on the anaspidacea, living and fossil

OX 'L'HE ANASPIDA.CBA, LIVING AND FOSSIL. 489

On the Anaspidacea, Living and Fossil.

By

Geoffrey Smith,Fellow of New College, Oxford.

With Plates 11 & 12 and 62 Text-figures.

CONTENTS.

1. HISTORICAL INTRODUCTION

2. EXTERNAL MORPHOLOGY(A) General Appearance(B) Segmentation(c) Appendages(D) Theoretical Considerations

3. INTERNAL ANATOMY

(A) Pei'icardiuni, Heart, and Vascular System(B) Alimentary Canal(c) Excretory System(D) Reproductive Organs(E) Muscular System.(F) Nervous System .(G) Theoretical Considerations

4. BIONOMICS . . . .

(A) Habitat and General Habits(B) Distribution(c) Breeding and Reproduction

5. T H E RELATION OF THE SYNCARIDA TO OTHERTRACA . . . .

6. SYSTEMATIC P A R T

PAGE. 489. 495. 495. 500. 502. 523. 528. 528. 52f». 537. 537. 542. 542. 545. 54<>. 54«. 548. 548

MALACOS-

. 550

. 555

1. HISTORICAL INTRODUCTION.

THE first members of the Anaspidacea to be discovered and

described were certain fossil forms occurring in the Permian

490 GEOFJPBEY SMITH.

and Carboniferous strata of Europe and North America, andit was not until long afterwards that a living representativewas found in the fresh waters of Tasmania and recognised asthe near relative of these very ancient fossils.

In 1,856 Jordan and von Meyer (1) described a fossilshrimp from the Permo-Carboniferous of Saarbriick whichthey named Grampsonyx finibriatus, and which was seento combine certain features of the Podopthalinate Crustaceawith the entire absence of a carapace.

In 1865 and 1868 Meek and Worthen (2 and 3) figuredtwo similar forms from the Coal-measures of Illinois whichthey named Acanthotelson stimpsoni andPaleeocaristypus.

The systematic position of these fossils remained obscureuntil Packard re-examined them, and in a series of papers(4, 5 and 6) did a good deal to elucidate their structure andaffinities. He instituted the use of the term " Syncarida" toinclude the three genera and to designate a group of thehigher Crustacea intermediate in its characters between theSchizopoda and the Bdriopthalmata. Packard's conceptionof the affinities of this group have been borne out byrecent investigations, and his name "Syncarida" has beenadopted by Caiman as one of the divisions of the Mala-coetraca.

In 1893 Thomson (7) gave an account of a remarkablefreshwater shrimp from the top of Mount Wellington, Tas-mania, which was pointed out to him by Mr. Rodway, ofHobart, and which had been known to, and even occasionallyeaten by, the settlers for some time. Thomson named theanimal Anaspides tasnianias, described the most importantpoints of its external anatomy, and decided that it belongedto the sub-order Schizopoda, of which he considered it themost primitive member.

The connection between Anaspides and the fossil Syn-cai'ida was first pointed out by Caiman (8), who revisedThomson's description in certain particulars, and drew acareful comparison between Anaspides and the Carboni-

ON TH10 ANASPIDAOEA, UVJXG AND FOSSIL. 491

ferous fossils, concluding- therefrom that they all belongedto the same order and are very closely allied.

In a subsequent paper (9) the same author amplifies andcrystallises the views of Boas and Hanseu on the classifica-tion of the Malacostraca, and proposes to do away with theorder Schizopoda and to redistribute its component families,uniting the Euphausiidic with the Decapoda to form adivision Eucarida, while the Eucopiidaa, Lophogastridas andMysidas are united with the Cumacea, Amphipoda andIsopoda as Peracarida. The Auaspididas are placed in aseparate subdivision, the Syucarida, together with the threefossil forms mentioned above.

Caiman gives the following diagnosis of the three divisions:Division Eucarida.—Carapace coalescing dorsally with

all the thoracic somites. Eyes pedunculate. Antennal pro-topodite with, at most, two distinct segments. Mandiblewithout lacinia mobilis in adult. Thoracic limbs flexedbetween foui'th and fifth, segments. No oostegites. Anappendix interna sometimes present on pleopods. HepaticctBca much ramified. Heart abbreviated, thoracic. Sperma-tozoa spherical or vesicular, often with radiating appendages.Development, as a rule, with metamorphosis.

Division Peracarida.—Carapace, when present, leavingat least four thoracic segments distinct. First thoracicsegment always fused with the head. Antennal protopoditetypically of three segments. Mandible with lacinia mobilis(except in parasitic and other modified forms). Thoraciclimbs flexed between fifth and sixth segments. Oostegitesattached to some or all of the thoracic limbs in female, form-ing a brood-pouch. No appendix interna on pleopods.Hepatic easca few and simple. Heart elongated, extendingthrough the greater part of the thoracic region, or displacedinto abdomen. Spermatozoa filiform. . Development takingplace within brood-pouch; young set free at a late stage.

Division Syncarida.—Carapace absent. All the thoracicsegments distinct. Eyes pedunculate. Antennal protopoditeof two segments. Mandible without lacinia mobilis. Thoracic

492 GEOEFBEY SMITH.

limbs flexed between fifth and sixth segments. No ooste-gites. No appendix interna on pleopods. Hepatic ceecanumerous. Heart elongated, tubular.

Largely as the result of Caiman's writings, the importanceof Anaspides , both as the sole survivor of a group ofCrustacea, otherwise known only from the Permo-Carboni-fei-ous seas of the northern hemisphere, and as the represen-tative of probably the most primitive Malacostracan division,became obvious, so that it was clearly desirable to learn moreabout its habits and internal anatomy, and to find out ifother allied forms were still existing in the freslnvaters of thesouthern hemisphere.

lu the autumn of 1907, at the suggestion and through theassistance of Professor G. C. Bourne, I went to Tasmania toinvestigate Anaspides . On arriving in Melbourne I metMr. O. A. Sayce, of Melbourne University, and learnt that afew weeks before my arrival he had obtained some specimensof a freshwater Crustacean, which he believed to be closelyrelated to Anaspides, from a small stream to the west ofMelbourne. Mr. Sayce has subsequently published anaccount of the animal, which he calls Koonungu cursor(10 and 11), belonging to a separate family, Koonungidas ofthe Anaspidacea. Perhaps the most interesting point aboutthis animal is the fact that, unlike Anaspides and all otherSchizopods, it possesses sessile eyes, a characteristic whichtends to break down the old distinction between Podopthal-mata and Bdriophthalmata, a distinction which Caiman'sclassification also ignores.

My own investigations in Tasmania were directed chieflytowards the elucidation of the obscure points in the habitsand internal anatomy of Anaspides tasmanias, and a pre-liminary account (12) of these matters was published on myreturn in June, 1908. I was also able to report the discoveryof a new species and genus of the Anaspididas, Par an-aspides lacust r is , from the great Lake of Tasmania. Asa result of my studies I inclined to the conclusion that theAnaspidacea, while possessing many peculiar features, were

ON THE ANASPIDACEA, LIVING AND FOSSIL. 493

related by certain characters—e. g. filiform spermatozoa andthe structure of the heart—to the Peracarida, and by othersto the Decapoda, a conclusion which has been subsequentlyconfirmed. The composite character of the Anaspidacea,which seem to be constructed by uniting characteristicstaken from the other divisions of the Malacostraca, appearedto me to point to the extremely primitive nature of the group,and to confirm Caiman's opinion that they should beseparated from the other Malacostracan divisions as a dis-crete group, the Syncarida.

In the September number of the 'Geological Magazine' for1908, Dr. Henry Woodward (13) describes for the first timesome specimens of a fossil crustacean from the Coal-measuresnear Ilkeston, Derbyshire, which must be considered as themost perfectly pi'eserved specimens of fossil Syncarida thathave as yet been found. Dr. Woodward names themPrasanaspides precursor , and there can be no doubtthat they represent an exceedingly close ally of the livingAnaspidacea. The details of segmentation, of the form ofthe limbs, and the general posture of the body in Prce-anaspides are exactly reproduced in the living Anaspidesor Koonunga, and we are amply justified in placing thisancient pakeozoic fossil together with the living genera inthe same order or even in a nearer relationship (text-fig. 3).

Our knowledge therefore of this interesting group ofprimitive Crustacea is beginning to take definite shape, andsince in the future it must always hold a prominent positionin Crustacean morphology a.nd classification, it is, perhaps,timely to bring together all we know about these animals ina systematic form, and to attempt to determine their placein classification, and the light which they throw upon theevolution of the higher Crustacea.

Before leaving the historical aspect of our subject,reference must be made to Professor Fritsch's views uponthe affinities of the fossils which he has described (17), fromthe carboniferous strata of Bohemia. In his admirablememoir he describes a fossil Malacostracan, Gasocaris

494 OEOFFRKY SMJTfl.

Krejci i , which he places together with Gampsonyx,Palaeocaris, Acanthote lson and Mectotelson in asub-order of the Podopthalmata, which he names Simpli-cipoda. Professor Fritsch believes that G-asocaris hadsimple tiniramous thoracic limbs, and he also ascribes thischaracter to the other genera, despite Packard's assertion inregard to Palasocaris, and the apparent condition ofGampsonyx. Professor Fritsch denies that any of theseforms is related to Anaspides, or to any other Schizopod,on the ground of their possessing uniramous limbs. AsCaiman (18) has pointed out, the mere fact, if it wereestablished, that some of these fossil forms had uniramouslimbs would not invalidate the conclusion that they arerelated to Anaspides . This relationship is established moreeffectively by such common characters as the lack of acarapace, the eight free thoracic segments, the pedunculatedeyes and form of the antennae, tail-fan and telson. Withregard to the possession of biramous limbs, we know nowthat Praeanaspides exhibited this character, and the sameis true of Palajocaris and perhaps of Gampsonyx. Itmust also be remembered that the exopodites of the thoraciclimbs iu Anasp ides and its living allies are exceedinglyslender and delicate structures, and that even in thebeautifully preserved fossil Preeanaspides they are by nomeans easy to be made out, though they are demonstrablypresent. We cannot therefore attach great weight toProfessor Fritsch's assertion that they axe altogether absentiu G-asocaris and G-ampsonyx, and even if this is the case, itwould not alter our conviction that these forms are closelyrelated to Anaspides . The fossil, Gasocaris (see text-figs.59, 60, 61), the details of whose structure Professor Fritschso beautifully illustrates, reproduces with great exactitudethe essential features of Anaspides. The pedunculatedeyes, the first antennas with three jointed peduncles, thesecond antennas with their scales, the entire absence of acarapace, the form of the telson and tail-fan, are all nearlyidentical with the con'esponding features in Anaspides.

ON THE ANASP1DAOKA, LIVING AND 1'OSSIL. 49.)

With regard to the segmentation of the body, ProfessorFritsch confesses that lie is doubtful, but the number ofsegments which he gives in his restored figure is plainlywrong. He only figures six thoracic segments in the restoredfigure, but it appears to be demonstrable from his figure ofan actual specimen in ventral view that there are eight freethoracic segments carrying eight similar limbs. It is im-possible not to observe that if only Professor Fritsch, at thetimeof writing, had been fiiniiliar with the living Anaspides,he would have interpreted his fossils otherwise. But whatshall we say of the restoration of Gampsonyx, in which,according to Professor Pritsch, there were seven abdominalsomites besides the telson, and two pairs of maxillipeds infront of the seven pairs of thoracic legs? As Caiman poiutsout, these characters are so exceedingly peculiar as to pre-clude direct comparison with any other known Crustacean,and would remove Gampsonyx from any immediate re-lationship to the Malacostraca at all. While gratefullyacknowledging, therefore, Professor Fritsch's careful descrip-tions of these interesting fossils, we find it impossible tofollow him in his general restorations of them, or in his denialof their relationship to the Anaspidacea. There is one otherpoint in Professor Fritsch's work which may excite a com-ment, and that is the alleged presence of an otocyst on theinner ramus of the uropod in Gasocaris and Gampsonyx.An otocyst in this position is only found elsewhere among theMysidacea; it is not present in the living Anaspidacea or inPrseanaspides.

2. EXTERNAL MOKPHOLOGY.

(A) General Appearance.

The Syncarida are rather small animals, the largest sizebeing attained by the living Anaspides tasmanise, excep-tional specimens of which may measure over two inches inlength. The smallest form known is the living Koonungacursor, which measures about a quarter of an inch in length.

496 GGOFFRKY SMITH.

We may describe, as a type, the general appearance ofA. tasmaniae, the mountain-shrimp of Tasmania, found inthe pools of rivers and in tarns at a high elevation. Thechitinous integument is soft and uncalcified, and of a straw-yellow colour j beneath it in the skin are numerous branchingblack chromatophores, arranged in a similar pattern on eachsegment. Along the dorsal middle line two dark lines arevisible, which are caused by the pigmentation on the floor ofthe pericardium.

hi the natural position the body is held straight and un-

TEXT-FIG. 1.

Pavanaspides lacustris, ?• Lateral view, x 4. a1. Firstantenna. «'•. Second antenna, ntd. Mandible, ep. Gills.Th. 8. Eighth thoracic segment. Ah. 1. First abdominalsegment. PI. 1. Eirst abdominal appendages. T. Telson.U. Uropod.

flexed with the limbs disposed in the characteristic mannersliown in PI. 11, fig. 1. The normal habit of the animal is towalk or run upon the stones at tlie sides or bottom of thedeep pools in which it lives ; this walking movement is effectedby the endopodites of the eight thoracic limbs, but it is alsoassisted by the long exopodites of the abdominal appendages.The exopodites of the thoracic limbs are kept in a continualwaving motion, and no doubt aid in respiration by agitating

0J\r THIS ANASP1DA0EA, LIVING AND i'OSSlL. 497

the water round the delicate leaf-like external gills attachedto the bases of the thoracic limbs.

The body consists of a head, bearing a pair of pedunculatedeyes, and there follow apparently eight free thoracic seg-ments, six abdominal segments and a telson. The sixthabdominal segment carries a pair of expanded, backwardlydirected pleopods, which form a powerful tail-fan.

Paranaspides lacustris (text-fig. 1, and PI. 11, fig. 2),from the Great Lake of Tasmania, although in its detailed

" TEXT-FIG. 2.

Koomingii cursor, (J, from a drawing by Mr. Suyce. X Hi.

structure very similar to Anaspides, differs very widely.from it in external appearance, and in this respect it isprobably the most aberrant of all the Syncarida, includingthe fossil forms. The body, instead of being deeply pig-mented, is of a transparent green colour, sparsely powderedwith black dotsj and there is a very marked dorsal flexure.The abdomen is elongated, the tail-fan enlarged, theexopoditic scales of the second antenna? also enlarged, andthe eyes are borne on elongated stalks. All these characters,which differentiate Paranaspides from the other Auas-pidacea, are correlated with the habits of the animal, which

498 GKOtfFKKY SJJ1TH.

lives among weeds in the littoral region of the lake, ratherafter the manner of a prawn, and pursues more of a swimminghabit than the rest of the order to which it belongs. Thishabit and the characters correlated with it are thereforemost probably a fairly recent acquisition.

The other living representative, Koonanga cursor, is alittle marbled-grey animal which differs from Auaspides illseveral important characters, such as the possession of sessilein place of stalked eyes, the entire absence of a scale on thesecond antennas, and the presence of only seven free thoracicsegments, but it closely resembles Anaspides in generalappearance, especially in its habit of running with the bodyheld straight and uuflexed.

Of the fossil forms we can say for certain that theyfollowed a similar mode of life to Anaspides andKoonunga. There is no trace in any of them of a truedorsal flexure, the fossils being in many cases preserved withthe body quite straight as in the normal walking position ofAnaspides. The tail-fan is small, the external scales notenlarged, and the eyes either shortly pedunculated or possiblyin some cases absent.

The most perfect resemblance to Anaspides is affordedby the English carboniferous fossil Prasanaspides, describedby Woodward. The segmentation and posture of the body,the detailed jointing of the limbs and autennas in this fossilso exactly reproduce the corresponding features of the livingAuaspides, that so far from there being any doubt as tothe two fonns being referable to the same order, it isjustifiable to include them in the same family. The entireabsence of the characteristic leaf-like gills in this and all theother fossil Syncarida is unfortunate, but we could hardlyhope that these extremely delicate and perishable structuresshould be preserved for us in a fossil state.

With regard to the other fossils, although there can besmall doubt that we are dealing with allied forms, it is difficultto be certain about details. Gampsonyx (text-fig. 53) hasa very similar body form and segmentation to Anaspides,


and apparently some of the thoracic limbs were biramous.The first thoracic limb was, however, raptorial and greatlyenlarged. There is less doubt about Pal aso car is (text-figs.56, 57), as the thoracic limbs are distinctly biramous, there

TEXT-PIG. 3.

Prseanaspides prascnrsor,after Henry Woodward, A. Lateralview. B. Dorsal view. c. Telson and ui-opods. D. Fourththoi'acic limb. B. Seventh thoracic limb,

are eight free thoracic segments, and the antennae and tail-fan ai-e very similar to those structures in Anaspides. Theeyes are unfortunately unknown.

Gasocaris (text-fig. 59), despite Professor Fritsch's asser-tion that the limbs are uniramous, was certainly a typicalmember of the Anaspidacea in all other respects.

500 GKOFPEKY SMITH.

Acan thotelson (text-fig. 62) appears to me to occupy adifferent position, and I am doubtful if it is rightly associatedwith the Syncarida at all. The thorax only possessed sevenfree segments, and the thoracic limbs show no trace of beingbiramous. The abdominal limbs are expanded, flabellatestructures, and the tail-fan is elongated and shai-ply pointed.The only resemblance of this creature to the Syncaridais thevery general feature that a carapace is absent, and there isreally no reason for supposing that this fossil is not ageneralised Amphipod. The condition of the eyes is un-known.

(B) Segmentation.

Perhaps the most striking feature of the Anaspidacea isthe entitle absence of acarapace. In Anaspides tasrnaniasthere appear to be eight free thoracic segments and there areundoubtedly six abdominal segments, without counting thetelson. The true segmeutal value of the first thoracicsegment behiud the head has, howevei1, been called in ques-tion by Caiman (8). He inclines to the view that the grooveseparating the head from what appears to be the first segmentcorresponds to the cervical suture in the Decapods., and thatthe apparent segment behind this suture really representsthree segments belonging to the two pairs of maxillae and thefirst thoracic limbs. The compound nature of the segment ispossibly indicated by the definite lateral suture which crossesit on each side in the position shown in text-figs. 1 and 4.

Whether this anterior segment represents three fusedsegments or not, a question which may remain open to doubt,it is certain that its freedom from the head is far morecomplete in Anaspides than in any of the higher Malacos-traca, and that the process of cephalisation has not gone sofar in Anaspides as in the latter. Since, also, there is nodoubt that the first thoracic segment is incorporated in,though it may not entirely represent, this segment, we willspeak of it as the first thoracic segment.

ON TilK ANASP1DACEA, LIVING AND FOSSIL. 501

The segmentation of Paranaspides corresponds exactlywith that of Anaspides, but in Koonunga cursor thereare only seven free thoracic segments, the anterior segmentbearing the first thoracic limbs being definitely fused withthe head, so that the segmentation agrees with the conditionin the more primitive Amphipoda and Isopoda.

Among the fossil Auaspidacea we meet with an interestingcondition. In Prasanaspides there are seven large thoracicsegments, and an extremely narrow segment in front, sepa-

TEXT-FIG. 4.

Paranaspides lacustris. Head with first and second antennaein situ.

rated from the head by a distinct groove (text-fig. 3). InPalseocaris, Gampsonyx, and probably Grasocaris thereare also eight thoracic segments, the most anterior segmentbehind the head being narrow as in Prasanaspides. Theextreme narrowness of this segment suggests that it reallydoes represent the single first thoracic segment which inAnaspides has invaded the head-region, and finally inKoonunga has become fused definitely with the head.Behind this rather problematical first segment the segmenta-tion of the body agrees perfectly in all the Anaspidacea, both

502 GEOFPEEY SMITH.

living and fossil. There is no trace of an extra segment inthe posterior part of the thorax, which has been supposed tobe present in the Euphansiidas.

(c) Appendages.

The first antennse in all the Syncarida present a veryuniform structure ; there is a three-jointed peduncle with twoflagella attached. In all the living forms, and probably inPrseanaspides, there is a definite flexure between the secondand third joints; this flexure does not appear in the fossilGampsonyx and Palseocaris. In all forms, exceptappai'ently Gr amps onyx, the inner flagellum is very much

TEXT-FIG. 5.

5.Koonnnya, ctirsor. First antenna.

shorter than the outer; in Gampsonyx the twoappear to have been of equal length.

In all the living forms an auditory organ has been dis-covered upon the upper surface of the basal joint of thepeduncle of the first antenna. This organ is roughly oval inParanaspides and Koonunga, kidney-shaped in Anas-pides.

It consists of a hollow sac opening on the dorsal innersurface of the basal joint of the first antenna by a narrowtransverse slit. The hollow of the sac is filled with fluid, butthere are no solid concretions of any kind. On the outer wallof the sac is a row of club-shaped chitiiious rods, arranged ina single antero-posterior series. If we study the histology ofthe sac by means of a transverse series (PL 12, fig. 1) we see

ON THil ANASPIDAOEA, LIVING AND i'OSSlL. 503

that the club-shaped rods are fixed into hollow sockets bymeans of a pedicel which is continuous into one of the columnarcells which form the outer wall of the otocyst. From thesecells muscular strands pass outwai-ds and are connected withthe pigmented ectodermal cells upon the outermost wall ofthe antenna. The internal chitinous wall of the otocyst isfurnished with short tooth-like setas. Below these setas areflattened cells which come into connection with fine nervousprocesses sent out from the nerve of the first antenna.

TEXT-FIG. fc>.

scale--J-

6-

Anaspides tasmanias. Second antenna.

The way in which this otocyst functions is not veryobvious. It is clear that the club-shaped rods are not thefinal sense elements since they are not connected directlywith the nerve-endings. The final sense elements areevidently represented by the short setas on the internal wall,which have not been hitherto observed. It appears, there-fore, that the club-shaped rods transmit the stimulus throughthe fluid of the sac to the sensory setas on the internal walland so to the nerve. The muscular apparatus connectingthe club-shaped i*ods with the external ectoderm suggeststhat the original stimulus comes from the exterior, and

VOL. 53 , PART 3 . NEW SERIES. 35

504 GEOFFREY SMITH.

impinges on the ectoderinal cells of the antenna, then thatthese transmit the stimulus to the muscles which pull upon

TEXT-FIG. 7.

Pavanaspides lacustris. Second antenna, bas. Basipodite.cxp. Coxopodite.

TEXT-FIG. 8

8 .

Koonnnga cursor. Second antenna.

the club-shaped rods. Theso in their turn transmit theimpulse to the setee and so to the nerve.

If this interpretation is correct the otocyst of the A.naspi-


dacea, although agreeing essentially with that of the Decapoda,differs from the latter in responding to stimuli from theexternal world, as well as to internal stimuli set on foot bythe position of orientation.

The second antenna consist of a two-jointed protopodite,which supports an exopoditic scale and a flagellate endopodite.The scale is comparatively small in Anaspides (text-fig. 6)

TEXT-FIG. 9.

9.

Auitspides tasmaniaj. Mandible.

and the fossil forms; it is considerably enlarged in Par an-aspides (text-fig. 7), probably in correlation with theswimming habit, and is altogether absent in Koonunga(text-fig. 8). The typical form of this antenna with itsexopoditic scale and flagellum is characteristic of the" Schizopoda " and Decapoda.

The inaudible in Anaspides tasmanite (text-fig. 9)

506 GEOPfltEY SMITH.

has the biting face divided into three regions—an uppertoothed portion, which differs slightly in the right and leftmandible, a lobe bearing a row of spines, and a lower molarsurface. The laciniamobilis, characteristic of the Peracarida,is absent. The palp is three-jointed. The mandible ofParanaspides lacustris (text-fig. 10) has the samestructure as the above, but the palp has a peculiar formation,which is particularly well marked in old specimens. In oldspecimens it appears to be distinctly four-jointed, and thebasal joint carries a very definite, little, external branch

TEXT-PIG. 10.

Paranaspides lacustvis. Mandible.tipped with two setae. In young specimens the extra joint,i .e . between segment two and three, may be absent, and theexternal branch is not so conspicuous. The external branchoccupies the position of an exopodite, and if the mandibularpalp in this form is really biramous, it would be unparalleledin Crustacea except among the Copepoda and Ostracoda.Considering, however, that Paranaspides is otherwise arather specialised form, and that the character in questionis best marked in old specimens, it seems doubtful if we arereally dealing with a primitive characteristic.


The mandible of Koonunga (text-fig. 11) possesses atoothed ridge and a single lobe beneath it bearing short

TEXT-FIG. 11.

Koonunga cursor. Mandible.

spines and a few small papillae. Sayce regards this lowerlobe as the molar surface, the spine row being, according to

TEXT-FIG. 12.

pafy-

1 2 .

Anaspides tasmanise. First maxilla.

him, absent, but since the molar surface is not normallyfurnished with setse in this manner, it seems better to regard

508 GEOPFEEY SMITH.

the lower lobe as corresponding to the spine row of the otherforms. The palp is three-jointed, with a very short terminajoint.

Apart from the abnormal condition of Koonnnga, the

TEXT-FIG. 13.

13.Paranaspides lacustris. First maxilla.

mandible of the Anaspidacea resembles that of the Mysi-dacea, except that the 1 acini a mobilis, characteristic of thelatter, is absent in the former.

In the Euphausiacea and Decapoda the mandible consists

14. \Koonunga cursor. First maxilla.

of biting lobe and molar surface without any spine row orlacinia mobilis. The Anaspidacean mandible is thereforeintermediate in structure between that of the Peracarida andEucarida.


The first maxilla of Anaspides (text-fig. 12) consistsof two biting blades, the upper one armed with stiff spines, the

TEXT-FIG. 15.

palp.

15.

Anaspides tasnianise. Second maxilia. ex. Exopodite.1, 2, 3, 4. Gnathobasic lobes.

TEXT-FIG. 16.

Paranaspides lactistris. Second maxilla, ex. Exopodite1, 2, 3, 4. Gnathobasic lobes.

lower with plumose setae; a palp is present in the form of asmall conical tubercle. In Paranaspides the structure isessentially similar, but the palp is rather more conspicuous. In

510 GEOFFREY SMITH.

TEXT-PIG. 17.paJp

11.Koonunga cursor. Second maxilla, ex. Exopodite. 1, 2,

3, 4. Gnathobasic lobes.

TEXT-FIG. 18.

-.-•9T

1 8 .

Anaspides tasmania'. First thoracic limb. ep. Gills ex.Exopodite. gn. Gnathobasic lobes.

ON THE ANASPIDACKA, LIVING AND FOSSIL. 511

Koonunga the palp is still more developed, and the lowerbiting blade is reduced, being tipped with only three setae.

The second maxilla (text-figs. 15, 16, 17) has a veryuniform structure in the three genera. It may be interpretedas consisting of four gnathobasic lobes, a palp which has takeaon the function of a gnathobase, and an exopoditic lobe, whichis well developed in Paranaspides, much reduced in Anas-pides, and absent in Koonunga.

The structure of this maxilla resembles that of the Mysi-TKXT-FIG. 19.

Paranaspides lacustris. First thoracic limb.

dacea more closely than that of the Euphausiacea, especiallyin the distinctness and arrangement of the gnathobases.

Among the fossil forms the structure of the mandibles andmaxillae has not been elucidated.

The first thoracic limb (text-figs. 18, 19, 20) differsonly in detail from those of the following segments. It con-sists of a pediform endopodite with a reduced lamelliformexopodite. In Anaspides and Paranaspides the limb iscomposed of eight distinct segments, a protopodite of twojoints and an endopodite of six. The coxopodite in both forms

512 GEOFFREY SMITH.

bears a pair of delicate leaf-like gills externally, whileinternally, i.e. towards the mouth, two gnathobasic lobes aredeveloped. The limb is flexed between the fifth and sixthsegments, so that there are three segments distal to the "knee-joint" and five pi'oxirnal to it. The terminal segment is short,and carries, as in all the thoracic limbs of the Anaspidacea,three enlarged setae, of which the central one is the largest.

In Paranaspides the first joint of the endopodite (i. e.

TEXT-FIG. 20.

20 .Koonunga cursor. First thoracic limb.

propodite) is expanded towards the mouth to form a definitelobe1.

In Koonunga the structure of this limb is aberrant.The coxopodite is entirely without gnathobases. There areonly seven segments in the limb, and since there are threesegments distal to the "kuee-joint" it is clear that a segmenthas disappeared from below, i. e. proximal to, this joint.There is no doubt that what has happened is that the propo-dite or first joint of the endopodite has fused with the basi-


podite, a process which can be observed to occur in some ofthe posterior limbs of Anaspides. If this is so, it is clearthat the exopodite no longer springs from the basipodite, asit normally should do, but from the fused basipodite andpropodite.

The second to the sixth' thoracic limbs may be treatedtogether, as they only differ in unimportant details. As a

TEXT-FIG. 21.

2 1 .

Koontinga cursor. Second thoracic limb. ex. Exopodite. en.Endopodite. ep. Gills.

type we may describe the fourth thoracic limb of Paranas-pides (text-fig. 22). This limb is composed of a pediformendopodite and a flagellate exopodite borne on a two-jointedbase. The coxopodite bears a pair of gills. We see inthis limb the process of fusion of the basipodite with thepropodite, the joint between them being repi'esented by agroove which does not completely traverse the fused seg-ments. In the more anterior limbs and especially in youngspecimens of both Anaspides and Paranaspides the

514 GEOFFKEY SMITH.

groove may be complete. In the second thoracic limb ofKoonunga (text-fig. 21), and indeed, in all the thoraciclimbs of this form, the fusion of the basipodite and propoditeis complete, so that the limb appears to be constantly formedof seven joints instead of eight. In all cases, nevertheless,there are constantly three segments above the " knee-joint."

la the female sex of all the three genera, the coxopoditeof the fifth, sixth and seventh limbs bears on its internal

TEXT-PIG. 22.

Paramispides lacustris. Fourth thoracic limb.

face a small setose lobe (text-fig. 24). The function ofthis lobe, which is entirely confined to the female sex, isunknown, but since the lobes are only present in the hinderlimbs in the neighbourhood of the oviducts and spermatheca,it may be suggested that they assist the process of fertilisa-tion in some way. The openings of the oviducts in the femaleare situated on the coxopodites of the sixth pair of limbs(text-fig. 27).

The seventh thoracic limb differs from the foregoing inthat in Anaspides and Paranaspides the exopodites are


reduced to small unsegmented lamellte, while in Koonungathey are absent altogether. In Anaspides the limb is eight-jointed ; the basipodite, carrying the exopodite, is very small,but separated from the propodite by a distinct groove. InParanaspides this groove is very incomplete, while inKoonunga there is no groove at all, the two segments being

TEXT-FIG. 23.

2 3 . '

Anaspides tasmanias. Seventh thoracic limb of male.

entirely fused. Attention has already been called to thepresence of the setose lobe on the coxopodite of this and thetwo preceding limbs in the female sex.

The eighth thoracic limb (texfc-fig. 25) in all threeliving genera is composed of apparently seven joints, thesecond and third segments having no doubt, from the analogyof the preceding limbs, fused together. The exopodite has

516 GKOFFBEV SMITH.

completely disappeared and there are no gills attached to thelimb.

In the male the openings of the vasa deferentia are situatedon the inner edges of the coxopodites of the eighth thoracicappendage (text-fig. 26).

In the females of the three living genera a very conspicuousspermatheca is to be seen, placed between the last pair ofthoracic limbs in the ventral middle line (text-fig. 27). Itconsists of a large conical papilla with a single median opening

TEXT-PIG. 24.

lobe

Paranaspides lacustris. Seventh thoracic limb of female,which leads into a wide tube. The tube bifurcates into twoslightly ramifying passages in which spermatozoa may fre-quently be found. The tube and its passages are lined withchitin secreted by a definite epithelium, and round theramifying bifurcations a cellular tissue is aggregated of aconnective or supporting character, probably with much thesame function as cartilage. This tissue is also found in thelabrum, and its peculiar histological character is shown ouPI. 12, fig. 17.

ON THE ANASPIDACEA, L1VIKG AND FOSSIL. 517

The presence of this spermatheca is of considerable taxo-nomic importance, as it appears to be entirely absent in theother Schizopods, viz. Mysidacea and Euphausiacea, but to

TEXT-FIG. 25.

Paranaspicles lacustris. Eighth thoracic limb.

be present in certain of the more primitive Decapods. Inthe lobster and certain prawns a similar spermatheca is

TEXT-PIG. 26

Paran spides lacustris. Basal segment of eighth thoraciclimbs, showing male openings.

present, and in the peculiar Eryonidea (Polycheles,V\rillemoesia) the presence of a spermatheca in the sameposition was pointed out to me by Mr. Gray, of the Oxford

518 GE01?PBEy SMITH.

University Museum. The investigation of this spermathecain the female of P olych eles has revealed a structure identical,with that of Auaspides. The spermafcheca of Polycheles isa shield-shaped chitinous structure with a median openingleading into a tube which bifurcates exactly as in Anaspides.

TEXT-FIG. 27.

Th.6

TK.8

2 7.

Paranaspides lacustris, ?. Ventral view of the last threethoracic appendages and first abdominal of left side in situ.Th. 6. Sixth thoracic limb. Th. 7. Seventh. Th. S. Eighth.Abel. 1. First abdominal. Spermth. Spermatheca. $ . Femaleopening.

There cau be no doubt that the structure in both cases isstrictly homologous, and that we have in the spermatheca ofAnaspidacea a Decapodan character, parallel to the presenceof the otocyst on the first antennas.

The thoracic limbs of the fossil Syncai-ida may now bedealt with as far as they are known, and those of Praaanas-

ON THE ANASPIDAOEA, LIVING AND FOSSIL. 519

pides (text-fig. 3), as being the best known, will receivefirst attention. The first thoracic limb is unknown. The secondlimb was apparently composed of three or four small basaljoints and an expanded segment below the "knee," andprobably three segments above the " knee." Except that noexopodite can be seen, it appears to have agreed with thecorresponding limb of the living Anaspidacea. The succeedingthoracic limbs agree very perfectly with those of living forms.There was a two-jointed protopodite from which sprang aflagellate exopodite and a stoat five-joinfced endopodite, threeof these segments being distal to the knee as in living Anas-pidacea. Iu the last two thoracic limbs it is impossible tomake out an exopodite, and this is again in agreement withthe structure of the living genera. In the most perfectlypreserved limbs it is only possible to make out seven segmentsin each limb, so that the fusion of the second and third segmentsmay liave already taken place in this form, but it is moreprobable that there were eight segments and that the con-dition of preservation does not permit us to see them all.

In Gampsonyx and Palseocaris we can only obtain avague idea of the structure of the limbs. In the former thesecond limb was apparently of a raptorial nature, beinggreatly enlarged and furnished with prominent spines ; thesucceeding limbs one can only describe vaguely as biramous.In Palasocaris, if we can trust the diagrammatic reconstruc-tion of Packard (text-fig. 56), the last six thoracic limbs wereall similar and all biramous, with stout endopodites and slenderexopodites.

In G-asocaris the endopodites are all similar and stoutlybuilt, but Fritsch denies the presence of exopodites at all, adenial about which we may suspend our judgment, owing tothe delicacy of the exopodites in the Anaspidacea and thedifficulty of making them out even in the best preservedfossils.

The abdominal appendages, 1-5 in the females ofAn as pides and Parauaspides, have all a very similar struc-ture, except the fifth, which is without an endopodite. The

VOL. 53 , PART 3. NEW SERIES. 36

520 GEOFFREY SMITH.

limb consists of an expanded protopodite of two segments,from which springs a long setose and many-jointed exopodite

TEXT-FIG. 28.

2 8 .Paranaspides lacustris. Third abdominal appendage.

En. Endopodite. Ex. Exopodite.

.-•Th.8

...-Ahdl.

Paranaspxdes lacustris. First two abdominal appendages ofmale in ventral view in situ. Th. 8. Eighth thoracic limb.Abd. 1. First abdominal. Abd. 2. Second abdominal appendage.

and a very small flabellate endopodite. This endopodite isabsent on all the first five abdominal appendages ofKoonunga .


In the males the endopodites of the first two pairs ofabdominal limbs are curiously modified as copulafcory styles(text-fig. 29). They are large, hollow, club-shaped organs,well armed with setae and hooks, disposed in a characteristicpattern, and the endopodites of the second segments fit intohollow spaces excavated in the first in a piston-like manner.The exopodites of these limbs are of a normal form.

The presence and structure of these copulafcory styles is adistinctly Eucaridan feature, recalling similar structures inthe Buphausiacea and Decapoda.

TEXT-FIG. 30.

3 0 .

Anaspides tasmanise. Telson and uropods.

The abdominal limbs of Prasanaspides show a longsetose exopodite, much as in the living forms; the endo-podite was apparently much reduced, as iu living Anas-pidacea. The reconstruction of Paleeocaris also showsuniramous abdominal appendages, so that they also apparentlyagreed well with the living forms. The vague reconstructionof Gampsonyx shows us apparently biramous limbs with thetwo branches of equal length, but it is very likely that thelong setas on the exopodites have been confused with endo-podites, an error that might occur in reconstructing Pras-anaspides.

The sixth abdominal limbs or uropods and the telson

522 GKOFFRKY SMITH.

have very characteristic forms, which are depicted in text-figs. 3, 30, 81, 32, 33, 55, 58, 6].

It will be noted that the uropods and telson of Paranas-pides are more elongated than in the other genera, incorrelation with the swimming mode of life.

Special attention must be called to the close correspondencebetween t*he nropods of Praeanaspides (text-fig. 3) andAnaspides (text-fig. 30), even individual seta?, e.g. those

TEXT-FIG. 31.

Paranaspides lacustris. Telson and uropods.

on the outer border oE the exopodites, being practicallyidentical in the two cases. The short setss also upon themargin of the last two segments in Anaspides andParanaspides are also represented in the fossil. Thedetailed correspondence in these and the other limbs ofAnaspides, Paranaspides and Preeanaspides permitsus to place them with confidence in the same family. Theuropods and telson of Gampsonyx, Palteocaris andGasocaris agree on the whole very accurately with the

ON THE ANASPIDACEA, LIVING AND FOSSir,. 523

other members of the Syncarida. The uropods and telsouof Acanthotelson appear to have possessed a very aberrantform.

Attention may again be called to the observation of Pro-fessor Fritsch. that in Gasocaris and Gainpsonyx an ovalswelling is present near the base of the inner raraus of theuropod, which he interprets as an otocyst (text-figs. 55, 61).Nothing of the sort is to be observed in any of the living

TEXT-FIG. 32.

Koonunga cursor. Telson and uropods.Anaspidacea nor in Prasanaspides. If this structure reallyis an otocyst we are evidently dealing with a character whichmust have belonged originally to the Syncarida and theprimitive Enmalacostraca, and has been retained in certainPeracarida (Mysidacea) but lost in the higher Syncarida,,Peracarida and Eucarida.

(D) Theoretical Considerations.

A theoretical consideration of the appenda.ges of theSyncarida with their associated organs will bring out several

524 GEOPFEBY SMITH.

points of importance. The nature of these structures isplainly of a generalised and primitive character. In thefirst place all the appendages, with the exception of the firstantennae, are either typically biramous or have departed verylittle from the biramous plan. In the next place it is impos-sible, on the character of the appendages, to place the Anas-pidacea in either the Peracarida or Eucarida, or any otherdivision of the Malacostraca. The mandible is of a Peracaridan

TEXT-FIG. 33.

End of telson. A. In Anaspides tasmanise. n. Paranaspideslacustris.

type, but lacks the lacinia mobilis; the maxillas are possiblynearer those of the Peracarida than of the other divisions,but a palp is still present on the first maxilla and the exopoditeof the second maxilla is greatly reduced.

On the other hand the otocyst on the first antennas is paral-leled only by the Decapoda, as is also the spermatheca at thebase of the Jast thoracic limbs, while the copnlatory styles arealso Eucaridan.

The segmentation of the thoracic limbs is also very inter-esting as affording us a primitive condition from which thatof the Peracarida on the one hand and that of the Eucaridaon the other can be severally derived. We have seen that

ON THE ANASPIDACIA, LIVING AND FOSSIL. 525

the primitive Anaspidacean condition of these limbs is thepossession of eight segments or joints, of which three areplaced distally to the " knee" while five are placed proxi-mally (e.g. text-fig. 18).

In the Peracarida and Eucarida we observe typicallyseven segments, but the " knee " in the two divisions, aspointed out by Hansen (14), is in a different position. In thePeracarida (e. g. Mysidacea) (test-fig. 34) there are two

TEXT-FIG. 34.

First thoracic appendage (maxillipede) of Macrouiysis flexuosa.

segments distal to the "knee," and primitively five proximalto it; in the Eucarida there are three segments distal to the"knee" and four proximal to it (text-fig. 35). Now wehave seen that in the Anaspidacea there is a tendency forthe second and third segments to fuse, and this process carriedto completion would give us the Eucaridan limb. In thePeracaridan limb we may suppose that a segment has dis-appeared distal to the knee, probably the terminal segment,which in the Anaspidacea is very small.

526 GEOFFREY SMITH.

In this manner it would appear that the "knee-joint" iuPeracarida and Eucarida is homologous, while a differentsegment has been suppressed in each case, in the Peracaridathe terminal segment, and in the Eucarida the secondsegment from the body, which has fused with the third.

Hansen has suggested that the claw usually present on theterminal joint of the Peracaridan limb represents the lost

TEXT-FIG. 35.

3 5 .

First thoracic appendage of Nyctiplianes.

teriuinal segment, and this may possibly be the case, thougha theoretical refinement of this nature must always remaindoubtful.

However this may be, the Anaspidacean limb with itseight segments gives us the generalised condition fromwhich the Peracaridan and Eucaridan types can be easilydeduced.


TEXT-FIG. 36.

Anasp ides tasmanise. Dissection from dorsal sui-face, to showheart and pericardium (per.). Ost. Ostium. Art. des. Place oforigin of descending or sternal artery from the lieart.

528 GEOFFBEY SMITH.

3. INTERNAL ANATOMY (Anaspides tasmanias).

(A) Per icard ium, Hea r t and Vascu la r System.

If a median dorsal incision be madein Anaspides , and theskin aud dorsal muscles be turned away on each side, wefind that we are in a spacious division of the body cavity witha very definite pigmented floor, which stretches from thefirst to the last segment. This space is the pericardium andits floor is pierced laterally in each segment by lacunarspaces which lead down into the cavities surrounding thebases of the appendages. The heart, which lies in the peri-cardium, is a long contractile tubular structure with ananterior expanded portion stretching from the first to theeighth, thoracic segment; in the first abdominal segment itnarrows considerably and is continued into a vessel whichappears to be contractile, but is perhaps more rightly to beconsidered as a posterior dorsal aorta. The heart and aortaare fixed to the floor of the pericardium by a series of inter-seginental short muscles, while the heart in the thoracicregion is also supplied with segmental dorsal alary muscles.The heart is constricted intersegmentally where the ventralmuscles attach it to the floor of the pericardium, but ostiaare only present in one place, namely at the base of thethird thoracic segment, where there appears to be a singlepair.

Anteriorly the heart gives off three arteries, a medianophthalmic artery and paired antennary ai-teries.

In the seventh thoracic segment a sternal artery leavesthe heart, and running obliquely downwards enters a ventralartery in the sixth thoracic segment, which runs forwards andbackwards just dorsal to the nerve cord (text-fig. 44, art.vent.). A small subneural vessel is also present.

The elongated tubular heart is of a strictly Peracaridantype, and recalls very strongly the heart of the Mysidse; thearterial system is of a generalised Malacostracan nature.The dorsal muscles which form the roof of the pericardium

ON THE ANASP1DACEA, LIVING AJSD FOSSIL. 529

are arranged in four lateral longitudinal bands constrictedintersegmentally.

The blood-corpuscles appear in sections as oval cells ofvarying size, but generally with similar nuclei. Above andto the sides of the cardiac division of the stomach there is aconspicuous mass of tissue with crowded darkly stainingnuclei and without definite cell outlines, in which verynumerous mitoses may be observed, even in an adult fully-grown animal. At the edges of this tissue, which lies free inthe hasmoccel, cells can be seen to be detaching themselveswhich have the appearance of blood-corpuscles. This tissue(PL 12, fig. 2) is present in the same position in " Schizopoda"and Decapoda which I have examined, and there appears tobe little doubt that it constitutes the blood-forming organof the higher Crustacea in which the blood-corpuscles arereproduced.

(B) The Alimentary Canal.

On removing the heart and the floor of the pericardium wefind the alimentary canal with its associated glands. Wewill first shortly enumerate the various parts of the alimentarycanal, beginning from in front backwards.

'The short dorso-ventrally directed cesopliagus leads into anexpanded stomach which is furnished with a series of com-plicated setose ridges. At the pyloric end of the stomach ashort median dorsal diverticulum marks the position wherethe mid-gut or endodermal portion of the alimentary canalbeginsaud thestomoda3um ends. At the same point, butventro-laterally, a great number of long, slender, liver cseca enterthe stomach, to the number of about thirty.

The mid-gut is continued downwards as a straight tubeuntil the second abdominal segment, where a large andconspicuous dorsal diverticulum is given off. This diverticulum,which is of a glandular nature, belongs to the mid-gut, but itdoes not mark the place where the pioctodteum begins. Thisposition is marked by a third diverticulum in the fifth abdo-

TEXT-FIG. 37.

ovdL.

CLLV. 3.

0 7.

Anasp ides tasmaniffi. Dissection from dorsal surface, heartand pericardium removed, to show ovary (ov.) on one side andalimentary canal. Left ovary removed. Ovd. Oviduct. Div. 1.First dorsal diverticulum. Div. 2. Second. Div. 3 Third.

ON THK ANASPIDACBA, LIVING AND FOSSIL. 531

initial segment, which also belongs to the mid-gut, butimmediately below it the character of the epithelium entirelychanges, and a chitinous lining, marking the proctodajalinvagination, covers the internal surface of the intestine.The anus opens ventrally on the telson.

We will now describe these various portions in moredetail.

The stomach may be divided into an anterior cardiac auda posterior pyloric portion. The division between them ismarked dorsally by the first diverticulum of the alimentary

TEXT-FIG. 38.

•cfcv.7...

3 8 .

Anaspides tasmanise. Stomach removed from body and viewedlaterally. For lettering see text.

canal, and ventrally by the entrance of the liver casca, sothat the cardiac portion is stomodeeum while the pyloric isendodermal, although chitinous pieces of ectodermal originare projected into the pyloric division.

The cardiac portion of the stomach is furnished with apair of lateral elevations carrying setae. Each lateral eleva-tion is in the form of an incomplete circle, the setas beingpresent in two main pieces of the ridge, viz. C and D in text-figs. 38, 39, 40. There is also present a median setose pro-minence (E), and a more posteriorly placed tooth (F) whichprojects into the pyloric cavity. In the median ventral linethere is a prominent pad (H) (text-iig. 40) which projects

532 GEOFFKEY SMITH.

into the cavity of the stomach and stretches into the pyloricportion (see also transverse section, PI. 12, fig. 3).

The pyloric portion of the stomach is furnished with apair of very long setose ridges on each side, the hindermostridge of each side being produced backwards far down theintestine (A and B, text-figs. 38, 39, 40).

TEXT-FIG. 39.

E—

39.

Stomach from dorsal surface, showing chitinous ridges, etc.

In comparing this armature of the stomach of Anaspideswith other Malacostraca, I make use especially of Gelderd'suseful work on the digestive system of the Schizopoda (15).

The lateral pieces of the cardiac division (c, D) correspondto Gelderd's ridges s, which he finds to be present in allMalacostraca; similarly the lateral pieces of the pyloricdivision (A, B) correspond to his ridges S2, which are also

ON THE ANASP1DACEA, LIVING AND FOSSIL. 533

universally present. The median pieces (E and F) correspondto the urocardiao ossicle and median tooth of the Decapoda.The piece (E) is present in most Malacostraca, but the posteriorpiece (p) possibly points to Decapodan affinities. The greatlength of the lateral pj'loric pieces (A and B) in Anaspidesresembles the condition of the EuphausiidEeandnotthatof the

TEXT-FIG. 40.

4-O.Stomach from ventral surface.

Peracarida, in which these pieces are short. Again, in thePeracarida there is a complicated median ventral l-idgein thepyloric division, which in the Euphansiidas is simpler andappears to be absent in the Anaspides, while the elongatedventral ridge (H) in the cardiac division of Anaspides recallsclosely that of the Euphausiidse. We may therefore concludethat in those points in which the gastric mill is not merely Mala-

534 GEOFFREY SMITH.

cosfcracau, it inclines to the Eucaridan rather than thePeracaridan type.

The f i rs t dorsal d iver t iculum to the gut is a thick-walled pocket which opens into the pyloric division of thestomach by two lateral passages on the outside of the lateralpyloric ridge (B). In the section shown in PI. 12, fig. 4,only one lateral opening is seen on the right side.

The histological character of this diverticulum is extremelypuzzling (PI. 12, fig. 4). It consists of thick walls in whichcrowds of nuclei are densely packed without definite cell-outlines round them, and many of the nuclei are seen to beeither actually undergoing mitosis or else in prepai*ation fordivision. This process of mitotic division is to be observedin fully grown adult specimens, so that we are evidently deal-ing with a tissue which l'emains in a permauently embryoniccondition. There are no cells of a glandular nature in thisdiverticulum, as might well be expected from its position onthe alimentary canal.

Ifc is impossible to do more than guess at the function ofthis remarkable orgau, but from the active reproduction ofthe cells composing it, it may be suggested that its functionis to supply new epithelial cells for the lining of the alimen-tary canal as the old cells wear out and become effete.

Laterally the walls of the fh'st diverticulum pass into theepithelium of the mid-gu t or endoderra (PI. 12, fig. 5).The portion of the mid-gut lying between the first andsecond divertioula is essentially glandular in nature, especiallyin the anterior and dorsal region. In this region the epithe-lial cells are tall and columnar. The majority of these cellshave lightly-staining oval nuclei, but wedged in between theseordinary columnar cells are nests of much-flattened cells withspindle-shaped nuclei, which stain intensely with hasmatoxylin.These cells are apparently special gland-cells of some kind,which pour a secretion into the gut. Veutrally and posteriorlythe liuiug epithelium of the mid-gut is composed of shortcubical cells with rather darkly staining nuclei.

The basement membraue of the mid-gut is remarkably

ON THE ANASPIDAOEA, LIVING AND FOSSIL. 535

thick and conspicuous; and it is thrown into markedsinuosities of outline; this membrane stops abruptly whenthe mid-gut passes over into stomo- or proctodasum.Exteriorly to the basement membrane is a thin submucouslayer with flattened darkly-staining nuclei (PI. 12, fig. 5).

The second dorsal di verticulum of the mid-gut is byfar the largest. Histologically (PI. 12, fig. 6) it is simplyformed as a pocket from the dorsal mid-gut epithelium, andits cells are elongated and columnar with very numerousnests of flattened special gland-cells. There can be no doubtabout the function of this diverticulum; it is simply a diges-tive gland which pours its secretions into the alimentarycanal.

The remaining portion of the mid-gut lying between thesecond and third diverticula is of a simple structure, and isevidently chiefly oE an absorptive nature (PI. 12, fig. 7).The epithelium is composed of short columnar cells withstriated outer borders; the basement membrane, charac-teristic of the mid-gut, is still to be observed, while thesubmucosa forms a thick reticular laj'er in which largehomogeneously staining nuclei are embedded. There are nogland-cells in this region.

The third dorsal diverticulum is exceedingly small,and is composed of columnar cells and numerous nucleiembedded in a common protoplasm. There are no specialgland-cells. A fair number of mitoses can always be observedin this diverticulum, though not so many as in the firstdiverticulum; its function may be similar to that suggestedfor the first, viz. to keep up a supply of epithelial cells forthe lining of the alimentary canal. The intestine behind thethird diverticulum is proctodteum, being lined internally withchitin, which is thrown into numerous folds. The epitheliumis columnar and hyaline, being very similar to that of thestomodteurn.

It remaius to describe the liver. This organ is composedof very numerous slender tubes, as many as thirty beingoften present, which open ventrally into the pyloric division

VOL. 53 , PART 3. NEW SERIES. 37

536 GKOFFEEY SMITH.

of the stomach by a wide common opening. The histologicalcharacter of these tubes varies greatly according to thecondition of metabolism. A transverse section through themiddle of a tube, when the animal is starved, shows a regularlining of tall columnar cells, the majority of which havehyaline, reticular cytoplasm, staining pink with eosin, andgranular nuclei (PI. 12, fig. 8). These cells are mainlyabsorptive in function. Scattered among these cells arenarrower cells with rather coarsely granular cytoplasm, whichis darkly coloured with liEematoxylin. These cells arespecial gland-cells and probably secrete a ferment.

At the ends of the tubes the nuclei are greatly crowded,but both kinds of cells can be recognised.

After feeding heavily, the histology of the tubes changes(PI. 12, fig. 9). The special gland-cells are no longer recog-nisable, and the absorptive cells lose their definite cell-outlinesand are distended with oily globules. Certain of the cellsconsist merely of au envelope containing an immense vacuolewith a darkly staining nucleus flattened on one side. Theliver of Anasp ides therefore has, at least, two distinctfunctions; it produces digestive ferments which are pouredinto the stomach, and it also plays an important part in theabsorption and storing of assimilated material.

If we compare the a l imen ta ry t rac t of Anaspides withthat of other Malacostraca we see that the structure of thestomach points rather to Eucaridan affinities. The presenceoE dorsal diverticula at the juncture of endodermnl mid-gutwith* stomodseum and proctodasum is a Decapodan character,since in the " Schizopoda/' i .e. Euphausiidae and Mysidse,etc., there is never a diverticulum between mid-gut andproctodceum. We have seen that there is also another diverti-culum in the middle of the mid-gut, and this character is, asfar as we know, peculiar to Anaspides and its immediateallies.

The liver, although in certain respects peculiar, is nearerto the Eucaridan plan than to the Peracaridan, since in thelatterthere is typically present a glandular ridge upon which the

ON THIS ANASPiDACEA, LIVING AND FOSSIL. 537

gland-cells are concentrated, while in the Eucarida andAnaspides the gland-cells are distributed about among theabsorptive cells.

On the whole, therefore, the alimentary tract of Anas-pides, while showing certain peculiar features, points, toEucaridan and especially Decapodan affinities.

(c) Excretory System.

The excretory organ of Anaspides is situated at the baseofthe second maxilla; it is a maxillary gland. ~We can dis-tinguish four chief regions : (1) A straight excretory ductwith rather thick walls and darkly staining somewhat flattenednuclei (PI. 12, fig. 10). This duct passes into the baseof the second maxilla on each side and opens by a pore onthe external border of the appendage. (2) The excretoryduct passes internally into a coiled excretory tube withstriated walls and greatly flattened darkly staining nuclei(PL 12, fig. 11). (3) This coiled tube passes insensibly intoanother coiled tube with an epithelium of a more glandularnature and with oval nuclei containing granules of chromatiu(PI. 12, fig. 12). The cytoplasm of these cells is moregranular but has a faintly striated appearance. (4) Theglandular tube is coiled into an expanded sac, the end-sacinto which it opens. The end-sac is lined with a flattenedepithelium, the cells of which contain globules of a yellowishcolour (PI. 12, fig. 13).

The presence of a maxillary gland, and the entire absenceof an antennary gland, is only found elsewhere in theMalacostraca among certain Isopods and in Nebalia. It isunknown either in the " Schizopoda" or Decapoda.

(D) Reproductive Organs.

Female.—The external sexual characters of the female,together with the spermatheca, have been described (pp.516—518). The ovary of an adult Anaspides is a lobed

538 GEOFFREY SMITH.

structure stretching from about the second thoracic segmentto the extreme hind end of the abdomen (text-fig. 37). If wetake a horizontal section through the ovary (PI. 12, fig. 14)we see that the lobes contain small immature ova while theextei-nal part of the tube is filled with large, nearly matureova filled with a purplish yolk. The external wall of thelobed portion consists of small undifferentiated cells of thegerminal epithelium. The inner wall of the ovary consistsof cells, most of which contain large nuclei which stain of anuniform dark colour with haematoxylin, while there is pi'esenta number oE granules which stain deep black. These cellsmay be called the trophic cells. On their inner borders theyare vacuolated, and it is evidently their function to elaboratefood material, which they supply in the form of yolk to thedeveloping ova. A certain number of these trophic cells canbe seen lying among the eggs in the middle of the tube.The oviducts are simple straight tubes lined with shortcolumnar cells; they pass below the ventral muscles to openon the coxopodites of the sixth thoracic limbs. They are notsupplied with any accessory glands.

Male.—The testes (text-fig. 41) are coiled tubes runningfrom the anterior thoracic region to the extreme hind end ofthe body. At the anterior end of the coiled tube a ductpasses posteriorly, and then turning anteriorly again opensat the bases of the eighth thoracic limbs.

In the glandular part of the testis of a young Anasp ideswe can study the process of spermatogenesis (PL 12, fig. 15).We see the primary and secondary spermatocytes under-going mitosis, and fully-formed spermatozoa. On the out-side of the tube we see large cells with darkly stainingnuclei of an exactly similar appearance to the trophic cellsfound in the ovaries. These cells are not, however, to beobserved in an adult testis. In an adult testis we only seenests of spermatocytes in various stages of spermatogenesis,or else groups of fully formed spermatozoa.

The upper part of the descending duct is sterile, so far asthe production of spermatozoa is concerned, and it is formed


purely of trophic cells (PI. 12, fig. 16). The lower portion ofthe duct has a thick epithelial wall with oval granularnuclei, and these cells produce an albuminous material thatis cast into the lumen of the duct and solidifies round the

TEXT-FIG. 41.

UsUsk

Anaspides tasinanise. Testis and vas deferens of one side.

spermatozoa to form the spermatophores. There is a thickmuscular layer surrounding the lower part of the vas deferens.In the section (PI. 12, fig. 16) the top of the speruiatophor,with its albuminous coat, is cut across.

The spermatozoa are filiform elongated bodies with a con-

540 GEOFFREY SMITH.

spicuous globular head and a long flagelluru (text-fig. 42).They resemble closely the spermatozoa of all the Peracarida.

The spermatophores are horseshoe-shaped bodies about halfan inch in length, with a constriction near the middle. Theypossess a fine chitinous investment on the outside, as well as

TEXT-PIG. 42.

4 2 .

Spermatozoon of Anaspides tasmanise. X about 50.

the albuminous material within. They are placed by themale into the spermatheca of the female, where they may besometimes found protruding their horn-like ends to the

TEXT-FIG. 43.

43 .Spermatopliores of Anaspides tasmaniiB. x 3.

exterior. After the spermatozoa have passed out of theminto the spermatheca they soon drop off.

The filiform character of the spermatozoa is a Peracaridancharacter; the presence of definite spermatopliores isEucaridan.

TEXT-FIG. 44.

Th.l.

Th.8,

Anaspides tasmaniie. Alimentary canal and gonads removedto expose nervous and muscular system. A. With abdominaloblique muscles (in.o.) in situ. Th.l. First thoracic ganglion.Th. 8. Eighth thoracic ganglion, cede. Calcareous band be-tween ganglia. art. vent. Ventral artery, m.v. Ventralmuscles, in.d. Dorsal muscles, cut through and turned back.•per. Edge of pericardium, which has been removed. B. Afterremoval of oblique muscles to show six abdominal ganglia.

542 GEOFFREY SMITH.

(E) Muscular System.

The dorsal muscles which lie above the pericardial spaceform four dorso-lateral, segmented, longitudinal bands runningthe entire length of the body. The oblique muscles aresegmented bundles running obliquely downwards in eaclisegment; they are very much larger in the abdominalsegments than in the thoracic. Segmentally arranged ventralbands are *ilso present, which again are larger in the abdominalthan in the thoracic i-egion. In order to see them and thenerve-cord in the abdomen, the oblique muscles have to beremoved.

(*') Nervous System.

We will deal first with the nerve-cord behind the sub-cesophageal ganglion. There are in the thoracic regioneight distinct ganglionio thickenings, one for each freethoracic segment: similarly in the abdomen there are sisganglia (text-fig. 44).

There are five bands of calcareous concretions situatedbetween the first six thoracic ganglia (cede, text-fig. 44).This is the only plnce where lime is present in the whole ofthe body.

Each thoracic ganglion gives off three chief pairs of nerves:an anterior thick pair which pass forwards to the appendagesof the segment; a more posterior slender pair which appearto supply the ventral muscles of each segment; and a stillmore posterior pair which innervate the oblique muscles. Ifwe examine a single thoracic ganglion more carefully bymeans of sections, we may obtain a diagrammatic reconstruc-tion, such as is shown in text-fig. 45. The inter-gauglioniccommissures (com.) are seen to fuse above and below theganglionic area, but their fibres are continuous right throughthat area on the dorsal surface. The large nerves (N. 1) tothe appendiiges send their fibres ventral to the commissural


fibres, and in the ganglion they form thick bundles oftransverse commissures which anastomose in the thick fibrousregion in the ventral middle line. A ganglionic mass isapplied to the dorsal surface of this nerve (G. 1) and also to•the ventral (G. 2). The two nerves to the muscles are seenissuing more posteriorly from the ganglion ; their fibres alsolie ventral to the commissural longitudinal fibres. A

TEXT-FIG. 45.

4 5 .

Reconstruction of a thoracic ganglion of Anaspides tas-manise. com. Longitudinal commissures of cord. N. 1.Nerve to appendage. N. 2. Nerve to ventral muscles. N. 3.Nerve to oblique muscles. G. 1. Dorsal ganglionic mass.G. 2 and 3. Ventral ganglionic masses.

ganglionic mass (G. 3) is situated only on the ventral surfaceat the exit of these nerves.

The brain or supra-cesophageal ganglion gives issue tothree nerves, the optic {opt.), antennulary (Ant. 1), andantennary (Ant. 2) nerves.

The optic nerves spring from the dorsal surface of thebrain, being the only nerves with this position of origin.

544 GEOFFREY SMITH.

Where they enter the brain there is a dorsal ganglionic mass(text-fig. 46, g. A.). The dorsal ganglionic masses (g. C.)belonging to the autennary nerves are applied anteriorly to theventral root of the optic nerve, .and it is possible that this partof the ganglionic insiss (g. C.) really belongs to the optic nerve.If this were so, the optic nerve would be supplied with a dorsal

TEXT-FIG.

gro.

4 6 - 1Reconstruction of brain of Anaspides tasmanise. Ant. 1.

Antennulary nerve. Ant. 2. Antennary nerve. Opt. Opticnerve. g.A. Dorsal ganglionic mass applied to root of opticnerve. g.F. Its posterior continuation. g.B. Ganglion appliedto root of Ant. 1. g.C. Ganglion applied dorsally to root ofAnt. 2. g.D. Ganglion applied ventrally to root of Ant. 1.g.]$. Ganglion applied ventrally to root of Ant,. 2. Com. Peri-cesophageal comniissiires. sub. gn. Suboesopbageal ganglion.

and ventral ganglionic mass, as in the case of a regulartrunk nerve. Lying dorsally in the brain there is a bundleof transverse commissural fibres forming a regular opticchiasma.

The autetmulary nerves {Ant. 1) enter the brain ventrally

ON THE ANASPJDACEA, LIVING AND FOSSIL. 545

and are supplied with a dorsal (G. -B.) and ventral (G. D.)ganglionic mass on each side. Their fibres form thick trans-verse longitudinal bands in the ventral part of the brain, withcomplicated branches passing posteriorly. The anteuuarynerves (Ant. 2) enter the brain laterally and ventrally, andare supplied with a very large dorsal ganglionic mass (G. c),which passes anteriorly underneath the fibres of the opticnerves, and a smaller ventral mass (G.E.). The fibres whichenter the brain from these nerves form a very complicatedand massive system, occupying the whole of the posteriorpart of the brain.

The peri-cesophageal commissures (com.) are short, andventrally to the oesophagus they come together in the sub-cesophageal ganglion (sub. gn.), which gives off three nervesto the mandible and two pairs of masillee.

(G) Theoretical Considerations.

IE we pick out the more important points in the internalanatomy of Anaspides, in order to compare it with theother Malacostraca, we perceive that its internal anatomyis of a generalised type resembling in some respects thePeracarida and in others the Eucarida, especially the Deca-poda.

The chief Peracaridan features are the elongated, tubularheart and the filiform spermatozoa, while the possession of amaxillary gland is paralleled by certain Peracarida (Isopoda).The alimentary canal, on the other hand, in so far as it isnot peculiar, recalls that of the Decapoda. The nervoussystem is of an unconcenfcrated primitive character, with adiscrete ganglion in each thoracic and abdominal segment.

Comparing these conclusions with those derived from a con-sideration of the external characters, we may observe thatthey are in complete agreement. From the external charactersof the limbs and of the segmentation we judged that theAnaspiilacea stood midway between the Peracarida and theEucarida, but on a more generalised and primitive plane,

546 GEOFJFREY SMITH.

from which either of the two divisions might be derived. Wealso saw that the Anaspidacea, in their external characters,approached nearer to the Decapoda, than to the Buphausiidtype of the Bucarida, and this, again, appears to be the caseiu the internal anatomy. This would indicate that theEuphausiacea are a late offshoot from the Decapodan stock,and if this conclusion is accepted it is evident that the oldgroup of the Schizopoda, including the Mysidacea, Euphau-siacea and Anaspidacea is an unnatural assemblage and mustbe abandoned (see phylogenetic tree on p. 551).

4. BIONOMICS : (HABITS, REPRODUCTION AND DISTRIBUTION),

(A) Habi ta t and G-eneral Habits .

Anaspides tasmanite inhabits deep pools of rivers ortarns on the mountains of southern and western Tasmania;the water in which it lives is always absolutely clear andcold, and the animal clambers about upon the stones oramong the submerged mosses and liverworts at the bottomof the pools. It very rarely swims, though it occasionallydoes so in a lazy fashion, and it will occasionally rise to thesurface of the water and turn over on to its back in themanner of a Phyllopod. Its usual mode of progression iswalking or running1 in the attitude presented in PI. 11, &g. 1;when alarmed it darts forwards or sideways by powerfulstrokes of its abdomen and tail-fan. I never observed it tospring backwards, a movement of which it appears to beincapable.

The exopodites of the thoracic limbs are not used inlocomotion to any appreciable extent, their function beingto keep the water agitated round the gills and so assist inrespiration. Even when the animal itself is stationary theexopodites can be seen to be in a continual waving motion.

As remarked before, the body is always held perfectly flatand unflexed whether the animal is walking, swimming, or atrest.

ON THE ANASPJDACEA, LIVING AND FOSSIL. 547

They appear to be omnivorous, as they will feed upon thedead bodies of insect larvae or even upon one another, buttheir chief food is the algal slime covering the rocks amongwhich they live, and they also browse upon the submergedshoots of mosses and liverworts.

The rivers and tarns in which they live are singularly freefrom any enemies such as predaceous fish which might preyupon them, the only fish inhabiting these highland watersbeing the little "Mountain Trout," G a l a x i a s t r u t t a c e u s .The English Trout, which have multiplied so wonderfully iuthe Tasrnanian streams and lakes, have hardly penetrated tothe mountain fastnesses where A n a s p i d e s dwells.

The only parasite found infecting A n a s p i d e s is a peculiarspecies of neogamous gregarine which lives in the free statein the alimentary canal and forms large associated cysts inthe liver-tubes, often in very great numbers. This gregarinewill shortly be described in this journal by Mr. J . S.Huxley.

P a r a n a s p i d e s l a c u s t r i s is known only from the speci-mens collected by me in the great Lake of Tasmania at anelevation of 3700 ft., where it inhabits the littoral region,living among the weeds and stones at a small depth ratherafter the manner of a prawn.

Its markedly humped back and translucent green colourgive it very much the appearance of a small prawn, e. g.Hippolyte, and it follows more of a swimming habit thanA n a s p i d e s , but in other respects it resembles the latterclosely. I t doubtless falls an easy prey to the great quantitiesof large English brown trout which inhabit the lake, and it isprobably in danger of an early extinction.

K o o n u n g a c u r s o r has been found hitherto only in asmall runnel issuing from the Mullum Mullurn Creek,Ringwood, to the west of Melbourne, and it is the only memberof the living Anaspidacea which lives at a low level. Ingeneral appearance and habits it resembles A n a s p i d e smore closely than P a r a n a s p i d e s does, although it is morpho-logically very distinct. The specimens which Mr. Sayce

548 GEOFFREY SMITH.

kindly showed to me ran about with great activity, keepingthe body perfectly flat and unflexed as in Anaspides.

(B) Dis t r ibu t ion .

We have seen that the living Auaspidacea are all confinedto the temperate part of the Australian region, called byProfessor Spenser the Bassian Subregion, where they inhabitexclusively fresh water, usually at a high elevation, where atany rate the winters are exceedingly cold. The fossilAnaspidacea, on the other hand, are, as far as we know,confined to the marine deposits of the northei'n hemisphere,being found in the Permian and Carboniferous deposits ofEurope and North America. I have suggested elsewhere(12) that animals with this type of distribution, viz. in thenorth temperate hemisphere and in the Alpine regions oftemperate Australia, probably have reached their presentposition in the southern hemisphere through South Americaand the submerged Antarctic Continent, and not through thetropics of Asia and Australia. Although this is little morethan a tentative suggestion, it is, perhaps, justifiable topredict that living members of the Anaspidacea may still befound in the temperate fresh waters of South America orNew Zealand.

(c) Breeding and Reproduction.

The breeding of A. tasmanise appears to go on throughthe early summer months (December to April) as the poolson Mt. Wellington were continually being replenished withyoung of a very small size. Since there is no brood-pouch,it was of interest to establish what the female does with hereggs, especially as this is a character of great taxonomicimportnnce. By keeping the animals in captivity itwas observed that the male deposits two very largespermatophores in the spermatheca of the female. Thesespermatophores are large curved structures with a thin chi-tinous coat (text-fig. 43), and they project outside the sperma-

ON THE ANASPIDACEA, LIVING AND FOSSTL. 549

theca. The spermatozoa pass into the spermatheca and thespermatophores drop off. Fertilisation appears to take placein the oviducts, since in some sections of Koonunga shownto me by Mr. Sayce, spermatozoa could be seen lying in thebasal part of the oviducts. As to how the spermatozoa reachthe oviducal openings from the spermatheca there is somedoubt, but it seems probable that they are assisted in tinsmigration by the peculiar setose lobes on the internal facesof the last three pairs of thoracic appendages, which areonly present in the female.

The female deposits and hides the fertilised eggs, whichare of a purple colour and measure about 2 mm. in diameter,singly and not agglutinated together, under stones andamong the roots of water plants (PI. 12, fig. 3). This peculiarhabit of oviposition is only found elsewhere among theMalacostraca in certain forms of Euphausiidfe which mayhave pelagic eggs; in all other Malacostraca they are eithercarried in a brood-pouch (Phyllocarida and Peracarida), or elseglued on to the abdominal limbs (Eucarida), or carried in aspecial chamber formed by the maxillipedes {Hiplocarida).Among EntomostVaca the only forms which deposit their eggsare the Argulidte.

As to the development of the eggs or the presence of anylarval stages my observations are unfortunately very few. Iam, however, convinced that no complicated metamorphosisis passed through, as the pools under my observation werecontinually being replenished by minute Anaspides of4—5 mm. in length, which already possessed the completeadult structure. It is also quite impossible that pelagiclarvae of such types as the Nauplius and Zoasa could beassumed, as they would at once be swept away in themountain torrents down to the lowlands, where Anaspides,as a matter of fact, never occurs. It is possible, however,that the young are hatched out from the egg, not with thecomplete adult structure, for Mr. Sayce (10) found a minutespecimen of Koonunga in which the abdominal appendageswere incompletely developed.

550 GEOFFREY SMITH.

5. THE RELATION OP THE SOTCARIDA TO OTHER MALACOSTRACA.

The examination which lias been made in the foregoingpages of the chief characteristics of the Syncarida has con-vinced us that they are the most generalised group of theEumalacostraca. We have also seen that the structure of theliving representatives of the division is practically identicalwith that of fossils which date, at any rate, from the Carboni-ferons period, so that they are among the oldest Malacostraca,with the exception of the Nebaliacea, known. Theirgeneralised and primitive nature is not only shown by theirgreat antiquity and by the possession of such characters asthe freedom oE all the thoracic somites, and the presence ofeight joints in the thoracic limbs, but also by the fact oftheir combining in their own structure many of the distinc-tive characters of the two divergent groups of the modernEumalacostraca, viz. the Peracarida and theEucai'ida. Thuswe have seeu that they possess the auditory organ, thealimentary canal and its glands, the spermatheca and copu-latory organs of the Eucarida, but at the same time the struc-ture of the heart and of the spermatozoa, the absence of acarapace, the direct mode of development and the maxillarygland point to Peracaridan affinities.

The Syncarida, therefore, represent to a great extentthe ancestral form from which the Peracarida and the Eucaridahave diverged. It has also been pointed out that the Syn-carida are closer to the Decapoda than to the other order of theEucarida, viz. the Euphausiacea, so that the latter, so far frombeing the primitive Eucaridan type, must be considered as aspecialised offshoot from the main Decapodan stem, whichlias lost certain primitive and ancestral ch'aracters. In fact,the only primitive character retained by the Euphausiacea isthe biramous structure of the thoracic limbs, and this hasalso been retained by the lower members of the Decapoda,and so is not particularly significant.

In the following phylogenetic scheme we have given ex-pression to this idea.

ON THE AtfA.SPIDA.CKA., I/IVING AND FOSSIL. 551

Eucarida.

Peracarida.

Phyllocarida.

Entomostraca.

Hoplocarida.Decapoda. Euphausiacea.

Syncanda.

Phylogenetic Tree, showing chief lines of descent in the Crustacea.

VOX,. 53 , PAET 3 . NEW SERIES. 38

552 GEOFFREY SMITH.

The position of the Hoplocarida (Squillidas) as an earlyoffshoot from the Decapod an stem must remain at presentdoubtful, but many of their characters point to this conclusion.The spherical spermatozoa, the ramified hepatic casca, thecomplicated metamorphosis, the absence of a lacinia mobilis,the presence of an appendix interna on the pleopods, allsuggest that the Hoplocarida have travelled some way in theBucaridan direction since their derivation from the commonEutnalacostracan ancestor. That this Eumalacostracan an-cestor which diverged from the Leptostracan stock was notidentical with the Syncarida must probably be conceded.But we may perhaps conceive of it as a straight-bodiedambulatory Crustacean without a carapace and with all thethoracic segments free and uncoalesced, with a tail-fan andwith all its thoracic and abdominal appendages biramous.The thoracic limbs consisted of eight joints, three of thesejoints being distal to the " knee." It possessed stalked eyes,an antennal scale, and two flagella on the antennules, and itpossibly lacked an otocyst on the antennules. Internally theheart was elongated, the alimentary canal had, perhaps, onlyan anterior dorsal diverticulum, and the liver-tubes were fewand simple. The spermatheca were filiform and there waspresent both an antennary and maxillary gland. It probablyhad a brood-pouch as in the Phyllocarida.

From this type it was but a step to the Syncarida. Directlyfrom this ancestral type sprang the Peracarida, with theircharacteristic brood-pouch ; a certain amount, of fusion ofsegments either with the head or with one another took place,and certain of them developed, independently of the Eucarida,a carapace.

The Eucarida were probably derived from an ancestorwhich had travelled some way in the Syncaridan direction,that is to say, it had an otocyst on the antennules, it had lostthe brood-pouch, and it had developed certain other characters,such as the spermatheca and copulatory styles and a morecomplicated liver and alimentary canal. As it divergedfrom the primitive Syncarida it acquired a carapace, spherical

ON THE ANASPIDAOEA, LIVING AND FOSSIL. 5 5 3

in place of filiform spermatozoa and a short triangular heart;it specialised an antennary gland and it possessed a compli-cated metamorphosis. After its divergence it threw off theHoplocarida and the Euphausiacea.

If this phylogenetic scheme for the Eumalacostraca beaccepted it is clear that we can no longer look upon the oldorder " Schizopoda " as a natural assemblage standing at thebase of the Eumalacostracan stock. This classification restssolely on the biranious structure of the thoracic limbs, andignores all the other organs, whether external or internal, inwhich the various divisions differ so fundamentally.

We must remain in some doubt as to the presence orabsence of certain characters in the ancestral Eumalacos-fcracan which gave rise to the Syncarida, Eucarida and Pera-carida. With regard to the brood-pouch it may well be thatthis has been independently acquired by the Leptostraca andthe Peracarida, and that its absence in the Syncarida repre-sents a primitive condition. Again, the otocyst on theantennules may have been possessed by the ancestralEumalacostracan and lost by the Peracarida. In this connec-tion we may mention the striking observation of ProfessorFritsch, according to which an otocyst was present on theinner ramus of the uropods in the fossil Syncarida, Gaso-caris and Gampsonyx. If an otocyst was really presentin this position in these forms we can only suppose that theprimitive Eumalacostracan possessed it, and that it has beenretained by only a few Peracarida (Mysidacea), and entirelylost by the higher Syncarida and Peracarida and by the entiredivision of the Eucarida.

The reconstruction of the phylogeny of the Malacostraca,therefore, leads to the reflection that the primitive ancestorsof the specialised groups are not distinguished from theirmodern representatives so much by simplicity of structure,but rather by combining in themselves the heterogeneouselements which have been segregated out in the course ofevolution and separated into the different streams of descentthat have given rise to the modern groups. It was the habit

554 GEOFFREY SMITH.

of morphologists, and perhaps still is, to imagine that a primi-tive ancestral form must have been simpler and haveexhibited less complication of • structure than its modernrepresentatives. In pursuance of this preconceived notionthe simplest organised members of a phylum or smaller groupof animals was always hit upon as representing the ancestralform ; but too often it has been shown that this simplicity isthe result of secondary degeneration or simplification ofstructure. We have only to mention the Archianelida andthe Marsupials to indicate what has been the trend ofmorphological opinion on this subject.

I t is often complained, especially by naturalists immersedin the positive details of what may appear more modern andfruitful branches of inquiry, that speculative morphology,as a reputable department of biology, is dead, killed by thewild speculations of its devotees. But it may have escapedtheir attention that the tendency of speculative morphologyto-day is to display a more cautious temper than heretofore,and instead oE attempting to link phyla with phyla by goldenbridges of aerial speculation to undertake the more modesttask of tracing the lines of descent within a smaller rangeof organisms, the history of whose evolution has been accom-plished at any rate somewhere within the period of timerepresented by the stratified rocks.

If the family Anaspididee was already fully differentiatedin the Carboniferous period and the family Nebaliidas in theCambrian, it may well be conceded that to look for theancestor of the Crustacea or to prove that Peripatus reallylinks the Arthropoda and Annelida together are tasks whichthe cautious morphologist may well abjure. But that specu-lative morphology, when content to deal with the phylo-genetic history of fairly confined and homogeneous groupswhose fossil ancestry have been preserved for us through along period of time, is altogether idle we need not admit.And if a general consensus of opinion might at some futuretime be achieved, that the process of evolution in such groupshas not been effected by the gradual complication of an

ON THE ANASPIDACEA, LIVING'AND FOSSIL. 555

originally simple structure and by the addition of new organsto comparatively undifferentiated organisms, but rather bythe segregation and separation of characters originally com-bined in one ancestral form into the various streams ofdescent which have emerged from it—if this opinion mightat any time be adopted and sustained, it would influence ourattitude to the philosophy of evolution as profoundly as anyconceptions deduced from the experimental study of livingorgauisrns, without reference to the history which they havepassed through.

6. SYSTEMATIC PAKT.

The Anaspidacea, liviug and fossil, are placed in a sepa-rate division of the Eumalacostraca. Thus :

Division Syncarida (Packard 5 and 6, Calmau 9).

A carapace is absent. The thoracic somites are either alldistinct, or the anterior one may be fused to the head. Theeyes are pedunculated or sessile. An otocyst is present onthe basal joint of the antennules. The antennal protopoditeconsists of two segments. The mandible is without a laciniamobilis. The thoracic limbs consist typically of eight seg-ments, and the " knee-joint" is between the fifth and sixthsegment. There ai'e no oostegites. A spermatheca is presenton the last thoracic segment of the female. There is noappendix interna on the pleopods. The endopodites of thefirst two pleopods of the male are modified to form copulatorystyles. The branchiae form a double series of leaf-likeplates on all but the last or last two thoracic limbs. Theheart is elongated and tubular. The alimentary canal isfurnished with three dorsal diverticula, one at the junc-tion of the stomodgeum and mid-gut, one in the middle ofthe mid-gut, and one at the junction of the mid-gut andproctodseum. The hepatic cseca are numerous, elongatedand unbranched, and without glandular ridges. The

556 GKOFFBEY SMITH.

excretory organ is a maxillary gland. The spermatozoaare filiform, and are transferred to the female in horn-shapedspermatophores. There is no concentration of ganglia in thethoracic or abdominal region.

The eggs are deposited immediately after fertilisation bythe female and hidden singly under stones, etc.

There is no complicated metamorphosis, the young hatch-ing out with the essential structure of the adult.

Order Ariaspidacea (Caiman 9).

Diagnosis of the single Order is the same as that of theDivision.

Family I. Anaspididas (Thomson 7).

The thorax is composed of eight distinct somites. Theeyes are pedunculated. First antennas of male withoutsensory modification. There is a well-developed antennalscale. The mandible has a cutting blade, a setose lobe anda molar expansion. The palp of the first maxilla is a non-setose papilla. The first thoracic limb has gnathobasic lobes,a slender lamellar exopodite, and two branchise attached tothe coxopodite. The anterior thoracic limbs are clearly com-posed of eight segments. The last thoracic limb is uniramous.The pleopods are all biramous with a small flabellate endopo-dite, except the fifth pair, which are without the endopodite.

Genus 1. Anasp ides (Thomson 7).

The thoracic segments are all nearly equal in length;the abdominal segments are slightly longer. The body iscarried straight without any dorsal flexure. The antennalscale is shorter than the first two joints of the endopodite.The mandibular palp is without an exopoditic lobe.

The first thoracic limb has two gnathobasic lobes on thecoxopodite. The sixth abdominal segment and thetelsonareas long as the two preceding segments. Thetelson is shorterthan the uropods.

ON THK ANASP1DAOEA, LIVING AND FOSSTL. 557

A. tasmanitB1 (Thomson 7). Plate I, fig. 1.

The frontal margin of the head is produced into a conicalprojection, the sides and extremity of the cone being furnishedwith setae (text-fig. 47). The eye-stalks do not project greatlybeyond the la-teral margins of the head. The body is carriedflat and unflexed. On the head segment a median triangularpiece is marked out by shallow grooves. The head segmentis slightly longer than the first thoracic segment. The firstthoracic segment is distinguished by the presence of twolateral sulci. The succeeding segments of both thorax and

TEXT-FIG. 47.

47.Anaspides tasnianiae. Head and eyes.

abdomen are sub-equal in length, but the sixth abdominalsegment is considerably longer.

The first antennae have the three basal joints short andstout; the internal flagellum consists of about twentysegments.

The otocyst is kidney-shaped.The second antenna has a short scale, not reaching the

top of the second joint of the peduncle.The palp of the mandible is without an exopoditic lobe on

its basal joint. The terminal segment is nearly half as longas the last but one.

The first maxilla has a small palp and three setae near it.1 In the description of the species, those characters which have been

mentioned or described in the diagnoses of the genera or families or inthe anatomical part are only lightly touched upon, or in some cases notmentioned.

558 GEOFFREY SMITH.

The second maxilla has the exopoditic lobe greatly reducedand fringed with a few minute setse.

The first thoracic appendage has two gnathobasic lobesattached to coxopodite. The terminal segment, as in all livingAnaspidacea, carries three enlarged setse, of which the middleone is the largest.

TEXT-FIG. 48.

Anaspides tasmanise, <J. A. First abdominal limb,abdominal limb.

B. Second

There are two gills attached to the coxopodite of this limb,and of all the succeeding thoracic limbs except the last.

The seventh thoracic limb bears a small exopodite upon adistinct segment of the limb.

The fifth, sixth and seventh thoracic limbs of the femalebear a small setose lobe on the inner face of the coxopodite.


The first abdominal appendage of the male possesses aspatulate endopodite furnished with a row of setae proximallyon its ventral internal face, and a pad of recurved hooksdistally. The tip of the endopodite is simple.

The second abdominal appendage of the male possesses abiarticulate endopodite furnished on its internal facewith a pad of recurved hooks. This pad is situated con-siderably below the joint separating the two segments of theendopodite. The terminal segment of the endopodite is withoutsetse or spines. There is no median spine on the sternum of thissegment. There is a row of rather long setae on the posteriordorsal mai'gins of the fifth and sixth abdominal segments.The margin of these segments has a moniliform ornamenta-tion owing to the bases of the setae being raised.

The telson has the form shown in text-figs. 30 and 33A. Ithas a row of short setas confined to the posterior border.

The uropods have a short basal segment with a few lateralseta>. The exopodite has three enlarged lateral spines on itsupper external margin. The setaa fringing the uropods areuniform in size and structure.

The adult animal may attain two inches in length.The ground colour is straw yellow, but the skin contains a

great number of black chromatophores disposed in a regularpattern.

Occurrence.—In isolated pools on and near the topof Mount Wellington, Tasmania; in the pools of the upperreaches of the North West Bay River on Mount Wellington,above the Wellington Falls; in the tarns on Mount Field,on the Harz Mountains and on Mount Read, West Coast ofTasmania. At an elevation of 2000 to 4000 feet.

Genus 2. Paranaspides (Smith 12).

The first thoracic segment is longer than the two succeed-ing segments put together; the abdominal segments aremuch longer than the mid-thoracic segments. The body hasa distinct dorsal flexure, the first abdominal segment project-

560 GEOFFREY SMITH.

ing dorsally as a hump. The antennal scale is longer than thefirst two joints of the endopodite. The mandibular palp hasa distinct setose exopoditic lobe. The first thoracic limb,besides the two gnathobasic lobes on the coxopodite, has theinner face of the first segment of the endopodite expandedinto a setose lobe.

The sixth abdominal segment and the telson are togetherlonger than the three preceding segments.

The telson is shorter than the uropods.

P. l a c u s t r i s (Smith 12). (Plats 11, fig. 2 ; text-fig. 1.)

The frontal margin of the head is produced into a conicalprojection, the cone being tipped with a bunch of setae.The eye-stalks project considerably beyond the lateralmargins of the head (text-fig. 4). The body is carried witha marked dorsal flexure. A triangular piece is not obviouslymarked out on the head segment, and the head segment is equalin length to the first thoracic. Lateral sulci are present on thefirst thoracic segment. The first thoracic segment is equalin length to the three succeeding segments. The abdominalsegments are all longer than the mid-thoracic segments. Thefirst antennae have the basal segments elongated and ratherslender; the internal flagellum consists of about twentysegments. The otocyst is oval.

The second antenna has a large scale, far exceeding inlength the two joints of the peduncle (text-fig. 7).

The palp of the mandible may be four-jointed; it possessesa distinct exopoditic lobe, tipped with setae, and the terminalsegment is equal in length to the last but one (text-fig. 10).

The palp of the first maxilla is larger than in A n a s p i d e stasmaniaa, and there are uo setas near it (text-fig. 13).

The second maxilla has a fairly well-developed exopoditiclobe fringed with long setae (text-fig. 16).

The first thoracic appendage has two gnathobasic lobes,and the third segment of the limb is expanded inwards into alobate biting blade (text-fig. 19)

ON THE ANASP1DACEA, LIVING AND 1'OSSIL. 561

The thoi-acic limbs are built upon a similar plan to those ofAnaspides, but they are more slender and the setse longer.The seventh thoracic limb has a small exopodite, but the seg-ment bearing it is incompletely marked off from the othersegments by a groove.

The first abdominal appendage of the male has a spatulateendopodite with an internal pad of recurved hooks, a

TEXT-FIG. 49.

4-9.

Paranaspides lacustris, $. A. First abdominal limb. B. Secondabdominal limb. en. Endopodite. ex. Exopodite.

proximal row of setae, and the tip is furnished with a numberof short spines (text-fig. 49A).

The endopodite of the second abdominal appendage of themale has an external pad of recurved hooks near the top ofthe first segment, and the terminal segment carries a numberof stout spines (text-fig. 49B). There is no median spine outhe sternum of this segment.

There is a row of short spine-like setas on the posteriordorsal margin of the fourth, fifth and sixth segments.

562 C4E0PFBKY SMITH.

The telson has an elongated form, with slightly concavesides. The posterior margin is produced into a number ofspines of very unequal length (text-fig. 33B and 34).

The uropods have a short basal segment; the exopoditehas six or seven stout spines on its external border. Theother setae fringing the uropods are unifoi'm.

The adult animal may attain an inch in length.The colour is transparent green, but with a few minute

black chromatophores scattered about, chiefly on the lateralportions of the segments.

Occurrence.—In the littoral zone of the Great Lake ofTasmania, among weeds and stones. Elevation 3700 ft.

Grenus 3. Praeanaspides (Woodward 13). (Text-fig. 3.)

The first thoracic segment is much shorter than the others ;the succeeding thoracic segments are sub-equal in size andthe abdominal segments are on the whole equal to the thoracic.There is no dorsal flexure. The antennal scale is apparentlyjust equal in length to the first two joints of the endopodite.The segment of the thoracic limbs immediately proximal tothe knee-joint is expanded especially in the anterior limbs.The sixth abdominal segment and the telson are together alittle longer than the two preceding segments. The telson isequal in length to tlie uropods.

P. pr as cursor (Woodward 13).

The characters of this very important fossil are exhibitedin text-fig. 3.

It is at once seen from these figures how close is theresemblance in all the essential characters between it andthe living Anaspides. Similar transverse striations on thesegments have been observed by Woodward in Palseocaris.With regard to the segmentation of the thoracic limbs itwould appear that there were the typical three segmentsdistal to the knee-joint and four segments proximal to it.

ON THE ANASP1DAOKA, LIVLNG AND FOSSIL. 563

the separate segment bearing the exopodite having probablyfused with the third segment, as in Koonunga and theposterior limbs of An as pi des. All the thoracic limbs exceptthe last were apparently furnished with exopodites.

The abdominal appendages were probably very much as inAnaspides,the exopodites being furnished with long, slenderhairs.

For the determination of specific characters probably thehind segments, telson and uropods are most important.

The dorsal posterior borders of the third, fourth, fifth andsixth abdominal segments have a moniliform ornamentationshowing where a row of spines was situated. The telson iselongated and bluntly rounded at the end. Its lateralborders are setose almost to the base, and at the posteriorlatei'al angles a few longer setas were present.

The uropods do not project further than the end of thetelson. The outer ramus has a row of about seven shortsetas on its outer border and two elongated spines at the lineof segmentation near the tips of this ramus.

Length.—Largest specimen 57 mm. in length.Occurrence. — In clay-ironstone nodules of the Coal-

measui'es near Ilkeston, Derbyshire.

Family II. Koonungidaa (Sayce 10 and 11).

The thorax is composed of seven distinct somites, theanterior one being fused with the head. The eyes are sessile.First antennas of male with sensory modification. There isno antennal scale. The mandible has a cutting blade and asetose lobe, but no molar expansion. The palp of the firstmaxilla is reduced, but distinct, and carries setas. The firstthoracic limb has a slender exopodite, but is without gnatho-basic lobes. The thoracic limbs are composed of only sevensegments. The last two thoracic limbs are uniramous. Thepleopods are all uniramous, except the first two pairs in themale, which are modi6ed as copulatory organs and retain theirendopodites.

564 GEOFFREY SMITH.

Genus ] . Koonunga (Sayce 10 and 11).

The cephalic segment is about equal in length to the twofollowing segments: there is a short transverse sulcus oneach side at about the middle distance, posteriorly to whichthe margins ai*e produced downwards and inwards. Thethoracic and abdominal segments are all subequal in size,the sixth abdominal segment not being longer than thesegments preceding it. There is no dorsal flexure. Thetelson is very short and does not equal in length the uropods.The basal joint of the uropods is nearly as long as therami.

K. cursor (Sayce 10 and 11). (Text-fig. 2.)

The frontal margin of the head is truncated and notproduced into a projecting cone (text-fig. 50). There are no

TEXT-FIG. 50.

50.Koonunga cursor. Anterior region of head with sessile eyes.

setae upon it. The minute sessile eyes are situated at the anglesof the frontal margin. The body is carried flat and unflexed.There is no median triangular piece marked out on the head.The head segment is as long as the two succeeding segmentsand there is a marked transverse sulcus on each side. Thesucceeding segments, both thoracic and abdominal, are allsubequal. The first antennas has the first segment distinctlystouter and longer than the two succeeding segments.The inner flagellum is six-jointed, and in the male a peculiarsensory appendage is present, furnished with spiral hairs(text-fig. 51). The otocyst is circular. The second antennais without a scale (text-fig 8).

The mandible is without a molar lobe, and the terminal

ON THE ANASPIDAOiSA, LIVING AND FOSSIL. 565

segment of the palp is much less than half as long as the lastsegment but one (text-fig. 11).

The first maxilla has a large palp, tipped with three setae.The lower biting lobe has only three setee. The exopodite ofthe second maxilla is not marked by the presence of anysetae. The first thoracic appendage is much stouter than the

TEXT-FIG. 51.

5 1 .

Koonnnga cursor,appendage (op.)- B.

. A. First antenna, showing sensorypiral sensory hairs froni appendage.

succeeding limbs; it is without gnathobasic lobes orexpansions.

The succeeding limbs are very similar to those of Anas-pi des save that the upper series of gills are much smallerrelatively, and the last two limbs are without exopodites.

All the thoracic limbs consist of seven, instead of eight,segments.

The exopodites consist of far fewer segments than inAnaspididae.

The first abdominal appendage of the male possesses abroad endopodite which ends in a cui*ved hook and a broad

566 GEOFFREY SMITH.

blade; it is, in fact, definitely bifid. The pad of recurvedhooks is borne on an internal projection (text-fig. 52A).

The endopodite of the second abdominal appendage of themale is two-jointed. The terminal joint is excavated at itstip and is clothed distally with very fine hairs. The pad ofrecurved hooks is situated internally on the proximal half of

TEXT-FIG. 52.

Koonunga cursor, (J. A. First abdominal limb. B. Secondabdominal limb. en. Endopodite. ex. Exopodite. sp. Spineon sternum.

the first segment. There is a large chitinous piece, with aposteriorly-directed spine, on the sternum belonging to thissegment (text-fig. 52B). The posterior dorsal margins of thehind abdominal segments are without setas or moniliformornamentation. There is a single large spine on each side onthe posterior dorsal margin of the sixth abdominal segment.

The telson is short and obtusely conical, and its posterior


and lateral margins are produced into a number of long stoutspines of equal size. Between the spines are a number ofshort fine bristles (text-fig. 32).

The uropods have large basal segments furnished dorsallywith three stout spines. The rami are short and truncated.The exopodite has abont six long spines on its externalborder, and a number of long compound setae terminally,which become shorter as we appi-oach the internal border.The setae clothing the endopodite are also much longer termi-nally than on the lateral borders.

The adult animal attains to about & inch.The colour is dark, marbled brown; ground colour yellow,

with numerous black chi'omatophores.Occurrence.—From freshwater reedy pools beside a tiny

runnel joining the Mullum-Mullum Creek, Ringwood, nearMelbourne.

Family III. Gampsonyehidas (Packard 5).

In this family may be included provisionally the threegenera Gampsonyx, Palasoearis, and Gasocaris.

The thorax is composed of eight distinct segments ofwhich the first is smaller than the rest. The succeedingsegments, both thoracic and abdominal, are subequal in size,except the sixth which is somewhat elongated. An antennalscale was apparently present. All the thoracic limbs appearto have been biramous.

The body was carried straight without any flexure. Theeyes were pedunculated.

Genus 1. Gampsonyx (Jordan and v. Meyer 1) ( = Gamp-sonychus = Uronectes) .

The flagella of the first antennas were apparently equalin length. The first thoracic limb was a powerful raptorialorgan armed with curved claws. The endopodites of thehinder thoracic limbs were slender and much elongated.

VOL. 53 , PART 3 . NEW SERIES. 39

568 GEOFFKIOY SMITH.

The telson was longer than the sixth abdominal segment.The abdominal appendages were apparently stout and flabel-late.

TEXT-FIG. 53.

5 3 .

G-ampsonyx fimbriatus. "Reconstruction after Jordan andv. Meyer.

Gr. f imbriatus (Jordan and v. Meyer 1). (Text-fig. 53.)

The structure of the antennae and eyes is shown in text-fig. 54. The scale of the second antenna appears to have

TEXT-FIG. 54.

5 4 .

Gampsonyx filnbriatus. Head; after Fritsch. Ant. 1. Firstantenna. Ant. 2. Second antenna with scale.

slightly exceeded in length the first segment of the peduncle.The posterior thoracic and the abdominal segments havewell-marked pleura,

ON THE ANASPIDAOEA, MVING AND FOSSIL. 569

The first thoracic limb is a powerful raptorial organ of thestructure shown in text-fig-. 53. The next two limbs wererather small and fairly stout, but the succeeding five thoracicendopodites are very long and slender. Fritsch denies thatany exopodites were present, but it may well be contendedthat the fine strife figured by Jordan and v. Meyer springingfrom the bases of the thoracic limbs represent the hair-likesetas present on slender exopodites.

The abdominal appendages, according to Fritsch, are flabel-TEXT-FIG. 55.

otooyst.

-•f-telson,

55.Gampsonyx fimbriatus. Telson and iiropod. After Fritsch.

lute in structure, but Jordan and v. Meyer made them out tpbe setose.

Fritsch gives a very finely detailed drawing of the telsonand uropods (text-fig. 55).

The telson was an elongated oval, with seise on its posteriorlateral margins, and two long and two short sette at the hindend. The outer ramus of the uropod was about equal inlength to the telson, and had a row of six short setse and oneenlarged spine on its outer border. Fritsch figures a con-spicuous sphere on the inner ramus which he interprets asan otocyst.

570 GEOFFREY SMITH.

Occurrence.—In the Carboniferous of Saarbriick, RhenishPrussia, and of Lebach, Bohemia.

Genus 2. Palasocaris (Meek and Worthen 2 and 3).

The flagella oE the first antennas were unequal in size.The. first thoracic limb was apparently not raptorial. Theendopodites of the thoracic limbs were stout and short,exopodites elongated. The pleura of the abdominal segmentsprojected backwards to end in a definite acute angle.The telson was shorter than the sixth abdominal segment.The abdominal appendages were slender, not flabellate.

•P. typus (Meek and Worthen 2 and 3). (Text-figs. 56and 57.)

The chief interest of this fossil is, perhaps, to be found in

TEXT-FIG. 56.

5 6 .

Palaeocaris typus. Reconstruction after Packard. Lateral view.

the fact that Packai'd definitely demonstrates the presence ofexopodites on the thoracic limbs. The eyes are unknown,but since in so many respects this fossil is close to Gramp-sonyx there can bo little doubt that they were pedunculated.The thoracic limbs, if we can trust Packard's restoration, wereall similar and biramous.

The telson was broad and short and apparently clothedwith a uniform border of setae. The uropods projectedbeyond the end of the telson and were also apparentlyclothed with uniform set£e.

In certain respects Meek and Worthen's figure (text-fig.57) iu dorsal view of P. typus is more interesting than

ON THK ANASP1DACEA, LIVING AND FOSSIL. 571

Packard's obviously diagrammatic restoration. The specimenis fiat and resembles to an extraordinary degree the dorsalview of Praeanaspides given by Woodward (text-fig. 3 B).

TEXT-FIG. 57.

Palaeocaris typus. Reconstruction after Meek and Worthen.Dorsal view.

TEXT-PIG. 58.

58

Palfflocaris typus. Telson and uropod. After Packard.

We see also in this figure the limbs spread out laterally in theexact position assumed by Auaspides when walking.

Length.—-78 inch.39§

572 GEOFFREY SMITH.

Occurrence,—In clay-ironstone concretions in lower Coal-measures of Mazon Creek, Morris, Illinois.

Woodward describes another species, P. Burne t t i i ,measuring 30 mm. from the middle coal-measures of Irwell,Lancashire. In this species he desci'ibes the transversestrife on the segments which he afterwards observed in Pra3-anaspides.

Genus 3. Gasocaris (Fritsch 17).

The flagella of the first antennas are unequal in size. Allthe thoracic limbs are similar, rather short and stoutly built(without exopodifces according to Fritsch). The telsou israther shorter than the uropods. There is a row of setaa onthe posterior dorsal margin of all the thoracic and abdominalsegments. The body is bi'oad anteriorly, tapering consider-ably toward the hinder end. The abdominal limbs are stout,but annularly segmented.

G. krejcii (Fritsch 17). (Text-figs. 59, 60, and 61.)

Fritsch establishes beyond doubt the pedunculated natureof the eyes and the structure of the first and second antennas.In his restoration, however, he makes the two flagella of thefirst antennae equal in size, which is contradicted by hisfigure of an actual specimen in ventral view (text-fig. 59). Thescale of the second antenna seems to have reached the top of thepeduncle on which the flagellum is inserted. With regard tothe segmentation of the body his reconstruction is impossibleto reconcile with the ventral view of an actual specimen. Inthe reconstruction he makes, besides a narrow head-segment,only six thoracic segments. In the ventral view, repi'O-duced here, it is easy to make out eight free thoracic segmentswith limbs, and a broad segment in front to represent thehead, and this structure would bring the specimen into linewith the other Anaspidacea. The thoracic limbs, as recon-structed by Fritsch, are all similar, being rather short and

ON THE ANASPIOA015A, LIVING AND FOSSIL. 573

stout and apparently without exopoilites, though with regardto this latter point we may well keep our judgment suspended,

TEXT-FIG. 59.

G-asocaris krejcii. Ventral view. Adapted from lYitsch. 1-8.Thoracic segments. Abd. Abdominal appendage.

574 GEOFFREY SMITJJ.

owing to the difficulty of making out the exopodites in eventhe best preserved fossil Syncarida. The abdominalappendages

TEXT-FIG. 60.

6 0 .

Gasocariskrejeii. Compound eye and second antenna. AfterFritsch.

appear to have consisted of astoutouterramus,segmentedin theannular way characteristic of such a form as Anaspides. Icannot observe in Fritsch's figures of actual specimens any

TEXT-FIG. 61.

Gasocaris krejcii. Telson and nropod. After Fritsch.

trace of endopodites, though Fritsch in his reconstructionfigures the appendages as consisting of endopodite andexopodite of equal length. It appeal's more probable thatthe endopodite was the reduced flabellate structure which weknow in the living Anaspides.

ON THE ANASPLDAOEA, LIVING AND FOSSIL. 575

The telson is ovate or conical, with a border of shortuniform setse round the posterior three quarters of themargin. The uropods projected slightly beyond the telsonand were fringed with uniform setae. Pritsch figures anotocyst at the base of the inner ram us.

Length.—About 13 mm.Occurrence.—In Coal-measures of. Nyran near Pilsen; in

older strata than those in which Gamps onyx occurs.

Genera of doubtful position.

Genus Acanthotelson (Meek and Worthen 2).

Two species are described, A. st impsoni and A. in-

TEXT-FIG. 62.

62 .Acanthotelson stimpsoni. Reconstruction after Packard.

Lateral view.

sequalis. Packard gives a restoration of the genus. Hefigures besides a head segment, seven thoracic and sixabdominal segments and a telson. The eyes are unknown.The first antennas had a three-jointed peduncle and a singleflagellum, according to the restoration, though it appears thatit may have had two. The second antenna was apparentlywithout scale. The seven thoracic limbs were long, fairlystout, and without exopodites. The first two limbs wei-eraptorial in structure and furnished with long spines on the

576 GEOFFREY &M1TH.

penultimate joint. The abdominal appendages had1 a longbasal segment and a terminal flabellate segment clothed withlong setaa.

The telson was an elongated pointed spine fringed withlong setae. The uropods, which consisted of external andinternal rami, Avere also long and pointed, and they werefringed with long setae.

Occurrence.—Coal measures of Illinois.

Genus Nectotelson (Brocchi 19).

N. rochei from the Permian of Autun, France. Only verybadly preserved specimens known.

LlTEKATORE.

1. Jordan and V. Meyer.—" Ueber d. Steinkolilen-forruatioii vonSaarbriicken,' ' Pakeontographica,' vol. iv, 1856.

2. Meek and Wortlien.—'Proc. Acad. Nat. Sci. Philadelphia,' 1865,p. 41.

3. ' Report G-eol. Survey Illinois,' 1868.4. Packard.—' Arnei'ican Naturalist,' vol. xix, 1885, p. 790.5. ' Mem. Nat. Acad. Sci. Washington,' vol. iii (2), 1886.6. ' Proc. Boston Soc. Nat. Hist.,' vol. xxiv, 1889.7. Thomson.—'Linn. Soc. Trans.,' London, 2nd ser., vol. vi, part 3,

1894, p. 285.8. Caiman.—' Roy. Soc. Edin. Ti-ans.,' vol. xxxviii, part 4, 1896,

p. 787.9. ' Ann. Mag. Nat. Hist.,' 7th ser., vol. xiii, 1904, p. 144.

10. Sayce.—'Victorian; Naturalist,' vol. xxiv, p. 117, 1907.11. ' Linn. Soc. Trans.,! London, vol. xi, Part I, 1908.12. Smith.—' Proc. Roy. Soc. London,' B, vol. lxxx, 1908, p. 465.13. Woodward.—' Geological Magazine,' decade v, vol. v, No. 9, 1908,

p. 385.14. Hansen.—' Ann. Mag. Nat. Hist.' (6), xii, 1893, p. 417i15. Geldert.—' La Cellule,' t. xxv, fasc. 1, 1906, p. 7.16. Woodward.—'Geological Magazine,'.1881, p. 533.

ON THK ANAS I'l DACE A, LIVING ANJ) FOSSIT.. 577

17. Fritsch.—' Fauna der Gaskohle und der Kalksteine der Perm-' formation Bohmens,' Bd. iv (1901), Heft, 3.

18. Caiman-.—' Zoologischer Anzeiger,' Bd. xxv, No. 661, 1902, p. 65.19. Brocchi.—' Bull, de la Soc. Geol. de France,' 1879, 3rd ser., vol.

viii, p. 1.

EXPLANATION OP PLATES 11 AND 12,

Illustrating Mr. Geoffrey Smith's memoir on " The Anas-piclacea, Living and Fossil."

PLATE 11.FIG. 1.—Anaspides tasmanioe. Dorsal view of a very large speci-

men. Natural size, in normal position for running.FIG. 2.—Paranaspides lacustr is . Lateral view. Natural size, in

position for running or swimming.FIG. 3.—Shoot of the liver-wort, Yungermannia, with two eggs of

Anaspides tasmanire. X 5.

PLATE 12.FIG. 1.—Transverse section through otocyst of Anaspides, x 50.

ec, ectoderm; n, antennulary nerve; in, muscle ; r, setose ridge.FIG. 2.—Portion of blood-forming organ of Anaspides, showing

mitoses and detached corpuscles.FIG. 3.—Transverse section through the cardiac portion of stomach of

Anaspides. C and J>, lateral ridges; E, dorsal median ridge; H,ventral ridge.

FIG. 4.—Transverse section through the alimentary canal of Anas-pides, where the first diverticulum (div. 1) opens into the mid-gut.The section is not quite transverse, so that the opening is only seen onthe right side. B, lateral chitinous ridge of pyloric division of thestomach extending into mid-gut; b. in., basement membrane of mid-gut.Div. 1, first diverticulum, showing crowded nuclei and mitoses.

FIG. 5.—Transverse section through anterior portion of mid-gut,showing basement membrane (b. m.), and special gland-cells (g.).

FIG. 6.—Section through a special gland of mid-gut, showing flattenednuclei of gland-cells.

578 GEOFFREY SMITH.

FIG. 7.—Section through wall of posterior part of mid-gut, showingabsorptive layer (m.), basement membrane (b. TO.), and submucosa (s.»».).

FIG. 8.—Transverse section through a liver tube of Anaspides,unfed condition, g, special gland-cells.

FIG. 9.—Transverse section through a liver tube, full-fed conditionwith fat vacuoles in cells.

FIG. 10.—Transverse section through external duct of maxillarygland.

FIG. 11.—Section through a more internal part of gland.FIG. 12.—Section through a more internal secretory part of gland.FIG. 13.—Section through a portion of end-sac.FIG. 14.—Longitudinal section through a piece of ovaiy of Anas-

pides, showing two large ova, small ova lying in the lobes on the right,and the trophic cells (jtr.) on the left.

FIG. 15.—Transverse section through testis of young Anaspides,showing trophic cells (tr.), primary (sp. 1), and secondary spermacytes(sp. 2), and spermatozoa (spt.).

FIG. 16.—Upper portion of vas deferens, with albuminous part ofspermatophore in lumen, and trophic cells (tr.) surrounding it.

FIG. 17.—Section through the supporting tissue of glandular appear-ance which surrounds the ducts of the spermatheca, and is also presentin the labrum.

$&$'M&.*

i?..•i* •• \ ' V

on the anaspidacea, living and fossil

Documents