[ 213 ] VII .
Lapworthura : a Typical Brittlestar of the Silurian with Suggestions for a New Classification of the Ophiuroidea .
By Igerna B. J. Sollas , B.Sc. , Lecturer on Zoology in Newnham , Cambridge , and W. J. Sollas , F.R.S. , Professor of Geology and in the University of Oxford .
( Received June 2 , \#151 ; Read December 7 , 1911 .
) [ Plates 9 and 10 .
] Contents .
Description of Lapworthura miltoni ... .
Palaeozoic Ophiuroidea other than Lapworthura Transition to modern forms ... ... ... ... .
Literature referred to ... ... . .
Note on Construction of Models . . .
The anatomy , of the Palaeozoic Ophiuroidea has been far from exhaustively studied , though they possess features of great interest .
The nature of the available material and the impossibility of obtaining conclusive evidence by the older methods of study are sufficient reasons for our imperfect knowledge of these primitive organisms .
In the following paper an account is given of the structure of the Silurian Ophiurid Lapworthura miltoni ( Salter ) , based upon a study of the fine series of this fossil in the Oxford Museum and of models made by the method described by one of us in 1901 ( 23 ) , 1903 ( 24 ) ; other Palaeozoic Ophiuroidea are described and compared with Lapworthura , and the bearing of the facts upon the interpretation of the structure of modern types is considered . .
It will be seen that in all essential points Lapworthura conforms to the type of structure discovered by Jaekel in Eophiura , Palaeura , and Bohemura ( 11 ) , though it has reached a more advanced stage of evolution .
And indeed there is evidence to show that all the Silurian and Devonian Ophiurids possess a fundamentally similar structure .
It is not possible to give exact detailed descriptions of all the named genera , but it is not difficult , in the light of the knowledge gained from the study of Lapworthura , to recognise in them a natural group differing in important respects from existing forms .
It follows from this that a revision of the classification of Ophiuroidea is necessary .
Carboniferous Brittlestars have been described by Meek and Worthen ( 18 ) .
The description of their genus Onychaster is not very complete but is sufficient to ( 288 .
) 21.2.12 .
PAGE .
214 .
222 .
227 .
230 .
231 show that in the Carboniferous era an advanced stage of evolution had been reached by this form ; and whether , as Meek and Worthen 's description would lead us to suppose , this form is extremely aberrant , or whether it approximates in structure to modern types , as Schondorf maintains , in either case its existence in the Carboniferous epoch precludes the division of the Ophiuroidea into Palmophiuroidea and Neophiuroidea , but we suggest the formation of a group Protophiuroidea for the Silurian and Devonian forms characterised by the completeness of the ambulacral series ( the first pair of ambulacral ossicles ( cq ) being retained ) , by the meeting of the ambulacral series of adjacent arms on the abactinal side of the jaws , and by the absence of upper and under arm-plates .
All the modern genera form a group Euophiuroidea , characterised : ( l ) by the specialised buccal armature consisting of oral angle pieces formed by the complete fusion of a2and adL , al having been lost ; ( 2 ) by the presence of under arm-plates , with the single exception of Ophioteresis .
With regard to the systematic position of Onychaster we refrain from offering an opinion till our own investigations are completed .
Eucladia has already been assigned to a special group , the Ophiocistia ( 25 ) , which is here retained but raised to the rank of a class equal in value to the Ophiuroidea , and possibly more nearly allied to them than to any other class of the Echinodermata .
Description of Lapworthura miltoni [ Salter)* This species was first described by Salter in 1857 , under the name of Protaster miltonii and re-described by him in 1861 ( 20 ) .
In his second paper Salter came to the conclusion that the madreporite is abactinal ( dorsal ) , and this led him to consider that Protaster miltoni should be transferred from the Ophiuroidea to the Asteroidea , its resemblance to the former group being merely imitative .
He gave a figure correct as far as it went ( Plate 18 ) representing the " real shape of the ambulacral bones .
" In 1886 Sturtz restored this species to the Ophiuroidea ( 26 ) .
In 1896 Gregory ( 9 ) created the genus Lapworthura for miltoni .
He gave a new figure of the arm skeleton , in which the anterior end of the ambulacral ossicles is incorrectly represented as the posterior .
He followed Salter in assigning the madreporite to the abactinal surface , and endeavoured to show that this is the primitive position of the madreporite in the class .
Gregory 's statements influenced the views of Sturtz .
In 1890 Sturtz had suggested that if the madreporite originally occurred on the same aspect of the body in both Starfish and Brittlestars , it was actinal in both and that it is in the Starfish and not in the Brittlestar that * Schondorf 's work came to our notice only after the completion of the present paper .
He has recognised , on the same evidence as our own , the true position of the madreporite .
With his main contention that the ambulacral ossicles were hollow half-cylinders and that consequently Lapworthura and all other Protophiuroidea must be removed from the Ophiuroidea , we are unable to agree ; on this point the evidence afforded by our sections leaves no room for doubt .
LAPWORTHURA : A TYPICAL BRITTLESTAR OF THE SILURIAN AGE .
a change has taken place .
He called attention to a prediction made by Sladen ( 22 ) in 1880 , who , commenting on Agassiz 's statement ( 1 ) that in very young starfish the madreporite is at first ventral ( actinal ) and only subsequently becomes dorsal ( abactinal ) , suggested that ancient fossil starfish would be found in which the madreporite retained its ventral ( actinal ) position in the fully adult state .
Sturtz added that this prophetic remark had not had long to wait for justification ; several fossil starfish were already known at that date in which the madreporite lies on the actinal surface of the adult .
In 1893 ( 28 ) Sturtz 's views on this subject were unchanged .
In 1900 he drew attention to the difficulty which is introduced into the phylogeny of Stelleroidea by Gregory 's description of Lapworthura and of Eucladia , and by his statements that Bury had found a dorsal madreporite in young Ophiuroidea ; for , he argued , while in Starfish we have modern forms with a dorsal ( abactinal ) madreporite and primitive forms with ventral ( actinal ) madreporite , and according to Agassiz , a ventral ( actinal ) madreporite in young starfish , in the Ophiuroidea on the other hand the converse appears to be the case .
Again , if in the original stock of Ophiuroidea the madreporite lay on the abactinal , and in the original stock of the Asteroidea on the actinal side , this fact would not lead to the view that the Ophiuroidea and Asteroidea were derived from the same stock .
Here it should be pointed out that Gregory was mistaken in his quotation of Bury , who was speaking not of the young Ophiuroid but of the larva .
In the same year ( 1900 ) Gregory in his article in Lantkester 's ' Treatise on Zoology , ' while still apparently holding the view that the madreporite is " dorsal " ( abactinal ) in some Ophiuroidea ( see p. 238 ) , did not turn his attention to the phylogenetic importance of this character .
The argument from embryology is considerably altered by modern work and is by no means decisive .
Bury 's figures ( 4 ) of newly metamorphosed Asterid larvae ( Plate 6 , figs. 21 , 22 , 24 , Bipinnaria asterigera)show the water pore as marginal and actinal .
MacBride(15 ) describes the madreporite in young Amphiura as near the edge of the disc , and the stone canal as nearly horizontal .
But he points out that " the arrangement in Amphiura might be obtained from that in Asterina by rotating the stone canal and accompanying structures outwards and downwards through an angle of 180 ' .
" He finds indications that this is what has actually occurred in the phylogeny , in the position of madreporite and stone canal of young Amphiura and in the undulating course of the genital rachis , which " points to the conclusion that the aboral parts of the inter-radii have greatly developed and have grown on to the oral surface ... .
and so the stone canal has been swung round ... .
In Asterina there is a trace of this process , " and MacBride is inclined to look on the condition of Asterina as the primitive one from which both Asterid and Ophiurid arrangements have been derived .
But the evidence which he gives of the primitive nature of Asterina tends to prove only that it is primitive among Asterids .
It is clear then that position of the madreporite in Lapworthura is a matter of importance and needs careful determination .
It has already been shown by one of us ( 25 ) that the madreporite of Eucladia is actinal .
An examination of the fine collection of Lapworthura miltoni in the Oxford Museum has convinced us that here , too , the madreporite occupies a position on the actinal surface .
When well preserved there is no mistaking the nature of this structure ; it is situated inter-radially near one of the arms , within the region of the disc , is almost hemispherical in form , measures about 0*78 mm. in diameter , and is covered with a characteristic labyrinth of fine ridges .
As to its recognition there can be no difficulty , but what is not so easy , on first studying the subject , is to determine which is the abactinal and which is the actinal aspect of the organism .
The distribution of the integumentary remains of the disc is a criterion which at once suggests itself , but one tedious in application and more useful for confirmation than discovery .
A much readier guidance is furnished by the arrangement of the plates of the buccal armature .
Slight inspection suffices to arrange a collection of Lapworthura in two series , one in which the jaws originate within the arms , continuing the direction of the vertebral ossicles , and the other in which the oral framework prolongs the direction of the lateral plates ; an examination of any modern Ophiurid will suffice to show that a similar difference in the relations of the oral armature is presented , according as the dissected disc is viewed from the abactinal or actinal aspect : and the actinal surface is that on which the lateral buccal shields are exposed as a prolongation of the series of lateral arm-plates .
This fact teaches that the second of the series into which we have supposed our collection of Lapworthura to be divided consists of specimens presenting the actinal aspect to view ; now it is on these and these only that the madreporite is found to occur .
The madreporite is consequently actinal in position , a conclusion which the examination of the integumentary remains fully confirms .
It may further be added that the position of the madreporite is as nearly as possible that which it still occupies in most existing Brittlestars ; in almost all cases it lies side by side with the second or third adambulacral plate , sometimes in contact with the base of the first adambulacral , which enters into the oral framework .
There are no oral plates in Lapworthura , but in the modern Brittlestars these scales cover precisely that part of the inter-radial area which corresponds to the angle in which the madreporite of the fossil is placed , while , to make the similitude more exact , the perforation of the oral plate by the stone canal in recent forms is always eccentric , and situated on the same side of the inter-radius as in Lapworthura .
Had Lapworthura possessed oral plates , one of them would have inevitably surrounded the madreporite in the fashion characteristic of living Ophiurids .
The Skeleton of the Arms.\#151 ; That the ambulacral ossicles are paired and opposite is well known , and it has been asserted by Sturtz , whose statement is confirmed by Dr. Gregory , that they are fused in the median line .
As this is a matter of some importance , it may be considered first .
The notion of a union of the ambulacral ossicles is not suggested by an examination of either the actinal or abactinal surfaces LAPWORTHURA : A TYPICAL BRITTLESTAR OF THE SILURIAN AGE .
217 of the arm ; a deep groove separates the adjacent ossicles on each face , and was doubtless occupied by a radial water vascular canal on the ventral side .
The existence of these grooves does not prove conclusively that the ossicles were separate ; the fact , however , that in some specimens a displacement of the ossicles has occurred , so that the members of a pair cease to be opposite , and become alternate in position is suggestive , and proves that , if any connection originally existed , it must have been of the slightest kind .
A few of the Leintwardine specimens are preserved , not as casts , but as solid calcareous residues , and these provided the material from which the models were made .
In the case of transverse sections , the relations of the parts were fairly well shown , though their form had been greatly distorted by earth pressure .
No evidence was obtained of any intimate union between the ambulacra !
ossicles .
A welcome confirmation of the statement that upper and under arm-plates are absent in these Ophiurids was met with : the lateral arm-plates were seen bearing spines , and articulated with the ambulacral ossicles , but no trace of any other structures was visible in the surrounding matrix , although a great number of sections were observed .
Horizontal sections gave similar results , the actinal and abactinal grooves are continued between the ossicles from one face to the other ; ambiguous cases were sometimes encountered , which might suggest a union , but these are more probably explained as due to the obliteration of the original fissure by pressure .
The form of the ossicles is very different as they are viewed from the actinal or abactinal side ; abactinally , they somewhat resemble the phalanges of a Plesiosaur , and are in contact at each end ; they are slightly rounded from side to side in the middle , thickened at each extremity , and produced into curved lateral processes at the outer corners .
On the actinal side they look somewhat like two capital letters " L " placed back to back ; the base of the " L " is broad and thick , and affords an articular surface for the lateral plates ; the limb is also wide , and immediately distal to the base is depressed , but further on becomes thicker ; this thickened part is excavated by a deep groove , which extends obliquely outwards and backwards ( in a proximal direction ) ; the edges of this are sometimes obviously thickened ; occasionally this groove is closed at the inner end , but more usually opens freely into the cavity of the arm .
There can be little doubt that it received the branch of the radial water canal that supplied the tentacles .
At its distal termination the leg of the " L " is depressed .
As the vertebral ossicles are traced from the oral end of the arm outwards , an interesting change may be remarked in their form ; at first they are somewhat broader than long , but subsequently these relations are reversed .
Thus a pair of ossicles situated about one-quarter of the length of an arm from its origin measured 1*3 mm. in length and 1*5 mm. in breadth , but at the broken extremity of ' the arm the breadth had diminished to 0'3 to 0'5 mm. , while the length had been but slightly reduced , viz. , to 1*1 mm. ; in another instance , an ossicle viewed from the abactinal VOL. CCII.\#151 ; B. side , at a point just outside the disc , had a length of 1*1 mm. , while its breadth taken across the middle was 1 '7to 1'8 mm. ; near the distal end of the arm an ossicle measured 1*7 mm. in length and 1*1 mm. in breadth .
As the ossicles become narrower they also become simpler , and finally are reduced to mere cylindrical rods .
The modern growth of the science of embryology has opened up a wide field of interest in the comparison of ancient forms of life with early stages in the development of the individual , but the warning becomes ever more emphatic not to trust to embryology as a guide until its indications are checked by palaeontology or comparative morphology .
A striking parallel is afforded by the forms through which the vertebral ossicles of a recent Ophiurid pass in the course of their development , and those which succeed one another in the chronological series of their fossil predecessors .
Thus , in existing species , one of the earliest forms of the vertebral ossicle is that of a simple rod , such as characterises a whole group of Palaeozoic Ophiurids , and which is represented in the youngest ossicles in the arm of Lapworthura miltoni .
In the next stage the ossicles of a modern species broaden out , and become united at their extremities , leaving a narrow bi-con vex opening in the centre , which is strikingly similar in form to the interspace between the bodies of a pair of ossicles in Lapworthura , though these , as we have seen , are in all likelihood not yet united together .
In the next stage of development in modern Ophiurids the ossicle becomes thicker and grows up about the branch of a water canal which supplies the tube feet , but does not at first wholly enclose it , or , in other words , a furrow is formed for the passage of the canal , resembling the furrow on the ventral side of the ossicles in Lapworthura .
The raised lips of this furrow in the fossil form seem to be a step towards roofing it over .
The lateral plates take a large share in the skeleton of the arms , they are scarcely inferior in size to the ambulacral ossicles ; each consists of a rod-like stem and a petal-like lamina ; the lamina is curved in conformation with the exterior of the arm , it extends forwards and backwards the whole length of a segment , and articulates or overlaps in each direction with the similar plates next to it in the series ; on its surface it bears a row of small granules with which the heads of the spines , four or five in number , are in contact .
The end of the stem abuts on the proximal thickening of the ventral surface of a vertebral ossicle .
The absence of upper and under arm-plates has already been referred to , and it only remains to add that in modern Ophiurids these plates make a later appearance than the ambulacral and adambulacral ossicles .
The difference in time of appearance is not great , being measured in the history of the young Brittlestar by days or even hours , though in the evolution of the race it was a matter of unknown aeons .
Buccal Armature.\#151 ; The buccal armature has hitherto received no attention ; a study of its structure shows that it is of fundamental importance , as it proves to represent , in all probability , a stage in the evolution of the buccal armature of modern Ophiurids .
The large mouth , which is guarded by exceptionally long teeth , LAPWORTHURA : A TYPICAL BRITTLESTAR OF THE SILURIAN AGE .
is bounded by five pairs of rods , and by the jaws with which these rods articulate at their inter-radial meeting place .
The variable size of the oral angles in most specimens shows that the creature could undergo a considerable amount of distortion without breaking , and that the rods must therefore have articulated by loose sutures in the middle of the radius ; indeed , the suture is occasionally visible as a groove in guttapercha casts taken from the better preserved specimens .
The two halves of a jaw may also be separated .
Inter-radially each rod overlaps the jaw abactinally , and thus comes into proximity with the corresponding region of one rod of the neighbouring oral angle .
Each rod is marked out into regions by constrictions .
According to our interpretation , these rods consist of au , and an abactinal crest of a3y aiy and possibly of o5 , these elements being fused end to end .
It is possible that \#171 ; 5 is merely overlapped by the distal end of the crest of a4 ( see text-figs .
2 and 4 d ) .
Text-fig .
4 d shows how the vertebral ossicles are related to the swellings of the rods .
It follows from the foregoing statements that the ambulacral pieces of the first four pairs ( or the halves of the first four vertebrae ) diverge more or less from each other on each side of the radial line ; ax and a2 have also shifted upwards so that they lie upon , instead of beside , the corresponding adambulacral pieces , and a3 has also shifted upwards , but to a less extent .
The above conclusions are based upon a study of the fossils and of the models constructed from serial sections of the fossils .
On the ventral aspect of the fossil the inverted L-shaped ridges of the ventral surface of the ambulacral ossicles can be traced up to the oral angle in parallel series ; at the angle , divergence of the members of the pairs of ossicles is seen to occur to a variable extent .
The oral angle may be anything between 30 ' and 110 ' in different angles of the same specimen .
When the fossil has a regular form these angles are between 30 ' and 40 ' ; the divergence starts ( that is the apex of the oral angle lies ) between the members of the pair a4 ; those of a3 are distinctly divergent , and in front of a3 no vertebrae can in general be distinguished on this aspect of the fossil .
In some specimens , however , a slender rod is visible bounding each of the tentacle spaces tx and t2 on their radial aspect ; these rods we believe to be the radial margins of and they continue the curve started by the divergence of a3 and ax .
In front of a3 and continuing the line of the adambulacrals and articulating or fused with ad3 is ad2 , which is simpler in form than the other adambulacrals and appears to carry no spines ; at its other extremity it articulates with the jaws adx .
Possibly its form is a survival of that which it originally had when the ambulacral pieces first moved upwards and outwards to their present position ; it resembles the adambulacral pieces of the simpler and earlier Eophiura .
That the rods bounding the oral angle on the abactinal surface are formed of fused ambulacral ossicles is suggested by the following facts derived from a study of the fossils and models : The main part of the rod is composed of an abactinal crest which runs the whole length of a3 and and of the radial border of ab .
This crest is 2 e 2 continued forwards over the tentacles to end upon the jaws .
It is natural to expect the forward continuation to have an origin similar to that of the main part of the rod , and it is naturally in the neighbourhood of the tentacle spaces that we should look for the two missing ambulacral pairs ( vertebrae ) , and , further , the inner margin of the rods continues , as already stated , the curve formed by the inner or radial surface of az and ct4 .
That two vertebrae are represented in this remaining part of the rod is suggested by the fact that the space intervening between the anterior end of \#171 ; 3 , and the oral extremity of the rod , is just twice the length of or a4 ; there are two tentacle spaces and adamhulacral pieces underlying the piece in question , so that all the evidence there is goes to prove that the number of ambulacral and adambulacral pieces is equal in this ancient Brittlestar .
Evidence from actual sutures is lacking , and also owing to the superposition of cq and a2 upon the adambulacrals , the useful , actinal L-shaped ridge no longer acts as a guide .
Fig , 1.\#151 ; One oral angle of Lapworthura miltoni as represented by a model composed of vertical sections .
Portions of the model are separated so as to show exposed to view at intervals the form of the sections composing it .
LAPWORTHURA : A TYPICAL BRITTLESTAR OF THE SILURIAN AGE .
Nevertheless , in the complete absence of any evidence to the contrary , and considering the beautiful regularity of structure revealed by the models , there remains no great doubt that the full complement of vertebrae are present .
The first of the two models is built up of 265 vertical sections of the disc .
The sections are at right angles to the long axis of one of the arms and consequently cut across the jaws at various angles .
Of these sections some of the most instructive pass through the jaw almost at right angles to its length ( fig. 1 ) .
Here we see ax lying above the first tentacle space and articulating with , but the sutures here , as elsewhere , are unfortunately obliterated .
On the actinal surface , is seen articulating with adx , and sending off an upward process by which it is connected with a2 ; this process represents the stalk of the adambulacral , but is now upwardly instead of inwardly directed , owing to the altered relative position of a2 .
On the other side of the gap left by the oral angle lies the other half of , and its ambulacral articulating with it , the tentacle space for tz now lying on the outer side of the ambulacral piece .
Behind this level we have parts of au a2 , az , a4 in succession , the sections running obliquely across the oral angle .
Passing forwards along the border of the oral angle , az with its adz is succeeded by a piece containing the tentacle space of t2 , of which the roof is a2 and the floor ad2 ; a2 is connected with ad2 by its outer border and also by the upward process already seen which forms the partition between the two tentacle spaces and Passing the tentacle space tx , we come to the jaw with ax well seen lying upon it , and separated from it by a suture both in this oral angle and the next one , reached after passing round the jaw .
To trace the sections of this model further would involve too much repetition , and would only afford confirmatory evidence .
The second model is composed of 37 sections parallel to the surface of the disc passing through one of the oral angles ( figs. 1 and 2 , Plate 10 ) .
The angle of the mouth represented in this model measures 90 ' , and has been forced wider open than it would be in its natural position of rest .
Owing to this fact , adz is more concealed than is the case in some of the actual arms of any well-preserved specimen , adx being the first adambulacral ossicle which is clearly visible ; adz is , however , to be found articulating with its neighbours on the under surface of the model , and underlying az ( fig. 2 , Plate 10 ) .
A certain amount of shifting of this kind was , perhaps , possible during life .
The perfect regularity of the ad^mbulacral series when there is no Fig. 2.\#151 ; Superposed tracings of three horizontal sections of Lapworthura miltoni .
displacement of this kind is shown in text-fig .
2 , which is obtained by superposition of tracings from photographs of Sections 10 , 16 , and 31 .
The type of buccal armature here described in Lapworthura miltoni is in its main features precisely similar to that described by Jaekel as occurring in Eophiura and Palaeophiura , both from the Lower Silurian ( D ) of Bohemia .
The chief differences are the fusion , end to end , of the halves of the abactinal portions of those ambulacral ossicles which diverge from their fellows in order to form the buccal fissures , the smaller number of these ossicles , and the greater vertical extent attained by the skeletal parts bounding the oral angle , particularly by and A further difference is the greater simplicity of as compared with its fellows .
In all these respects , and in the opposite arrangement of its ambulacral ossicles , Lapworthura shows a more advanced stage of development than the Lower Silurian forms , Eophiura and Palaeophiura .
In discussing the transition from the Eophiuran jaw to the jaws of modern forms , MacBride says : \lt ; l We may suppose that the gaping ' angles ' of these old forms have , so to speak , healed up , except at their innermost portions .
" In Lapworthura we seem to be actually watching this process of healing up .
MacBride also observes that in Eophiura : " The actual 'jaw ' ... .
seems , as in modern forms , to consist only of the first adambulacral fused to the second ambulacral .
" In Jaekel 's paper there is no statement to this effect , nor can we agree that his figures even suggest the loss of the first ambulacral ossicles .
Palaeozoic Ophiuroidea other than Lapworthura .
It has already been stated that in its main features the structure of Lapworthura is typical of Silurian and Devonian Ophiuroidea , and that for this reason we propose to institute a new group , Protophiuroidea , to include all Silurian and Devonian Ophiuroidea with the exception of Eucladia and its ally , Euthemon .
These genera were placed by one of us ( 2 , 1899 ) in an order of Ophiuroidea , the Ophiocistia .
Bather ( 3 , 1907 ) has recently made the very probable suggestion that the structures regarded as muscular arms are , in fact , gigantic scale-covered tube-feet .
Thus the Ophiocistia , as far as we know them , agree with Ophiuroidea only in the possession of an actinal madreporite and of a buccal armature ; these characters alone are not sufficient to justify the inclusion of these fossils in the Ophiuroidea .
The Ophiocistia are , therefore , regarded as a class of Echinodermata , possibly allied to the Ophiuroidea .
The diagnostic characters of the group Protophiuroidea are , as already stated , ( l ) the completeness of the ambulacral series ; ( 2 ) the meeting of the ambulacral series of adjacent arms on the abactinal surface of the jaws ( ) ; and ( 3 ) the absence of upper and under arm-plates .
In respect of these three characters the structure of the Protophiuroidea is Asteroid in nature .
The actinal position of the madreporite is also shared by some of the early Asteroidea , so that the sharp LAPWORTHURA : A TYPICAL BRITTLESTAR OF THE SILURIAN AGE .
223 limitation of the arms from the disc is the chief distinguishing character between the early Starfish and Brittlestars .
It is natural to ask what purpose was served by this outward and upward shift of the proximal ambulacral ossicles .
Greater firmness of the buccal armature was obviously one advantage gained thereby .
It is also clear that as long as the adambulacral ossicles formed a continuous series up to the jaws , and the vertebral ossicles were of no greater vertical extent than the adambulacral pieces , the latter would form an obstacle in the course of muscles running ( like the external inter-radial muscles of the oral skeleton of modern forms ) between the vertebrae of adjacent arms in the inter-radii behind the jaws .
Within the Protophiuroidea a gradual evolution can be traced .
In the most primitive genera the ambulacral ossicles are neither completely alternate nor .completely opposite .
As examples of this most primitive structure Eophiura , Pakeura , and Bohemura may be quoted .
They are further characterised by the large number of ambulacral ossicles which bound the oral angles and by the absence of fusion with one another in a longitudinal series of these ossicles .
For the forms with alternating ambulacral ossicles Gregory ( 9 ) proposed the order Lysophiurse .
We were at first inclined to retain it as a division of the Protophiuroidea , referring the remaining members of the sub-order to a division Synophiurse characterised by opposite ambulacral ossicles .
But the state of knowledge at the present time hardly permits of this .
For while it is certainly clear from Jaekel 's work ( 11 , see , for instance , his figs. 1 and 6 ) that in some of the simplest genera the ambulacral ossicles were not opposite , it is also certain that free opposite ambulacral ossicles are easily displaced either during life or after death , and it is difficult in some cases to be sure whether alternation is natural or due to displacement .
We therefore think it wisest for the time merely to recognise the fact that there has been a progression within the Protophiuroidea from forms with free , partly alternating , ambulacral ossicles , on the one hand to forms with completely alternate ambulacral ossicles , and on the other to those in which these ossicles are opposite , at first free , afterwards having slight connections , and increasing in vertical extent , and that in both groups the buccal armature has advanced in complexity .
Through the kindness of Prof. T. McKenny Hughes we have had an opportunity of examining Forbes ' Protaster sedgwicki , the type specimens of which are in the Sedgwick Museum .
This species , together with a new species found by Dr. Groom in the Middle Ordovician beds of Shropshire , for which we propose the name Protaster groomi , and the little P. leptosoma which we rename Rhodostoma , form an interesting group for comparison with Lapworthura , Protaster sedgwicki ( Forbes ) was described by Forbes in 1849(8 ) , and was found by him in Kirkby Moor Flags ( Upper Ludlow ) of Docker Park , Benson Knot .
The type specimens include two casts , one of the actinal and one of the abactinal surface , of an Ophiurid which agrees with Forbes ' description , and which were figured by him , and , on a separate piece of rock , casts of the abactinal surface of a very different species having opposite ambulacral ossicles .
It is rather badly preserved , but in those features which can be made out it resembles Rhodostoma Gregory re-described P. sedgwicki in 1896 ( 9 ) , and correctly described the L-shaped ridges , crossed by a groove , seen on the actinal aspect of the arms .
It is curious that he does not mention the fact that the specimens labelled as type specimens belong to two different species , and that he gives a description of the ossicles at the distal end of the arms .
The distal ends of the arms are broken off in the specimens having alternating ossicles ; the specimens having opposite ambulacral ossicles agree neither with Forbes ' nor Gregory 's description .
Protaster sedgwicki , as represented by the specimens figured by Forbes , is an Ophiurid with alternating ambulacral ossicles .
The relative proportions of the disc and buccal armature contrast with those of Lapworthura ; while the diameter of the disc in both species is 1*8 mm. , that of the circle formed by the apices of the oral angles is 0*6 mm. in P. sedgwicki and IT mm. in Lapworthura ; on the other hand , the form of the ambulacral ossicles is closely similar in the two species .
Gregory has considered a Y-shaped appearance of some of the adambulacral ossicles near the mouth as one of the diagnostic characters of the genus , but we believe that this form , when it occurs , is simply due to the way in which the specimen is fractured .
No spines are visible ; Forbes , when he speaks of spines , seems to allude to the adambulacral plates .
The scales of the disc are large and hexagonal .
The first two adambulacral ossicles and the first three ambulacral ossicles are involved in the formation of the buccal armature .
The first two adambulacral ossicles are fused with one another .
The main bar bounding the oral angle appears to be formed by a great development of a2 , a piece which is probably ax lies parallel to this bar , from which it is marked off by a groove , and along its proximal margin , while a crest of az lies alongside its distal margin , the greater part of the ossicle a3 being overlapped by the large a2 .
The genus Protaster thus needs a fresh diagnosis , and from the evidence at our disposal we offer the following :\#151 ; Protaster : Protophiuroidea with alternating ambulacral ossicles .
The second ambulacral ossicles attain a large size and form the main ambulacral constituent of the buccal armature ; they are elongated in a direction parallel to the oral angle ( text-fig .
3 ) .
The first ambulacral ossicles are distinguishable as separate bars .
Protaster groomi agrees in all essential respects with Protaster sedgwicki , the same ossicles contributing to the buccal armature and ax being distinguishable as a separate bar .
The greater complexity of form of the buccal armature , however , justifies the establishment of a new species for this form .
As a full description of this species will be published by Dr. Groom , we confine ourselves to the statement that ax appears to have grown inwards towards the middle line from LAPWORTHURA : A TYPICAL BRITTLESTAR OF THE SILURIAN AGE .
its position on the abactinal side of the jaw ( ac^ ) , and a large crest , which may be either wholly formed of a2 , or of a2 and a3 , exists as in Rhodostoma leptosoma , from the same horizon as Lapworthura ( .
, from the Lower Ludlow beds of Leintwardine ) , is represented by numerous specimens in the Oxford collection .
One or two specimens presenting the actinal aspect are preserved in the original material .
The whole calcareous skeleton is exceedingly slight and delicate .
As in the two species of Protaster above described , the buccal armature is formed by abactinal crests of ambulacral ossicles which have elongated in a direction parallel to the borders of the oral angles .
The first ambulacral ossicle , ax , bounds the oral angle and is free from its successors ; a2 is the largest of the ambulacral ossicles contributing to the buccal armature , and in worn specimens it may appear to be the only one taking an important share in its formation .
The laminae of the adambulacrals are separated by a wide interval from the bodies of the vertebrae , the connecting pedicle , composed partly of their own stalk and partly of processes of the vertebrae , being long .
It is no doubt owing to this fact that ad2 , instead of articulating at its distal end with , which could not afford it much support , has acquired a connection with , thus giving extra firmness to the oral skeleton .
The typical condition of the ambulacral ossicles is opposite ; nevertheless , alternation may occur throughout the length of an arm and in all the arms of one specimen .
The facts observed certainly suggest that this character is a variable one in the species .
To obtain an idea of the form of the ambulacral ossicles , evidence from many specimens must be pieced together .
They were apparently more or less rectangular , the members of a pair being in contact only at their anterior and posterior ends ; their inner ( radial ) borders were thickened , appearing as a raised triangular area on the dorsal surface of the ossicle .
The adambulacral pieces consist of a stalk lying in the plane of the actinal surface of the ossicle , and a spine-bearing bar , which curved back over the dorsal surface , so that the spines lay over the thinner regions of the ambulacral ossicles .
On the actinal surface , the ambulacral ossicles show the usual L-shaped ridges .
Protaster leptosoma ( Forbes ) was placed in the genus Sturtzura , created for it and P. brisingoidesby Gregory in 1898 .
Chapman ( 6 ) restored P. brisingoides , the type species , to Protaster in 1907 .
Consequently , as Bather ( 2 ) points out , the name Sturtzura becomes a synonym of Protaster .
We propose to place the species in the new genus Rhodostoma , which is defined as follows :\#151 ; Rhodostoma : Protophiuroidea VOL. CCII.\#151 ; B. 2 F Fig. 3.\#151 ; Buccal armature of Protaster sedgwickithe figure is obtained by superposing camera lucida drawings of the dorsal and ventral aspects .
Dorsal aspect dotted outline .
with either opposite or alternating ambnlacral ossicles .
Three ambulacral ossicles are modified as elements of the buccal armature , and are elongated in a direction parallel to the oral angles , thus giving to this part of the skeleton a petaloid appearance .
It is to this that the name of the genus refers .
It will be seen that , according to the above interpretation , the buccal armature of the three species in question , while agreeing in the main features with that of Lapworthura , has , nevertheless , been developed in a somewhat different manner , but the difference is perhaps more apparent than real .
In Lapworthura we have supposed that the bars bounding the oral angles have been formed by the fusion , end to end , of abactinal crests of the first four or five ambulacral ossicles , so that , if we could restore the original lines of fusion , they would be more or less transverse to the bar , instead of running along its length ( fig. 4 ) .
This difference , however , will be seen to depend simply on the extent to which the component ossicles had elongated in a Fig. 4.\#151 ; Series of schemes of buccal armatures of Palaeozoic Ophiurids .
A. Schematic figure of an early form based upon Jaekel 's drawing of Eophiura .
B. Hypothetical form in which the proximal ambulacral ossicles elongate , sending outgrowths towards the radial line .
C. Bhoclostoma D. Lapworthura miltoni .
E. Eoluidiadecheni .
direction oblique to their long axis previous to fusion with one another ( when this fusion has occurred ) .
With the exception of Miospondylus , the most advanced genera are Furcaster , Eoluidia , Eospondylus .
For these three genera Sturtz gives figures of the buccal armature which are almost interchangeable , and , indeed , the structures figured are remarkably similar .
We believe that fig. 4 E gives the correct interpretation of the buccal armature in Eoluidia decheni .
As in modern Ophiurids , a2 forms the oral angle , but it is not yet indistinguishably fused with adl , and al is still present .
On the ventral surface of the arms , the familiar L-shaped ridges of the ambulacral ossicles are visible bounding an open groove .
No trace of upper or under arm-plates is present .
Miospondylus rhenanus , of which a single specimen is known from the Lower Devonian of Bundenbach , has been described as having under-arm plates , a character LAPWORTHURA : A TYPICAL BRITTLESTAR OF THE SILURIAN AGE .
227 which was also attributed to Eoluidia and Eospondylus .
The plates so named are visible in one arm only of the single specimen on its abactinal surface beyond the edge of the disc .
This arm appears to have undergone rotation , so that , for the greater part of its length , its actino-lateral aspect is uppermost .
Near the disc , where the rotation undergone is less , the so-called under arm-plates are clearly paired .
What their real nature is remains problematic .
Curious jointed structures , looking like jointed and flattened spines , are present on the lateral aspect of the arm and are named by Sturtz side shields .
Transition to Modern Forms : Nature of the Plates .
It will be readily recognised that if the interpretation given in fig. 4 is correct , the step is not a long one from such a form as Eoluidia to a modern Brittlestar .
The acquisition of upper and under arm-plates , the increase in vertical extent of the vertebrae , the complete fusion of ad\ and a2 , comprise the necessary changes .
That these changes had already taken place in the Lias is as shown by the sections and models which have been constructed by Miss Wright ( now Mrs. Hallam , formerly of Somerville College , Oxford ) , of the Liassic Ophiurid , As is well known , Ludwig 's scheme of the homologies of the parts entering into the formation of the buccal armature of recent Ophiurids has met with general acceptance , with the exception of the homology of the peristomial plates , which , in the absence of evidence , is either denied or doubtfully accepted .
Whatever may be the final decision upon this debatable point , we cannot fail to admire the insight which led Ludwig , as early as 1878 , to construct his well-known schemes of the Asteroid and Ophiurid mouth parts ; by these , both types were reduced to a common plan , in which the ambulacral series of the neighbouring arms meet upon the abactinal surface of the jaws .
The essence of his views has received a remarkable confirmation in the discovery of the structure of the Early Palaeozoic Ophiuroidea , though he was almost certainly mistaken when he pressed the peristomial plates into service to fulfil the scheme .
Lyman , in 1882(10 ) , objected to Ludwig 's view that the peristomial plates were the homologues of the first pair of ambulacral ossicles , because , he said , " the plate is in no way connected with either of the mouth tentacles , and because it may be composed of one , two , or three pieces , or be altogether wanting .
" In his valuable and now well-known paper , Dr. sir Strassen brought forward , in 1901 , further evidence against the homology .
He considers that the morphological evidence is , of itself , almost sufficiently adverse ; for when , as in the majority of species , peristomial plates are present , their number is not two , as Ludwig 's hypothesis would lead us to expect ; but , further , when absent , their absence occurs not in the more specialised , but in the more primitive genera , in many species , for instance , of Ophiomusium , Ophioglypha , and Ophioplinthus .
Palaeontology completes this argument , for in Ophioderma egertoni from the Lias of 2 p 2 Lyme R-egis , which agrees closely in skeletal structure with the more primitive of modern Zygophiurae , peristomial plates are wanting .
Sir Strassen 's embryological evidence taken alone was almost sufficient to prove that the peristomial plates are not ambulacral ossicles , for he shows that in Amphiura squamata ossicles , which are undoubtedly au are still recognisable at a time when the peristomial plates are laid down .
These are the " Loffelplattchen , " which lie in the oral angles and afterwards disappear .
At the same time , he describes certain other small ossicles , which he maintains are a pair of ambulacral ossicles , and a ventral plate .
From this he concludes that not " 2 , but 3 segments take part in forming the oral region , of which the most anterior lacks lateral plates .
" Jaekel seems to follow zcjr Strassen in this view , but MacBride points out that it requires that a number of tentacles have disappeared ( 11 , p. 502 ) .
It also requires that in the first segment the lateral shields have disappeared , and in the second segment the under arm-plate .
Further , though it cannot be objected that these ossicles are " in no way connected with the mouth tentacles , " as Lyman said of the peristomial plates , still they are associated with the tentacles of sir Strassen 's second segment , while they are themselves , according to him , the ambulacral ossicles of his first segment .
In studying the development of Amphiura , one of us found that these little ossicles are not constant in number ; as a rule , there are two of them , each lying abactinally from the root of the first mouth tentacle .
Either one of each pair , or both , have a downwardly directed elongation between the mouth tentacles reaching to the under arm-plate of the second segment .
We have never observed the Fig. 5 ( 1 ) .
Fig. 5 ( 2 ) .
Fig. 5.\#151 ; ( 1 ) Horizontal section of an oral angle of Amphiura squamata , showing a pair of small ossicles ( 2 ) Vertical section of an oral angle of Amphiura squamata .
One of the ossicles sends down a process reaching the under arm-plate % number 3 , but in one and the same specimen we have seen a variation between 1 and 2 in the different oral angles .
In Ophiactis halli , in the corresponding position , there is a single T-shaped ossicle .
LAPWORTHURA : A TYPICAL BRITTLESTAR OF THE SILURIAN AGE .
229 Now , as is well known , a vertebra has a definite characteristic method of development , arising as a pair of simultaneously formed spicules between the terminal and the first-formed vertebra and abactinal to the main nerve trunks .
The ossicles in the oral angles develop late , after the formation of a great part of the skeleton ; if they were vertebral vestiges they should appear before all the other vertebrae .
They are not more developed in their younger than in their final stage .
If we assume them to be reduced vertebrae we are not at liberty to attribute their retardation to their vestigial nature , for Ludwig 's " ax , " although they completely disappear at last , yet originate in their proper order .
It is true that Ludwig accounts for the smaller size of oq at a time when and a2 are the only ambulacral ossicles present , by stating that a2 was the first to develop ; but we have not found this to be the case .
Bather ax owes its smaller size to its slower growth .
Finally , the ossicles in the oral angles are ventral to the nerve trunks .
These ossicles of Amphiura squamata are , as n Strassen points out , represented in other forms by similarly placed but more substantial ossicles .
We have observed them in microscopic sections of lepida as hoop-like ossicles embracing the tentacles , articulating below with the under arm-plate of the second segment , above with adx , and we find that a precisely similar pair of ossicles is present in the Liassic species Ophioderma egertoni .
There can he little doubt that they are " etwas neu erworbenes , " as sir Strassen says , of the peristomial plates , and possibly they are the equivalent of an under arm-plate , the fact that they are double being due to their position .
It should be pointed out that Ludwig figures an under arm-plate in the vertical wall of the oral angle ( 9 , Plate 3 , fig. 1 ) .
Sometimes each of these little hoop-like plates is divided by a transverse suture .
But too much value must not be attached to sutures , for they may be secondarily acquired .
This is well shown in the case of the oral angle pieces .
These pieces are formed , as is well known , by the fusion at an early stage of a2 and adv no line of demarcation being visible and processes of the two growing ossicles becoming inextricably entangled ; later a transverse suture appears dividing this compound piece into two parts which are sometimes spoken of as a2 and adv a proceeding which is clearly inaccurate .
This suture is doubtless adaptive and is already present in Ophioderma egertoni .
I 1 is satisfactory to find that there is no conflict between the evidence , such as it is , from palaeontology , embryology , and morphology as to the history of the buccal armature and arm-skeleton of Ophiuroidea .
Literature Referred to .
1 .
Agassiz .
' Proc. Amer .
Acad. Arts and Sci. , ' 1863 , p. 109 .
2 .
Bather , F. A. ' Geol .
Mag. ' 1907 , dec. 5 , vol. 4 , pp. 313 and 523 .
3 .
Idem .
' A Guide to the Fossil Invertebrates in the British Museum , ' 1907 .
4 .
Bury .
' Quart .
J. Micro .
Sci. , ' 1889 , vol. 29 , n.s. , pp. 409-450 .
5 .
Idem .
Ibid. , 1895 , vol. 38 , pp. 45-134 , Plates 3-9 .
6 .
Chapman , F. ' Melbourne P.R. Soc. , Victoria , ' 1907 , ser. 2 , vol. 19 , p. 21 .
7 .
Idem .
'Geol .
Mag. , ' 1907 , dec. 5 , vol. 4 , p. 479 .
8 .
Forbes .
' Figures and Descriptions Illustrative of British Organic Remains , ' December 1 , 1849 .
9 .
Gregory .
' Proc. Zool. Soc. , ' 1896 , p. 1033 .
10 .
Idem .
' A Treatise on Zoology , ' edited by E. Ray Lankester , 1900 ; Echinodermata , Stelleroidea .
11 .
Jaekel .
' Zeitschr .
d. Deutschen Geol .
Ges .
, ' 1903 , vol. 50 , p. 106 ( Protokolle ) .
12 .
Ludwig .
'Zeitschr .
f. Wiss .
Zool. , ' 1878 , bd .
31 , pp. 347-394 , Plate 15 .
13 .
Idem .
' Bronnis Klassen und Ordnungen des Thier-reichs , ' II , 3 .
14 .
Lyman .
' Challenger Reports , ' " Ophiuroidea , " 1882 , vol. 5 , p. 6 .
15 .
MacBride .
' Camb .
Not .
Hist. , ' " Echinoderms , " 1906 , p. 501 .
16 .
Idem .
'Roy .
Soc. Proc. , ' 1907 , p. 440 .
17 .
Idem .
' Quart .
J. Micro .
Sci. , ' 1895 , vol. 38 , pp. 339-411 , Plates 18-29 .
18 .
Meek and Worthen .
'Geol .
Surv .
L. Illinois , ' 1869 , vol. 3 , p. 526 , and 1873 .
vol. 5 , Plate 16 , and ' P. Ac .
Not .
Sci. Philad .
, ' 1869 , p. 83 .
19 .
Salter .
'Ann .
Mag. Not .
Hist. , ' 1857 , ser. 2 , vol. 20 , p. 331 , Plate 9 .
20 .
Idem .
Ibid. , 1861 , ser. 3 , vol. 8 , p. 484 , Plate 18 .
20a .
Schondorf .
' Wiesbaden , Jahrb .
Ver. Natk./ 1909 and 1910 .
20b .
Idem .
' Palaontographica , ' 1910 .
21 .
Sedgwick .
'Student 's Text-book of Zoology , ' 1909 , p. 206 .
22 .
Sladen .
' Proc. Geol .
and Polyt .
Soc. West Riding Yorks , ' 1880 .
23 .
Sollas , W. J. 'Rep .
Brit. Ass .
, ' 1901 , p. 643 .
24 .
Idem .
' Phil. Trans. , ' B , 1903 , vol. 196 .
25 .
Idem .
' Quart .
J. Geol .
Soc. , ' 1899 .
26 .
Sturtz .
'Palaontographica , ' 1886 , vol. 32 , p. 75 .
27 .
Idem .
Ibid. , 1890 , vol. 36 , p. 238 .
28 .
Idem .
' Verhand .
Not .
Ver. Rheinl .
, ' 1893 , pp. 1-92 .
29 .
Idem .
Ibid. , 1900 , p. 198 .
30 .
Sir Strassen .
'Zool .
Anz .
, ' 1901 , vol. 24 , p. 609 .
LAPWORTHURA : A TYPICAL BRITTLESTAR OF THE SILURIAN AGE .
Note on the Construction of the Models .
By Prof. Sollas .
Plaster models such as we have used in this study have many advantages over wax ones ; they are more enduring , and can be readily reproduced .
To obtain a plaster model is , however , by no means an easy problem ; the obvious solution would seem to be to cast from a wax one , but if the latter be at all complicated , this is a task which even an expert will not undertake .
The only way is to build up the plaster model in successive sheets , a method which is rendered possible by the fact that plaster cast upon freshly-set plaster will adhere to it .
The steps in the process are as follows :\#151 ; Tracings of the photographed sections are first made on transparent paper , and round each a rectangle is drawn to provide for exact registration .
Sheets of wax are prepared of uniform thickness ; placing one of these over a tracing the outline of the section is cut out with a fine pointed scalpel and the sides of the sheet are cut to correspond with the surrounding rectangle .
When all the sections have been thus treated , the rectangular sheets of wax are superposed in order and the model , which exists now as a cavity in the wax , is examined to see if the registration is true .
If this is found to be the case we may next proceed to reconstruct by casting .
The first sheet in the series is attached to a plain sheet of wax which serves as base , plaster is then poured into the hollow of the section and allowed to set ; directly afterwards the sheet with its contained plaster is placed in a specially constructed machine by which both sheet and plaster can be planed down to a definite thickness .
The sheet is then removed from the machine and the next sheet in the series is registered upon it and attached , plaster is poured into the hollow , and wherever this rests on the plaster of the first sheet adherence takes place ; the fresh surface is planed down , a third sheet added , and so on till the series is complete .
It then only remains to remove the wax and set free the plaster model .
This is an operation which requires great care .
I have tried various methods , all of them depending on melting or melting and dissolving the wax ; but none of them are quite satisfactory .
The difficulty in all cases arises from softening of the wax before it melts , the weight of this softened wax pulling upon the slender processes of a model is often sufficient to break them .
The process is no doubt laborious , but it has the great advantage that the model , once obtained , may be handed over to an expert in making plaster casts , who will be able to take any number of copies from it .
To obtain sheets of wax of uniform thickness , the following method may be used : Two rectangular shallow tin dishes are constructed , one to fit inside the other .
I find 50 mm. by 30 mm. a convenient size for the inner dish .
The larger one is filled with water , to serve as a water bath , the inner with glycerine , and the whole is heated to above the melting point of the wax .
A large quantity of melted wax is prepared , and a measured quantity , sufficient to give a sheet thicker than is required , is ladled out and poured over the hot glycerine .
It is then allowed to cool and solidify .
As it solidifies it must be cut away from the sides of the dish or it will be torn by contraction .
The sheet is lifted from the glycerine by means of a false floor of perforated zinc , washed with water , and allowed to dry .
It is then covered by a thin sheet of paper on each side , and run through a roller press ( with hot rollers ) such as is used by photographers for burnishing photographic prints .
The interval between the rollers is adjustable , and jiniform sheets of any desired thickness can be obtained in this way with a perfection not to be met with in those put upon the market by German firms .
The planing machine is a steel bed resting upon four millimetre screws at the corners , it works vertically up and down in a frame , the upper sides of which provide a sliding surface for the planing tool .
The importance of planing to a measured thickness is due to the fact that it is impossible to attach* the sheets of wax one to another in such a manner as to secure perfect contact between them ; a small but measurable interval always exists between the two opposed surfaces , and the sum of these intervals in a long series of sections would give a markedly greater thickness to the model than it ought to possess .
By planing down after the addition of each sheet the model is kept to its true thickness .
* The simplest way to attach the sheets is to stick a hot needle through them , the melted wax makes a firm connection when solid ; unfortunately , however , it intrudes by surface tension between the sheets , and thus establishes an interval .
DESCRIPTION OF THE PLATES .
Plate 9 .
1 .
Lapivorthura miltoni ( Salter ) , abactinal surface .
X about 4 .
2 .
L. miltoni , actinal surface .
X about 5 .
3 .
Rhodostoma leptosoma , abactinal surface .
X about 4 .
4 .
R. leptosoma , actinal surface .
X about 5 .
Plate 10 .
1 .
Model composed of horizontal sections of one oral angle of Lapworthura miltoni , abactinal aspect .
X 20 .
2 .
Same model as in fig. 1 , actinal aspect .
3 .
Models of skeleton of two arm segments , actinal and abactinal aspects .
X 20 .
4 .
Model composed of vertical sections of three complete oral angles .
Solid'S \amp ; Sollas .
Phil Trans. B , Vol. Plate 9 .
Grout intaglio et imp .
Sottas and Sottas .
Phil. Trans. , B. vol. 202 , PI .
10 .

