IMAGE EVALUATION TEST TARGET (MT-3) 'V. / O {./ •i^ MS) y. ^ ^ i/i u. 1.0 I.I iWIIIM 125 S IIIIIM |||m i ill 112.0 1.8 1.25 1.4 1.6 .4 6" — ► V] #m ^. °m ■w 4:

:h9 upper left hand corner, left to right aiiu .op to bottom, as many frames as required. The folowing diagrams illustrate the method: Les cartes, planches, tableaux, etc., peuvent §tre filmds d des taux de reduction diffdrents. Lorsque le document est trop grand pour dtre reproduit en un seul clich6, il est film6 A partir de Tangle supdrieur gauche, de gauche d droite, et de haut en bas, en prenant le nombre d'images ndcessaire. Les diagrammes suivants illustrent la mdthode. 1 2 3 1 2 3 4 5 6 \_Fm ON f J. / [From the Qitarterly Journal of (he Geological Society /or February 1879.] ON THE MICROSCOPIC STRUCTURE OE STROMATOPORIDiE, etc. > * . « It • .•. » e -■•o » 4 »» , « ' < t ') a ,, t 1 i •MMM 48 J. W. DAWSON ON THE MICROSCOnO On the Microscopic Structuue of Stromatoporid,i5, and on Paleozoic Fossils mineralized witJi Silicates, in illustrqtion of Eozoon. By J. W. Dawson, LL.D., F.ll.S., F.G.S., &c. [Plates III.-V.] Among the collateral subjects which have arisen in the discussions with res])0ct to Eozoon canadcnse, two of the most important are : — Ist, the question of itn structural relations with the PaUcozoic fossils of the genus ISti'omatojiora and allied genera ; 2nd, the oc- currence in PaUcozoic rocks of fossils mineralized with hydrous silicates akin to tlie sorpontino and loganite found in some of the best-preserved specimens of the Laurentian fossil. For several years I have taken advantage of every opportunity to make collec- tions illustrative of tlioso questions, and to subject the specimens obtained to microscopic oxamiiuition. In this 1 have been greatly aided by friends wdio will be mentioned in the sequel, and by the large number of excellent sections ])repared by Mr. AVeston, of the Geological Survey of Canada, for the late Sir W. E. Logan and for myself. In the following ])apcr I purpose to state the conclusions arrived at as the result of these observations, with such portions as may be necessary of the large accumulations of facts on which they are based. I. SlROMATOrORIDJE. 1st. Microscopic Structure. — The Stromatoporida3 have long been a zoological stumbling-block, and have been referred to Corals, Sponges, Foraminifera, and even to Hydractinia^. I do not purpose to review these diverse opinions, most of which are undoubtedly based on imperfect acciuaintancc with the microscopic structure of these curious fossils, but to give an intelhgible account of the struc- ture of some ty]ncal species, preparatory to the consideration of their relation with Eozoon. The genus Stromatopora, properly so called, is founded on the species Stromatopora concenfrica, Goldfuss, and its allies, which range from the Upper Cambrian to the Devonian inclusive. Avoiding, for the present, com])lcxities arising from the various states of preserva- tion and of weathering, I may refer in the first instance to remark- ably -well-preserved specimens from the Cornifcrous Limestone of Ohio, and from the island of Marblohead in its vicinity, which have been placed in my hands by Mr. A. E. Walker of Hamilton, and Dr. Newberry of New York, in those the concentric laminai and pillars of the fossil are in the condition of opaque calcite, apparently re- taining its minute structure and not affected by crystallization ; and the interspaces or chambers arc occupied by transparent calcite, • • . • •• •• • • • • • • >• • • • • • • ••••• •,••• • ••» •• ••••II*.* • « •• • • •• • • ,• • •• •• • !• • .. permitti surfaces In th^ in the s interspa width ol or roun they are accident; which ai texture finely gri the mass minute t The lami of a mil pillars ac through 1 theraselv( thicken t to travers the hollo\ to the or( divided b attached i laminae in extending In son than a n several in often covt no connes them as o attribute t in other p The abc Stroniatop Other spe( in the thic size of the faces. In appear re1 that the p in size and examined, porous the cidental. * These s careouB test, Q. J. G. ..:.•••.*.•:•.* •••." •••••• ••• • • • • •• • • •• • • * < • • 8TRUCIURE OF STROMATOPORID^. 49 VLiEOZOlC EOZOON. SCUSSIOUS riportaut .*alicozoic 1, the oc- hydroiis 10 of the r several le coUec- pccimens u greatly id by the in, of' the 1 and for inclusions >rtions aa lich they ong been Corals, purpose oubtedly ucture of le struc- of their on the ich range iding, for preserva- remark- cstone of lich have and Dr. id pillars cntly re- ion ; and calcite, permitting all the structures to be very well seen, either on polished surfaces or in transparent slices *. In these specimens about three interspaces and two laminae occur in the space of a millimetre ; and though neither the laminoe nor the interspaces are uniform in thickness, the latter are about twice the width of the former. In some places the laminae rise into conical or rounded eminences with corresponding depressions ; in others they are nearly flat and concentric, this difference being apparently accidental. The laminae are connected with each other by pillars, which are either round or somewhat flattened (PI. III. fig. 1). The texture of the laminae is not spicular, but peifvctiy continuous and finely granular, as if made up of minute fragments of calcite. When the mass is broken parallel to the laminae, the piUars appear as minute tubercles (PI. IV. fig. 5), but a true exterior surface is smooth. The laminae are pierced with numerous round pores about one tenth of a millimetre in diameter. Some of thes pass through hollow pUlars across one interspace into another ; but the greater part pass through the laminae from one interspace into the next. The laminae themselves are here and there pierced with horizontal tubes which thicken the laminae where they pass (PI. Ill, fig. 3, b) ; they appear to traverse the laminae obliquely from one space into another, or from the hollow pillars laterally. They may be called canals. In addition to the ordinary laminte, some of the chambers or interspaces are sub- divided by very thin secondary laminae. In a few cases these are attached to ordinary laminae as a sort of inner wall. The ordinary laminae in the more regular specimens are often of great continuity, extending without interruptions for several square inches. In some specimens there are a few rounded perforations, less than a millimetre in diameter, which extend vertically through several interspaces. Their walls are densely calcareous. They are often covered up with the growth of the lamiuoB, and seem to have no connexion with the other parts of the structure, I do not regard them as oscula, but as perf(..rations of some parasitic animal ; and I attribute to the same origin certain rounded cavities, similarly waUed, in other parts of the mass. The above is an accurate description of the most common type of Stromatopora when, as nearly as possible, in its natural condition. Other species of the genus, as now usually limited, differ principally in the thicknesses and distances of the lamina?, in the number and size of their perforations, and in their more or less tuberculated sur- faces. In some species the pores are so numerous that the laminae appear reticulated ; in others the laminae are so much thickened that the pores appear as tubuli. The pUlars also differ somewhat in size and form. In the very numerous specimens which I have examined, I have convinced myself that while the laminae are always porous they are never spicular, and that the so-called oscula are ac- cidental. They are due either to the causes above referred to or to * These specimens are associated with a beautiful Sacoammina, having a cal- careous test, which I have described as Saccammina criana. Q. J. G. S. No. 137. 1 60 J. W, DAWSON ON THE MICROSCOPIC branches of Syrimjopora and similar corals included in the mass of ^troraatopora. It is evident from the above description that the animal matter of 8tromatopora must have occupied the chambers or interspaces, and must have extended from chamber to chamber through tho pores and hollow pillurs. Such a structure is obviously that of a rhizopod rather than of a sponge. Furtlier, tho arrangement of tho laminae and pillars is very nearly allied to that of Parleria and Loftiisia as described by Carpenter and Brady, which T have myself studied in s])ecimens kindly given to me by Professor T. 11. Jones *. In so far as the hollow pillars and perforated plates are concerned, it has some points of correspondence, though more remote, with Iteceptaculites. The supposed oscula on which lias been based a reference of those forms to sponges are certainly not constant. I have seen large masses of the form above described, presenting more than 30 s(]uarc inches of surface, without a trace of an oscu- lum ; and in those specimens where tubular orifices appeared, I have found that they cut like perform tions made by a boring instrument through the mass, irrespective of its structure, and that they were lined with continuous calcareous walls different from the laminaj of the fossil. It is scarcely necessary to say, after the above descriptions, that I attach no scientific value to the ingenious and elaborate attempt of Mr. H. J. Carter ('Annals and Magazine of Natural History,' ser. 4, vol. xix. p. 44) to prove that ^:itromatoporce are skeletons of hydroids allied to Hjidractiaia. The resemblances of StromatoporcB to those hydroids arc altogether superficial, and depend on both having a parasitic and concentric habit of growth. In every essential cha- racter they differ entirely, and can have no close zoological affinity. In comparison with Eozoon^ the ^oneral appearance and habit of growth are so similar that specimens cannot easily be distinguished by the naked eye, or where the minute structures are not preserved. In microscopic structure the thin lamina? of StromatoporcB correspond to the proper wall of Eozoon. The thickening of the walls corre- sponds to the supplemental skeleton, and the horizontal tubes to tho canals, while the interspaces and the pillars correspond to the chambers and connecting walls of the older fossil. The main struc- tural difference is, that while Eozoon has a delicately tubulated proper wall of Numrauline tj'pe, that of Sfromatopora has coarser perfora- tions and pores. Stromatopora and Eozoon may both be regarded as largo sessile laminated calcareous Rhizopods ; but the former pre- sents a less generalized type than the latter, which combines struc- tures that were usually separated even in the Palaeozoic period. Stromatoporce of the type above described are abundant in the Corniferous Limestone. They occur throughout the Upper Silurian and are especially abundant and of large size in the Niagara Lime- stone, where they abound even in those Dolomitic beds that contain * More recently I have also stiulied the remarkably beautiful species of Loftusia from British Columbia described by Mr. G. M. Dawson, which covifirm the resemblance of these specimens to Stro^natoporw (see his paper read before this Society, infra, p. 69), STRTTCTTTRE OP STROMATOPORID^. ffi few other fossils. They arc also abundant in the Dolomite of the Guelph Liraestone ; and it is perhaps not accidental that both hero and in the Laurentian, fossils of this structure are associated ■with magnesian rocks. They occur also in the Lower Silurian, though less abundantly ; and the oldest specimen I have seen is in the Pots- dam Sandstone ; and this, its structure not being preserved, may have belonged to Eozoon rather than to Stromatnpora. The Lower Sihi- rian species have usually very thin and continuous walls. In the great Niagara Limestone, as seen at Niagara Falls, the masses of Stroma tojyor a occur precisely as Eozoon occurs in the Laurentian limestones, and are mineralized with quartz and dolomite, and often almost entirely converted into crystalline masses, though occasion- ally showing their structure in great perfection. Certain beds of the Niagara formation, near Hamilton, contain not only Strcmatojiorce, but multitudes of sponges ; and through the kindness of Lieutenant-Colonel Grant, of that place, I have been enabled to examine a number of specimens of these, and to compare them with Stromatojporce. These sponges are all siliceous and spicu- late, and belong chiefly to two or three species of Astylosiponyia of Homer, and to Aulocopina of Billings, of which his A. Grantii is the type. The species of Asti/lospomjia present a most regular and beautiful hexactinellid structure, as perfect as that in the sponges of the Cretaceous, showing even the hollow nodes, which have been supposed to be absent in the Pala3ozoic Hexactinellida). Auloco- pina has a different structure, presenting series of hexagonal tubes built up with interlaced spicules, and giving off bundles of spicules in a radiating manner. These sponges are thus entirely distinct, both in material and structure, from the contemporary Stromato- j)orcB, and there is no link of connexion whatever. The species included in the genera Caunopora of Phillips and Ccenostroma of Wincholl, and in part in Si/nnqostroma of Nicholson, and which may be represented by the Stromatopora pohjmorpha of Goldfuss, have the horizontal canals largely developed in laminae thickened by supplemental deposit, and traversed by an infinity of minute canaliculi or ramifications of the canals opening at their surfaces. The horizontal canals radiate from central points whore they are connected with vertical tubes or groups of tubes penetrating the whole thickness of the mass (PI. IV. fig. 9, and PI. V. fig. 10). The whole organism thus becomes divided into a series of vertical systems, which often very much obscure the concentric lamination, and in different states of preservation give very perplexing appear- ances. They may all be explained by bearing in mind that the hori- zontal canals, like those of IStromrtto^'ora proper, pass in the substance of the laminae, now much thickened, and that at the centres of the systems they descend through the chambers by vertical tubes or groups of tubes which correspond to the hollow pillars of Stromatopora. A great number of specimens of Caunopora^ Cosnostroma, and allied forms, both European and American, have passed through my hands ; but I was unable to decide, except inferentially, as to their minute structure, till I was so fortunate as to obtain, through the e2 52 J. W. DAWSON ON THE MICROSCOPIC kindness of Mr. Selwyn, Director of the Geological Survey of Canada, a specimen collected by Professor li. lioll on the Albany lliver, Hudson's liay, in rocks of Upper Silurian age. In this specimen the skeleton remains as carbonate of lime, while the canals and tubes arc in great part empty, so that their minute ramilications are in the condition of a recent specimen, and can be injected with colouring- matter. The actual structures tlius presented are as follows : — Lamina) thin and obscure. Chambers almost entirely filled with supplemental deposit, traversed by innumerable microscopic hori- zontal canaliculi, which are tortuous and anastomose frequently. They are connected with systems of radiating canals which termi- nate centrally in vertical tubes traversing the whole thickness of the specimen, and opening at the surface in round pores visible to the naked eye and placed on the summits of slight eminences. The pores are about 4 millimetres apart horizontally. The upper surface is smooth and does not show the radiating canals, but these are disclosed by erosion or by horizontal fracture. This species closely resembles Hall's Caunopora incrustans, from the Devonian of New York ; but the pores are more regular and less than half as far apart, and the radiating tubes are more numerous. For the above species I have proposed the name Caunopora huclsonica (PI. lY. fig. 9 and PI. V. fig. 10). I have described in a former publication* a fossil preserved in a similar way, but less perfectly, and which has vertical tubes in groups instead of singly ; it is from the Gait Limestone of Ontario, and belongs to the genus Coenostroma, as limited in the sequel. I would propose for it the name O. (jaltense. The structures in Caiuiojjoni and Cceaostroma are unquestionably, at first view, more akin to those of sponges than arc those of the typical Stromatoporo}, as the vertical tubes may be taken for oscula, and the extremities of the fine tubes for the incurrcnt pores. On the other hand, the solidity of the calcareous walls and supplemental thickenings is at variance Avith such a view ; and in many respects they more nearly resemble Eozoon than any of the Palajozoic fossils with which I am acquainted. The canal-system in both is, indeed, so much alike that it would not be easy to distinguish it, except that Eozoon wants the continuous vertical tubes and possesses a true nummuline wall. The minute structures of such species as the Sbwnatopora noda- Icita of Nicholson (>S'. sanduslajana of Riiminger, PI. V. fig. 11), con- nect the true Stromato2yorce with the species of the genus Ccenostroma ; and these, by confluence of the separate tubes, pass into those of the genus CauHopora. Of the species separated by Nicholson in the genus Syringostroma, that which he has named S. columnaris is a very peculiar type. It is penetrated vertically by what seem to besohd columns, and which, on microscopical examination, prove to result from upward bending and fusion of the laminae along certain vertical lines. The effect is obviously to give much additional strength to the skeleton. Between * ' Life's Dawn on the Earth,' p. 160. 8TRTJCTTTRE OP STROMATOPORTBJ!. the columns the laminsD are supported by pillars as in Stromatopora. They are also penetrated by horizontal canals which ramify radially, and are connected with vertical tubes as in Ccenosti'onui, to which this form is very closely allied ; I have seen only one species from the Corniferous Limestone, specimens of which have been kindly given to me by Dr. Newberrj' and Mr. Hinde of Toronto. Dictyostroma of Nicholson includes species in which the con- necting pillars arc formed by upward bending of the laminae them- selves in conical points. The only species described by Nicholson (7). nndulatum) is from the Niagara formation of Louisville, Ken- tucky. I have, however, seen an imperfectly preserved specimen with this structure from the Black-]liver Limestone of Port Claire. Mr. Hinde has sent me another from the Corniferous of Port Col- borne. These seem to be different from Professor Nicholson's species in the distance between the laminae, which is much less than in the coarsely constructed species which he has described. The laminae are porous in these specimens ; but I have seen no vertical tubes or oscula. The species Stromatopora compacta, from the Trenton Limestone, described by Billings, and which is not uncommon, does not appear to belong to this group of organisms. It consists of very minute hexagonal tubes with extremely thin walls and well-developed tabulae, which, from their strong development and continuity, give in some specimens an appearance of concentric lamination. The species seems to belong to the genus Stenopora, but its cells are excessively minute. Corals of the genus Fistulijwra, with small tubes imbedded in a cellular ccenenchyma, may readily, in certain states of preservation, be mistaken for Stromatopora. Stromatoporce seem not unfrequently to have overgrown corals of different species ; and, in the case of Syringopora, the tubes of these projecting through the mass often simulate oscula. Mr. Hinde in- forms me that, in the case of one species, this association is so com- mon that it suggests the idea of a case of " commensalism." As connected with Stromatopora, it may be well to remark that some misapprehension still appears to exist respecting Archceocy- athns. a fossil of the Cambrian rocks of Mingan, Labrador, and of which several species have been described by Billings. Of these the only one I have studied is A. profundus. This is certainly a calca- reous, chambered organism, with pores connecting the chambers, and must have been the skeleton of a Rhizopod. The other species have similar structures. It is true, however, that on treating them with acids, Billings obtained siliceous spicules in the matrix, which I have myself examined. I regard them as having belonged to lithistid sponges of the genus Trichosponr/ia of Billings, accidentally asso- ciated with the ArcJiceocyathvs. Some specimens of Stromatopora present remarkable lines of growth, caused by the appearance of two or three layers of smaller cells at intervals of seven or eight interspaces (PI. IV. fig. 4). The preservation of these without the intervening portions may, in some cases, account for the abnormally wide interspaces sometimes seen H J. W. DAW801T OK THE MICROBCOPIO in imperfectly preserved Bpccimens. I have not been able to satisfy myself whether these linos of growth are of specific value. In one epecimeu from the Devonian of Iowa, a Stromatopora of this typo presents largo vertical tubes which extend from one growth-line to the next, but are sparsely distributed througli the mass, and not connected with radiating tubes as in Caimopora. Another interesting structure, soen in a sjjecies from the Comife- rous Limestone, usually, tliough perhaps incorrectly, identified with /S*. concent7'ica, is the division of the pillars at their summits into branches (PI. III. fig. 2), so as to support at many points the layer above, whicli in this case is thin and not much strengthened with supplemental deposit. 2nd. State of Preservation. — Stromatoporce have apparently always boon calcareous when recent. They are sometimes preserved in tho state of calcito with the chambers cither tilled with tho same mate- rial or with silica. Sometimes they are entirely silidfied, or tho lamina) and pillars arc silicifted and tho chambers filled with calcite. Occasionally tho chambers aro filled with dolomite or the wholo structure is dolomitizcd. A specimen of tho type of S. conceutrica from the Devonian or Upper Silurian of James's Bay is now before mo, and affords a good illustration of modes of preservation in silica. In some places tho lamiiiDD and pillars have been silicificd, while tho chambers are filled with limestone. When weathered or treated with acid, those portions show tho wholo structure very clearly, including tho perforations of tho lamiuoe and tho hollow piliars (PI. IV. fig. 7). Other portions have tho chambers also tilled with silica, tho lamina) being distinguishable by thtur loss transparent and ])orous character. In these portions the laminoB and jnllars have usually been first coated over with minute crystals of quartz. A layer has then been dei)osited of chalcedony with botryoidiil surfaces, and finally the remaining cavities have been fiUed with crystalline vitreous quartz. In the greater part of the specimen, however, the chambers have been filled with silica, while the lamina) have remained as calcite, and those portions, when weathered, present tho ap])oarance of thick structureless laminae separated by thin spaces and penetrated by numerous round holes representing the pillars. Portions in this state might be mistaken for a coral of the type of Fhtidipora, but in certain aspects they present that lobated amoeboid form which is so characteristic of similarly preserved specimens of Eozoon. In sjjeeimens of Stromatopora from the Xiagara Limestone, it not unfrequently happons that certain layers or groups of layers are silicificd, while others alternating with them remain as calcite. In this case, when the specimens are weathered, they present distant concentric layers very difterent in appearance from the actual structure. As with other fossils, crystallization plays strange freaks with Stromatoporce, reducing them to such a condition that, but for the partial preservation of portions hero and there, thoy might be mis- taken for inorganic bodies. This is well seon in the abundant Stro- 8TRUCTTJRB OP STROMATOPORIDiE, ts matoporce of tho great dolomito-beds of the Niagara Liraostono. Of these many are entirely reduced to erystallino masses of quartz or dolomite, except small portions at the surface, while others havo become hollow and resemble cavities lined or filled with crystals. In tho Upper Silurian dolomite of Guelph, in like manner, there aro specimens which have been converted into a granular dolomite, in which, however, the lamina) and, in some cases, the canals aro moro or less apparent. The study of Stomatoporce in these different conditions throws great light on the appearances presented by Eozoon in various states of preservation, and forms a guide to the interpretation of these, which should bo before tho mind of every one who desires to form correct opinions on the subject. Since writing the above, I have seen the remarks of Dr. Nicholson on the calcareous nature of titromatopora, and Zittel's observation of the occasional calcification of the spicules of siliceous sponges, as reported in the ' Geological Magazine' for January 1878. It had not occurred to me that any one acquainted with Stromato- porce would doubt their calcareous nature ; but Nicholson has suf- ficiently disposed of such doubts by tho consideration that the Stro- matoporce are found silicificd only in beds in which corals and shells have suffered the same change. Nor had it seemed necessary to refer in this connexion to tho replacement of siliceous spicules with calcite. It is, or should be, well known from the behaviour of sili- ceous spicules with alkalies, and when heated, that many of tlicm aro not purely siliceous, but contain animal matter. This, with tho more soluble character of their silica, enables them to bo changed or removed without affecting tho siliceous matiix. Hence in tho sili- ceous sponges from the Niagara Limestone the spicules aro some- times opaque and granular in appearance, or have disappeared alto- gether, or have been replaced with calcite or with iron pyrite. These changes are, however, rare, and have no bearing on the calcareous nature of Stromatoponi. 3rd. Class IJlc at ion of Stromatoporidie. — It is not my purpose to enter into any revision of tho numerous species of this family, or to attempt to summarize the work which has been done with refereme to the American species by Hall, Wincbell, Nicholson, liillings, Rominger, and others. It may, however, Lo useful to state tho results at which I have arrived with reference to the leading generic forms. 1. Stromatopora (Goldfuss, 1827). — In tho original definition by Goldfuss the genus is characterized as exhibiting ''alternating strata of a solid and porous character." The porous strata in this defini- tion are the real lamina?, the solid strata are the fillcd-up chambors ; and according as one or tho other is preserved, we have in this type thin laminae connected by pillars, or thick Inmina) perforated with round holes and separated by vacant spaces. The typical species is S. concentrica of Goldfuss, and the genus may bo held to include all the species with thin or moderately thick lamina) connected with solid and hollow ])illar3 and perforated with minute pores, or having 56 J. W. DAWSON ON THE MICROSCOPIC a reticulated texture. Stromatocermm of Hall is a 8}Tion5Tn. The genus ranges from the Upper Cambrian to the Devonian inclusive, and it is not easy to separate the species w hich have been described. 2. Caunopora (Phillips, 1841). — The typical species, C. placenta, is defined by Phillips as " amorphous," composed of concentric or nearly plane masses perforated by flexuous or vermiform small tubuli and by larger straight subparallel or radiating open tubes, persistent through the mass. This definition includes those species ■with simple tubes giving origin to radiating tubuli passing through the thickened laminae. Coenostroma (Caunopora) incrustans of Hall is a typical American species, as is also Caunopora hudsonica above referred to. 3. Coenostroma (Winchell, 1867). — This genus, as defined by its author, includes those species in which the radiating tubes diverge from tho surfaces of little eminences raised in the concentric lamellae ; but, as Hall has well remarked, the presence or absence of eminences is a trivial character. The real distinction should be based upon the absence of the central simple radiating tubes, which in these species are represented by a group of more or less divergent ascending tubuli, so that the surface of the last layer presents eminences not with a single large pore at the summit, but with several small pores diverging from their sides. My Coenostroma galtense above referred to and Hall's Caunopora {Coenostroma) plamdata are typical forms. T have been obliged to reverse the generic names as used by Hall, in the twenty-third Report of the Regents of New York, in the manner above stated, as Phillips's name certainly applies to the species with sinffle vertical tubes. Stromatopora nodidata of Js'icholson (Ohio Report, vol. ii.) pro- bably belongs to this genus. 4. Sifrinf/ostroma (Nicholson, 1875). — This genus is defined by the author as composed of concentric laminae and vertical pillars which are so thickened and so amalgamated with one another as to leave nothing but the most minute rouuJed cells. Laminated tissue traversed by numerous large irregularly disposed horizontal canals. The species S. densa included under this genus in the Ohio Report is undoubtedly to be referred to Coenostroma as above defined, being a species with the vertical tubes small and the radiating tubes very large. The species S. columnans has, however, a very special cha- racter in the apparent want of vertical tubes, and in the coalescence of the laminae along certain vertical lines, giving solid columns ter- minating in imperforate or microscopically perforate tubercles at the surfaces. 5. Bictyostroma (Nicholson, 187')). — In this genus the upper sur- face of each lamina is developed into conical points which support the lamina above instead of pillars. The laminae have horizontal canals, and the upper surface is apparently solid, though, no doubt, minutely perforate ; the irregular oscula referred to by Nicholson are probably accidental. D. undulata, Nicholson, is the typical species ; but I think we may add to these several others in which BTBtTCTXJRE OP STiiOllATOPORID^. &7 ' the thin laminae extend upwards in conical forms instead of pillars, and in some of which the laminse are thin and apparently destitute of the horizontal tubes. It is not impossible, though the specimens in my possession are not sufficiently perfect to render this certain, that Lahecliia conferta of Edwards and Haime, from the English Wenlock, may be allied to this or the last genus. [The above descriptions of Stromatoporidao were written before the publication of Nicholson and Murie's excellent memoir in the Journal of the Linnean Society, which reached me only a few days before the proof of the above pages. Their descriptions of the structures, and views as to the classification and affinities, agree in the main with those above given, and where they differ deserve careful consi- deration. They do not seem to have met with so good examples of the hollow pillars and perforated laminie as those I have described, nor to have so distinctly observed the relation of the horizontal canals to a supplemental deposit of calcareous matter. In their comparison with FarTceria too much importance is, I think, attached to the arenaceous character of that fossil — a character which we find in living Ehizopods associated with forms not dissimilar to those which are calcareous. It is also not improbable that some Stroma- toporidae are built up of microscopic calcareous grains. Loftusia likewise presents points of comparison of some importance ; and the Carboniferous species of that genus described and figured by Dr. G. M. Dawson (see p. 69) is especially instructive. In the memoir in question the genus Syringostroma of Nicholson is divided into Stylodictyon and Pachystroma, Stromatoceruon of Hall is re- tained as a separate genus for some peculiar Stromatoporidse of the Lower Silurian, and a new genus (Clathrodictyon) is formed for vesicular species without pillars. The separation of forms contained in Syringostroma I have myself suggested above ; and I think the grounds for retaining Stromatorerium and adding Clathrodictyon may be sustained. The authors should, however, I think, have retained Ccenostroma of Wincholl, and placed in it some forms which they have distributed in other genera. The new facts stated respecting Lahecliia are important with reference to that somewhat problematical fossil.] The geological distribution of the American Strom atoporida) known to me may be stated as follows, though the species, no doubt, require some revision : — Totsdam formation Stromatopora, sp. Trenton formation Stromatopora rugosa, Hal'. Dictyostrcnia ? sp. magara formation Stromatopora concentrica, Goldfuss. Ccenostroma constellatum, Hall. Caunopora hudsonica, n. sp. Dictyostroma iiiidulatum, Nicholson. Gvelph formation Stromatopora ostiolata, iV^ Ccenostroma galtense, n. sp. 58 J. W. DAWSON ON PAIiEOZOIC P0881L8 Coniiferous formation Hamilton formation Chemung formation Stromatopora granulata, Nicholson. S. iiiamraillata, N, S. Hindei, N. S. perforata, N. S. ponderosa, N. S. substriatcUa, N, S. tubcrculata, N. Syriiigoslroma columnare, N. Cocnostroma dcnsuin, N. Stromatopora mix, Winchell. S. civspitora, W. Coenostrouia monticulifera, W. 0. pustulifera, W. Stromatopora expansa. Hall, S. erratica, H, S. alternata, H. Caunopora incrustans, H, Ccenostroma soliduluin, H, 0. planulatum, H. II. Paleozoic Fossils associated with Serpentine and otheb Hydrous Silicates. Fossils having their cavities and pores infiltrated with hydrous silicates are much more abundant in Palaeozoic limestones than is usually imagined. In some instances serpentine itself is found to have been concerned in sucL infiltration : while in other cases the infiltrating hydrous silicates are found to approach to poUyte, fah- lunite, and other minerals which have usually been regarded as pro- ducts of decomposition or metamorphism, but which, as Dr. Sterry Hunt has justly remarked, cannot reasonably be referred to such an origin when they are found tilling the pores of Crinoids and other fossils in strictly aqueous deposits. In this case they must surely be the results of original dcpot.itiou in the manner of glauconito ; and, as we shall find, they sometimes appear to be strictly the represen- tatives of that mineral, which occurs under similar conditions in other parts of the same formations. 1. Serpentine of Lal-e Chebo(/amo7ir/. — Mr. Richardson, of the Geological Survey, has observed, north of the Laurentian axis, on tlie Sagucnay Paver, certain rocks which a])pcar to bo similar in mineral character to the Quebec group of Sir William Logan, and occupy a geological position intermediate between the Laurentian and the Trenton formation. Among these he describes a band of serpentine associated with limestone at Lake Chebogamong, which lies about 200 miles to the N.E. of Lake St. John, in a little-explored region. Among the few specimens which Mr. Richardson was able to bring back with him was one of extreme interest — a specimen ap])arently from the junction of the limestone and serpentine, and containing a portion of a tabulate coral, of which some of the cells are filled with a mixture of serpentine and calcite, and some with calcite. The serpentine seems to have been weathered; it has a granular, uneven a])pearance, and under the microscope shows patches with fibrous structure like chrysotile. There aro also whitish serpentine veins, fringed with chrysotile or a mineral re- MINERALIZED WITH SILICATES. 59 sembling it under the microscope. The cell-walls of the coral are perfectly black and opacjue, and probably carbonaceous. The coral found thus mineralized was examined by Mr. Billings, who had no doubt of its nature, though uncertain as to its generic affinities . After careful study of it, I am disposed to refer it to the genus Astroceriwn of Hall, and it is not distinguishable in structure from A. pyriforme of that author, a species very common in the Upper Silurian limestones of the region in which the specimen occurs. The genus Astrocerium is specially characteristic of the Niagara forma- tion ; and though Edwards doubts its distinctness from Favosites, I think there are constant points of difference, especially in the micro- scopic characters of the cell-walls, which entitle it to be separated from that genus. In such specimens oi Astrocenum as are well pre- served, the walls of the hexagonal colls seem to have been of corneous texture, with minute corneous spicules instead of radiating septa. They have pores of communication, and there are also occasional larger pores or tubes in the angles of the cells. The tabulae are very thin and apparently purely calcareous. This accounts for the sin- gular fact, mentioned by Hall, that the cell- walls are sometimes entirely removed, leaving the tabulae in concentric floors like those of SU'omatopora. I think it likely that the typical species of As- trocerium may have been inhabited by Hydroids, and may have been quite remote from Favosites in their affinities. The formation in which the serpentine and limestone of Lake Chebogamong occur is described as consisting of chloritic slates, in some places with hornblende crystals, dolomites, and hard jaspery argillaceous rocks. Upon these rest conglomerates and breccias with Laurentian fragments, and also fragments of the rocks before mentioned, and on these lie the limestone and serpentine. Tlio ser- pentine has been analysed by Dr. Hunt, who tiuds it to contain chromium and nickel, and in this respect to be similar to that of the Quebec group, and not to that of the Laurentian*. The fossil would give evidence of a much later date than that usually attributed to rocks of the character above stated ; but it is quite possible that there may be two series of different ages in the region, the lower being Lower Silurian or perhaps older, and the upper of Upper Si- lurian age. If the serpentine belongs to the newer formation, its association with a coral of the genus Astroceriam would of course bo quite natural. If it belongs to the older formation, and the over- lying limestone to the newer, the serpentine in the latter may bo a remanie silicate derived from the older rocks and mixed with the limestone at their junction. 2. Serpentine of Melbourne. — The serpentines of this place belong to a great series of more or less altered roclcs extending through the province of Quebec, and referi'od by Sir A\'illiam Logan, on stra- tigraphical grounds, to his Quebec group, equivalent to the Arenig or Skiddaw series of England f. In ascending order these rocks at Melbourne present first a thick series of highly plumbaginous schists * Eeport of Geological Survey of Caniida, 1870-71. t Hunt, however, holds that these rocks are in part Iluronian, w 60 J. W. DAWSON ON PALEOZOIC FOSSILS or shales, with thin bands of limestone holding fragments of Lower Silurian corals and crinoids. These pass upwards into a thick series of slaty rocks characterized by the prevalence of a shining crystal- lino hydro-mica, and known as nacreous or hydro-mica slates. They are associated with quartzose bands, and also with lenticular layers of crinoidal limestone. Parallel with these beds and, according to Logan's observations, overlying them, is the series containing the serpentine, which is associated with layers of limestone and nacre- ous slate, and also with brecciated and arenaceous beds, probably originally tufacoous, with beds of auorthito, steatite, and dolomite, and also with red slates, the whole forming a miscellaneous and ir- regular group, evidently resulting from the contemporaneous action of igneous and aqueous agencies, and affording few traces of fossils. The serpentines, which occur in thick and irregular beds, are differ- ent in colour and microscopic texture from those of the Laurentian system, and also present some chemical differences, more especially in the presence of oxides of nickel, chromium, and cobalt, and of a larger percentage of iron and a smaller proportion of water*. These serpentines are undoubtedly bedded rocks and not eruptive ; but they may have originated from the alteration of volcanic mate- rials f. They appear, shortly after their original deposition, to have been broken up, so as in many places to present a brecciated ap- pearance, the interstices of the fragments being filled with limestone and dolomite, which themselves are largely mixed with the flocculent scrpentinous matter, and traversed by serpentinous veins some- times compact and sometimes fibrous. Besides the very impure limestone thus occurring in the serpentinous breccia, there are also true layers or beds of limestone and dolomite included in or near to the great serpentine band. No well-preserved fossils have been found either in these beds or in the brecciated serpentine ; but on treating the surfaces of slabs with an acid or making thin slices, fragments of organic bodies are developed which well illustrate the manner in which serpentine, whatever its origin, may bu connected with the mineralization of such fragments. It is to be observed here that the irregular bedding of the serpen- tine, and the apparent passage on the line of strike into dolomite and red slate, might accord either with a purely aqueous and oceanic mode of deposition like that of glauconite, or with deposition as beds of volcanic sediments, afterwards altered and partly redeposited by water. The association with ash rocks and agglomerates would, how- * Under the microscope the Laurentian serpentines are usually homogeneous and uncr3-8talline, but with the structure of netting veinlets which I have elsewhere called septariform. The Melbourne serpentines usually present a confused mass of acicular crystals or a fibrous structure, and, where structure- less, polarize more vividly than those of the Laurentian. t '^" • Ibergei- (Essay on Metallic Veins) quotes many German chemists to the lat " olivine rock and the serpentine formed from it always contain cop ' ickel, and cobalt." This origin might thus apply to the serpentines in the v^uebec group in Canada, but not to those of the Laurentian, as I have already urged on other grounds in my reply to Hahn, in the 'Annals and Magazine of Natural History,' 1876, vol. iviii. pp. 32, 33. MINEilALIZED WITH SILICATEa. 61 ever, tend rather to the latter view, as would also the chemical cha- racters of the serpentine already referred to; hut the aasociation with fossils mentioned below tends to show that at least a part of the mineral is an ordinary aqueous deposit. It is also to he ob- served, with reference to the superposition of serpentine on fossilife- rous Lower Silurian rocks, that a similar relation is atiirraed by Murray to occur in Newfoundland, where massive serpentines overlie unaltered fossiliferous rocks of the Quebec group *. No fossils have been found in the compact serpentine, but only in the limestone paste of the brecciated masses and in the Hmestone bands inter stratified. The limestone of the breccia contains not only angular fragments of serpentine but disseminated serpentine and small veins of the same mineral. Its fossils are limited to small tu- bular bodies, crinoidal joints, and fragments, apparently oi Stenopora, very imperfectly preserved. The tubular bodies may be portions of Hyolithes or TJieca. Their interior is usually filled with dolomite ; their walls are in the state of calcite ; and they are incrusted with an outer ring of serpentine. In some instances the calcareous organic fragments are seen to be filled in the interior with serpen- tine. The crinoidal fragments are in a similar condition, the ser- pentine having apparently surrounded them in a concretionary manner after the cavities had been filled with dolomite. Frag- ments of calcite, dolomite, or older serpentine included in the lime- stone, and of no determinate form, are enclosed in the new or re- manie serpentine in like manner, and in some cases this newer or coating serpentine was observed to have a fibrous structure. The serpentine thus coating and filling fossils and fragments is of a lighter colour than the serpentine of the fragments themselves, and in this respect resembles that of the small veins traversing the lime- stone. Such traces of fossils as exist in the layers of limestone are similar to those in the breccia, but not, so far as observed, coated with serpentine. It would thus appear that, contemporaneously with the original deposition of the serpentine, thin bands of limestone were laid down, with a few fragments of crinoids, corals, and shells ; that subse- quently, but perhaps within the same geological period, and while the deposition of serpentine was stiU proceeding, portions of the surface of the serpentine were broken up and imbedded in limestone ; that the fissures of this limestone were penetrated with serpentine veins, and its few fossils coated with that mineral, which also forms flocculent laminae in the limestone. The mode of deposition of this Palaeozoic serpentine is thus con- siderably different from that of the Laurentian, which forms layers intimately interstratified with great limestones, and also nodules, concretionary grains, and fiUings of fossils in these limestones. This difference in mode of occurrence is, no doubt, connected with the dif- ference in composition of the two varieties of the mineral already noticed. In both eases, however, the serpentine has been so depo- * Bedded serpentines also occur in unaltered Silurian dolomites at Syracuse in New York (Hunt, Chem. and Geol. Essays, p. yiO). J. W. BAWaON ON PXTJEOZOta FOSSILS sited that it could take part in the mineralization of marine organic remains. The condition of the fragments of Silurian fossils in the limestones associated with the nacreous or hydro-mica slates is of interest in connexion with this subject. The shining laminated mineral asso- ciated with these fossils has been regarded from its chemical com- position as a hydro-mica. Under the microscope, however, it shows a Wiint of homogeneity which suggests the presence of two or more silicates, or the association of crystals of hydrous mica with minute grains of siliceous matter of some other kind. Though now highly crystalline, it must originally have been a fine sediment, since it fills the finest colls of Stenopora and Ptilodictija. Nor can its present state luive been produced by any extreme mctamorphism, as the uudistortcd state of these fossils amply testifies. Further it is in- teresting to observe that though tlie hydrous silicate is little mag- nesiau, the fossils themselves are not infrequently converted into dolomite. In tliese fossilifcrous beds there are also tabular crystals, apparently of anhydrous mica, little groiips of crj'stals of tremolite, cavities filled with (]uartz, and crystalline grains of a mineral ha\'ing the microscopical characters of olivine ; and these have been de- veloped or included in the mass without injury to the structures of the most delicate corals. Similar appearances are presented by limestones from other parts of the Quebec group, of which a great series of slices has been pre- pared by Mr. Weston under the direction of the late Sir W. E. Logan, who, in his later researches in this group of rocks, gave much atten- tion to the microscopic fossils in the more altered beds, as a means of determining their ages. Eesides large series from Melbourne and its neighbourhood, I have examined slices from Stanford, Farnham, Cleveland, Bedford, Orford, Arthabaska, Point Levi, lliviere du Loup, and other places, in most of which Lower Silurian fossils occur as- sociated with hydrous silicates. The fossils above referred to occur in rocks undoubtedly of Lower Silurian age, and regarded as altered or metamorphosed members of the Quebec group. In the unaltered representatives of these rocks at Point Levi and the island of Orleans there occur considerable quantities of a true glauconite, which has been analyzed by Dr. Hunt, and wliich is without doubt an oiiginal deposit in the sandy and argillaceous beds in which it occurs, which in many cases are pre- cisely similar to Cretaceous greonsands. Dr. Hunt's analysis shows that this glauconite contains alumina, iron, potash, and magnesia, and thus approaches to the Laurentian loganite. In the forms of its little concretions it resembles the serpentine grains in the Laurentian limestones ; and like modern glauconite it has moulded itself in organic forms. Some of these are spiral or multilobate, as if casts of minute univalve shells or of spiral and textularine Fora- minifera*. Others are annular or are arcs of circles, and some pre- * Ehrenborg lins found casts of rotaline and textularine Foraminifera in Lower Silurian beds in Eussia; and such forms occur in Upper Silurian lime- Btones iu Nova Scotia, MINERALIZED WITH SILICATES. 6d sent a delicate fibrous or tubulated appearance, as if they had moulded themselvea on porous shells or very raiuutely-cclled corals, spicules of sponges, &c. Shreds of corneous Polyparies, perhaps of Graptolites, abound in the matrix, but are not connected with the glauconito grains. Unfortunately there arc no Stromatoporce in these beds, otherwise we might have an almost precise recurrence of the relations of serpentine with Eozoon in the Laurontian*. Another appearance which may be mentioned in this coimcxion occurs in certain beds of Utica Slate in the vicinity of the trappoan mass of Montarville, and converted into a hard sonorous rock. In one of these are stems of crinoids Avhich have retained their external form, while the calcareous material has been entirely removed and replaced by a soft green crystalline mineral whose physical and mi- croscopical characters are those of chlorite, and which in any case may be regarded as one of those hydrous silicates sometimes termed " viridite." 3. Limestone of Pole Hill, New Bninswiclc, and of Llanrjnn/Uor/ in Anglesey. — In a paper in the Transactions of the Koyal Irish Academy, and subsequently in * Life's Dawn on the Earth,' I noticed a remarkable limestone discovered by Mr. C. Robb, of the Geological Survey, at Polo Hill in New Brunswick, and believed to be of Upper Silurian age. It is composed of fragments of crinoids and shells, the cavities of which are finely injected with a hydrous silicate of alumina, iron, and magnesia, the composition of which, according to Dr. Hunt, approaches to that of the pollyte of Von Kobell, and also to that of a hydrous silicate djscribed by Hoff'mann as filling the cavities of specimens oi ^r:oon found in Bohemia. It has also some resemblance to the loganite which mineralizes the Eozoon of Burgess, in Canada. At the same time I mentioned a specimen of limestone of similar character which I had found in the !NIcGill-College collec- tion, and which I supposed to be from "Wales. It is labelled " Llau- golloc," and belonged to the collection of the late Dr. Holmes, of Montreal. Professor Ramsay, to Avhom I have applied for informa- tion as to the locality, kindly informs me that the name is probably " Llangwyllog," that the place so named is in Anglesey, and that limestone of Lower Silurian or Cambrian age occurs in its vicinity. A portion of this specimen was submitted to Dr. Sterry Hunt, from whose analysis it appears to be of similar character with that of Polo Hill, and like it injects in the most beautiful manner the pores and cavities of crinoids, shells, and corals f. The limestone containing this silicate is of subcrystallinc texture, with occasional bright cleavage-faces which belong to crinoidal fragments. Its colour, owing to the included silicate, is dull olive, and it shows oc- casional small deep green and reddish specks. Its aspect is so waxy, that at a little distance it might bo mistaken for an impure ser- pentine. When examined with the microscope, the flocculont olive-greon silicate is seen to penetrate the mass exactly in the manner of the * Eeport of Geological Survey of Canada, 18(')<). t As the analyses of these specimens by Dr. Hunt have not been published 64 J. W. DAWSON ON PALEOZOIC FOSSILS serpentine in ophiolite, and it has a polariscope appearance approach- ing to that of serpentine ; while greenish by reflected light, it appears reddish when seen in thin slices with transmitted light. It pene- trates the finest pores of crinoids, and at the same time fills the cavi- ties of shells and the colls of corals. The larger fillings of this kind give the deep green spots above mentioned, while the red spots are apparently caused by the partial oxidation of the iron of the mineral. In one shell, apparently a small OrtJwceras or Theca, the dark green fiUiug has cracked in the maimer of Septaria, and the fissures have been filled with carbonate of lime. In some places the mineral has penetrated the pores of shells of Brachiopods or crusts of Trilobites, producing a tubulated appearance not unlike the proper wall of Eozoon. From the characters of the fragments, I should imagine that this limestone is Lower Silurian rather than Cambrian. It affords an excellent instance of the occurrence of hydrous silicates infiltrating organic fragments, and it deserves the attention of collectors having access to the locality. A curious point of coincidence of this lime- stone with some of those in the Lower Silui-ian of Canada is the occurrence of a few bright green specks, probably of apatite or vivianite, giving on a small scale that association of phosphates with hydrous silicates which wo find on the great scale in the Lau- rentian. The above facts I intend to be supplementary to my papers on Eozoon and on the graphite and phosphates of the Laurentian already in England, or in such a manner as to be readily compared with each other, I reproduce them here : — Pole Hill, New Brunswick. Llangwyllog, Wales. Silica 38-93 28-88 18-86 4-25 1-69 0-48 6-91 35-32 22-66 21-42 6-98 1-49 0-67 11-46 Alumina Protoxide of iron Maffneaia Potash Soda Water 10000 10000 In the Llang\vjllog specimen the silicate amounted to three per cent, of the whole, the remainder being carbonate of lime with a very little siliceous sand and fine clay. In the Pole-Hill specimen the silicate amounted to about five per cent., the remainder being limestone with a few quartz grains. It will be seen that these two silicates, evidently deposited from solution in such a manner as to fill the finest organic pores, are remarkably similar in com- position ; and the fact that they closely resemble Hoffmann's mineral found in Bohemian Eozoon, and also the loganite filling the Burgess Eosoon (Quart. Journ. Geol. Soc. vol. xxi. 1865), gives them additional interest. MINERALIZED WITU SILICATES. 65 publishod in tho Journal of this Society, and as illuetrativo more cspecittlly of tho ufHnity of Kozoon with its Hucccssors in t'unctioii, fhc Silurian Stromaloporci'^ and of the abundant occurrence of ser- pentine and other hydrous silicates in association with fossils in the Lower Silurian as well as in tho older Lauren lian. III. Imitative Forms resembling Eozoon. It is easy for inexperienced observers to mistake laminated con- cretions and laminated rocks either for Strontatopora or for Eozoon, and such misapprehensions are not of unfrequent occurrence. As to concretions, it is only necessary to say that these, when they show concentric layers, are deficient altogether in the primary require- ments of lamina) and interspaces ; and under the microscope their structures arc either merely fragmental, as in ordinary argillaceous and calcareous concretions, or they have radiating crystalline fibres like oolitic grains. Laminated rocks, on tho other hand, present alternate layers of different mineral substances, but arc destitute of minute structures, and are either parallel to tho bedding or to the planes of dykes and igneous masses. In the Montreal mountain there are beautiful examples of a banded dolerite in alternate layeis of black pyroxene and white felspar. Those occur at tho junction of the dolerite with the Silurian limestone through which it has been erupted. Laminated gneissose beds also abound in the Laurentian. Still more remarkable examples are afforded by altered ro(-ks having thin calcite bauds, whether arising from deposition oi from vein- segregation. One of these now before mo is a specimen from the collection of Dr. Newberry, and obtained at Gouverneur, St. Law- rence County, Nov; York. It presents thick bands of a j)eculiar granitoid rock containing highly crystalline felspar and mica with grains of serpentine ; these bands are almost a (juartcr of an inch in thickness, and arc separated by interrupted parallel bands of calcite much thinner than tho others. The whole resembles a mag- nified specimen of Eozoon, except in the absence of the connecting chamber- walls and of the characteristic structures. A similar rock has been obtained by Mr. Vennor on the Gatineau : but it is less coarse in texture though equally crystalline, and appears to contain hornblende and pyroxene. These are both Laurentian, and I con- sider it not impossible that they may have been organic ; but they lack the evidence of minute structure, and differ in important details from Eozoon. Another specimen from the Horseshoe Mountain in the Western States (I regret that I have mislaid the name of the gentleman to whom I am indebted for this specimen) is a limestone with perfectly regular and uniform layers of minute rhombohedral crystals of dolomite. The layers vary in distance regularly in the thickness of tho specimen from two millimetres to one, and must have resulted from the alternate deposition in a very regular manner of dolomite and limestone. These are but a few of the examples of imitative structures which might readily bo confounded with Eozoon, or which, if resulting from organic growth, have lost all decisive evidence of the fact. Perhaps still more puzzling imitative forms are those referred to 66 /. W. DAW80N ON PALEOZOIC FOSSILS \* by Hahn, which occur in somo felspars, and wliich I have found in great beauty in certain crystals of orthoclase from Vermont. Thoy are ramifying tubes resembling the canal-systom of Eozoon, and aro evidently a peculiar form of gas-cavities or inrlusions. Similar ap- pear; i ices are, however, often presented by the more minute and mi- croscopic varietiesof graphic granite, in which the little plates might readily be mistaken, in certain sections, for organic tubulation. In the present state of knowledge, it ia perhaps more excusable to mistake such things for organic structures ihan to deny the exist- ence of true organic structures because thoy resemble such forms. Those who have examined moss-agates are familiar with the fact that while some show merely crystals of peroxide of iron or oxide of manganese, others present the forms of Vaucherice or Confervce. So if one were to place side by side some fibres of asbestos, spicules of Teihea, and coniferous wood, preserved, like some from Colorado, as separate white siliceous fibres, they might appear alike ; but, even if thoroughly mixed together, the microscope should be able easily to distinguish them. I have specimens of fossil wood, collected by Hartt in Brazil, which have been mineralized by limonite in such a manner that no one, without microscopic examination, could believe them to be other than fibrous brown • haematite. Such difficulties the micro-geologist must expect to fiud, and by patient observation to overcome. EXPLANATION OF THE PLATES. Plate III. Fig. 1. Vertical section of Sfromafopora from the Niagara formation, showing the larairae and pillars, without supplemental! matter, X 20. 2. Vertical section of Stromatopora from the Corniferous Limestone, showing pillars ramifying and thickened at the ends, and lamines without supplemental matter, x 20. 3. Vertical section of Stromatopora from the Corniferous Limestone, with much supplemental matter, but showing unthickened laminaj at a a, also horizontal canals at i J, X 20. Plate IV. Fig. 4. Vertical section of iS'j'rowrt^'o/jom from the Corniferous Limestone, X 2, showing lines of growth, aa; c, vertical section of part of the same, X 20; rf, surface of lamina, X 20, showing solid and hollow pillars. 6. Portion of lamina of another specimen, X 20, showing large pores and bases of two pillars. 6. Portion of another specimen, X 20, showing hollow and solid pillars and a pore at a. 7. Portion of silicified Stromatopora, weathered, and showing laminae and pillars in relief, X 20. 8. Portion of Stromatopora resting on a tabulate coral and showing acer- vuline cells at base, X 2. 9. Vertical section of Caunopora hudsonica, showing Tertical tube and horizontal canals, X 20 ; o, horizontal section of part of the same, showing canals and canaliculi ; b, vertical section, more magniaed. Plate V. Fig. 10. Horizontal section of Caunopora hudsonica, showing canals radiating from a central tube, X 20. 11. Vertical section of Ccenostroma nodulata, Corniferoua Limestone, show- ing canals and concentric laminae, with much supplemental matter. X 20. 12. Horiaontal section of the same, showing large radiating canals, X 20. MINERALIZED WITH 8ILICATE9, m Discussion. Prof. Duncan expressed his belief that many difierent forms wore united under the one head of Strowatupora, and that the confusion was often due to the mode of mineralization. He called attention to"a Smithia on the table, which, by destructive mineralization, had assumed a deceptive resemblance to IStromatopora. He thought this had been the case in some of Mr. Lonsdale's specimens. The tubules in the lamina) of Stromatojwra certainly had much resem- blance to the tubules of Milhpora. Some of the specimens on the table seemed to have openings like calicos ; as they opened into the coenenchyma they could not be corals. The cross tubules excluded them from Polyzoa. They showed no true supplemental skeleton or nummulino layer like Eozoon, and so he doubted their Forami- niferal character. With regard to the mineralization, he had some years before received specimens of fossils from Canada, which Dr. Dawson's description had recalled to his mind. Mr. CiiAMPERNowNE dcscribcd the tubular structure which he had observed in some of the Stromatoporida) from Devonshire, both in the horizontal and vertical sections, and felt certain that the group contained many dififerent forms. He had never seen Eozoonal struc- ture in the Devonshire fossils. Dr. MuBiE stated that some specimens which he had seen re- sembled the Hoxactinellidoe, and he thought they represented sponges, not precisely Hexactinellids. T Quart Journ . Geol Son Vol XXXV PI III H Olson Kth Mintorn Bros imp STROMATOPORID^, , X20 Quart Journ . Geol Soc .Vol XXXV. PI . IV. A. 4a « 20 * 2 4 i. 6. .^i'£^? (« ♦ * ♦ ^ > 20 7. a. ] ^ 9a. *>.». H OU.iiiliUi STROMATOPURID/F. 2« !n> Mintwii BroB itnp . '^ ^ I I i Quart JoLim . Geo] Sos .Vol . XXXV. PI .V. 10. > j«) \ _ -W ;C* * V'''* > U* X 2 12 X W H Olson lull Mmterri Bros iin T GTROMATOFORID^.