on the ciliary mechanisms and interrelationships of ...hirundo, spondylus gaederopus, anomia...

On the Ciliary Mechanisms andInterrelationships of Lamellibranchs.

PAET VII: Latero-frontal Cilia of the Gill Fila-ments and their Phylogenetic Value.

By

Daphne Atkins, B.Sc.Marine Biological Laboratory, Plymouth.

With Plate 29, and 12 Text-figures.

CONTENTS.PAGE

INTRODUCTION 346MATERIAL AND METHODS 348A. THE LATERO-FRONTAL CILIATED TRACTS OF THE

GILL FILAMENTS 350THE OCCURRENCE OF LARGE OR EU-LATERO-FRONTAL CILTA,

TOGETHER WITH SUBSIDIARY OR P R O - L A T E R O - F R O N T A L CTT.TA . 3 5 0

THE OCCURRENCE OF ONLY SMALL OR MICRO-LATERO-FRONTALCILIA 357

THE OCCURRENCE OF MODERATE-SIZED OR ANOMALOUS LATERO-FRONTAL CILIA, TOGETHER WITH SUBSIDIARY OR PARA-LATERO-FRONTAL CILIA IN ONE FAMILY ONLY, THE OSTREIDAE . . 365

LAMELLIBRANCHS WITH EU-LATERO-FRONTAL CILIA . . . 368LAMELLIBRANCHS WITH MICRO-LATERO-FRONTAL CILIA . . 376LAMELLIBRANCHS WITH ANOMALOUS LATERO-FRONTAL CILIA . 380THE ARRANGEMENT OF THE VARIOUS CILIARY TRACTS ON THE

GILL FILAMENTS 380(a) In Lamellibranchs with Eu-latero-frontal Cilia . . 380(6) In Lamellibranchs with Micro-latero-frontal Cilia . . 383

B. LATERO-FRONTAL CILIA AND PHYLOGENY . . 383DISCUSSION ON THE RELATIONSHIPS OF THE VARIOUS TYPES OF

LATERO-FRONTAL CILIA 383THE TAXONOMIO VALUE OF THE LATERO-FRONTAL CILIATED

TRACTS 388REVIEW OF PELSENEER'S AND RIDEWOOD'S CLASSIFICATIONS OF

T H E T.AATRT.T.TRBATTflTTTA . . . . . . . 388

RELATIONSHIP OF LAMELLIBRANCHS WITH MICRO-LATERO-FRONTALCILIA 398

DISCUSSION ON THE ORIGIN OF THE GROUP CHARACTERIZED BY THEPOSSESSION OF MICRO-LATERO-FRONTAL CILIA . . . 404

3 4 6 D- ATKINS

T H E E V O L U T I O N O P T H E EULAMELLTBRANCHIATE OR SYNAFTO-

RHABDIC GLLL I N T H E G R O U P CHARACTERIZED B Y T H E POSSESSION

OI1 MLCRO-LATBRO-FRONTAL CTTJA . . . . . . 4 0 6

COMMON CHARACTERS O F LAMELLTBRANCHS POSSESSING M I C R O -

LATERO-FRONTAL CTT.TA . . . . . . . 4 1 5

SUGGESTED MODIFICATIONS O F T H E CLASSIFICATION O F T H E

LAMELLIBRANCHIA 424

SUMMARY 428

R E F E R E N C E S F O R P A R T S V, VI , A N D V I I . . . 430

I N T R O D U C T I O N .

THE value of ciliary structures in the classification of theLamellibranchia must depend largely on a knowledge of theirform and of their distribution in the class of animals underconsideration, and to some extent of their function. Someciliary mechanisms, such as tracts of cirrus-like frontal cilia onthe gills are obviously adaptive, being correlated with certainhabitats and modes of life, namely sand dwelling and rock andwood boring (Atkins, 1987). The presence of fan-shaped groupsof long cilia—guarding cilia—along the marginal food groovesof the gills also seems to be adaptive, being correlated with acertain amount of mud or silt in the soil (Atkins, 1987). Suchadaptive ciliary structures may have little or no taxonomicvalue.1 Those instanced are not restricted to closely related forms,and in a family one member may possess and one lack them.

Certain tracts of cilia2 forming an essential part of the ciliary1 It is not implied that because a character is useful and adaptive it

cannot serve in classification, but that among the Lamellibranchia, whichappear to be very adaptable, adaptive characters must be used withcaution.

2 The terms frontal, latero-frontal, and lateral, which have been appliedto the three tracts of food-collecting cilia on the gill filaments (see Ridewood,1903, p. 163) are descriptive of the position these cilia occupy in the greatmajority of Lamellibranchs, but they have come to denote cilia withcertain characteristics, especially as to function, and are therefore appliedin instances where they are not descriptive of position, at least in preservedmaterial, and in some instances most probably on the living gill. Forexample, the frontal cilia may extend well round on to the lateral faces ofthe filaments, and the latero-frontal cilia may be clearly lateral in positionas in Nucu lana m i n u t a (Text-fig. 8) and Tr igonia marga r i -taoea (Text-fig. 1 c).

THE CILIARY MECHANISMS OF LAMELLIBRANCHS 847

feeding complex, though independent of habit and habitat, areso universally present on the gills, with little or no modificationof their form, that these again are useless in classification. Suchare the lateral cilia of the gills, which create the inhalent current,found in all Protobranchia, Filibranchia, Pseudolamellibranchia,Eulamellibranchia, and even on the few filaments forming thebranchial sieve in P o r o m y a among Septibranchia, as wellas in certain aspidobranch and pectinibranch Gastropods:analogous tracts of cilia occur in several groups of ciliary feedersoutside the Mollusca. The mere presence of such cilia thereforehas no classificatory value, but the arrangement or pattern ofthe cells bearing them may be of generic, and family, andpossibly of wider, value; for instance, the same type of cell-pattern is characteristic of a large number of eulamellibranchfamilies (Atkins, 1938 a).

The frontal, food conveying, cilia of the gills are also anessential element in the feeding complex, and too widely presentto be of value in classification. Modification of these cilia seemsto be largely in relation to habitat (Atkins, 1937). But themodification in which adjacent antagonistic tracts of frontalcilia on the same gill filament occur on a l l lamellae, is foundchiefly in one group of Lamellibranchs, 'the Aviculidae andtheir allies', but is not entirely restricted to this group—it isfound also in the Solenidae (Atkins, 1936)—and is not found inall its members, for instance not in P i n n a (Atkins, 1937).

The latero-frontal cilia of the gills form an additional partof the ciliary feeding complex, but, apparently, are not absolutelyessential, as are the lateral (current producing) and frontal (foodconveying) cilia, for they have been described only in Lamelli-branchs, not in Gastropods; nor have analogous cilia beendescribed in ciliary feeders outside the Mollusca. They arestraining cilia, preventing the unimpeded passage of particlesbetween the gill filaments to the supra-branchial chamber. Theythus contribute to the efficiency of the method of feeding, pre-venting the escape and wastage of perhaps desirable foodparticles, but would not seem to be indispensable.

Two chief types of latero-frontal tracts have been found toexist, the one efficient, composed of a row of large, with also

848 D. ATKINS

a second row of small latero-frontal cilia, on each side of thefrontal surface; the other apparently rather inefficient, composedof one row only of tenuous and small latero-frontal cilia on eachside. The types are unconnected with habits or habitats. Thedistribution of the second type among Lamellibranchs is notirregular, but is distinctive of a particular group. It may seemridiculous to consider that the composition of a certain tractof gill cilia can be of value in determining genetic affinity, butthe fact remains, that without any preconceived ideas, andsetting out merely with the intention of spending odd minutesnoting what proportion of Lamellibranchs had large latero-frontal cilia, and what proportion had small ones—or, as wasat first thought, were without them—it was gradually realizedthat the occurrence of only small latero-frontal cilia is character-istic of members of a group ' the Aviculidae and their allies' orthe ' sedentary' branch of Lamellibranchs established by palae-ontologists (Jackson, 1890; Douville, 1912 a) mainly, or entirely,on shell characters. One desideratum of a character for use indetermining phylogeny, namely conservatism, seems to bepossessed by the latero-frontal ciliated tracts, for, so far as thepresent work has gone, there appears to be little variation withinthe two types, and they therefore afford an important andreliable character for purposes of broad classification.

In addition to the acknowledgements already made in Part Iof the series, I wish to thank Professor A. Morley Davies forcriticizing the present paper from the palaeontological view-point, and to renew my thanks to Professor J. H. Orton formost helpful general criticism of this paper in particular. Pro-fessor E. S. Goodrich has not only edited it, but has madesuggestions for which I am much indebted to him.

MATEEIAL AND METHODS.

The great bulk of the material used for this, as for the otherpapers of the series, was brought in by 8. 8. Salpa of the MarineBiological Association of Plymouth. Living specimens ofN u c u l a n a m i n u t a and M u s c u l u s d i s c o r s were ob-tained from Millport Marine Station by purchase, as were alsoliving specimens of P e t r i c o l a p h o l a d i f o r m i s from Whit-

THE CILIARY MECHANISMS OF LAMELLIBBANCHS 349

stable, and Bouin-Duboscq-preserved specimens of P t e r i ah i rundo, Spondylus gaederopus, Anomia ephip-pium, Solenomya toga ta , and Nuculana pella fromthe Zoological Station of Naples.

My thanks are due to a number of zoologists who have mostkindly furnished me with valuable material: to Dr. B. Prashadof the Indian Museum, Calcutta, for alcohol-preserved specimensof Amussium pleuronectes , P inc tada vulgar is ,P inc tada margar i t i fe ra , Malleus a lbus , Isogno-mon i sognomon var. canina, and Area (Scaphula)celox; to Mr. G. C. Eobson of the British Museum for frag-ments of the gills of Trigonia margar i t acea , P lacunaplacenta (?), Vulsella sp., Isognomon a l a t a , Ae-ther ia e l l ip t ica , and Miilleria dal l i ; to Mr. A. G.Lowndes for living specimens of Anodonta ana t ina andSphaerium corneum; to Mr. C. Oldham for living Dreiss-ensia polymorpha; to Mr. G. A. Steven for a formalin-pre-served specimen of Ch lamys v i t r e a ; and to ProfessorC. M. Yongefor Bouin-preserved specimens of Spondylus sp.from the Great Barrier Eeef.

I am indebted to Professor J. H. Orton for the loan of Sir W. A.Herdman's slides of the gills of P inc tada vulgar is andPlacuna placenta from Liverpool University, and toMr. H. H. Bloomer and Mr. G. C. Eobson for the loan of slidesof the gills of Mutela bourguigna t i .

The form of the latero-frontal cilia of the gills is discoverablemost rapidly and reliably from the examination of livingmaterial, and this was done whenever possible. When preservedmaterial only was available, entire filaments were examinedunstained in alcohol, or formalin, with a drop or two of glycerineadded, and the results obtained verified by the examination ofstained sections.

Material was fixed in Bouin-Duboscq's fluid (see Atkins, 1937 b,p. 424). Sections were cut 5/u, or &{i thick. The stains chieflyemployed were Heidenhain's iron haematoxylin, either aloneor counter stained with acid fuchsin, and Mallory's triple stain.

All drawings, except Text-figure 2 c, were made with the aidof a camera lucida.

850 D. ATKINS

A. THE LATEBO-FKONTAL CILIATED TEACTSOF THE GILL FILAMENTS.

THE OCCUBRENCE OF LARGE OR EU-LATEEO-FEONTAL CILIA,

TOGETHEB WITH SlJBSIDIAEY OB P E O - L A T E R O - F B O N T A L ClLIA.

It has been known for a number of years that some Lamelli-branchs have large latero-frontal cilia (eu-l-f.c, Text-fig. 1), orcilia of the corner cells, on the gills (Peck, 1877; Engelmann,1880; Janssens, 1893, &c), though some of the earlier workersdid not discern the direction of their effective beat or discovertheir function. These cilia are large, cirrus-like, and impossibleto overlook, and, as described by Orton (1914), have 'theappearance of flexible combs working along the sides of thefilaments'. They are present on each side of the tract of frontalcilia and 'stand out from the sides of the filaments, forminga sort of grating between them, and lash relatively slowly acrossthe length and towards the middle of the frontal face of thefilament' (Orton, 1912). The latero-frontal cilia are strainingcilia, and throw particles on to the frontal face of the filamentwhence they are transported by the frontal cilia.

The structure and movement of these large cilia have beeninvestigated by numerous workers (Engelmann, 1880; Janssens,1893; Wallengren, 1905, I ; Gray, 1922, 1928; Carter, 1924;Grave and Schmitt, 1925; Bhatia, 1926).

Carter's (1924) account is the most detailed. He investigated

Transverse sections of filaments of marine Lamellibranchs havingeu-latero-frontal cilia. A, Modiolus modio lus ; B, Kel l iasuborb icu la r i s ; c, Tr igonia m a r g a r i t a c e a . Fig. l cwas composed from several sections, abf.c, abfrontal cilia; c.d.,ciliated disc; c./.c, coarse frontal cilia; c.r., calcified rods;eu-l-f.c, eu-latero-frontal cilium; /.a, frontal cilia; f.c.l., longfrontal cilia borne on three or four rows of long cells; f.f.c, finefrontal cilia; I.e., lateral cilia; tn.g., mucous gland; pro-l-f.c,pro-latero-frontal cilium; t.m., transverse muscle-fibre. Thechitinous skeleton is shown in black, except in Tr igoniam a r g a r i t a c e a where'it is shaded and the calcified rods shownin black, A and B Bouin-Duboscq's fixative; c, alcohol fixation;A-O, iron haematoxylin and acid fuchsin. X 735. For this, andother figures of gill filaments, sections have been chosen whichwere as free as possible of mucous glands, as these interfere withthe orderly arrangement of the ciliated cells.

o-L-fic.

-L-fic.

[For description see previous page.}

352 D. ATKINS

these large, complex cilia by means of the micro-dissectionneedle and found that though they appear homogeneous whenliving, they are composed of a series of triangular plates (10 to15 in Myt i lu s ga l lop rov inc i a l i s ) set one behind theother in the plane of the beat, with the shorter in front (Text-fig.2 c, a, b, p. 855), that is on the side which is directed forward inthe effective stroke (Text-fig. 2 c, c). These plates together forma blade-shaped compound cilium with the flat side of the bladein the plane of the beat (Text-fig. 2 c). When the cilium dies,or on fixation, the plates forming the cilium break up, and thereis left a double row of fibres (individual cilia) formed by theiredges. Below each fibre lies a basal granule; there is thereforea double row of granules below each latero-frontal cilium.

A ciliary structure of this complexity cannot rightly be calleda cilium, but as the term latero-frontal cilium is well establishedin the literature it is retained here.

Gray (1928, p. 18) has described how the latero-frontal ciliumis perfectly straight when at rest and that 'movement occursby a flexure which begins at the tip and passes down to the basethereby bending the cilium into a hook-shaped structure', while' during recovery the process is reversed for the cilium straightensfrom the base to the tip' (Text-fig. 2 c, c) .

In Lamellibranchs with large latero-frontal cilia there is, inaddition, along the frontal side of these, a row of cilia which maybe considered either as specialized frontals, or, perhaps morecorrectly, as a second row of small or subsidiary latero-frontalcilia (pro-l-f.c, Text-figs. 1, 2A). They are difficult to observe,being over-shadowed by the large ones, but it has been deter-mined that they are in a single row, and that the beat is towardthe frontal surface—as is that of the latero-frontal cilia—thoughsomewhat obliquely in the direction of beat of the frontal cilia.They are more closely set than the large cilia, and are borne anumber on a cell. The cells which bear them are narrow andelongated in the direction of length of the filament. In Ens iss i l iqua one cell covers the same length as do about eight largelatero-frontal cells (Text-fig. 2 B) ; in A n o d o n t a as six or seven,and in Cyclas cornea ( = S p h a e r i u m corneum) asabout ten (see Engelmann, 1880, PL V, figs, 2,4). The number of

THE CILIARY MECHANISMS OF LAMELLIBEANCHS 353

these subsidiary latero-frontal cilia to a cell could not be deter-mined, but, as they are more closely set than the large ones,there would be more than eight, six or seven, and ten respec-tively to a cell, and possibly double or more than these numbers,in Ens is s i l i qua , A n o d o n t a , and Sphae r iumcorneum.

The presence of these cilia was first observed in transversesections of the gill filaments of Modiolus modio lus (Text-fig. 1 A), when the appearance was assumed to be due to slightobliquity of the sections, so that part of a second large latero-frontal cell was cut. The appearance was so consistently present,however, that the correctness of this interpretation seemedhighly doubtful. Careful examination of living filaments re-vealed the presence of a row of cilia, stouter than the frontalcilia, with certain of the characteristics of the latero-frontalcilia. These cilia have been found generally in bivalves witheu-latero-frental cilia, including Myt i lus edu l i s . They aremore clearly discernible in some gills than in others, accordinglargely to the transparency of the gill. For instance they wereseen more clearly in the living gill of Kel l ia s u b o r b i c u l a r i sthan perhaps in any other, except T h y a s i r a f l exuosa ,though they are not particularly clear in sections of that gill(Text-fig. 1B) . In Lamellibranchs in which the outer demibranchis without a marginal groove, and the filaments are merely bentround, the various ciliary tracts are seen in profile at the bend,and the subsidiary latero-frontal cilia may occasionally beclearly seen. They are extraordinarily clear along the free edgeof the outer demibranch of T h y a s i r a f lexuosa (Text-fig.2 A, pro-l-f.c). These cilia are distinguishable from the frontalcilia in successfully stained transverse sections of well-fixedmaterial (fixed in Bouin-Duboscq's fluid: stained Heidenhain'siron haematoxylin) because they are stouter and stain moredarkly, and by their darker staining ciliary rootlets. The nucleiof the cells are narrow and elongated in the direction of thelength of the filament, and in cross-section appear small. Theyhave no greater affinity for basic dyes than have the oval nucleiof the frontal cells.

So far as I am aware the only reference to these ciliated cellsNO. 319 A a

354 D. ATKINS

is in a footnote in Engelmann's paper of 1880 (p. 511)—of whichI was ignorant until after their discovery—in which he stated:

pro-l-fx.

tu-l-fx

.c.

p/o-l-fxc.

eu-l-Fc.'c.TEXT-EIG. 2 A and B.

A, Thyas i ra f lexuosa . Two living filaments in optical sectionat the free edge of the outer demibranch, to show the pro-latero-frontal cilia (pro-l-f.c). The frontal cilia (/.c.) were beating towardthe observer, and therefore are not seen at their full length.eu-l-f.c, eu-latero-frontal cilium. I.e., lateral cilia. X537J. B,Ensis s i l iqua . Sketch to show the shape of the pro-latero-frontal cells (pro-l-f.ee). eu-l-f.c.c, eu-latero-frontal cell; f.c,frontal cilia; f.c.n., nucleus of frontal ciliated cell. The doublerows of basal granules of the eu-latero-frontal cilia are shown.Bouin-Duboscq's fixative: iron haematoxylin and acid fuchsin.X 918J. See p. 355 for Text-fig. 2 o.

' Zwischen den Eckzellen und den gewohnlichen Flimmerzellendes Kuckens ist noch eine, bisher wie es scheint ubersehene,Lage sehr schmaler, in der Langsrichtung der Leistchen lang-gestreckter Flimmerzellen eingeschaltet. Sie mogen Neben-


zellen heissen. Ihr eigenthiimlicher Bau stimmt wesentlich mitdem der " Seitenzellen " iiberein. Speciell ist die Anordung undEinpflanzung der Cilien auf der Oberflache bei beiden die nam-

TBXT-FIG. 2 c.

Diagrams illustrating the structure and movement of the compositeeu-latero-frontal cilia, represented as being each composed of fiveplates only, (a) Frontal view of three cilia: plates are indicatedin the middle one. (6) Oblique view of one cilium. (c) Ciliumin side view to show the type of movement (after Gray, fig. 12 b,1928, somewhat modified). The cilium is straight when at rest.The plain arrows indicate the direction of the effective stroke;the broken arrows the direction of the recovery stroke.

liche.' He figured these cells in A n o d o n t a and Cyclascornea .

No observations on the arrangement of the basal granules ofthese cells have been made for the present work, but from theliving gill the cilia appear to be in a single row, and to resemblethe large latero-frontal cilia in this, and the direction of theeffective beat, position of rest and type of metachronal wave,rather than the lateral cilia, which they were said to do by

356 D. ATKINS

Bngelmann. The subsidiary latero-frontal and the lateral ciliado agree, however, in being borne on narrow cells, elongated inthe direction of length of the filament. The subsidiary latero-frontal cilia probably act as a second sieve, preventing the escapeof small particles between the bases of the eu-latero-frontalcilia.

It is possible that Janssens (1893, p. 66) saw something ofthese cells in sections, for describing the latero-frontal cells ofA n o d o n t a a n a t i n a he wrote: 'Sur les coupes, on croitsouvent avoir sous les yeux la section de deux rangees decellules, Fig. 15 en bas, mais c'est une apparence produite parl'obliquite de la section ou d'autres causes.'

Thus many Lamellibranchs may be considered as possessingtwo rows of latero-frontal cilia on each side of the frontal sur-face, a row of very large and another of small or subsidiarylatero-frontal cilia. The large ones may be termed e u - l a t e r o -f r o n t a l c i l i a , and the subsidiary ones p r o - l a t e r o -f r o n t a l c i l i a .

I n N u c u l a n a and N u c u 1 a the nuclei in a row on the frontalside of the eu-latero-frontal cells are long and narrow—in thedirection of length of the leaflets—indicating the presence ofa row of long, narrow cells, such as bear the pro-latero-frontalcilia in higher Lamellibranchs, but it is extremely difficult toidentify with certainty pro-latero-frontal cilia on the livingleaflets of Protobranchs, owing to the thickness of the leaflets.The closeness to one another of the eu-latero-frontal cilia (about1 //. or less apart) and in N u c u l a n a the position of the latero-frontal tracts, well down on the lateral faces of the filaments(see Text-fig. 8, p. 873), add to the difficulty. A renewed andcareful attempt at the definite identification of pro-latero-frontal cilia in N u c u l a n u c l e u s was made in April 1937,and it was finally concluded that there is a regular row of ciliaon the frontal side of the eu-latero-frontal cilia, correspondingto the pro-latero-frontal cilia of the higher groups, and thatthese cilia are most probably pro-latero-frontals. Their positionof rest, however, could not be observed, and this was unfortunate,as latero-frontal cilia of whatever kind, and frontal cilia, differin their attitude when brought to rest, latero-frontal cilia being

THE CILIARY MECHANISMS OF LAMELLIBBANCHS 357

straight, while frontal cilia are curved. Pro-latero-frontal ciliahave not been labelled in Text-figs. 7 (p. 372) and 8 (p. 373),which are reproductions from Part I (1936), but may be dis-cerned in the sections of Nucula and Nuculana, thoughnot in those of Solenomya of which the fixation was notparticularly good. In Nucula a pro-latero-frontal cell coversabout the same length as do four or five eu-latero-frontal cells.

THE OCCURRENCE OP ONLY SMALL OR MICRO-LATERO-FRONTAL CILIA.

Eu-latero-frontal cilia, such as described in the precedingpages, are absent in certain families of Lamellibranchs, whichare characterized by the possession of small and tenuous latero-frontal cilia. Small cilia of this kind may be termed micro -la tero- f ronta l cilia (mi-l-f.c, Text-fig. 3A-E). Suchfamilies have been generally considered as lacking latero-frontalcilia, with the exception of the Anomiidae in which they wereobserved by Orton (1914), though in his paper he does notdistinguish between their size in this family and in Mytilus,for instance. In his note-book of 1912—which he kindly allowedme to see—he noted, however, that 'the straining cilia ofAnomia aculeata Miiller (=He te ranomia squamula(L.)) are extremely fine'.

Latero-frontal cilia have generally been considered absent inthe following bivalves: Glycymeris and Area (Orton,1914); Pecten (Kellogg, 1892; Janssens, 1893; Dakin, 1909;Orton, 1914; Setna, 1930; Gutsell, 1931); Lima (Studnitz,1931); Pinna (Atrina) rigida (Grave, 1911) and Pinc-tada ( = Margaritifera) vulgaris (Herdman, 1905).In the work of several of these authors, however, the absenceof latero-frontal cilia is inferred, as they are not mentioned, ornot shown in the figures, but Janssens—who gave beautiful anddetailed figures of the gill structure of a number of Lamelli-branchs—definitely stated that 'dans les diverses especes dePecten que nous avons eues a l'etude, nous n'avons jamaispu decouvrir les cellules des coins', and Gutsell that' the elongatelatero-frontal cilia described for Myti lus , Ostrea, and

358 D. ATKINS

fie.. \._mi-L-fic. f.c.--

TEXT-FIG. 3.

Transverse sections of filamenta of Lamellibranchs having micro-latero-frontal cilia, and of Ost rea edul is having anomaloustogether with para-latero-frontal cilia, A, Area (Scaphula)celox; B, Malleus a lbus ; c, Pec ten maximus (one ofthe apical filaments); D, Spondylus gaederopus (one ofthe two apical filaments); B, Lima hians (ordinary filament,third from principal); F, Ost rea edul is (ordinary filament,fifth from prinicpal). abf.c, abfrontal cilia; an-l-f.c, anomalouslatero-frontal cilium; c.d., ciliated disc; c.f.c, coarse frontal cilia;

[For remaining description see opposite.]

THE CILIAKY MECHANISMS OF LAMELLIBRANCHS 359

various lamellibranehs have not been found in the scallop(Pecten irradians), nor have I succeeded in demonstratingthat the most lateral of the frontal cilia (in a latero-frontalposition) function as would typical latero-frontal cilia'.

Curiously enough Pelseneer (1891) depicted large latero-frontal cilia in his figures of Anomia ephippium, Pectun-culus (=61ycymeris) glycymeris , Area ba rba t a ,Peeten opercular is , and Lima hians , and indeedshowed them as large as in a figure of Modiolaria (=Mus-culus) marmora tus for instance. Hornell (1909) alsofigured and described large latero-frontal cilia in Placunaplacenta . In Sir W. A. Herdman's slides of Placuna Ifind the latero-frontal cilia to be small as in Anomia (Atkins,1936). Bourne (1907) did not show or mention latero-frontalcilia or cells in Anomia (Aenigma) aenigmat ica .

In the living gill the micro-latero-frontal cilia when more orless motionless appear as a fine palisade viewed from the frontalsurface. In Glycymeris glycymeris and Area t e t r a -gona they are about 14 to 17/x long; in Heteranomia andMonia about 12jn long. Measurements of the width andbreadth at the base were not made, but the cilia are very slender.In lateral view of the filament they appear as a row of shiningdots, no doubt owing to bending during the stroke.

Certain details of the form of the micro-latero-frontal cilia,and the cells bearing them, have been gathered from entirefilaments and from sections, which were not, however, cut forthis purpose.

In the Arcidae it has been found that the cells (mi-l-f.c.c,Text-fig. 4) bearing the micro-latero-frontal cilia are narrow,elongated in the direction of length of the filament, and withlong narrow nuclei (mi-l-f.c.n.). In Glycymeris glycy-

/.c, frontal cilia; f.f.c, fine frontal cilia; I.e., lateral cilia; m.g.,mucous gland; mi-l-f.c, micro-latero-frontal cilium; n., nerve;p.c, pigment cells; p.ch., pale-staining chitin; para-l-f.c, para-latero-frontal cilium; t.m., transverse muscle-fibre. The chitinousskeleton is shown in black. A, B, alcohol fixation; c-F, Bouin-Duboscq's fixative; A-F, iron haematoxylin and acid fuchsin.X735. Transverse sections of gill filaments of the Arcidae,Anomiidae, and Pteriidae have already been given (Atkins, 1936).

860 D. ATKINS

mi.-l.-ftcc.

l.c.n i-L.—Pen.

TEXT-ITO. 4.

Glycymeris g lyoymer i s . Lateral surface of part of a filamentto show the shape of the micro-latero-frontal cells (mi-l-f.c.c).The lateral ciliated cell rows (l.c.c.) are also shown, though indi-vidual cells are not indicated. As the lateral ciliated cells areconsiderably larger basally than peripherally, it will be observedthat the rows of nuclei do not lie directly beneath the rows ofcells—as these appear at the surface—to which they belong, f.c,frontal cilia; /.cm., nuclei of frontal ciliated cells; l.c.n., nucleiof lateral ciliated cells; m.g., mucous gland; rni-l-f.c.n., nucleus ofmicro-latero-frontal cell. Two per cent, osmic acid only, x 980.

meris there are roughly about eighteen cilia to a cell: ciliaryrootlets were not seen in longitudinal sections of the filamentsin this species, nor in Area t e t ragona , but were seen intransverse sections.

In Pecten the number of micro-latero-frontal cilia to a cellcould not be determined.


In P i n n a f r ag i l i s there are four small latero-frontal ciliato a cell (Text-fig. 5 A) : these will be termed provisionally miero-latero-frontal cilia, though there is some doubt as to theirhomology, owing to the possibility of there being a second rowof latero-frontal cilia present (see p. 363). Each cilium, or more

TEXT-FIG. 5.

A, P inna f ragi l i s . Section, nearly sagittal, through the latero-frontal epithelium. Bouin-Dubosoq's fixative; iron haematoxylin.X 1,470. B, Myt i lus edul i s . Sagittal section through thelatero-frontal epithelium. Fleming without acetic, iron haema-toxylin, counterstained with orange G and silver nitrate. -^ objec-tive and No. 18 (Zeiss) eye-piece. B, after Bhatia, 1926.

probably each of the fibres of the cilium, has two rootlets, orrhizoplasts, which diverge widely from a basal granule to passon opposite sides of the nucleus. The rootlets of the four ciliacross one another, except the outer ones of the two outer cilia.A similar disposition of the rootlets is found in the cells of eu-latero-frontal cilia, for example in M y t i l u s , where the tworootlets of the fibres forming the sides of a triangular plate ofthe large complex cilium pass on opposite sides of the nucleus,the two inner ones crossing each other (Text-fig. 5 B). Accordingto Lucas (1931) the ciliary rootlets extend to the nuclear zoneonly, and not to the base of the cell, as described by Bhatia(1926). Apparently two of the micro-latero-frontal cilia ofP i n n a f r ag i l i s correspond to a single large complex latero-frontal cilium, of M y t i l u s for instance. In P i n n a f r a g i l i s ,however, the four cilia of each cell are entirely separate, and not

362 D. ATKINS

connected, as are the two fibres forming the sides of a triangularplate of the large complex cilium of the majority of Lamelli-branchs. It is not impossible, though I think it unlikely, thateach cilium of P i n n a may consist of two rows of fibres sideby side as in eu-latero-frontal cilia, but so close together thatat a magnification of 1,470 diameters two sets of ciliary rootletscould not be distinguished.

The anomalous latero-frontal cilia of O s t r e a (see p. 365)could be derived from micro-latero-frontal cilia as seen inP i n n a by division of the cell; in such a hypothetical divisiontwo sets of basal granules with their rootlets would pass intoone cell, and the other two sets of basal granules with theirrootlets into the other. The two cilia of each cell might thenapproach each other and finally their fibres become connectedby membranes. Possibly eu-latero-frontal cilia may have ariseneither in this way from cilia of the micro-latero-frontal type, oras fully formed eu-latero-frontals.

The intra-cellular fibre system of epithelial cells may servea function of co-ordination, as suggested by Worley (1934), butit is noteworthy that there is a similar complicated arrangementof the ciliary rootlets of the eu-latero-frontal cilia (e.g. in My-t i 1 u s) where the ciliary fibres linked up by these rootlets mustbeat synchronously—for they form the edges of a triangularelement—and of the micro-latero-frontal cilia of P i n n a , wherethe four separate cilia of a cell linked up in a similar mannerprobably beat metachronously. Unfortunately the beating andmetachronism of the latero-frontal cilia of P i n n a were notespecially noticed owing to the difficulty of observing them inthis form. If in P i n n a the cilia on individual cells should befound to beat metachronously, then it would seem possible thatthe arrangement of ciliary rootlets may not be related to thetype of unicellular co-ordination. Under the influence of certaindrugs the metachronous beating of the cilia of individual epithelialcells may change to synchronous, but this is attributed to theharmful effect of the drugs on the regulatory mechanism, thecilia becoming uncontrolled and usually beating simultaneouslydue to their mechanical effect upon each other (Worley, 1934).

Although cilia with a structure closely comparable to that of

THE CILIARY MECHANISMS OF LAMBLLIBEANCHS 363

eu-latero-frontal cilia are known, for instance the velar cilia ofthe Nudibranch veliger, where each cilium consists of aboutfifteen plates and there are two, three, four, and sometimes moresuch complex cilia to a cell (Carter, 1926), and the very longabfrontal cilia or cirri of My t i l u s , each consisting of four orfive plates in M y t i l u s g a l l o p r o v i n c i a l i s (Carter, 1924),it is not known whether they have a similar arrangement of theciliary rootlets, or if the arrangement described above is peculiarto cilia with a straining function. Carter (1928) held that aciliary rootlet is not a definitely differentiated portion of thecytoplasm, and doubted the existence of such fibres in thelatero-frontal cells of M y t i l u s and in those of the velar cellsof Nudibranch veligers.

In thin transverse sections of the filaments of P i n n af rag i l i s there is occasionally an appearance as of two latero-frontal cilia on each side of the filament. This may be due toslight obliquity of the sections. In most bivalves an examina-tion of the living gill is the best way in which to settle such aquestion, but the living gill of P i n n a is extremely difficultto deal with owing to its deep plications, organic junctions, andabove all, extreme sensitiveness, the slightest stimulation caus-ing it to contract to a contorted mass of filaments. If two rowsshould prove to be present, then the arrangement of the ciliaryrootlets in the latero-frontal cells of P i n n a as compared withthat in the Ostreidae (see p. 365) is all the more interesting. Itis possible, however, that diverging ciliary rootlets from anentire cilium, passing on opposite sides of the nucleus as inP i n n a , or from a ciliary fibre of a triangular element as inM y t i l u s , may be linked up in some way with a special typeof behaviour, and so are likely to be found in all cells the cilia ofwhich have this characteristic. It is therefore perhaps notimpossible that all types of straining cilia will be found to havediverging rootlets.

In Ma l l eus a l b u s each latero-frontal cilium has a shortrow of basal granules, rather less than 1 p long, in the plane ofthe beat; this indicates that the cilia are slightly blade-shaped,that is, they are wider in the plane of the beat than in the planeperpendicular to it. The material from which the sections were

364 B. ATKINS

cut had been preserved in alcohol, and probably because of theresulting poor fixation the ciliary rootlets could not be dis-tinguished in longitudinal sections through the latero-frontalepithelium. There appeared to be about four cilia to a cell,but this could not be definitely determined. The width of thesecilia at the base—unusual for micro-latero-frontals—suggeststhat there is just a possibility, as in P i n n a f r ag i l i s , of tworows of latero-frontal cilia, though a second row was not dis-cernible in sections, and in the state of preservation this washardly to be expected.

The latero-frontal cilia themselves are less close together inP i n n a f ragi l i s and Mal leus a lbus than in Glycy-mer i s and Area , and are fewer to a cell, and somewhatlarger. Owing to the possibility of there being two rows oflatero-frontal cilia in P i n n a f rag i l i s and Mal leus a lbusdetails of the structure of the latero-frontal cilia in these twospecies have been omitted from Table I, p. 370.

Lucas (1931) noted for Myt i lu s eclulis and Amblemac o s t a t a that in fixed and stained preparations the nucleusof the eu-latero-frontal cells has a stronger affinity for basicdyes than have the nuclei of other cells of the gill. This was foundto be so in most of the gills sectioned for the present work, butnot invariably. In most bivalves with micro-latero-frontal cellsthere seems to be little, if any, difference in affinity for basicdyes between the nuclei of these cells and others, but the Indianfresh- and brackish-water form, Area (Scaphula) ce lox ,is an apparent exception in which the nuclei in cross-section arenarrow and stain darkly (Text-fig. 3 A). The micro-latero-frontal cilia in this species of Area (museum material preservedin alcohol) showed more clearly than in well-fixed material ofother forms.

Lamellibranchs with micro-latero-frontal cilia, with thepossible exception of P i n n a , appear to be without a subsidiaryrow, such as is present in bivalves with eu- and in those withanomalous latero-frontal cilia.

Micro-latero-frontal cilia have been found to possess certainof the characteristics of eu-latero-frontal cilia, namely:

(1) their arrangement in a single, regular row;


(2) the direction of the effective beat, which is toward thefrontal surface;

(3) their position of rest;(4) the same type of metachronal wave. The metachronal

wave was especially noted in the Arcidae, Anomiidae,and Pectinidae.

They differ from them, however, in their (a) smaller size,especially in the width and breadth at the base; (b) apparentlysimpler structure, being probably composed of fibres and notof triangular plates; and (c) in that more than one is borne ona cell, while there is only a single eu-latero-frontal ciliumto a cell.

Micro-latero-frontal cilia agree closely with the pro-latero-frontal cilia of bivalves possessing eu-latero-frontal cilia (seeTable I, p. 370): they may be actively straining cilia, or mayperhaps act merely as a guard to prevent loss of food particlesfrom the frontal tract.

THE OCCURRENCE OF MODERATE-SIZED OR ANOMALOUS LATERO-

FRONTAL CILIA, TOGETHER WITH SUBSIDIARY OR PARA-

LATERO-FRONTAL CLLIA IN ONE FAMILY ONLY, THE OSTRE-

IDAE.

In one family, the Ostreidae, the latero-frontal cilia of themain row are only moderately developed, especially as to theirsize at the base, and may be termed a n o m a l o u s l a t e r o -f r o n t a l c i l i a {an-l-j.c, Text-fig. 3 F, p. 358). They are how-ever about 14 to 25 p. long.

The cells from which these cilia arise are small and bear onecilium each, as may be seen when the cells are becoming dis-sociated under the influence of the narcotic stovaine. Thesecilia appear to have the same structure, and the same arrange-ment of their ciliary rootlets as in M y t i l u s , but, while beingblade-shaped, they are much less wide at the base in the pianoof the beat (roughly about \\yi), than those of M y t i l u sg a l l o p r o v i n c i a l i s (4-5/z, see Carter, 1924) and other bi-valves having eu-latero-frontal cilia (see eu-l-f.c, Text-fig. 1,p. 351, and an-l-j.c, Text-fig. 3 F). The effective beat is in thesame direction and the position of rest is the same as that of the

366 D. ATKINS

other forms of latero-frontal cilia, and the metachronal wave isof the same type.

The eu-latero-frontal cilia of bivalves are characteristically,closely, and evenly spaced; those of Ensis siliqua forexample are about 2/u apart, those of Nucula nucleus about1/x apart. In Ostrea edulis , Ostrea virginica, andOstrea angulata it has been found that the length of theanomalous latero-frontal cilia and their distance apart varieson different filaments. They are longest and closest togetheron the principal and transitional filaments, and shortest andfarthest apart on the filaments forming the plical crests. Fre-quently they are closer together on that side of the transitionalfilament next to the principal than on the side away from it.In Ostrea edulis and Ostrea virginica the distanceapart of these cilia varies between about 1-5 to 3-7\L, whileon the filaments of the plical crests of Ostrea angula tathey are about 6 or 7/x apart.1 In the genus Ostrea the frontalcurrents on the principal and transitional filaments are dorsalin direction, and particles intended for consumption are mainlycarried along these. That the latero-frontal cilia, which arestraining cilia preventing particles from being swept through tothe exhalent chamber, should be longer and closer together onthese filaments than on those forming the plical crests, which areconcerned chiefly with rejection (Atkins, 1937), is indicative offunctional correlation.

While the anomalous latero-frontal cilia of Ostrea are easilyseen in fresh material, they are difficult to distinguish in sections

1 An interesting instance of wider spacing of eu-latero-frontal cilia onone side of a filament than on the other was observed over stretches of thefilaments of Pe t r i co l a pho lad i fo rmis , due to the presence of aprotozoon 35 to 80/a, long, adhering by the whole length of its body parallelto the filament in the region between the lateral and latero-frontal cilia,and suppressing some of the latero-frontal cilia, so that they were abouttwice as widely spaced as normally; while in some places they were en-tirely wanting for short stretches. The parasites, which were very numerous,were frequently attached along one side only of a filament, the oppositeside where the latero-frontal cilia were normally spaced being free fromthem. This parasite though allied to the Ciliata appeared, in the stageseen, to be without cilia.

THE CILIAEY MECHANISMS OF LAMBLLIBRANCHS 367

(see also Yonge, 1926): this is no doubt owing to the fact thatunder the action of fixatives they tend to separate into theirconstituent fibres, and, as the cilia are only of moderate size,the tufts of fibres are not very noticeable. They were found toretain their form better in sections of material fixed in 2 per cent,osmic acid than in Bouin-Duboscq's fluid.

Subsidiary latero-frontal cilia are present in the Ostreidae,but are very difficult to distinguish even in the living gill. Theyare possibly somewhat smaller than those of bivalves possessingeu-latero-frontal cilia; however, the gills are rather opaque andthe cilia therefore difficult to observe even in a preparation ofa single lamella. Apart from the organic junctions, the presenceof numerous blood cells in the lacunae of the demibranch tendsto make it opaque, especially when these are greenish.

As will be seen from the second half of this paper (p. 383), itis probable that the anomalous latero-frontal cilia of theOstreidae are not homologous with eu-latero-frontal cilia, andas the homology of the subsidiary latero-frontal cilia of thisfamily with the pro-latero-frontal cilia of forms possessing eu-latero-frontals is doubtful, it is proposed to call them p a r a -l a t e r o - f r o n t a l c i l i a . In Text-fig. 1 of Part II (1937),therefore, these cilia should have been labelled para-l-f.c. andnot pro-l-f.c. Para-latero-frontal cilia have been omitted fromTable I, p. 370, for, from the little that is known of them, theyappear to agree in structure and arrangement with pro- andmicro-latero-frontal cilia.

To summarize: anomalous latero-frontal cilia agree with eu-latero-frontal cilia in being borne singly on small cells, of whichthe greater diameter is at right angles to the length of the fila-ment ; in having, so far as is known, a similar structure and asimilar arrangement of their ciliary rootlets; and in the directionof the effective beat, position of rest, and type of metachronalwave. They also agree in being accompanied by subsidiarylatero-frontal cilia.

They differ from eu-latero-frontal cilia, however, in theirsmaller size, especially as to the width at the base in the planeof the beat, inconspicuousness in sections, and marked variationin length and spacing on different filaments of the same individual.

368 D. ATKINS

LAMELLIBRANCHS WITH EU-LATERO-FRONTAL CILIA.

In 1891 Pelseneer depicted large latero-frontal cilia in allthirteen of the lamellibranch gills he figured in his ' Contributiona l'etude des Lamellibrancb.es', including those of five species,Anomia e p h i p p i u m , P e c t u n c u l u s (=Glycymer i s )g lycymer i s , Area b a r b a t a , Pec t en o p e r c u l a r i s ,and Lima h ians in which large ones are undoubtedly absent.Kellogg (1892) soon after stated that they were absent inSolenomya ve lum and P e c t e n irradians—in the firstinstance almost certainly incorrectly—and expressed the opinionthat latero-frontal cilia are ' not so widely found among Lamelli-branchs, I believe, as seems to be so generally supposed'. Kide-wood (1903), in the course of his extensive work on the gills ofLamellibranchs, found them to occur in the great majority (hisremarks on these cilia are based presumably on preservedmaterial only), and this has been borne out in the present work,mostly on living gills. As late as 1930, however, Nicol referred tospecies possessing latero-frontal cilia—that is eu-latero-frontals—as ' aberrant Lamellibranchs'.

During the course of the present work, large or eu-latero-frontal cilia were found in members of the three families ofProtobranchia (they were previously recorded by Orton, 1912,1914, in Nucula and Solenomya t o g a t a ) ; in all themarine families of Eulamellibranchia obtamable at Plymouth,and in the fresh-water families Dreissensiidae, Sphaeriidae,Unionidae, Mutelidae, and Aetheriidae. In the Filibranchiathey were found in the Mytilidae and Trigoniidae only. Theywere absent in all the Pseudolamellibranchia examined. A listof the species of Lamellibranchia found to have eu-latero-frontal cilia is given on p. 374. Pro-latero-frontal cilia areperhaps always present also, though they have not always beenrecognized owing to difficulties of observation or imperfect pre-servation of material.

A visual impression is that there is little or no direct correla-tion between the size of the bivalve, gill, or filament, and thatof the eu-latero-frontal cilia. There is no great difference in thesize of these cilia in Modiolus mod io lus , which may reach

THE CILIARY MECHANISMS OF LAMELMBBANCHS 369

a length of 11 cm. or more, and in K e l l i a s u b o r b i c u l a r i s ,of which the largest specimen seen was about 1 cm. long. Thedepth of a filament, from the frontal to the abfrontal surface,of M o d i o l u s m o d i o l u s is more than four times that ofone of K e l l i a s u b o r b i e u l a r i s , while the frontal surface isabout three times wider, antero-posteriorly, as shown in Text-fig. 1 A and B (p. 351).

While a certain amount of variation in the size of eu-latero-frontal cilia is found when actual measurements are taken,there is no possibility of confounding them with micro-latero-frontal cilia. (A provisional comparison of the different typesof latero-frontal cilia is given in Table I.)

Among fresh-water forms, L a m p s i l i s and Q u a d r u l a arestated by Grave and Schmitt (1925) to have latero-frontal ciliawhich often exceed 15/x in length, and are about 1 to lj/x indiameter at the base. While this is unusually small for eu-latero-frontal cilia, the size at the base is sufficient to preventtheir being confounded with micro-latero-frontal cilia. Measure-ments were made of these cilia in a few fresh-water forms duringthe present investigation. D r e i s s e n s i a p o l y m o r p h a wasfound to have eu-latero-frontal cilia about 24/u. long, androughly about 9ju wide at the base in the plane of the beat,measured living. Those of A e t h e r i a e l l i p t i c a are about27ju. long (measured from preserved entire filaments), thoughthey appear somewhat less than this in transverse section (Text-fig. 6 A), and are about 6 or 7/* wide at the base in the plane ofthe beat (measured from sections); those of M u t e l a b o u r -g u i g n a t i (Text-fig. 6 B ) , judging from sections, are about20fj, long, and about 6/x wide. Measurements made on sections,however, are unreliable, as apart from the occurrence of shrink-age, the entire length of the cilium may not be cut.

Among marine forms Carter (1924) gave their length in livingM y t i l u s g a l l o p r o v i n c i a l i s as about 30/*, and at the base| to 1 ft broad in the plane perpendicular to the beat, and 4 to5fi wide in the plane of the beat. The only measurements I havemade of eu-latero-frontal cilia of marine forms were in G a s t r o -c h a e n a d u b i a , in which these cilia are roughly about 26filong and 6 to 8ja wide at the base in the plane of the beat

NO. 319 B b

Cells.

Number of cilia toa cell.

Structure of cilium.

Basal granules.

Ciliary rootlets.

Length of cilium.

Width of cilium atbase in the planeof the beat.

TABLE I. Provisional

Eu-latero-frontal Cilia.

Small: longer diameter at rightangles to length of filament.

1

Composed of a number of triangularplates (10-15in Mytilus accord-ing to Carter, 1924), bounded byfibres on each side of the cilium.Surface of plates at right angles toplane of beat: plates shorter onside of cilium which is directedforward in effective beat. At thebase the cilium is considerablywider in the plane of the beat thanat right angles to it, so is character-istically broadly triangular in sideview.

Each plate associated with a pair,so that in surface view of the cellthere is a double row of basalgranules at right angles to thelength of the filament.

Two diverge from each basalgranule, and pass on oppositesides of the nucleus, there beingtwo pairs to each plate.

About 20-30*1 (Lampsilis andQuadrula often exceed 15 M).

About 4-9 M, excluding those ofLampsilis and Quadrulawhich are said to be 1-11M indiameter, and of Xylophagawhich are about 2 /i wide.

Comparison of Types of Latero-frontal Cilia.Anomalous latero-frontal

Cilia.

As eu-Iatero-frontalcilia.

As eu-Iatero-frontalcilia.

Probably as eu-latero-frontal cilia, thoughthey are not as large.

Probably as eu-Iatero-frontal cilia.

Probably as eu-latcro-frontal cilia.

14-25 n.

Roughly about 1 i ii,measured from sec-tions.

Micro-latero-frontal Cilia. *

Long, narrow: elongated inthe same direction as fila-ment.

More than 1: about 18 inGlycymeris glycy-meris.

Probably no plates, buteach cilium composed of afew fibres, probably placedone behind the other inthe plane of the beat, sothat the cilium is slightlywider in this plane, thanin that at right angles to it.

The arrangement was notobserved.


About 12M in Hetera-notnia and Monia:14-17(i in Glycymerisand Area.

Not measured, but probablynot much greater than atright angles to the planeof the beat.

Pro-latero-frmtal Cilia.*

As micro-latero-frontal cilia.

A number to each cell (seep. 352).

Probably no plates, but eachcilium composed of a fewfibres, probably placed onebehind the other in theplane of the beat. Thedifference in the width andbreadth at the base isslight.



About 10 u in Modiolusmodiolus, and 6JJ inKellia suborbicularis(measured from sections).

Not measured, but probablynot much greater than atright angles to the planeof the beat.

CO

3

Para-latero-frontal cilia appear to agree with micro- and pro-Iatero-frontal cilia in the little that is known of them.

THE CILIAEY MECHANISMS OF LAMELLIBRANCHS 371

ro-l-F.c.pro-L-fic

eu-l-F.c.JJv}eu-t-f.c

L.c.

p.ch.

TEXT-FIG. 6.Transverse sections of the filaments of two fresh-water Lamelli-

branchs, having eu-latero-frontal cilia. A, Aether ia e l l i p t i ca ;B, Mutela b o u r g u i g n a t i (inner demibranch). abf.c, ab-frontal cilia; c.r., calcined rod; eu-l-f.c, eu-latero-frontal cilium;f.c, frontal cilia; g.c, gland-cell; I.e., lateral cilia; m., muscle-fibres; p.ch., pale-staining fibrous chitin; pro-l-f.c, pro-latero-frontal cilium (?). The chitinous skeleton is shown in black inAe the r i a , but in Mutela it is shaded and the calcified rodsshown in black. Alcohol fixation; iron haematoxylin and acidfuchsin. Owing to the condition of fixation pro-latero-frontalcilia could not be identified with certainty. X 735.

(measured living), and in Xylophaga dorsalis , in •whichthey are roughly 20/u. long and only 2//, wide at the base in theplane of the beat (measured from sections). This narrownessof the eu-latero-frontal cilia in Xylophaga is not found inthe allied Barnea parva, and seems exceptional amongmarine forms.

872 D. ATKINS

B

TEXT-FIG. 7.

Transverse sections of the gill leaflets of two Protobranohs. A,Nucula ; B, Solenomya t o g a t a . abf.c, abfrontal cilia;c.k., ciliated knob, or glandular organ; eu-l-f.c, eu-latero-frontal cilia ; / .c , frontal cilia; I.e., lateral cilia; m.f., muscle-fibres;s.c, slight ciliation dorsal to the lateral cilia. Bouin-Duboscq'sfixative; iron haematoxylin. x 735.

In the Protobranehia there is a certain amount of variationin the shape of the eu-latero-frontal cilia. In Nucula they are

THE CILIABY MECHANISMS OF LAMELLIBRANCHS 373

-t.m.f

TEXT-FIB. 8.

Transverse section of a gill leaflet of the Protobranch Nucu lanam i n u t a . The whole is not figured because of the great depth(frontal to abfrontal) of the leaflets. The involution of the wallsnear the lateral ciliated tracts is due to contraction of the muscles,c./.c, coarse, and f.f.c, fine frontal cilia; eu-l-f.c, eu-latero-frontal cilia; I.e., lateral cilia; m.f., radiating striated muscle-fibres; s.c, slight ciliation dorsal to the lateral cilia; t.m.f., trans-verse muscle-fibres. Bouin-Duboscq's fixative; iron naematoxylin.X 735.

rather short, but very broadly triangular in side view (Text-fig. 7 A); while in Nuculana minuta (Text-fig. 8) and

374 D. ATKINS

Solenomya togata (Text-fig. 7 B) they are longer, but lessbroadly triangular. In Nuculana minuta they are about30/J, long, and in Nucula nucleus about 20/x long, measuredliving.

What is especially characteristic of the size of eu-latero-frontal cilia, is not so much the length, which apparently mayvary from 15 to 30/* at least, but their breadth at the base ina plane perpendicular to the beat, and more especially theirrelatively great width in the plane of the beat; this last dimen-sion giving them their distinctively triangular shape in side view.

List of Lamellibranchs found1 to haveEu-latero-frontal Cilia.

PROTOBEANCHIA.

Solenomyidae: Solenomya togata.2

Nuculidae: Nucula nucleus (L.), Nucula radiataHanley, Nucula ni t ida Sowerby.

Nuculanidae: Nuculana ( = Leda) minuta (Muller),Nuculana pella.2

FILIBRANCHIA.

Trigoniidae: Trigonia margaritacea.2

Mytilidae: Mytilus edulis L., Modiolus modiolus(L.), Modiolus adriat icus Lamarck, Modiolusphaseolinus (Philippi), Musculus (=Modiolaria)marmoratus (Forbes), Museulus ( = Crenella)discors (L.).

EULAMELLIBRANCHIA.

Dreissensiidae: Dreissensia polymorpha (Pallas).Astartidae: As ta r te sulcata (da Costa).Thyasiridae: Thyasira flexuosa (Montagu).Lucinidae: Myrtea spinifera (Montagu), Phacoides

borealis (L.).Ungulinidae: Diplodonta ro tunda t a (Montagu).

1 The list is comprehensive, but does not include the examination ofmembers of all families of Lamellibranohs.

2 Preserved material only examined.

THE CILIABY MECHANISMS OP LAMELLIBRANCHS 375

Erycinidae: Kellia suborbicular is (Montagu),Lasaearubra (Montagu).

Galeommatidae: Galeomma tu r ton i Sowerby.Leptonidae: Lepton squamosum (Montagu).Montacutidae: Montacuta ferruginosa (Montagu),

Mysella b iden ta t a (Montagu), Entova lva per-rieri (Malard).

Cyprinidae: Oyprina is landica (L.).1

Sphaeriidae: Sphaerium corneum (L.).Unionidae: Anodonta ana t ina (L.).Aetheriidae: Aetheria el l ipt ica,1 Miilleria daleyi.1

Mutelidae; Mutela bourguignat i . 1

Cardiidae: Cardium echinatum L., Cardium ovaleSowerby, Cardium edule L., Cardium crassumGmelin (=norvegicum).

Veneridae: Dosinia exoleta (L.), Dosinia lupinus(L.), Gafrarium minimum (Montagu), Venusverrucosa L., Venus easina L., Venus ovataPennant, Venus fasciata (da Costa), Venus s t r ia-tula (da Costa), Paphia ( = Tapes), rhomboides(Pennant), Paphia pu l las t ra (Montagu), Paphiadecussata (L.).

Petricolidae: Pe t r ieo la pholadiformis Lamarck,Mysia unda ta (Pennant).

Donacidae: Donax v i t t a t u s (da Costa).Tellinidae: Tellina tenuis da Costa, Tellina fabula

GmeHn, Tellina donacina L., Tellina crassaPennant, Macoma bal th ica (L.).

Semelidae: Scrobicularia plana (da Costa) ( = pi-perata), Abra alba (S. Wood), Abra n i t ida(Miiller).

Asaphidae: Gari fervensis (Gmelin) (=ferroensis),Gari te l l inel la (Lamarck).

Solenidae: Solen margina tus Montagu (=vagina),Ensis sil iqua (L.), Ensis a rcua tus (Jeffreys),Ensis ensis (L.), Cultellus pellucidus (Pennant).

Solecurtidae: Solecurtus scopula (Turton) (=candi -1 Preserved material only examined.

376 D. ATKINS

dus), Solecurtus chamasolen (da Costa) ( = a n t i -quatus).

Mactridae: Mactra corall ina (L.), Spisula el l ipt ica(Brown), Spisula solida (L.), Spisula subtrun-cata (daCosta).

Lutrariidae: Lu t ra r i a lu t r a r i a (L.) ( = elliptica).Myidae: Mya t runca ta L.Erodonidae: Aloidis gibba (Olivi) (=Corbula nu-

cleus).Hiatellidae: Hia te l la arct ica (L.), Hia te l l a galli-

cana (Lamarck) ( = rugosa).Gastrochaenidae: Gastrochaena dubia (Pennant).Pholadidae: Barnea Candida (L.), Barnea parva

(Pennant), Pholadidea loscombiana Turton, Xylo-phaga dorsalis Turton.

Teredinidae: Teredo navalis L.Periplomatidae: Cochlodesma praetenue (Montagu).Thraciidae: Thracia vi l losiuscula (Macgillivray),

Thracia d is tor ta (Montagu).Lyonsiidae: Lyonsia norwegica (Gmelin).Pandoridae: Pandora pinna (Montagu) ( = obtusa).

LAMELLIBBANCHS WITH MICRO-LATERO-FBONTAL CILIA.

In Lamellibranchs in which micro-latero-frontal cilia occur,the frontal cilia frequently extend round on to the lateral facesof the filaments, and, therefore, the micro-latero-frontal cilia areactually lateral in position, thus increasing the difficulty ofdiscerning these tenuous cilia when a filament is viewed fromthe side. A certain amount of shrinkage, however, occurs onfixation, so that in transverse section they are more or lesslatero-frontal in position (Text-fig. 3, p. 358).

In successfully stained transverse sections of well-fixedmaterial the micro-latero-frontal cilia themselves can frequentlybe seen, and nearly always they can be distinguished from thefrontal cilia by their more conspicuous basal granules and darkerstaining ciliary rootlets. But by far the best way to determinetheir presence is in the living gill. When living material wasunobtainable, and where fixation was far from perfect, as in

THE CILIAEY MECHANISMS OF LAMELLIBBANCHS 377

alcohol-preserved museum specimens, it seemed justifiable todeduce the presence of micro-latero-frontal cilia when eu-latero-frontals were clearly absent. Entire filaments were examinedunstained in alcohol, or formalin, with a drop or two of glycerineadded, and the results obtained verified by sectioning. In entirefilaments the presence of eu-latero-frontal cilia is generally easilydeterminable, unless the material is in very poor condition.The arrangement and size of the nuclei of the latero-frontal cellswas also taken into consideration. When eu-latero-frontal ciliaare present the row of large, closely set nuclei is conspicuous, andthe absence of such a row was taken as confirmative evidenceof the absence of eu-latero-frontal cilia. The nuclei of the micro-latero-frontal cells, which are elongated in the direction oflength of the filament, and are widely spaced, are inconspicuousin the entire filament. It is not impossible that the presence ofmoderate-sized or anomalous latero-frontal cilia, as in O s t r e a ,might be overlooked, though the presence of a row of small,closely set nuclei, as in that genus, should be evident at leaston wide principal filaments.

Micro-latero-frontal cilia are difficult to distinguish even inliving gills until one has become accustomed to them, and itis understandable that their presence escaped workers restrictedto preserved material. In 1931 on a first examination ofH e t e r a n o m i a s q u a m u l a (L.) ( = Anomia a c u l e a t aMiiller) latero-frontal cilia were thought to be absent, and it wasonly in view of Orton's (1914, p. 299) published statement oftheir presence (his notebook in which they are described as'extremely fine' had not then been seen, and was not untilSeptember 1932), that a careful search was made and they werediscovered in a reduced form. The difference in size of the largelatero-frontal cilia of many Lamellibranchs and the tenuousones of H e t e r a n o m i a was so striking, that it seemed possiblethat small ones had been overlooked in Pec t e n , and otherspecies, by workers familiar with them in their fully developedform, and it was decided to examine the bivalves available, withthe results given in this paper.

Micro-latero-frontal cilia would seem to be straining cilia, orat least to act as a guard to prevent loss of food particles from

378 D. ATKINS

the frontal tract, but their effectiveness as such is presumablyless than that of the large ones, and very much less in H e t e r a -n o m i a and Monia (Atkins, 1986).

Micro-latero-frontal cilia have been found in the followingfamilies: Arcidae, Anomiidae, Pteriidae, Pectinidae, Spondy-lidae, Limidae, Pinnidae, and are inferred to be present inthe Amussiidae (Atkins, 1938). They are also inferred to bepresent in the Isognomonidae and Vulsellidae (see p. 376). Alist of species is given on p. 379.

The distribution in the Lamellibranchia of species havingmicro-latero-frontal cilia is of great interest. The form of thesecilia does not appear to be correlated with habitat, for bivalvespossessing them are adapted to a variety of environments. Someare shallow water forms; others have a wide vertical range.Some live in comparatively clear water; others in situationswhere sediment is liable to be heavy. A number are attachedby a byssus, temporarily (Area, P i n c t a d a , M a l l e u s ,I s o g n o m o n , P i n n a , and certain Pectinidae and Limidaeto rocks and stones; P t e r i a to Eun ice l l a ) , or permanently(A no mi a, H e t e r a n o m i a , and Monia) ; others are ce-mented to rock (Chlamys d i s t o r t a , S p o n d y l u s , andP l i c a t u l a ) . Among these attached forms P i n n a is perhapsalone in being able to burrow by means of a strong water-current(Grave, 1911). P l a c u n a and G l y c y m e r i s are unattached.According to Hornell (1909)' the bottom favoured by P l a c u n ap l a c e n t a is a fairly stiff or pasty greyish-black mud. On thisthe shells generally lie prone upon their convex left valves, thehinge region sometimes slightly sunk in the mud, which maylightly cover the dorsal third of the shell.' G l y c y m e r i sg l y c y m e r i s burrows by means of its foot in sand and shellgravel, lying more or less hidden beneath the surface (Atkins,1936). The two burrowing members of the group do not employthe same methods: the free living, circular valved G l y c y m e r i sburrows chiefly by means of its foot, though with the aid of acurrent expelled from the anterior region of the shell; the byssi-ferous, triangular shaped P i n n a by means of a water-currentexpelled from the pointed anterior end of the shell (see Grave,1911); this sharply pointed extremity is no doubt of assistance


in penetrating the mud and soft sand in which the animal lives.Many of the Pectinidae and Limidae, and the Amussiidae arefree swimming; Lima hians is nidamentous. Vulsellalives embedded in sponges.

There thus appears to be no correlation between habitat ormode of life and the possession of micro-latero-frontal cilia, but,as will be seen later (p. 383), it is considered that these ciliaindicate the genetic relationship of families possessing them.

List of Lamell ibranchs found to have Micro-la tero-f ronta l Cilia.

FlLIBRANCHIA.

Arcidae: Area te t ragona Poli, Area lactea L., Area(Scaphula) celox Benson,1 Glycymeris glycy-meris (L.).

Anomiidae; Anomia ephippium L.,1 Heteranomiasquamula (L.), Monia patel l iformis (L.), Moniasquama (Gmelin), Placuna placenta . 1

PSEUDOLAMELLIBRANCHIA.

Pteriidae: Pter ia hirundo (L.) (=Avicula taren-tina), P inc tada (=Margaritifera==Meleagrina)margar i t i fera (L.),1 P inc tada vulgaris (Schu-macher),1 Malleus albus Lamarck1 (see, however,p. 363).

Vulsellidae: Vulsella sp.1 (most probably).Isognomonidae: Isognomon ala ta1 (most probably).Pinnidae: Pinna f ragilis Pennant (see, however, p. 361).Pectinidae: Pecten maximus (L.), Chlamys dis-

to r t a (da Costa) ( = pusio), Chlamys vi t rea(Gmelin),1 Chlamys opercularis (L.), Chlamyst iger ina (Muller).

Amussiidae: Amussium pleuronectes L.1 (mostprobably).

Spondylidae: Spondylus gaederopus,1 Spondylussp.1 (from Great Barrier Reef).

1 Preserved material only examined.

380 D. ATKINS

Limidae: Lima hians (Gmelin), Lima loscombiSowerby.

LAMELLIBEANCHS WITH ANOMALOUS LATERO-FRONTAL CILIA.

Anomalous latero-frontal cilia, as previously mentioned, havebeen found in only one family of Lamellibranchs, the Ostreidae,which were placed by Pelseneer (1911) in the Pseudolamelli-branchia. The species examined were Ostrea edulis L.,Ostrea v i rg in ica Gmelin, and Ostrea angu l a t a(Lamarck).

THE ARRANGEMENT OF THE VARIOUS CILIARY TRACTS ON

THE GILL FILAMENTS.

(a) In Lamel l ib ranchs with E u - l a t e r o - f r o n t a lCilia.

The common arrangement of the various ciliary tracts of thegill filaments in bivalves with eu-latero-frontal cilia is that shownin transverse sections of Modiolus modio lus , Kell iasuborb icu la r i s (Text-fig. 1 A-B, p. 351), Mutela bour-gu igna t i and Aether ia e l l ip t ica (Text-fig. 6, p. 371).Abfrontal cilia may or may not be present (abf.c, Text-fig. 6).

In two Protobranch families, Nuculidae and Solenomyidae,the arrangement is closely comparable, except that the tractsof lateral cilia are very wide, 26JU.1 in Nucula r a d i a t a and30/A in Solenomya toga t a (Text-fig. 7 A-B, p. 372). Awide tract of lateral cilia is possibly primitive, for it is found incertain Gastropods, for instance in the Ehipidoglossae, Fis-surel la graeca L., Trochus c ine ra r ius L. (Pelseneer,1891), P l e u r o t o m a r i a beyr ich i i (Woodward, 1901),and Ha l io t i s t u b e r c u l a t a (Crofts, 1929), and in thePectinibranch, Crepidula forn ica ta (Orton, 1912). InNucula the unciliated tract between the lateral and eu-latero-frontal cilia is of some width;in Solenomya toga t a it isnarrow, and the tips of the lateral cilia extend to the level ofthe frontal surface (Text-fig. 7 B). Kellogg (1892, PI. XCL

1 Measured from sections parallel with the surface of the cells; in livingfilaments they are probably wider.

THE CILIAEY MECHANISMS OF LAMELLIBEANCHS 381

fig. 77) showed the lateral and frontal cilia as continuous inS o l e n o m y a , but this was corrected by Orton (1913).

In the Protobranch family Nuculanidae the lateral ciliatedtracts, six or seven cells wide (measuring 18 to 22/x in sectionsof N u c u l a n a mi n u t a), are far from the frontal surface(I.e., Text-fig. 8, p. 373), but owing to the great depth (frontalto abfrontal) of the leaflets, they are yet much closer to it thanto the abfrontal surface. Close to the eu-latero-frontal cilia(eu-l-f.c.) is a tract of fine frontal cilia (f.f.c.): between these andthe coarse frontal cilia (c./.c), the main tract of which is actuallymore or less on the frontal surface, is an unciliated region. Adescription of the ciliation of the gill leaflets of N u c u l a n ami n u t a has already been given (Atkins, 1936).

The arrangement of the ciliated tracts in T r i g o n i a m a r -g a r i t a c e a (Text-fig. 1 c, p. 351) is unlike that found in anyother Filibranch or Eulamellibranch, resembling rather thatin the Protobranch N u c u l a n a m i n u t a (Text-fig. 8), thoughwith certain differences. It is not suggested, however, that thereis any close relationship between T r i g o n i a and N u c u l a n a ,which is a considerably specialized Protobranch (Atkins, 1936).The four rows of lateral cilia (I.e.) are far distant from the frontalsurface, being in fact slightly nearer the abfrontal than thefrontal surface (as in the Gastropod C r e p i d u l a (Orton,1914)). Separated from the lateral cilia by one or two unciliatedcells are the large latero-frontal cells, bearing large and longcilia (eu-l-f.c.). Both cells and cilia have the characteristic ap-pearance found in numerous Lamellibranchs. Determination ofthe presence or absence of pro-latero-frontal cilia was im-possible owing to poor fixation. The cilia clothing the frontalsurface are remarkably long (c.f.c.); they are shown in the figureat about their full length, though they are generally cut shortin sections. In a latero-frontal position are long cilia (J.c.l.),which, judging by their attitude in sections, beat mainly acrossthe length of the filament, though it was impossible to deter-mine the direction of the effective beat. From an examinationof the entire filament it was seen that they are borne onthree or four rows of long narrow cells, the cells being elongatedin the direction of length of the filament. These cilia may

382 D. ATKINS

possibly send particles toward the frontal surface (that is if theeffective beat should prove to be toward that surface), as dothe scattered patches of coarse cilia in Nuculana minuta(Atkins, 1936); or (if the effective beat should prove to be ab-frontal in direction) they are perhaps additional current pro-ducing, that is functionally lateral, cilia: an examination ofliving material is the best way to determine this. Between thesecilia and the true latero-frontal cilia, the surface of the filamentappears to be uniformly ciliated (f.f.c), though it is impossibleto be certain of this owing to the imperfect fixation. The ciliaare finer, shorter, and not as closely set as those of the frontalsurface. By analogy with certain other bivalves, it seemspossible that the long, stout cilia of the frontal surface will befound to transport unwanted particles ventrally, while the fineones, clothing the sides of the filaments between the coarsefrontal and the latero-frontal cilia, will be found to conveyparticles which are destined to be eaten. The frontal currentson the gills of this species are likely to be of much interest.

Ridewood (1903, p. 202) stated that the frontal and lateralcilia of Trigonia lamarki are normal. Unless the arrange-ment of the ciliated tracts of the gill filaments differs widelyin the two species, Trigonia margar i tacea and Tri-gonia lamarki , he evidently included the true latero-frontal cilia (eu-l-f.c, Text-fig. 1 c) with the lateral (I.e.), andassumed the long cilia in a latero-frontal position (f.c.l.) to bethe true latero-frontal cilia. The arrangement of the cilia is sounusual that I have ventured to give a figure, although thematerial was preserved in alcohol for museum purposes and thefixation is imperfect. Examination of well-preserved material,and above all of the living gill is very desirable.

Pelseneer (1891, PL XIII, fig. 43) gave a figure of a transversesection of a filament of Trigonia pec t ina ta in which nolateral cilia are labelled; it is possible that what he has namedciliated discs are the lateral, together with the latero-frontal,cilia. In Trigonia margar i t acea , and also in Trigonialamarki (see Eidewood, 1903), the ciliated discs are closeto the interlamellar edge of the filament, and not as shown inPelseneer's figure.


An interesting feature of the filaments of Trigonia mar-gar i tacea (alcohol preservation) is the presence of calcifiedrods (c.r., Text-fig. 1 c), which were previously thought to occuronly in Miilleria and the Unionidae (see p. 394).

In Xylophaga dorsalis the frontal cilia extend wellround on to the lateral surfaces of the filaments, and the un-ciliated space between the latero-frontal and lateral cilia is ratherwide. These peculiarities are not found in the allied Barneaparva .

(b) In Lamell ibranchs with Micro-latero-frontalCilia.

In these Lamellibranchs the arrangement of the variousciliary tracts appears to be more or less constant (see Text-fig. 3,A-E, p. 358), though the width of the frontal tract may vary,the cilia extending round on to the lateral faces; abfrontal ciliamay, or may not, be present.

B. LATEBO-FEONTAL CILIA AND PHYLOGENY.

DISCUSSION ON THE EELATIONSHIPS OF THE VABIOUSTYPES OF LATERO-FRONTAL CILIA.

From the work recorded in the foregoing pages it is now knownthat three variations of the latero-frontal tract of Lamelli-branchs exist, namely: (1) eu-latero-frontal cilia, together withpro-latero-frontal cilia; (2) anomalous, together with para-latero-frontal cilia; and (3) micro-latero-frontal cilia. Thesecond variation is known from one family only, the Ostreidae.

It is known that eu-latero-frontal and—though with lesscertainty—pro-latero-frontal cilia (see p. 356) are present inliving Protobranchs, so that it is possible that the latero-frontaltract of the higher Lamellibranchs, with eu- and pro-latero-frontal cilia, may have been inherited from that order, andis not necessarily a new development. The presence of eu-latero-frontal cilia on gills, which, certainly in Nuculanaand Nucula, are not used greatly in feeding, points to theantiquity of these cilia in the Protobranchia.

The relationship of the various types of latero-frontal cilia

884 D. ATKINS

is speculative whilst so little is known of the structure of thetenuous kinds of latero-frontal cilia (namely, the micro-, pro-,and para-latero-frontal cilia) and their cells: a detailed studyof the histology of these was outside the scope of the presentwork, but a provisional comparison of the structure of thevarious types of latero-frontal cilia is given in Table I, p. 370.So far as is known the micro- and pro-latero-frontal cilia areclosely comparable in structure; they are both of small size,are borne a number on a cell, and the cells and their nuclei arenarrow and elongated in the direction of length of the filament.It is probable that these cilia are not composed of triangularplate-like elements, as are the eu-latero-frontal cilia of M y t i 1 u sand others, but of a few simple fibres. Whether, however, themicro- and pro-latero-frontal cilia are homologous, or onlyanalogous, is not certain. The possibilities are that micro-latero-frontal cilia (a) are homologous with pro-latero-frontalcilia; (b) have given rise by specialization to eu- and to anoma-lous latero-frontal cilia; (c) are distinct structures, and so onlyanalogous to pro-latero-frontal cilia; or (d) are reduced eu-latero-frontal cilia, though this is unlikely (see p. 387).

The anomalous latero-frontal cilia of the Ostreidae appearto be closely comparable in structure with eu-latero-frontalcilia, though of smaller size. They agree in that a single complexcilium arises from a cell, the cells are small, with the greaterdiameter at right angles to the length of the filament, and thearrangement of the ciliary rootlets appears to be similar in thetwo instances. However, it is extremely probable from a con-sideration of phylogenetic relationships, based on other char-acters (see p. 409), that the anomalous latero-frontal cilia of theOstreidae are not homologous with eu-latero-frontal cilia, butthat these cilia of similar structure have arisen independentlyin the Lamellibranchia, and that their similarity is due to con-vergence: one has not attained the size of the other.

In the Ostreidae not only are there anomalous but also para-latero-frontal cilia. It is therefore necessary to consider theorigin of the two rows. One possibility is that micro-latero-frontal cilia, such as are found in other members of the group,have persisted as the para-latero-frontal cilia of the Ostreidae,

THE CILIARY MECHANISMS OP LAMELUBKANCHS 385

the anomalous arising as new structures not derived by modifica-tion from any pre-existing cilia. A second possibility is that theanomalous latero-frontal cilia have been derived from micro-latero-frontal cilia, perhaps by the approximation of these inpairs, and the formation of connecting membranes, the para-latero-frontal cilia being new structures, perhaps modifiedfrontals. In this event there would have been progressivespecialization of frontal cilia, the outermost of these beingmodified to form micro-latero-frontal cilia, and these again toform anomalous cilia, while a subsidiary row of latero-frontals(para-latero-frontal cilia) arose by modification of the nowouter frontal cilia. An investigation into the composition of thelatero-frontal tract in the Pteriacea to determine whether asecond row of cilia is actually present in P i n n a , and may bein other members of the sub-order, might throw light on thequestion.

It has already been shown how the anomalous cilia of 0 s t r e acould have been derived from the type of cilia of the main rowin P i n n a (see p. 362), and if two rows should occur not onlyin P i n n a , but in other members of the Pteriacea, it seemsvery possible that the anomalous latero-frontal cilia haveoriginated in some such manner.

The appearance of latero-frontal cilia in the Lamellibranehia,whether evolved independently in the different lineages, or asmodifications of one ancestral type must be considered. To thelateral and frontal cilia inherited probably from ancestorscommon to both Gastropods and Lamellibranchs (Pelseneer,1891) have been added latero-frontal cilia. The latero-frontaltracts, however, are not all of the same type, and the possi-bilities are either that the variations have been derived fromone original type, or that the various branches of Lamellibranchshave independently evolved the same straining device, that is,the same sort of latero-frontal cilia on each independent lineof evolution. The evolutionary tendency has been directed inthis event toward the attainment of complex cilia composed oftriangular plates, namely eu- and anomalous latero-frontal cilia.

Pour suggestions may be made: (a) The ancestral Lamelli-branchs may have had a single row of micro-latero-frontal cilia

NO. 319 C o '

386 D. ATKINS

on each side of the frontal tract, probably arising by modifica-tion of the outer frontal cilia. This constitution of the strainingapparatus persisted in most members of a branch representedto-day by the Arcidae, Anomiidae, and Pseudolamellibranchia,but in certain members there was an attempt at the productionof a more efficient apparatus by the addition of larger cilia,possibly in P i n n a , and certainly in the anomalous cilia of theOstreidae, the highest, though by no means the most recent,member of the group. In another branch, or branches, theaddition of large complex cilia to the original micro-latero-frontal tract took place early, such branch or branches beingrepresented by the Protobranchia, Mytilidae, Trigoniidae, andEulamellibranchia, in which the straining apparatus consistsof a row of pro- ( = micro ?) and a row of eu-latero-frontal ciliaon each side of the frontal tract. There would thus have beenparallel evolution of complex cilia, eu-latero-frontal cilia in onebranch or branches, and anomalous in another. As suggestedpreviously, when two rows are present, the micro-latero-frontalcilia may have persisted as para- and pro-latero-frontals (thatis, all are homologous), the anomalous and eu-latero-frontal ciliabeing new structures, not derived from any pre-existing cilia;or the two rows may have arisen by progressive modificationof the outer frontal cilia, in which case the eu- and anomalouslatero-frontal cilia would have arisen by specialization of micro-latero-frontal cilia.

(b) Secondly, the micro-latero-frontal cilia type and the eu-,together with pro-latero-frontal cilia type, may have been inde-pendently evolved from forms in which no latero-frontal ciliawere present, the micro- and pro-latero-frontals being ana-logous.

(c) Thirdly, the original type of latero-frontal tract may havebeen eu-, together with pro-latero-frontal cilia, this type per-sisting in the branch or branches represented by the Proto-branchia, Mytilidae, Trigoniidae, and Eulamellibranchia. Theloss of eu-latero-frontal cilia would then have to be postulatedin the group represented by the Arcidae, Anomiidae, andPseudolamellibranchia, with a fairly successful attempt at theregaling of complex cilia in the anomalous latero-frontal cilia


of the Ostreidae. This loss and then recovery of complex ciliais unlikely. If this has occurred then the micro- would behomologous with the pro-, but not necessarily with the para-latero-frontal cilia.

(d) Fourthly, the original type of tract may have been eu-latero-frontal cilia alone. To these pro-latero-frontal cilia wereadded in Protobranchia (probably), Mytilidae, Trigoniidae(possibly), and Eulamellibranchia. It would then be necessaryto postulate the reduction of eu- to miero-latero-frontal ciliain the Arcidae, Anomiidae, and Pseudolamellibranchia, andeither their enlargement to anomalous cilia with the additionof para-latero-frontal cilia in the Ostreidae, or the persistenceof micro- as para-, and the addition of anomalous latero-frontalcilia.

Of these four possibilities (c) and (d) seem the most improbable.The evolution of straining cilia has apparently occurred only

in the Lamellibranchia, no other group of ciliary feeders, sofar as is known, having acquired this refinement of the ciliarymethod of feeding.

Essential ciliated tracts of the food-collecting mechanism ofthe gills are the lateral, as current producing, and the frontalcilia, for conveying material, these being found in both Gastro-pods (Streptoneura) and Lamellibranchs: analogous tracts occurin several groups of ciliary feeders outside the Mollusca. Thelatero-frontal or straining cilia increase the efficiency of theciliary method of feeding, but apparently are not essential.Latero-frontal cilia have not been described in any Gastropod,so that it may be taken for granted that eu-latero-frontals areabsent. The only Gastropods that were examined for micro-latero-f rontal cilia were the Pectinibranch C r e p i d u l a fo m i -cat a and the Ehipidoglossate Aspidobranch C a l l i o s t o m az i z y p h i n u m , and in these species I was unable to find them:C r e p i d u l a is a ciliary feeder (Orton, 1912), while Ca l l i -o s t o m a possibly does not consume the material collected bythe gills. Though this is very slight evidence indeed, it is perhapssafe to presume that latero-frontal cilia of any kind are absentin Gastropods, and that these ciliary structures have arisenin, and are characteristic of the Lamellibranchia. It is possible

388 D. ATKINS

that the appearance of these structures in the Lamellibranchiamay be connected in some way with the loss of the radula; itis hoped to go into this in a later paper.

THE TAXONOMIC VALUE OF THE LATERO-FRONTAL

CILIATED TRACTS.

Gill structure as a means of classification has been criticizedin that it is liable to progressive modification, and thus relianceon it tends to produce 'horizontal' rather than 'vertical'divisions (Douville, 1912a; Davies, 1933). The compositionof the latero-frontal ciliated tracts seems to be a more or lessstable character, in fact, except for the Ostreidae, it was foundthat bivalves (117 species representing 51 families investigated)have one or other of two types, namely: (a) eu-, together withpro-latero-frontal cilia; or (b) micro-latero-frontal cilia alone.With the exception of the Ostreidae, no bivalve has been foundin which the latero-frontal ciliated tracts cannot be easilydetermined to belong to one or other of these types. Thischaracter is therefore not open to the above-mentioned objec-tion. But while the composition of the latero-frontal ciliatedtracts may be used to separate a large group with micro-latero-frontal cilia from a much larger group with a row of eu- and arow of pro-latero-frontal cilia on each side of the frontal surface,it is useless, so far as my knowledge goes, for the determinationof smaller divisions: this, however, may not prove to be so inthe group characterized by the possession of micro-latero-frontalcilia when more is known of the structure of these cilia.

In the following discussion it is claimed, or assumed, that thecharacter of the latero-frontal cilia affords a real clue to thebroad affinities of bivalves, and offers a test of existing classi-fications.

EEVIBW OF PELSENEEB'S AND EIDEWOOD'S CLASSIFICATIONS

OF THE LAMELLIBRANCHIA.

The two classifications of the Lamellibranchia best knownto zoologists are perhaps those of Pelseneer (1906) and Bide-wood (1903), based largely on gill structure; that of Pelseneerbeing the more widely accepted. Pelseneer, in 1911, returned


to his older classifications of 1889,1891, and 1892 retaining theOrder Pseudolamellibranchia, including the Aviculacea with thePinnidae and Ostreidae, and the Pectinacea with the Limidae.

The two classifications are given briefly on p. 390, in relationto, and so far as they affect the present work. The compositionof the latero-frontal ciliated tracts in the various families isindicated, so that it may be seen at a glance how the differenttypes are dispersed through the sub-orders and orders of theseclassifications.

In Eidewood's classification, forms previously considered alliedare widely separated, and dissimilar forms associated. He(p. 184), himself, stated that 'It is not claimed that the schemeof classification set forth in the following pages represents thegenetic affinities of the forms included; but while disinclinedto inflict upon a long-suffering world of zoologists a new classi-fication of the Lamellibranchia in which I myself have no greatconfidence, I have, for reasons similar to those stated in aprevious paper, arrived at the conclusion that, for purposes ofready reference, a key to the species examined based on theparticular feature under consideration is not only justifiable,but even useful.'

Examining Eidewood's classification in some detail we findthat in the Protobranchia the two families, Nuculidae (in whichhe included the Nuculanidae) and Solenomyidae, have eu-latero-frontal cilia, most probably also with pro-latero-frontalcilia at least in the Nuculidae (see p. 356).

His order Eleutherorhabda can be shown to be not less thandiphyletic, some families possessing eu- together with pro-latero-frontal cilia, and others micro-latero-frontal cilia only.He defined the sub-orders as follows: Dimyacea; gill lamellaeflat and homorhabdic, with no ascending filaments: Mytilacea;gill lamellae flat and homorhabdic, with ascending filaments:Pectinacea; gill lamellae plicate and heterorhabdie, with ascend-ing filaments.

In the Dimyacea he placed Dimya a r g e n t e a andA n o m i a a c u l e a t a ( = H e t e r a n o m i a squamula) . Het-e r a n o m i a has micro-latero-frontal cilia: whatever the formof the latero-frontal cilia of Dimya—which I have not had

EID

BW

OO

D'S

A

ND

PE

LS

EN

EE

R'S

C

LA

SS

IFIC

AT

ION

S

OF

TH

E L

AM

EL

LIB

EA

NC

HIA

.

Eid

ewoo

d (1

903)

.O

rder

1.

PEOT

OBEA

NCHI

A.

Nuc

ulid

ae,2 S

olen

omyi

dae.

2

Ord

er 2

. E

LE

UT

HE

BO

EH

AB

DA

.Su

b-or

der

1. D

imya

cea:

D

imy

a,5

An

om

ia

ac

u-

lea

ta (

= H

ete

ran

om

ia

squ

amu

la).

4

2. M

ytila

cea:

Ano

miid

ae (

excl

udin

g A

no

mia

acu

leat

a),4 A

rcid

ae,4 T

rigo

niid

ae,1 M

y-ti

lida

e,1 M

elin

idae

,4 Am

ussi

idae

.4

3. P

ectin

aeea

: Sp

ondy

lidae

,4 Pec

tinid

ae,4 A

vie-

ulid

ae.4

Ord

er 3

. SY

NA

PTO

BH

AB

DA

.Su

b-or

der

1. O

stre

acea

: Pi

nnid

ae,4 L

imid

ae,4 O

stre

idae

.3

2.

Subm

ytila

cea.

1

3.

Tel

linac

ea.1

4.

Ven

erac

ea.1

5.

Car

diac

ea.1

6. M

yace

a.1

7.

Phol

adae

ea.1

8. A

natin

acea

.1

9. P

orom

yace

a.5

Pels

enee

r (1

906,

mod

ified

191

1).

Ord

er 1

. PR

OTOB

BANC

HIA.

So

leno

myi

dae,

2 Nuc

ulid

ae,2

Led

idae

.2

Ord

er 2

. FI

UB

RA

SCTT

TA .

Sub-

orde

r 1.

Ano

mia

cea:

Ano

miid

ae.4

2.

Arc

acea

: A

rcid

ae,4 P

ectu

ncul

idae

,4 Phi

lo-

bryi

dae,

5 Tri

goni

idae

.1

3.

Myt

ilace

a: M

ytili

dae.

1

4.

Dim

yace

a: D

imyi

dae.

5

Ord

er.

3.

PSET

TOO

LAM

ELLI

BR

AN

CH

IA.

Sub-

orde

r 1.

Avi

cula

cea:

Avi

culid

ae,4 V

ulse

llida

e,4 P

ern-

idae

,4 Pin

nida

e,4 O

stre

idae

.3

2.

Pect

inae

ea:

Pect

inid

ae,4

Am

ussi

idae

,4

Spon

dylid

ae,4 L

imid

ae.4

Ord

er 4

. E

UL

AM

EL

UB

KA

NC

HIA

.Su

b-or

der

1. S

ubm

ytila

cea.

1

2.

TeU

inac

ea.1

3.

Ven

erac

ea.1

4.

Car

diac

ea.1

5.

Cha

mac

ea.5

6. M

yace

a.1

7.

Ade

smac

ea.1

8. A

natin

acea

.1

Ord

er 5

. SE

PTT

BE

AN

CH

IA.5

6 S izi

03

1 Eu-

tog

ethe

r w

ith p

ro-l

ater

o-fr

onta

l ci

lia (

but

see

p. 3

68 )

. 2 E

u- t

oget

her

with

pro

-lat

ero-

fron

tal

cilia

appa

rent

ly p

rese

nt (

see

p.

356)

. 3 A

nom

alou

s to

geth

er w

ith

para

-lat

ero-

fron

tal

cilia

. 4 M

icro

-late

ro-

fron

tal

cilia

onl

y.5 N

o m

embe

rs e

xam

ined

.

THE CILIARY MECHANISMS OP LAMELHBRANCHS 391

an opportunity of examining—there seems no justification forseparating Heteranomia from the rest of the Anomiidae.

Of the six families placed in the Mytilacea, the Anomiidae.Arcidae, and Melinidae have micro-latero-frontal cilia, and theirpresence is inferred in the Amussiidae owing to the closesimilarity of the gills of Amussium pleuronectes to thoseof the Peetinidae in which these cilia occur, although latero-frontal cilia could not be distinguished because of poor preserva-tion (Atkins, 1938).

In his family Melinidae Bidewood placed the genera Melina( = Perna = Isognomon) and Malleus. Of these twogenera, miero-latero-frontal cilia have been distinguished inMalleus albus (mi-l-j.c, Text-fig. 3 B) (see, however, p. 363):the gills of Isognomon a la ta and Isognomon isogno-mon were not sufficiently well preserved to distinguish micro-latero-frontal cilia, but eu-latero-frontals were almost certainlyabsent. It might here be observed that Pelseneer in his classi-fication of 1906 placed Malleus in the Aviculidae, but followedEidewood in assigning P e r n a (Pernidae) to the Mytilacea: in1911 he replaced this latter family in the Aviculacea.

In the Amussiidae Eidewood included Pl ica tu la aus-t r a l i s , which Pelseneer (1906) placed in the Spondylidae, andWatson (1930) concluded should either be a distinct family inthe Pectinacea, or the Peetinidae should be divided into not lessthan four sub-families—the Amussiinae, Plicatulinae, Pectininae,and Spondylinae. I have been unable to obtain material ofP l ica tu la .

There can be no doubt but that Eidewood was wrong inincluding the Amussiidae in the Mytilacea even on a considera-tion of the form of the gill. It has been shown in a previous paper(Atkins, 1938) that the gills of Amussium pleuronectesare plicate and heterorhabdic, and in fact very closely resemblethose of certain of the Peetinidae, and unless it is to be separatedfrom members of the family with flat and homorhabdic gills(Amussium dall i , Amussium meridionale , Amus-sium lucidum), then the natural position of the Amussiidae,from a consideration of general anatomy, is near the Peetinidaein the sub-order Pectinacea.

392 D. ATKINS

The Amussiidae are not alone among the Eleutherorhabda incontaining some members with flat and homorhabdic, andothers with plicate and heterorhabdic gills. In the Pectinidaea species, Pecten groenlandicus , with flat and homo-rhabdic gills is known (Noman, 1882). In fact in this familythere is considerable variation in the structure of the gills (seeText-fig. 12, p. 414). The simplest is P e c ten groenlandicuswith flat and homorhabdic lamellae, lacking interlamellarsepta, and with a single row of ciliated discs, those at the loweredge of the demibranch; most known species appear to haveplicate and heterorhabdic lamellae, with interlamellar septa,and with a number of ciliary interfilamentar junctions as inPecten maximus, Pecten i r r ad ians , Chlamysopercular is , Ohlamys d i s to r t a , Chlamys t iger-in a, and Chlamys v i t rea ; the most highly developedwould seem to be Pecten t enu icos ta tus Mighels (= P e c -ten grand is Solander) in which organic interfilamentar junc-tions occur, though ciliary ones are present near the free marginof the demibranch (Drew, 1906, 1907 a). In the Pteriidae,Stempellar ia magellanica (see Clasing, 1921) andMalleus albus, 1 have flat and homorhabdic gills, whilePter ia and P inc tada have plicate and heterorhabdic gills.In the Vulsellidae, within the same genus, one species Vulsellarugosa, has flat gills, and another, Vulsella l ingulata,plicate gills, (see Pelseneer, 1911). It will thus be seen that thedifference between flat and plicate gills does not warrant theseparation of related species and genera, as Pelseneer recognizedin 1911.

Of the sub-order Mytilacea there now remains the Mytilidaeand Trigoniidae to be considered. These two families alone ofEidewood's sub-order have eu-latero-frontal cilia, together withpro-latero-frontals in the Mytilidae (the filaments of Trigoniamargar i tacea were not well enough preserved to makeidentification of pro-latero-frontals possible) and most probably

1 There seems some doubt as to the family in which Malleus shouldbe placed. Jackson (1890) considered it as closely allied to Avicu la ,and provisionally placed Vulsel la (Vulsellidae), and Malleus on thesame branch from Av icu l a : Pelseneer (1906, 1911) placed Malleusin the Aviculidae itself.


belong to a distinct phylogenetic group, or groups, from thatof the rest of the families in the sub-order. It seems very im-probable that there is any close relationship between theMytilidae and the Trigoniidae, for the ciliation of the filamentsis markedly dissimilar (see p. 381; Text-fig. 1, A and c,p. 351). The position of the Trigoniidae will be discussed later(p. 394).

Little need be said of the three families, Spondylidae, Pectin-idae, and Aviculidae, placed by Bidewood in the sub-orderPectinacea: all three possess micro-latero-frontal cilia.

Eidewood (1903, p. 181) stated his reasons for the separationof forms previously classed in the Pectinaeea and their inclusionin the Mytilacea as follows: 'Of the genera of the Pseudo-lamellibranchia which it is now proposed to associate with theFilibranchia under the title Eleutherorhabda, some have well-developed principal filaments and plicate lamellae (e.g. Avi-cu la , Meleagr ina , P e c t e n , Spondylus) , whereasothers have flat lamellae and filaments undifferentiated (e.g.P l i c a t u l a , Mal leus , Mel ina , Amuss ium, andDimya). So sharply marked off are these first four generafrom all the rest of the Eleutherorhabda by reason of theirstrongly differentiated principal filaments and their plicatelamellae, that it is well to let them constitute a sub-order bythemselves.' It is difficult to follow why Eidewood should haveascribed such importance to the difference between flat homo-rhabdic, and plicate heterorhabdic lamellae in the Eleuthero-rhabda, while considering (p. 161) that in the Synaptorhabdasuch a difference was of not more than specific, or at most sub-generic, value; and after drawing attention to the fact thatprincipal filaments are ontogenetically a secondary differentia-tion, in P e c t e n for example (p. 162). Pelseneer in 1911 justlycriticized Eidewood's use of this character, and also his use ofthe structure of the interfilamentar junctions, whether ciliaryor vascular, to divide the Pseudolamellibranchia among theEleutherorhabda and Synaptorhabda.

It is necessary to consider Ridewood's sub-order Ostreacea,which he placed in the order Synaptorhabda along with seven sub-orders now known to possess eu- together with pro-latero-frontal

394 D. ATKINS

cilia, and the sub-order, Poromyacea, which included bivalvesgenerally classed in a separate order Septibranehia. Ofthe three families of the sub-order Ostreacea, the Limidaeand Pinnidae possess micro-latero-frontal cilia, the Pinnidaepossibly with an additional row (see p. 363), while the Ostreidaehave anomalous, together with para-latero-frontal cilia. Thusthe composition of the latero-frontal tract indicates that theSynaptorhabda, as the Eleutherorhabda, are at least diphyletic,and the Ostreacea should be removed from the Synaptorhabdaas Pelseneer contended in 1911.

We have now to see how Pelseneer's classification is affectedby a consideration of the composition of the latero-frontalciliated tracts, and it may be said at once that it is only in theFilibranchia and Pseudolamellibrancbia that any change issuggested.

In the Protobranchia the three families Solenomyidae,Nuculidae, and Ledidae ( = Nuculanidae) all have eu-latero-frontal cilia, most probably with pro-latero-frontal cilia also,as found in the Nuculidae (see p. 356).

His order Filibranchia, as Eidewood's order Eleutherorhabda,is almost certainly diphyletic, if not triphyletic. The Anomiidae,Arcidae, and Pectunculidae have micro-latero-frontal cilia; theTrigoniidae and Mytilidae eu-latero-frontal cilia, together withpro-latero-frontals certainly in the Mytilidae. Pelseneer (1889,1891, 1892, 1906) derived the Trigoniidae from the Arcidae andplaced the two families together in the sub-order Arcacea. Thedifference in the composition of the latero-frontal tracts, as wellas the arrangement of the various ciliary tracts of the filaments,in the two families, indicate that they belong to distinct phylo-genetic groups, though at about the same level of gill and mantleevolution. From a consideration of the different arrangementof the ciliary tracts in the Trigoniidae and Mytilidae it is un-likely, as previously mentioned, that these two families are atall closely related. The position of the Trigoniidae is of muchinterest, and in this connexion a peculiarity of the filaments ofT r igon i a m a r g a r i t a c e a may be mentioned here; that is,the presence of calcified rods1 embedded in the chitinous sup-

1 These were obvious in entire filaments of alcohol-preserved material,


porting structure (c.r.) (Text-fig. 1 c, p. 351). The presence ofthese rods in Trigonia is curious, as, according to Bidewood(1903, p. 168), calcified rods are peculiar to the filaments of theUnionidae and Miilleria. They have been described inAnodonta, Unio, Miilleria, and Monocondylaea:from the appearance of sections of the filaments of Mutelabourguigna t i (Text-fig. 6 B, p. 371) it is evident that theyare also present in that species. Calcified rods are thus charac-teristic of a number of the Naiadacea: they appear, however, tobe absent in Aetheria; Eidewood (p. 230) did not mentionthem in his description of the gills of Aetheria plumbea,and there is no indication of them in the filaments of Aetheriae l l ip t ica . Eidewood may have overlooked their presence inTrigonia lamarki , unless this is a character which variesin related species.

Neumayr (see Douville, 1912 a) concluded from a study ofthe hinge that the Unionidae and Trigoniidae were closelyrelated. Douville (1912 a, p. 448), however, has pointed out thatthe similarity between the hinges of Unio and Trigonia isdelusive, for the tooth of Trigonia is not originally doubleas that of the Unios, but is derived by division from a primi-tively simple tooth. Douville, himself, derived the Trigoniidaefrom the Preheterodonts by way of the Myophoriidae, and con-sidered that they gave rise to no higher forms. Prashad's(1931, p. 48) objection to the association of the Unionidae (orsuper-family of the Naiadacea) and the Trigoniidae in the sameorder, Schizodonta, is based on the difference in the structureof the gills in the two families, the former being eulamelli-branchiate and the latter filibranehiate. The difference asbetween the two types of gills seems to me not of great impor-tance in view of the occurrence of both ciliary and organic

owing to their having been fractured at intervals. The two main rods occurtoward the interlamellar ends of the filaments and are continuous through-out the length of the filament. Two other, but much smaller, rods aresituated beneath the lateral ciliated cells; they appear to be only inter-mittently present. Material prepared for sectioning was soaked for ashort time in weak hydrochloric acid alcohol; the part of the rods remainingafter the treatment stains intensely with iron haematoxylin, as indicatedin Text-fig. 1 c (p. 351).

396 D. ATKINS

interfilamentar junctions in the same individual in Pec t e nt e n u i c o s t a t u s , of compound ciliary and organic junctionsin certain species of Pteriidae, Limidae, and Pinnidae (seep. 406), and of the apparent ease with which extensive inter-filamentar and interlamellar junctions are produced in abnormalgills of the filibranchiate species, M y t i l u s e d u l i s (Atkins,1930). A greater difficulty is the arrangement of the ciliatedtracts of the filaments, for this is likely to be a more stablecharacter, few variations being known (see p. 380). Though thethree families of the Naiadacea, the Unionidae, Aetheriidae, andMutelidae, have eu-latero-frontal cilia as have the Trigoniidae,yet the arrangement of the various tracts of cilia is that commonto the Mytilidae and the Eulamellibranchs possessing eu-latero-frontal cilia (Text-figs. 1 A, B, p. 351; 6, p. 371), while thearrangement in T r i g o n i a m a r g a r i t a c e a is entirely dif-ferent (Text-fig. 1 c). Although this would not be an insuperabledifficulty in the way of the derivation of the Unionidae from theTrigoniidae, or from a common ancestor, it would seem toindicate that the two families are not closely related, in thisagreeing with the conclusion reached by Douville. The calcifiedrods of the two families may quite possibly have been developedindependently. It must not be overlooked, however, that thereis nearly as great a difference in the arrangement of the ciliarytracts in two families of Protobranchs, the Nuculidae, andNuculanidae, which at one time were classed as one family,as in the Trigoniidae and Unionidae.

In Pelseneer's order Pseudolamellibranchia are included anumber of families with micro-latero-frontal cilia, and one, theOstreidae, with anomalous together with para-latero-frontalcilia. My material representing the families Vulsellidae andPernidae ( = Isognomonidae) was not sufficiently well preservedto allow of the identification of micro-latero-frontal cilia, buteu-latero-frontals are almost certainly absent; it may perhaps beassumed that the former kind of cilia are present (see p. 376).Micro-latero-frontal cilia were identified in Ma l l eus a l b u s ,but there seems some uncertainty as to whether Ma l l eus isa Vulsellid or a Pteriid (see p. 392). The presence of micro-latero-frontal cilia is assumed in the Amussiidae, owing to the

THE CILIABY MECHANISMS OF LAMEIiMBRANCHS 397

close similarity of the gills of Amussium pleuronectesto those of the Peetinidae (Atkins, 1938).

The families of the Pseudolamellibranchia are closely related,and, as will be seen later, in all probability form with theArcidae and Anormidae a monophyletic group.

In his classification of 1906 Pelseneer placed the Avieulidae,Vulsellidae, Amussiidae, Spondylidae, and Peetinidae (andcertain extinct families) together in the sub-order Pectinacea.The last three families mentioned are more closely related toone another than they are to the first two, and this he recognizedin 1911 in creating a sub-order Aviculaeea for the first twofamilies together with the Isognomonidae and Pinnidae, andwith the Ostreidae as an offshoot, and placing the last threetogether with the Limidae in the Pectinaeea.

Pelseneer's view of the close relationship of the families of thePectinacea is borne out by work on the gills. Apart from thecommon possession of micro-latero-frontal cilia the members ofthe Amussiidae, Peetinidae, Spondylidae, and Limidae in-vestigated (which were all forms with plicate and heterorhabdiclamellae) agree very closely in the structure of the gills. Theyhave: (1) a suspensory membrane to the gill; (2) a similararrangement of the muscles of the gill axis; (3) a similar arrange-ment of the horizontal muscles of the principal filaments; (4)few or no vertical muscles in the lamellae; (5) respiratory ex-pansions on the principal filaments; and (6) chitinous supportingstructure of the principal filaments of similar form in all exceptthe Limidae. In the forms with flat and homorhabdic gills,for example Amussium dal l i , Amussium lucidum,Amussium meridionale , and P l ica tu la aus t r a l i s ,respiratory expansions are apparently absent.

Pelseneer's order Eulamellibranchia contains only families,which, so far as my work has gone, possess eu- together withpro-latero-frontal cilia. Palaeontologists, however, consider thatthis order is not a monophyletie one, and Douville (1912 a) hasdivided it into two groups, a' normal' and a' burrowing' branch.The possession of the same type of latero-frontal tract by bothgroups may be explained by their probable origin, the 'normal'branch from a nuculoid form, and the 'burrowing' branch from

398 D. ATKINS

a form of which Solenomya may be considered representa-tive (Davies, 1983): both Nucula and Solenomya haveeu-latero-frontal cilia (Text-fig. 7, p. 372), and, as it has beenseen (p. 356), Nucu la , if not So lenomya , probably haspro-latero-frontal cilia as well. In the phylogenetic division ofthe Eulamellibranchia the form of the latero-frontal cilia istherefore apparently of no assistance. It is curious that acertain type of pattern of the lateral ciliated cells is commonamong both ' normal' and' burrowing' groups of Lamellibranehs(Atkins, 1938 a). Differences in the detailed arrangement ofthese cells would seem to be of doubtful utility, and the samepattern occurring in different bivalves might be expected toindicate relationship. This character may repay a fuller investi-gation than I was able to give to it (1938 a). It is perhapspossible that the ' normal' and ' burrowing' groups of Lamelli-branehs are not as widely separated as suggested by Douville'stable (1912a, p. 466).

BEIIATIONSHIP OF LAMBLLIBBANCHS WITH MICRO-LATERO-

FEONTAL CILIA.

Bivalves with micro-latero-frontal cilia, those previously con-sidered as lacking such cilia, are not scattered haphazard throughthe Lamellibranchia, but are found in families which are moreor less closely related (see Text-fig. 9), forming, except for myinclusion of the Arcidae, a group designated ' the Aviculidae andtheir allies' by Jackson (1890). This is the same group asDouville's (1912a) 'sedentary' branch, except for his inclusionof the Mytilidae. That a group established by palaeontologistsalmost entirely on shell characters should be related also by thecondition of a certain tract of cilia on the gill filaments is mostinteresting.

Jackson (1890, p. 362) found the Aviculidae and their alliesto be linked in one great group by characters of anatomy andshell structure which connect the several members, and to becharacterized 'by possessing p rod i s soconchs of homo-geneous l a m i n a r s t r u c t u r e , bu t no t p r i s m a t i c ,and wi th umbos d i r ec t ed p o s t e r i o r l y ' . 1 The sue-

1 Professor A. Morley Davies informs me by letter that '"umbos


ceeding dissoconch has 'an external layer of prismatic cellulartissue which is more or less developed but exists at least in theearly nepionic stages of one valve' (1890, p. 377).

Douville (1912a, p. 460) characterized his 'sedentary' branchas ' essentiellement byssiferes'. He considered that fixation bythe byssus has impressed certain characters on the group whichtended to persist through various changes in the mode of life.This method of fixation may be accomplished in very differentways according to the length of the byssus. When it is so shortthat fixation is practically by the foot, as in A r e a , it resultsin an inequilateral shell, the posterior region, where the exhalentand main inhalent apertures are situated, being more developedthan the anterior, but with little or no difference in the size ofthe adductor muscles. As the byssus lengthens pressure isexerted by the foot on the anterior adductor, and as the pressureincreases this muscle diminishes in size, as in the Pinnidae, andfinally disappears (many forms of the group). The reduction ofthe anterior adductor is accompanied by that of the anteriorregion generally.

Byssal fixation and cementation, however, are not confinedto the 'sedentary' branch, and in two families at least, theAetheriidae and Tridacnidae, may be accompanied by themonomyarian condition. According to Yonge (1936), however,in the Tridacnidae this condition was brought about in anentirely different way—by rotation of the mantle and shellround the visceral mass—from that in the 'sedentary' branch.Douville, of course, included the Mytilidae, in which byssalfixation has resulted in reduction and disappearance of theanterior adductor muscle in some species, in his 'sedentary'group: the position of the Mytilidae will be discussed later.

The group (Arcidae, Anomiidae, Pteriacea with Pinnidae,Ostreidae, Pectinacea with Limidae) characterized by the pos-session of micro-latero-frontal cilia would appear in all pro-bability to be a monophyletic one (Text-fig. 9). As previously

directed posteriorly" is not, as it might seem at first sight, an embryonicor juvenile character, but is an effect of later unequal growth. Possiblyits general presence in "the Aviculidae and their allies" may be connectedwith the reduction of the anterior region of the shell'.

OST

REA

CEA

PT

ER

IAC

EA

PE

CT

II^

AC

EA

wIS

OG

N0

M0

NID

AE

PT

ER

IID

AE

s tl

MID

AE

PE

CTI

NIO

AE

OS

TRE

IDA

E

VUL5

ELLI

DA

E

PIN

N1D

AE

SP

OND

YLID

AE

AM

USS

IIDA

E

RE

CE

NT

ME

OS

EN

E

PALA

EOG

ENE

CR

ETA

CEO

US

JUR

AS

SIC

TRIA

SS

IC

PE

RM

IAN

CARB

ONIFE

ROUS

DE

VO

NIA

N

SIL

UR

IAN

OR

DO

VIC

IAN

>

GU

jcjjm

eris

Are

a

^ y

V y

Vulse

Ua

Isog

nom

on

Ost

rea

x , *-

^

^

--

--

Vla

Ueu

s.'

5in

cta

da

Pbe

ria

/

Pin

na

.

Spon

dylus

Lim

a

El DAE

£'"

UJ

Pec

ten

—-—

iAm

us

siu

m

ChL

amys

lete

raho

mia

Mon

ia

'"

'P

lacu

na

Anom

ia

TE

XT

-EIG

. 9.

Geo

logi

cal o

ccur

renc

e an

d ph

ylog

enet

ic in

ter-

rela

tions

hips

of t

he g

roup

of L

ameU

ibra

nchs

cha

ract

eriz

ed b

y th

e po

sses

sion

of m

icro

-lat

ero-

fron

tal

cilia

(on

ly f

amili

es a

nd g

ener

a ex

amin

ed i

nclu

ded)

. G

eolo

gica

l occ

urre

nce

obta

ined

lar

gely

from

Dal

l (i

n Z

ittel

, 19

13);

phy

loge

netic

rel

atio

nshi

ps l

arge

ly f

rom

Jac

kson

, 18

90.

(Arc

idae

is

used

in

a w

ide

sens

eto

inc

lude

Gly

eym

eris

an

d al

lied

gene

ra,

and

Pa

rall

elo

do

n an

d al

lied

gene

ra).

THE CILIAEY MECHANISMS OF LAMELMBRANCHS 401

stated, it differs from Jackson's group in the inclusion of theArcidae, but this ancient family appears from palaeontologicalevidence to be connected with the Pteriacea, for according toArkell (1930, pp. 306-7) it can scarcely be doubted that C y p r i -c a r d i t e s ( = P a l a e a r c a ) , which ranges from the Ordovicianto the Devonian, and P a r a l l e l o d o n , a genus of Arcidae,were descended from a common stock. C y p r i c a r d i t e s wasplaced in the Arcacea by Dall (in Zittel, Eastman edition,1913, vol. i reprinted in 1927 and 1937), but Zittel (Germanedition, 1924, see Arkell) classed it with the Ambonyehiidae,a family of the Pteriacea. Douville (1912 a) associated it withthe Pterineidae. He held that the ancient forms of the Arcidaeapproach much more to the Pterineidae than to the Nueulidae,but thought that the Arcidae really constitute a branch parallelto that of the Aviculidae (=Pteriidae), but fixed only by thefoot, and in consequence with a much less great reduction ofthe anterior adductor muscle, and of the anterior portion of theanimal generally. In the living genus L i m o p s i s (sometimesclassed with the Arcidae, sometimes in a separate family, andwhich according to Pelseneer (1911, pp. 8, 9) should, withP e e t u n c u l u s ( = Glycymer i s ) , be placed in a familyPectunculidae, apart from the Arcidae), however, the anterioradductor muscle is frequently much reduced, and in an alliedfamily, the Philobryidae, is wanting (Pelseneer, 1906, p. 258).

Palaeontologists differ as to the position of the Mytilidae.This family was included by Douville in his ' sedentary' branch,but omitted by Jackson from 'the Aviculidae and their allies'.Douville (1912 a, p. 460) distinguished two great groups in the' sedentary' branch: (1) the Pterineidae with straight hinge, andamphidetic ligament, at first simple and inserted on a moreor less developed area; and (2) the Mytilidae with curved hinge,opisthodetic ligament, remaining marginal, and without anarea. Jackson (1890, p. 364), on the other hand, stated that'The striking differences in the prodissoconch and nepionicstages of the Mytilidae and Aviculidae are sufficient, I think, toseparate these groups, and the Mytilidae should be put in agroup distinct from the Avieulidae and their allies, which I haveshown are all bound together by important features as one

NO. 319 D d

402 D. ATKINS

group'. His conclusions are supported by the difference in thecomposition of the latero-frontal tracts in the Mytilidae and inthe Pteriidae and their allies (see pp. 368-380).

Comparing the group with micro-latero-frontal cilia withPelseneer's classification of 1911 it will be seen that this groupcorresponds to his order Pseudolamellibranchia (see p. 390)together with the Arcidae and Anomiidae, which last twofamilies he associated with the Mytilidae and Trigoniidae in theFilibranchia. As previously stated, the Arcidae and Anomiidaehave a type of latero-frontal tract quite distinct from that ofthe other two families of the Filibranchia, with which theyappear to have no close relationship.

Pelseneer abandoned the order Pseudolamellibranchia in1906, but in 1911 (p. 119) he stated: 'je suis ramene a mon ideeancienne et premiere qu'il faut reconstituer un groupe entreles Filibranches et les Eulamellibranches—groupe representantun stade phylogenetique posterieur au premier et formant unebranche, globalement moins specialised que le second et orienteedans une autre direction.' He (1911, pp. 120-1) defined it asfollows: (1) the gills have ciliary or cellular interfilamentarjunctions (or both) and are free or attached to the mantle byciliary junctions, but (2) the pallial lobes are without suture(whilst all Eulamellibranchs have always non-ciliary and vas-cular interfilamentar and interlamellar junctions, and alwaysone or several pallial sutures; (3) the auricles inter-communicate;(4) the presence of abdominal sense organs1 on the posterioradductor muscle; (5) the anterior adductor muscle is wanting;(6) the byssus is normally well developed.

Pelseneer's order Pseudolamellibranchia is not sufficientlyinclusive, for the present work indicates that the Arcidae andAnomiidae are related to the families forming that order.Palaeontological evidence for the inclusion of the Arcidae hasbeen briefly given (p. 399); it remains to consider the Anomiidae.

1 Paired abdominal sense organs are present on, or near, the posterioradductor muscle in the Arcidae, whilst a single organ is present on theright side only in P lacuna p l a c e n t a (Anomiidae), as in Pec ten(Willey, 1911, pp. 158-61). Paired abdominal sense organs also occur inthe Trigoniidae (Pelseneer, 1906) and in the Mytilidae (Field, 1922).


The present research certainly does not support Pelseneer's(1911, p. 16) contention—in opposition to Jackson, Bernard,Bice, and Stenta—that the affinities of An o mi a are with theFilibranchia and especially with the Mytilidae. Althoughthe form of the latero-frontal cilia cannot help in solving theproblem of the immediate ancestry of the Anomiidae, whetherthey are derived from the Pectens, as Jackson supposed, or fromsome other member of the group, it at least indicates that theyare members of' the Aviculidae and their allies'. Jackson (1890,p. 362) thought it possible that A n o m i a was derived fromthe A m u s s i u m or H e m i p e c t e n group of the Pectens, asthese resemble Anomia in having thin nacreous shells. Ac-cording to DouviUe (1912 a) the nacreous type of shell precedesthe porcelaneous in evolution, and forms derived from por-celaneous types are always porcelaneous; it therefore seems thatA n o m i a , whatever its ancestry may prove to be, must havebeen derived from forms "with nacreous shells. The Anomiidaeare known certainly from the Jurassic: there is a record ofAnomia in the Devonian, but this occurrence seems doubtfulas there is no other record before the Jurassic.

To summarize the various views considered here: Jacksonsaw in 'the Aviculidae and their allies' a natural group fromwhich the Mytilidae are clearly separated; he did not includethe Areidae. For Douville his 'sedentary' branch is 'essen-tiellement byssiferes'. It contains not only the forms includedin Jackson's group, but also the Areidae (which he consideredreally constitute a branch parallel to that of the Pteriidae) andthe Mytilidae. He distinguished two great groups in the' sedentary' branch, the Pterineidae and the Mytilidae, separatedby certain characters of the hinge. For Pelseneer his orderPseudolamellibranchia is a group between the Filibranchia andthe Eulamellibranchia, though not ancestral to the latter, itsmembers connected by a number of characters in common. Itcorresponds to Jackson's group less the Anomiidae, whichfamily Pelseneer contended has affinities, not with the Pseudo-lamellibranchia, but with the Mytilidae.

The group characterized by the possession of micro-latero-frontal cilia corresponds to Jackson's group of ' the Aviculidae

404 D- ATKINS

and their allies' with the addition of the Areidae; to Douville's'sedentary' branch less the Mytilidae; and to Pelseneer'sPseudolamellibranchia with the addition of the Areidae andAnomiidae. The sub-orders and families in the group are then(so far as my material goes) as follows:

Arcacea: Aroidae.Anomiacea: Anomiidae.Pteriacea: Pteriidae, Isognomonidae, Vulsellidae, Pinnidae.Pectinaeea: Pectinidae, Amussiidae, Spondylidae, Limidae.Ostreacea: Ostreidae.The geological occurrence and phylogenetic interrelationships

of the group are shown in Text-fig. 9, p. 400.

DISCUSSION ON THE OBIGIN OF THE GBOUP CHAEACTEBIZED BY

THE POSSESSION OF MICEO-LATEBO FBONTAL CILIA.

Eu-latero-frontal cilia are present in living representativesof all three existing families of the Protobranchia, together withpro-latero-frontal cilia at least in Nucula. The presence ofthese large straining cilia on gills which, certainly in Nuculaand Nuculana, play only a subsidiary part in feeding, pointsto the antiquity of eu-latero-frontal cilia, and incidentallysuggests that the gills of ancient Protobranchs may perhapshave played a greater part in feeding than do those of mostof their modern representatives. This has already been sug-gested for the Nuculanidae on the evidence of the complicatedfrontal currents present on the small gills of Nuculanaminuta (Atkins, 1936). It may very well be that in theProtobranchia, suspension feeding (as in S o 1 e n o m y a) is moreprimitive than deposit feeding (as in Nucula and Nucu-lana).

It has been generally considered that the Nuculidae,1 or

1 Schenck (1934) in his paper on 'Classification of Nuoulid Pelecypods'though not doubting that the family Nuculidae has Devonian representa-tives, yet insisted that he had seen no Nucu la , sensu stricto, in rocks ofPalaeozoic age, the Palaeozoic specimens he had studied not being closelyrelated to the type species of Nucu la , s.s.


nuculoid forms, are the stock from which sprang the higherLamellibranchs (Jackson, 1890; Pelseneer, 1891, 1911; Eice,1897, &c), with the exception of Douville's 'burrowing' branch,or Desmodonts, of which Davies (1933) suggested S o l e n o m y aas representative of the ancestral form. If this be so, and if eu-together with pro-latero-frontal cilia were present in primitiveNuculidae, as they are in their living representatives in thegenus N u c u l a , then it would seem to follow that from theNuculidae have arisen at least two lines, one in which this typeof tract has persisted (Mytilidae, Trigoniidae, Eulamellibranchia)and the other in which eu-latero-frontal cilia have been lost(Arcidae, Anomiidae, Pseudolamellibranchia), excluding thepossibility of their reduction. It seems very improbable, how-ever, that such useful structures in the ciliary method of feedingas eu-latero-frontal cilia once possessed would be lost, so longas the method of obtaining food remained the same, and thenan attempt made to regain them, as exemplified by the ano-malous latero-frontal cilia of the Ostreidae. This is the moreunlikely as the mainly deposit feeding Protobranchs, N u c u l aand N u c u l a n a , have retained eu-latero-frontal cilia, thoughhaving no great use for them. The fact that in the second line(with micro-latero-frontal cilia) the most highly developedlatero-frontal cilia—the anomalous—are found in the highestfamily of the group, the Ostreidae, points to the evolution ofthese structures within the group, and it would seem that thesimilarity both of structure and function between the anomalousand eu-latero-frontal cilia is due to convergence.

It may thus be conjectured that the remote ancestors of the'sedentary' branch are to be sought in forms other than theNuculidae, or any Protobranch family with living representa-tives, for these all have eu-, if not also pro-, latero-frontal cilia.Judging by the persistence of eu-latero-frontal cilia in theNuculanidae and Nuculidae on gills which play only a sub-sidiary part in feeding, as well as by the uniformity of composi-tion of the latero-frontal tract within existing families ofLamellibranchs, it seems improbable that extinct members of theNuculidae, Nuculanidae, and Solenomyidae differed much fromliving members in the form of their latero-frontal cilia. But it

406 T>. ATKINS

may be that some'ancient, extinct families of Protobranchs werewithout eu-latero-frontal cilia, and gave rise not only to formswith such cilia, but also to those with micro-latero-frontal cilia.Douville (1912 a, pp. 443-4) noted that three families ofPalaeoconehs, the Cardiolidae, with equivalve shells, the Anti-pleuridae and Vlastidae, with inequivalve shells, would be per-haps better placed in the 'sedentary' group, but he thought itprobable that these are not to be considered as primitive forms(p. 421): it is therefore likely that they are not sufficientlygeneralized to have given rise to the 'sedentary' branch, andbesides they are not known until the Silurian, while the Pteri-neidae occur in the Ordovician.

Thus, though the ' sedentary' branch of Lamellibranchs canbe traced back to the Pterineidae (Ordovician to Carboniferous)yet their origin and relationship with other groups of Lamelli-branchs remains obscure. It is an unfortunate truism that thetaxonomic value of any character drawn from the soft parts ofan animal is greatly restricted in that the position of fossilscannot be tested by such characters.

THE EVOLUTION OF THE EULAMELLIBRANCHIATE OE SYNAPTO-

BHABDIC GILL IN THE GROUP CHARACTERIZED BY THE

POSSESSION OF MICRO-LATERO-FRONTAL CILIA.

Within the group characterized by the possession of micro-latero-frontal cilia are two families, Pectinidae and Pteriidae,which, though characteristically filibranchiate or eleuthero-rhabdic, yet contain members with the synaptorhabdie ten-dency ; one of these, the Pectinidae, and the forerunners of theother, the Pterineidae—and probably the Pteriidae themselves—have given rise to forms with eulamellibranchiate or synapto-rhabdie gills (see Text-fig. 12, p. 414).

Most known members of the Pectinidae have gills with purelyciliary interfilamentar connexions, but the giant scallop Pec -t e n t e n u i c o s t a t u s Mighels has both organic and ciliaryjunctions, the latter occurring near the free margins—theyoungest parts—of the gills (Drew, 1906,1907 a)', where, as willbe seen later, in L i m a vestiges of ciliated discs are found.

THE CILIAEY MECHANISMS OF LAMELMBBANCHS 407

From the Pectinidae in Carboniferous times arose the eu-lamellibranchiate family Limidae (Jackson, 1890, pp. 388, 391).The gills of L ima h i a n s and L i m a l o s c o m b i are curiouslyPecten-like, the organic interfilamentar junctions having theappearance of spurs—such as are found in P e c t e n , S p o n d y -lu s , and Amussium—which have fused. In L i m a h i a n s(fig. 1, PI. 29) and L i m a l o s c o m b i remains of ciliary junctionspersist, being found mostly toward the ventral margins of thedemibranchs, but the junctions are mainly organic. Bidewood(1903, p. 217) noted that in L i m a i n f l a t a obsolete ciliateddiscs were merely suggested by the regularity of the prismaticepithelium.

An apparently unique use of interlocking cilia is found in theLimidae (Lima h i a n s and L i m a loscombi) , in which theinterfilamentar junctions do not run across the plicae as hori-zontal septa, as in P i n n a and O s t r e a , for instance. Theinterfilamentar junctions occur at regular and rather closeintervals; they arise from the side of the principal filament andpass across the abfrontal surface of the ordinary and apicalfilaments to the next principal filament. The junctions are notdirectly transverse; they run in a series of V's, the apex of eachbeing directed dorsally, and situated on an apical filament(Text-fig. 10). Long interlocking cilia, with the characteristicrotary movement of such cilia, are present on the intraplicalface (i. e. facing the exhalent chamber) of that part of the inter-filamentar connexion which extends across the filaments formingthe crest of the plica (in Text-fig. 10 the apical and three ad-jacent filaments on each side). Under normal conditions thetwo arms of the V interlock (see fig. 1, PI. 29), and hold thesides of the crests together, so that the plicae are steep-sided.Flattening of the plicae can only take place in that region devoidof interlocking cilia adjacent to the principal filaments, unlessthe cilia are torn apart, as shown in Text-fig. 10, which probablydoes not normally occur. Movement of the plicae is mostlya bending sideways of one fold toward the next, so that theypartly overlap (see Text-fig. 11): only to a slight extent doeswidening and smoothing out of the plicae occur. It wouldseem that these ciliary junctions act in the manner of organic

408 D. ATKINS

i.e.

p.fi— -

i.e.

VentraL

TEXT-FIG. 10.

Lima h i a n s . Sketch from life of the abfrontal surface of a plica..The plica has been opened out, so that the cilia normally inter-locking on opposite limbs of the V-shaped interfilamentar junc-tions, are shown in a non-interlocking position. (The number offilaments to a plica in Lima h ians varies between about 14 and19.) a.f., apical filament; g.c, gland-cells of abfrontal surface ofprincipal filament; i.e., interlocking cilia; p.f., principal filament.

3i

intraplical septa—such as occur in P i n n a and Ostrea—inpreserving the form of the plicae, and this method is a possibleprecursor of organic union, a step in the tendency to progressivefirmness of the gill. The lacunar tissue of an organic horizontal

THE CILIABY MECHANISMS OF LAMELLIBEANCHS 409

septum probably allows of greater play in increasing and de-creasing the depth of the plicae than is possible by the ciliarymethod, and is "without its weakness.

In the Pteriidae there are at least three species which showthe synaptorhabdic tendency. In the genus P te r i a , P t e r i ahirundo has ciliary interfilamentar junctions (an instance ofa principal and adjacent filament in organic union was observed,but this is unusual, and may have been due to proximity to thebase of the gill), but in P te r i a argentea the junctions areof a compound nature, the intraplieal edges of the spurs thatbear the ciliated discs having fused (Eidewood, 1903, pp. 212-13). A similar condition has been described in P inc tadavulgaris (Herdman and Hornell, 1904, p. 60; Herdman,1905, pp. 227-8), and has been found during the present workin P inc tada margar i t i fe ra .

The Pinnidae, which arose in Devonian times, probably fromLeptodesma, one of the Pterineidae (Jackson, 1890), hasliving members in which the gills have extensive organic inter-filamentar junctions, they are in fact eulamellibranchiate, yetsections of Pinna fragilis have shown that vestiges ofciliated discs still exist alongside the organic unions (fig. 2,PL 29), as in Lima hians and Lima loscombi.

In P te r i a hirundo the interfilamentar junctions areentirely ciliary; in P te r ia argentea , P inc tada vul-garis , and P inc tada margar i t i fe ra they are mainlyciliary with some little organic union; in Pinna fragilismainly organic with vestiges of ciliary junction. In the Pteri-idae and Pinnidae there are therefore species the gills of whichform a graded, though not a direct evolutionary, series.

A third eulamellibranchiate or synaptorhabdic family, theOstreidae, has arisen within the group, but is somewhat apartfrom the rest—possessing anomalous together with para-latero-frontal cilia—though it obviously should be included on accountof its derivation on independent grounds from either the Pteriidaeor Pectinidae. Its gills are more highly evolved than those ofeither of the other two families with eulamellibranchiate gills,Limidae and Pinnidae, not only in the degree of developmentof the latero-frontal ciliated tract, but in the presence of organic

410 D. ATKINS

h.m.p.F.

TEXT-PIG. 11.

Lima h i ans . Transverse section of three plicae and four principalfilaments of a lamella to show the bending sideways of the plicae.a.f., apical filament; ch,, darkly staining chitin; g.c, gland-ceD

[For remaining description see opposite.]

THE CILIARY MECHANISMS OF LAMBLLIBBANCHS 411

fusion between the dorsal edges of the ascending lamellae of theouter demibranchs and the mantle, and between those of theinner demibranchs in ter se; and in the absence of any ves-tiges of ciliary interfilamentar junctions, such as exist in Limaand Pinna (the first interfilamentar junctions of the gills ofthe spat of Ostrea virginica (Jackson, 1890, p. 304) andof Ostrea edulis (Yonge, 1926, p. 320) are organic). lathegroup under consideration the gills of Ostrea are the mostdefinitely eulamellibrancbiate in structure.

The immediate ancestors of the Ostreidae are doubtful; it isnot known with certainty whether they have arisen from thePteriidae or the Pectinidae. Jackson (1890, p. 307) thought itprobable that Ostrea was derived from the Pteriidae, sug-gesting that it was descended either directly from Pern a(= I s o g n o m o n) or a close common ancestor of the two genera,or from Avicula (=Pter ia) . Ostrea appeared in Car-boniferous times, Isognomon is not known until the Trias(Dall in Zittel, 1913, pp. 450, 447), so that Ostrea is unlikelyto have descended directly from Isognomon. The Pteriidaeare present in the Silurian, P te r i a itself in the Devonian; itis possible therefore for Ostrea to be descended either fromPte r i a , or some other member of the family. Dall (in Zittel,1913, p. 455) followed Jackson in deriving Ostrea from thePteriidae; Pelseneer (1911, p. 119) placed it as an offshoot fromthe Pteriacea: Davies (1933), on the other hand, considered thatOstrea shows by its hinge structure and muscle-plan itsgeneral affinity t oPee t en . In 1853 Forbes and Hanley (vol. ii,p. 261) placed Lima, Pee ten , Ostrea, and Anomia inthe Ostreidae.

Gill structure unfortunately does not afford any help indetermining the immediate ancestry of the Ostreidae. The gillaxes are free for much of their length, but differ from those of

of abfrontal surface of principal filament; h.m., horizontalmuscle-fibres of the interfilamentar junctions; h.m.p.f., horizontalmuscle-fibres of the principal filaments; i.e., interlocking cilia onintraplical faces of the interfilamentar junctions; m.g., mucousgland; p.ch., pale-staining chitin; p.f., principal filament. Bouin-Duboscq's fixative; Mallory's triple stain. X200.

412 D. ATKINS

the other two families with eulamellibranchiate gills, Limidaeand Pinnidae, in the slight development of the longitudinalmuscles. Vertical muscles in the demibranchs of O s t r e ae d u l i s are poorly developed: absence or paucity of suchmuscles is characteristic of the Pectinacea, including theLimidae, while the Pteriidae and Pinnidae usually have suchmuscles well developed. It is doubtful whether any relationshipof the Ostreidae to the Peetinacea can be considered as indicatedby the state of development of these muscles, for in Vu l se l l a sp . ,belonging to the family Vulsellidae of the Pteriacea, verticalmuscles are poorly developed, extending only a short distancefrom the gill axes into the principal filaments. The slightdevelopment of vertical fibres in V u l s e l l a may possibly becorrelated with the peculiarly sheltered life, embedded insponges, and in O s t r e a with a life of fixation from a veryearly age. Frontal currents similar to those of O s t r e a arefound on the gills of both Pec t e n and P t e r i a , in factopposed frontal currents on all lamellae, frequently on the samefilament, are common in the group with micro-latero-frontalcilia, the only known exception being P i n n a (Atkins, 1936,1937,1937a).

In the group with micro-latero-frontal cilia, eulamelli-branchiate gills, so far as is known, are always plicate andheterorhabdic, in contrast to those of the group with eu-latero-frontal cilia, which are frequently flat and homorhabdic. Itseems that in the former group the growth of organic junctionshas only occurred in gills which had already developed plications,while in the latter group it has occurred also in the flat-gilledstage. Plat homorhabdic and plicate heterorhabdic gills havebeen separated in Text-fig. 12, but though it seems possible,or probable, that the simpler condition generally precedes themore complex, yet the fact that the two conditions are foundnot only in the same family, but in the same genus (Pelseneer,1911), suggests that the step from the one condition to the otheris small. In the eulamellibranchiate genus Don ax, whichcontains species with flat, with slightly plicate, and with stronglyplicate lamellae, Eice, with what justification I do not know,regarded the simplicity of the flat forms as retrogressive (quoted

THE CILIARY MECHANISMS OF LAMELLIBKANCHS. 4 1 3

by Ridewood, 1903, p. 162): the possibility cannot be excludedthat in filibranehiate gills the flat condition may in some in-stances be secondary.

To sum up, in the group with micro-latero-frontal cilia somefamilies are still in the filibranehiate or eleutherorhabdic stage(Arcidae, Anomiidae, Pectinidae, Amussiidae, Spondylidae,Plicatulidae, Pteriidae, Vulsellidae, Isognomonidae), thoughshowing an early stage in the transition to the eulameUibranchiateor synaptorhabdic condition in the compound junctions ofPter ia a rgen tea , P inc tada vulgar is , and P inc tadamargar i t i fe ra , and considerable organic junction in P e c -ten t enu icos t a tus ; others have attained the eulameUi-branchiate stage (Limidae, Pinnidae, Ostreidae) but certain ofthem (Lima hians , Lima loscombi, Pinna fragilis)show in vestiges of ciliary junctions, signs of a passage througha filibranehiate stage (see Text-fig. 12).

In the group with micro-latero-frontal cilia there is clearevidence of the passage from the filibranehiate to the eulameUi-branchiate condition. In the group with eu-latero-frontal cilia nosuch evidence has yet been demonstrated, though it is probablethat existing EulameUibranchs had ancestors with filibranehiategills.

I think it is evident from the foregoing pages that giUs atabout the same stage of evolution occur in but distantly relatedgroups of LameUibranchs—that in the different lineages thereis the same tendency towards consolidation of the giUs—andtherefore that the terms Filibranchia and Eulamellibranchia,Eleutherorhabda and Synaptorhabda, should only be used asdescriptive of stages in gill evolution, and not logicaUy as thenames of orders. GiU structure is essentiaUy a progressivecharacter as already pointed out by DouviUe (1912 a) andDavies (1933). Pelseneer (1911) held that the evolution of thegill may be considered as symbolizing the phylogenetic evolutionof the LameUibranchia: on this DouviUe (1912 a, p. 423) com-mented : 'Les caracteres tires de cet organe sont bien eertainementdes caracteres evolutifs, mais . . . leur valeur phylogenique esttres douteuse; les differents rameaux evoluent en effet d'unemaniere analogue, et ils doivent presenter la meme succession

PROTO-BRANCH-IATEGILLS

FILIBRANCHIATE GILLS

FLAT AND HOMORHABDIC PLICATE AND HETERORHABDICInterf i tamentar junctions. Entirety :2.Ctfiarg with

EULAMELLIBRANCHIATEGILLS

PLICATE AND HETERORHABDICJnterftlmentarjuncbions

2. EnbireLu• ••biimuwi i1

vestiges

TEXT-FIG. 12.

Diagram showing the condition of gill evolution in the group ofLamellibranchs characterized by the possession of mioro-latero-frontal cilia. When families have living members with the gillsin a certain stage this is indicated by a thick line.


de caraeteres; ceux-ei ne peuvent done permettre de recon-stituer Ies rameaux.'

COMMON CHARACTERS OF LAMELLIBRANCHS POSSESSING

MlCRO-LATERO-FRONTAL ClLIA.

Lamellibranchs within the group characterized by the posses-sion of micro-latero-frontal cilia, have, in addition, a numberof characters in common as given below. Not all these charactersare restricted to members of the group, certain of them, whichare probably primitive, occurring for example in those formsnoted below which are apparently only distantly related. Threesuch characters are: (a) freedom of the posterior region of thegill axes (as also in aspidobranch Gastropods); (b) freedom ofthe dorsal edges of the ascending lamellae of the gills, or ciliaryconnexion only in ter se and with adjacent parts; and (c) theabsence of pallial sutures.

The Trigoniidae and Mytilidae, both of which were placedin the same order as the Arcidae and Anomiidae by Pelseneer,have some or all of these characters. In the Trigoniidae all threeare found; in fact this family seems to be at about the samelevel of evolution as the Arcidae as regards gill and mantlefreedom, though of different lineage. The Mytilidae, with cer-tain exceptions (Mytilus ovalis , Musculus marmora-tus , Septifer bi locular is (Eidewood, 1903)), agree withmembers of the group in the entirely ciliary method of divisionbetween the infra- and supra-branchial chambers, and in therelative freedom of the mantle margins, there being only anarrow pallial suture, but at least in Mytilus edulis andModiolus modiolus, differ from them in the shortness ofthe free portion of the gill axis and its reattachment at theextremity.

Certain Eulamellibranchs exhibit some of these characters;in As ta r te su lca ta is found freedom of the posterior portionof the gill axis, and in the same region ciliary connexion of thegill i n te r se and with adjacent parts (Atkins, 1937a); thereis also but one pallial suture. In the Solenidae and Solecurtidae(as represented by Solen marg ina tus , Ensis s i l iqua,Ensis a reua tus , Cultellus pel lucidus, Solecurtus

416 D. ATKINS

scopula, and Solecurtus chamasolen), the gill axesare free for a considerable distance, and this possibly occursin a number of Lamellibranchs, but masked by the fusion ofthe dorsal edges of the ascending lamellae with some adjacentparts. In the genus Venus (Venus verrucosa, Venuscasina) the gill axes are free for a short distance and thenreattached at the extreme tip as in Mytilus edul is .

As mentioned previously these are probably primitive lamelli-branch characters, the retention of which has not been restrictedto any one group, though certain of them have tended to persistto a greater degree in some groups than in others.

The common characters of the group with micro-latero-frontal cilia are as follows:

1. ' Prodissoconchs of homogeneous laminar structure, butnot prismatic, and with umbos directed posteriorly'; the suc-ceeding dissoconch with an external layer of prismatic cellulartissue (Jackson, 1890). In this description Jackson did notinclude the Arcidae, in different genera of which the umbos inthe adult may be directed forwards, inwards, upwards, or back-wards (Eeinhart, 1935). I have been unable to ascertainwhether the remainder of this definition is applicable to theArcidae.

2. Essentially byssiferous, leading to the greater developmentof the posterior than the anterior region of the animal (Arcidae),to reduction of the anterior region generally, and especially toreduction (Pinnidae) and disappearance (many forms) of theanterior adductor muscle (Douville, 1912a). Byssal fixation isnot restricted to the group.

3. Considerable free posterior region to the gill axes. In someforms (Malleus, Isognomon, P t e r i a , P inna , Ostrea)the gill axes are free for much of their length; in the Arcidae,Anomiidae, and Pectinacea to a less extent. According toPelseneer (1911, p. 29) in the Pectens the extent of the freeposterior portion of the axis varies in different species, beinglong in those with a stout byssus, and very short in free andswimming forms. In the Ostreidae the dorsal edges of theascending lamellae of the outer demibranchs are fused with themantle, and those of the inner demibranchs in ter se, so that


the condition of the gill axes is not seen without dissection. InOstrea also the united dorsal edges of the two inner demi-branchs are fused with the visceral mass anteriorly, for a dis-tance varying between about a third and two-thirds of the totallength in different species (Ostrea edul is , Ostrea vir-ginica, Ostrea angulata).

In certain members (Arcidae, Anomiidae, Pectinacea) the gillaxis is usually borne on a more or less deep suspensory mem-brane; such a membrane is absent generally in the Pteriacea(Pteria, Malleus, P inc tada , P inna , Isognomon)and in the Ostreidae.

4. Considerable development of muscles in the gill axes,Ostreidae (0strea edulis) excepted. While characteristic ofthe group, this character is by no means restricted to the group.

5. Method of division of the supra- from the infra-branchialchamber: the division of the mantle chamber is either merelyby the touching of the upturned edges of the demibranehsagainst adjacent parts, or by interlocking cilia, and rarely byorganic junction.

It is advisable to discuss this character in some detail. Wherethe division is brought about merely by the touching of thedorsal edges of the ascending lamellae of the outer demibranehsagainst the mantle, and by those of the inner demibranehsinter se and against the foot or visceral mass, interlockingcilia are frequently present on the gills, though not on the partsthey come in contact with: where the two inner demibranehsare in contact there is probably weak, easily dissolved, ciliaryunion. According to Bourne (1907) the short stiff cilia on thepallial faces of the velar filaments—that is the long reflecteddorsal ends of the ascending filaments—of Anomia(Aenigma)aenigmatica function as ciliated discs, and give a sufficientamount of friction against the mantle—which is without corre-sponding cilia—to prevent the whole of the ascending lamellafrom slipping down. Division by mere touching of the gillsagainst adjacent parts is found in the Arcidae (Area, Glycy-meris), Monia (Anomiidae), and Pectinacea. In Anomiaephippium and also in Anomia aenigmatica (Bourne,1907) this method prevails between the outer demibranehs and

NO. 319 E e

418 D. ATKINS

the mantle, but the inner demibranchs are in organic union.In a fragment of gills labelled P l a c u n a p l a c e n t a ( ? ) fromthe British Museum, the union of the dorsal edges of the innerascending lamellae was of a compound nature, being mostly-organic, but with an exceedingly short ciliary junction to theventral side of the organic fusion (fig. 3 A, PI. 29). This type of junc-tion has been described also i n P l a c u n a p l a c e n t a b y Hornell(1909, p. 70), while Eidewood (1903, p. 199) has stated for thesame species that the two inner lamellae of the inner demibranchsare not united at their upper edges, the gills in this differingfrom those of A n o m i a e p h i p p i u m , with which theyotherwise agree exactly. It is possible that if material is alcoholpreserved, the two edges might very easily be torn apart, orperhaps the method varies in different regions; one or other ofthese suggestions may possibly be the explanation of the con-flicting statements.

Eetention of the primitive condition of freedom both of theposterior region of the gill axes, and of the upturned edges ofthe demibranchs from adjacent parts, in the otherwise highlyevolved Pectinacea has been possibly dependent on the activehabits of many of them; it allows of the supra- and infra-branchial chambers being thrown into one during swimming,thus preventing possible injury to the gills. It is known thatpreparatory to the clapping of the valves in swimming theextremities of the gills swing forward by contraction of thelongitudinal muscles of the gill axes, and this obviously necessi-tates freedom from connexion of the gills with adjacent parts.

Division of the mantle chamber by ciliary contacts is foundin H e t e r a n o m i a s q u a m u l a (Anomiidae) (Atkins, 1936),and in the Pteriacea, though not in the Ostreidae. The con-tradictory statements of various authors as to whether thelamellae are free from, or attached to, adjacent parts in membersof the Pteriacea may be explained by the ciliary nature of thejunction allowing of fairly easy separation of the opposed sur-faces. Grobben in 1900 (pp. 493-5) had already recognized theciliary nature of the junction in P t e r i a and P i n e t a d a ,and considered it to be universal in the Pteriidae, in which heincluded I s o g n o m o n , C r e n a t u l a , and V u l s e l l a , but


his paper seems to have been overlooked by Eidewood (1903)and Herdman and Hornell (1904,1905). To the forms examinedby Grobben may be added Mal leus a l b u s . According toEidewood (p. 206) in this species' the upper edges of the ascend-ing lamellae are free from adjacent parts', bu t ' the apex of theetenidium is fused with the mantle edge'. In a specimen re-ceived from the Indian Museum, Calcutta, the dorsal edges ofthe ascending lamellae of the two inner demibranchs were inciliary connexion: those of the outer demibranchs were joinedto the mantle by the same means, but the left outer demibranchhad become partly detached, probably in opening the animal.In the region where the gill axes are free the interlocking wasespecially strong: sections through the gill and mantle failedto reveal any organic union. On the mantle margin, in theposition of the posterior tip of the gill, fragments of lamellaremained attached to the mantle after removal of the gill;sections, however, showed that the union was entirely ciliary,and that the tips of the gills are not fused with the mantlemargin as Eidewood supposed, unless this occurs in someindividuals and not in others.

In certain of the Pteriidae and in the allied family Isogno-monidae there is undoubted slight organic fusion, as well asciliary junction, between the two inner demibranehs. InP i n c t a d a v u l g a r i s while the junction is mostly ciliary, tothe ventral side of this there is a narrow organic bridge (Herd-man, 1905, p. 227): this is also found in I s o g n o m o n a l a t a(fig. 3 B, PI. 29). A junction of a compound nature has alreadybeen described in P l a c u n a p l a c e n t a (Anomiidae), but inthat instance the relative importance of the two kinds of unionwas reversed (see fig. 3, PI. 29). In P t e r i a m a c r o p t e r a acompound ciliary and organic junction is found between thetwo inner gills, but apparently only for a certain length behindthe visceral mass (Pelseneer, 1911, p. 25).

In the Pinnidae the division between the two mantle chambersis generally by ciliary contacts. This was found to be so inunspecified P i n n a by Grobben (1900, p. 495) and Stenta (1903,P- 228), in A t r i n a r i g i d a by Grave (1911, pp. 418-19), andin P i n n a f r a g i l i s by myself. In the latter species, at least,

420 D. ATKINS

there is a marked difference in the strength of the union betweenthe gills themselves in the middle line, and between the gillsand the mantle: the former junction is easily dissolved in theliving animal, the latter most difficult to separate; indeed itwas thought that the union was organic until sections weremade. Ridewood (1903, p. 214) stated that in Pinna nobilisthe upper edge of the outer ascending lamella is fused with themantle, though in the four other species he examined (Pinnapect ina ta ( = fragilis), Pinna nigra , Pinna zea-landica, Pinna virgata), it was free; it is not impossiblethat he may have been misled by a strong ciliary junction suchas occurs in Pinna fragil is .

The gills being long and the axes free for much of their lengthin Malleus, Pinna, and others, it is understandable thatthe junction of the gills with the mantle is necessarily strong.If such gills become separated from the mantle it would probablybe difficult for them to regain their position. In Pinna theweakness of the ciliary junction in the middle line will allow thesupra- and infra-branchial chambers to be thrown into one—thus preventing injury to the gills—during burrowing, whenwater is violently expelled from the shell anteriorly.

In Ostrea alone is the division between the two mantlechambers entirely organic. Ostrea, which is in all probabilitythe most highly evolved member of the group, at least as regardsthe gills, for there are no vestiges of ciliary interfilamentarjunctions, has moderate-sized or anomalous latero-frontal cilia:these differ in certain respects from the eu-latero-frontal ciliaoccurring in bivalves outside the group (see p. 365).

To summarize: division of the mantle chamber is (a) bytouching of the upturned edges of the demibranchs againstadjacent parts in the Arcidae and the Pectinacea, and inMonia among the Anomiidae. While this condition is foundbetween the outer demibranchs and the mantle inPlacunaand Anomia, between the two inner demibranchs there isa compound ciliary and organic junction in the former, and anentirely organic junction in the latter, (b) By interlocking ciliain Heteranomia squamula (Anomiidae) and the Pteri-acea, though there is also some slight organic fusion between the


two inner demibranchs in certain Pteriidae and Isognomonidae.(c) By entirely organic junction in the Ostreidae.

Forms with the first method of division of the mantle chamberhave each gill, or the united gills, free from adjacent parts: suchgills are generally borne on more or less deep suspensory mem-branes. Forms with the second and third methods of divisionhave the gills more or less firmly attached to adjacent parts:such gills generally lack suspensory membranes (see p. 417).

6. Gills with the outer and inner demibranchs similar, in thatthere is no supra-axial extension to the outer demibranch,though the ascending lamella of both demibranchs may be asdeep as the descending. In the Ostreidae, however, the ascend-ing may be rather deeper than the descending lamellae, andthis is especially noticeable in the outer demibranch of 0 s t reaangu la ta .

In the unrelated family Mytilidae the outer demibranch isalso •without a supra-axial extension, and possibly also in theTrigoniidae.

7. In those species of which it has been found possible toobtain living material, namely, members of the Arcidae, Ano-miidae, Pectinidae, Limidae, Pteriidae, Pinnidae, and Ostreidae,it has been found that longitudinal currents are present at thefree ventral edge of both inner and outer demibranchs (exceptthe outer demibranch of Heteranomia), though these maybe posterior in direction as in the Arcidae and Anomiidae. Alsocharacteristic of the group, with the exception of the Pinnidae,is the presence of opposed frontal currents on all lamellae,frequently on the same filament (Atkins, 1936, 1937, 1937 a).Outside the group this arrangement of frontal currents is foundin the Solenidae (Atkins, 1936). In the Mytilidae investigatedlongitudinal currents are present at the free ventral edges ofboth demibranchSj but frontal currents are entirely ventralward.The presence of opposed frontal currents on all filaments butof certain lamellae only in Barnea Candida, Pe t r ico lapholadiformis , Spisula sub t runca t a , Spisula el-l ip t ica , and Cultellus pellucidus has been shown to bea special adaptation (Atkins, 1937, 1937 a).

8. Absence of pallial sutures. The mantle lobes are charac-

422 D. ATKINS

teristically free throughout their extent in the Arcidae, Ano-miidae, Pteriidae, Isognomonidae, Amussiidae, Pectinidae,Spondylidae, Plicatulidae1 (see Watson, 1930, p. 25), andLimidae. The right and left vela are fused for a certain distancefrom the hinge ventrally in the posterior region in H e t e r a -nomia s q u a m u l a (Atkins, 1986), and anteriorly in certainspecies of L i m a without byssal fixation (e.g. L i m a h ians ) .Slight fusion of the opposite mantle lobes in the region of theposterior tips of the gills occurs in the Pinnidae and in theOstreidae. In L i m a h i a n s and L i m a l o s c o m b i there isno fusion in this position, but the opposite vela are greatlyincreased in depth locally to form projections, which approacha pointed membraneous process of the postero-ventral regionof the visceral mass, and which, even when the valves are widelygaping as usual in these species, make a partial, if not completedivision between the exhalent and inhalent currents.

The degree of fusion between the two mantle lobes is probablymore an adaptive character correlated with certain habits ratherthan a progressive one. The retention of the primitive conditionof an open mantle in the group with micro-latero-frontal ciliais probably dependent on their being typically surface forms.Surface forms and mobile burrowers have retained a largelyopen mantle in spite of one or two sutures, though many of theseamong the Eulamellibranchs, have developed siphons. Suchsiphons, however, rarely attain more than a moderate length,and are generally partially or entirely united; when the siphonsare very long and separate this is an adaptation to depositfeeding, as in the Tellinidae and Semelidae. An extensivelyclosed mantle, mostly though not invariably accompanied bysiphons, is found generally among burrowers and borers occupy-ing more or less permanent holes, and perhaps has been developedlargely as a means of increasing the strength of the expelledcurrent on sudden closure of the valves, the fused mantlemargin being withdrawn into, and thus causing reduction of,the infra-branchial chamber as described by Drew in S o leno-m y a (1900) and Bns i s (1907). It may be that such forms

1 No member of this family was examined, but there is no reason todoubt that it is correctly placed in the Pectinacea.


need to be able to expel especially strong currents to keep theirburrows clean and increase the depth: perhaps the arrangementcompensates for the tendency of the adductor muscles towardcontiguity to the hinge and to reduction in size, with consequentloss of power, evident in at least certain forms living in shelteredpositions where there is no need for the shell to be tightly closed.

9. The inner fold of the mantle margin is commonly welldeveloped, especially in swimming members where it forms adeep velum. The velum is deep in the Pectinidae, Amussiidae,Spondylidae, and Limidae; moderately deep in the Anomiidae,Pteriidae, and Ostreidae; narrow in the Isognomonidae; inPl ica tu la aus t ra l i s it is no more than 'a small ridgescarcely J mm. in height' (Watson, 1930, p. 25).

10. The mantle is capable of withdrawing a considerabledistance from the shell edge—that is, the retractor muscles ofthe mantle margin are inserted far from the shell edge—butto a much less extent in the Arcidae.

11. There is a tendency for members to he on, or be attachedby, the right valve.

Forms cemented by the right valve are, Spondylus,Chlamys d i s to r t a , and Pl ica tu la aus t r a l i s .

Forms permanently attached by the byssus passing througha sinus in the right valve are, Anomia, Heteranomia ,and Monia.

Forms temporarily attached by the byssus with the right sidenext the surface of attachment are, Isognomon, Malleus,P t e r i a , P inc t ada , and certain Pectinidae.

Unattached forms lying on the right side are the free andswimming Pectinidae and the Amussiidae.

There is a difference of opinion as to the valve on whichPlacuna p lacenta (Anomiidae) lies: according to Hornell(1909, p. 46) it is the left convex valve; according to Fischer(1887, p. 933) it is the right. Jackson (1890) refers to Wood-ward as mentioning that, when young, Placuna has a byssalsinus in the right valve.

Apart from the uncertain case of P lacuna , Ostrea seemsto be alone in the group in lying on the left valve, which iscemented to the underlying surface. It is doubtful, however,

424 D. ATKINS

whether much importance should be attached to the habit oflying on, or being attached by, the right valve (for see Pelseneer,1911, p. 86). O s t r e a is then exceptional in the compositionof the latero-frontal ciliated tract; in the method of divisionof the mantle chamber; in the poor development of the longi-tudinal muscles of the gill axis; and in making attachment bythe left valve.

Some free and some attached forms generally maintain thevalves in a vertical position, for example, L ima h i a n s ,L ima l o s c o m b i , and A r e a . When at rest beneath thesurface G l y c y m e r i s appears to lie indifferently on the rightor left valve, and has been found occasionally more or lessvertical.

12. The presence of abdominal sense organs on the posterioradductor muscle (see p. 402). The position of these organs,however, appears to be correlated with the absence of siphons,abdominal sense organs occurring in this position in asiphonateforms outside the group with micro-latero-frontal cilia, forexample, in the Trigoniidae (Pelseneer, 1906, p. 237) and in theMytilidae (Field, 1922, p. 175).

13. The auricles intercommunicate; the Anomiidae excepted.Intercommunication of the auricles is also found in mostMytilidae (Pelseneer, 1911, pp. 95, 120).

SUGGESTED MODIFICATIONS OP THE CLASSIFICATION OF

THE LAMELLIBEANCHIA.

I am loath to add further names to the classification of theLamellibranchia, while unable to divide phylogenetically thegreat group with eu-latero-frontal cilia, but I am inclined tothe opinion that two groups, Macrociliobranchia and Micro-ciliobranchia should be provisionally introduced. In the Macro-ciliobranchia are placed the Protobranchia, the Pilibranchia(emended to include the Mytilacea and Trigoniacea only), theEulamellibranchia (Pelseneer, 1911), and the Septibranchia. Itis very probable, however, that the emended Filibranchia is stilldiphyletic, as already pointed out (p. 394). The Septibranchiamay provisionally be classed as an order of the Macrocilio-

THE CILIARY MECHANISMS OF LAMBLLIBEANCHS 425

branchia: this seems warranted by their probable origin fromthe Anatinacea (with eu-latero-frontal cilia). The disappearanceof latero-frontal cilia in this order is no doubt to be correlatedwith the change in the mode of feeding. Small latero-frontalcilia have been noted in P o r o m y a o r e g o n e n s i s by Ride-wood (1903, p. 274), but it is probable that these should beregarded as vestiges of eu-latero-frontal cilia, rather than astrue micro-latero-frontal cilia.

As previously mentioned (p. 413), the terms Mlibranch andEulamellibranch indicate stages in gill evolution, likely to occurin different lineages, and therefore their permanent retentionas names of orders, Filibranchia and Eulamellibrancbia, doesnot seem desirable, but until the Macrociliobranchia can bedivided according to genetic affinities they must he retained,though lamellibranehs with filibranchiate and eulamelli-branchiate gills occur in the other group, Microciliobranchia.

In the Microciliobranchia the Anomiacea, Pteriacea, Pectina-cea, and Ostreacea appear to be more closely related to oneanother than they do to the Arcacea, but whether this is suffi-cient to warrant the creation of two orders, the order Pseudo-lamellibranchia (emended) lapsing, I am unable to determine.In order to introduce as few new names as possible the orderPseudolamellibranchia has been retained in an emended sense,in spite of its unsuitability, leaving the introduction of a newname until such time as the revision of the Macrociliobranchiais undertaken. Pelseneer recognized the Pseudolamellibranchiaas a natural group, in spite of filibranchiate and eulamelli-branchiate members, but he missed the link that connectsthem, namely, the characteristic composition of the latero-frontal ciliated tract, and omitted to include the Arcidae andAnomiidae.

The suggested classification of the Lamellibranchia is thenas follows:

Class L A M E L L I B R A N C H I A .Group I. MACEOCILIOBRANCHIA. Latero-frontal tracts of thegill filaments or leaflets consisting of a row of eu-latero-frontalcilia, with also a row of pro-latero-frontal cilia in all or mostmembers.

426 D. ATKINS

Order 1. PEOTOBRANCHIA (Pelseneer): Nueulidae, Nucu-lanidae, Solenomyidae.

Order 2. FILIBRANCHIA (emended). Eestrieted to Filibranchswith latero-frontal tracts of the type described for the group.This order is the order Filibranchia of Pelseneer, less theAnomiacea, the Arcidae, and allied families.

Sub-order 1. MYTILACEA (Pelseneer, 1911). This is theMytilacea of Pelseneer, 1906, less the Pernidae.

Sub-order 2. TRIGONIACEA. Gill filaments with distinctivearrangement of the ciliary tracts, see p. 381. Trigoniidae.

Order 3. EULAMELLIBRANCHIA (Pelseneer, 1891, 1892, 1911,not 1906).

Order 4. SEPTIBRANCHIA (Pelseneer).Group II. MICRO CILIOERANCHIA. Latero-frontal tracts of thegill filaments consisting characteristically of a row of micro-latero-frontal cilia.

Order 1. PSETJDOLAMELLIBRANCHIA (emended). This orderis the order Pseudolamellibranchia of Pelseneer, 1911,emended to include those Filibranchs (Anomiacea, Arcidae,and allied families) with latero-frontal tracts of the typedescribed for the group.

Sub-order 1. ABCACEA (emended). This sub-order is theArcacea of Pelseneer less the Trigoniidae and allied families.

Sub-order 2. ANOMIACEA (Pelseneer).Sub-order 3. PTERIACEA, including the Pinnidae ( = A v i c u -

l acea of Pelseneer, 1911, less the Ostreidae).Sub-order 4. PECTINACEA, including the Limidae (Pelseneer,

1911).Sub-order 5. OSTREACEA. Latero-frontal tract consisting of

anomalous together with para-latero-frontal cilia. Ostre-idae.

Pelseneer (1911) and Douville (1912a) have both given dia-grams illustrating their conceptions of the phylogeny of theLamellibranchia as a whole; Jackson (1890) gave only that ofthe Aviculidae and their allies.

In Pelseneer's diagram there is a single tree with manybranches, and the divisions (orders) are horizontal and indicatestages in the evolution of the gills. The Filibranchia are shown

THE CILIARY MECHANISMS OP LAMELLIBRANCHS 427

as derived from the Nuculidae; the Pseudolamellibranchia fromthe common ancestor of the Mytilidae and the Areidae; theEulamellibranchia from the Mytilidae by way of the Astartidae;and the Septibranehia from the Anatinacea. Comments on thisclassification have been made on pp. 394-398.

Douville, on the other hand, recognized three divergentbranches from the beginning, corresponding to the three princi-pal modes of life of the class, which he held early impressed onthe branches certain characters which tended to persist throughlater secondary modifications. The three branches are:

(1) the 'normal' branch, more or less active and free living,descended from ' formes primitives nacrees normales' by way offorms like A c t i n o d o n t a ;

(2) the 'sedentary' branch affected by byssal fixation,descended from 'formes primitives nacrees fixees' by wayof the Pterineidae; the Areidae being connected with thesethrough P a l a e a r c a ( = Cypr i ca rd i t e s ) ; and

(3) the 'burrowing' branch modified by a protected life ina more or less permanent burrow, descended from 'formesprimitives nacrees cavicoles' by way of the Solenopsidae,Protomyidae, and Grammysiidae. His divisions are thus verti-cal. Davies (1933) has recently given a clear exposition, withdiagrams, of Douville's scheme of classification and comparedit with several other well-known classifications, so that there isno need to enter into it fully here. At present the only modifica-tions of Douville's phytogeny that I am able to suggest are two:(a) the exclusion of the Mytilidae from the ' sedentary' branch.Though the condition of the latero-frontal tract shows that thisfamily belongs to the Macrociliobranchia, yet it gives no indica-tion of its allies, nor does the pattern of the lateral ciliated cells,which is not of the common type found among the 'normal' and'burrowing' branches. Dall (in the Eastman edition of Zittel,1913, p. 462) suggested that the prototypes of the Mytilidaeare to be found in the Modiolopsidae: Douville placed Modio-lops is in his 'normal' branch (though placing the Mytilidaein the 'sedentary' branch): this perhaps indicates where theallies of the Mytilidae should be sought. And (b) that possiblythe 'burrowing' branch is not as widely separated from the

428 D. ATKINS

'normal' branch as is indicated in his diagram, a conclusionsuggested by the character of the latero-frontal tract and thepattern of the lateral ciliated cells. A certain pattern of thelateral cells is common among both 'normal' and 'burrowing'branches, though it is not the only pattern found among these ;the Mytilidae, Trigoniidae, Astartidae, Unionidae, and Aetheri-idae having patterns different from the common one (Atkins,1938 a). I think the work recorded in this paper also shows thatN u e u l a cannot be regarded as broadly ancestral to the'sedentary' branch.

SUMMAEY.

Certain Lamellibranchs have the latero-frontal tract composedof large complex 'cilia', here called e u - l a t e r o - f r o n t a lc i l i a , together with subsidiary ones, termed p r o - l a t e r o -f r o n t a l c i l i a . This type of latero-frontal tract occurs insome or all of the three families of Protobranchs (there is somedoubt as to the presence of pro-latero-frontal cilia in all thefamilies), and in the Mytilidae and probably the Trigoniidae(fixation too imperfect for the identification of pro-latero-frontal cilia) among the Hlibranchs, and in all the marinefamilies of Eulamellibranchs obtainable at Plymouth, and inthe fresh-water families, Dreissensiidae, Sphaeriidae, Unionidae,Mutelidae, and Aetheriidae. A list of the species investigatedis given.

Other Lamellibranchs, which were previously considered aslacking latero-frontal cilia, have been found to possess smallones only, difficult of observation, termed m i c r o - l a t e r o -f r o n t a l c i l i a . These occur in the Arcidae, Anomiidae,Pteriidae, Pectinidae, Spondylidae, Limidae, Pinnidae, and areinferred to be present in the Amussiidae, VulseUidae, andIsognomonidae, in which eu-latero-frontal cilia are certainlyabsent. A list of the species examined is given.

In one family, the Ostreidae, moderate-sized latero-frontalcilia, termed a n o m a l o u s l a t e r o - f r o n t a l c i l i a , togetherwith subsidiary ones, termed p a r a - l a t e r o - f r o n t a l c i l i aare present.

In bivalves having eu-latero-frontal cilia the arrangement of


the various ciliary tracts, frontal, latero-frontal, and lateral isfairly constant, notable exceptions being a Protobranch,N u c u l a n a , and a Filibranch, T r i g o n i a . In bivalves havingmicro-latero-frontal cilia the arrangement of the various tractsseems more or less constant.

The homology of the various types of latero-frontal cilia isdiscussed. The composition of the latero-frontal ciliated tractshas been found to be a stable character, and, as it is correlatedwith other characters, has taxonomic value.

It is suggested that the variations in the constitution of thelatero-frontal tracts tend to show that Eidewood's (1903)classification does not express genetic affinities, as he himselfconceded, nor does Pelseneer's (1911) entirely, and that Pel-seneer's order Filibranchia, and Eidewood's orders Eleuthero-rhabda and Synaptorhabda are not monophyletic.

Families possessing micro-latero-frontal cilia appear to beclosely related, and form a group, which, with certain modifica-tions, corresponds to 'the Aviculidae and their allies', or the'sedentary' branch of Lamellibranchs, previously establishedby the palaeontologists, Jackson and Douville respectively,largely on shell characters. Thus the constitution of the latero-frontal tracts of the gill filaments supports the findings ofpalaeontologists with regard to this group. Unfortunatelyneither Jackson nor Douville proposed a formal name for thegroup.

The relationship of forms with micro-latero-frontal cilia, andthe evolution within the group of the eulamellibranchiate or syn-aptorhabdic gill are discussed. One family, the Ostreidae, whichmust be included on account of its relationship with either thePteriacea or Peetinacea (based on other evidence) has moderate-sized, or anomalous latero-frontal cilia together with para-latero-frontal cilia. The anomalous latero-frontal cilia differ incertain respects from the large cilia characteristic of the majorityof the Lamelhbranchia, and are presumed to have arisen in-dependently.

Common characters of the group characterized by the posses-sion of micro-latero-frontal cilia, in addition to the form of thelatero-frontal cilia, are: (1) shell characters of the prodissoconch,

430 D. ATKINS

Arcidae excepted; (2) byssal fixation; (3) considerable freeposterior region to the gill axes; (4) considerable developmentof muscles in the gill axes, Ostreidae excepted; (5) method ofdivision of the pallial cavity, Ostreidae excepted; (6) gills with-out a supra-axial extension to the outer demibranch; (7)presence of longitudinal currents at the free ventral edge of bothinner and outer demibranchs; and of opposed frontal currentson all lamellae and frequently on the same filament, Pinnidaeexcepted; (8) absence of pallial sutures, Pinnidae and Ostreidaeexcepted; (9) inner fold of the mantle margin characteristicallywell developed, especially in swimming forms; (10) insertionof the retractor muscles of the mantle margin at a considerabledistance from the shell edge, Arcidae excepted; (11) tendencyfor members, except the Ostreidae, to lie on the right valve;(12) abdominal sense organs on the posterior adductor muscle ;and (13) intercommunication of the auricles, Anomiidae ex-cepted.

Two groups of the Lamellibranchia are proposed provision-ally, namely Group I, Macroeiliobranchia, including the ordersProtobranchia (Pelseneer), Hlibranchia (emended to includeonly the Mytilacea and Trigoniacea), Eulamellibranchia (Pel-seneer, 1911), and Septibranchia (Pelseneer); and Group II,Microciliobranchia, with the order Pseudolamellibranchia,emended to include the sub-orders Arcacea (excluding theTrigoniidae), Anomiacea, Pteriacea, Pectinacea, and Ostreacea.The Macroeiliobranchia will need revision, for it is very probablethat the Filibranchia (emended), if not the Eulamellibranchia,are still not monophyletic.

BEFEEENCES FOE PAETS V, VI, AND VII.

Arkell, W. J., 1930.-—"Generic Position and Phytogeny of some JurassicArcidae. I", 'Geol. Mag.', vol. lxvii. II, ibid., vol. lxvii.

Atkins, D., 1930.—"Abnormal Conditions of Gills in Mytilus edulis.Part I. Permanent Natural Reversal of the Frontal Cilia",' Journ. Mar.Biol. Assoc.', vol. xvi, N.S., 1929-30.

1932.—"Ciliary Feeding Mechanism of Entoproct Polyzoa, and acomparison with Ectoproct Polyzoa " , ' Quart. Journ. Micr. Sci.', vol. Ixxv,N.S.

1936.—"Ciliary Mechanisms and Interrelationships of Lamellibranchs.Part I. Sorting Mechanisms", ibid., vol. lxxix, N.S.


Atkins, D-, 1937.—" Part II. Sorting Devices on Gills", ibid., vol. lxxix, N.S.1937a.—"Part III. Types of Lamellibranch Gills and their Food

Currents", ibid., vol. lxxix, N.S.19376.—"Part IV. Cuticular Fusion, &c", ibid., vol. lxxix, N.S.1938.—"Part V. Note on Gills of Amussium pleuronectes L.", ibid.,

vol. lxxx, N.S.1938a.—"Part VI. Pattern of the Lateral Ciliated Cells of Gill

Filaments", ibid., vol. lxxx, N.S.1938 b.—" Part VII. Latero-frontal cilia and their Phylogenic value,'ibid., vol. lxxx, N.S.

Bhatia, D., 1926.—"Structure of Latero-frontal Cells of gills of Mytilusedulis", ibid., vol. lxx, N.S.

Borg, F., 1926.—"Recent Cyelostomatous Bryzoa", 'Zool. Bidrag.Uppsala', Bd. x.

Bourne, G. C, 1907.—"Structure of Aenigma aenigmatica, Chemnitz",'Quart. Journ. Micr. Sei.', vol. li, N.S.

Biitschli, 0., 1934.—"Atemorgane: Lamellibranchiata", in 'Vorlesungenuber vergleichende Anatomie', Bd. ii, Lief. 6.

Carter, G. S., 1924.—•" Structure and Movements of Latero-frontal Ciliaof Gills of Mytilus", 'Proc. Roy. Soc. Lond.', vol. xcvi, B.

• 1926.—"Nervous Control of Velar Cilia of Nudibranch Veliger",'Brit. Journ. Exper. Biol.', vol. iv.

1928.—"Structure of Cells bearing Velar Cilia of Nudibranch Veliger",ibid., vol. vi.

Clasing, M., 1921.—"Beitrage zur Kenntnis der Aviculiden", 'Jen. Zeit.fur Naturwissensehaft', Bd. Ivii, Heft 3.

Crofts, D. R., 1929.—'Haliotis. L.M.B.C. Memoir XXIX.' Univ. Pressof Liverpool.

Dakin, W. J., 1909.—'Pecten. L.M.B.C. Memoir XVII.' Univ. Press ofLiverpool.

1928.—"Anatomy and Phytogeny of Spondylus . . . LamellibranchNervous System", 'Proc. Roy. Soc. Lond.', B, vol. ciii.

1928a.—"Eyes of Pecten, Spondylus, Amussium and Allied Lamelli-branchs, Discussion on their Evolution", ibid., vol. ciii, B.

Dall, W. H., 1885-6.—"Reports on Results of Dredging, in Gulf of Mexico(1877-78) and in Caribbean Sea (1879-80), by U.S. Coast Survey Steamer'Blake' . . . Mollusca. Part I. Brachiopoda and Pelecypoda", 'Bull.Mus. Comp. Zool.', vol. xii.

Dautzenberg, P., and Bavay, A., 1912.—'LameUibranches de l'expeditiondu Siboga. Partie Systematique. I. Pectinides.' Monographie liiib.Siboga-Expeditie.

Davies, A. M., 1933.—"Bases of Classification of Lamellibranchia",'Proc. Malac. Soc, Lond.', vol. xx, 1932-3.

Douville, H., 1898.—"Classification phylogenique des LameUibranches",'C.R. Acad. Sci., Paris', tome cxxvi.

432 D. ATKINS

Douville, H., 1912.—"Classification phylogenique des Lamellibranohes",ibid., tome cliv.

1912a.—"Classification des Lamellibranches", 'Bull. Soc. Geol.France', ser. 4, tome xii.

Drew, G. A., 1900.—"Locomotion in Solenomya and its Relatives", ' Anat.Anz.', Bd. xvii.

1901.—"Life-history of Nucula delphinodonta (Mighels)", 'Quart.Journ. Micr. Sci.', vol. xliv, N.S., 1900-1.

• 1906.—'Habits, Anatomy, and Embryology of Giant Scallop (Pectentenuicostatus)' Univ. Maine Studies, no. vi.

1907.—"Habits and Movements of Razor-shell Clam, Ensis direotus",'Biol. Bull. Woods Hole', vol. xii, 1906-7.

1907 a.—"Circulatory and Nervous Systems of Giant Scallop (Pectentenuicostatus . . . ancestry of Lamellibranchiata", ibid., vol. xii, 1906-7.

Engelmann, T. W., 1880.—"Anatomie und Physiologie der Flimmer-zellen", 'Pfliiger's Archiv. f. ges. Physiologie', Bd. xxiii.

Field, I. A., 1922.—"Biology and economic value of sea mussel, Mytilusedulis", 'Bull. Bur. Fish.', vol. xxxviii, 1922.

Fischer, P., 1887.—'Manuel de Conchyliologie et de Paleontologie conchylio-logique.' Paris.

Forbes.E., and Hanley, S., 1853.—'History of British Mollusca, and then-shells', vols. i-iv, London.

Grave, B. H., 1911.—"Anatomy and Physiology of Wing-shell Atrinarigida", 'Bull. Bur. Fish.', vol. xxix, 1909 (issued 1911).

Grave, C, and Sehmitt, F. O., 1925.—"Mechanism for Coordination andRegulation of Ciliary Movement as Revealed by Microdisseetion andCytological Studies of Ciliated Cells of Mollusks", 'Journ. Morph.',vol. xl.

Gray, J., 1922.—"Mechanism of Ciliary Movement", 'Proc. Roy. Soc.Lond.', vol. 93, B.

1928.—'Ciliary Movement.' Cambridge Univ. Press.Grobben, C, 1900.—"Expedition S. M. Schifi 'Pola' in das Rothe Meer,

XX", 'Morphologie und Anatomie von Meleagrina sowie der Aviculidenim allgemeinen.' Wien. Also in 'Denkschriften der math.-natur. Classeder K. Akad. der Wiss. Wien', Bd. Ixix, 1900.

Gutsell, J. S., 1931.—"Natural History of the Bay Scallop", 'Bull. Bur.Fish.', vol. xlvi, 1930 (1931).

Haas, F., 1934.—" Bivalvia." Bd. iii, Abt. 3, in Bronn's 'Klassen undOrdnungen des Tierreichs'. (Gives a full bibliography of Mollusca.)

Haren-Noman, D. van, 1881-2.—"Lamellibranchiaten gesammelt wahrendder Fahrten des 'Willem Barents'", 'Nied. Arch, fur Zool.', Suppl.,Bd. i, Leipzig.

Herdman, W. A., and Hornell, J., 1904.—"Anatomy of the Pearl Oyster(Margaritifera vulgaris)", 'Rept. to Gov. Ceylon Pearl Fish. Gulf ofManaar', pt. ii (Roy. Soc).

THE CILIABY MECHANISMS OP LAMELLIBKANCHS 433

Herdman, W. A., 1905.—"Structure of Gill of Ceylon Pearl-Oyster",' Journ. Linn. Soc, Zool., Lond.', vol. xxix, 1903-6.

Hornell, J., 1909.—"Anatomy of Placuna placenta", 'Kept. Gov. Baroda,Mar. Zool.', pt. i.

Jackson, R. T., 1890.—"Phytogeny of Pelecypoda, Avieulidae and theirAllies", 'Mem. Boston Soc. Nat. Hist.', vol. iv, no. 8.

Janssens, F., 1893.—"Branchies des Aeephales", 'La Cellule', tome ix.Kellogg, J. L., 1892.—"Morphology of Lamellibranchiate Mollusks",

'Bull. U.S. Fish. Commission', vol. x, 1890 (1892).Lucas, A. M., 1931.—"Investigation of Nervous System as a possible factor

in regulation of Ciliary activity of Lamellibranch Gill", 'Journ. Morph.',vol. li.

1932.—"Coordination of Ciliary Movement. I. Relation of CiliaryCoordination to Ciliary Inhibition", ibid., vol. liii.

1932a.—"Coordination of Ciliary Movement. II . Effect of tempera-ture upon ciliary wave length", ibid., vol. liii.

March, M. C, 1912.—"General Classification of the Pelecypoda", 'Ann.Mag. Nat. Hist.', ser. 8, vol. x.

Montgomery, T. H., 1902.—"Phylogenetic Classification", 'Proc. Acad.Nat. Philadelphia', pp. 187-232.

Neumayr, M., 1889.—"U. d. Herkunft der Unioniden", 'Sitzungsber. K.Akad. Wiss, Wien', vol. xcviii. Also in 'Ann. Mag. Nat. Hist.', vol. iii,1889.

1891.—"Morphologischen Eintheilung der Bivalven", 'Denkschr.d.K. Akad. d. Wiss. Wien', Bd. lviii (not seen personally).

Nicol, E. A. T., 1930.—"Feeding Mechanism, Formation of Tube, andPhysiology of Digestion in Sabella pavonina", "Trans. Roy. Soc. Edin.%vol. lvi, part 3.

Noman, D. van Haren, 1882.—"Lamellibranchiaten gesammelt wahrendder Fahrten des 'Willem Barents', 1878 und 1879", 'Niederl. Areh. f.Zool.', Suppl., Bd. 1.

Orton, J. H., 1912.—"Mode of Feeding of Crepidula, &c", 'Journ. Mar.Biol. Assoc.', vol. ix, N.S., 1910-13.

1913.—"Ciliary Mechanisms on Gill and Mode of Feeding in Am-phioxus, Aseidians and Solenomya togata", ibid., vol. x, N.S., 1913-15.

1914.—"Ciliary Mechanisms in Brachiopods, &c", ibid., vol. x, N.S.,1913-15.

1924.—"Causes of Unusual Mortality among Oysters in EnglishOyster Beds during 1920 and 1921. Part I" , 'Ministry of Agricul. andFish., Fishery Invest.', ser. ii, vol. vi, 1923 (1924).

Peck, R. H., 1877.—"Minute Structure of Gills of Lamellibranch Mollusca",'Quart. Journ. Micr. Sci.', vol. xvii, N.S.

Pelseneer, P., 1888.—"Anatomy of Deep-Sea Mollusca." 'Challenger'Reports, part lxxiv, vol. 27.

NO. 319 F f

434 D. ATKINS

Pelseneer, P., 1889.—"Classification phylogenetique des Pelecypodes','Bull. Sci. France et Belgique', tome xx, ser. 3, vol. 2.

1891.—"Contribution a l'etude des Lamellibranches", 'Arch. Biol.',tome xi.

•—— 1892.-—"Introduction a l'etude des Mollusques", 'Ann. Soo. Roy.Mai. Belg.', tome xxvii. Published separately 1894.

1906.—'Mollusca. Treatise on Zoology', edited by E. R. Lankester,Part V, London.

1911.—'Lamellibranches de 1'expedition du Siboga, Partie Ana-tomique', Monographie liiia, Siboga-Expeditie.

Posner, C, 1875.—"Bail der Najadenkieme", 'Archiv. f. Mikr. Anat.',vol. xi.

1877.—"Histologische Studien iiber die Kiemen der AcephalenMollusken", ibid., vol. xiv (not seen personally).

Prashad, B., 1931.—"Parallel Evolution in the Molluscan Faunas ofSouth-eastern Asia and South America", 'Proc. Roy. Soc. Edin.',vol. li, 1930-1.

1932.—'Lamellibranchia of Siboga Expedition. Systematic Part IIPelecypoda (exclusive of the Pectinidae)', Monographie liii c, Siboga-Expeditie.

Rabl, C, 1877.—"Bau der Najadenkieme", 'Jen. Zeitschr.', vol. xi (notseen personally).

Reinhart, P. W., 1933.—"Classification of Pelecypod Family Arcidae",'Bull. Mus. roy. Hist. nat. Belg.', tome xi, no. 13.

Rice, E. L., 1897.—" Systematisehe Verwertbarkeit der Kiemen b. d.Lamellibranchiaten", 'Jen. Zeitschr.', vol. xxxi (seen in abstract' Journ.Roy. Micr. Soc.', 1897).

Ridewood, W. G., 1902.—"Structure of Gills of Lamellibranchia", 'Proc.Roy. Soc. Lond.', vol. lxx.— 1903.—"Structure of Gills of Lamellibranchia", 'Phil. Trans. Roy.Soc, Lond.', B, vol. cxcv.

Saguchi, S., 1917.—"Studies on Ciliated Cells", 'Journ. Morph.', vol. xxix.Schenck, H. G., 1934.—"Classification of Nuculid Pelecypods", 'Bull.

Mus. roy. Hist. nat. Belg.', tome x, no. 20.Setna, S. B., 1930.—"Neuro-museular mechanism of gill of Pecten",

'Quart. Journ. Micr. Sci.'. vol. lxxiii.Sharp, B., 1888.—"Phytogeny of the Lamellibranchiata", 'Proc. Acad.

Nat. Sci., Philadelphia', pp. 1 2 H . Also 'Ann. Mag. Nat. Hist.', vol. ii,ser. 6, 1888.

Smith, E. A., 1885.—'Report on Lamellibranchiata.' 'Challenger' Reports,pt. xxxv, vol. 13.

Stenta, M., 1903.—" Stromungen im Mantelraume der Lamellibranchiaten",'Arb. Zool. Inst. Wien', Bd. xiv, Heft 2.

1908.—"Classificazione dei Lamellibranchi", 'Boll. Soc. Adr. Sci.Nat., Trieste', vol. xxv.

THE CILIABY MECHANISMS OF LAMELLIBEANCHS 435

Studnitz, G., 1931.—-"Morphol. und Anat. von Lima inflata, nebst biolog.Unterauch. an Lima hians", 'Zool. Jahrbiicher', Bd. liii.

Thiele, J., 1926.—"Mollusca", in Kukenthal, 'Handbuch der Zoologie',Bd. v, Berlin.

Wallengren, H., 1905.—"Biologie der Muscheln. I. Die Wasserstromun-gen", 'Lunds Univ. Arsskrift', N.F., Afd. 2, Bd. i, nr. 2.

Watson, H., 1930.—"Anatomy and Affinities of Plicatula", 'Proc. Malac.Soc, Lond.', vol. xix, pt. 1.

Willey, A., 1911.'—'Convergence in Evolution.' London.Winckworth, R., 1931.—"Nucula nitida", 'Proc. Malac. Soc, Lond.',

vol. xix, pt. 6.1932.—"British Marine Mollusca", ' Journ. Conch.', vol. xix, no. 7.1934^—"Names of British Mollusca II" , ibid., vol. xx, no. 1.1934a.—"Names of British Mollusca III", ibid., vol. xx, no. 2.

Woodward, M. F., 1901.—"Anatomy of Pleurotomaria beyrichii", 'Quart.Journ. Micr. Sci.', vol. xliv, 1900-1.

Worley, L. G., 1934.—-"Dual nature of metachronism in ciliated epithe-lium", 'Journ. Exp. Zool.', vol. lxix.

Yonge, C. M., 1926.—"Structure and Physiology of Organs of Feedingand Digestion in Ostrea edulis", 'Journ. Mar. Biol. Assoc.', vol. xiv,N.S., 1926-7.

1928.—" Structure and Function of Organs of Feeding and Digestion inSeptibranchs, Cuspidaria and Poromya", 'Phil. Trans. Roy. Soc,Lond.', B, vol. ccxvi.

1932.—"Feeding and Digestion in Pterocera and Vermetus . . .Crystalline Style in Gastropoda." Great Barrier Reef Expd. 1928-9,'Sci. Repts.', vol. i, no. 10. Brit. Mus. (Nat. Hist.)

1935.—"Digestion in Ciliary Feeding Animals", 'Journ. Mar. Biol.Assoc.', vol. xx, N.S.

1936.—'Mode of Life, Feeding, Digestion and Symbiosis withZooxanthellae in Tridacnidae.' Great Barrier Reef Expd. 1928-9.Brit. Mus. (Nat. Hist.), vol. i, no. 11.

1936a.—-"Evolution of Swimming Habit in Lamellibranchia",'Mem. Mus. Roy. Hist. Nat. Belg.', ser. 2, fasc. 3.

1937.—" Biology of Aporrhais pes-pelecani and A. serresiana",'Journ. Mar. Biol. Assoc.', vol. xxi, N.S.

Zittel, K. A. von, 1913.—'Text-book of Palaeontology.' English editionedited by C. R. Eastman, vol. 1, 2nd edition. Macmillan & Co., Ltd.(reprinted in 1927 and 1937)

EXPLANATION OF PLATE 29Fig. 1.—Lima h i a n s . Transverse section of a plica to show vestiges

of ciliary interfilamentar junctions (i.e.), ciliary junction between twolimbs of a plica (c.j.), and muscle fibres, o./., apical filament; ch., darklystaining chitin; h.m., horizontal muscle-fibres of interfilamentar junctions;

on the ciliary mechanisms and interrelationships of ...hirundo, spondylus gaederopus, anomia...

Documents