceans develop from the bases of appendages, while those of insects and myriapods develop from entire appendages. She also concluded that the appendages of crustaceans are primitively biramous, while those of insects and myriapods are uniramous. Manton's proposal requires that the overall arthropodan body plan, including a chitinous exoskeleton, segmentation, and ventral nerve cord, would have evolved independently three times. Consequently, few zoologists accepted the proposal of three arthropodan phyla, but many did accept the clade Uniramia (minus Onychophora) as a taxon within Arthropoda.
• RECENT MORPHOLOGICAL STUDIES: Willmer (1990, chap. 11) accepted Manton's proposal as follows: "uniramians may be derived from a proto-annelid group, whilst crustaceans probably diverged from the stem spiralians earlier, from a flatworm-like stage; chelicerate origins are still enigmatic." The paleontologist Jarmila Kukalová-Peck (1992) challenged the concept of Uniramia by pointing out that numerous fossil and living crustaceans and insects have polyramous appendages. She also noted that Manton's evidence for "whole-leg mandibles" was from studies of myriapods and onychophorans but not insects. Nielsen (1995, pp. 171, 173) listed a number of synapomorphies that "clearly demonstrate that the Arthropoda are a monophyletic group."
• The expression of homeotic genes during development shows that insect mandibles develop from only a limb base, as in crustacea, contradicting the whole-leg-mandible hypothesis (Popadic et al. 1996). Several different molecular-phylogenetic studies (summarized in Regier and Shultz 1997) also suggest that crustaceans are closer to insects than myriapods are, making Uniramia paraphyletic.
• A study using 12S rDNA sequences supported the monophyly of Arthropoda (Ballard et al. 1992). Another study combining evidence from 18S rDNA and ubiquitin sequences, as well as morphology, also concluded that Arthropoda are monophyletic (Wheeler, Cartwright, and Hayashi 1993). Arthropodan monophyly is also supported by evidence from the order of genes in mitochondria (Boore et al. 1995).
• Although it is clear that Arthropoda are monophyletic, relationships within Arthropoda remain unresolved. Among the disputed conclusions from these studies are the findings that Onychophora belong in Arthropoda, Crustacea are polyphyletic, and Insecta branches from within Crustacea.

Pentastomids are crustaceans.

• HISTORY: The unique body plans of adult tongue worms led to the erection of the phylum Pentastomida. Since the 1970s, however, similarities of sperm morphology, embryology, and cuticle have suggested that pentastomids are crustaceans.
• RECENT MORPHOLOGICAL STUDIES: Willmer (1990, pp. 298-299) continued to accord Pentastomida the status of a separate phylum derived from "proto-platyhelminthes" in parallel with Crustacea. Nielsen (1995, p. 164), however, considered pentastomids to be crustaceans.
• Abele, Kim, and Felgenhauer (1989) found molecular evidence that Pentastomida are crustaceans. Like rhizocephalans, they appear to be barnacles that are highly adapted for parasitism.

Lophotrochozoa, the major protostomate clade that includes Mollusca and Annelida, also includes other animals with trochophore larvae, the lophophorates, and all descendants from their most recent common ancestor.

• HISTORY: Hyman (1951, vol. 2, p. 16) cautiously accepted the view that "the trochophore is indeed a reminiscence of the common ancestor of the eucoelomate Protostomia and perhaps also of the pseudocoelomate groups." Many protostomate larvae with little or no resemblance to a trochophore have been said to be modified trochophores, but the phyla in which a trochophore larva is evident are Mollusca, Sipunculida, Annelida, Pogonophora, Echiurida, and perhaps Cycliophora, Entoprocta, and Rotifera. Hyman (1959, vol. 5, p. 600) stated that "the common possession of a lophophore of similar anatomical and histological construction and similar positional relation to the body certainly proves an affinity between Phoronida, Ectoprocta, and Brachiopoda, but it is impossible to define this affinity in specific terms." She (pp. 603-605) placed the lophophorates among the Protostomia, based on their supposed trochophore larvae. Because of the enterocoelous origin of the coelom in brachiopods and other similarities to deuterostomes, however, Hyman concluded that the lophophorates were "a connecting link between the Protostomia and the Deuterostomia, but the details of this connection cannot be stated." It is now known that lophophorates have either direct devel-opment or larvae that are not trochophores. It is also known that except in Phoronida the mouth does not originate from the blastopore, and most lophophorates have a radial cleavage pattern. As a result of these findings, many authors since the 1970s have considered the lophophorates to be deuterostomates.
• RECENT MORPHOLOGICAL STUDIES: Although Willmer (1990, p. 355) noted nearly as many characters linking lophophorates to protostomes as to deuterostomes, she considered it "sensible to keep the lophophorates as a quite separate super-phylum of tripartite coelomates, linked to deuterostomes just above the acoelomate flatworms... from which they can most readily be derived." (See summary list.) Willmer (p. 121) noted the flexibility with which the term "trochophore" has been used and concluded that "the point may have been reached where the trochophore seems to be almost totally devalued, and is a useless catch-all term." Nielsen (1995) found several synapomorphies uniting the Ectoprocta with Entoprocta as a clade in his Spiralia (p. 206), and several synapomorphies placing the Phoronida and Brachiopoda in Deuterostomia (p. 333). Thus he did not regard the lophophorates as a natural (monophyletic) group. (See indented list.) He (p. 86) regarded the trochophore larva as one of the apomorphies defining the Protostomia (Spiralia plus Aschelminthes), even though its occurrence is scattered.
• One of the earliest results of 18S rDNA analyses is that Mollusca, Annelida, Sipunculida, Pogonophora, and Brachiopoda belong to a clade within the protostomates that is distinct from that of the Arthropoda (Lake 1990.) Later studies showed that the clade includes other phyla with trochophore larvae as well as the other two lophophorate phyla, Phoronida and Ectoprocta. 18S rDNA analyses also suggested that the "lophophorates" were not monophyletic within this clade, since only Brachiopoda and Phoronida appeared to be closely related to each other.
• Halanych et al. (1995) confirmed that the polyphyletic lophophorates are protostomates related to annelids, molluscs, and others with trochophore larvae, and they proposed naming the clade Lophotrochozoa. They formally defined Lophotrochozoa as "the last common ancestor of the three traditional lophophorate taxa, the mollusks, and the annelids, and all of the descendants of that common ancestor."
• As will be described in the following sections, subsequent molecular studies added Platyhelminthes and other groups with neither a lophophore nor a trochophore larva to the clade.
• The validity of the Lophotrochozoa is independently supported by the finding that annelids, nemerteans, flatworms, gastropods, and brachiopods all share unique Hox genes (De Rosa et al. 1999).
• Relationships of clades within Lophotrochozoa remain largely unresolved, perhaps because they diverged rapidly at about the same time (Halanych 1998).

Lophotrochozoa includes spiralians.

• HISTORY: Hyman did not refer to a group "Spiralia," but she considered the protostomes to be a cohesive group. Since Annelida, Mollusca, and some other coelomate protostomes have spiral cleavage, many authors assume that all protostomates have spiral cleavage as the plesiomorphic condition, even arthropods and other groups where the actual cleavage pattern is usually not spiral. Other authors, however, restrict the term "Spiralia" to phyla in which spiral cleavage is actually and unequivocally observed, though not necessarily in all species. These phyla are Gnathostomulida, Platyhelminthes, Mesozoa, Entoprocta, Mollusca, Sipunculida, Pogonophora, Nemertea, Annelida, and Echiurida. Spiral cleavage is not seen in Nematoda or in Arthropoda except for a few crustaceans.
• RECENT MORPHOLOGICAL STUDIES: Willmer (1990, p. 222) concluded that the term "Spiralia" should "only be retained with reservations, accepting that we do not know how far it tells us of shared ancestry." Nielsen (1995, p. 96) divided the Protostomia into two sister groups: Aschelminthes and Spiralia. (See indented list.) His Spiralia comprised Sipunculida, Mollusca, Annelida (including Gnathostomulida, Pogonophora, and Echiurida), Onychophora, Arthropoda, Tardigrada, Entoprocta, Platyhelminthes, and Nemertea.
• 18S rDNA analyses to be discussed later indicate that the phyla that do in fact have a spiral cleavage pattern all belong to Lophotrochozoa, which also includes the lophophorates in which the cleavage pattern is usually radial (Fig. 17). Other phyla that occur within Lophotrochozoa have a cleavage pattern that may be primitively spiral but is distorted or unobservable because of yolk.

Flatworms appear to be lophotrochozoans rather than basal to other Bilateria.

• HISTORY: Hyman (1940, vol. 1, p. 36) regarded acoelomates (Platyhelminthes and Nemertea) as a distinct branch within the Protostomia. She therefore dismissed the "alleged degradation of flatworms from annelids" as an example of "theoretical vaporizings." Undaunted, some authors have continued to seriously consider the possibility that flatworms originated from an annelid or some other spiralian coelomate. (See papers by Ax, Ehlers, and especially by Smith and Tyler in Conway Morris et al. 1985.) The morphological evidence for this conclusion includes the fact that, like annelids, flatworms have a classical spiral cleavage, and their serially repeated protonephridia and gonads are reminiscent of segmentation.
• RECENT MORPHOLOGICAL STUDIES: Willmer (1990, p. 361) concluded that the bilaterally symmetric animals--both pseudocoelmate and coelomate--had diverged along many lines from Platyhelminthes or flatworm-like animals. Nielsen (1995) placed Platyhelminthes within his Spiralia as shown in the indented list.
• Some analyses of 18S rDNA support the traditional position of Platyhelminthes as basal to the other Bilateria. (See, for example, Van de Peer and De Wachter 1997.) When sequences with high rates of base substitution are eliminated to avoid long-branch attraction, however, the majority of flatworms fall within Lophotrochozoa (Aguinaldo et al. 1997; Carranza, Baguña, and Riutort 1997; Ruiz-Trillo et al. 1999). Giribet et al. (2000) found that Platyhelminthes, as well as other acoelomates, may form a clade (Platyzoa) in Lophotrochozoa.
• The inference from 18S rDNA sequence analyses that flatworms are within Lophotrochozoa is supported by additional evidence from the type of intermediate-filament proteins (Erber et al. 1998) and the similarity of Hox genes in flatworms to those of annelids (Balavoine 1998; De Rosa et al. 1999).
• Myzostomids-incompletely segmented animals with trochophore larvae-may represent a link between flatworms and annelids. Myzostomids have generally been considered to be an-nelids, but evidence from 18S rDNA and EF-1∀ indicate that they are apparently flatworms (Eeckhaut et al. 2000).

Acoela may or may not be basal to other Bilateria.

• HISTORY: Identification of a basal bilaterian group has long been a goal of phylogenetics, since it might provide insight into the transition from diploblasts to triploblasts and would provide a more suitable outgroup for cladistic analyses than the usual diploblasts. The Acoela, with their simple, solid bodies and densely ciliated epidermis, fulfilled the expectations of many systematists that the ancestral bilaterian would be planula-like. In her early writings Hyman accepted the view that the Acoela might be basal Bilateria, but in 1967 (vol. 6, p. v) she noted that they did not appear to be as primitive as she and others had formerly thought.
• RECENT MORPHOLOGICAL STUDIES: Willmer (1990, p. 462) continued to accept the view that Platyhelminthes were basal bilaterians derived from a planula-like ancestor, with acoel-like flatworms being one of several groups from which many lines of bilateria diverged. Nielsen (1995, pp. 221-222), however, considered the idea of a planula-like ancestor of the Bilateria improbable, since, unlike acoels, cnidarian larvae have a permanent gut and lack a syncytial endoderm. He placed Platyhelminthes, including Acoela, within his Spiralia.
• Analyses of 18S rDNA sequences support the basal position of Acoela with respect to the majority of Platyhelminthes (Katayama, Nishioka, and Yamamoto 1996, summarized in Figure 12.) A study by Carranza, Baguña, and Riutort (1997), however, cast doubt on the monophyly of Platyhelminthes, as well as their position as basal Bilateria. The analysis of 18S rDNA data by Zrzavý et al (1998) divided flatworms into several phyla, with most well within the Bilateria. They placed the Acoela at the base of the Bilateria.
• One criticism of these 18S rDNA studies is that all the sequences from Acoela had rates of base substitution several times higher than those for most Metazoa, creating the potential for long-branch attraction. Ruiz-Trillo et al. (1999) undertook a new analysis using only a se-quence from an acoel species with a slower rate of substitution. Their much-heralded study found that this acoel nevertheless appeared at the base of the Bilateria, while the bulk of flatworms occurred in Lophotrochozoa. Separating the Acoela from Platyhelminthes can be justified on the basis of the following morphological differences: Acoela have a unique duet-spiral cleavage that lacks the second pair of cells that form in the typical quartet-spiral cleavage, Acoela have a highly regulative development rather than the determinative development of spiralians, and Acoela have only endomesoderm and no ectomesoderm.
• Adoutte et al. (2000) doubted the conclusion of Ruiz-Trillo et al., however, partly because the acoel sequence was saturated with mutations and therefore still liable to long-branch attraction. In addition, they noted research suggesting that Hox-gene sequences in acoels are similar to those of lophotrochozoans. Moreover, in the 18S-rDNA study by Giribet et al. (2000), in which long-branch attraction was avoided by excluding diploblasts, the Acoela did not occur at the base of the bilateria, but close to other flatworms within a clade Platyzoa that was sister to Lophotrochozoa. Analysis of EF-1∀ gene sequences suggest that acoels branch within Platyhelminthes (Berney, Pawlowski, and Zaninetti 2000), but the adequacy of EF-1∀ for such analyses has been doubted (Littlewood et al. 2001).
• In short, molecular phylogenetics has so far been unable to resolve the position of Acoela.

Nemertea may be closer to "coelomates" than to flatworms.

• HISTORY: Although the nemertean body plan is essentially acoelomate, the rhynchocoel is technically a coelom. Consequently, there has been a long debate about whether nemerteans belong with acoelomates or coelomates. Hyman (1951, vol. 2, pp. 473 and 528) acknowledged that the rhynchocoel is a true coelom, but nevertheless she accepted the prevailing view that nemerteans were acoelomates that had evolved from a flatworm.
• RECENT MORPHOLOGICAL STUDIES: Willmer (1990, pp. 204-207) referred to nemerteans as "acoelomates that nevertheless possess a coelom" and concluded that they were not closely allied to the line of flatworms that led to coelomate spiralians, but were an "early and specialized independent branch derived from some other group of flatworms." Nielsen (1995, p. 211) cautiously accepted a sister-group relationship of Nemertea and Platyhelminthes in the clade Parenchymia within Spiralia. (See indented list.) He felt that the rhynchocoel was not homologous with the coeloms of other spiralians (p. 231).
• The molecular-phylogenetic evidence suggesting that acoelomates are derived from coelo-mates has rendered this issue largely moot. Still, it is interesting that even 18S-rDNA studies that failed to place Platyhelminthes within Lophotrochozoa place Nemertea within Lophotrochozoa, as shown in Figure 17 (Winnepenninckx, Backeljau, and De Wachter 1995). A study using base sequences for the EF-1∀ gene also places Nemertea in Lophotrochozoa-in fact, within Mollusca (McHugh 1997).

Gastrotrichs are not closely related to nematodes.

• HISTORY: Hyman included the gastrotrichs with nematodes in the phylum Aschelminthes on the basis of "slight spaces" between the body wall and viscera, which she characterized as "presumably of the nature of a pseudocoel as they have no definite lining but their embryonic origin is as yet unknown (1951, vol. 3, p. 158)." Most authors continue to ally gastrotrichs with nematodes and other "pseudocoelomates." Hummon (1982 and personal communication) has found, however, that this pseudocoel is an artifact of fixation--a pseudo-pseudocoel. In life, gastrotrichs are as acoelomate as flatworms, and there are more morpho-logical characters linking them with flatworms than with nematodes.
• RECENT MORPHOLOGICAL STUDIES: While acknowledging doubts about the existence of pseudocoels in gastrotrichs, Willmer (1990, p. 245) concluded that nematodes and nematomorphs "probably derived from gastrotrich-like ancestors." (See summary list.) Without using a body cavity as a character, Nielsen (1995, chap. 33) placed Gastrotricha among the aschelminths within his clade Cycloneuralia as the sister group of the clade Introverta (Nematoda plus others). (See indented list.)
• Comparisons of 18S rDNA sequences by Wirz et al. (1999) indicate that Gastrotricha are not closely related to either Nematoda or Rotifera. The exact position of Gastrotricha remains unresolved, but some analyses of 18S rDNA sequences place them within Lophotrochozoa near Platyhelminthes. See, for example, Garey et al. (1996).

Acanthocephalans are closely related to rotifers.

• HISTORY: Hyman (1951, vol. 3, pp. 47-50) elevated Acanthocephala to a separate phylum simply because she could not decide between the conflicting arguments for putting them with Platyhelminthes or with Aschelminthes. Most of the morphological evidence placed them in Aschelminthes with rotifers and other pseudocoelomates. She noted, however, that the pseudocoel in acanthocephalans does not form in the same manner as in other pseudocoelomates, and serological studies suggested that among the intestinal parasites acanthocephalans were closer to cestodes than to nematodes. The argument between flatworm and aschelminth affinities continued for another decade (Hyman, 1959, vol. 5, p. 739), when her most definitive statement on the subject was the following: "Astonishingly, [O. von] Haffner [1950] arrives at the conclusion that the Acanthocephala are closer to the Rotifera than to any other aschelminth group. The author finds the arguments for this strange conclusion very unconvincing."
• RECENT MORPHOLOGICAL STUDIES: Willmer (1990, p. 245) concluded that rotifers and acanthocephalans are "probably related." Nielsen (1995, p. 252) considered it "clear that the acanthocephalans and rotifers must be sister groups, and that the acanthocephalans therefore cannot be 'parasitic rotifers'."
• Analyses of 18S rDNA sequences indicate that Acanthocephala are closely related to Rotifera within Lophotrochozoa (Wallace, Ricci, and Melone 1996; Winnepenninckx et al. 1995). Garey et al. (1996) found that Acanthocephala arises from within Rotifera, which would make them derived rotifers presumably highly modified by parasitism. A later study using sequences from more species, however, indicated that Acanthocephala are merely the sister group of the monophyletic Rotifera (García-Varela et al. 2000).

Cycliophora appear to be related to rotifers.

• HISTORY: Symbion pandora was first collected in the 1960s from the mouthparts of the Norway lobster, but it was assumed to be a rotifer and stored in a museum drawer. The spe-cies was then rediscovered by Peter Funch and Reinhardt Kristensen, who, after studying its many unique features and complex life cycle, erected the new phylum Cycliophora in 1995. Funch and Kristensen proposed that Cycliophora was close to Entoprocta and Ectoprocta.
• Comparisons of 18S rDNA suggest that Symbion is in Lophotrochozoa and is closer to Ro-tifera than to Entoprocta or Ectoprocta (Winnepenninckx, Backeljau, and Kristensen 1998).

Echiura and Pogonophora may be polychaete annelids.

• HISTORY: Hyman planned to cover both Echiura and Annelida in the same volume, indicating that she considered them to be closely related. Most authors continue to ally the echiurans (together with sipunculans) to the annelids, as well as to molluscs, on the basis of their trochophore larvae and spiral cleavage. In 1959, when Hyman wrote about them, Pogonophora were still a new and little-known phylum usually allied with the hemichordates. "It is not open to doubt," she wrote (vol. 5, p. 224), "that the Pogonophora belong to the Deuterostomia." Since then, the clearly segmented opithosome has been found, their cleavage has been determined to be spiral, and their larvae have been found to be trochophores. Most authors now therefore consider the pogonophorans to be most closely related to annelids. Some authors have accepted the proposal by Meredith Jones (1985) that the phylum be divided into two: phylum Vestimentifera and phylum Pogonophora (= Frenulata). Most, however, continue to regard both groups as members of the monophyletic Pogonophora.
• RECENT MORPHOLOGICAL STUDIES: Willmer (1990, p. 215), after discussing the evidence that echiurans show tentative segmentation during development, concluded that they "should not be placed within the same phylum as segmented annelids, [b]ut the two key features of development to identical trochophores and the presence of identical chaetae (together with a number of other similarities...) should still be enough to keep the two phyla very closely allied in any phylogenetic scheme." She considered it reasonable to place the pogonophorans "not too distant from the main annelid/echiuran lineage" (p. 216). Nielsen (1995, p. 140) concluded that "at present the pogonophorans must thus be regarded as a specialized polychaete group." He also (p. 142) tentatively included the echiurans within Annelida.
• Using base sequences for the EF-1∀ gene, McHugh (1997) concluded that vestimentiferans and echiurans arose separately from within the polychaete annelids. Using the amino-acid sequences for EF-1∀, Kojima (1998) found the same result for vestimentiferans. Halanych, Lutz, and Vrijenhoek (1998), using sequences from genes for mitochondrial cytochrome c oxidase subunit 1, 18S rRNA, and 28S rRNA, found that vestimentiferans and other pogonophorans arose from within the polychaetes relatively recently. These findings, as well as evidence that oligochaetes and leeches also arose from polychaetes, renders Polychaeta paraphyletic.
• Pogonophorans appear to arise from within the order Sabellida and are now often referred to as the family Siboglinida.

Chaetognatha may not be closely related to other deuterostomates.

• HISTORY: Among the many hypotheses for chaetognath affinities was the suggestion first made in the 1860s that they are related to nematodes on the basis of the thick cuticle, similar arrangement of body-wall musculature, and similarities of the grasping spines to the adhesive bristles on the heads of certain marine nematodes (Hyman, 1959, vol. 5, pp. 1, 3). According to Hyman, that view was still current through the 1950s. Hyman acknowledged the similarity of adult chaetognaths to "aschelminths," but she gave more weight to the radial, indeterminate cleavage and deuterostomy. She also noted that in the juvenile the coelom develops enterocoelously, although not in the same way as in Echinodermata, Hemichordata, and Chordata. Her final conclusion was that she could not relate Chaetognatha to any other phylum, and that they were perhaps derived from the early bilateria. Because of "the possibility that Chaetognatha are remotely related to the dipleurula ancestor of the other Deuterostomia," however, she (1959, vol. 5, p. 66) placed them among the deuterostomates. This practice has generally been followed since.
• RECENT MORPHOLOGICAL STUDIES: Willmer (1990, p. 319) regarded the association of Chaetognatha with deuterostomates as "extremely tenuous" and decided that the most prob-able origin was from "the acoeloid or proto-platyhelminth form that may be at the roots of the Metazoa." Nielsen (1995, p. 235) tentatively placed the chaetognaths within Aschelminthes in an unresolved trichotomy with (Rotifera + Acanthocephala) and his clade Cycloneuralia, which includes Nematoda and others. (See indented list.)
• Several analyses using the base sequences of 18S rDNA from several species indicate that chaetognaths are not closely related to deuterostomates (Giribet et al. 2000; Telford and Holland 1993; Wada and Satoh 1994b). There is some indication that chaetognaths evolved as a distinct clade from the base of the Bilateria, but this may be a consequence of long-branch attraction. One analysis using 18S rDNA sequences suggests that chaetognaths are related to Nematoda in Ecdysozoa (Halanych 1996).

Molecular evidence supports the conventional phylogeny of echinoderm classes.

• HISTORY: Paleontological evidence as well as morphological characters have long supported the view that crinoids are the oldest extant echinoderms, followed by ophiuroids and aster-oids, and finally echinoids and holothuroids.
• Evidence from both 18S rDNA and mitochondrial gene rearrangements support this phylog-eny (Smith et al. 1993; Wada and Satoh 1994a).

Concentricycloids may be asteroids

• HISTORY: Baker, Rowe, and Clark (1986) described a small, disc-shaped animal found in wood collected from kilometer-deep ocean near New Zealand and named it Xyloplax medusiformes. A second species of Xyloplax was discovered later. Although Xyloplax has pentaradial symmetry and podia, it does not have the test of an echiuroid, the cucumber-shape of a holothuroid, or the arms of a crinoid, ophiuroid, or asteroid. For this and other reasons, Baker et al. proposed the new class Concentricycloidea.
• Janies and Mooi (1998) found that both 18S rDNA sequences and morphology placed Xyloplax within Asteroidea.

Hemichordata may be closer to Echinodermata than to Chordata.

• HISTORY: Until the 1950s Hemichordata had frequently been included in the phylum Chordata, and Hyman (1959, vol. 5, p. 74) considered it "impossible to deny" that the phylum Hemichordata was related to chordates. On the other hand, the tornaria larva of enteropneusts resembles an asteroid larva so closely that Hyman (1959, vol. 5, pp. 197-199) was moved to write that, "There appears no escape from the conclusion that hemichordates and echinoderms stem from a common ancestor…. In other words, the common ancestral stock gave off the echinoderms as a blind branch, then continued along its main line of evolution to hemichordates and chordates." Most authors have cited the presence of pharyngeal slits and a dorsal, hollow nerve cord as evidence for a sister-group relationship of Hemichordata to Chordata rather than to Echinodermata.
• RECENT MORPHOLOGICAL STUDIES: Willmer (1990) expressed no strong conviction about the relative position of Hemichordata to Chordata versus Echinodermata, but in her summary figure (p. 361) she placed the Hemichordata as a branch below the Echinodermata on the line terminating with Chordata. Nielsen (1995, pp. 333, 385) divided hemichordates into two phyla, Pterobranchia and Enteropneusta. He placed Enteropneusta as the sister group of Chordata in a clade he named Cyrtotreta, and he placed Pterobranchia in an unresolved trichotomy with Echinodermata and Cyrtotreta. (See indented list.)
• One early study that used 18S rDNA sequences from four deuterostomate species suggested that the acorn worm Saccoglossus was closer to vertebrates than to the echinoderm (Holland, Hacker, and Williams 1991). More recent studies using more 18S rDNA sequences and a variety of analytical methods almost invariably show Hemichordata to be monophyletic and more closely related to Echinodermata than to Chordata (Cameron, Garey, and Swalla 2000; Halanych 1995; Turbeville, Schulz, and Raff 1994).

Vertebrates apparently did not evolve from an echinoderm.

• HISTORY: Hyman (1959, vol. 5, p. 201) unequivocally rejected the idea, then "widely spread," that vertebrates originated directly from an echinoderm independently of cephalochordates and urochordates. That view was resurrected, however, by R. P. S. Jefferies (1986), a paleontologist who claims to have identified gill slits, a brain, notochord, dorsal nerve cord, and other chordate features in fossils of extinct "calcichordates," which are considered by most paleontologists to have been echinoderms. Although most paleontologists do not accept Jefferies identification of these features, the calcichordate theory is still often treated seriously. (Refer to Gee 1996 for further discussion.)
• RECENT MORPHOLOGICAL STUDIES: Nielsen (1995, pp. 377-378) also rejected Jefferies analysis of calcichordate morphology and the calcichordate theory.
• Analysis of 18S rDNA sequences show that Urochordata, Cephalochordata, and Vertebrata all belong to the same clade (Chordata) that is sister to, but clearly separate from, the clade that includes Echinodermata and Hemichordata. (See, for example, Wada and Satoh 1994b; Zrzavý et al. 1998.) This is not the result that would be expected if Chordata or Vertebrata evolved from an echinoderm.

Cephalochordata, rather than Urochordata, may be the sister group of Vertebrata.

• HISTORY: In the first half of the 20th century the amphioxus Branchiostoma was considered to be the closest extant relative of vertebrates. Like vertebrates, cephalochordates have myomeres, a ventral pulsating blood vessel that may be homologous with the vertebrate heart, an intestinal diverticulum that resembles the embryonic precursor of the vertebrate liver, and separate dorsal and ventral roots of the spinal cord. One vertebrate feature that cephalochordates apparently lack is a head. Instead the notochord extends to the front of the animal, and there is only an anterior enlargement of the nerve cord in place of a brain. Largely for that reason, many zoologists prefer Garstang's suggestion that vertebrates evolved from a larval urochordate by paedomorphosis, even though there is little direct evidence for the idea. (See Gee 1996 for further discussion.) Northcutt and Gans (1983; Gans and Northcutt 1983) have attempted to revive the cephalochordate theory by noting that vertebrate cranial structures, such as sense organs and muscles, originate from neural crest, unlike similar structures in the rest of the body. They argue, therefore, that the head is apomorphic in vertebrates and evolved in an amphioxus-like ancestor during the transition from filter feeding to predation.
• RECENT MORPHOLOGICAL STUDIES: Nielsen (1995, p. 396) distinguished between the stiffening rod ("urochord") in the tails of Urochordata and the more-anterior notochords of cephalochordates and vertebrates. This and other characters led him to divide the clade Chordata into the phylum Urochordata and its sister group Notochordata, with the latter comprising the two phyla Cephalochordata and Vertebrata (pp. 385, 418). Thus Vertebrata would have shared a more recent ancestor with Cephalochordata than with Urochordata.
• Analysis of 18S rDNA sequences suggests that Branchiostoma is the sister clade of Vertebrata (Wada and Satoh 1994b). That study also showed that Urochordata, represented by an ascidian, a larvacean, and a salp, form a monophyletic clade outside Cephalochordata + Vertebrata.
• The affinity of Cephalochordata to Vertebrata is further supported by the similarity of the Hox genes in Branchiostoma to those in humans and mice (Garcia-Fernàndez and Holland 1994).

Turtles may be the sister group of Crocodilia + Aves rather than basal Reptilia.

• HISTORY: The fossil record for turtles goes back only to the Jurassic period, long after the diversification of the main reptilian lines. It is only because turtles lack skull fenestrations that they have been assumed to be derived from the anapsids of the Permian period, which are assumed to have been basal reptilians. Turtles are therefore generally assumed to be the most basal of extant reptiles.

Using several unusual features of the mitochondrial genome, as well as sequences from genes for mitochondrial rRNA, Zardoya and Meyer (1998) found that turtles (Testudines) are the sister group of Archosauria (Crocodilia + Aves). Tuataras and lizards form a clade that is basal to Archosauria + Testudines. Turtles are therefore derived diapsids.

Placental mammals may be divided into four superordinal clades.

• HISTORY: Relationships among orders of placental mammals are largely unresolved, partly because of their rapid diversification in the late Cretaceous and early Tertiary periods and partly because of extremes in convergent and divergent adaptation to a wide range of habitats.
• By synthesizing several hundred morphological and molecular phylogenetic trees, Liu et al. (2001) identified nine well-supported clades. Their clades are generally consistent with two purely molecular-phylogenetic trees published by Madsen et al. (2001) and Murphy et al. (2001).
• These molecular studies support four superordinal clades of placental mammals that di-verged in the following sequence: Afrotheria, Xenarthra, Glires + Euarchonta (flying lemurs, tree shrews, and primates), and a clade named by Madsen et al. Laurasiatheria. Traditional orders within these four clades are shown in Figure 19.

Figure 19. Traditional orders of placental mammals arranged into four major clades based on analysis of gene sequences. Adapted mainly from Murphy et al. (2001).

 

Home | Preface | Table of Contents | Section I | Section II | Section III | Section IV | Glossary |
| References | Downloadable Word Document |