Transcriber Note

Magnificatons are denoted as 30/1. The » is used in Swedish for quotes and dittos. Numeric abbreviations N:o, 1:o, 2:o, etc. for Number, primo, secundo, etc. Commas are used here as decimal separators (ex., 0,06 = 0.06.)

KONGL. SVENSKA VETENSKAPS-AKADEMIENS HANDLINGAR. Bandet 34 N:o 8.

RESEARCHES

ON

THE VISUAL ORGANS

OF

THE TRILOBITES

BY

G. LINDSTRÖM.

WITH SIX PLATS

COMMUNICATED TO THE ROYAL SWEDISH ACADEMY OF SCIENCES

SEPTEMBER 12TH 1900.

STOCKHOLM

KUNGL. BOKTRYCKERIET. P. A. NORSTEDT & SÖNER

1901.

CONTENTS.

At the outset of this memoir it must be said, that it is in fact a joint work by the first discoverer of the remarkable hypostomic structure to be described, Herr G. Liljevall, and by myself. A considerable portion of this work belongs to him as he with great discernment searched for a large amount of the material, prepared it as to be fit for the many microscopic sections which he executed, and at last with his well known artistic skill drew the figures. My share, again, in the common work consists in the supervision of the figures, in the arrangement of the observations and details in a coherent form, in which suggestions and remarks of Herr Liljevall have been of great use, further to write the descriptions and to make all the necessary researches in literature, so whatever may be faulty in these respects must be laid to my charge.

September 12th 1900.

G. LINDSTRÖM.

INTRODUCTION.

It is well known amongst palæontologists that the hypostoma of the trilobites has been the particular object for research and description by a few authors, as Barrande, Novák and Brögger and some passing notes concerning it are given in the descriptive works of other authors. In the following pages reference is often made to them and I may here give a list of the chief works and the abbreviations with which they are cited below in this paper.

Angelin Palæontologia Scandinavica = Ang.

Barrande Système silurien de Bohéme = Barr.

Brögger Die silurischen Etagen 2 & 3 im Christianiagebiet = Br. 1.

—— Die Ausbildung des Hypostoms bei einigen Asaphiden Br. II.

Hall, James, Palæontology of N. York, vol. I, II, VII = H.

Holm Trilobiter ur Dalarnes graptolitskiffer = Hm I.

—— Ostbaltische Trilobiten, Illæniden, = Hm II.

Novák, Studien an Hypostomen Böhmischer Trilobiten. Three different papers between 1884-1886. = N. II, N. III, N. IV.

Salter Monograph British Trilobites.

Schmidt Ostbaltische Trilobiten I, II, IV. = Schm.

Woodward, Henry, Carboniferous Trilobites. Pal. Soc. 1883-84.

Hypostoma of Bronteus, front and side views.
a. anterior margin. b. posterior margin. c. anterior pair of wings. d. posterior pair of wings. e. anterior groove. f. posterior groove. g. lateral grooves, h. maculæ.

Before proceeding further I shall give a general representation of the shape of the hypostoma. This enigmatic part of the trilobite skeleton resembles as to its outline a heraldic shield being broad at its anterior margin, gradually tapering towards the posterior extremity, which generally is acuminated or, as in the Asaphidæ, bifid. In a few, as Deiphon, Remopleurides, it is nearly square. Its exterior surface may consist of a single field or he divided transversally into two or three fields, according to the presence of one or two shallow grooves. These are transverse, curved forwards near the lateral margins and may be designated as the anterior groove and the posterior groove. Beside these grooves there are in some instances others, one on each of the lateral margins, parallel with these. On both sides of the anterior margin a protuberance, sometimes of considerable length, stands out. They are bent backwards towards the interior side of the hypostoma and in some genera, as Dysplanus, assume the shape of long, acuminated, halfways hollow horns. These protuberances were called »ailes» by Barrande and it is best to retain this term, in english »wings», in spite of its being not quite expressive. A smaller posterior pair lies nearer to the posterior margin and these wings are so much bent backwards, that they are seldom visible from the outside. They are prolongations from the »duplicature» or the narrow fold of the posterior margins round the interior side of the hypostoma.

For the rest the whole exterior surface is covered with terrace lines running concentrically and varying according to the different genera and even species. Many other genera, again, as Acidaspis, Calymmene, Homalonotus, Dalmanites etc., are devoid of them and instead granulated or smooth.

In the plurality of species there are two tiny patches or maculæ, sometimes elevated above the surrounding surface like tubercles and so they have also been called by some authors. But I have preferred to use the name »macula» for them, as the plurality does not form tubercles. They are generally entirely smooth and glossy and situated next to the anterior groove, either above it or in it, at a regular distance from each other and the lateral margins. They may form a sunk spot or, as commonly, an ovoid or elliptic area surrounded by a linear elevated border. Thus amongst the Asaphidæ. In others, again, (Bronteus, Illænus etc.) they form a moderately elevated tubercle. The direction in which they are oriented in relation to the longitudinal axis of the hypostoma is quite as variable. It is to the closer consideration of the nature of these macula that this memoir will be chiefly devoted. They have been delineated several times, but very seldom has any accurate attention been paid to them by previous authors. It is highly doubtful if Barrande ever alluded to them, when he in his »Système Silurien de Bohéme» vol. I, p. 156, in describing the hypostoma, says: »Cette plaque bombée porte souvent des saillies et des empreintes creuses, dans le voisinage de la bouche. Leurs formes varient suivant les espèces mais en conservant toujours les caractères génériques. Nous les considérons comme les points d'attaches des muscles et des màchoires.» In the numerous specific descriptions in his grand work there is never any mention made of the macula and in contemplating the figures and reading the explanations of these, we shall find that he meant something more comprehensive with his statement, to wit, the entire groove where these tubercular macula are situated.

In Novák's third paper, p. 4, we see the figures of two Phillipsiæ with tubercles on the hypostoma and he says only that in the anterior groove there are two such symmetrically placed ... glandular intumescences.[1]

[1] »In der Mittelfurche sitzen 2 kleine symmetrisch gelegene, nicht immer dentliche drüsenartige An schwellungen.»

Brögger again in his paper on the »Ausbildung des Hypostoms» devotes a page (p. 19) to explain the nature of these maculæ and regards them as the exterior marks of two muscular impressions or points of attachment on the inside for muscles which have attached the hypostoma to the walls of the glabella or the head. The real nature of these so variable tubercular maculæ is, however, a quite different one, as we hope to be able to demonstrate in describing all their varying shapes in several genera.

The interior surface of the hypostoma shows all the protuberances of the exterior surface (tubercular maculæ, spines and granulations) as shallow pits, the exterior grooves being on the contrary elevated ridges. For the rest the interior surface is smooth and its margins are posteriorly covered with a more or less broad duplicature, from the lateral margins of which the posterior wings project.

The manner in which Liljevall became aware of the interesting structural features of which I am going to give an account, is as follows. We were about to describe and delineate some new or not sufficiently known Upper Silurian trilobites from the island of Gotland. In order to draw a specimen of Bronteus polyactin Angelin, which is rarely found in the Wenlock and Ludlow strata of Gotland, Liljevall was preparing and cleaning its hypostoma. The shape of this hypostoma, as represented in [pl. II] fig. 22, is clypeiform, tapering towards the posterior margin, which is regularly tongue shaped. The anterior margin is faintly curved and at both sides elongated in a broad, pointed wing. Its exterior surface is divided in three transverse fields, defined through two shallow grooves, the anterior and largest field occupying more than the moiety, the median one is crescent-shaped and narrow, and the posterior one is nearly of the same size and confluent with the lateral borders of the hypostoma, which near the median transverse axis of the latter project in a blunt angle on each side. The surface, for the rest smooth, is covered by concentric, irregular terrace lines. On the inferior edge of the upper groove two small tubercles are situated, one on each side closer to the lateral margins than to the central axis. The size of their longitudinal axis amounts to 0,96 mm. in the shortest and to 1,07 mm. in the longest specimens. This axis is parallel with the groove and consequently oblique to the longitudinal axis of the hypostoma. They are oblong or ellipsoid, their inferior apices bluntly pointed or rounded, varying a little as seen in the three specimens figured, (Pl. II figs. 23, 24, 25). For about two thirds their surface is perfectly smooth or rather glossy, and the lower third is covered with a compact accumulation of small granules. The extension of this peculiar accumulation varies, being somewhat larger in some specimens, and the limit which runs oblique towards the glossy part is more or less curved. Such a granule taken single is of the extremest minuteness, measuring 0,055 millimeter at the highest. Their outline is perfectly circular and semiglobose.

When Liljevall saw this group of granules he was at once struck with the perfect, outward similarity between them and the facets of the compound eyes on the upper surface of the head of this same species, and I fully concurred in the same view. It now became of importance to ascertain whether the intimate structure also corresponded in both, and microscopic sections were accordingly prepared. Although not so clear as in other species of this genus a vertical section of a cephalic eye (Pl. II fig. 21) shows a median row of black spots which compared with other sections (Pl. II fig. 9) must be the centres of the badly preserved lenses. This is confirmed by the horizontal section (Pl. II fig. 20) where the well known image of the polyedric facets appeared. In the same manner the granulated spot of the hypostoma shows in a vertical section even more clearly ([pl. II] fig. 27) the elongate, coherent lenses with a black centre and the polyedric form of the lenses in a horizontal section (Pl. II fig. 26). The assumption of a pair of small adventive eyes on the exterior side of the hypostoma became thus based on the clear evidence of the perfect structural agreement between them and the eyes of the head. But it was necessary to strengthen this evidence through extended researches on the hypostoma of as many trilobite genera as possible and also to study the intimate structure of the cephalic eyes in as many genera as could be accessible.

We shall now give first a general account of the structure of the eyes in the trilobites as far as we have been able to study them, and then proceed to describe the maculæ of the various genera, which we have observed.

With regard to the presence or non existence of visual organs, we must remark that a considerably greater number of genera than those which unanimously have been regarded as blind, also are devoid of eyes though they still by many authors are ranked amongst the oculate ones and we then had better first to make a review of them and thus to eliminate them from the number of the oculate species.

Blind Trilobites.

In the chaos of generic forms and in the great disagreement which prevails as to the systematizing of the trilobites of the Cambrian time, there is a thorough revision of them highly needed by a person having access not only to the literature, but also to the original specimens. It is almost impossible in the present state of things to tell with any degree of certainty how many well established genera had been living during that period. Hence the difficulty of fixing the systematic names of many specimens the visual organs of which are to be described.

My researches on the visual organs of the Cambrian trilobites are founded on the Angelinian Collection in the Swedish State Museum together with collections of foreign species, but also largely on the waste European and American literature, though we have to deplore the often occurring inexactitude of the figures, especially in the older works, and constructed or schematized figures in some of the newer ones, which give a quite false notion of the structure. There is no lack of figures to show how it ought to be, according to preconceived notions and, on the other hand, a great scarcity of representations, to show how it really is. In spite of all this there is a sufficiently great number of well established facts to demonstrate the organization of the Cambrian genera.

The trilobites of this division may be called blind only in so far as they have no eyes on the upper surface of the head, but they may have been provided with visual organs, though more imperfect, on the hypostoma as really seems to have been the case with some of them.

According to the structure of the head shield the blind trilobites may be subdivided into the following well characterized groups. These are:

I. Without facial ridge:
1) The Archæan Trilobites.
II. With facial ridge:
1) The Olenellidæ.
2) The Olenidæ and related.

I. Blind trilobites without facial ridge (= »eye lobe»).

Group 1. The Archæan Trilobites.

In these the head shield is in one piece without any facial suture and facial ridge, and without the least trace of anything that might be called a visual organ and they must consequently be considered as totally blind. In contradistinction to the following groups, excepting the oldest Olenellidæ, the head consists of only three parts, 1) glabella, and 2-3) the two fixed cheeks. These genera range from the oldest zone in which hitherto trilobites have been found, that of Olenellus (Holmia) Kjerulfi, to the zone of Paradoxides Forchhammeri, and some, as Agnostus, even continue as high in the Lower Silurian series as in the Brachiopod schists. Beside Agnostus the other genera are Conocoryphe (seven species in the Swedish Cambrium), Toxotis, Ctenocephalus (?), Elyx, Aneuacanthus, Conophrys and Microdiscus.

»Harpides» breviceps Ang., also belongs here. Anopocare of Angelin should also be regarded as one of this group. But it cannot be retained any longer because it is founded on two other, well known forms, being, according to Linnarsson, Peltura scarabæoides (pl. 27 fig. 1, a in Pal. Scand.) and young specimens of Sphærophthalmus alatus (ibid. figs. 1 & 2).

It is remarkable that some of the Conocoryphidæ have an imperfect facial ridge, to be compared with the commenced one in Sao Barr. I (pl. 7 fig. 9). So the American Con. trilineata and reticulate. Walcott U. S. Geol. Survey 10th Rep. pt. I, pl. XCV f. 5 & 6. It is, as it were, arrested in the development and these adult trilobites had stopped, where the larva of Sao was proceeding in its second stage. They are the forerunners of the blind trilobites with facial suture, belonging to the third group. It needs scarcely be mentioned that the genera now enumerated have hardly anything in common, beside the general character of the head, and that real affinity exists only between Agnostus and Microdiscus, and probably also between Conocoryphe, Ctenocephalus and Elyx.

Beecher[2] asserts that there is a suture in Agnostus, but in vain we have searched for it in numerous well preserved specimens and Dr Holm also denies its presence. Nor are there any signs of closed up sutures, which also could not possibly be expected in so early a stage of evolution. It may then be taken as well settled that a fundamental character in these the oldest[3] of all known trilobites is the total want of a facial suture and a compactness of the whole head shield which later is broken up in several parts through the disjunction of the free cheeks. In the Lower Silurian formation there are a few genera sharing in the same structure of the head shield, though by no means else related. Such are Dindymene, Areia, Carmon and in the U. Silurian Cromus. The two species forming Barrande's genus Dindymene are so dissimilar that Dind. Friderici Augusti had better to be removed to a new genus and the first described one to be retained as type of the genus Dindymene. The same is the case with Carmon, where the type species C. mutilus is blind and without free cheeks while the other species, known only by its fixed cheeks and glabella is one of the Olenidæ.

[2] Nat. Classification of the Trilobites, p. 183.

[3] Oldest in that sense that they are the descendants of an archaic precambrian stock, the chief characteristics of which they have retained in the main unchanged and persisting long ages after the close of the Cambrian times, some, as Agnostus, continuing high up in the Lower Silurian.

II. Blind trilobites with facial ridge.

This large division embraces the second and third groups or, with a few exceptions, all the rest of the Cambrian trilobites on account of a feature in the cephalic sculpture common to them all, though widely different as to its first origin in both. What forms the prominent and common characteristic of these two groups is the presence of the facial ridge, which emanates from the basis or the front of the foremost segment of the glabella and in a great variety of different shapes continues backwards near to the posterior border of the head. It has received several names as eye-line, palpebral lobe, ocular ridge, eye-lobe, ocular fillet (Matthew). In German it is named Augen-leiste, in French filet (Barrande) and in Swedish ögonlist.[4] Some authors make a difference between the more narrow part, calling it eye line, and the thicker posterior node, which they name the palpebral lobe proper.

[4] That name is the most current amongst the swedish authors, together with »palpebrallob»; Holm says ögonlob and frontallob.

As this peculiar ridge exists before any facial suture has made its appearance and separated the head shield in five parts, viz. the median glabellar part, the two fixed cheeks and the two free cheeks, and as it occurs in genera which never possessed any facial suture, and where no eye ever was formed, it is not adequate to call it an ocular ridge etc. the more so, as it, at least during a long series of genera succeeding each other, has had no connection whatever with any eye. I therefore propose to call it facial ridge (in swedish faciallist). It occurs on the head of almost all Cambrian trilobites, excepting the archaic ones, and it is retained in the later Cambrian Peltura, Sphærophthalmus etc., which have real, compound eyes, as well as in a few Lower Silurian genera as Triarthrus, Pliomera,[5] Euloma, in the Upper Silurian Arethusina and Acidaspis and most persisting in Harpes, ranging from the Lower Silurian Lower Red Orthoceratite Limestone into the middle Devonian beds.

[5] Pl. Törnquisti Holm.

It presents itself in the most variable shapes, and as it in fact can be followed through its development in the oldest known species, it is suitable to begin its description together with the characteristics peculiar to the oldest or second group in this large subdivision, that of the Olenellidæ.

Group 1. The Olenellidæ.

In this group we have two families of different age, the older, less developed the Olenellidæ proper, and the younger the Paradoxidæ. The former consists of the genera Olenellus, Holmia, Mesonacis and Schmidtia, and we shall attend to them first. They have no facial suture[6] and consequently a tripartite head shield like the Archaic ones, no eyes, but there is that strongly developed and most characteristic facial ridge. As seen in Olenellus Thompsoni[7] the crescent like ridge starts as a direct outflow from the base of the first segment of the glabella, and is in direct continuation with it as an integral part. It is roundbacked, regularly curved and at its starting point as broad as one of the posterior glabellar segments. It is regularly semilunate, tapers posteriorly and ends near the occipital segment. From the narrow second segment of the glabella, inside the just described larger ridge, a smaller ridge emanates, broader and flatter than the former, slightly curved, and ends between the third and fourth glabellar segment. Already in Holmia Kjerulfi the ridge is modified. There it is only one ridge, the anterior one, nearly as doubled through a shallow groove running along its back. The second one has dwindled away so as to be seen only as a narrow stripe near the occipital segment and ending outside this in a point. The anterior ridge, represented by Holm as consisting of two nearly independent parts, is indeed in one piece, though its dorsal groove sometimes is deep, and it is with its total breadth joint to the first glabellar segment. Along its outer edge, where it lies close to the cheek, a narrow slit runs, and I suppose that it is to be considered as the first indication of the forming of a facial suture, which however does not reach longer than the ridge. As in Paradoxides it is so tight to the cheek, that there has been no place for an eye. It is no accidental break, its edges being too regular and unbroken in all specimens. This ridge is by far much shorter than in Olenellus and terminates opposite the third segment.

[6] Beecher Natural Classification of the Trilobites, p. 191, pretends that in Olenellus and Holmia real sutures »in a condition of symphysis» occur. He seems to deny the facial sutures and to accept as »real sutures» the »internal sutures» described by Holm in Olenellus Kjerulfi. It is highly doubtful if these interior elevated lines are to be regarded as sutures. They are indeed no sutures, but in reality elevated linear ridges, inclosing, as it were, narrow canals. The real sutures known, the facial sutures, never form elevated lines, be it on the outside of the head or on the inside. Probably these lines are derived from some now unknown interior organization and it may be fit to remind of the somewhat similar though more numerous linear canals of the branchiæ on the interior surface of the great head shield of Apus. (Zaddach De Apodis cancriformis Anatome ... pl. II fig. 1) or what Huxley (Anatomy Invertebr. Animals p. 281) calls the convoluted »shell gland» in the carapace. A quite different structure is what I suppose to be the incipient facial suture in Holmia.

[7] Walcott in 10th Ann. Rept. U. S. Geol. Survey, pl. 82 f. 2.

Larva of Olenellus asaphoides, chiefly according to
Ford and Walcott.

Already Linnarsson observed the second ridge as he tells in a paper where he describes Olen. Kjerulfi for the first time.[8] It has been called an »ornamental» spine, but in the following we shall learn what it really is. This ridge in connection with its spine has not been observed in any other of the Olenellidæ, at least not in their adult stage.

[8] Öfversigt Vet. Ak. Förhandl. 1871, tafl. XVI fig. 1.

Thanks to the discovery of very early larval stages of the American Olenellus asaphoides, which Ford[9] and Walcott[10] have described and figured, we can through combination of these decipher the development and signification of the facial ridges. To facilitate my explanation I here join a series of cuts from the earliest stage to the more developed, with addition of two schematic stages to complete in a certain degree the series of Walcott and Ford.[11] See figures [p. 13].

[9] In American Journ. of Science 1877 p. 265 and 1881 p. 250.

[10] Bullet. U. S. Geol. Survey N:o 30, 1886, pl. 17 fig. 5-6.—Tenth Annual Rept. U. S. Geol. Survey, 1888-89, printed 1890, plate 88, fig. 1, 1 a.

[11] In his paper on »The larval stages of Trilobites» p. 175 Beecher gives a new figure (f. 6), original from Ford's collection, of the larva of Ol. asaphoides, but it is so sketchy that I cannot with certainty make out what it means. I cannot agree with him when he speaks of free cheeks and eyes in these and he is completely wrong when he says that the outer pair of spines belong to the free cheeks etc. (p. 176).

The [figure 1] has been hypothetically composed as a deduction from [fig. 2], which presupposes an earlier stage of development like that in [fig. 1], when there existed three or four pair of lateral appendages in the larva. This then consists of a central portion of five segments. The large anterior crescentlike segment does not, however, show any distinction between a central part and lateral appendages, it is nearly as large as the three next taken together and its backwards bent side horns embrace the next two posterior segments and attain with their narrow pointed tips the back of the fourth segment. The central portion consists of five segments, when the somewhat not well definite posterior marginal segment is taken in account. Each of these segments excepting the fifth one has lateral appendages, those of the second and third segment being quite as broad as the central part and bent backwards in a curve ending in a small pointed spine. The lateral appendages of the fourth segment are largest of all, more than double the length of the two next in front, triangular and standing out beyond the posterior border of the shell as a broad spine.

In [fig. 2] a great change has set in. There is no distinction between the lateral appendages of the third and fourth segments at left. These two have been fused together, they have united, so as to make the left triangular spine look larger than it was originally. The appendages of the right side are still in the same state as before. But this fusion of the lateral appendages also takes place in another direction, as shown in another specimen ([fig. 3]), so that the second and third appendages on both sides coalesce into one piece. Now it is easy to imagine that at last a complete fusion has set in between all lateral appendages and that instead of the original three on each side, there is only one large piece, reaching beyond the shell as a broad spine, as represented in the hypothetical [figure 4].

In the [figure 3] a progressive change is also seen in the transformation of the first central segment. From occupying the whole foremost space of the shell it has been lessened in size, more distantiated from the anterior border of the shell and rounded off, more prominent and definite from the lateral appendages, which have become narrow, though of the same length as before. The two first central segments seem to have been united into one.

Between the hypothetical [figure 4] and [fig. 5], there is evidently a great lacuna, not yet filled up. In the interval of time the two appendages, which we saw in [fig. 4], have been much modified, the posterior one having lost so much in bulk, and the anterior being lengthened and stretching out beyond the border of the shell in a narrow spine alongside the posterior one. The central segmented part has now assumed a shape, which on comparing it with the following stages of development makes its true nature evident and that it indeed is what in the adult animal becomes the glabella. The meaning of the previous stages then also is easily understood. The central segmental piece in them is the glabella or we may, as Bernard already has done, call it for rhachis and the side appendages for pleura, as this little larva represents the whole body of the future trilobite, and embodies all its parts in nuce. Through the great changes which these pleura undergo, it results, as we have seen, that two pair vanish, being incorporated with the large fourth pair and that only two rest for a while, the anterior one being the so much renowned frontal-lob or eye lobe and the second one the so called »ornamental spine», which in fact is a compound of the original second third and fourth pleura of the corresponding segments.

It is to be borne in mind that this larva, which represents the future trilobite in its earliest stage, is nothing but the head, or what in the adult takes the place of the head, and especially its dorsal surface and that it thus solely consists of the future head.

In the [figure 5] (Ford's fig. 1.) there are the first signs of the pygidium coming and in the [fig. 6] it is well developed, but the thorax is still non apparent. In the stage [fig. 6] both pleura have increased in length and the compound one also in bulk. They project with spiny points beside the beginning pygidium and the anterior pleura have united across the first segment of the glabella through a narrow ridge, which seems to cut that segment in two. The sequence in the order of development or growth thus is first the head, then the pygidium and last the thorax. At least it is so in these the oldest of all trilobites. But in nearly all trilobites of which there are good data, the head is the part first developed.

Evidently a large hiatus exists between the stages represented in figs. [6] and [7] (Ford's figs. 2 & 3, 1877), in the latter of which the animal, though not adult, has had the thorax and pygidium added to the head. The modifications in the size of the pleura are the chief changes. The anterior pair is reduced and retired within the posterior border of the head forming a semilunar arch joined with the occipital ridge in an angular bend. The posterior pair is enlarged and its spine is by and by reduced (figs. [8], [9], [10]) till it quite disappears and only the wide semicircular field between the first pleuron and the glabella is left behind as a remnant of its dilated body. It is to be remarked, that while in the plurality of the adult Olenellidæ all traces of the spiny projections of the second pleuron have vanished, they are still retained in the adult Holmia Kjerulfi, though not in the American Holmiæ, and thus give it at the same time a more ancient and a more larval stamp. The shallow groove along the back of its first pleuron indicates strongly its pleural nature, as the thoracic pleura commonly are divided through such grooves. The same peculiarity is also observable in several of the American Olenellidæ. It is much the same with the posterior pleuron, the pleural nature of which is revealed through its spine, that is homologous and identical with the spiny terminations of the thoracic pleura. We have thus through the remarkable finds of Ford and Walcott combined received an explanation of the morphological origin and nature of the facial ridge, the so called eye lobe and found that it has nothing whatever of the character of a visual organ. But it must be borne in mind, that these developmental changes are peculiar only to the Olenellidæ, the origin of the facial ridge in the later trilobites is, as we shall see, quite a different one.

The Olenellidæ belong chiefly to the oldest of the Cambrian beds with trilobites, and none of them has as yet been found higher up.

Next in order of evolution we have the important tribe of the Paradoxidæ. These are preeminently distinguished from the Olenellidæ through the well developed facial suture, which without exception in them all runs outside the facial ridge and separates this from the free cheek.[12] This is a great step forwards in the evolution and establishes the fact, demonstrable also by other evidence, that the formation of the facial suture is subsequent to the appearance of the facial ridge. This preexisting ridge seems to have had no small influence on the development of the suture, it checked its progress from the front or from the sides toward the fixed cheeks and directed its course against the genal angles. It lay as a protection for the glabella against this disseverance, causing the separating line to run along its outside.

[12] They have thus a quinquepartite cephalic shield, as the later trilobites.

This group consists of the genus Paradoxides proper, as well as of Centropleura, Metadoxides and Hydrocephalus, if this is an adult form and not the larva of an unknown Paradoxides. Perhaps such forms as »Conocephalites» Emmrichi Barrande, as well as Anomocare limbatum, An. excavatum, Bathyuriscus and Dolichometopus may on account of the shape of their facial ridges be considered as related to the Paradoxidæ. But this must be left for coming researches to decide. Some American Cambrian forms also share in this characteristic and may upon closer inspection be ranged here. So Zacanthoides. In these as in the true Paradoxidæ the facial suture follows the ridge along its whole length, while in the trilobites of the third group the facial suture is in contact only with the posterior end of the ridge, the so called eye lobe. Remopleurides does not show characters, that as Beecher thinks, could unite it with the Paradoxidæ. These are blind and Remopleurides has well developed eyes and an organization that gives it an isolated position in the system.

The facial ridge continues in a great variety of shapes, short or long, but always forming the fraction of a circle, always of nearly equal thickness, only slightly tapering towards one of the extremities, and always when in direct connection with the glabella, starting from the base of its foremost, largest segment. As a rule the ridge is more developed in the young or larval individuals, continuing from the glabella to near the posterior cephalic border in an uninterrupted arch[13] quite as in several of the adult Olenellidæ of the oldest Cambrian. It can be taken as granted that its origin is the same as in the Olenellidæ though at present the only evidence at hand is the small larva of Paradoxides oelandicus, which Linnarsson called Parad. aculeatus.[14] In this we see the anterior pleuron or the facial ridge alone present, elongated downwards like the same pleuron in the figure 6 of Olenellus asaphoides and terminating like this in a fine spine stretching backwards outside the posterior border. Of the second pair of pleura there is nothing to be seen. This must then have been aborted at an earlier stage than in the Olenellidæ.

[13] G. F. Matthew had before me, as I now find, pointed out this distinction in his memoir »Illustrations of the Fauna of St. John» N:o IV, p. 163. When he speaks of »the embryonic stage» in this and other passages he evidently means »larval stages», as the embryonic stages of necessity must remain unknown to us. (Later remark.)

[14] Om faunan i lagren med Paradoxides ölandicus (1877), p. 359, pl. 14, f. 11.

The connexion between the free cheeks and the middle part of the head has been very lax not only in the Paradoxidæ, but on the whole in nearly all Cambrian trilobites with free cheeks. When the free cheek is dissevered it shows no trace of the ridge, there is only a large scallop on the spot where it embraced the ridge. In the Paradoxidæ the rim of the indenture and the ridge are in so close contact that there is not the least place for an eye between them, as can be seen in the few specimens with a complete head. In all oculate trilobites again without any exception the facial suture separates that part of the eye which is the real visual organ with corneal facets, from the interior often elevated portion, opposite it, the so called palpebral lobe. The eye is always placed on the free cheek,[15] the lobe again always on the fixed cheek of the head shield. No real eye exists without the palpebral lobe, and, on the other hand, that part of the facial ridge which later in the development changes to a palpebral lobe, occurs alone without any eye in a great number of Cambrian trilobites, and consequently these are blind and such is the case with the Paradoxidæ and a great number of the succeeding.

[15] Excepting in Harpes, which has no free cheeks.

There is not the least evidence to support the suggestion that the »ocular ridge» is homologous, with the eye of Apus[16] and that the real crystal cones lay sunk beneath the surface in a »water sac». As we, for instance, in Peltura have an »ocular ridge» (= facial ridge mihi) on the fixed cheek and opposite its posterior extremity, the »eye lobe», a real eye with facets on the free cheek it is not likely to suppose that the »ocular ridge» nor the »eye lobe» ever functioned as a visual organ or that two widely different sorts of eyes were placed in closest vicinity opposite each other.

[16] Bernard, The systematic position of the Trilobites. Qu. Journ. Geol. Soc. 1894, p. 411.

It may be worth while here to remind of the great similarities, whether analogous or homologous, in the formation of the superior surface of the head in the trilobites and the embryons and the newly hatched larva of Limulus. The latter have the head shield separated into five portions, partly through a facial suture which, as in Peltura, Dalmanites and others, forms a continuous line around the glabella. This suture divides the eye node in a similar way as in the trilobites, that is, sectioning it in two parts, of which one adheres to the central fixed cheeks and the other to the free cheeks.[17] And to judge by the figures of Kingsley the former, the whitish moiety is the first developed and sometimes for a while quite alone as the facial ridge of the trilobites and probably also anterior to the suture, as this is not complete at this stage. This white node reminds of the small facial ridge in Arionellus ceticephalus Barr.

[17] Packard, Development of Limulus polyphemus. Memoirs Boston Soc. Nat. H. vol. II, pl. V, fig. 25. Nothing is said about the exterior structure of the eyes.

Dohrn, Zur Embryologie and Morphologie des Limulus Polyphemus. Jenaische Zeitschrift 1871. A very good figure (pl. XIV, f. 4) shows clearly the two parts of the eyes, the interior one being larger.

Kingsley also (Devel. of Limulus, Journal of Morphology, vol. 7, 1892, pl. VI, fig. 34) has in the last larval stage the suture and the eye in two parts, of which one is white lying inside the suture and the eye proper, black, outside.

As Kishinouya[18] has already pointed out and as I have anticipated above the head of the most developed trilobites in their adult state, and the head of the larval Limulus consists of five parts, viz. 1) the glabella in the centre, 2 & 3) the fixed cheeks, 4 & 5) the free cheeks. An elevated ridge in the adult Limulus shows where the suture once lay and it is on the outside thereof that the eye of the adult is placed. What other authors call the ridge or the eye ridge in Limulus, Kishinouya rightly names a suture.

[18] Journal College of Science Tokio, vol. V, p. 53, 1892.

It was the renowned Swedish naturalist Wahlenberg, who first recognized the importance of the facial suture, which he called »linea ocularis»,[19] but to another Swedish palæontologist Dalman[20] the exact definition of this suture is due, to which he gave the name still in use. He expressly remarks that the suture crosses the ocular node and limits the outside of the »lobus palpebralis» and he makes a clear distinction between that lobe and the »tuberculi and eminentiæ oculares» (= facial ridge) of which he says that they are the more or less evident elevations situated in the blind Palæades on the place of the eyes (in which he is wrong) »and which perhaps are an indication of such organs». But then he says doubtfully (p. 255) »tuberculorum ocularium veram naturam determinare haud ausi sumus, etsi oculorum formam sat bene exhibere videantur» and he adds concerning Paradoxides »oculi nulli, eorum loco autem tuberculi duo».

[19] Additamenta quædam ad petrificata telluris Suecana, in Acta Upsaliensia, vol VIII, 1821, p. 294.

[20] Vet.-Akad. Handl. 1826 p. 126.

Group 2. The Olenidæ and related families.

Next in the ascending order we have the largest group of trilobites in the Cambrian, of which the greatest part is formed by the Olenidæ. A facial ridge different in shape and different in development from that in the former group characterizes them. Unlike the semilunate ridge of the Olenellidæ it issues mostly from the front of the first segment of the glabella and goes generally backwards till it meets the facial suture. It is narrow and fine as a thread, but for the rest assumes a great variety of forms. It may be curved as a circle segment as in Sao, Liostracus, it may be long and straight, standing out in a right angle to the glabella as in Eurycare, it may be short and straight in an acute angle to the glabella as in Parabolina and so forth. In the same genus, as for instance Olenus, straight and curved ridges occur in the various species, and consequently the form of this sort of ridge cannot be used as a generic character. It moreover differs from the ridge in the Olenellidæ and the Paradoxidæ in having the posterior extremity widened as a tubercle or node, which commonly has been called the »palpebral lobe», while in the older groups the ridge is only at the most a little pointed in both ends or of equal thickness.

Some persistent archaic genera as Conocoryphe and Elyx inform us how this ridge has originated and how widely different it is from that in the Olenellidæ. On the surface of their head shield no ridge is visible, but on the interior side of it we behold, as the figures show ([Pl. VI, figs. 43 & 44]) on each side of the glabella a ramifying system as of the most minute vessels, which spread and cover the whole surface of the cheeks. This reticulation issues from a main trunk that goes nearly straight in an oblique direction from the foremost segment of the glabella and emits narrow branches forwards and backwards on its both sides. And these branches go on dividing till they occupy the entire interior surface of the fixed cheeks. The figures representing this are taken from Elyx laticeps [fig. 43], and an almost similar image has been obtained from a species of Conocoryphe [fig. 44].

Now the question may be asked, what does this network of branching canals mean? I think we cannot gain a more plausible answer than that given us through the inspection of the soft parts that lie hidden behind the glabella of Limulus and its fixed cheeks, the only living crustacean which offers the greatest homologies with the head shield of the trilobites. Next below the shell of the glabella and the cheeks of Limulus there is a complicated stratum of muscles and behind this the heart and the great central circulatory system spreading from that centrum towards the sides of the head shield, the vessels being the more fine and minute, the more they are elongated from the centre. I now suppose that likewise in these trilobites the narrow and prominent glabella has been the receptacle not only of a strong mass of muscles, but also for the central part of the circulation. This centre has there sent out two relatively strong ducts or canals, one on each side of the glabella feeding all soft tissues inside and near the cheeks and probably also other important parts of the body. It is also clearly seen in Elyx how the orifice of the main trunc opens in the hollow of the glabellar apex.

In all crustaceans, as far as is now known, their more or less hard calcareous or chitinous skeleton is moulded by the subjacent tissues and glands. The sculpture of the surface consequently is an outcome of the fashioning procedures showing what has been going on below and what is still going on. If we then on the exterior surface of the head shield of the trilobites see the radiating lines in relief we must conclude that they are due to a subjacent system of almost capillary vessels causing hollow impressions on the inside and elevated ridges exteriorly. In Elyx the vessels have made no strong impression as to be visible on the outside. In a couple of American Conocoryphæ again (C. reticulata and C. trilineata) the main trunc of the vessels has formed a short faintly elevated ridge. In Solenopleura ([Pl. VI, f. 45]) the facial ridge has been fully developed, and by the casts of the inside it is clearly seen that the main trunc of the vessels makes the inside of the ridge and has been much incrassated. On the other hand the smaller branches issuing from it have been reduced in size and number, but are still visible. Thus, if I am right, the two, the facial ridge and the main trunc, are in causal connection and the former has been moulded by the later, when it had gained in size sufficiently, and when a richer affluxion of nutritive fluid was directed backwards towards the point where the eye at last originated. The further changes in this ridge thus are related to the development of the eye. As far as I am aware, there are as yet no data to tell us whether the appendages or pleura of the glabella in the larval (or primordial) Olenellidæ have the same origin as the ridge of the Olenidæ or are homologous with it. Perhaps it may be so in respect to the foremost one, which also is persisting. It is, however, not known how the pleura in the olenellid glabella have been developed, whereas it is well ascertained by the growth of the glabella in Sao, Liostracus and others that such pleura never have been developed in them.

What the phylogenetic evolution has taught us concerning the formation of the facial ridge is confirmed by the ontogenetic development of individuals of some species as for instance Sao hirsuta Barr., Liostracus sp. Brög. and others.

This development can be followed in detail through the excellent figures of the larval stages of Sao hirsuta which Barrande has given in his magnificent work, vol. I, pl. 7. In the first stage figured (figs. 1 d-e) the whole animal consists only of the head shield which is completely smooth, the glabella scarcely segmented, no facial ridge, no facial suture. In the second stage, according to me (figs. 2-4 a), the glabella has become distinctly segmented, and the pygidium and partly the thorax have been added. In the third stage, 4 c, d-9, one small narrow string exits from each side of the front of the glabella, making the first faint beginning of the facial ridge. They form a right angle with the glabella. In continued growth they become by degrees a little more curved (figs. 5 c, 6 b) and the lengthened ridge bends parallel to the outer lateral margin of the head (fig. 9 b). In fig. 9 d it has become so far complete that it reaches nearly back to the posterior cephalic margin, but is still of the same narrowness all along. In the fourth stage, in a specimen (fig. 10 a) of 3 mm. in length, the facial suture makes its first appearance, setting in from the posterior margin of the head and meeting in its forward progress the posterior extremity of the facial ridge which now begins to swell out. It seems that both have a mutual influence on each other, the suture being deviated from a straight course[21] to take a bend outside the ridge, and the extremity of the ridge again at this contact to increase in size so as to form the elongated tubercle, often called palpebral lobe (fig. 14 b). This now augments in the same rate as the whole body. The characters of the four stages of the development of the larva consequently are: 1:o The archaic stage, only head shield with ridgelike glabella. 2:o The coming and growth of the pygidium and the complete segmentation of the glabella. 3:o The coming and growth of the facial ridge. 4:o The coming and development of the facial suture.

[21] In Trinucleus and Ampyx where there is no facial ridge, the suture has a straight direction along the lateral margins.

Barrande regards the whole ridge as a prolongation of the eye and the tubercle at its posterior extremity as the eye itself.[22] But, again, in page 399 he says »Sa surface (of the eye) est toujours mal conservée, pour nous permettre de voir si elle était réticulée.» And he also confounds the ridge, »filet» as he names it, with the eye itself, and the first faint beginnings of this ridge in his fourth stage he considers as the eye.[23] By a partly schematical figure of the free cheek (fig. 25) he places the eye on this cheek, and in the same manner in fig. 29, »restaurée d'après divers fragments» he figures a reticulated surface of the eye on the free cheek, outside the tubercle. I have sought for a reticulated surface on sufficiently good specimens, but never found any, and I must consider Sao as one of the blind genera. Barrande himself also in the table on the eyes of the trilobites places Sao in the group with »Surface visuelle inconnue», p. 131.

[22] Page 383 »l'œil argue est prolongé par un filet en relief, vers le front de la glabelle».

[23] p. 389.

(0 is new, added here to supplement fig. I.)

Brögger has also succeeded[24] in finding a series of small larva, which he considers as belonging to a species of Liostracus. As the figures drawn by Brögger twenty five years ago may now be very little known, I here reproduce them with the kind permission of the Editors of »Geologiska Föreningens Förhandlingar», where they were published in 1875. This development proceeds nearly upon the same plan as in Sao. The first stage, however, (I) seems to be much earlier than any of Sao, the rhachis or glabella consisting of an unsegmented ridge of more primitive appearance. Before this first stage of Brögger's larva a still older phase of development can be imagined, (0) a simple rounded, smooth head shield without any indication of a glabella at all. This stage might correspond to the head of certain species of Agnostus, as A. glandiformis, A. nudus, which have no glabella. This stage 0 is also valid for the larva of Sao. Several stages are evidently wanting between I and II, in the later of which the thickened glabella is divided in four segments. In III we have six segments, all these three stages consisting only of the ovate head with narrow fixed cheeks. In IV the pygidium has been added to the primitive head, but the segments of the glabella have been reduced to four and in V slightly altered in shape In VI, again, we see the head with five glabellar segments and scarcely the first sign of the facial ridge. Between VI and VII there must be links missing, as the change can not be so abrupt, and likewise between VIII and IX as in VIII there are still no free cheeks nor any facial ridge. This interesting discovery of Brögger confirms, together with those of several other authors[25], the supposition that the development of the later Cambrian and older Silurian forms is a quite different one from that of the Olenellidæ and the Paradoxidæ. They have a rhachis, but no pleura proper, as the single facial ridge has a quite different signification and appears at a comparatively much later stage than the facial ridge of the Olenellidæ, which is present from the earliest stages known.[26]

[24] Fossiler fra Öxna og Klettna, Geol. För. Förhandl. 1875, p. 572, pl. 25, fig. I-X.

[25] Foremost among these stands Matthew in »Illustrations of the Fauna of St. John» IV, where he, p. 143, pl. II figs. 1 f. etc., describes a few stages very like those given by Brögger, so the glabella as an unsegmented, narrow ridge etc.

[26] A deviating form of the ridges is shown by »Liostracus» tener Hartt, Acad. Geol. 2d Ed. p. 652 (see also Matthew Illustrations of the Fauna of St. John [1887] p. 132 p. 1 f. 3 a-3 c). Beside the usual facial ridge there is a second pair of ridges between the first and the glabella, arching in an opposite direction. See also Walcott, Bulletin U. S. Geol. Survey, N:o 10, 1884 pl. V, figs 6, 6 a, 6 b, new figures and copy of Hartt's description.

This form of facial ridge, although prevalent in the middle and later Cambrian times, dates back so far that genera coeval with Holmia, viz. Ellipsocephalus and Arionellus show it along the facial suture. This early occurrence of the ridge coeval with the less developed Olenellidæ leads to the assumption of a different origin of these trilobites as a branch, which already far back in the oldest Cambrian or precambrian times had deviated from a common ancestor and I have tried to give a view in tables of these two coordinated lines of evolution further on, at pages 24 and 25. The most remarkable genera which, as far as I have found, belong to this fourth group are Ellipsocephalus, Arionellus, Liostracus, Olenus, Leptoplastus, Parabolina, Corynexochus, Parabolinella, Sao, Ptychoparia, Doropyge, Oryctocephalus etc.

I have not been able with an absolute degree of certainty to recognize whether some of these genera now enumerated, have been oculate or blind like Sao, like the Paradoxidæ and similar. The precarious state of preservation prevents all definite conclusions in that respect. It seems, however, that the evidence gathered through the examination of numerous specimens rather points in a negative direction. As the free cheeks in these old Cambrian trilobites have been in a very loose connection with the fixed cheeks and generally deciduous, contrary to the condition in the Silurian ones, it is in many instances very difficult to tell whether species with facial ridge, especially those from the earlier Olenus schists have been blind or provided with eyes.

The order of succession of the genera in the Swedish uppermost Cambrian, in the Olenid slates is, according to S. A. Tullberg's researches[27] at Andrarum in Scania as follows:

1. Parabolina (oldest division of the Olenus slates).
Olenus.
Liostracus.
2. Eurycare.
Leptoplastus.
3. Peltura.
Sphærophthalmus.
Ctenopyge.
4. Cyclognathus (uppermost).
Acerocare.

[27] Om Agnostusarterna vid Andrarum.

It is already in the second division that we find the earliest oculate genus in Eurycare and in the higher strata. Sphærophthalmus and Ctenopyge with enormous hemispherical eye balls, while Olenus and Parabolina, their earliest predecessors, probably were blind. We have not amongst hundreds of specimens of these genera found a single specimen showing an ocular globe covered with facets. Olenus and the nearly related genus Parabolina are found in innumerable specimens in the thinly laminated alumschists of Scania and other provinces of Sweden. But rarely a perfect head shield, or nearly so, is to be found with the free cheeks in place. If so, the semicircular scallop in the free cheek is entirely filled up by the posterior lobe of the facial ridge and there is no place left for any eye.—If we, again, find a non compressed glabella and fixed cheeks likewise, the facial lobe (-»eye-lobe») is in some elevated so much as to leave a little space between it and the scallop of the free cheek, which space must have been an empty lacuna if not filled up by an eye ball. But it may also be that the free cheek has been somewhat put out of its order and that consequently some space has been left between it and the free cheek. It is quite as much with older genera, as Solenopleura especially, in which the posterior lobe of the facial ridge (vulgo »eye lobe») has attained a great development, and in which one just could expect to find a sphærical eye resting between the elevated lobe and the scallop. The elevated rim of this scallop does not in fact constitute a proof for its having clasped an eye as the elevated scallop in Sphærophthalmus did. In Paradoxides, again, in Ellipsocephalus, where there is absolutely no trace of an eye ever having been present, whenever you succeed to find the free cheek in juxtaposition with the fixed cheek it is evident that the elevation of the scallop rim is due to the impact of the posterior lobe of the facial ridge. In many specimens of Dolichometopus and Corynexochus etc. no free cheeks have ever been found and to judge by the shape of the facial ridge it may be concluded that these also were deprived of eyes.

On seeing this great number of trilobites, that on account of their organization must be considered as blind, the first suggestion that strikes the mind, is that they must have lived in abyssal depths of the Cambrian sea, where the most intensive darkness prevailed. Nor does the nature of the strata contradict such an assumption as evidently this fine sediment must have been deposited far beyond the reach of the influence of the wave motion, a depth amounting to more than a thousand metres as now calculated, for else it could not have preserved unbroken such delicate parts of organisms as that free cheek of Ctenopyge pecten with its extraordinary long and delicate horn, figured on plate III fig. 27 and many others. But it seems incredible that such a state of things should have prevailed during the deposition of all the Cambrian strata, although they in Sweden amount to only 160 feet in thickness, according to the evaluation of S. A. Tullberg, not considering what has been lost through denudation. The length and duration in time can in this instance not be measured by the thickness of the beds, but by the great changes in the faunas which there have succeeded one another. The physical conditions, to judge by the composition of the rocks, seem in the main to have lasted during immeasurable periods and still the fauna has changed in no little degree. During that enormous length of time, embracing in Sweden eight well separated periods, there must, however, have been minor changes in the conditions of depth and consequently in the nature of the depositions. To a certain extent the physical agents must have influenced the organization of the animals, but not essentially. There are sure evidences of another factor being the chief agent and that is the evolution.

We have from the lowest Olenellus beds to the lowest Lower Silurian strata followed the clear traces of the changes from the eyeless Olenellidæ past the Paradoxidæ with the facial suture to the great tribe of the Olenidæ in which the facial ridge with its lobe is so prominent, and to the oculate genera at the top of the formation.

As stated above there seems to be evidence enough for accepting two different lines of evolution in these the oldest trilobites. For the oldest, Olenellus is the type and for the second Sao may be taken as the representative.

These two evolutional lines could be represented with their phases in chronological succession as follows:

A. The Olenellus line.

1. The trilobite consisting only of the head with rhachis and pleuræ, no sutures, no ridge, no eyes. Adult forms precambrian. Corresponding larval forms (Olenellus asaphoides) in the Lowest Cambrian. Still older and simpler forms may be presupposed as preceding these.

2. Semilunate facial ridge on the tripartite head-shield, no sutures, no eyes, but a hypostoma fused to the rostrum of the head and provided with maculæ. The adult animal with thorax and pygidium in the oldest Cambrium, the Olenellus zone. Genera: Olenellus, Holmia (with beginning suture), Mesonacis, Schmidtia.

3. The head quinquepartite with facial suture, short semilunate ridge in the adult, a long ridge in the young, no eyes. The genera occur in the Paradoxides-zones of the Cambrian formation. These genera are Paradoxides, Centropleura, Metadoxides, Hydrocephalus, which are direct descendants of the Olenellidæ.

This line of evolution, in which the species never acquired eyes, was extinct in the sixth zone of the Swedish Cambrian, Centropleura Lovéni being the last.

B. The Sao line.

I call this so because we see in the development of Sao, as represented by Barrande, most of the different' phases in evolution, which the ridge bearing trilobites of this second group have experienced during the Cambrian times.

I.

The body consists only of the head, neither pygidium, nor thorax yet developed.

1. Psilocephali (ψιλος, bald), primeval, precambrian, adult stage not found, supposed to be like the head shield of Agnostus glandiformis and A. biplicatus quite round and bare without any glabella nor sutures. The corresponding stage in development also wanted.

2. Glabellate. The entire animal only head with a mesial ridge, which at first is entire or unsegmented and later is metamorphosed into the segmented glabella.

a. unsegmented., Precambrian, adult form unknown, supposed to be like the first larval stage of Liostracus according to the figure I of Brögger and also as the head of Agnostus parvifrons Linnarsson. Corresponding larval stages Liostracus (I).

b. segmented. Probably precambrian, not represented as adult in the Cambrian strata. Supposed to be like the larval forms II and III of Brögger's Liostracus, consisting of the round or oval head shield and the segmented glabella. Also Barrande's first stage of Sao (Barr., fig. 1 pl. 7).

As a transition to the next phase forms may be imagined having head and the beginning pygidium, nearly as Agnostus atavus, minus the thorax. The corresponding larval stages are Liostracus (Brögger's figures IV, V, VI) and Sao (Barrande, pl. 7 fig. 2-4 a, b). Nearly so, though the pygidium is more developed, are the larva of Agn. bibullatus and Agn. nudus (Barr., pl. 49) both without thorax.

II.

These have the three chief parts of the body developed.

3. The facial ridge. Cambrian, adult with glabella and facial ridges, short or long, emanating from the top of the glabella, thorax and pygidium. Several American Conocoryphæ. Corresponding larval stage, Sao, stage 3, Barr. pl. 7 figs 4 d-9.

4. Suture. A suture dividing the fixed cheeks in two pair viz. two fixed cheeks and two free cheeks. Fully developed facial ridge. Oldest known adult forms are Ellipsocephalus, Arionellus, already in the Lowest Cambrian, the Olenellus beds, what presupposes a long, antecedent lineage far back in the precambrian times. Corresponding larval stage in Sao, Barr. pl. 7 figs 10-13. The plurality of Cambrian trilobites belong here. An intermediate stage leading to the next is seen in such forms in which the ridge posteriorly is widened into the »eye lobe», which rests in the scallop of the free cheek. So it is in Liostracus and many others besides. Solenopleura possibly oculate.

III.

5. Globular eyes. Cambrian, in the youngest zone, the Olenus schists. The oldest at present known oculate trilobite Eurycare is found in the second division of these schists.

In the lowest Lower Silurian division, the Ceratopyge limestone, Euloma and Ceratopyge occur as the last survivors of the blind, partially ridge bearing genera. The Trinucleidæ and Ampyx belong to another group of trilobites.

Even among other exclusively Lower Silurian genera, in which the plurality of the species is oculate, there are species entirely blind. So with Illænus, in which genus Dr Holm has not found any eye in Ill. Angelini, I. leptopleura and Ill. cæcus. The free cheek in these three species is much narrow, as the facial suture lies near the margin of the head.

The eyes of the Trilobites.

If, as is probable, to judge by the conformity of their cornea with that of recent crustacea, the trilobites like these were provided with crystalline cones beneath the corneal lenses or facets, only the latter have been preserved in a fossil state. Although the crystalline cones in consequence of their solid consistence might have been petrified as well as the cornea, they must, imbedded as these tiny cones lie, entirely wrapped up in delicate tissues, fall away and be lost, when the dissolution of the dead body had set in. Consequently the curious appendages on the inferior side of the lenses in Dalmanites vulgaris ([Pl. III f. 50]) or Phacops quadrilineata ([Pl. V fig. 38]) can noways be considered as belonging to the original structure of the eye, apart from their great dissimilarity with anything appertaining to the eyes of the Arthropoda. The cornea on the contrary cohered with the integument of the body, and it has been well preserved in a great number of trilobites.

I subdivide the trilobites in respect to the form of their eyes in the following manner:

I. Genera with compound eyes.

1. With prismatic plano-convex cornea facets.

Acidaspis, as a transitional form to the next group.
Asaphus.
Bumastus.
Cyphaspis.
Dysplanus.
Encrinurus.
Illænus.
Megalaspis.
Nileus.
Phillipsia.
Niobe.
Proetus.
Ptychopyge.
Symphysurus.

2. With round or biconvex transversally elongate lenses.

Acerocare.
Bronteus.
Chirurus.
Ctenopyge.
Cyrtometopus.
Eurycare.
Peltura.
Sphærophthalmus.

II. Genera with aggregate eyes of biconvex lenses.

Acaste.
Chasmops.
Dalmanites.
Phacops.

III. Genera with isolated eyes, one or several stemmata at
the extremity of a straight facial ridge.

Harpes.
Harpides.
(?) Trinucleus in the larval state.

Only a few authors have before now occupied themselves with the intimate structure of the trilobite eye. Packard gave in 1880, in the »American Naturalist» a note on the structure of the eye of trilobites (p. 503). There are some rough and inexact sketches of the eyes of Limulus and Asaphus, and although he seems to have known the beautiful researches of Grenacher he still »claims that the trilobite eye was organized on the same plan as Limulus». This statement is altogether wrong, and as I hope to show the trilobites have had eyes entirely different from that of Limulus and instead agreeing with those of the Isopoda and perhaps also with a few other Crustacea. In 1889 J. M. Clarke published an account[28] on the »Structure and development of the visual area in the trilobite Phacops rana Green». The aggregate eye described by him are of the type forming my third group. His holochroal division embraces my first and second groups and the schizochroal my third.

[28] Quart. Journ. of Morphology, vol. II p. 253.

The latest contribution to the knowledge of these eyes is found in Exner's »Physiologie der facettirten Augen von Krebsen and Insecten», 1891, where he gives good figures of the lenses of Phacops fecundus, pl. II figs 18, 19. He says that the palpable difference in the structure of these eyes and those of Limulus point to a change in the function of these eyes.

I. Compound eyes.

1. Eyes with prismatic, plano-convex lenses.

A pellucid, smooth and glossy integument, a direct continuation of the common test of the body covers the corneal lenses, quite as is the case in so many of the recent crustacea. In the plurality it is, however, difficult to discern the lenses from the outside.

The lenses, as seen in a vertical section of the eye of Asaphus expansus, ([pl. I fig. 12]), are columnar prisms, like the pillars of basalt, attaining a length of 0,2 mm and at the point where the eye joins the test of 0,3 mm. At their interior extremity they have a breadth of 0,066 mm. On that point the surface is convex and at the exterior surface plane. They are closely packed and in a transverse section resemble a pavement of regular hexaeders. But they also assume other shapes and become rhombs or even quadrates, as seen in a specimen of Asaphus fallax ([pl. I fig. 18]), where the hexaeders and quadrates lie side by side without transitional forms. As a rule the lenses become more and more irregular in the vicinity of the surrounding frame or near the suture, nearly blotted out, as it were, and without any definite border line mingled with the confused, spongy mass that like a belt or a frame surrounds they eye in Asaphus and is sharply limited from the other part of the free cheek. This remarkable zone which is almost only present amongst the Asaphidæ (Asaphus, Megalaspis, Ptychopyge, Isotelus) retains in a confused manner somewhat of the prismatic structure of the eye as shown in the section ([Pl. I, fig. 11, b]). The eye of Bumastus also is environed by a similar zone, with a structure like that of the eye ([Pl. II fig 35, 41]).

In an undetermined species of Asaphus the lenses, although somewhat apart, are of an elongated hexaedric outline, which passes into a regular circular one farther away and on the surface of the eye they are slightly convex ([Pl. I figs 27-29]). In other genera belonging to this group the shape of the lenses are like those of Asaphus, so for instance in Illænus (I. chiron and I. Esmarki) and in Niobe. In Dysplanus centrotus they are shorter and broad, and their interior or lower surface strongly convex. It is likewise so in Nileus, where Nileus armadillo has an exceedingly thick exterior integument above the lenses. Such an integument has in a still higher degree increased in Bumastus sulcatus so as to exceed in thickness the stratum of the corneal prisms and it may in fact be doubted if the eyes of this species ever were able to function as visual organs. Proetus nearly resembles Bumastus in the thickness of the integument covering the prismatic lenses, which are interiorly convex, with a diameter of 0,03 mm.

In all genera belonging to this group a horizontal section gives the image of the hexaeders as in Asaphus with some change to squares or rhombs.

In scrutinizing a horizontal, somewhat extensive section of an eye in this group of trilobites, it will be perceived, as for instance in the figures ([Pl. V fig. 16, 22]) that the regular and evidently homogenous and intact prismatic lenses by and by have been altered and in a part of the section, a little distant from the intact ones disintegrated in their interior, showing various aspects of alteration. I cannot but think that this is a destruction which has set in long after the fossilization. It has revealed certain states of the intimate structure, certain delicate details, that now with an astonishing regularity come in sight and probably also lie hidden in the intact prisms. In the specimens of Asaphus, which we have studied, the alteration has taken the shape of a concentric stratification forming the body of the prisms, which is well discernible in a horizontal section, but not easy to detect in the longitudinal one ([Pl. I figs 9-10, 11]). It is likewise so in Niobe. In the other genera again the decomposition makes the prisms look like empty tubes in which a few irregular traverses and trabecular remains of their solid mass radiate towards the interior. They thus assume the aspect of a composite coral with its septa in the calicles ([Pl. VI fig. 31]). This is also evident in Nileus palpebrosus and Dysplanus.

2. Eyes with biconvex lenses.

The surface of the eye is, as in Chirurus glaber Ang., a mass of contiguous hemispherical lenses, probably once covered with a membrane, as is still to be seen in well preserved specimens of Bronteus laticauda. Both in Chirurus and Bronteus the lenses seen vertically are globular and ordinated beside each other either continuous or separated only through a faint dividing line. In a horizontal section passing right through the point of contact they show the common hexaedral shape and when somewhat corroded the interior radiate structure also comes forth, the radii directed towards a little black point in the centre. The lenses of the Brontei have the same stellate structure as in Bumastus ([Pl. II fig. 7]). In Cyrtometopus the lenses are in size the fourth of those in Chirurus and they form an extremely thin stratum in strongest contrast with the adjoining cheek, which surpasses them more than six times in thickness ([Pl. III fig. 19]). The lenses of Cyrtometopus are more flattened and irregular than in the former genera. The free cheek around the eyes does not form a border zone, somewhat imitating the eye structure as in Asaphus, but is more compact, composed of vertical elements which give to the test of the trilobites in general a tendency to split up in vertical prisms.

Of a peculiar interest are the eyes in the oldest of all oculate trilobites, at present with certainty known, Eurycare, Peltura, Sphærophthalmus and Ctenopyge.

Of these genera Eurycare is the oldest (see table [p. 22]). Amongst the many free and detached cheeks only a single, very little one has been found with the eye ball fixed. It seems to be of the same structure as in Sphærophthalmus. In Sphærophthalmus and Ctenopyge the eye globes are enormous, considering the size of the cheek in which they are set and occupy more than a third of the length of the free cheek ([Pl. III f. 26, 31]). They are hemispheric, blackish and glossy, more so in the former genus. The spheroidal lenses, projecting on the surface, are in Ctenopyge larger near the facial suture and small at the opposite side where the eye is fixed in the free cheek. For the rest, in both genera ([Pl. III fig. 34]) the lenses form a thin stratum, where they in a vertical section lie elongated, flattened and biconvex, slightly joined with each other at the point of contact. The fine form which they exhibit reminds of the lenses of Sphæroma.[29] They are in diameter thrice as long as they are high. Seen in a horizontal section passing through the point of contact they show hexaeders with a curiously jagged outline ([Pl. III fig. 33]).

[29] Bellonci Atti dei Lincei. Memorie, vol. X, 1881, Sphæroma, pl. II fig. 11.

Peltura which is coeval with these, has a narrow semiglobose visual field ([Pl. III figs 35-41]), the superior surface of which is quite smooth and evenly rounded. On its interior side there stand out, somewhat distantiated, in a low relief semiglobular facets, quite as regular incrassations of the cornea, thus not forming free lenses, but rather reminding of the for the rest differently formed quasi-lenses of Limulus. In a vertical section they appear as the inferior moiety of real ovate lenses ([Pl. III f. 40-41]).

The much younger Acerocare has a similar cornea. A very little specimen, the head of scarcely more than one millim. in length, retains both eyes, of which one shows the slightly convex lenses and the other a cast of the interior side as in Peltura. These both genera should in consequence of their peculiar limuloid cornea be ranged for themselves apart from the real lenticulate genera, but any material sufficient for doing this properly, is at present not at hand.

II. Aggregate eyes.

These are found solely in the family of the Phacopidæ, unless the Lichadidæ were also provided with this sort of eyes, but we have had no opportunity to study them. It seems, however, not likely that they had aggregate eyes. Barrande has represented them quite as finely reticulate as the eves of any Asaphid. We have sectioned and figured the eyes of Dalmanites vulgaris and D. obtusus from the Silurian of Gotland and found that these have truly aggregate eyes, each consisting of a regular biconvex lens, lying enclosed in a socket of its own and covered by a cornea of its own. The distance between the eyes is much variable and in a few instances they are nearly contiguous. Extremes are seen on [pl. III figures 43, 47]. The lenses are comparatively large, and have always had a covering membrane, though this in many instances has been lost. This membrane which is an immediate continuation of the general integument of the body covers the lenses all round their superior moiety. In its prolongation downwards between the lenses ([Pl. VI fig. 3, 4]) it is free from the contact with them and hangs alongside and around much incrassated, so as to take in a section a lengthened lancet like shape. It lies thus alongside the other interstitial test, and is like this perforated by longitudinal canals. In a horizontal section taken a little below the surface it encircles the lens as a wall like ring ([Pl. VI fig. 1, 2]). In a vertical section the lenses lie in direct contact with the cheek without any intervening zone and the cheek has the structure so common amongst the trilobites, being perforated by vertical tubes going straight down from the surface ([Pl. III fig. 44], [Pl. VI f. 5]).

In Dalm. vulgaris and also in Phacops quadrilineata there is as already before mentioned a peculiar structure beneath the lenses, consisting of narrow, threadlike, straight lines, twice as long as the lenses ([Pl. III 49-50], [pl. V fig. 38]). In a horizontal section they are found to be irregular prisms closely packed. It can not be any structure peculiar to the eyes or the lenses, rather some parasitic growth added since the death of the animal. The lenses are in several specimens composed of clear calcareous spar. In others again they have been filled with a dark muddy calcareous rock excepting in the lower moiety where there is left a residue of the white spar, having in all lenses assumed a regular shape which I consider as organic ([Pl. VI fig. 5]). This spar covers the whole bottom and its upper rim is incrassated and bent inwards. In horizontal sections this residue is a whitish ring close inside the interior ring wall ([Pl. VI f. 2]). I would suggest that this curious conformation is due to the original structure of the lens, supposing that it in these crustaceans has been built upon the same plan as in several other Arthropoda. In Cymothoa[30] and in Sphæroma[31] for instance the lenses are built up of thin strata, which are parallel with the convex outside, so that on the inferior surface of the lens they are arched downwards and on the superior side upwards, being not strictly concentric. In the spiders they are constructed upon this same plan[32] perhaps more evidently. If now in Phacops the lens consisted of such semiconcentric strata and the upper moiety has been destroyed, the rest must have taken the shape as we find it. It is moreover peculiar that the destruction has been exactly similar in all lenses of that specimen. Can it be due to the circumstance that the power of resistance in the inferior strata has been greater?

[30] Bullar Philos. Transact. 1878, pt. II, pl. 46, fig. 12.

[31] Bellonci Atti dei Lincei, Memorie X 1881, pl. II, fig. 11.

[32] See Grenacher pl. II f. 18 Epeira.

In Phacops quadrilineata the lenses are more elliptic than in the former. On their interior surface beneath the spiny tufts mentioned large hexaedral prisms of clear calcareous spar issue, one prism for each lens ([Pl. V fig. 38]), having thus a very deceptive appearance, but no doubt of inorganic origin with the lenses as a basis for their crystallization, quite as in the Cystoids where the interior often is converted into a mass of crystalline prisms, issuing from the interior surface of the plates.

III. Genera with stemmata and ocelli.

In a little group that has retained larval or ancestral characters during a great part of the palæozoic period, the genus Harpes stands as a type. It has ranged from the oldest Lower Silurian, if we join the related Harpides, to the middle Devonian. From near the top of the glabella, though not so much forward as the facial ridge of the blind trilobites (Olenus, Liostracus etc.) a straight ridge of much varying length stands out on both sides and at its extremity two or three globular stemmata with glossy surface lie encased. Probably this ridge has the same origin as in the Olenidæ, the more so, as there are indications of an extensive circulatory system. On [plate IV fig. 18-19] the right hand ocelli of Harpes vittatus Barr. from Lochkow, Bohemia, are represented. They are two, lying isolated near each other, quite globular with circular outline, smooth and glossy as to exhibit a shining surface. Their size is 0,4 mm in diameter. Being cut vertically in the direction of the longitudinal axis of the head they resemble elongated hemispheres, convex on the exterior surface, slightly concave on the interior. The test of the head lies between them as a saddle and covers them only partially and on the outer sides they lie with their margins encased in the head shield. Seen in thin sections of the right lens and magnified the whitish mass is traversed by vertical, blacker streaks, standing somewhat radiating towards the sides and cancellate. The other lens has a horizontal row of black dots. All this is probably not of any structural value, only due to later changes.

The remarkable genus Harpides from the lowest Lower Silurian belongs also to this group and has beside the fixed peduncle a peculiar elongated ridge going from the eyes to the lateral margins of the head[33], a ridge which is also present in some of the true Harpes.

[33] Harpides breviceps cannot belong to this genus and is rather related to Erinnys Salter, as also Matthew holds it.

As Beecher has shown[34] the larva of Trinucleus possesses quite the same transverse ridge with intumescent eyelike extremities, and although the smallness of the specimens has not permitted to ascertain the presence of a true eye, it may be apposite to suppose it on the homology with the eyes of Harpes, a genus with which Trinucleus is related. But as well known, in the adult Trinuclei there is no trace of these ocular ridges nor of real eyes so that Beecher in his paper »Blind Trilobites»[35] numerates Trinucleus amongst these. Out of the nine Scandinavian species of Trinucleus no less than seven have a well marked little tubercle on each side of the glabella placed exactly on the same spot where the larval Trinuclei had their much larger eyes placed. In this case it may be allowed to suppose that the tubercles are the direct successors of the larval eye and that they are true ocelli. Reede seems to be willing to regard them as possessing a visual function.[36] Beecher[37] holds the eye nodules in the larva and the ocelli in the adult to be identical.

[34] Structure and appendages of Trinucleus.

[35] Geol. Magazine 1898. pp. 439, 493, 552.

[36] l. c. p. 447.

[37] l. c. p. 309.

But, as Barrande has shown,[38] there is a certain species of Trinucleus, the larva of which wants a facial ridge and eyes, as there also are several adult forms without ocelli. These have remained on a much ancestral stage, while the larva with eyes are more highly developed in such species, where the adult have been subject to a retrograde development. Ampyx and Dionide, though completely blind, evidently belong to this group, and once, as is to be hoped, larval forms may be discovered showing their development. In a certain way Arethusina shows characters proper for this group, in having the straight ocular ridge, quite as in Harpes, but eyes of the reticulate type and probably prismatic. It thus like Harpes conserved an ancestral characteristic long periods since it had disappeared in most of the other genera.

[38] Sys. Sil. de Bohéme I, pl. 30, figs 41-50.

Of the groups, in which Joh:s Müller[39] long ago classified the Crustacean eyes, his second »Hauptgruppe» (»Aggregate von einfachen linsenhaften Augen») and the fourth »zusammengesetzten Augen ... facettirte Hornhaut» the former corresponds with my third and the latter with my two first divisions. In so far as the cornea and its facets or lenses are to be regarded, there is the greatest analogy with the Isopoda. In vertical sections of Sphæroma we have the same sort of elongated, flattened biconvex lenses as in Sphærophthalmus and others. Since Grenacher and Exner and others have published their excellent works on the eyes of the Arthropoda, there can be no foundation for speaking of the resemblance of the trilobitic eye with that of Limulus, as this genus stands completely isolated amongst all Arthropoda in that respect. There is, as stated above, a certain resemblance between the cornea of Peltura and that of Limulus, but this is not yet ripe for a discussion. Nor is there any evidence for correlating the eyes of the trilobites with the eyes of the Phyllopoda. Bernard thinks that the so called eye of the Paradoxidæ has been formed upon the same plan as that of Apus. There is nothing to prove this hypothesis that the facial ridge or any part of it ever had been a visual organ, and the evidence at hand rather tends in a contrary direction.

[39] In Merkels Archiv 1829 p. 46 and in Treviranus Zeitschrift für Physiologie Bd IV p. 97.

There are signs of long physiological and anatomical efforts to prepare the development of the eyes on the free cheek, as revealed through the long series of blind trilobites. A system of radiating blood vessels, similar to those described above as covering the inside of the head in some older genera, all issuing from the scallop in the free cheek, where later the eye had to find its place, have left their stamp, their mark on the surface of the free cheek. They attest the great vital activity which was so intense at the point were the eye was to be formed. We give the figures of two such cheeks of different types. One from Parabolina spinulosa ([pl. V fig. 31]) is the more common, where six or more isolated trunks radiate from the semilunar ridge round the indenture and subdivide in branchlets which cease near the lateral margin of the cheek. It may be that it is an annular vessel near the indenture that feeds them all and that this probably is in connection with the great central circulatory system. In Olenus ([pl. V fig. 29]) the vessels are partly anastomosing and form a reticulate system and they are studded with minute wartlets. Another sign, which may be taken as a preparation, is the elevated rim around the scallop, which is so prominent in several of the Cambrian genera, but which does not embrace any facet bearing cornea..

From what has been stated above the following conclusions have been arrived at.

1. The plurality of the genera living during the Cambrian period were blind and it was first at the close of that period, in the Olenus schists, that genera with real visual organs appeared. There may have been oculate trilobites earlier, as Solenopleura, but we know nothing of their eyes.

2. The primordial glabellar pleuron which was metamorphosed into a facial ridge is no visual organ. It is in the Olenidæ nothing but the elevated line made in the test by the subjacent main trunk of the circulatory system. It swells out in a node, »palpebral lobe», but not before the facial suture has been formed. In the genera where there is no facial suture, there is no node. In the Paradoxidæ where the ridge is of a different origin, there is no node, though there is a suture.

The four types of eyes in the trilobites have probably succeeded one another in the following chronological order:

1) with stemmata or ocelli; 2) biconvex or lentiform; 3) prismatic; 4) aggregate. The oldest known representatives for each type are for 1) Harpides rugosus in the Ceratopyge limestone of the Lower Silur., for 2) Eurycare, in the Cambrian Olenid schists, division 2, for 3) Megalaspis, in the Ceratopyge limestone of the Lower Silurian, for 4) Phacops in the Lower gray Orthoceratite limestone. The eyes of the trilobites show the greatest conformity with those of the recent Isopoda.

The most perfect eyes amongst all the trilobitic eyes may be those of the Phacopidæ, which are also geologically the youngest, the least developed again those of the Proetidæ or rather of the Bumasti. The great thickness of the cornea in these must have weakened their power of vision and they had probably only a faint perception of light.

On the maculæ of the hypostoma.

We shall now turn our attention to the visual organs which Liljevall discovered on the hypostoma of Bronteus. In doing this, we may bear in mind, that the genera in which we really have found lenses on the maculæ are relatively few, but that we shall review the maculæ in all genera, which we have been able to examine, and try to show that even most of these may, although in an inferior degree, have acted as visual organs. At the same time some more details shall be given about the cephalic eyes for comparing them with the hypostomic ones.

Acidaspis Murch.

The hypostoma is of a peculiar type, deviating from that of the other genera, squarish or rectangularly transverse, entirely without terrace lines and no maculæ proper. Barrande has given no less than nine different samples of these hypostomas.

Acidaspis crenata Emmr.

([Pl. I figs 1-6.])

There are certainly no maculæ of the same sort as in so many other trilobites and the interior side of the hypostoma does not bear the least traces of impressions which might be taken as the reverse of the maculæ and still more less as muscular impressions. In fact, the total absence of such in Acidaspis militates against the interpretation of the macula impressions as muscular scars. When the maculæ fail, also the so called muscular impressions fail. Moreover we are able to see still clearer in this matter through what we know about the structure and position of the extremely similar hypostoma of Apus, of which we have given a description in the end of this memoir. This skeletal part of Apus is without any connection with the surroundings, excepting at its anterior margin, and is consequently movable in a direction outwards and upwards. Along that anterior margin it is fixed to the outside of the ventricle through three pair of muscles, three muscles on each side of the margin. Zaddach De Apodis cancriformis anatome (pl. II fig. XIV p. 68).

For the rest the marks of the attachment of the muscles are as a rule in the Crustacea, at least those of the head, elevated small platforms, so in the Trilobites, of which I have excellent specimens in a Bumastus and others. On the inside of the head of Limulus they are faintly elevated patches. The small hollows on the inside of the hypostoma formed by the maculæ are indeed the sockets in which the soft parts of these more or less developed hypostomic eyes were sheltered. But there is still a feature in the hypostoma of Acidaspis which merits our attention and which perhaps may have a significance akin to that of the maculæ.

The hypostoma is square, with two short pointed wings, one on each side of the slightly bent anterior margin and likewise two smaller ones at the corners of the posterior margin. A groove follows on a short distance the lateral margins and the posterior margin and disappears a little below the anterior one. In the same direction, distally, though a little more inside and unconnected with them there are two small grooves, the bottom of which consists of a shell substance of different colour and structure than the other parts. Having been a little ground and seen in transmitted light it exhibits the shape of a club and a homogenous yellow spot, tapering posteriorly and swelling out distally ([Pl. I fig. 4]). It must be left an open question whether these maculæ share in the nature of visual organs as the quite different maculæ of the other trilobites, but it may be possible that it is so. It must, however, be remarked that there are two types of hypostoma in the genus Acidaspis as shown by the illustrations of Barrande. One has the small grooves, possibly all sheltering the claviform maculæ, disposed as in the now described A. crenata. This group embraces five species of the Bohemian Silurian formation. The other group of three species again has the hypostoma of the same quadratic or rectangular shape, but the two short grooves, which may be expected to contain the maculæ, are placed midways between the anterior and the posterior margins, nearly as the maculæ bearing grooves of other trilobites. We have however not had material for pursuing our researches in this genus, the other species of the Swedish Acidaspidæ being unknown as to their hypostoma.

It may here be added an observation concerning the ornamentation of the exterior surface of the hypostoma of Ac. crenata. It is covered by a great number of diminutive circular or oblong wartlets occupying the whole surface excepting the lower third of the central field just above the posterior groove which is smooth. These wartlets seen through transmitted light ([Pl. I fig. 5]) show in their interior something like a peculiar black spiculum rising from a bifid rootlet and confined within the wartlet and not extruding from it. In a longitudinal section ([Pl. I fig. 6]) the spicula perforate the wartlets reaching through their whole length. As in Calymmene these interior pseudo-spicula are tubes, filled with iron-pyrites. It is probable that these tubes were once bearing setæ and quite as in Apus formed a fur of bristles.

The structure of the cephalic eyes ([Pl. 1 f. 1-2]) is prismatic, but the separate prisms are rather short and broad. Their lower or interior end is convex. The separating lines between the single prisms are not always distinct.

Agnostus Dalm.

[Pl. 1 fig. 7.]

Agnostus glandiformis Ang. Although there is not the slightest evidence of eyes in this the largest of its genus, nor any free hypostoma hitherto has been found, we may here give a little account of our researches into this species. The scantiness of material forms a chief obstacle to our knowledge. Only three entire, rolled up specimens have been found, and it is by sectioning and preparing such that any hope can be entertained to gain reliable results.

A rolled up specimen from Andrarum Scania was sectioned lengthwise. The tail-piece closed tight against the head-shield, so there had been little chance for foreign matter to penetrate into the interior which, however, is filled. Close below the cephalic shield there is a remarkable structure, mostly resembling an elongated intestine with swellings joined by more narrow ducts and anteriorly the coherence is interrupted. As in crustaceans and Arthropods in general the stomach and the intestine are situated on the dorsal side of the body, there is nothing unlikely in assuming that this in reality may be the remnants of the intestine. This may also be compared with the observation made by Vollborth in his memoir »Ueber die mit glatten Rumpfgliedern versehenen Trilobiten», 1863, p. 46, tab. 1 fig. 12 where the heart-tube probably is delineated. Barrande has also given figures of what he considers as the intestine in Trinucleus.

Below this organ in Agnostus, there lies a section of a vaulted calcareous plate with its convexity turned against the dorsal side of the head shield, that is to say quite the reverse what might have been expected if it had been the hypostoma in its true position. This may, however, not be any objection against considering it as a sectioned hypostoma, loosened from its connection with the cephalic shield, disturbed in its original position and turned round, when the shell became filled with mud. At the distal end there is a much distinct duplicature.

Asaphus Brongn.

The wellknown hypostoma of this genus has the same characteristic shape, though more pronounced, as in Ptychopyge, with its posterior margin deeply indented, so as to form two large, pointed lobes. There is properly only one median field, surrounded by flat, lateral borders, continuing down into the posterior lobes, from which it is separated through a shallow groove. The macula are situated in this groove, on each side of the inferior border of the central field. They are more or less prominent, but whatever their form may be, their surface is always entirely smooth, lying well circumscribed amidst the surrounding terrace lines.

The following list enumerates all the species in this genus, of which previous authors have delineated the tubercles, though they in the descriptions only in very few instances have mentioned their presence.

A. acuminatus Nieczkowski, »Zusätze zur Monogr. der Trilobiten 1859», tab. I, f. 6. Maculæ most prominent, but no mention made of them in the description.

A. (Isotelus) canalis J. Hall, Pal. N. Y. I, pl. 4 bis, f. 18-19. Though this and a following species probably on account of the deviating conformation of their body, belong to a different genus or subgenus, I mention them along with the Asaphi, as there is the greatest similarity in their hypostoma. Fragment of an interior cast; still more strange is the fragment showing part of the interior surface and the duplicature.

A. expansus L. Br. I pl. VII f. 3, BR. II pl. I f. 2, 2 c. In the two first figures there are no tubercles nor maculæ marked, in the last figure there are distinct tubercles.

A. fallax A. pl. XXVIII f. 3, c. BR. II pl. I, f. 3, both nearly congruent.

A. (Isotelus) gigas J. Hall, Pal. N. Y. 1, pl. 60, f. 7 g, pl. 66 f. 5, the inside of an entire, uncommonly large hypostoma gives a general good view with the macula; which are large and evident. He mentions the maculæ as »two circular spots». »These probably indicate the points for the attachment of muscles and tendons upon the inside,» he adds.

A. ingens Barr. Novák II, pl. I, f. 7, a good figure showing two semiglobular maculæ. Barrande's figures pl. 33, f. 7, 8 are not distinct.

A. ludibundus Tqt. Br. II, pl. 1, f. 7. Two semilunar sulci on the cast of the interior surface, being only the posterior borders of the maculæ which are a little more raised and distinct than the rest.

A. nobilis Barr. pl. 32, f. 6. Indistinct traces of maculæ. In the figure 6, pl. 31 representing the hypostoma of a young specimen, there are no tubercles at all. If the figures were to be relied upon, it might be assumed, that the tubercles appear at a more mature age.

A. Powisii Salter, pl. 23, f. 6. Two semilunar, narrow tubercular maculæ, their interior apices converging towards the anterior border of the hypostoma.

A. raniceps A., pl. XXVIII, f. 2 c., BR. II, pl. 1, f. 4. A good figure.

A. raniceps var. maxima Br. II, pl. 1, f. 6. Semilunar impressions on a cast and thus far incomplete.

A. striatus Boeck. To this belongs probably A. expansus M. Sars »Ueber einige neue oder unvollständig bekannte Trilobiten» in Isis 1835 p. 333, and especially p. 340 and the following (Bemerkungen über die untere Seite von einigen Trilobiten). Sars there gives a very good description of the exterior side, where he also mentions the two maculæ, pl. IX, f. 9 a, b. (»2 kleine Knoten»), Br. I, pl. VIII, f. 4 a, and Br. II, pl. I, f. 9. These latter figures differ in so far, that the former has the maculæ excavated or rather tubular below, in the latter again they are regular.

A. trinucleorum Br. II, pl. I, f. 16 a large specimen with linear maculæ arranged rectangularly in respect to the longitudinal axis of the hypostoma and both on the same level.

A. tyrannus Murch. Salter in Mem. Geol. Survey, Brit. Foss. Dec. II, 1849 pl. V, f. 4, excellent figure. On p. 2 he says: »there is an oval circumscribed tubercle at the origin of each (fork) most distinct on the inner surface». This figure is again reproduced in the »Monograph», pl. 22, f. 6, where it is said »Two linear tubercles with their interior apices converging towards the posterior margin of the hypostoma».

Asaphus sp., probably A. raniceps, Pompeckj, Trilobitenfauna Ost- and West-Preussens, taf. VI, f. 7. p. 80 »an seinem Hinterrande liegen zwei kleine, flache Höckerchen, die als Reste des Hinterlappens aufzufassen sind»(!?).

We shall now describe the Asaphi which we have been able to examine more in detail.

Asaphus expansus L.

[Pl. I figs. 8-17.]

In the numerous specimens of which we have examined the hypostoma, there are always small, nearly circular maculæ which by their lighter colour are marked out from the surrounding smooth space of the inferior sinus of the lateral grooves where they are situated. They do not rise at all above the surrounding surface from which they are separated only by a fine, scarcely perceptible line. On the interior surface of the hypostoma they are better seen, and the enclosing line is deeper and more distinct. Their surface is there quite as smooth and even as on the exterior side of the hypostoma. They do not consequently in the least manner resemble muscular impressions as observable on the inside of the glabella of other trilobites. These are on the contrary elevated above the surface and finely striated in various ways.

The peculiar nature of these macula is revealed through the various sections we have made. In [pl. I fig. 15] a vertical section across the entire hypostoma is represented. The sectioned, dark maculæ (a) lie on both sides of the faintly curved central field, in the sinus of the lateral grooves, and the strongly developed terrace lines continue sidewards just a little on the interior side, where the duplicature turns round. As seen in a magnified vertical section ([fig. 16]) the maculæ consist of horizontal, whitish, straight lines, probably lines of successive growth, and these are crossed rectangularly by more irregular whitish lines separated from each other through dark spaces. The whole thus gains the aspect of a dark surface cancellated by white lines. This reticulated or spongious macula is enclosed as to its superior as well as to its inferior part in the compact and homogenous test of the hypostoma and occluded from the influence of the light. Its value as a visual organ consequently is insignificant. In a horizontal section ([fig. 17]) the structure is still more bewildering. There the whole macula is a confused, spongious white mass with dark spaces between the white meshes, and only at one side, the left one, some obscure indications as of polygones are visible. It would indeed have been impossible to interprete what this means, had not the study of the cephalic eye given a clue thereof.

The eye of this species is covered with a delicate perfectly smooth and glossy integument ([f. 12]) which is reposing immediately on the prismatic lenses of the cornea. In a few instances, depending on the colour or the state of preservation of that integument the lenses are indistinctly translucent. Owing to their state of preservation their aspect is greatly variable. They are in many instances, as seen in fig. 9, six-sided or some rhombic, foursided or even quadratic ([f. 8, 10]). They are all of the same length, 0,2 mm. along the surface proper of the eye, but are lengthened to 0,3 mm. towards the border of the eyes, [f. 12], which will be described further on; they are of equal breadth, amounting to 0,066 mm. Their inferior ends are slightly convex or nearly plane when well preserved, else, when as often is the case, disintegrated as to be scooped out and vaulted. Their interior structure, as revealed by sections, shows a cylindrical core, [f. 8-10], composed of concentric strata. This cylinder fills nearly the whole interior space of the prisms, there being, however, in many instances a compact dark mass between the cylinder and the walls of the prisms. There are also sections in which the prisms are filled with a uniformly black mass without any concentric structure. In longitudinal sections the outlines of the individual prisms are not clearly discernible ([f. 11]). There are longitudinal, white lines of varying thickness with lateral irregular offshoots, which may join with those from opposite walls, and give the interior a sort of spongious or cancellate appearance.

Towards both sides of the ocular surface, towards the superior and inferior side, a change sets in as to the shape of the lenses, as best seen in horizontal sections ([figs. 8, 10]) they are lengthened and become more and more indistinct, and at last in the upper and lower marginal zones pass over into a reticulate, spongious mass, which seen in a longitudinal section presents almost the same aspect as in the regular prismatic surface of the eye. [Fig. 11 b], [a] being the visual field. It is, however, more densely reticulate, but a prismatic arrangement is quite as much evident as in the ocular surface proper. In the rule the passage from the prismatic surface to the reticulate is gradual, the prisms becoming by and by irregular in their outline and diminishing in size ([fig. 12, a] the eye proper, [b] the border zone), but there are also instances where the distinction between these two fields is sharp and without any gradual transition. In the inferior reticulate zone there are generally some oblongue, funnelshaped pits. I am uncertain whether they are to be regarded as regular parts of the eye structure or rather as burrows of some parasite. They do not continue deep down.

If we now compare the reticulate zone of the eye with the maculæ of the hypostoma, for instance the vertical section [fig. 16] with [fig. 11 b], we find the most complete identity in structure. In the same manner the horizontal section of the border zone of the eye, [fig. 8 a], [fig. 10 b], and of the macula, [fig. 17] are similar. The same chaotic, spongious mass in both, with some tendency to form prisms more evident in the longitudinal sections, where the same reticulate structure with predominant white, longitudinal streaks is so palpable. That there is a complete identity in structure between the two, the macula of the hypostoma and the border zones of the eye, is as evident as anything can be, but as to the functional identity or what this function may have been it is difficult to decide anything with certainty, at least it seems to me that the capacity of vision must in both have been far more restricted than in the eye proper. They rather give the impression as of rudimentary visual organs.

Asaphus raniceps Dalm.

[Pl. I f. 23-26.]

The maculæ of the hypostoma are placed obliquely in the saline manner as in the preceding species. They are more prominent and the oblong macula, with the longest diameter of 1,2 mm. is on the top of a little mound and surrounded by a fine, elevated marginal line ([fig. 23]). Its somewhat convex surface is entirely smooth and if sectioned horizontally exhibits the same sort of spongy texture as A. expansus ([pl. I f. 24]). In a vertical section ([fig. 25]) the macula does not occupy so large a space as in As. expansus, but rather lies as a lenticular disk in the hypostomic test closer to the superior surface than the inferior. With sufficiently high power the same sort of pillars, divided by horizontal strata is seen. Upon the whole the vertical section is not so clearly developed. The reticulate zone of the eye is more definitely separated front the prismatic zone than in A. expansus. There can be no doubt that there is a correspondence in this species between the structure of the macula! and the spongious zone of the eyes.

Asaphus cornigerus Schloth. (A. Kowalewskyi Lawrow).

Of this strange species with its enormous eyestalks, more than 2 centimeters in length I have through the kindness of Akademiker Friedr. Schmidt in St. Petersburg had occasion to study some specimens.

The rather large maculæ, obovate-circular, are like those of several other Esthonian species oriented inwards and upwards instead of inwards and downwards as in the Swedish species described. As the specimens have suffered through corrosion of the surface there are only faint traces of a marginal line. Their microscopic structure is badly preserved, and they look pale and transparent with only few indications of the spongious texture. The lenses of the cephalic eye are nearly square prisms sometimes with a slight approach to hexaedral pillars. There is no clear transition into a spongious or reticulate border zone. But this may depend upon the bad preservation.

Asaphus fallax Dalm.?

[Pl. I fig. 18-22.]

It is very difficult to distinguish this species or rather variety from A. expansus, but if we have found genuine specimens, there are some points in the shape of eyes and maculæ which make it different. The hypostoma ([f. 21]) is rather more broad and the duplicature is large with an upturned margin. The little macula is placed on the top of a smooth rounded elevation and enclosed by an elevated rim.

The cephalic eves are rather short, regular hexaedral prisms ([f. 18-20]) and as shown in [fig. 18] change into irregular squares near the border of the eve.

Asaphus sp.

from Brunsby kanal, Segerstad parish, isle of Öland. The maculæ are large pale whitish and of an uncommonly fine-meshed reticulation, well limited from the surrounding hypostoma.

Asaphus sp.

[Pl. I fig. 27-30.]

from the islet of Sandö, north of Gotland. Although we cannot give any account of its hypostoma, not having had sufficient material, the structure of the eyes is so peculiar that it seems worthy of being recorded. The integument is extremely thin and transparent and the subjacent lenses are clearly seen, and through their impact on the thin integument they make this to stand out in a very low relief above them ([fig. 27]). These lenses are uncommonly short, forming at the surface rather oblong, slightly hexaedral prisms with a narrow interspace between them. In a section lower down they have the shape of hollow, white rings filled with black mud and in a longitudinal section the white walls of the lenses look like short pointed spikes and interiorly they are completely empty. We here find also the same gradual change from regular cones to the spongious border zone as in the above mentioned species [f. 29]. The border zone is finely reticulated.

Asaphus (Isotelus) gigas J. Hall.

[Pl. II figs. 1-3.]

The enormous hypostoma of this giant resembles in a high degree that of the Asaphi, but is at the same time the most evident verification of the experience that the hypostoma per se cannot be regarded as the sole criterion for determinating the generic affinities of different species. The whole structure of this trilobite in other respects gives it a quite independent position, distinct, from Asaphus.

We have had at our disposal several specimens of the large hypostoma, the dimensions of the largest, fragmentary in its anterior margin, being as follows: breadth 48 mm. length 41 mm., probably 47 mm. when entire, breadth of each of the posterior lobes at their bases 19 mm., length of the same 23 mm. The maculæ which are placed on the flat surface of the hypostoma without being at all elevated, are prominent through their great size and their pale, whitish colour. They are somewhat oblongue having their longest axis directed inwards and downwards. They attain a diameter of four millimeters. Horizontally sectioned, [f. 2], they show the spongious texture and vertically the quasi prismatic reticulate structure richly developed, [f. 3]. The pillars are very distinct and continue without interruption all through the macula and in the interstices there are traverses joining between two pillars or partially filling the darkish interspaces, thus giving the whole the aspect of some »tabulate» coral sectioned. By the inspection of the horizontal section alone, presenting the irregular spongy texture it would have been impossible to imagine the ordinated arrangement which the vertical section reveals to us. The eyes consist of regular hexaedral prisms, and there is a very sharp boundary line between them and the spongious zone, which is very narrow and without distinct separating lines joins with the cephalic test. For the rest there are in all probability at least two different species sent from America under the name of Isotelus gigas, of which only the largest, almost like a Homalonotus, and the hypostoma of which has been partly delineated in this memoir, seems to be the real one. They differ both in the shape of the hypostoma as well as in other respects.

Barrandia Mac Coy.

According to Salter Monogr. pl. 19 fig. 9 his B. Portlocki, of which there is a fragmentary and broken hypostoma, shows feeble traces of oblique macula, nearly in the fashion of Asaphus. In the description at page 139 is told about »the usual pair of tubercles».

Bronteus Goldfuss.

Barrande delineates the maculæ on the hypostoma of

B. palifer, Tab. 45 fig. 17,

B. planus, Tab. 48 fig. 7, and

in the supplementar volume of

B. rhinoceros, Tab. 9 f. 16,

B. furcifer, Tab. 11 fig. 16, but there is not the slightest indication of granulation on any of them, nor is there in the descriptions, generic or specific, the least mention made of the tubercular maculæ. It is also remarkable that in the works of Angelin and Novák where several hypostomas belonging to species of this genus are delineated, not a single one shows these tubercles. We shall now continue the descriptions of the Swedish Brontei, already begun with Br. polyactin in the introductory part of this memoir.

Bronteus irradians Lindstr.

[Pl. II figs. 4-5.]

has a hypostoma that much resembles that of Br. polyactin. In its general shape it is similar and the two concentric grooves with the two maculæ placed in the same way, just below the superior groove. These maculæ are much larger than in Br. polyactin, nearly thrice their size. They are also more ovate or rather like a bean, the smooth surface is larger and the granulated spot restricted to a more narrow space forming an oblique patch. The granules or lenses are also individually larger than in the allied species, double their size or 0,06 millimeters. We have not succeeded in making sections of the cephalic eye nor of the maculæ.

Br. platyactin Angelin.

[Pl. II f. 14-19.]

The hypostoma has a transversally triangular form, and is divided only in two fields through a shallow semicircular groove near the posterior margin. The two maculæ are situated above the groove near its superior sinuses. They are elongated, [fig. 17], elliptic with the narrow pointed end directed outwards and the broad rounded end inwards. The chief surface is scooped out as a shallow depression. The granulated spot is situated on the broader end and covering it completely. The relatively large lenses are arranged in five regular rows, the uppermost one being the longest. On the interior surface of the hypostoma there are the corresponding sockets of both maculæ with smooth surface. The horizontal sections of the granules [figs. 18, 19] present the image of white rings in close contact, without, however, to occasion a prismatic structure, a dark interspace lying between each ring. These lenses are filled with a dark mass, and in some the same sort of radiated structure is perceptible as in the lenses of the cephalic eye. In horizontal sections of the cephalic eye the lenses approach the polyedral shape. In another section near to the surface of another specimen the lenses are decidedly hexaedral. The vertical sections, [fig. 14], reveal their real nature as lenses where they lie as a string of beads with a dark nucleus enclosed within a thin whitish shell. They are covered by a thin membranous lining. When seen in transmitted light the lenses proper are dark, and the shell white and in reflected light the lenses are lighter than the rest.

Br. laticauda Angel.

[Pl. II figs. 6-13.]

Lower Silurian from Dalecarlia. The hypostoma, [f. 10], is of a broad clypeate shape, the anterior margin rounded without any large projecting wings. On the exterior surface there are two grooves parallel with the rounded inferior margin. As in Br. polyactin the two elliptic tubercles are situated on the inferior edge or slope of the upper groove and they are deepened by a shallow depression as in Br. platyactin. The granulated spot situated along the posterior margin of the macula deviates much in shape from that of the other species. It consists of a long and narrow stripe ending in a fine point outwards and widening inwards, where it is rounded, forming thus a claviform, curved elevation. Around the granulated area the maculæ are quite smooth, and show in a horizontal section an irregular structure of tiny black dots nearly resembling the structure of the hypostomic shell. We have not been able to obtain any good vertical section, but by casts of the interior side of the tubercles it is found that the granules form polyedric facets like those of the eyes though perhaps not so regular. There is a specimen with the granules intact, and the polyedric shape is then not so distinct.

Sections of the eyes elucidate the structure, which partially is obscure in other species of Bronteus. There is a thin membranous coating covering the subjacent well formed lenses ([fig. 8]). When this membrane in some instances has been peeled off the lenses lie bare ([fig. 9]). These have a dark nucleus and in some instances it seems as if there were two. In a vertical section they look spheroidal, in a horizontal section again they are polyedric, especially when taken somewhat below the surface or near the middle line ([fig. 6]). In [fig. 7], some lenses are delineated in a horizontal section, highly magnified and the corallian appearance is evident. The lenses are well separated by distinct lines, and from their inner tubes whitish reticulations radiate towards the centre.

If we now compare these sections, where the lenses are so distinctly seen, with the more obscure sections of other Brontei, it is evident that the dark points in them are nothing but transformed and deteriorated centres of the lenses. In Br. platyactin they are a little more distinct than in the other species. And it is with these that the granulated spots on the maculæ of the hypostoma are most concordant.

Bronteus sp. indet.

[Pl. II figs. 28-30.]

This species, of which only the hypostoma is known, found at Lansa, Gotland, is likely, to judge by that, to be nearly related to Br. polyactin. The shape of this hypostoma is exactly the same and the maculæ are placed in the same position, but they are larger, a little sunk on the blind surface and the group of the lenses is different. Its superior margin forms an ingoing arch and the lenses themselves are not convex, but rounded or slightly polyedric, and separated through thick interspaces.

Bumastus Murchison.

The only recorded hypostoma with maculæ belongs to B. insignis Hall as described by Salter Monogr. p. 208 pl. 27 fig. 7. He says: »A pair of compressed tubercles occurs at the lower third: they are transverse-ovate, and more than their own diameter apart.» This is nearly in concordance with what is seen on the species to be described below.