On the Origin of Molluscs, the Coelom, and Coelomic Segmentation

ON THE ORIGIN OF MOLLUSCS, THE COELOM, AND COELOMIC SEGMENTATION

JOSEPH VAGVOLGYI

Abstract Molluscs had a common origin with the annelids, as shown by remarkable ontogenetic similarities. Their common ancestors were non-segmented (non-eumetameric), acoelomate Downloaded from https://academic.oup.com/sysbio/article-abstract/16/2/153/1647154 by guest on 20 April 2020 animals. Molluscs developed from the common stock through incorporation of the distinguishing molluscan characters such as the shell, mantle, mantle cavity, radula and ctenidium and annelids by the adoption of coelomic segmentation and of changes accompanying it. The assertion that molluscs hand ancestors with coelomic segmentation is contradicted by ontogenetic and anatomical facts as well as by theoretical considerations. The common moHuscan-annelid stock most likely originated from ancestral flatworms; this is indicated by the developmental, and to a lesser extent, structural similarities between the flatworms, molluscs and annelids, and by the lower level of organization of the flatworms. The coelom is a phylogenetically new structure in the molluscs and annelids, and was acquired independently in the two phyla. Coelomic segmentation also is a phylogenetically new acquisition of the annelids; it came about through developmental changes affecting the growth of the mesoderm bands. These views are the corollaries of ontogenetic considerations.

A widely held hypothesis concerning the crawling mode of locomotion. And finally, origin of molluscs is the so-called annelid Lemche (1957) reported the discovery of a theory. According to Hyman (1951:4), the "living fossil," a primitive, segmented mol- "annelid theory" states that ". . . annelids, lusc, Neopilina galatheae Lemche, 1957. arthropods, and chordates are closely re- Soon two more species of Neopilina were lated and [that] lower bilateral groups may found, N. ewingi Clarke and Menzies, 1959, have arisen from annelids by degeneration." and N. veleronis Menzies and Layton, 1963. In the present paper the term is restricted, These discoveries were accepted by many referring to the theory that annelids gave as evidence for the segmented origin of rise to molluscs. Early proponents of this molluscs (Lemche, 1959a, 1959b, I960; theory (Pelseneer, 1899; Heider, 1914; Lemche and Wingstrand, 1959b; Portman, Soderstrom, 1925; Naef, 1926) believed that I960; Fretter and Graham, 1962). Lemche the serial repetition of certain organs in went so far as to postulate that molluscs are some molluscs represented vestigial segmen- more closely related to the arthropods than tation, and claimed on this basis that mol- to the annelids, and that the latter arose luscs descended from the segmented worms, independently of the former two groups. annelids. They also believed that the far- Fretter and Graham supported this argu- reaching similarities in cleavage and early ment by claiming that the coelom is primar- development were further indications of this ily small in the arthropods as well as in the relationship. More recent investigators molluscs. added other viewpoints to these arguments. Opponents of the annelid theory have Thus, Garstang (1928) theorized that mol- repeatedly pointed out the weaknesses of luscs evolved from the trochophore larvae it. Thus, Nierstrasz (1922), Hammersten of the annelids by neoteny. Johansson and Runnstrom (1925), Ivanov (1928, 1944), (1952) aimed at giving a functional expla- Beklemishev (1958a, 1958b), Boettger nation to the annelid theory, when he pro- (1959), Steinbbck (1962) and Hunter and posed that the molluscs had lost the segmen- Brown (1965) argued that molluscs cannot tation of their annelid ancestors in turning be considered as annelids with reduced to living on hard, rocky surfaces of the segmentation, because of fundamental ana- intertidal zone, and assuming a creeping, tomical and developmental differences in 153 154 SYSTEMATIC ZOOLOGY

segmentation. Odhner (1961) pointed out The claims of Lang, Nierstrasz and Gra- that molluscs attained segmentation inde- ham that molluscs originated independently pendently of the annelids. Beklemishev of the annelids overemphasize the similari- (1958b), Boettger (1959), Steinbock (1962) ties between the molluscs and flatworms and Hunter and Brown (1965) have argued and underemphasize those between the mol- convincingly that Neopilina cannot be con- luscs and annelids. Also, it is not certain sidered proof for the annelid origin of mol- whether some of the molluscan-flatworm Downloaded from https://academic.oup.com/sysbio/article-abstract/16/2/153/1647154 by guest on 20 April 2020 luscs, because of differences in the segmen- similarities are due to close relationships or tations, because Neopilina is not at the root to parallel evolution. For these reasons, I of the molluscan phylogeny, and for various reject these claims. It also seems doubtful other reasons. Yochelson (1963) reached a that the construction by Beklemishev of similar conclusion from paleontological hypothetical larval forms satisfying superfi- studies; he maintained that the monoplaco- cial morphological requirements is of any phoran molluscs (to which Neopilina be- help. The ideas expressed by Hammarsten longs ) do not constitute the basic molluscan and Runnstrom and Boettger (that the an- stock; rather, they are a primitive group of cestors of aschelminths came from ances- the more advanced molluscs. The results of tors of the flatworms), on the other hand, Schmidt's microscopic and crystallographic account both for the molluscan-flatworm studies (1959) on the shell of Neopilina and the molluscan-annelid similarities, and are neatly consistent with this conclusion. avoid making unwarranted and unnecessary Further objections, not expounded in the assumptions. Therefore, they appear to rep- literature, also can be brought up against resent "the truth," according to our present the annelid theory. The theory does not knowledge. explain why we must consider molluscan Thus, it seems that the problem of the segmentation secondary and reduced, or origin of molluscs already has been solved. why we must accept the ontogenetical sim- Many zoologists, however, perhaps even the ilarities between molluscs and annelids as a majority, are either unaware of, or unwilling proof of the annelid theory, when alterna- to accept, this answer. The present paper tive assumptions also are possible. Weak- has been written with the purpose of sum- nesses in Garstang's and Johansson's theory ming up the arguments, elaborating upon also are found, as I will show in a later sec- those that were not explicit or expounded in tion. due detail, adding some new arguments, To supplant the annelid theory, Lang and thereby presenting a unified and hope- (1896), Nierstrasz (1922), and Graham fully convincing argument for the case. (1955) proposed that molluscs evolved from flatworms or ancestors to the flatworms, DISCUSSION completely independently of the annelids. Hammarsten and Runnstrom (1925), Boet- It seems profitable to arrange the argu- tger (1959 and Beklemishev (1963, on the ment around the three main points of the other hand, argued that molluscs and anne- theory followed here. These are: 1) mol- lids evolved together but separated from luscs and annelids had evolved together; each other before the annelids acquired 2) the molluscs separated from the annelids coelomic segmentations. Whereas, however, before they acquired the distinguishing mol- Hammarsten and Runnstrom believed that luscan criteria such as shell, mantle, mantle the common molluscan-annelid stock origi- cavity, radula and ctenidium (Yonge, 1960; nated from the ancestors of the flatworms, Morton and Yonge, 1964), and the annelids Boettger thought that the common stock acquired coelomic segmentation; 3) the evolved from a group near the root of the most likely ancestors of the common mol- aschelminths, and Beklemishev thought luscan-annelid stock were the forms ances- it evolved from "larval" coelenterates. tral to the present-day flatworms. ON THE ORIGIN OF MOLLUSCS 155 Downloaded from https://academic.oup.com/sysbio/article-abstract/16/2/153/1647154 by guest on 20 April 2020

FIG. 1. Segmentation (pseudometamery) of the monoplacophoran mollusc Neopilina galatheae; diagrammatic. Notice that the various segmented organs occur in different numbers, and that there is only partial harmony in their arrangement. The coelomic cavity is represented by the pericardial (not shown) and nephridial cavities. (From Lemche and Wingstrand, 1959b, Galathea Rep. 3:9-71, with permission of the editor.) 1, mouth; 2, pedal nerve cord; 3, nephridium; 4, pedal retractor muscle; 5, commissure; 6, gonad; 7, gill; 8, aorta; 9, ventricle; 10, auricle; 11, lateral nerve cord; 12, anus.

The common origin of molluscs and an- and the annelids have spiral cleavage; the nelids.—This contention is based upon the fate of certain cells is, with some exception, similarities between the molluscs and anne- the same in the two groups; mesoderm lids in cleavage, embryonic development, bands arise from cell 4d in both groups; and early larval stage. Both the molluscs and the early larval stage, the trochophore, 156 SYSTEMATIC ZOOLOGY is nearly identical. These similarities are but they do not support the claim that mol- accepted as indications of a genuine and luscs developed from annelids. close relationship by practically all zoolo- The separation of molluscs and anne- gists (Hyman, 1951; Barnes, 1963). A con- lids.—The separation of molluscs and anne- sensus, admittedly, does not constitute a lids must have happened at or before the proof; it does give, however, the best sup- time molluscs and annelids acquired their port to the above view. distinguishing features (Hammarsten and Downloaded from https://academic.oup.com/sysbio/article-abstract/16/2/153/1647154 by guest on 20 April 2020 The developmental similarities in ques- Runnstrom, 1925; Boettger, 1959), but not tion were interpreted by some followers of after. There is no evidence whatsoever for the annelid theory (Naef, 1926) as support- the assumption that annelids had ancestors ing the contention that the molluscs evolved with characteristic molluscan features, or from the annelids. It is quite clear, how- that molluscs had ancestors with character- ever, that this claim is unwarranted. Sim- istic annelid feature. Therefore we have to ilarities in the ontogeny only "prove" that assume that each line acquired its own char- the groups in question are closely related, acteristic features after the separation of but they do not "prove" the direction of the two lines. evolution. On the basis of the developmen- Most zoologists would agree with the tal similarities alone, the direction could first part of this conclusion, that annelids equally well be from molluscs to annelids, did not come from molluscs; but many from annelids to molluscs, and from a com- would also maintain that molluscs did have mon ancestor toward both. Only differ- segmented ancestors. They argue that a) ences in the ontogeny could indicate direc- primitive molluscs are segmented, and that tion; a primitive condition in one group, b) this must mean that molluscs descended and a more advanced in the other, would from ancestors with a more complete seg- indicate that evolution proceeded from the mentation. former to the latter. The decision whether the primitive mol- There are two obvious, although slight, luscs, or the molluscs in general, are or are differences between the molluscan and an- not segmented depends upon one's defini- nelid ontogenies. One is the formation of tion of segmentation. Adherents of the an- the crosses, the other that of the coelomic nelid theory consider segmentation to be cavities. The cells forming the major part any repetition of organs along the main axis of the equatorial ciliary band of the troch- of the body. Thus, they say that the mol- ophore larva are arranged, at an early luscs are segmented. Evidently, these stage, in the shape of a cross centered on workers consider the segmentation of mol- the animal pole; the arms of the cross are luscs of the same kind as that of the anne- radial in molluscs, interradial in annelids, lids. Many others, however, consider the and they are derived from cells Ia12-ld12 in serial repetition of certain organs, without molluscs, from Ia112-ld112 in annelids (Nier- corresponding coelomic subdivisions and strasz, 1922). Now, if we could ascertain without the developmental characteristics which type of cross formation were the more of the annelid segmentation (see below) primitive, we could possibly draw some merely regularization metamery or pseudo- conclusion as to the direction of evolution; metamery; and these others distinguish be- but, unfortunately, nothing is known on this tween this and the mesodermal or coelomic subject. In the formation of the coelom or true segmentation, referred to henceforth (p. 158), the direction seems to have been as eumetamery, of the annelids (Ivanov, from molluscs to annelids. Thus, summing 1928, 1944; Beklemishev, 1958a, 1958b; up, the developmental similarities indicate Steinbock, 1962; Clark, 1964; Hunter and close relationship between the molluscs and Brown, 1965). In this view, then, the mol- annelids, they suggest that in some respects luscs lack true segmentation or eumetamery molluscs are more primitive than annelids, and have only pseudometamery. ON THE ORIGIN OF MOLLUSCS 157 Downloaded from https://academic.oup.com/sysbio/article-abstract/16/2/153/1647154 by guest on 20 April 2020

FIG. 2. Segmentation (pseudometamery) of a polyplacophoran mollusc (chiton); diagrammatic. No- tice that the number and arrangement of the various segmented organs, as in Neopilina, is not matching. The coelom is represented by the pericardial (not shown) and nephridial cavities. (Redrawn from Barnes, R., 1963, Invertebrate zoology, after Haller, Lang and Yonge, with permission of the author and the publisher.) 1, mouth; 2, shell plate; 3, gill; 4, foot; 5, girdle; 6, anus; 7, ventricle; 8, auricle; 9, kidney; 10, gonad; 11, aorta; 12, commissure; 13, pedal nerve cord; 14, lateral nerve cord; 15, brain.

The differences between the pseudo- pilina galatheae has two pairs of atria, five metamery of molluscs and the eumetamery pairs of gills, two pairs of gonads, six pairs of annelids are anatomical, ontogenetical and of nephridia, eight pairs of pedal retractor phylogenetical. There are three main ana- muscles and ten pairs of nerve commissures; tomical differences (Beklemishev, 1958a; these numbers may slightly be different in Boettger, 1959; Steinbock, 1962). First, the N. ewingi and N. veleronis. Chitons have molluscan coelom is not segmented (figs. 1 eight shell plates with attached muscles, and and 2; also see p. 158 about the nephrocoel), a varying, sometimes very great, number of whereas in annelids, there is typically a pair gills and nerve commissures. Furthermore, of coelomic sacs in each segment (Fig. 3). the gills of chitons are not even paired, The molluscan coelom is not even fully strictly speaking; their number may be dif- homologous with the annelid coelom; this ferent on the right and left sides in as many topic may be discussed more profitably as 48% of the specimens in some species later, after the ontogeny has been described. (Hunter and Brown, 1965). In contrast to Second, in molluscs the various segmented this, the segmented organs of the annelids, organs may occur in different numbers. the parapodia, ganglia, commissures, circu- Thus, the monoplacophoran mollusc Neo- lar blood vessels, nephridia, coelomoducts 158 SYSTEMATIC ZOOLOGY and coelomic sacs occur in identical num- mesodermal bands. Therefore, the mollus- bers (or multiples), and are strictly paired. can coelom is not entirely homologous with Third, the arrangement of the various seg- the annelid coelom. It qualifies as a coelom mented organs is less regular in the molluscs since it is surrounded by mesodermal tissue, than in the annelids. It is true that Neo- but this is the extent of its homology with pilina exhibits an almost completely orderly the annelid coelom. arrangement; but in chitons the gills are We turn now to the development of other Downloaded from https://academic.oup.com/sysbio/article-abstract/16/2/153/1647154 by guest on 20 April 2020 unevenly distributed between the body seg- segmented organs. The shell plates first ments, as determined by the shell plates, appear in the early larva; they originate by and the commissures are arranged discord- the simultaneous subdivision of the origi- antly from both the gills and the shell plates. nally undivided shell gland into six plates. Thus, we have to maintain that molluscs as Later, one more plate will be added anteri- a group have a less orderly arrangement orly and another posteriorly, and thereby than do annelids. the definitive number of eight will be estab- The ontogenetic differences between mol- lished (Beklemishev, 1958a; Hunter and luscan and annelid segmentation are as fol- Brown, 1965). The gills appear weeks later, lows. In molluscs, specifically in the chiton after thirty days of life; first the posterior Acanthochiton discrepans Brown (Hammar- gills appear, and, as the animal grows, more sten and Runnstrom, 1925; Raven, 1958), gills are added anteriorly (Hunter and coelomic cavities do not develop inside the Brown, 1965). The nerve commissures orig- mesodermal bands, and the latter do not inate in a different manner again (Hammar- become metamerized; instead, they differ- sten and Runnstrom, 1925; Raven, 1958). entiate into various organs, among them Isolated cells from the cerebral ganglia mi- the pericardial-nephridial complex. It is grate posteriorly, dividing mitotically as true that Naef (1926), following earlier au- they go, and form thereby the lateral and thors, speaks of coelomic cavities in the pedal connectives and the ganglia there- mesodermal bands, but Hammarsten and upon. From the connectives, the commis- Runnstrom categorically deny the existence sures grow out presumably by this same of these when they write: ". . . die von method. Hammarsten and Runnstrom men- Kowalewsky beschriebenen Cblomsacke tion, however, that in other molluscs the nicht existieren . . ." (1925:277). These au- ganglia in the body and the foot arise from thors assume that Kowalewky mistook the local ectodermal thickenings, and the con- nerve strands, the liver anlage, etc., for the nection between them is established secon- mesodermal bands. Raven, in his compre- darily (1925:311). They presume, how- hensive work on molluscan development ever, that the manner of growth observed (1958), follows Hammarsten and Runns- in chitons is the primary, primitive method. trom. According to these authors, a single We may conclude this section saying that coelomic cavity appears in chitons, in the the development of the various segmented anlage of the pericardial-nephridial com- organs is not coordinated in molluscs; the plex. This cavity has, at the beginning, a segmented organs develop from different single pair of ducts, which will become the rudiments, at different times and different nephridia; the genital ducts appear only rates. later, and primarily from ectodermal The ontogeny of the annelid segmenta- sources, not from the mesoderm which tion is completely different (Ivanov, 1928, makes up the pericardial-nephridial com- 1944; Beklemishev, 1958a; Dales, 1963). We plex. The coelomic cavity may become par- will consider the presumably primitive mode tially subdivided into pericardial and ne- of development, found in the oligochaetes phridial cavities, but it is the only cavity to and some polychaetes; but it must be appear. In the annelids, on the other hand, pointed out that the other polychaetes also a series of coelomic cavities appear in the follow a basically similar course of develop- ON THE ORIGIN OF MOLLUSCS 159 ment, the differences being inconsequential from the point of view of the present argu- 7 ment. In this primitive ontogeny of segmentation, the mesoderm of the early embryo, derived from 4d, becomes metamerically subdivided into a small number of segments, depending upon the number of the Downloaded from https://academic.oup.com/sysbio/article-abstract/16/2/153/1647154 by guest on 20 April 2020 ectodermal metameric organs. The subdivision of the mesoderm of the early larva is a simultaneous process. Not so the segmentation of the mesoderm of the metatro- chophore stage. This mesoderm also is derived from the cell 4d (in oligochaetes and some polychaetes); it forms two bands, one on each side of the body. Inside the mesodermal bands, a series of cavities appear, in an anterior-to-posterior order. Each pair of cavities plus the surrounding tissues develop into a segment; thus, each segment arises as a unit. The cavities become the coelomic sacs, while their lining differen- tiates into the peritoneum, the musculature of the gut and the body wall, the blood vessels and the coelomoducts; from the ecto- derm of each segment, the epidermis, the ganglia, the nerve strands and the chaetae develop. At the posterior end of the mesodermal bands, cell division continues and new coelomic cavities appear, and thereby new segments are continuously produced. Clearly, molluscs do not have this kind of segmentation. FIG. 3. Segmentation (eumetamery) of the polychaete annelid, Nereis virens. Only the an- The presumed phylogenetic differences terior region is shown; the total number of seg- between the molluscan and annelid segmen- ments may reach 200. The dorsal body wall has tation are as follows. The molluscan seg- been partially removed; the ventral nervous system is not shown. Notice that the number and arrange- mentation (pseudometamery) was prob- ment of the various segmented organs are match- ably achieved in two stages. These can be ing, and that there is a pair of coelomic sacs in sharply distinguished for didactic purposes each segment. (From Brown, F. A. [ed.], 1950, but could have been concomitant in real- Selected invertebrate types, with permission of the ity. In the first stage, there was an increase publisher.) 1, palp; 2, prostomium; 3, eyespots; 4, peristomium; 5, parapodia; 6, coelomic sacs; 7, sep- in the number of certain organs; this may tum, separating coelomic sacs; 8, stomach; 9, ven- have come about by actual multiplication tral blood vessel; 10, nephridium; 11, lateral blood of an organ, or by the breaking up of one vessel; 12, esophageal coecum; 13, esophagus; 14, large organ into several smaller ones (the dorsal blood vessel; 15, pharynx; 16, cirri; 17, nephridia of Neopilina may exemplify this). tentacle. The increase was not necessarily equal; certain organs were multiplied to a greater ex- became arranged according to a unified tent than others. In the second stage, there plan. The second stage is now often incom- was a regularization of the multiplied or- plete. The annelid segmentation (eumeta- gans in number and position, so that they mery), on the other hand, came about in 160 SYSTEMATIC ZOOLOGY one stage, by the production of identical the molluscs and the annelids, and tried to sets of organs in great numbers; thereby the derive the former from the latter. And, number and arrangement of the various similarly, Lemche and Wingstrand wrote segmented organs immediately exhibit a (1959b: 66): "We do not know whether the unified, segmented pattern (Beklemishev, dorsal coelom is metamerically subdivided 1958a). or not, but the metameric tendency is pres- The above presumptions are supported ent in the gonads which may be regarded as Downloaded from https://academic.oup.com/sysbio/article-abstract/16/2/153/1647154 by guest on 20 April 2020 by ontogenetic considerations. As ontogeny part of the coelomic system." And (p. 67): shows, the segmented organs of the annelids "The complete organization and the pres- develop in sets, each set differentiating out ence of a heart in the last segment makes it from an ectodermal and mesodermal block [the last segment] very different from an of cells. The segmented organs of the mol- arthropod or annelid telson, indicating that luscs, on the other hand, develop from dif- the metameres in 'Neopilina are not formed ferent rudiments, at different times and by a rhythmic activity of a terminal growth different rates from one another. Further- centre. The more probable explanation is more, in the various molluscan groups that that the mesoderm in the body region of exhibit segmentation (monoplacophorans, Neopilina is simultaneously subdivided in polyplacophorans, some gastropods and the manner characteristic of the most an- cephalopods), different organs may be mul- terior segments in annelids and arthropods tiplied, and the regularization may be on [i.e., those formed in the early larva]." But, quite different levels. Thus, the mono- nevertheless, they conclude (p. 66): "The placophoran mollusc Neopilina galatheae metamerism of Neopilina is so regular and has six organs in multiple numbers (Fig. 1), is present in so many organ systems that which are quite harmoniously arranged; there is no reason to regard it as different chitons have their shell plates, musculature, from that of annelids and arthropods." gills and commissures multiplied, but only Portman (1960) repeats this argument al- the shell plates and the muscles are har- most word for word. Extending this argu- moniously arranged; some prosobranch ment, one could maintain that there is no snails have but one structure, the commis- reason to regard the segmentation of the sures, multiplied; tetrabranch cephalopods cestodes or the vertebrates as "different" have their gills and atria multiplied, which from that of the annelids and arthropods, are in complete harmony. Had the mol- because all are very regular. luscan segmentation evolved in the same We may conclude this section saying that manner as the annelid segmentation, this the molluscan segmentation differs from the chaotic pattern would make no sense. annelid segmentation in anatomy, ontogeny Most of the differences between mollus- and phylogeny; the molluscan segmentation can and annelid segmentation were known is pseudometamery, the annelid eumeta- by the adherents of the annelid theory. mery. Thus, Naef wrote (1926:41): "Man muss The second assumption of the annelid nun zugeben, dass ein solcher Prozess [seg- theory is that molluscan segmentation de- mentation of the mesoderm bands] bei Mol- veloped through reduction from a previous lusken nicht in typisch unverkennbarer fuller segmentation. Generally this is in- Form beobachtet ist. . . . Weiter ist zuzuge- terpreted as meaning that the molluscs de- ben, dass von einer Segmentierung des veloped from annelids that lost their seg- Cb'loms in zahlreiche und regelmdssige Ab- mentation. This assumption is not only schnitte und weiter von einem Colom vom hypothetical (Hammarsten and Runnstrom, funktionellen Character dessen wohlausge- 1925) but also erroneous. First of all, the bildeter Anneliden bei Mollusken keine segmentation of molluscs is different in kind Rede sein kann." But, in spite of these ad- from that of the annelids, and thus irreduc- missions, he equated the segmentation of ible from the latter. A second considera- ON THE ORIGIN OF MOLLUSCS 161

tion is as follows. An atypical or incomplete segmentation such as that of the molluscs may be interpreted in two different ways: either as primarily incomplete or as secondarily reduced. The former is the simpler interpretation, the latter the more complicated. On the principle of parsimony Downloaded from https://academic.oup.com/sysbio/article-abstract/16/2/153/1647154 by guest on 20 April 2020 we should accept the simpler interpretation as valid unless some evidence compels us to do otherwise. No such evidence has been brought forth by the annelid theory. There- fore, we should consider valid the simpler interpretation, that the molluscan segmentation is primary. In Johansson's opinion however, this is not so. He maintains that the segmented coelom of the presumed ancestors of molluscs was reduced when the latter became ". . . adapted to hard bottoms in the surf region and has developed a foot and a shell .... In this connection the segmented coelom, being no longer of importance for locomotion, was reduced to form a single cavity with one or two pairs of ducts" (1952: 290). He considers the constancy in the number of the shell plates of chitons as the proof ". . . of a formerly complete true metamery as in the Arthropoda" (p. 287). He fails to explain why the constant number of the shell plates has such significance. He 12 did not explain either why he disregarded the facts that, during ontogeny, the segmentation and the subsequent reduction of the coelom can be followed step-by-step in the 8 arthropods, not at all in the molluscs, or why life on the rock bottom in the surf zone had to lead to the reduction of segmentation. The truth is that coelomic segmentation is fully compatible with life in such habitats, as shown by the fact that". . . poly- FIG. 4. Segmentation (pseudometamery) of the chaetes are extremely numerous in the rocky triclad flatworm Procerodes lobata. Excretory sys- regions along the shore . . ." (McGinitie tem omitted. Notice that the number and arrange- and McGinitie, 1949:194). ment of the commissures does not match that of the other segmented organs, and that there is no coelom Contrary to Johansson's opinion, a com- at all. The segmentation of flatworms is of the parison of the locomotory mechanisms of same kind as that of the molluscs. (From Lang, A., 1881, Mittheil. Zool. Stat. Neapel 3:187-251, with permission of the publisher.) 1, brain; 2, commissure; 3, yolk gland; 4, proboscis; 5, circular nerve 12, sperm duct; 13, testes; 14, salivary glands; 15, of proboscis; 6, mouth; 7, penis; 8, genital opening; intestinal diverticulum; 16, intestine; 17, oviduct; 9, albumen gland; 10, uterus; 11, genital atrium; 18, nerve cord; 19, ovary; 20, eye. 162 SYSTEMATIC ZOOLOGY the molluscs and annelids makes it appear itive ciliary-mucous-muscular locomotory unlikely that the molluscs ever had coelomic mechanism that they still own today. segmentation. The argument is as follows. Actually, evolution of the molluscs pro- The locomotion of the lower molluscs, such ceeded not toward but away from this prim- as the polyplacophorans and primitive itive mechanism. Cephalopods developed a snails, is rather primitive in kind; these ani- very advanced jet-propulsion system, and mals slide over the substrate on a sheet of the bivalves evolved a mechanism which, Downloaded from https://academic.oup.com/sysbio/article-abstract/16/2/153/1647154 by guest on 20 April 2020 slime deposited by the mucous glands of similarly to that of the annelids, utilizes the the foot, by ciliary action or by locomotory the body fluids as hydraulic skeleton. The waves executed primarily by the longitud- latter will be considered more closely. The inal muscles of the sole (Elves, 1961; Clark, foot in the bivalve molluscs contains a large, 1964). The annelids have a more efficient blood-filled cavity, the contents of which, locomotory mechanism, namely, a form of when put under pressure by the contraction peristaltic motion. These animals have two of the powerful adductor muscles, cause the muscle layers beneath the epidermis: an dilation of the stretched-out foot; the latter outer layer of circular muscles and an inner is thereby anchored in the sand, and the longitudinal layer. By contracting the cir- rest of the body can be pulled to it (True- cular muscles in a certain segment or seg- man, 1960). Thus, bivalves developed a ments, pressure is exerted against the hy- way of digging, burrowing in the sand, draulic skeleton (coelomic fluid), and this which is superficially similar to that of the in turn stretches the relaxed longitudinal annelids. The differences are that, whereas muscles, leading to the elongation of the the cavity being utilized is coelomic in na- segment or segments in question. Con- ture in the annelids, it is haemocoelic in the versely, the contraction of the longitudinal molluscs; the cavity is segmented in the an- muscles and the resulting stretching of the nelids, unsegmented in the molluscs; and circular muscles will cause the shortening the antagonistic muscles are located dif- and thickening of the particular bodily ferently. The two animal groups thus have region. By stretching out a certain region, evolved, in response to similar environmen- anchoring it, and then pulling up the re- tal demands, superficially similar locomo- mainder of the body, the animal will be tory mechanisms; they represent a case of able to move. It should be emphasized that convergent evolution, rather than regres- peristaltic motion could be carried out with sion. a single, unsegmented coelomic cavity, or According to Morton (1958:26, and in even without one (simply by using the tis- litt., 1966) Garstang in prose and verse sues of the body as a hydraulic skeleton, as (1928, 1962) and conversation adumbrated in nemertines); but the division of the coe- the idea that the molluscs developed from lomic cavity into many, at least partially the annelids through neoteny; that is to say, isolated, compartments makes the system molluscs are annelid larvae which became more controllable, maneuverable, and thus sexually mature. According to Morton this more economical (Clark, 1964). can explain why the segmentation (eumetamery) of the presumed annelid ancestors The loss of such an efficient system is is not observable in molluscs: because the understandable in arthropods, which ac- post-trochophoral stage, in which eumeta- quired exoskeleton, some endoskeleton, skel- mery ensues, is omitted from the life cycle. etal muscles, and, most of all, jointed ap- Garstang's article (1928) indeed implies pendages, availing themselves of a loco- that the molluscs developed from the larvae motory mechanism superior to peristalsis. of the annelids when he says (p. 88): But it is difficult to see why the primitive ". . . the Veliger [a post-trochoporal larval molluscs should have abandoned peristaltic stage, characteristic of the molluscs] ... is locomotion and regressed to the more prim- a trochosphere [trochophore] transformed ON THE ORIGIN OF MOLLUSCS 163 by the incorporation of Molluscan charac- The similarities in cleavage are widely ters ..." (p. 88). It is also true, however, acknowledged (Nierstrasz, 1922; Naef, 1926; that this is far from proposing a theory on Hyman, 1951); typically, flatworms are said the origin of molluscs. But whether Gar- to conform to the spiral cleavage pattern, stang elaborated upon it or not, I doubt that found in molluscs and annelids. Deviations neoteny could be the means by which the from this pattern occur—e.g., having two molluscs developed from the annelids. First, blastomeres instead of four—but these are Downloaded from https://academic.oup.com/sysbio/article-abstract/16/2/153/1647154 by guest on 20 April 2020 because molluscs have, in their life cycle, a regarded as secondary modifications (Ax, genuine and sexually immature trochophore 1963: Steinbock, 1963). Further similarities stage; one who assumes that the adult stage between the flatworms and the molluscs also corresponds to a sexually mature trocho- have been noted; thus, the structure of the phore, has to explain the presence in the life nervous system, the organzation of the mus- cycle of this "supernumerary" trochophore. culature, body wall and the manner of loco- Second, to claim neoteny, one ought to have motion based thereupon, and the mode of a good agreement between the larva of a digestion (primarily intracellular) are re- certain group and the adult of another, to markably similar in the two groups (Nier- which the former supposedly gave rise. In strasz, 1922; Graham, 1955).* As Hartman some more widely accepted cases of neo- (1963) emphasized, however, similarities in teny this is actually the case; thus, the co- lower levels of organization can easily owe elenterates' planula larva is remarkably sim- to analogous solutions of the same problems ilar to the flatworm adult, and the diplopod by unrelated groups. Thus, the manner of larva to the insect adult (Hyman, 1951; locomotion, or the prevalence of intracellu- Hand, 1963; de Beer, 1946). In the annelid lar digestion are not unequivocal proofs of trochophore and the molluscan adult, how- relationship. The structure of the nervous ever, the similarity could not be less. Thus system may be a better indicator of the as- it seems safe to say that molluscs are not sumed relationships (figs. 1, 2, and 4). neotenic annelids. The flatworms are said to be on a lower We may conclude this section saying that level of organization than molluscs and an- evidence is lacking for the assumption that nelids: they lack a coelom, anus, circulatory the segmentation of molluscs (pseudome- system, and have a very simple nervous sys- tamery) arose from the segmentation of the tem and primitive locomotory mechanisms. annelids (eumetamery) through reduction. The molluscs, and particularly the annelids, The two kinds of segmentation are funda- are definitely on a higher level in most or mentally different, one cannot be reduced all of these features. Some authors argue, from the other. The explanations that the though, that the simplicity of the flatworms reduction and loss of segmentation in mol- is secondary, the result of regression from luscs was due to functional reasons or neo- coelomate animals (Ax, 1963; Remane, teny are groundless. 1963); but their theories are emphatically The origin of the molluscan-annelid rejected by many (e.g., Beklemishev, 1963). stock.—The third topic to be discussed is: Thus, we may conclude that evolution pro- where did the common molluscan-annelid ceeded from the flatworms toward the mol- stock originate? One possible answer is, luscs and annelids. from the flatworms, or, more accurately, from the ancestors of flatworms (Hammer- sten and Runnstrom, 1925; Boettger, 1959). * The fact that the annelids do not share these The evidence for this view comes from the similarities does not contradict the flatworm origin similarity in cleavage between the flat- of the molluscan-annelid stock; the annelids, with the "invention" of coelomic segmentation, have worms, molluscs and annelids, and from the radically departed from the flatworm-molluscan fact that flatworms are on a lower level of type of organzation, in structural and functional organization than are the other two groups. aspects. 164 SYSTEMATIC ZOOLOGY

The common ancestors of the molluscs bellaria, have developed from the larvae of and annelids may be reconstructed from the the Coelenterata .... [which had] a slit- features common to, or typical of, the prim- shaped mouth, an intestine with an epithe- itive molluscs and annelids, as follows. lial lining, and a pair of aboral tentacles .... They were small, probably pseudometa- by means of 'progressive' neoteny." meric, animals without a coelom (p. 156); I believe that Beklemishev has insuffi- with a ciliated body, a one-way alimentary cient ground to claim an independent origin Downloaded from https://academic.oup.com/sysbio/article-abstract/16/2/153/1647154 by guest on 20 April 2020 canal and a primarily intracellular digestion; for the flatworms and the Trochozoa. First, probably with no circulatory system; with as he himself emphasized, the protrocho- sexes separate, cleavage spiral, and trochophore is, in various respects, similar to the phore larva; living in marine, benthic habi- turbellarians (adult stage), not only differ- tats. ent from them. Second, he had to create a A different origin for the molluscan-anne- hypothetical larval form to accommodate lid stock was proposed by Beklemishev his theories; this larva, a coelenterate with (1963). He began from a comparison of an intestine, seems quite unrealistic to me. the planula stage of the coelenterates with Third, there is a huge difference between the adult stage of the flatworms (turbel- a coelenterate larva with a slit-shaped larians) and the protrochophore stage of mouth, and a trochozoan adult in which the the annelids (the latter he took as repre- gut was formed by the closure of the middle sentatives of the trochophoric animals or portion of the blastopore; it seems that Bek- Trochozoa, to which annelids, molluscs, lemishev too easily bridged this gap. We etc., belong). The first two structures are have to add that the cleavage pattern of the so well known that they need not be de- turbellarians is sufficiently similar to that scribed here. The protrochophore, accord- of the molluscs and annelids for most zool- ing to Beklemishev (p. 240), ". . . is a pro- ogists to consider them as having a common taxial larva with an aboral sense organ and origin (Hyman, 1951), and that, were Bek- a primary mouth opening on the oral pole lemishev's theory true, we would have to .... [it] has a nervous system of the ortho- accept a dual origin for bilateral symmetry gon type, situated radially around the main and cephalization, which is possible but not axis of the body. ... [In] all these respects proven. On all these grounds it seems sim- . . . [it is] similar to the Turbellaria and, pler to assume that the flatworms and the on purely promorphological grounds, to the molluscs and annelids did not evolve inde- planula of the Coelenterata .... Whereas pendently, as Beklemishev suggested, but the mouth in the Turbellaria is a small open- had a common origin, and that the differ- ing, . . . the primary mouth opening or ences Beklemishev emphasized arose after blastopore of the protrochophore forms a the trochozoan line branched off from the narrow slit .... When the blastopore flatworms; possibly the changes in pro- closes, the definitive mouth and anus are morphology themselves played a role in the formed at its ends and the physiologically separation of the two lines. ventral side is formed along the whole With this, we conclude the discussion of blastopore. The circumblastoporal nervous the origin of molluscs. Two other problems plexus gives rise to the ventral nervous seem pertinent to the subject. One is the stems of the Trochozoa .... The Turbel- origin of coelomic segmentation (eumeta- laria do not possess anything homologous mery), and the other that of the coelom. to the blastoporal side of the body or to the These are the two major "innovations" ac- longitudinal nervous stems of trochophoric quired about the time when the divergence animals." of the molluscan and annelid lines took From the above findings, Beklemishev place. concluded (loc. cit.) that the ". . . trocho- The origin of the coelom and coelomic phoric animals, independently of the Tur- segmentation in molluscs and annelids.— ON THE ORIGIN OF MOLLUSCS 165

Several theories have been put forth to ex- nephridial complex and in annelids inside plain the origin of the coelom and coelomic the mesoderm bands. Neither the gonocoel, segmentation. Since they have been dis- the enterocoel or the nephrocoel theories cussed by recent authors (Hartman, 1963; have enough flexibility to explain these dif- Hyman, 1951; Clark, 1963, 1964; Remane, ferences with ease. It is simpler to believe 1963), I will consider them very briefly. that both groups acquired the coelom inde- The gonocoel theory suggests that the coe- pendently and anew, and by their own Downloaded from https://academic.oup.com/sysbio/article-abstract/16/2/153/1647154 by guest on 20 April 2020 lom arose from the expansion of the cavity methods (this also means that their com- of the gonads; the enterocoel theory says mon ancestors did not have a coelom at that it arose from the cavities of digestive all, p. 156). diverticula, and the nephrocoel theory that Several theories have been proposed to it came about by the expansion of the flame explain the origin of coelomic segmenta- bulbs of protonephridia. The first two tion (Hyman, 1951; Clark, 1964). The theories assume that the organs giving rise gonocoel and enterocoel theories, as men- to the coelom were serially arranged (pseu- tioned above, combine the origin of seg- dometameric), and thus their transforma- mentation with that of the coelom. The tion produced not one single coelomic cav- theory of cyclomerism, which says that seg- ity, but a series. In other words, these the- mentation arose by the bilateral, serial re- ories propose to solve the problem of the arrangement of originally radially situated origin of segmentation (eumetameric) to- gastric pouches, is another form of the en- gether with that of the coelom. This may terocoel theory. Criticism of these theories appear as an advantage, but in actuality it was given previously. The fission or corm is one of the most serious objections to these theory, which says that segmented animals theories. Because, having once accepted arose from incomplete fission, i.e., fission them, one also must assume that animals without subsequent separation of the pro- with unsegmented coelom, such as molluscs, ducts of fission, has also been understand- phoronids, ectoprocts, etc., all had to de- ably rejected by Hyman. She finds that the scend from ancestors with segmented coe- origin of segmentation can best be ex- lom, which is not certain at all. Another plained by a combination of the theories of serious objection to all of the above theories pseudometamerism and locomotory mecha- is that they consider the coelom homologous nisms. According to the first, the ancestors throughout the animal kingdom, which of the segmented animals were pseudometa- seems absurd in the case of independent meric, with a series of gonads in the body phyletic lines such as the protostomes and the deuterostomes. Further objections can (note similarity to the gonocoel theory). be found in the papers cited. The conclu- The swelling of the gonads during sexual sion is that none of these theories is accept- maturity made the bending of the body dif- able. ficult except at the sites between the gonads, and this supposedly led to the serial The coelom in the molluscs and annelids constriction—segmentation—of the body. seems to be a "new" structure, not only in The second theory also presupposes a long the sense of appearing for the first time, and pseudometameric body, but it empha- but also in being additional to any one of sizes the impact of locomotion on the mus- the already existing structures, not homol- culature; it was, according to the theory, ogous to them. This mode of origin of the serpentine swimming motions that led to coelom was suggested by the schizocoel theory, upheld by Hartman (1963). Ontoge- breaking up of the musculature into seg- netic data support this contention, roughly ments. I doubt that either of these explana- as follows: In molluscs and annelids the tions is valid. First, both are decisively coelom is formed in different manners; in Lamarckian. Second, and mainly, because molluscs it develops inside the pericardial- segmentation is primarily a developmental 166 SYSTEMATIC ZOOLOGY process, it needs a developmental explana- kinds of changes were achieved in a single tion. mutation, or that they occurred simultane- Such an explanation has come forth from ously. I believe that these aspects of Ber- Berrill, concerning the origin of the segmen- rill's theory do not apply to the origin of tation of the chordate musculature. He annelid segmentation. wrote (1955:171): "Segmentation ... is primarily a phenomenon of development ACKNOWLEDGMENTS Downloaded from https://academic.oup.com/sysbio/article-abstract/16/2/153/1647154 by guest on 20 April 2020 and growth and requires analysis in terms I am most indebted to Dr. Carl W. of growth activity before we can consider Schaefer for his constant help in preparing how it may have arisen during the course of this article. I am also very grateful to Drs. evolution." And (p. 173): "Segmenta- William K. Emerson, William E. Fennel, tion ... is not something that can be grad- Robert Robertson, and Ruth D. Turner, who ually attained—it appeared suddenly as the have read and criticized the manuscript or result of a critical modification of a devel- an earlier version of it, and to Mrs. Stephani opmental process and could well have ap- Schaefer, who prepared the illustrations. peared as a single mutant type, as readily as almost any other kind of innovation or REFERENCES Ax, P. 1963. Relationships and phylogeny of mutation. The magnitude of the structural, the Turbellaria. Pp. 191-224, in The lower functional, and evolutionary consequences Metazoa (E. C. Dougherty, ed.), University of is somewhat beside the point. Admittedly California Press, Berkeley. it would be a form of macro-evolution, but BARNES, R. D. 1963. Invertebrate zoology. Saun- as the developmental outcome of just as ders, Philadelphia, 632 pp. BEKLEMISHEV, V. N. 1958a. Grundlagen der small an initial change as most others." vergleichenden Anatomie der Wirbellosen. Vol. Except for one reservation to be mentioned I. Promorphologie. Hochschulbiicher fur Biolo- below, I agree with this theory and believe gie, 6. VEB Deutscher Verlag Wissensch., Ber- that the coelomic segmentation of the anne- lin, 441 pp. BEKLEMISHEV, V. M. 1958b. On the early evo- lids has had a similar origin. This conclu- lution of the molluscs (in Russian, with English sion is derived from a comparison of the summary). Zoologicheskii Zhurnal, 37:518-522. molluscan and annelid ontogenies as fol- BEKLEMISHEV, V. N. 1963. On the relationship lows. of the Turbellaria to other groups of the animal kingdom. Pp. 234-244, in The lower Metazoa In molluscs, the growth of the mesoderm (E. C. Dougherty, ed.), University of California bands is short, differentiation sets in soon, Press, Berkeley. and the mesoderm bands will give rise to BERRILL, N. J. 1955. The origin of the verte- various mesodermal organs. Rhythmicity in brates. The Clarendon Press, Oxford, 257. pp. BERRILL, N. J. 1961. Growth, development, and growth is not evident. This condition is pattern. Freeman, San Francisco, 555 pp. taken as primitive, and the following one, BOETTGER, C. R. 1959. Comments to H. Lemche: found in the annelids, as advanced and de- Protostomian interrelationships in the light of rived. The latter arose presumably through Neopilina. Proc. XVth Interntnl. Congr. Zool., two changes from the former. First, the London, pp. 386-389. CLARK, R. B. 1963. The evolution of the celom differentiation of the teloblasts, which by and metameric segmentation. Pp. 91-107, in mitotic divisions produce the mesoderm The lower Metazoa (E. C. Dougherty, ed.), bands, is postponed; the teloblasts retain University of California Press, Berkeley. their capacity to divide, usually throughout CLARK, R. B. 1964. Dynamics in metazoan evolution. The Clarendon Press, Oxford, 313 pp. life (Berrill, 1961:312). Second, the rate of CLARKE, A. H. AND R. J. MENZIES. 1959. Neo- growth, or the rhythm of the differentiation pilina (Vema) ewingi, a second living species of of the products of growth, has changed from the Paleozoic class Monoplacophora. Science, non-rhythmic to rhythmic. It seems that 129:1026-1027. both kinds of changes can easily be ex- DALES, R. P. 1963. Annelids. Hutchinson Univ. Library, London, 200 pp. plained by mutations. It is neither neces- DE BEER, G. R. 1940. Embryos and ancestors. sary nor profitable to assume that the two The Clarendon Press, Oxford, 108 pp. ON THE ORIGIN OF MOLLUSCS 167

ELVES, M. W. 1961. The histology of the foot Geol. Congr. XXIst Session, Norden 1960, pp. of Discus rotundatus, and the locomotion of gas- 92-101. tropod molluscs. Proc. Malacol. Soc. London, LEMCHE, H. AND K. G. WINGSTRAND. 1959a. 34:346-355. The comparative anatomy of Neopilina gala- FRETTER, V. AND A. GRAHAM. 1962. British theae Lemche, 1957. prosobranch molluscs. The Ray Society, London, LEMCHE, H. AND K. G. WINGSTRAND. 1959b. 755 pp. The anatomy of Neopilina galatheae Lemche,

GARSTANG, W. 1928 (1929). The origin and 1957. Galathea Rep., 3:9-71. Downloaded from https://academic.oup.com/sysbio/article-abstract/16/2/153/1647154 by guest on 20 April 2020 evolution of larval forms. Rep. British Assoc. MCGINITIE, G. E. AND N. MCGINITIE. 1949. Adv. Sci., Sect. D., pp. 77-98. Natural history of the marine animals. McGraw- GARSTANG, W. 1962. Larval forms and other Hill, New York, 473 pp. zoological verses. Blackwell, Oxford, 76 pp. MENZIES, R. J. AND W. LAYTON, JR. 1963. A GRAHAM, A. 1955. Molluscan diets. Proc. Malac. new species of monoplacophoran molluscs, Neo- Soc. London, 31:144-159. pilina (Neopilina) veleronis from the slope of HAMMARSTEN, O. D. AND J. RUNNSTROM. 1925. the Cedros Trench, Mexico. Ann. Magaz. Nat. Zur Embryologie von Acanthochiton discrepans Hist. Ser. 13., 5:401-404. Brown. Zool. Jahrb. Abt. Anat. Ontog., 47(2): MORTON, J. E. 1958. Molluscs. Hutchinson Univ. 261-318. Library, London, 232 pp. HAND, C. 1963. The early worm: a planula. MORTON, J. E. AND C. M. YONGE. 1964. Classifi- Pp. 33-38, in The lower Metazoa (E. C. Dough- cation and structure of the Mollusca. Pp. 1-58, erty, ed.), University of California Press, Berk- in Physiology of Mollusca, Vol. I. (K. M. Wilbur eley. and C. M. Yonge, eds.), Academic Press, New HARTMAN, W. D. 1963. A critique of the en- York. terocoele theory. Pp. 55-77, in The lower Meta- NAEF, A. 1926. Studien zur generellen Morphol- zoa. (E. C. Dougherty, ed.), University of Cali- ogie der Mollusken. 3 Teil: Die typischen Be- fornia Press, Berkeley. ziehungen der Weichtierklassen untereinander HEIDER, K. 1914. Phylogenie der Wirbellosen. und das Verhaltnis ihrer Urformen zu anderen Sect. 4, part 4, 453-529, in Die Kultur der Ge- Colomaten. Erg. Fortsch. Zool., 6:27-124. genwart (P. Hinneberg, ed.), Teubner, Berlin and Leipzig. NIERSTRASZ, H. F. 1922. Die Verwandtschafts- beziehungen zwischen Mollusken und Anneliden. HUNTER, W. R. AND S. C. BROWN. 1965. Cteni- dial number in relation to size in certain chitons, Bijdr. Dierkunde, 22:33-42. with a discussion of its phyletic significance. ODHNER, N. H. 1961. Some notes on the classi- Biol. Bull. Marine Biol. Lab. Woods Hole, 128: fication of the Gastropoda. Proc. Malac. Soc. 508-521. London, 34:250-254. HYMAN, L. H. 1951. The invertebrates. Vol. II. PELSENEER, P. 1899. Recherches morphologi- The acoelomate Bilateria. Platyhelminthes and ques et phylogenetiques sur les mollusques ar- Rhynchocoela. McGraw-Hill, New York, 555 pp. chaiques. Mem. Curonnees et Mem. Savants IVANOV, P. P. 1928. Die Entwicklung der larval- fitrangers Acad. Belgique, 57:1-113. segmente bei den Anneliden. Zeitsch. Morph. PORTMAN, A. 1960. Mollusca. Pp. 1623-2164, Okol. Tiere, 10:62-161. in Traite de Zoologie (P. P. Grasse, ed.), Tome IVANOV, P. P. 1944. The primary and secondary V. Fasc. II. metamery of the body (In Russian, with English RAVEN, C. P. 1958. Morphogenesis: The anal- summary). Zhurnal Obshchei Biologii, 5:61-95. ysis of molluscan development. Pergamon Press, JOHANSSON, J. 1952. On the phylogeny of the New York, 311 pp. Mollusca. Zool. Bidr. Uppsala, 29:277-291. REMANE, A. 1963. The enterocelic origin of the LANG, A. 1896. Textbook of comparative anat- celom. Pp. 78-90, In The lower Metazoa (E. omy. Vol. II. MacMillan, London, 618 pp. C. Dougherty, ed.), University of California LEMCHE, H. 1957. A new living deep-sea mol- Press, Berkeley. lusc of the cambro-devonian class Monoplacoph- SCHMIDT, W. J. 1959. Bemerkungen zur Schal- ora. Nature, 179:413-416. enstruktur von Neopilina galatheae. Galathea LEMCHE, H. 1959a. Molluscan phylogeny in the Rep., 3:73-77. light of Neopilina. Proc. XVth Internatl. Congr. SODERSTROM, A. 1925. Die Verwandtschaftsbe- Zool. London, pp. 380-381. ziehungen der Mollusken. Uppsala. 1-54 pp. LEMCHE, H. 1959b. Protostomian interrelation- STEINBOCK:, O. 1962. tiber die Metamerie und ships in the light of Neopilina. Ibid., pp. 381- das Zolom der Neopilina galatheae Lemche, 1957. 389. Verhandl. Deutsch. Zool. Ges. Zool. Anzeiger, LEMCHE, L. 1960. A possible central place for 26 Supple., pp. 385-403. Stenothecoides Resser, 1939 and Cambridium STEINBOCK, O. 1963. Origin and affinities of the Horny, 1957 (Mollusca Monoplacophora) in in- lower Metazoa: the "aceloid" ancestry of the vertebrate phylogeny. Part XXII. Rep. Internatl. Eumetazoa. Pp. 40-54, in The lower Metazoa 168 SYSTEMATIC ZOOLOGY

(E. C. Dougherty, ed.), University of California Soc. London, ser. B, 232:443-518. Press, Berkeley. YONGE, C. M. 1957. Reflections on the Mono- TRUKMAN, E. R. 1966. Bivalve molluscs: fluid placophoran, Neopilina galatheae Lemche. Na- dynamics of burrowing. Science, 152:523-525. ture, 179:672-673. YOCHELSON, E. L. 1963. Problems of the early YONGE, C. M. 1960. General characters of Mol- history of the mollusca. Proc. XVIth Internatl. lusca. Pp. 13-136, in Treat. Invert. Paleont. Congr. Zool., Washington D. C, 1963. Vol. 2, (R. C. Moore, ed.), University of Kansas Press, Lawrence.

p. 187. Downloaded from https://academic.oup.com/sysbio/article-abstract/16/2/153/1647154 by guest on 20 April 2020 YONGE, C. M. 1947. The pallial organs in the aspidobranch Gastropoda and their evolution Department of Biology, Brooklyn College, throughout the Mollusca. Philos. Trans. Roy. Brooklyn, N. Y. 11210.