- 4 i) e 5 a r ‘ we i 2 i= m4 iy 7 BA a} Le Laat ANNALS OF THE NEW YORK ACADEMY OF SCIENCES VOLUME XXIV 1914 Editor EDMUND OTIS HOVEY New York Published by the Academy 1914, 1915 THE NEW YORK ACADEMY OF SCIENCES (Lyceum or Naturat History, 1817-1876) Orricers, 1914 * President—Grorert Freperick Kunz Vice-Presidents—CHARLES P. BerkEy, RaymMonp C. Ospurn, CHARLES BASKERVILLE, CLARK \WISSLER Corresponding Secretary—HeEnry EK. Crampron, American Museum Recording Secretary—Epmunp Oris Hovey, American Museum Henry L. Donertry, 60 Wall Street Librarian—Rauen W. Tower, American Museum Treasurer Yditor—EpMUND Otis Hovey, American Museum SECTION OF GEOLOGY AND MINERALOGY Chairman—Ciuartes P. Berkey, Columbia University Secretary—A. B. Pacint, 147 Varick Street SECTION OF BIOLOGY Chairman—RayMonp C. Ospurn, 557 West 124th Street Secretary—Witiiam K. Grecory, American Museum SECTION OF ASTRONOMY, PHYSICS AND CHEMISTRY Chairman—Cuar Es BaskerviILLE, College of the City of New York Secretary—ERNeEstv E. Smiru, 50 Hast 41st Street SECTION OF ANTHROPOLOGY AND PSYCHOLOGY Chairman—Cuark WissterR, American Museum Secretary—Rosert H. Low1e, American Museum CONTENTS OF VOLUME XXIV Page ARMAS TOROS His tency Sina DOG SCV ISIE B Gacy OS Oe CLORG CLR OG CLERCR ORI ere RATE EEE Ee i KOPTTN COTS spare yee ev ayrot stows) octiay/oi he Ne en lewne crete fs aravepetsaty sista ic aS einy siete suns Aa dadhget evaiaresoate li WOmMPemlsmeerspvershayercie re renciere ustetevorenscete parade reieds [age eieve cvelevcls alobercharvis;avensiiecaye Shsichelehate ili DaAtesrolLublication land Hdittons of Brochures... sc ckeeees seen iii PATS CeO Pe UL UTS GT LL OMS repeyen cacy tore te cote ver eretate tayaial se Malloteveleis lors! cna eek ovelisralete late. cicharavers caciets iv ‘The Use of Crinoid Arms in Studies of Phylogeny. By Etvira Woop. (GIEAICY eye DESMO) Vee eacie Ges pus GO OCG CORSO CG CRO AIO RID ERIC TOR aE ERE Cee eas ae 1 INoteston Camarasaurus Cope. By Cs C2 Mook... oe-es cece csc seee 19 The Genesis of Antigorite and Tale. By ALrexis A. JULIEN. (Plate VI).. 23 A Study of the Changes in the Distribution of Temperature in Europe and North America During the Years 1900 to 1909. By HenryxK Arc- TONKS gg ab lidatag enol wale UNIS S Cine aey OR RENCE CUP On ca COL M CIERONCITS CRT DORE) ce RARE ene airs 39 The Genesis of Certain Paleozoic Interbedded Iron Ore Deposits. By VAY MOND > AR TEM TED BARGE (ETaAtes: Vill Xexel)ies os eee eiaiscaces oe ala loss Chimateyandeb volition assy. We Ds VUAT TED Wee) eteescis ose cin circa 171: Development of the Neuraxis in the Domestic Cat to the Stage of Twenty- one Somites. By H. von W. ScHULTE and FREDERICK TILNEY. (Plates ERENT NON) Maerpat Re cvecaeepencn che Pe havea ae Me ater eter owl ara tees ce sobate Wy austen Ne 319 Records of Meetings, 1914. By EpmMuNp Oris HOVEY............---ccees 347 herOrecamizawdonworpliewNCcademys cece aera atsloieer ace eee cee 407 AM rey" OreieTu oka U ClM ATE ENE 6 ke cas Sere CRORE SIE cn: OItEy CE SE REPO ENG I eR oes 407 OTe reO tal COME sree sewers eters Se ecope nee icee ten eek eteca cal oy he wit Svote idcasc dhe cheno 409 FATING TUG E Clin W INANE CIgerspa ete geeeste cy nes a ercltoy el sete areca theta gis oes ices MeLe lobe teeey ose aS 410 Wom SEUGU CLOTS ene patet penetek oror ter atovoicis ot fo Gt orate eet otk eee MSE aa tan ee 413 ESV VARW Sleeper ch artasvonreu renee neret ere Reve ares vareue pecs Ven) tee soe hcp omnia ar ei cies avehaves craven 414 MembershipsWistss 31s) ecember sl 914 yess 5 2.0 scicineio ccs oaie ide aera 421 MITRE CNA eat crencexs crete srate arse MC MEy Se LeTRG tel ete ete: ah ae wr aren ep eile s. Shoe Siatewithe ns wise eens 433, DATES OF PUBLICATION AND EDITIONS OF THE BROCHURES. Edition. Pp. 1-17, 1 May, 1914. 1100 copies. Pp. 19-22, 21 May. 1914. 1150 copies. Pp. 23-38, 25 July, 1914. 1500 copies. Pp. 39-113, 27 June, 1914. 1200 copies. Pp. 115-170, 4 August, 1914. 1250 copies. Pp. 171-318, 18 February, 1915. 1250 copies. Pp. 319-346, 31; March, 1915. 1850 copies. Pp. 347-443, 14 May, 1915. 1000 copies. LIST OF ILLUSTRATIONS Plates I.—Cactocrini, Individual specimens. 11.—Cactocrini, Details of plates. I11.—Cactocrini, Details of plates. IV.—Cactocrini, Details of plates. V.— Cactocrini, Details of plates. VI.—Laminated Chrysotile-Asbestos from Thedford, Canada. VII.—Map showing the Distribution of Clinton Hematites in the United States. VIII.—Fossil Ore. [X.—Fossil Ore—Photomicrographs. X.—O6litie Ore. XI.—Cavernous Consolidation. XII.—Special Consolidations. XIITI.—Special Consolidations. XTV.—Partial Impregnation. XV.—Cherty Limestone Seams. XVI.—Fossil Ore—Photomicrographs. XVII.—Contact Specimens. XVIIT.—Mud Cracks. XIX.—Contact Specimens. XX.—Contact Specimens. XXI.—O56litic Ore—Photomicrographs. XXII.—Sections of Embryos before and after Segmentation. XXIII.—Sections of Embryos with Two and Three Somites. XXIV.—Transverse Section, Embryo of Four Somites. XXV.—Transverse Section, Embryo of Four Somites. XXVI.—Transverse Section, Hmbryo of Four Somites. XXVII.—Reconstruction of Neuraxis of Embryos of Four and fight Somites. XXVIII.—Transverse Section, Embryo of Eight Somites. XXIX.—Transverse Section, Embryo of Eight Somites. XXX.—Transverse Section, Embryo of Hight Somites. XXXI.—Reconstruction of Neuraxis of Embryo of Nine Somites. XXXII.—Reconstruction of Neuraxis of Embryo of Ten Somites. XXXIII.—Reconstruction of Neuraxis of Embryo of Twelve Somites. XXIV.—Reconstruction of Neuraxis of Embryo of Fourteen Somites. XXXV.—Reconstruction of Neuraxis of Embryo of Sixteen Somites. XXXVI.—Reconstruction of Neuraxis of Embryo of Seventeen Somites. XXXVII.—Reconstruction of Neuraxis of Embryo of Nineteen Somites. XXXVIII.—Reconstruction of Neuraxis of Hmbryo of Twenty-one Somites. XXXIX.—Reconstruction of Portions of Neuraxis of Embryos of Fight and Nine Somites. XL.—Reconstruction of Portions of Neuraxis of Embryos of Twelve and Twenty-one Somites. XLI.—Transverse Section of Embryo of Sixteen Somites. iv Text Figures Page Relationship of species of Cactocrinus and Teleiocrinus...... 002200000 srl FRESTOLALONN OM OCMAnASOILIUSs DV) | COMC a aieiersieie ciclclete leis ele oreieice w/e ssieleeyslcl oe « 2 Diagrams of consecutive annual and consecutive monthly means of tem- SSrRIKD ‘en AECL o na cooodooo6 bon dDo goo SG oud ODODE dye aisles letersret sheets 43, Mica eNO, WEIS SIUSSI 55 octinadoo pobdomouboonnMedodG UOoo DOS OD Ee Dodo oC 47 Wineraaleion, asirntsailststss aocadcccocddeuccs cp odbnoou Nn doo cdbooFoDUGouOOdG AT MACROMELONE TL SSO LESSON coe chasieiete cae iere eiavelsiettyalions) mleh easelayenchisvereccislsieenajerereealersiete 47 IM Daxerxoncoenoyo,~ AUST ILATNSHI) (Et Sao dio Goo 010 b bind UG OU Clin OIsIO-G.0 OG OD Odo toi CON AT Wirveraalencm, IESBBEI IE, 3.65 g4¢cucssondbaccbobsoouucod ban CUO ODO UQSINOOUL 48 IMECODLELOT lS O4— 1 OOS eres terete tepeteneteieys eveols. a1 re/ eheilelieVxreusl olo\lelistavelal sic! opeleieile) « 48 MIAICEOPLELO ML gO — MOO Ae acta crakear eer iste yet irene speleyeve lovelci shel ats alctereyetaliereie es cle ekaile 48 Wilreeoyollerorn, TESTES: Ga Gb aaobucoo0 op udhooedodugeoouUoUUoUdcrodoE ONGC 48 Curves of the consecutive means at Bucharest, Odessa, Warsaw, Kazan enol Angee fon CH). Ga motta cio tomo Od Clo cio Jind Ue Ole DOO EL aaOORp SOO Oe see 62 Consecutive temperature curves at Geneva, Aachen, Vestervig, Bodo, ley oeNe Gua Syohmumiikyey tio chooks ouoeadanDOnCoUnoob oso oanoO uC 63 Temperature departures: form che, year ml GOO Me ere ceteusrele¥eice cc lelaieiele a oleie eielere 66 Memperavunrerdepanhunes for the year OOM ier cals cieie ciel -lels ici alelel siecle s 66 Nemperaturerdepartures stor thel year DOOD wesw yaralacy='<) hc «iste isle s) sisicre srieisc 67 Memperatunre.departures! fom the year LOOSE: cs cies siewicisis cece s sieve cvsl oi eele oe 67 Temperature departures tor! they years L904 ss 25 i535 oes ctelsteisic cialn ats eels). oe = 69 Memperatune;departunes: for they year 905. . 2a)... os). sverne clas sisis se elo leuers 69 Remperatunendepartures) for ithe yea L9OG Es. c1cc ses che lernialotel alete te atcvsiclsteleisie le 70 RNemperatunesdeparthurespror tne yieat OOM cia crates iclele ble 38 2 os Gee OOO mace DOING IgE gon Co ole Onan Chieinr sori ats ‘ INTRODUCTION In studying the phylogeny of Paleozoic crinoids, the worker is greatly hampered by the difficulty of obtaining information about the early stages in ontogeny. By the time the young crinoid is sufficiently calcified to be preserved in the fossil state, the calyx has nearly, if not quite, all the plates which are to be present in the adult. This fact is illustrated by a calyx of Batocrinus subequalis only 4 millimeters in height which pos- sesses all the plates of the adult, with the full number of arms and a well-developed tegmen and anal tube. The present paper is concerned with results which were obtained from a study of adult, or of late neanic stages of camerate crinoids. In order to determine the exact amount and character of the varia- tion occurring in the number and arrangement of calyx plates, about one hundred specimens of Batocrinus and Cactocrinus were examined and the position of each plate carefully recorded. From this investigation, it was found that there is very little variation in the plates concerned in the ‘support of the arms. In the specimens of Cactocrinus examined, the only variation in the radial series was found to be due to the presence of an extra arm or one arm less than the normal number for the ray, necessi- tating a greater or smaller number of plates in the calyx. In each case, ‘the plates present followed the normal order for a similar ray or half ray 1 Manuscript received by the Editor, 4 February, 1914. (1) LIBRARY NEW YOR BOTANICA GARDEN 2 ANNALS NEW YORK ACADEMY OF SCIENCES in another species. In Batocrinus, the most frequent variation in the radial series was due to the absence of the first costal, or rarely, to the presence of an extra distichal below the axillary one. A greater amount of variation occurs in the number of interbrachials, as these are simply space fillers, and their number depends mainly on the relative height and width of the cup. The most marked variation was found in the anal area, Of the seventy-five specimens of Batocrinus studied, four had two plates in the second row of the anal series, as in the Actinocrinide, and one of twenty-five specimens of Cactocrinus had three plates in the second row, after the manner of the Batocrinidew. As will be readily seen, this amount of variation furnishes little evidence which can be used for working out phylogeny, except in the most general way. The ornament on the surface of closely related species differs in degree rather than in kind, and hence does not show changes sufficiently definite to furnish satisfactory results. The stems of crinoids often show very definite and well-marked changes from the proximal to the distal portion, but it is so comparatively rare to find the crinoid stems intact for any considerable distance from the calyx that they are only occasionally helpful. A study of crinoid arms has shown that in some genera, at least, more satisfactory results may be obtained. It has long been recognized that the uniserial condition at the base of many arms which later become biserial is reminiscent of ancestral species whose arms were uniserial throughout. Applying the same principle to other characters, it is be- lieved that changes in the form and ornament of the arm are indicatious of changes through which the ancestors of a species have passed and that, taken in connection with other characters, they may be used to determine the phylogeny of the group to which the species belongs. This method of study will obviously be most useful in genera having highly modified arms, and the genus Cactocrinus has furnished the material upon which the present paper is based. No single character can be used alone in working out relationship, and in this study constant reference has been made to the characters of the calyx as well as of the arms. The columns, in nearly all the specimens available for study, were not preserved- In the descriptions of species of Cactocrinus which follow, a reference is given to the original description and to a later full description and fig- ures. Only such facts are added here as have a bearing upon the present investigation, except in the case of new species and Cactocrinus probosci- dalis, which may be used as a standard of comparison for other species and is fully described. WOOD, CRINOID ARMS IN STUDIES OF PHYLOGENY eS) DESCRIPTION OF CACTOCRINUS AND SIX SPECIES OF THE GENUS Cactocrinus Wachsmuth and Springer 1897. Cactocrinus WacHSMUTH and Sprincer, North American Crinoidea Camerata, p. 600. The species included in the genus Cactocrinus were, with few excep- tions, originally described under the genus Actinocrinus, and for the present purpose, the genus is sufficiently defined by enumerating the fea- tures by which it is distinguished from’ Actinocrinus. The most notice- able of such distinguishing characters seems to be the arrangement of the arms, which form a more or less continuous row around the calyx; that is, the interbrachial and interambulacral plates do not meet between the arms as is the case in Actinocrinus. Another distinguishing feature is that, in species having more than four arms to the ray, the third bifurca- tion takes place on the second or third plate above the distichals in Act- nocrinus, while the axillary palmar immediately follows the axillary distichal in Cactocrinus. The two genera are also said to differ in the structure of the pinnules, but the spines on the proximal pinnules of Cactocrinus are represented also on many species of Actinocrinus, as well as on other genera as far removed as Dorycrinus and Hretmocrinus. Hence this cannot be used as a distinctive feature. The type of the genus is Cactocrinus proboscidalis (Hall). Cactocrinus proboscidalis (Hall) Plates plates hess 2 2a) 20 1858. Actinocrinus proboscidalis Hall, Rept. Geol. Sury. Iowa, Vol. 1, pt. 2, p. 584, pl. 10, fig. 13. 1897. Cactocrinus proboscidalis Wachsmuth and Springer, North American Crinoidea Camerata, p. 601, pl. 58, figs. 3, 4, 5, 6, 7 a—d. The calyx of this species has a width slightly greater than its height. There are, as usual in the genus, three basals, five radials and five each of first and second costals. The axillary distichals rest directly upon the second costals, and each of these gives rise to two arms, making four to the ray. The surface of the calyx is ornamented by a node at the center of each plate from which simple carine radiate across the margins and become continuous with the carine from adjacent plates. The entire surface of both cup and arms is covered with extremely fine granules. The arms are long, three or four times the height of the calyx. Be- yond the first palmars, from one to three plates pass entirely across the diameter of the arm, hence the biserial condition is attained early in this 4 ANNALS NEW YORK ACADEMY OF SCIENCES species. In form, the arms are cylindrical at the base, but at a distance from the base varying from the third to the tenth plate, they become flat- tened laterally, and the individual plates are elongated, producing a marked expansion of the arm at about half its height which may be taken to represent the adult stage. Beyond this, the arm tapers gradually to a point, at the same time becoming more nearly circular in cross section than at its middle portion. In the growth of the crinoid arm, new plates are formed at the tip, and these small, nearly cylindrical plates are in an immature condition. Their resemblance in form to the plates near the base of the arm probably indicates that they are passing through a stage comparable to an early stage in the development of the entire arm. That such localized stages occur in other genera has already been pointed out by Grabau.? Immature plates were observed only in Cactocrinus proboscidalis and C. baccatus. Others of the species studied had the arms strongly in- curved and the tips concealed by the matrix, except in the C. multi- brachiatus series where the arms were imperfect at the ends. Kach plate of the arm of (. proboscidalis bears a distinct transverse ridge running the entire width of the plate and situated about one-third of its height from the upper margin. The ornament appears on early plates, at distances from the base differing somewhat in different speci- mens, and continues to the extreme tip of the arm. The proximal pin- nules, as in other species of the genus, bear strong overlapping spines. Column cylindrical. The proximal nodals project but slightly beyond the internodals, but at a distance of about 20 millimeters from the calyx, the nodals have twice the diameter of the internodals and have their margins extended into a thin; knife-like edge. At a distance of 85 milli- meters from the body, the number of intercalated plates has increased until there are seven between successive nodals, and at this point the nodals have blunt margins projecting but little beyond the internodals. -'Tegmen moderately high, covered with numerous nodose plates. Anal tube long and slender. Horizon AND LOCALITY: Lower Burlington limestone, Burlington, Iowa. No. 415, Museum of Comparative Zodlogy collection. Cactocrinus baccatus sp. nov. Plate I, fig. 3; plate IT, figs. 3, 4, 4a, 4b Calyx similar to that of Cactocrinus proboscidalis in form. Basals variable in size, sometimes minute with the calyx resting upon the radials, sometimes as large as in Cactocrinus proboscidalis. 7A. W. GraBau: Amer. Jour, Sci., 4th ser., Vol. 16, pp. 289-300. 1903. WOOD, CRINOID ARMS IN STUDIES OF PHYLOGENY 5 In arrangement of the plates and surface ornament, this species does not differ essentially from C. proboscidalis, there being the usual number of plates, five basals, five each of radials, first and second costals and ten axillary dis- tichals each of which bears two arms, making four to the ray. The arms of the two species are similar in form, being cylindrical at the base, strongly flattened laterally throughout the greater portion of their length, tapering and becoming more nearly cylindrical near the tips. The most characteristic difference is in the ornament of the arms, for while the arm plates of C. proboscidalis bear a single transverse ridge throughout, the arms of C. baccatus have this simple ridge only near the base of the arm. At a distance from the base varying from 6 to 20 millimeters in different specimens and varying to some extent. in different arms of the same specimen, the transverse ridge is broken up into a row of small nodes of which there are five or six on each plate at the greatest diameter of the arm. The column was not preserved with any of the specimens found. HorizoN AND LOCALITY: Lower Burlington limestone, Burlington, Iowa. No. 558, Museum of Comparative Zoology collection. Cactocrinus platybrachiatus, sp. nov. Plate I, fig. 2; plate III, figs. 1, 2, 2a, 2b The basals of the only specimen representing this species are not preserved: Of the plates above the radials, only those of two rays and one interradial area are preserved. Their arrangement seems to be the same as that usual in the genus for five or six armed rays; that is, in one of the half-rays present, the axillary distichal is followed, without intervening plates, by an axillary palmar which gives rise to two arms. The palmar resting on the other axillary face of the distichal bears one arm giving three arms to this half ray. The same arrangement is seen in another half ray, but whether there were two or three arms in the other half of the same ray cannot be determined. There is one small interdistichal, and the formula for the interbrachials is 1, 2, 2, 1. The ornament on the calyx is essentially the same as that of Cactocrinus proboscidalis, but since there are more plates due to the greater number of arms, and a ridge crosses each suture line between the plates, the cost ap- pear more crowded than on the latter species. The node at the center of each plate is also less prominent. The arms are cylindrical near the base but expand rapidly in the median portion. They are flattened laterally but less strongly so than the arms of Cactocrinus baccatus. The form of the arms changes gradually, until in the upper portion they are flattened dorso-ventrally. Their tips are incuryed toward the anal tube and buried in the matrix. The bhiserial condition is at- tained early, only one or two plates at the base passing entirely across the arm. The surface of the arm is smooth at the base. A little later a transverse ridge appears which soon breaks up into a row of nodes like those of Cacto- crinus baccatus. This type of ornament persists for the greater portion of the length of the arm, but by the time the dorso-ventral flattening is established, some of the lateral nodes become confluent, reducing their number until there are but three on each plate, and on some of the latest plates visible, there are 6 ANNALS NEW YORK ACADEMY OF SCIENCES ouly two, a shorter node near the median line of the arm and an elongated one placed laterally. The column is unknown. Horizon AND LOCALITY: Lower Burlington limestone, Purlington, lowa. No. 568, Museum of Comparative Zodlogy collection. Cactocrinus platybrachiatus is distinguished from C. baccatus by the greater number of arms, the dorso-ventral flattening of the arms and the confluence of their surface nodes in the upper part of the arm, probably representing a late stage of development. It is distinguished from C. reticulatus by the many nodes on the arms at the maximum differentia- tion in structure, representing the adult stage, the less strongly flattened arms and the absence of lateral spines on the arms near their tips. Cactocrinus reticulatus (Hall) Plate III, figs. 3, 4, 4a, 4b, 4c 1861. Actinocrinus reticulatus Hall, Description of New Species of Crinoidea, Preliminary notice, p. 2. 1897. Cactocrinus reticulatus Wachsmuth and Springer, North American Crinoidea Camerata, p. 605, pl. 58, figs. 2a, 2b. The arrangement of the calyx plates in this species is somewhat vari- able, owing to the fact that it has sometimes five and sometimes six arms to the ray. Of six specimens selected at random, three had 28 arms, one 27, one 24 and one 22 arms. When there are five arms to the ray, it is always one of the median palmars which is axillary and bears two arms, while with six arms to the ray, both median palmars become axillary and the lateral palmars bear a single arm. On the calyx of large individuals, the primary coste are sometimes bordered by a second series producing a smaller triangle within a larger one. The nodes at the centers of the plates are inconspicuous or sometimes absent. In form, the arms are cylindrical at the base, but soon their plates are elongated as in Cactocrinus proboscidalis (see Plate II, fig. 1). A little later, the plates are curved, forming an arm equal in lateral and dorso- ventral diameters. At a slightly higher point, the arms are flattened dorso-ventrally, and the latest of the exposed plates are strongly flattened in the same direction. ‘The ornament begins on the early plates of the arm as a strong transverse ridge or elongate node near the lateral margin, and a few plates later, a new node appears near the median line of the arm. As growth continues, these two nodes appear on successive plates nearer and nearer to the lateral margin, and when they have receded far enough to leave a plain space near the median line, a new node appears in WOOD, CRINOID ARMS IN STUDIES OF PHYLOGENY q that position. This node continues to increase in size at the same time that the lateral node diminishes. Later, as the arm approaches its dorso- ventrally flattened form, the lateral node disappears, the former median node is elongated into a projecting spine and the newly introduced median node is now a prominent feature of the ornament. The remainder of the arm is incurved and buried in the matrix, hence the character of the latest formed plates is unknown. The column is not preserved on any of the specimens at hand. Horizon AND LocaLity: Lower Burlington limestone, Burlington, Iowa. No. 527, Museum of Comparative Zodlogy collection. Cactocrinus denticulatus Wachsmuth and Springer Plate IV, figs. 1, la, 2, 2a, 2b 1897. Cactocrinus denticulatus Wachsmuth and Springer, North American Crinoidea Camerata, p. 606, pl. 57, figs. 5a, 5b. Cactocrinus denticulatus is closely related to C. reticulatus. The ar- rangement of the calyx plates is the same, except that in C. denticulatus six arms to the ray form a constant feature necessitating a greater num- ber of plates for their support. The ornament is somewhat more elabo- rate from the fact that small nodes are present on the coste of the upper part of the calyx, and these sometimes extend as irregularly placed nodes over the base of the arms. The arms pass through the same series of changes in form as those described for Cactocrinus reticulatus, except that the early condition with elongate plates is not present, and the successive changes up to the dorso- ventral flattening appear at a relatively earlier period in the development of the arm than in the preceding species. The latest exposed plates of the arm have a more extreme form than the corresponding plates of C. reticulatus. ‘They are more strongly flattened dorso-ventrally, have longer spines, and the nodes are so high and pointed that, in some specimens, they might almost be called spines. Still further differences appear in the lateral spines, which are often alternately longer and shorter, and when this is the case, the nodes also alternate in size, the larger node occupying the plate with the shorter spine. This alternation in size of the spines and nodes is not perfectly regular, but it is a pronounced tendency which manifests itself to some extent on all the specimens studied. The nodes are not, as in the preceding species, close to the median line but have receded to some distance from it. The column is missing from all the specimens studied. Horizon AND LOocALITY: Lower Burlington limestone, Burlington, Iowa. No. 534, Museum of Comparative Zodlogy collection. 8 ANNALS NEW YORK ACADEMY OF SCIENCES Cactocrinus opusculus (TIlall) Plate I, fig. 4; plate IV, figs. 3, 4, 4a, 4b, 4c, 4d 1860. Actinocrinus opusculus Hall, Suppl. Geol. Rept. Iowa, see description of 0) 48 1897. Cactocrinus opusculus Wachsmuth and Springer, North American Crinoi- dea Camerata, p. 607, pl. 56, figs. 5a, 5b. The calyx of this species bears a close resemblance to that of Cacto- crinus reticulatus, except in the greater number of plates necessary for the support of six arms to the ray, which is the normal number for C. opusculus. The surface of the calyx is ornamented by a single node at the center of each plate with connecting carine, as in C. reticulatus. At the base, the arms are, as usual in the genus, cylindrical and smooth, but they soon become flattened laterally, as in C. proboscidalis, and in retarded specimens bear the strong transverse ridge on the arm plates characteristic of that species. One specimen retains the C. proboscidalis type of ornament for 25 or 30 plates, while in accelerated individuals this condition is represented by only 2 or 3 plates, or may be absent altogether. In average individuals, at a distance from the base varying from the tenth to the twentieth plate, a slight angulation appears on the arm near the median line. This slight elevation increases in size on succeeding plates until it forms a distinct node. At the same time, it recedes far- ther and farther from the median line. The lateral node continues to increase in prominence until it becomes a distinct spine, and accompany- ing this change in the node, and partly in consequence of it, the form of the arm changes until it is strongly flattened dorso-ventrally. Mean- while, another row of nodes has come in close to the median line on each side, as in Cactocrinus reticulatus. At the highest point observable the arm is strongly flattened dorso-ventrally with a row of spines along each lateral margin and a row of nodes on each side of the median line. From the time the angulation appears until it develops into a distinet node, the surface of the plates is distinctly corrugated, although the strength of the corrugation varies greatly in different specimens. The specimen figured in Plate I, fig. 4, and Plate IV, fig. 3, is a highly accel- erated individual showing all the characters at the acme of their develop- ment. HorIzoN AND LOCALITY: Lower Burlington limestone, Burlington, Iowa. No. 523, Museum of Comparative Zodlogy collection. WOOD, CRINOID ARMS IN STUDIES OF PHYLOGENY 9 COMPARISON OF THE PRECEDING SIX SPECIES A study of the first five of the species just described shows a series of gradations in structural characters which is here interpreted to mean that they form a continuous phylogenetic series in which the tendency of evolution has been from the simpler forms to the more complex. Throughout ‘the comparisons which follow, changes in structural fea- tures are assumed to represent stages in development, and the complete series of such changes to express the evolution of the arm as a whole. The calyces from Cactocrinus proboscidalis to C. denticulatus show a progressive increase in the number of features to be considered. he greater number of calyx plates is due to the increase in the number of arms developed, from four in C. proboscidalis to six in C. denticulatus. The elaboration of surface features is expressed in additional carinze and fine nodes covering them in C. denticulatus. The arms furnish more conclusive evidence of relationship. In study- ing the arms of crinoids, we have to consider several distinct characters, such as the stage at which the biserial condition is introduced, the form of the arm as expressed in its transverse section, the thickness of the in- dividual plates and the surface features or ornament of the arms. Each of these characters may develop at a different rate of evolution in differ- ent specimens or even in different arms of the same specimen, but in the same phyletic series new features for each will appear in the same order but not, as already stated, necessarily at the same time. For example, we may find in one arm of Cactocrinus denticulatus the median row of nodes well developed on the twenty-fifth plate, while in another arm they are not distinct until the thirty-fifth. The lateral row of nodes may be de- veloped on the sixth plate or not until the fifteenth plate, but the median row never appears before the lateral row. Comparing the arms of the five species in detail, we find that Cacto- crinus proboscidalis has a laterally flattened arm with simple transverse ridge. (C. baccatus retains the same form and, according to the interpre- tation of the facts here given, passes through the same early stages as its ancestor, C. proboscidalis ; that is, first cylindrical, then laterally flattened with a transverse ridge on each plate, but this species goes a step farther in the breaking up of the transverse ridge into a row of nodes. In both these species, the arms taper to a point. The arms of CO. platybrachiatus pass through the same early stages as its ancestors, repeating the trans- verse ridged stage of C. proboscidalis, the nodose stage of C. baccatus and adding a feature of its own in the confluence of the nodes at a late stage of development. In form, the arms present entirely new features in their 10 ANNALS NEW YORK ACADEMY OF SCIENCES expansion near the point of curvature and their dorso-ventral flattening. In the arms of C. reticulatus, the stage of the cylindrical smooth arm is followed by one in which each plate bears a short prominent ridge near the lateral margin which is believed to represent the confluent nodes of its an- cestor, now reduced to one elongate node, or short ridge, placed close to the lateral edge of the plate. A few plates later,a small node appears near the median line. This stage is represented on Plate ITT, fig. 3. In successive plates of the arm, these two nodes appear nearer and nearer to the lateral margin until a plain space is left into which a new row of nodes is intro- duced near the median line. The median nodes increase in strength, while the outer ones diminish in size as they recede toward the lateral margin until they disappear. Meanwhile, the form of the arm has changed, be- coming flattened dorso-ventrally so that the former median node occupies the lateral margin and is elongated into a spine. The line of nodes of latest origin remains near the median line. These changes are illustrated on Plate III, figs. 3,4, 4a-4c. It thus appears that, in this group at least, new features arise near the median line of the arm and on successive plates seem to move laterally until they disappear and are replaced by features of later origin. This fact has led to the conclusion, stated above, that the elongate node on early plates of Cactocrinus reticulatus repre- sents confluent nodes in a late stage of their evolution and soon to disap- pear, rather than that it has any relation to the transverse ridge present in varly stages of its predecessor, C. platybrachiatus. The smooth space thus left on the median half of the plate, in the preceding species, becomes a field for the introduction of new features which appear successively as lines of nodes. In the arm of C. denticulatus, the earliest stage to appear is that with an elongate node and a shorter one, both near the lateral margin of the plate. ‘The ancestral features are, in this species, somewhat obscured by the presence, on early plates, of nodes which are the continuation over the base of the arms of the irregularly placed nodes present on the calyx of this species. They constitute a feature of later origin quite distinet from the two nodes near the lateral margin of thé arm plates. These irregularly placed nodes are present only on early plates of the arm, and by the time the eighth plate is reached, they have disappeared, as shown on Plate IV, fig. 1. The two nodes remaining after the disappearance of the irregularly placed nodes are, in the specimen figured, sharply pointed, but this is not a constant feature for the species. Beyond this point, the evolution of the arm for the greater portion of its length is the same as that described for C. reticulatus, except that it is more accelerated, new features appearing at an earlier period than in the latter species. The WOOD, CRINOID ARMS IN STUDIES OF PHYLOGENY 11 final stage in the evolution of the arm of C. denticulatus is more extreme than that of C. reticulatus in that the arms are more strongly flattened, the spines are longer and there is an alternation in the size of both spines and nodes. The lateral movement of surface features is further illus- trated in this species by the fact that on the latest plates observable, the line of nodes last introduced does not remain near the median line as on Q. reticulatus but has receded to some distance from it. This series of five species of crinoids appears to constitute an excellent illustration of the principle of recapitulation, each member repeating the life history of its ancestor until, in the later members, early stages are crowded out of the ontogeny to be replaced by characters of later origin. Cactocrinus opusculus, the sixth of the species described above, bears a strong general resemblance to C. reticulatus, but it has always six arms to the ray and consequently more plates in the calyx. The arms at the latest stage observable are closely similar to those of C. reticulatus at the same stage, but they have arrived at this condition along a different path from that traversed by the latter species, as shown by a comparison of the figures on Plate IV, fig. 3, with those on Plate ITI, fig. 3. The early stages lack the strong lateral node seen on C. reticulatus, and there is no indica- tion in the ontogeny of the species that it has passed through the stage with rows of nodes present on C. baccatus and C. platybrachiatus. C. opusculus seems to have been descended from C. proboscidalis but as a lateral branch, following a different line of evolution from that of the C. reticulatus series. The resemblance between the final stages in the arms of C. opusculus and C. reticulatus may be considered a case of parallelism. Another line of evolution from C. proboscidalis, divergent from that of the reticulatus series, is represented by C. clarus. This species closely resembles C. proboscidalis in the calyx and in the strong lateral flatten- ing of the arms, but it is a much larger species and has five or six arms to the ray. The arms have on their early plates a transverse ridge which is strong near the lateral margin and is faint or absent near the median line. On successive plates this ridge becomes shorter and shorter, 1. e., apparently moves laterally on the arm like the nodes of C. reticulatus until, between the thirtieth and fiftieth plates, it disappears altogether, and the arm plates are smooth. These facts are here interpreted to mean that C. clarus is descended from C. proboscidalis but diverges from other lines of descent in the direction of loss of the ornamental feature repre- sented by the transverse ridge. Having followed certain lines of descent from Cactocrinus probosci- dalis, it would be interesting to trace its ancestry, but I have, as yet, seen 12 ANNALS NEW YORK ACADEMY OF SCIENCES no specimen which seems to fulfil all the requirements for such an an- cestor. We may, however, reasonably infer what were some of its char- acteristics. The arrangement of the calyx plates was probably the same as that of C. proboscidalis, and the plates were nodose, either with or without connecting carine. The arms were cylindrical, smooth, tapering at the tips and the biserial condition was attained late, 7. e., more than two or three plates passed entirely across the diameter of the arm. Such an ancestor would be expected to occur in strata older than those con- taining C. proboscidalis, and we should naturally look for it in the Kin- derhook, but the species of Cactocrinus recorded from the Kinderhook, C. nodobrachiatus, C. ornatissimus and C. arnoldi, have ornamented arms of a type quite different from those of C. proboscidalis or any of its descendants. This indicates that the ancestor of Cactocrinus must have lived at a period considerably earlier than the Kinderhook. Cactocrinus thetis, of the Lower Burlington, has arms which in form and surface are like those of the hypothetical ancestor of C. proboscidalis, but they are biserial almost from their point of origin, and there are six arms to the ray, while C. proboscidalis has but four. C. thetis was prob- ably descended from the same ancestor as C. proboscidalis, but while the latter has developed in the direction of surface ornament and modifica- tion in the form of the arms, (. thetis has been retarded in surface orna- ment and has advanced in the direction of number of arms and in the early attainment of the biserial condition. The two species represent divergent lines of evolution. Another species of Cactocrinus which probably occupies a relation to C. proboscidalis similar to that of C. thetis is C. thalia. The latter spe- cies has long, slender, cylindrical, smooth arms, only four to the ray, but it does not seem to be an ancestor of C. proboscidalis, since it is a larger species with longer and more slender arms which become biserial at an early stage. CACTOCRINUS MULTIBRACHIATUS SERIES Cactocrinus multibrachiatus (Hall) 1858. Actinocrinus multibrachiatus Hall, Rept. Geol. Sury. Iowa, p. 580, pl. 10, fig. 10. 1897. Cactocrinus multibrachiatus Wachsmuth and Springer, North American Crinoidea Camerata, p. 617, pl. 56, figs. 6, 7; pl. 58, fig. 8. The calyx of this species resembles that of C. proboscidalis except in the greater number of plates necessary for the support of eight arms to the ray, this being the normal number for the species, although a smaller WOOD, CRINOID ARMS IN STUDIES OF PHYLOGENY aR number is frequently present. The post-palmars, palmars and axillary distichals rest one upon another without intervening plates, following the usual plan in the genus Cactocrinus. ‘The nodes at the centers of the plates are not prominent, and the costw are simple, except on the lower half of the radials, where there are sometimes two or three passing to the basals. The arms are long and slender, tapering very gradually to the tips and but slightly incurved. They are cylindrical for a distance of about twenty to thirty plates from the base and then become somewhat flattened dorso- ventrally, developing an obtuse angulation along the lateral margin. At about half their length, the arms are somewhat expanded laterally, and at this point or a little higher, they develop a narrow transverse ridge close to the upper margin of each plate. These ridges give the arm an appearance of being serrated along its lateral margin with each plate slightly inset above its predecessor. The arm of C. multibrachiatus is well represented by the figures of C. cwlatus spinotentaculus on Plate V, fig. 1, except that it is all on a smaller scale. HORIZON AND LOCALITY: Lower Burlington, Burlington, Iowa. No. 548, Mu- seum of Comparative Zodlogy collection. Cactocrinus ceelatus var. spinotentaculus (Hall) Plate V,. figs: 1, 2, 2a 1860. Actinocrinus spinotentaculus Hall, Suppl. Geol. Rept. Iowa, .p. 86. 1897. Cactocrinus caelatus var. spinotentaculus Wachsmuth and Springer, North American Crinoidea Camerata, p. 619, pl. 59, fig. 10. This species is closely similar to the preceding, except that it is much larger and the calyx is proportionally higher. The proportion of height to width in C. cwlatus var. spinotentaculus is about 1: 1'/, as compared with 1: 114 in C. multibrachiatus. In arrangement of plates and surface ornament, the two species are the same. The arms are eight to the ray and so similar to those of C. multibrachiatus that the same drawing serves to represent the character- istics of both, keeping in mind the fact that the present species is more than twice the size of C. multibrachiatus, and the corrugations of the surface, in common with other features, are much coarser. Hor1IzoN AND LOCALITY : Lower Burlington, Burlington, Iowa. No. 552, Mu- seum of Comparative Zodlogy collection. ; 14. ANNALS NEW YORK ACADEMY OF SOIENCES Cactocrinus limabrachiatus (Hal!) Plate V, figs. 3, 4, 4a, 4b 1861. Actinocrinus limabrachiatus Hall, Description of New Species of Crinoi- dea, Preliminary notice, p. 2. 1897. Cactocrinus limabrachiatus Wachsmuth and Springer, North American Crinoidea Camerata, p. 608, pl. 58, figs. 9, 10a, 10b. The arrangement of plates in the calyx of this species is the same as that already described for species having six arms to the ray. The sur- face is highly ornamented with strong nodes and carine, which, on the larger calyces, are of two series. The carine leading to the arms are much stronger than the others. | The arms of each ray are grouped together, suggesting the arrange- ment in Actinocrinus, although the spaces between the rays are still nar- row. ‘The arms are long and slender, cylindrical at the base, but they soon become flattened dorso-ventrally. They are slightly expanded at a distance of about half their length from the base and taper very gradually to the tips. The biserial condition is reached late in the development of the arm, there being from four to seven plates at the base which pass entirely across its diameter. At a distance of from 5 to 8 millimeters from the base, varying on different arms, each plate is ornamented by a projecting transverse ridge near its upper margin, and the surface is coy- ered by strong vertical corrugations. This type of ornament persists to the tip of the arm, while its form changes from cylindrical at the base to strongly flattened above, as shown by the transverse sections, Plate V, figs. 4, 4a, 4b. HorizoN AND LocALITY: Lower Burlington, Burlington, Iowa. No. 528, Mu- seum of Comparative Zodlogy collection. COMPARISON OF THE PRECEDING THREE SPECIES A comparison of Cactocrinus multibrachiatus and C. celatus var. spinotentaculus shows the relationship between the two to be so close that they might be considered the same species, were it not that in a large series of specimens, C. multibrachiatus shows all the characters of an adult individual, while it is only about half the size of C. cwlatus var. spinotentaculus.. The proportionally much higher calyx of the latter is also a distinctive feature. The arms of the two species pass through the same structural changes, which are interpreted as stages of development and are closely similar, except that in C. ca@latus var. spinotentaculus, they are biserial nearer the base and are larger. WOOD, CRINOID ARMS IN STUDIES OF PHYLOGENY 15 It seems reasonable to assume that C. cwlatus var. spinotentaculus is a direct descendant from C. multibrachiatus, differing but little from the latter, its immediate ancestor. I have not seen the arms of Cactocrinus celatus,; hence comparisons are made with its variety, of which well-pre- served material is available. Cactocrinus limabrachiatus resembles C. multibrachiatus in the form of the calyx and in the changes in form through which the arms pass. The ornament on the surface of the calyx and the arms of the former species is similar in kind to that of the latter, but is more extreme, the cost being stronger and more numerous and the corrugations of the arms coarser. In this respect, it is more highly differentiated than C. multibrachiatus, but in the number of arms and in the late attainment of the biserial condition it is more primitive. My interpretation of these facts would be that both are descended from a common ancestor but rep- resent divergent lines of evolution, Cactocrinus limabrachiatus having developed in the direction of a high degree of surface ornament, while C. multibrachiatus has advanced in number of arms and earlier develop- ment of biserial plates. RELATION OF CACTOCRINUS TO T'ELEIOCRINUS The derivation of the genus Teleiocrinus from Cactocrinus has already been suggested by Wachsmuth and Springer,’ and it is interesting to note that the development of the arms confirms the evidence derived from the calyx and the mode of branching of the arms. In Teleiocrinus umbrosus (Hall), the type of the genus, the arms are more numerous and more slender than those of Cactocrinus celatus var. spinotentaculus. Their slenderness is perhaps due to their greater num- ber, both on account of economy of material and the crowding due to fre- quent branching near the base. The method of branching in T'eleiocrinus follows the Cactocrinus plan, each axillary being succeeded by another without intervening plates of the same order until there are fifteen or sixteen arms to the ray. In form, the arms are cylindrical at the base, or in accelerated individuals somewhat flattened even at this point. Higher up on the arms, they become strongly flattened dorso-ventrally and de- velop a narrow transverse ridge near their upper margins. The corruga- tions on the surface of the arms are much finer than those of Cactocrinus celatus var. spinotentaculus, as might be expected from the more delicate structure of the whole arm. The arm of Teleiocrinus umbrosus is figured on Plate V, figs. 5, 5a. A comparison of these figures with those of 27 3 WACHSMUTH and SprinGer: North American Crinoidea Camerata, p. 627. Cam- bridge, 1897. 16 ANNALS NEW YORK ACADEMY OF SCIENCES Cactocrinus calatus var. spinotentaculus on the same plate shows the close similarity between the arms of the two species throughout their entire development. T'eleiocrinus umbrosus seems to have been developed from Cactocrinus cvlatus var. spinotentaculus by an increase in the size and thickness of the calyx plates and a more frequent branching of the arms near their base, which produced the expanded rim of the calyx char- acteristic of Teleiocrinus. Telerocrimus wiabrosys t G& coclatus ; : SPIMOECTILEOC UTS C denticu/atius ie C opusculus Creticulatus | C wrultibrachiotus Ss platybra chralts | C. Ninabrachia- C. thetys C haccatus Pele | C. clarus G: prohoscidalis - Cthalva Rg S Ww S is ce iw R N ae) 8 ma) SQ Be Smooth arvim7ed anceslor Fic. 1.—Relationship of species of Cactocrinus and Teleioerinus Televocrinus altha@a (Hall) is represented in the collection of the Mu- seum of Comparative Zodlogy by only one specimen which preserves the arms. From this, it appears that the numerous arms are flattened dorso- ventrally at the base but become larger and cylindrical in form above. This suggests that we have in Teleiocrinus althea an actual advance in evolution expressed in a simplification of form rather than in greater poe The general relations of the species mentioned above are expressed in diagrammatic form in Fig. 1. WOOD, CRINOID ARMS IN STUDIES OF PHYLOGENY 17 CONCLUSION The considerations presented in the above paper constitute only a be- ginning in a line of investigation which seems to promise good results, if followed out in genera which have highly modified arms. From the proximal to the distal portion of the arm, we find a series of changes in structural features which succeed one another in a definite order. These changes may be interpreted as stages in development, each individual re- peating the stages present in its immediate ancestor and adding, in the distal portion, new characters of its own until the number of characters becomes too great for representation in the life history of a single organ- ism, and certain characters, usually the earlier ones, are greatly abbre- viated or are omitted from the ontogeny of highly modified descendants. When thus interpreted, the arms of crinoids furnish evidence from which the phylogenetic relations of different species and genera can be inferred. With the attention once drawn to the subject, it will probably be found that the number of genera which may be studied by this method and the degree of modification existing are greater than would appear at first thought. I PLATE XXIV, = RR Ot a NALD RS Sim, 6888 eer pen on ea sa VOLUME SOL. ACAD. ANNALS N. Y. | Poy Pe peor” teers ane ei ees 2 » WOT iting, + amy wet J ae” ; as SS tanlf E. ROWE FROM PHOTOGRAPHS BY BOSTON HELIOTYPE CO., ANNALS N. Y. Acan. Scr. VoLuME XXIV, PLATE II Bo @ * )* 2b a 4b ELVIRA WOOD, DEL. ANNALS N. Y¥. ACAD. SCI. VoLumeE XXIV, PrareE III 96 ELVIRA WOOD, DEL. .“s ANNALS N. Y. Acap. Sci. ELVIRA WOOD, DEL. Votume XXIV, PLATE IV VOLUME XXIV, “PLATE V INGAD. OCI. Me ANNALS N. ELVIRA WOOD, DEL. ’ PUBLICATIONS OF THE NEW YORK ACADEMY OF SCIENCES (Lyceum or NaturaL History, 1817-1876) The publications of the Academy consist of two series, viz. : (1) The Annals (octavo series), established in 1823, contain the scientific contributions and reports of researches, together with the rec- ords of meetings and similar matter. A volume of the Annals coincides in general with the calendar year and is sold at the uniform price of three dollars per volume. The articles composing the volume are printed separately, each in its own cover, and are distributed in bundles on an average of three per year. The price of the separate articles depends upon their length and the number of illus- _ trations, and may be learned upon application to the Librarian of the Academy. The author receives his separates as soon as his paper has been printed, the date of issue appearing above the title of each paper. (2) The Memoirs (quarto series), established in 1895, are issued at irregular intervals. It is intended that each volume shall be devoted to monographs relating to some particular department of Science. Volume I is devoted to Astronomical Memoirs, Volume II to Zodlogical Memoirs, ete. The price is one dollar per part as issued. All publications are sent free to Fellows and Active Members. The Annals are sent to Honorary and Corresponding Members desiring them. Subscriptions and inquiries concerning current and back numbers of any of the publications of the Academy should be addressed to THE LIBRARIAN, | New York Academy of Sciences, care of American Museum of Natural History, New York, N. Y. Pas eee et! a | AAR S582" A : rae « ~a ge ap ' \ See rom s q _ ANNALS OF THE NEW YORK ACADEMY OF SCIENCES Vol. XXIV, pp. 19-22 Editor, Epmunp Otis Hovey NOTES ON CAMARASAURUS COPE aise 818g : C. C. Moox “~ NEW YORK PUBLISHED BY THE ACADEMY ®1 May, 1914 , THE NEW YORK ACADEMY OF SCIENCES (Lyceum or Naturat History, 1817-1876) OFFICERS, 1914 < President—Gnrorcr FrepERICK Kunz, 601 West 110th Street Vice-Presidents—CHARLES P. Berkey, RAyMoNnD C. OsBuRN, CHARLES BASKERVILLE, CLARK WISSLER Corresponding Secretary—HeEnry EH. Crampron, American “Museum Recording Secretary—Epmunp Otis Hovey, American Museum Treasurer—HeEnry L. DoneErty, 60 Wall Street Tibrarian—Ratrpu W. Tower, American Museum Editor—Epmunp Otis Hovey, American Museum SECTION OF GEOLOGY AND MINERALOGY Chairman—CHARLES P. Berkey, Columbia University Secretary—A. B. Pacrni, 147 Varick Street SECTION OF BIOLOGY Chairman—Raymonp C.. Osspurn, 557 West 124th Street Secretary—WiLu1aM K. Grecory, American Museum SECTION OF ASTRONOMY, PHYSICS AND CHEMISTRY Chairman—CuHariEs BAsKERVILLE, College of the City of New York Secretary—Ernest E. Smiru, 50 East 41st Street SECTION OF ANTHROPOLOGY AND PSYCHOLOGY Chairman—C.Lark WissLEeR, American Museum Secretary—Rosert H. Lowir, American Museum _ The sessions of the Academy are held on Monday evenings at 8:15 o'clock from October to May, inclusive, at the American Museum of Natural History, 77th Street and Central Park, West. — ~ [ANNALS N. Y. Acap. Scr., Vol. XXIV, pp. 19-22. 21 May, 1914] NOTES ON CAMARASAURUS COPE By C. C. Moox (Read before the Academy, 9 March, 1914) CONTENTS Page Origin ve dESCLIp MOM eters erasers a oelostereievoletei ereleloieve oleisle ee sree svsials ayers lel eve 19 IEG HONS TES 5 Soe ob Ob OOO COI DID REE ODO b.ceNe Eoin OIG oI cO CO CICS Oran ares 21 COPE SMECSLOLAIGLOMAv st ntevotare che eh ottrstencreceicco) ober nicte ay ole syeneloley dccsve ei ivdiele/enatar eis dtele cts 22 ORIGINAL DESCRIPTION Camarasaurus was originally described by Edward D. Cope in “‘Pale- ontological Bulletin 25,” published August 23, 1877. The type species is Camarasaurus supremus, and the type specimen consists of a cervical, three dorsal and four caudal vertebre. ‘These bones were found near Canyon City, Colorado, and in the same quarry a considerable number of bones were excavated, belonging to three or more individuals. These bones were more or less associated with the type and it is impossible to say which belongs to one individual and which to another. The vertebre of the original type may not all belong to the same individual. The various remains are of the same general character and there need be no hesitation in referring them to the same genus and species. Some of these later bones were described in a subsequent paper in the American Naturalist for February, 1878, and figures of vertebra, scapula and pubis were given. All of these remains together now constitute numbers 5760, 5760’, 5761, 5761’, 5761”, 5761a, of the collections of the American Mu- seum of Natural History. The original description by Cope confounds to some extent the generic characters of Camarasaurus with the characters of the Sauropoda as a whole. The hollow centra, and lightly built, laminated neural arches and spines are possessed by all the Sauropoda, some members of the group possessing the lightening structures to a much greater degree than does Camarasaurus. The general characters of Camarasaurus, without giving detailed de- scriptions, are as follows: Cervicals: Number probably thirteen, of moderate length, of considerable height, with spines double, without a median tubercle. (19) 20 ANNALS NEW YORK ACADEMY OF SCIENCES Dorsals: In the restoration made by Cope the number of dorsals was placed at twenty. Later the series was studied at the American Museum, and a com- posite column was made up by placing together vertebrie showing progressive fore-and-aft characters. At this time the number was estimated to be four- teen, of which thirteen were actually represented, dorsal two being absent. In the fall of 1913, opportunity was given the present writer by Professor Henry Fairfield Osborn to restudy these vertebrie in preparation for his monograph on the Sauropoda. It was then found that by the elimination of duplicate bones the number is probably ten. RELATIONSHIPS The close similarity of Camarasaurus with Morosaurus has long been considered ground for placing the two genera in the same family. At the present time, it appears that this similarity is close enough to force the conclusion that the two animals belong to the same genus. Among the characters common to Camarasaurus and Morosaurus, the following may be mentioned : 1. Centra of dorsals increasing gradually in opisthocelianism from the posterior to the anterior region. 2. Principal laminze supporting the transverse processes strong, with little development of accessory lamin. 3. Spines low and broad, with only one cavity of any importance on their sides. 4. Caudals short, with inferior surfaces of centra convex in transverse direction. 5. Scapulie short, greatly expanded at both proximal and distal ends. 6. Humerus short and stout, index of maximum length into minimum cir- cumference about .440. 7. Ulna slightly twisted at the distal end. 8. Femur very stout, index about .440. Ratio of length of femur to length of humerus about .600. 9. Metacarpals long and slender. 10. Sacral spines low and broad. 11. Ischium slender, tapering distally. The only characters in which the two forms differ are those which may be taken as individual variations or specific characters, such as size, posi- tion of capitular rib facets on anterior dorsals, presence or absence of a median tubercle between the two spines of the anterior vertebre, or slight differences in the laminar supports of the transverse processes. It is concluded, therefore, that Camarasaurus and Morosaurus are generically identical, and as Camarasaurus has a priority of about one month, the species now under Morosaurus should be referred to the former genus. MOOK, NOTES ON CAMARASAURUS COPE About 1/100 natural size Restoration of Camarasaurus by Cope. FIGuRE 1. The position of the fore limb in relation to the vertebral column was not indicated by Cope if De ANNALS NEW YORK ACADEMY OF SCIENCES Corr’s RESTORATION A life-size restoration of Camarasaurus was made by Dr. John A. Ryder under the direction of Professor Cope about 1878, parts of several individuals being assembled to make a composite individual. The material on which the restoration of the skull was based was very incomplete, only the posterior portion of the cranium and the anterior portion of the mandibles being represented. The restoration of the skull was, therefore, almost entirely hypothetical. The teeth were restored as of carnivorous rather than herbivorous type, and were placed along the sides of the jaws instead of in the front as is now known to be the case in the Sauropoda. The teeth extend posteriorly behind the orbit, some of them even appearing to be rooted in the jugal bone. The cervical and dorsal vertebrae are not distinctly separated in the restoration, nor are the dorsal and sacral. No ribs are represented. The cervical series as restored contains ten or twelve vertebre, no atlas being represented. The dorsal series contains sixteen, seventeen or nineteen vertebre, according to the interpretation of vertebra eleven and twelve as dorsals or cervicals, and vertebra twenty-nine as dorsal or sacral. Sixty caudals are present in the restoration. According to our present knowl- edge of Camarasaurus, the number of cervicals should be twelve or thir- teen, the number of dorsals ten, of sacrals five, while the number of cau- dals is doubtful. In the restoration, there are too many anterior caudals and too few small distal ones. The bones of the fore-limb are too long in the restoration. Four hypo- thetical carpal bones are represented. The phalangeal formula of the restoration is 4, 5, 5, 5, 5. The ischium is represented as slightly ex- panded at the distal end as in Brontosaurus, instead of tapering slightly as it does in the type. The tibia and fibula are each about seven inches longer than the actual bones. Three tarsal bones, of which at least one is hypothetical, are represented. The phalangeal formula as restored is %, 3, 35 5, 4. It is interesting to observe that, at this early date, Professor Cope con- cluded that the Sauropoda walked upright, instead of crawling, as was contended a few years ago by Tornier and others, and denied by Matthew and Holland. PUBLICATIONS OF THE NEW YORK ACADEMY OF SCIENCES (Lyceum or NATURAL History, 1817-1876) “The publications of the Academy consist of two series, viz.: _(1) The Annals (octavo series), established in 1823, contain the scientific contributions and reports of researches, together with the rec- ords of meetings and similar matter. A volume of the Annals coincides in general with the calendar year _ and issold at the uniform price of three dollars per volume. The articles composing the volume are printed separately, each in its own cover, and are distributed.in bundles on an ».zrage of three per year. The price of the separate articles depends upon their length and the number of illus- trations, and may be learned upon application to the Librarian of the Academy. The author receives his separates as soon as his paper has been printed, the date of issue appearing above the title of each paper. (2) The Memoirs (quarto series), established in 1895, are issued at irregular intervals. It is intended that each volume shall be devoted to monographs relating to some particular department of Science. Volume I is devoted to Astronomical Memoirs, Volume II to Zodlogical Memoirs, ete. The price is one dollar per part as issued. All publications are sent free to Fellows and Active Members. The Annals are sent to Honorary and Corresponding Members desiring them. Subscriptions and inquiries concerning current and back numbers of : any of the publications of the Academy should be addressed to THE LIBRARIAN, New York Academy of Sciences, care of American Museum of Natural History, New York, N. Y. ANNALS OF THE NEW YORK ACADEMY OF SCIENCES Vol. XXIV, pp. 23-38, pl. VI Editor, Epmunp Otis Hovey THE GENESIS OF ANTIGORITE AND TALC BY ALEXIS A. JULIEN NEW YORK PUBLISHED BY THE ACADEMY 25 JuLy, 1914 THE NEW YORK ACADEMY OF SCIENCES (Lyceum or Natural History, 1817-1876) Orriorrs, 1914. President—Grorar FREDERICK Kunz, 601 West 110th Street Vice-Presidents—CHARLES P. Berkey, RayMonD C. OsBuRN, : ‘ CHARLES BASKERVILLE, CLARK WISSLER Corresponding Secretary—HeEnry E. Crampton, American Museum Recording Secretary—EpmunpD OtTIs Hovry, American Museum Treasurer—HENRY L. Donerty, 60 Wall Street TAbrarian—Rateu W. Tower, American Museum Editor—EpuuNb OT1s Hovey, American Museum: SECTION OF GEOLOGY AND MINERALOGY - Chairman—Cuar.eEs P. Berxey, Columbia University Secretary—A. B. Pacin1, 147 Varick Street SECTION OF BIOLOGY Chairman—RaymMonpD C. Ossurn, 557 West 124th Street Secretary—Wi.u1AM K. Grecory, American Museum SECTION OF ASTRONOMY, PHYSICS AND CHEMISTRY Chairman—CHARLES BASKERVILLE, College of the City of New York Secretary—ERNEsT E. SMiTH, 50 East 41st Street SECTION OF ANTHROPOLOGY AND PSYCHOLOGY Chairman—-CLakk WIsSLER, American Museum Secretary—Roxgert H. Low1z, American Museum The sessions of the Academy are held on Monday evenings at 8:15 o’clock from October to May, inclusive, at the American Museum of Natural History, 77th Street and Central Park, West. [Annas N. Y. AcaAp. Sci., Vol. XXIV, pp. 23-38, Pl. VI. 25 July, 1914] THE GENESIS OF ANTIGORITE AND TALC By Auexis A. JULIEN (Presented by title before the Academy, 7 April, 1914) CONTENTS Page THAR TONE gigs ao ooo maed b9.5 Or 18s GG COBO BO BOO Od CORD. OD DD ICO OIO. 23 TMG GAIT OF Talked oencadesadeahasee saatoe Go capo Gidmy > que BP Sites MRE 25 TAS CENA Oe MiMniTEs 6 aa6go00do co vOdoS MONO OoboOo Go COO UG EU oO lao 26 Formation of antigorite from olivine by hydration.................. 26 Formation of antigorite directly from olivine by thermal alteration... 28 Direct hydration by agencies between two belts...............+-.04- 29 Dual processes in genesis of tale and antigorite........................- 30 Genesis of chrysotile and retinalite........................ Fit RNR Scr 33 WOM POSTUCMVEITSE ele tress tcl srehalsia ors) aie otere oie ro fat ctailokavioucrs shatacehe elo: oteis elei tel 'elig, vcyere he 36 COGiVGIENGIIES Sta aa Shao cote Malte «cod coer ROice no hCrca CEE ete tence eee eee 38 INTRODUCTION The avowal of Delesse in 1865—“Above all other eruptive rocks, ser- pentine has hitherto remained a veritable enigma !”—still pertains to all prevalent hypotheses concerning its origin. Although indeed the existence of a “serpentine group,” as complex as Schweizer’s series (picrosmine, picrophyll, substance e, antigorite, serpentine, chrysotile and other sub- stances a, b, c, d, f) is no longer in common acceptance, there are few mineralogical authorities even now who are not inclined to favor at least the dual distinction, “serpentine” and chrysotile, with differentiation as allomorphs in physical and optical characteristics if not in chemical com- position. The solution appears to have been long delayed by ignorance of certain facts: 1). The impurity of specimens.—The term “serpentine” has been indifferently applied to all forms of the mineral, and, with the same freedom, to the massive rock, often designated as “ordinary serpentine,” in which the proportion of the mineral rarely exceeds 60 to 80 per cent. and may even fall to 35 per cent. or less. In opposition to this loose practice, Lacroix has long ago urged the restriction of the term “antigo- rite” to all forms of the mineral proper, and of the term “serpentine” exclusively to the rock occurrences. Analogy with the precision obtained by discrimination of calcite from limestone, of dolomite (or miemite, ac- (23) a4 ANNALS NEW YORK ACADEMY OF SCIENCES cording to some) from magnesian limestone, of tale from steatite, etc., confirms the advantage of such distinctive use of the two terms, if antigo- rite as a definite mineral shall be held to comprise all substances of the contposition H,Mg,Si,0,. In most of its specimens, however careful their selection for apparent purity, as judged by uniform texture, color and translucency, such as “noble serpentine,” retinalite, etc., the evidences of large intermixture with other substances are readily established. As this impurity mainly consists of other magnesian salts, the usual method of identification of antigorite by deduction of certain molecular ratios from the analysis is by far too rude and unreliable. Only by recasting of the analysis,’ with precise reference to the percentage of combined. water and, if possible, with control by microscopic and optical examination of the very material used for the analysis, can the true constitution be determined for the aggregate present in almost every specimen of the presumably pure min- eral. An unfortunate consequence of disregard of these precautions has been the partial vitiation of many physical and chemical investigations of the mineral. For example, it is easily determined, by recasting of the stated analyses, that specimens of the “dark green serpentine” from Newburyport, Massachusetts, selected as typical in experiments for de- termination of constitutional formula,? actually contained 11 to 22 per cent. of deweylite, etc.; and that the foliated antigorite from Antigorio, Piedmont, used for determination of the form of silicic hydrate existing in the constitution of true antigorite,’ contained 15 per cent. of prochlo- rite, deweylite, ete. It may be fairly suspected that this impurity of material may have led in part to uncertainty attending conceptions of that constitution. In regard to tale, its ordinary intermixture with quartz, chlorite, antig- orite, tremolite, ete., is well known. 2). The obscurity of the products of decay in laterite——In past dis- cussions, the grains of antigorite and scales of shining tale detected on weathered outcrops, though merely ancient elements residual from their insolubility, have been commonly mistaken for new generations. This misleading presumption has hindered recognition of the actual abundant derivatives from rock decay, magnesia, its hydrate, carbonates and soluble hydrosilicates. The resulting discordance of inferences from the numer- ous proposed genetic hypotheses for tale and antigorite, with facts even 1 Ann. N. Y. Acad. Sci., XVIII, 129-146. 1908. 2? CLARKE AND SCHNEIDER: Am. Jour. Sci. (3), XL, 308. 1890. 3S. HILLEBRAND: Sitz.-ber. d. math.-naturw. Kl. d. r. Acad. d. Wiss., Berlin, CXV, Abt. I, 697-712. 1906. Or JULIEN, GENESIS OF ANTIGORITE AND TALC 2} then known concerning the products of weathering of ferro-magnesian minerals and rocks, may be briefly reviewed in connection with each mineral. THe GENESIS OF TALC In regard to tale, T. S. Hunt* in 1860 made the following suggestion, without further elaboration: “While steatite has been derived from a compound like sepiolite, the source of serpentine was to be sought in another silicate richer in magnesia.” In this, his conjecture concerning tale was a happy one and was ap- proved by Delesse in 1861. By neither was there ever advanced any explanation or proof and the fleeting suggestion dropped from view. Taking for example a single mineral, olivine, as the source of talc, as in the peridosteatites of Maryland and North Carolina, the following genetic equation, for direct alteration of olivine into tale, has been pro- posed :° 4(Mg, Fe), Si0,—5(Fe, Mz)O+H,O=H,Mg,(Si0,), Olivine Iron oxide and Tale magnesia So also the derivation of tale from tremolite or enstatite has been at- tributed to attack by carbonated waters, as explained by the reactions® CaMg;Si,O,. + H,0 + CO,= H,Mg,Si,0,, + CaCO, Tremolite Tale Calcite Mg,Si,O,. + H,0 + CO, = H,Mg,Si,0,, + MgCO, Enstatite Tale Magnesite According to another authority :7 “Tale forms in the upper zone of metamorphism. In this respect it is like ehlorite and serpentine. It is especially likely to form under conditions of weathering. . . . It appears to be one of the end products of rock alteration in the belt of weathering.” Yet in the decay of olivine, for example, on weathered outcrops of dunite or other peridotite, while there can be no doubt of the removal of iron oxide and magnesia and of absorption of water, not a trace of newly formed tale has ever been distinguished among the products of decay. Furthermore, the above equations take no note of the free silica which has universally separated in abundance during development of *+Chem. and Geol. Essays, Boston, 296. 1875. 5 J. H. PRATT AND J. V. Lewis: N. C. Geol. Survey, I. 1905. °C. H. SmytH, Jr.: Sch. of Mines Quart., XVII, 333. 1896. 7™C. R. VAN Hisp: Treatise on Metamorphism, U. S. Geol. Sury. Monogr. 351. Wash- ington, 1904. : 26 ANNALS NEW YORK ACADEMY OF SCIENCES tale and has become either a prominent constituent of the resulting quartz-talc aggregate, steatite, or a prominent associate in veins or seams in close vicinity to a tale-rock. THE GENESIS OF ANTIGORITE For the purpose of this paper it will be unnecessary to discuss all the hypotheses which have been devised, or to consider but one important source, olivine. “y. \ % e poor Genthite, garnierite.............. ; { carbonates and hydrosilicates (connarite). drocarbonate, carbonate, hydrosilicate, in part amorphous; siderite. Colloid ferric or ferroso-ferric hydrates, hy- ‘Siderite, magnetite, hematite ..... hydrocarbonates and hydrosilicates; cal- cite, dolomite. Colloid or amorphous caleium carbonate, ‘Calcite, dolomite............ Teoestishe Amorphous magnesia; amorphous magne- Brucite, magnesite, hydromagnesite. sium hydrate, hydrocarbonates, carbon- ates; brucite, hydromagnesite, magnesite. Colloid magnesium hydrosilicates (dewey- Antigorite, talc, deweylite, sepiolite. { lite, sepiolite), in part amorphous. In the equations above given to illustrate the supposed direct conver- sion of olivine into antigorite, the calculated volume changes varied from + 12 to + 37 per cent. To this expansion and subsequent shrinkage, the phenomena of fracture, gliding, slickensiding, etc., observed in many bodies of serpentine, have been attributed by G. P. Merrill and others. DuAL PROCESSES IN GENESIS OF TALC AND ANTIGORITE The object of the present paper is to distinguish and define my con- clusions (without the evidences) concerning the dual processes as well as dual regions of alteration—first, the belt of weathering, and later the lower region, connected with development of both tale and antigorite from olivine. Three other minerals, hitherto treated merely as interesting accessories during development of tale and “serpentine”—viz., brucite, sepiolite and deweylite—now offer their claim as essential elements, in amorphous or colloid form, to the genesis of the two minerals in question. The key to JULIEN, GENESIS OF ANTIGORITE AND TALC 31 that genesis, I believe, lies in the relationship in each case of a colloid magnesium hydrosilicate (Type I), originating from decay of olivine or other ferro-magnesian mineral, during weathering, to a complementary hydrosilicate (Type IL), containing more silica and magnesia and less than about half as much water, into which the former has been afterward converted in a lower region of metamorphism. The four known magnesium hydrosilicates may be thus arranged to show this relationship of the two types: Percentage Type. Derivation. Product. Formula, composition (disregarding n aq.). SiO, | MgO | H,0 Iie From decay of | Sepiolite | H,Mg,Si,0,)+7 aq...| 60.80 | 27.10] 12.10 olivine. (colloid). it Fromalteration of | Tale....... le Mies Ones cab oan 62.00 | 33.10] 4.90 sepiolite. Te From decay of | Deweylite | H,,.Mg,Si,0,,+ n aq..| 40.20 | 35.70] 24.10 ; olivine. (colloid). JL From alterationof | Antigorite..| H,Mg,Si,O,.......... 43.50 | 48.52 | 12.98 deweylite. The processes involved in the development of these four minerals in nature may be represented in part by the following equations, confining our attention to the single mother-mineral, olivine, out of the twenty- three known to pass into sepiolite and deweylite. For tale: 4 MgFeSiO,+ 8 H,O+2 O+7n aq. =(H,Mg,Si,0,) +n aq. ) Olivine Colloid sepiolite + 2 (H;Mg0,-+-n aq.) + (HgFe,0,+-n aq.) + (H,Si0,+-n aq.) Amorphous magnesium hydrate Colloid ferric hydrate Colloid silicic hydrate Kssential volume change (disregarding nm aq.) = -+ 67.40 per cent. Then in a lower region: 3 (HyMg,Si,0;)+ ” aq.) + A=2 H,Mg,Si,O,, +Si0, +4 H,O+n aq. Colloid sepiolite Heat Tale Quartz In massive form the normal rock aggregate, steatite, has thus become developed, a mixture of tale and quartz. Essential volume change*® = — 32.96 per cent. For antigorite: 8 MeFeSiO, +21 H,0+4 O-+n aq. =(H,,Mg,Si,0,,-+n aq.) +4( H,Mg0, +n aq.) Olivine Colloid deweylite Magnesium hydrate + 2(H,Fe,0,+ n aq.) +5 (H,SiO, +n aq.) Colloid ferrie hydrate Colloid silicie hydrate 16 Without regard to n aq. 32 ANNALS NEW YORK ACADEMY OF SCIENCES Kssential volume change’® = + 80.31 per cent. It will be noted that incipient development of both sepiolite and dewey- lite from mineral decay has been attended by separation of a certain amount of magnesium hydrate. This was not understood, except by Roth and Teal, or included in the formulas previously given; it may serve as a test of the truth of the reaction here set forth. Later, with subjection of deweylite to thé thermal conditions in a lower metamorphic belt, the complementary process of alteration has taken place : 3 (H,Mg,Si,0,_ +7 aq.) +A—4 H,Mg,Si,0,+Si0, + 10 H,O-+n aq. Colloid dewey lite Heat Antigorite (94.8 Hyalite per cent.) or quartz, Hssential volume change’® —= — 32.79 per cent. Reference has already been made to commonly accepted views concern- ing dynamic effects upon bodies of serpentine by the changes of volume in progress during passage of minerals into talc and antigorite. It is now apparent that admission into the equations of the hydrated colloids of seplolite, deweylite, etc., actually found in nature, would involve an early hypothetical expansion far greater than hitherto estimated. On the other hand, the later physical changes which have preceded the birth of tale and of antigorite have generally culminated in notable contraction of the rock mass. We have to do here, however, with more than chemical reactions. The attendant physical processes of solution, leaching, trans- port and migration of soluble constituents, and their later alteration in a deeper thermal zone, have resulted in a complex fissuring, and often in an amount of contraction which has decidedly offset the expansion from early chemical changes. The observed evidences of internal disruption and movement in bodies of serpentine may be therefore everywhere ex- plained, I judge, by successive throes of expansion and contraction—e. g., at Staten Island and New Rochelle, New York; Montville and Hoboken, New Jersey, and Thetford, Canada—and also by local strains and faults produced by orogenic disturbances. For precise definition of processes above considered, I think we need differentiation of the following terms: Decay of rocks, to express the result of operations within the belt of weathering, disintegration, oxidation and extreme hydration. Among the more important products are the colloid magnesium hydrosilicates of the first type (colloid deweylite, sepiolite), magnesium oxide, hydrate and giobertite, besides various forms of ferrous and ferric hydrates, hy- drocarbonates, ete. 16 Without regard to n aq. JULIEN, GENESIS OF ANTIGORITE AND TALC 33 Alteration, to express the interchanges and consequent new formations, with great loss of water, which take their birth in a more deeply seated region. The common products are the magnesium hydrosilicates of the second type (talc, antigorite), hardened deweylite, forms of limonite, gothite, turgite, hematite, etc. Decomposition (Zersetzung of Roth), to express the molecular disso- ciation, still more complex interchanges, and still greater to complete dehydration, which have ensued within the zone of anamorphism. Ex- amples of these products are periclase, spathic magnesite, dolomite, siderite, breunerite, regenerated olivine (boltonite, forsterite), specular iron, magnetite, etc. In regard to the term “hydrometamorphism,” whether in the sense of Lindgren, referring to the action of meteoric or vadose waters, or in that of G. P. Merrill, referring to the action of waters from deep-seated sources or from magmas, I find no application for it below the belt of weathering. There only has originated the highest hydration; below it, every change has been attended by progressive loss of water. 5 GENESIS OF CHRYSOTILE AND RETINALITE In Plate VI, a well-known laminated variety of asbestos-rock from Thetford, Canada, is presented. Here lie the leaves, silver and green, in long succession, of the book of the history of asbestos, waiting for inter- pretation of the mystery of its origin. If “serpentine,” as long believed, is a colloid, incapable of crystalliza- tion, is this fibrous chrysotile but an alteration product from asbestiform amphibole or bronzite? Or are these fibers only “serpentine” wires, pro- truded through pores in the vein walls, like those of metal in the arts? Or, along fault planes, has the serpentine been rolled out and sheared into these silky threads? Or, if there be a crystalline paramorph of amorphous “serpentine,” is this its fibrous deposit from lateral infiltra- tions into rock fissures? Is it possible that these have been generated by diamagnetic secretion along the vein walls, expelling into the median fissure of the vein the feebly magnetic brucite, poor in iron, and the diamagnetic calcite? Or are the fibers in fact capillary or acicular crys- tals either of “serpentine” itself or of its paramorph, thrust from one wall to the other, or grown simultaneously inward from each wall? It is doubtful whether any one of these conjectures has proved satis- factory even to its author. Toward solution of this part of Delesse’s enigma, in my turn, it re- mains to sketch some of the migrations and transformations of the magnesian derivatives from rock decay, as they oozed downward from 34 ANNALS NEW YORK ACADEMY OF SCIENCES laterite into fissures and occupied them as vein deposits. Magnesium hydrate, the most soluble and mobile, was the earliest to form a coating on each wall, sometimes filling up the entire fissure. So originated the veinlets of brucite, crystalline at Hoboken, New Jersey, and crystallized at Hopansuo, Finland, and Texas, Pennsylvania. The next stage in the process has been connected with the tendency of brucite, when subjected to rock-strains, to molecular rearrangement in direction of the pressure. Its grains become shot through with parallel lines, without regard to the cleavage, and at last transformed into aggre- gates of fine fibers. Thus brucite has frequently passed into its fibrous allomorph, nemalite, with fibration normal to the vein walls, well shown at Hoboken and Montville, New Jersey, Xettes in the Vosges, etc. From solution in carbonated waters, veins of the less soluble carbonates, hydro- magnesite, magnesite, etc., have been also produced, or from the action of such waters on brucite already deposited, as at Hoboken, many localities in California, ete.; or where nemalite has occurred, coating each wall of a vein, the interspace remains sometimes filled up with laminated brucite, as at Hoboken, or with magnesite, as at Montville. Next, by passage of siliceous waters, crystalline brucite has been con- verted into its antigorite-pseudomorph, marmolite (as shown by Volger and others) at Hoboken and elsewhere, and its crystals into “thermophyl- lite” at Hopansuo. In the marmolite of Hoboken, pearly flakes of un- altered brucite can be sometimes plainly distinguished. This again im- plies the intervention of deweylite, and there is abundant evidence of its generation by the following process—reaction of free magnesium oxide, hivdrate or carbonate, or of dolomite, with percolating solutions of silicic hydrate or of alkaline silicates. Of the resulting equations it will suffice here to offer the following: 4 H,MgO,+3 (H,SiO;-+n aq. ) = (H,.Mg,Si,0,,+ aq.) + H,O Magnesium Silicie hydrate Colloid deweylite hydrate Volume change (disregarding n aq.) == — 3.75 per cent. Deweylite of this origin, subjected to thermal conditions, passed into antigorite by the reaction already explained. Where silicification of nemalite took place, it was converted into dewey- lite with pseudomorphous fibration, and this, by later thermal action, into its antigorite-pseudormorph, chrysotile. The passage of nemalite into chrysotile, supposedly direct, was detected by G. H. O. Volger'® in specimens from Hoboken in his cabinet, but the intervention of deweylite was not suspected. 16 Entwicklung der Mineralien der Talk-Glimmer Familie, Ziirieh, 254-270, 1855. JULIEN, GENESIS OF ANTIGORITE AND TALC 85 The office of deweylite has not ever been recognized, doubtless in part on account of the difficulty of detection of a colloidal amorphous sub- stance, and in part of its general alteration into antigorite. Evidence of the latter change is revealed by the frequent partial survival of dewey- lite grains in intermixture, and also by the very chemical composition of many specimens of antigorite. For example, T. 8. Hunt made among others the following analysis'* of chrysotile “from a narrow vein traversing the Eozoon rock of Petite Nation seignory, Quebec: silica, 43.65; magnesia, 41.67; protoxyd of iron, 1.46; water, 15.48; 100.16.” He commented thus, with surprise, on his results: “these serpentines from the Laurentian limestones are remarkable for their freedom from iron oxide, for their large amount of water, and their low specific gravity.” ® These anomalies are explained by the results of my recasting of his analysis: antigorite, 95.13; deweylite, 4.63; hyalite, 0.40. In develop- ment of the pseudomorphs, marmolite from brucite and chrysotile-asbes- tos from nemalite, a steady progression in contraction is shown, to about one-third of the volume, without disturbance by expansion, from the original magnesium hydrate to the final product, antigorite. This seems to be correlated with the perfect preservation of all structural details, even to the most delicate features of nemalite. This genetic history of chrysotile, if accepted, enables us to use its occurrence as a test of conditions which have always prevailed during genesis of antigorite from decay. Its general association with the other forms of that mineral, even at the “stubachite” locality, establishes iden- tity of origin through the dual processes already explained. Colloid deweylite, the magnesian companion of brucite in migration from laterite, has likewise been concentrated in simple veins, as at Texas, Pennsylvania, Bare Hills, Maryland, etc. Where a portion of the dewey- lite has escaped the subsequent alteration, its intermixture has produced the waxy, translucent variety of antigorite, retinalite, common at many localities. Its analyses invariably reveal an unusually high percentage of combined water, due entirely, as shown by the recasting, to the pres- ence of several per cent. of unaltered dewevlite. Moreover, specimens are not uncommonly sprinkled with visible grains of that mineral. “Rpt. Prog. Geol. Sury. Can., 205. Ottawa, 1866. LAIN SOUL Cle (2) ekki GS.) L864: 36 ANNALS NEW YORK ACADEMY OF SCIENCES CoMPOSITE VEINS While the separate deposition of both magnesium hydrate and colloid deweylite has frequently taken place, as described, in simple veins of each mineral, nevertheless their normal and probably more common mode of conveyance from laterite downward has been in intermixture. Com- posite veins have resulted by separation of “successive deposits of each from this mixture, and not, as might first be judged, by a series of de- posits upon each wall, now of one mineral, now of the other, in alternation. A simple form has consisted of a vein with wall coatings of brucite or nemalite, with a middle sheet of deweylite. By silicification, the wall coatings have passed into fibrous deweylite, and this, by later alteration, into chrysotile-asbestos, with a sheet of massive antigorite or retinalite intervening, as at Portchester, New York, ete. The reverse order of arrangement has been also observed, with sheets of massive antigorite or retinalite (7. e., originally deweylite) coating the walls, and a central sheet of brucite, nemalite and sometimes calcite, as in the Vosges; or with a central sheet of nemalite, in part chrysotile, as at Hoboken. A proof of the above suggested intermixture of the two magnesian components is yielded from study of analyses of retinalite. A specimen “associated with eozoon” at Calumet, Quebec, gave T. S. Hunt the follow- ing results: silica, 41.20; magnesia, 43.52; ferrous oxide, 0.80; water, 15.40; 100.92. My recasting of this reveals the following constitution : antigorite, 83.90; deweylite, 11.76; brucite, 5.24. That is to say, a nota- ble portion of each of the original magnesian components has escaped alteration and remains intermixed with the antigorite. An interesting example of such intermixture has been observed in a symmetrical asbestos vein, two inches in width, in dark green serpentine containing particles of chromite, on lot 13, Range V, Thetford, near Rob- ertson station, Canada.‘® The first deposit on each wall has been a thin layer of dark blue antigorite (originally deweylite) “with grains of chromic iron” ; then a layer of chrysotile (originally brucite), with fibra- tion normal to the wall; then a thin layer of pale-green retinalite (orig- inally deweylite) ; and a central sheet, about +§ inch thick, of dark blue antigorite (originally deweylite mixed with magnesium hydrate), along the middle of which run minute seamy partings of chrysotile (originally brucite) parallel to the plane of the vein. A succession of four passage solutions of magnesia is here indicated : first, the colloid hydrosilicate; then the hydrate; then again the hydro- 19 f, CIRKEL: Asbestos, p. 28, Fig. 6. Ottawa. 1905. JULIEN, GENESIS OF ANTIGORITE AND TALC ey silicate ; and then the main solution, or mixture of the two components. During consolidation and contraction of the last deposit, disassociation of the hydrate took place by diffusion into the shrinkage crevices near the middle of the vein. In fact, however, the separated deposits here found on the walls are probably, like those next to be described, derivatives from mixed solutions, by disassociation higher up the vein. The most complex variety of composite veins is that represented in the illustration (Plate VI) and ordinarily found in proximity to laterite rich in magnesian silicates. It consists of a lamellation, in abundant repetition, of thin alternating sheets of chrysotile and retinalite, the thickest near the vein-wall and thinning outwardly; the first very thick layer of retinalite on the vein-wall is absent, having been broken from the specimen. In structure and development the variety is essentially identical with the lamellation of antigorite (“eozoon”) in dolomitic lime- stone at Grenville, Canada, and other localities, although there the ma- terial of alternation with retinalite is calcite in place of chrysotile. In each case, I have concluded, a rhythmical process of unilateral vein depo- sition from laterite solutions has originally taken place—every pair of lamellee comprising a film of colloid, magnesium hydrosilicate, with one of crystalloid, magnesium hydrate here and calcite in the Canadian oc- currence, separated from the colloid by dialysis. The rhythm of deposition has apparently been due to limitation of the flow into the vein fissure of the mixed solution of the two magnesian salts in meteoric waters to a certain period of accumulation, perhaps the rainy season of the year. After spreading upon the surface of the wall, dis- association began, the colloid being left clinging as a new coat upon the wall, while from its outer boundary—perhaps through a dried film serv- ing as a septum—the crystalloid magnesium hydrate became diffused more or less completely by dialysis and so formed the companion coat of each pair of alternations. The amorphous magnesium hydrate readily crystallized into brucite, and this, by subsequent pressure—perhaps by rock strains, through expansion in neighboring portions of the mass— was converted into its fibrous variety, nemalite. Other fissures have been opened by contraction of the rock more or less transversely to this lamella- tion, but these have been generally filled with magnesium hydrate, amor- phous and crystalline, as simple veins, changed in turn into nemalite by rock strains. By later silicification or alteration under thermal conditions, all these lamellz and transverse veins have become altered—nemalite into chryso- tile-asbestos and deweylite into massive antigorite, in part retinalite. 38 ANNALS NEW YORK ACADEMY OF SCIENCES CONCLUSIONS In review, then, magnesia, in hydrated or carbonated condition, and deweylite and sepiolite, in colloid form, have always been the only mag- nesian derivatives from laterite, with tendency to early migration and transport, in virtue of their solubility. Antigorite and tale, on the other hand, ctystalline and never colloid, have merely served as insoluble fixatives to harden and record the trans- formations of their mobile and protean predecessors. Chrysotile is but a pseudo-fibrous variety of antigorite—in fact, a pseudomorph in antigorite after a pseudomorph in deweylite after nemalite, the fibrous form of brucite. | To the list of rock-making minerals, brucite, deweylite and sepiolite need to be added as important accessories. The evidences in confirmation of these views from field observations, optical examinations, etc., together with a review of the literature of brucite, serpentine, antigorite, and the hydrous magnesian minerals, have been gathered for presentation in a separate monograph. DEPARTMENT OF GEOLOGY, CoLuMBIA UNIVERSITY. ANNALS N. Y. ACAD. SCI. VOLUME XXIV, PLATE VI PUBLICATIONS OF THE . NEW YORK ACADEMY OF SCIENCES (Lycrum or Natura History, 1817-1876) \ The publications of the Academy consist of two series, viz. : (1) The Annals (octavo series), established in 1823, contain the scientific contributions and reports of researches, together with the rec- ords of meetings and similar matter. A volume of the Annals coincides in general with the calendar year and is sold at the uniform price of three dollars per volume. The articles composing the volume are printed separately, each in its own cover, and are distributed in bundles on an average of three per year. The price of the separate articles depends upon their length and the number of illus- trations, and may be learned upon application to the Librarian of -the Academy. The author receives his separates as soon as his paper has been printed, the date of issue appearing above the title of each paper. (2) The Memoirs (quarto series), established in 1895, are issued at irregular intervals. It is intended that each volume shall be devoted to monographs relating to some particular department of Science. Volume I is devoted to Astronomical Memoirs, Volume II to Zodlogical Memoirs, ete, The price is one dollar per part as issued. All publications are sent free to Fellows and Active Members. The Annals are sent to Honorary and Corresponding Members desiring them. _ Subscriptions and inquiries concerning current and back numbers of | any of the publications of the Academy should be addressed to THE LIBRARIAN, New York Academy of Sciences, care of American Museum of Natural History, New York, N. Y. PRESS OF JUDD & DETWEILER, INC., WASHINGTON, D. C. ANNALS OF THE NEW YORK ACADEMY OF SCIENCES Vol. XXIV, pp. 39-113 Editor, Epmunp Ottis Hovey A STUDY OF THE CHANGES IN THE DISTRI- BUTION OF TEMPERATURE IN EUROPE AND NORTH AMERICA DURING THE YEARS 1900 TO 1909 BY HENRYK ARCTOWSKI NEW YORK PUBLISHED BY THE ACADEMY 2% JUNE, 1914 THE NEW YORK ACADEMY OF SCIENCES (Lyceum or NAturRAL History, 1817-1876) Orriorrs, 1914 President—GuorcE FREDERICK Kunz, 601 West 110th Street Vice-Presidents—CHARLES P. Berkey, RaymMonp C. Ospurn, CHARLES BASKERVILLE, CLARK WISSLER Corresponding Secretary—HeEnry E. Crampton, American Museum Recording Secretary—EpMuUND Otis Hovey, American Museum Treasurer—-Henry lL. Donerty, 60 Wall Street Librarian—Ratpu W. Tower, American Museum Editor—Epmunp Otis Hovey, American Museum SECTION OF GEOLOGY AND MINERALOGY ~ Chairman—Cuartes P. Berkey, Columbia University Secretary—A. B. Pacrnt, 147 Varick Street SECTION OF BIOLOGY . Chairman—Raymonp C. Ospurn, 557 West 124th Street Secretary—WiLuL1aAM K. Grecory, American Museum SECTION OF ASTRONOMY, PHYSICS AND CHEMISTRY _ Chairman—Cuar.es BaskERVILLE, College of the City of New York Secretary--Ernest E. Smiru, 50 East 41st Street SECTION OF ANTHROPOLOGY AND PSYCHOLOGY - Chairman—C.Lark WIsSLER, American Museum Secretary—Roxsert H. Lowir, American Museum The sessions of the Academy are held on Monday evenings at 8:15 _ o'clock from October to May, inclusive, at the American Museum of Natural History, 77th Street and Central Park, West. \ [ANNALS N. Y. Acap. Scr., Vol. XXIV, pp. 39-1138. 27 June, 1914] A STUDY OF THE CHANGES IN THE DISTRIBUTION OF TEMPERATURE IN EUROPE AND NORTH AMERICA DURING THE YEARS 1900 TO 1909? By Henryk ARCTOWSKI (Presented in abstract before the Academy, 2 February, 1914) CONTENTS , Page MrdtMO MUCHO Mrs r Ga O men OM Date see Oste G0 ste eee O o> en sel on Ot : e'o— T'0— O°I+ o'O+ T'o+ F'O+ oie T'Oo+ ics e"o— ON 6°0O+ T'o+ m1) F'O+ (62 | T'o+ eo) 9°0- Gules 6°0+ o O+ T'O+ F'O+ (oe 169) —= a 9°0— co S°O+ T°O+ ¢°ot+ $-o+ OS e-o+ ~ ’ ’ 4 co L: + t-) i e 2 | oO —) et ID HAR re 19 190 ooo +44 AIO DS Hk191 rin ll | Nelle sinele ol ar] =) —) S | | oO o Se miei St Ehih E eee tea ++4+++4++ BHHSonnHS Geer | o a ~ - | Neji Oo YORK ACADE ‘o— TT oT oo “¢ oe 205i Cilla AO == G20 Cale 406 (Gb b0os eS S06T GOGL Oe i) rere C20= 9°O+ EO -°O+ T'Oo+ JE ClO ste 2} g‘0 = L20— —90-- tO) S305 Osi “0 On Ose 0 = F'O— e'o+ 1 ¢°o+ €°0o+ L.0— o°O+ 9°0—- 0 i G‘o+ C20 Lot €°0o+ AX) Oster 6:0 = ‘0 8°0=— 9°0+ e0= S'O+ G°o+ EOS 3°O+ Sa Oe 0 9°0-— e°-o+ £°0— L°O+ g‘o+ GaQlee Ga Oete 6°0- 0 e°o+ A (= 6°0+ G20 ss C20 0 | f F°O+ 6 “o+ 0 ok - + ANNALS NE ] | re reer —— 1 +tt+Ft+4+ en) ooococeo +444 10 pa Oe OS < OU @'o+ o- F 606T 80Gt ocoocoooso FFFHETt ttt Ht nOm 0° 2057 6 3 OSz 0° € moc or > o + NATATAT ra ND 4 —) + al —) | oleate ro HSH 6 19 0D 10 I HD HD SH = Manon COA AI OD HAIN —) ri OO =H —) -f ‘9— FOGT nr) SO6T TO6T or) San LOG6T 90 ID Be £5 109 £ 6 109 SHAD SH NI SH ° | Sali Tex oo NI 118 - ~ OO HS 00 09 4 1013 OO OD OMMOMUS~rOOCOINSION On ~ lo) 19 OO SH hh A oO~rseootird IIAIDIN = Oe Ol nN ~ HONHOONHOMOR ~ ORO OOD 1 SH I= =| 3 o a a eae is penunu0j—suoynjis unadoing sof vyop aunqosadwmat— J] ATavL, €)e_s\ ee ee) os) esle © eis) eeene tee ee “o* pespRey “* mpoT yoo BO POF Car takeh(o) ‘* peysulodysS punsi9ay4sy ‘ SurdoyaAn “+ Surdoyury] BABYS ** S.10qs190e A “++ surdoyuosr Ox€ ioe Be 56 eM OLK OA a len el one Doren 0 oD peiaieheseltn lols) lehes heanyen SITET, enews oto Deo od Omen o DOCU OUOOOG icjihips . eee ee eee erence Em UMRYS[IRy ORI O OOOO COG FONG DONG" FRITH trees DONT] * MIUIBYOABIY ** SLOTSUIS[OT] * S1OJIOUIUIE, TL, “** BTRARDIOg teens BSBA 2 OLCOnSy S1OGRET() 6 ua 6 ep ues19¢q ** SuUBAsuglif}) “"** somepnys sess) Tepueyy ee ee au ee eee [350A IIpsR “s* ppeddeiy Pees opi “* BIUBISTAT) “'* TasoqRy Continued TABLE II.—Temperature data for Huropean stations 1900 ee 1902 1908 1904 1905 1906 IU Bett) 1909 Mean CO 1D 1D 1D 1 SH SO CO OT et 0 et ed tH LO HO RH OO HOO SH SH 1D CO SHOT ooocooceoooooocosooo“eooeonoocoocococesosooeso MOV SS an eer ra CUU TH SH HS 109 69 AV 109 £9 19 C9 B= 0.00 AAU Tal HO CUTTS C0 IDOI SH =H SSSSSSCSsSOSSSSSSSSSSsSosssscessssesssssess F4FFFFFFFFFFFF THE $4447 144° $444 0° 4+ ICN iri CIGNA Cia) St SH CSCO OCoTSICN orlsslG: SNe AON NN MOIS ei oses SSSSSSSSSSHSHSS Sosessossosessssssoee P11 ++4+444444444+447° T1111 Tid 1 +444 4+ =~ a = HUISAOSSSOM 2 Om DOAKBSAGWOO™ SOSH Onn H On OSHnHOHHOOHHS +4444 ttt t+ DNOHONS DHT DD at Sononsns SeS99090 ++t++4++4++4++4++4+4++ elias TACOS oi MHC) SH 109 CO 169. SH ACD 4H 4 CD Ssossosesesesssssssssssssco Tt +t444+ + F444 Ht SOD COND CIA 4 41.9 FEttstid CASH HCD CO 109.00 HCO 169 B= B= CO CH OO @ \s) 10 . * @« . OO. Cts ADC Ds OR 6 Ch x CARESS ca och hs : afc Se Wises ane tore ee Be o Shwe thesacoe : ieee Aap? 2S bie pee oe Shs: 55: AoC : Aa ie acer Se ses mS ses Bey, . oe! ere, . esa . On CIC er ium eo (@ . SQ . aes . sane ae im ct oa BO ee cane ie a igs Ba a esis aS ed 338 Se ore -gU- Tmag Ka 0 hap oD qg- ay. M+- 2G -aA, o AOL eoeS gas : Gey ofl Sp2 2 oS og Md ROG = -SOR85820 mia —Gieit mS gna S:o iS Be ig tO 2D ASS ROO SBD TASSHDS awe pO Sie secR wisn 4g an Ses Rela soa eke eas owe =a 4 ¢ — * — CS ad DoS SRO OH asian SS mea on SA ahora ats Ze PROAntoOnPnwteoOMt ene oe eRe pndnene DENMARK CO HH HO EERO NOG: Et it= Sesesseso 44444444 MH Medes Soooosso tT +T+T HOD MOS AIAIN e660 S950 FEPEtH+ 169 C1 CN eosesocs Ftttt444 fo ot —0.9 —0.8 —0.9 —1.0 —0.9 —0.9 PD | re | —0.8 .4 3 m5) » +0. +0. +0.2 9 5 +0.3 0.1 oS ys - Geo ks| Mim sesessso li +++ ++ MOND AMO SeeeeEert Bim Veliu/ielelsel ee ej\e\iele ieee 6: 0,18) /« aj Lara a a aunes| wil ig ce AP -g ‘9 3 Og, aD @ 5:0 A 2 s:0 g anmAat aio w 4OVGY Smo 8 NBO pM mam It ~ NID OND rie nosso seit Te T ooooocr FFFtttttt+ CES 4 4d Ane Cin tty OWCor a~=-- ~o~— =r ocooceo ~ ae FFttt YORK ACADEMY OF SCI Or°xw =H co 0019 © 00 QUA NO HOO rS coooo ++4+4++ ANNALS NEW cf) Of Ne MW 6O 1D 1D HK CO OD HHLON ocooosS oo. Vt OD ra OU OD 1D HH SO coococossos +44444+444+4 A Qe rol © | b oe Vis Ci O ies sO = ‘O “O iO “O CO6T Oshidg co 20.00 00 op = t= CO DWMHAMC I~ OAs t 1D 61D CY 61D SH OD ea ON ed DABDOAONR ACS Baa Cs ah oes ft Con) Crs Ai od 00 fd of ° THA SHH mS Deol OOO £1 1m t= CO u pe soe 1D Oo) Orie LS mtr DrHAAOrmrna So ccoocsoe +4+444+ Son eos i000 1d Scoosess Ft44+444 Soa hee | FOr OM Net ~~ So” o> —_ (=) So for) —_ 8) O56 16. w ya) 0 610 <0: 8) 0 0)\6 6) 6) 618 6) 0d) .0) 60, aia ie) a. ela), S1sqmoag 60:6 ae ches wie (9.\616)0) m6 )0!s vi 0\ eels (sie elele «eels 9p01948O S606). 60) 0) © ee) 0/0 Siebel esol eceaevele si claidas! RAOUL ae eee eee eee ee DRO OIOIS OG ODER OO OOD CRG (Far syy silos (eile s @)n).6) se sheug) elas) ee) 9s)\sisile) «061s se reees TOMTIAT Raw s (ure foye ae. e).6 bon OME POSTON IIE OTN O-IO8 VEY hil RCT BOK ECK ONS) eh ensiieuenaie wien ieene Peas ich) TONS PATE TIVO Ni Sains sole p) Genie Sellen sei. oieetpele Sek emma OT) 1 ONTOS TS . . . a | IPUNWIUIMGS #tatgvenee-eaoge ore ackeyoveleiel suelienyisuslelais) shee heaPeae ANC TS TANG a s. 0,0 0 enue 9 0.5 pele. 6 SOR IEMA 4 YT °.0\(s} 5 peloxe e's) see) eke.e (fd ADEPT SOTO COO. Uap ence | wiv euele ee ee lee lelers, ee ZOO DOO Vii jar ys] oe le 9) elle ois (fisisuene sreceaeisiete. ote rire aT ABUSED OU TON €:(0.0 0001/0, '6,\9) elele, enexelatelere #0, e 0, pie) ae78 0p OOO ra cb: Coys ANVWUAD NYAHLYON elena ie ayane) elpxaie als te (areca wiwketdleselion Gia o2252° MOMMOTA @ el wise) st elier sie eats ecnlisite DI GO OO Os rN ey LaLa): mien sin 2eljeieh she feholioxol che snakotchenshcs ithe OS TACOM CHI? DIRS O ISS ODOC OUI SCOT DCD GOOKOOOT Tir aviniy (| POO ODIO OCOD COD OOO DODD OOOO OO OU TNT uN aNvV10d Sig Sinieueie tcc eevieat snes teuecelo aye chine eee ee TIN SS EIT apie eat dee Ps EO OPO DUS UIO G Fri STL i (4 \ pkey hie ke cere eae 2 Us eeeee Id ed "** TJeplaH @ ie 10 <0 0/68 (6 (0a (06,16, ave, @ 6) 6.01.) 190818 weer ele e166 * UesuluOds) uo[ry "** 9[90f) Santos Porte nett eet ee reste esses ss apuaqso see eee ORONO OE LON OTS DICE OLS OOO GOO IO iwig S| eee) a ¢ © 2.8 0)\6) 6e)e) enelel6, ola) p ial ee RTE OCOD, FORO TONEY FH OP Cet Chi w) OOO. GOD OHO. TOL OOO O08 oye s hs fala | eNelis\-0, ease ellexeleloun OSI O CTE GOOD O00, 0 LiCGNT OD Tay A) see eee ERT OREO CORI CUETO OOOO AE BERT ES | DIS SIONS CHO. OS CIC SOO OOOO GIO VO COO CODE ays} PATO OOOO OM OD WOOGCDI OOOO OER 00 By NIT, GNVTYAZLIMS penulju0pg—suoynjis uvadoung sof vyop aanqgosadwat—y]] aTavL TABLE II1.—Temperature data for European stations—Continued ARCTOWSKI, CHANGES 1900 1901 ~- 1902 1908 1B0¢ 1905 1906 1907 1908 1909 Mean TID Te OH HOM OOO IDIOM ADO OD Ol9 OOM OOH 10 9 1 €O 109 CO 6D C19 XH v4 C9 109 SSscssHsssssesssssssss Sosscceéoscssssdses FEFEFEEAFFTTFTTTTFTTFE F444 44-441 H+ NCD NNNAARS wes MAAR ANN ay sh Di Sen ok ek PERE eee | oo 5 Ags aS ae er a COC) A COUT CIN sossossc])] TI) FFtt4t+° $44 Nn oso = OH Sy rH 9 CCID CD 1D CU HCV 10. OU LCD RH Bi AD CO) HD eH HN SHON O19 019 CNH CHD CO 6 H.00 NHAC) SSesssssssosssssssesssssésssdo I eS | FEF EEEEF TFET EFT F FFT STEFF ttt t+ MOO MOM HS OID KOKORO 09 SH U9) KO CoH CICCOLO LO SH HCD 1H 019 2 £0 60 100 b= SHI SSSSSssssssosssssssssssssssso Ssssososso FHEFFFEFEFFTFTF FF H ttt FEF FFE ETT E EHH FAM AONMANMANNOAS DANA WO SD AW AR A= HW Hl DIS 19 O19 Sn OnNSSCSHnRHO SS SSS SSSSSSSSS LETT rs ST PST ARTI CTP Has TTA Ft TTT TT a te Td VW se AANA rai Tir MicdOo ES Ac} N of TANGY SANA ra CUCU ee eS ee eT +ttt+ 144 $4417 TTT 1 oT ttt TFi itil CO SENS A UD) CD UDI) 1S) SAD COMO HE SH $0 9 1S SO XH he 19 09 1D 1D 1G 1D AG 2 oC HH SsesssssssossHsesssossssssssessssosssss — FEFEFEFEFTTTTTTFTFTFFT FF ttFtttttt tte Ft NAD DDADANANAW6 SOM NAONM OMANI OMAHA O HA OOH REE DEERE SDHDDEDODDE HONG DODD OO EI Bet WS ODS eet Sy nets CE, ame eat = Mer ae mea BO BE 8 Roe eR Onn a ated eae Brags wep iat eek a Dc oe Ca ES SR Or ae teeta eon autou ioe am ledtd ta my ocd (al del sale te VG (of id) “epacecuser e eM ieen were eileacen Nepkopkela's) col teyuou so] ca ay bad Wag fey are lee oh sa) sale GO: fo, Bue! cele se oD Ga © Oy OsOUn Holle ae ee Aw Saye Os sO ot og cles oe as 5 oS 0 oo Se 2g epietie Samay ese)” Ceiicyiale let estsm®) Ko) 4) sia) = 2) iggy 5 tS, Aad ec eres otal) -th Jleir=tt fo DE Ten fol teh .e) se? Ce SDs oy tarot est ve ° = Eat oe OG) og oe Ae Pa oe Sigg te eh Shs - si LES ORC Rha p ete Neer molsthee eats Bt pe ae aD EA ca rs gy Ne B ad Eset aismaes tals ais tes Dey al area oe --3 i ci) Banse! omen beane hoy atotee Ske on & Fee su N ciety eee ee eed ae in eee (aes 2 ce ‘ ; os 5 AD ; Bano Suge sas eeS srs Seogssassa 28 8-23 OSS eS SSR BSA OF BSS SB PSC Agog ee oe oop aHEeA GaSe hee abe Bocas we ke ema ke S aac agasses BOVE RS oOPONORAOR SOC ONO BERS THO CAB HiS SOM HMAnmewsMeqoneenDomnoMastnaZeeaeaMs q0duzk IN DISTRIBUTION OF TEMPERATURE BAVARIA 55 on oe cooceoo ++ + bal PO EAICO cy CS soosso +44444+ 19.00 UII Sossso FETT BDAC O26 OO ososso ++4+44+ Ee Ore 1o)o18 ososse FEFFFF . Bam bergiteccncia sims omnes WALD ZDIIT Be keyere oerarcirencte 2 Ludwigshafen Kahl-am-Main eee) tiie toe miele nara. cte tlecel EV ENUANOU LLIN eeree creda sce otae wits terrer'ei/uitay Boe aielie: aesvenremerete Bad-Kissingen ae vas ANNALS NEW YORK ACADEMY OF SCIENCE © ww Sa a 910 LO = 90> Osa S70 1s S | —) Ou oo 1 | ee Hi CYS a9 BS | OAS = 0 [(8 CeO -— = je 1} 9°0- 9°0— <'0- ¢°0- c°0— T'o-— G-0— PaO 0 a Oi 0 .05- P20 9°0—- L0->- G‘0o— e0= LE0-- cl0= O— “0 70 — 9°0- L0— 10> 30 G.0-- C0 Os Go Ot+ oc0 = 9:0 T'O+ N= Lo CAA Na ae : 9 P = (= ‘0 i) | ririo THO OMX OO oo ht =~ —) men 10=— Le -~o = | o- ‘0- Ga0= 10m" (920-5 S30 *= T°O+ —— s0== —) Ns HO SO HON 1O= “O 0 Jen Oz AO cer (Os gc" 0 I! On 9°0— G 105 920 IT WW 9°0— ee € G af ook Yer) at 7A sH sh el = aR 0= 90= (0— S0= 6061 806T bo Te} LO6T ‘0O- c'0o+° 0+ Try 1 CY) OD b + i ~~ on a —) i> A SO $ t-t-19 S D rrr rte ++ ws ce IMmee A L + oo ++ HO Hie CNV tT I SH CO 013 C13 CVD YD SH AT CVD OV SH IC o.0rtr c°o+ 2°0- ¢° m1) g°o+ “0 L°O+ (Aaa | oe c'o+ S30 L°o+ 10) 9°O+ x() L201 Geis 870-- ee Oats ha O ste i) O rat et et non) ateatcal wD + nA) S990 ff 1 +4+4444444+444444+ G Sc db CO HOO ON XH HIG BH eooooscoessoo SH OD COLO NAD St ri PN Si f=) ‘o+ =69°0+ CO6T FOGL 606 So 2 ON GYD SH SH LS SH XH WA o x | TAD NI OD OD | “0= ‘0- “O- oe on ZO6L CVAD SH rt =H OVD OD matted o OO SH SH rt C8 + Xo oro SH os . - SIO MN SH Het et tt oO | —) + Yer) =) + | ++4++4+ ° | Nee S | a) + re | S HIDHOLININOOAMDAOOOL ODE SSSSSOSOSSSHSSSOSOSHOHS —) | [Yer >) | be oe ho | ooocoooeo UD DUD UD OV 69 GYD SHO OD ON ooo: ) Te | oo + CO GS 1D et SH SHAT Ht CO 2 =) + =) ++ MORDAAA rr ooo If Wt NOMA o | ra + ~ So + VIALSNV earn DOT UR GoegEAS HN oD Ry Re gene ca anode : Peete eee ee eee TTA nee ‘ 11BS}INIS Lt ee et ee Oe OupAy ey andi sp ty oO OXCL Od On) qoopdoyqog B RO LOO Mgio ol DO B06 Bee's S1aqmoyas a SONO Kel PECIOTOIO OO. Sntenahenons Petseeeeess rpayquesaayy 0 OO Od OnMmooO OD ee BPE OD BORD er HONG NT Aegelana taeda ph iyeraec is : ‘reese maqueqoy “8 * THO. [To] SSE COO ONE IPO OS ny \atase|9}6} UIJVYSYO psy ** UeyeyUIYOL yy] Sapo rrr ese eeceeer ses ees ses anpisuopnedg sels gieiain ¢ oie eco lei sieve ole. oimteliniel oi lot ATER) “7 ** mesuIyj0g "os * Goede S + HRA S + S | Sssooessoessesssssssss Krooor~ S | = . . . . . . . . . . —) == i VAI st rt 1 OD —) + A) —) | HS 19 29 XD SE 00 LO b= oo © —- rt 9 1 oof irate Do =H oo 2° eel . . | IDOHHOHOS HH HOD S | =) + = 1) ail + | S | ooo lala ra OUAT SoS leteats ooo + NO rat rt rt o°o° } I =) | re =) + im + 4 =) | S + al =) + nr) » MOUOUOMOOOMmOUOOD OOOO OODO OOOO ©) (e{ Je) (o\/e) |e/(6y1e 0 eelisliv, oetelale) siekelslenellalerery oO 1D =H OO NI SHO B= 00 10 AI be AY HD b= 19 19 ‘ —) re | —) + >) THe MORRO KHOR Or FEFEEFETETTT Tt © OD 4 SS TH YD SH b= HOO HOD 9 2 2 OOS ODO OHS WO-DHDreOSDWOr~O CD Be 9 6 1S 1 SLD LS HS o) a) ait nN =) ae Worle) —) | 110) 8 SH AC =) | SAHA NALLYNM vA Te te th tte ra iek er eee nage tiene es LOC TON OTH WS Om ae: Die PODI BVOC OODROOHOOOROOGO OOD OCOS jrriaqnayay els) Oe 0 Gis 7 “G Coado 0 BT GOSS DITO OG (0 0.010.001.0500 FarsimeInyiy 0 o— T0— ot 5°()= 1e9(N= ‘G TOPCO ODD OOOO GOOG OCOD GSS OOO Tah iKOlty5) 606T 806T LO6T 906T SO6T FOGT SOG6T GOGT TO6T O006T Uvoy -o= qm OS $°0-- nO “O KeOok Or tod hid SHH ++4+++ tor oo at S> 00: oo at NA [—) | 69101010 osoo ++++ Ld Tet bl td AROTOWSKI, CHANGES IN DISTRIBUTION OF TEMPERATU re S | Ole ts ponuyjuoj—suoynjzs unadowng sof dppp aimyosadway—]L wavy YORK ACADEMY OF SCIENCES Ww 7 v) ANNALS NE 58 TAINO ° | Neos ae | matt ° | “oO Gi0-- “oO G0 = Ee Wi o50 « O— 10 G20 20 L220 = 0 L:0-- oz 0 L20-- ree — Gu0 = 6O6T a1) ik 6°0- gto— Qeoa €°0-— Cie LO b°O— 6° O+ S20 60> 610 — 9°0— 20S c.0= Os 9°0— SOs Se0 me (Ga S20 9: 0=> JES Yo 930s 6-0 = (5, == SDV oe 60c- We Uies ¢°O- o0= i Ole S20 (Gos ¢*0O— SO ae c.0 = Go) a sa ry 0= 20> SOG6T o- oF 0m 70 — 9°0- S20— Hee 0 60+ P10 9°0— 0 G0 Lys a0 *-O 0 = mo— Cs0e 5-0— Ot [e20 = aT) (f°) m0) o- mI OE t= an an i in Son I oo oo i a) . . i Pn) | = + LOGT © OA) SoxAAAAOoOOonnHSo DANSOSOWS $4+4+F44+4+4+ oO OD ON i of VD srs HORI HOH HOOD —) ZO+ C2 Ot CO HH Ost th RM CNC ASN Ipnparnal aunt +1 +4+44+ fe + AMNANRA Ate —) Om “O Ee PO Gains GO se Ge G20 cOGT Wo HOO | VO 1D SHO a-ss occe (4 Vier CO OD CD ved CY 6D CO ASS SH YS SH SYD SH CVD 00 o Ot oo a UH TOD 1D o + m —) | oooooo Cats eT Ftt44+ ( 6°0+ 8°O+ FOGT or 1 fon) punisure TID OVNI LS HD OD o lO HOWDMION DAD o TOTHAN CONRAN —) SO6T NUON ret HOO rst ace ne | DFG CaO a OM ees 850 oO ae 6n0 = G20 CAO Om ae 90 90> 8I1O aa ole! LS _~ rE DRACODARSE bao | 1O=te 10— at) Oa. o= 10> re 0— ZOGT onion! SHAN TO6GT cont 5 | ‘o- D1 SO rt HOD SH EGE OP UPEEe o BHOOnOnOHSoSSHHSSSHSSOSSSO Se tet to LO CAI MONO Re ONO RIG He Ol O10 WOr~inoone oooooocorn FH4+ttt4+ ~ io.) =| oS S = IMOAMOCOS COMO AM DIDMONSMDA AM NAAIOC CWA 2 OO CID HINO OMHID OO t= OD DMN AOWDAAMAHASrAONoOW eel ts Varitek ake) Vey re = = rie be oe De Do ‘a penunu0oj—suoynjs uvadoing 40f Dyop aénposodwat— TI eee aaa ee ee ee . ueZzaJY os) 0 eee MIO ON GELS (=H eV OD | Cece ewer ees YoyoOTOA TAuYOAA cress WBSIBAA "ho" THTBANOG +2221 BABDULA teeees nRUdadT "S.ANQSI90T YS thee eee Hayy cele Cs ekslieus) eke lene . POU OO ODS ORMO UO) (Ely Puc) Cont Chai C eC CQ) Ont ria Ch kit yt GeO ath Crore oe IBASU) sale, oiuieleds ellos telielini eine 6b Wel 0! ella EO OES GG, ee ed Od oaeo . . tosses NagazaG euotajate (aletelege le) svelelelaunis se eae eee ecuRquemyos @hbieje::ee)elene7eis' «1b ellepelsice o. © \0 fe) o)'e) ele! sisialie AUOSA0g COR OVOmC OT como mcd od aoc POD OE EEO Si 9 5 o* see oa Geo b Oud OO, OC OD OG eee BLRLD,.O . . Cid Cech Of Cis we . Peres ce ere. ee . yeldezs¢Q of 62a) [o el'e).0l\6 91 Wie; © 6 \ejlacsi'b) vlie)'s\ : viiniiele [OQ OUIEN je \nsielle DOI IOSCIOOH OHIO DO'S 210'0 WOOD'S DO CONT Ys BV ASE INI €\0.0| 06 0 ele 101.9) 08 0 «6's 8 (6,0 Sseieve) eerie AAS ZOTOS| HO cen Oc ry ist) see Piro Grn an A]TOUIZSeY . ©) s) eases ie) \slisie sie Suelb) el miele! vten eis ee ee wwe . RSoo[Ry Oca ke xelefeielisisin susie cietegexel stelle . eas) AT OTOE e)e)\ece' 0] (eles sets/isL slim One Cc chon: Of Oe) Heel) Ato s10%) ele) ere) oie /6, «\0, 4) 01 e're).e! elie\Py.si.elie (sib) ¢felellsielle) stele \« eieus oun yy OD OOO 5 Mao ol ew ore) o:\e'0).e (eile levels ve..0)\0,16 1) \'e\/s\ ee ieliginiie 1030] . shecet were Jol Neila cele [61 .ee.veLerie| wae MEE er ene UIZdaI1q aq] «) shale ev) ereve ele «) 6) «, eenbevel6 alesis 6 wiels pIByRBAISNG a eve. 6) 6 Ge) BU) eLeve wee 8:8) «10, S)6\o).cieileice [eitereielvitel. by.) ercehe 9dT[OS}} 04) 2B \s).0\0)(e eel ©) \e//s)\e/le)ce) e/a) 8i.ei(njieie oire Lelia! ae\lellelsjieliwsieie yoeqie’y] alav L 59 7 a RATURE 7 4 SOSrPDOeresao~t CIN HCO ORO IID nate A ION He AAS A TLD NS Os LO et oH O89 DO. Nt 19 OO ¢ ™= ~ =) Str uc +FFFH o 1D DS rt§ S BO 1D 1D SH DO SSSSOSnnHSSHOSSSSSSSS Inn A NADWOHAMMIOSOSOM-ASW a o VO Od PAOD 9D SL Be tN SH st Sonodoonn ++EFTE+EEEELLTT EFF F THT FEFFTFHTH F444 | MD HHDARDNDDOMNAHAMDHSRIKOS HHOnnOn On RHOOCOSCOHS ARCTOWSKI, CHANGES IN DISTRIBUTION OF TEMPE ho fon S coal i?) > lor) an S =} o> m i) al il Cleve. 6 eleLe s eue\le * BIQRIOD “IL DIDS OPO OO DOM OOOO HOODOO TOO O00 25. 4:yAnaqzi=hagoyg) °3 ME Nk ain Ls CROCE GOIN Cr STEOTOH AO TG gy AaKU Neen: fa) “OL Dace Utena) Ce OG no 0 sMovielsey ses excuo? The Naar Sa Ta TG a) Fay) Wee ODER O-E IAC DOO OOS CO OSLO OOO ary A ata ‘OL QHele\ts e\le!ta\le (eae eb le lle\is/isiellbicsiiaielelsie JAIBI[[ I -JSeIByON| HY] ep en ONTI ONE OF: C390... 050000 Oo ooo OG 0 10] olfejye e878 BBA °3 POC OSLO =O. OF 0/2) OLOLO CWE O00 9 Co (eu fsTosMoys | 6 TOW Od Goce O SET G.D WOOO. HOOD OO OO EELS 4G BECERRA LOND Cs OOOO Ayn racy oth acte & Oe ee ce sits) (eloliets OO God oma Od Ot ‘** Dsa1BULIy (GY) i eam ROO CLONE OOO! OuDsOAGyGud-C 0.0 GO /A0°OUN POG FOUR ED MCHA: VINVIWNA S | Te) oo Ne | 10 i=} xe 2009 =) > £2 Of =~ — as So Se rN oe BO Ssesssseesssso rererererer or ererererer Pell all fe ~— ~~ = — oo [1 IS ++ Ko int = VNU CD De OY rt 0D HEAT CO et | 11 | Yer) n S = ++++ as Si | | ANA eee 5 a) =F | Crna Ccon i) | =) | + 1D | an | i) | t=b =) + 1D S | =) ROA +H O1O Oh AN On Oey Ost C0 OOOO a GOD EOC COMO ODS ONop ihel SHE OOOO OOO TOO OG O-0.000 O0 soe kenont Lac PAT TOA) [et EE OUI ROO COO VOO OKO CG DO00 0.000 ¢ “') XTaRzey Crap, D0 G © ee 0 60 ele 0 8 see sc Ureirevisiiele PHOO DIO GS Heyy ifs OT ii POCO OC OCD OUOOOCO ODIO CO OHO OO SOO CO Ty y EN Can CRORO rCHOeCCeD Sec Osteo BLSIEIS Aji Soro ooaovTsoncongnoooDeDodO00C ** Jos}soaqg 6 OL ORDICEONG EMO cn ct cen eee ee eee OOOO OO UT Tay \oy gel! 9) Mit SP OROUEO OH ov oowo DO KEK n og OOOO FOU SOO psa, 6 e. ©)\0) 0) «) efe) eee 6) 60) 6) vre)/efe).e) velo 5) vn SORE sue hacen TOW oO al Ned o°ooo Lee 19 6 b= AD +4+++ cl UA oooo eal 1D OO 4 o + ee sessesesss Ft+t+ +4++ Nod wIOR o o | | —) | Breas leroy eae Te S + QU OT | Te) + <=) | merioe WN i) oa OO~OM- Ano inl ce Immo = —) | —) m=) <=) + =) oa Si ees | = + Oil Sou soUoa man oD Oooo UOC OOK pHemetierteferels “+ Bssapo e9Q tees PROP OIGO CEOICL GAC) DatniO O-0.Gec0 O sivele retetelteitels URzZRYy "eT DIDI DORIC D aaIOTI G-G1OI GOGO COOH OOO SET ay TeV ON, De @) Gre) /6.(e) 18, (6. e 10 10,067 e676) 81.4 ee w6@ bebe) ee see Jodoyseqes Oat yay OP Om O40 TO.) Oe CO ot oh OO GON JNOSUBY YAR, aje LeAvOMIOMIoo GOOG ODDO DODO DaO004 “555 ARIONIN * . See . . eee . . PBLSIOARST [OT DIDO OID DOOD DOD OOM SO DUO ROOD Tal nq] oMenenewensnetieme os teeee oly lee) ss biemyel sheae Win eA OR SON , - * UBTYRBIISY pee Ne SOR CA CUO Cee OCS A VIO ACH) UL) 6) Sse .6) 0) 6/6) §\e)s) #0) eres eis) a) One) See! 6) a) ae SANQUL1I9} BA eo DOIN is hey HOMIDIT OOO GU GODOT COD DU OOOO OD Gyr K Guay e)a\[0 7s Foaie (0) ev) altel (eye) seus oe sree enh TS TOS TAIN GT Tomo Uo mond clo cd-d 700 ee cece esc ce ** BUTIOT, CD LOuD. Olan Oro Ly rote 0.00 0/0 Ooo 0OnD ih eB OTON cece ee DIO TO OCD Cree O1Or0 IanO 0.0.18 PeOO + jodos1ey OPO OO Oe OO ONG Peo oO CO Orla 7 ORO Or YS[esuvyyay 1 a1 1 = + | > | 19 | PIS Bo all oa SO Ot= 00 0 Jo | | 1 S ae 2o1s + co —) | + 5 onl re —) oA aolVe o OID 1D ANS t= =) + —) | ~ Orig. —) YD ACD SH NU SHAT CVD BCD CD CY =) | mS | FFttt+t+Fttd44 a > OD ret Sd vat SH SH ONS. SoscssonsHsssseHss rPOMnR Hen teS SCANNSSOMAANNTAAOMr-D or) Sol | =) + AMOS Be DHODM-MOOr~ i=) + 1D 9 4 OO hes e°o+ 9°0= “o+ g° (a= S10 D,°O= FOGT S061 OGL TO6T 006T uBoyl ponulyuo0j—suoynjys uvadoung of vIDp ainjosadwaft— J], ITavVL +4 F4+444444+44 —) | FH EPTH+TTt Ht ttt Ft a ° | Srmotienassenonase mn OnnnnOOonSS: 9 G z Oct L I I 19 4 =) for) on BIBp Isoq} Udy ‘OTGT ‘129q01900 Ur peyst;qnud useq Jet JOU Surarvy sarod R Q Sa aa Q ~ =) R Ry ce) ~ ’ S ? ica} Q =< ¢& a" 610 = 82055 Do £ rai ey AX) x 0 0 TeObts ue ais ee OnO at: < ; ‘O- mT) Re [ee Si Oe 8°O+ oo) [Be oF 20 — E20 =te ma T° ; 20 = GalOste _ us “j= 0st = es Tae 20 SAO ica) 9° Oe (\= tO > ¢ : (= -°O+ ; ¢ (bo (i oe 2°o+ R a 1Ak= 6 ie o re ~*~ = — re DOAN S Co UveTY ana O-Diaui sesso IISBA verse? TT1aAVG-GIny, ess QFOINSBIY-WITL FOODS Aho aygyauiriaey y, sire iS}ielleheheyeheuece! eye) pie edefellelelenelelic) ole) =) sceleynin DIE-nS0e OO OanO OO chuck ChOLOtr Oo te Co OO g isle .6\[o)¢ pre) eels RUI[OS * JaIBYLUS ayevareleneacnonsite BIRYOS POCO IOTNO a OOO © moro tho red s)s),e oh eveneneyaWals BIRUIN F9-OL00 }BAVG-NolUWBY ean eer ts s801d teense eee Tsaqty “'* (LOSBIG-Se0U1k | arayeyels ian ecie BIULLOWIRIY eee ee “i-)- le BODOBS |! Sree Choi MoO O Oo oo MO Now & SEO ODOO ARSE suoypys Uvadoing 40f vypp ainposodwaf—JI 2 T4¥L er ee ee ee 1)sa.l0]{ ee ee eee see ULS.ITN*) 6 1e)\6 8/0. 6.9.68 . iyedwmiys) e)Je je)(8/\6 > e 8) pte. 7m . ZYB[B) CHO. aOR eyo JUBSDOST ee ee ee . TUdI [Ro] On Ofce Oh cr fuesnseBig SPist oe TOU OLOGL feels vo) s)s sae BAoled,) ARCTOWSKI, CHANGES IN DISTRIBUTION OF TEMPERATURE 61 The records of France, Spain, Portugal, Italy, Servia, Turkey and Greece have not been taken into consideration. For Russia, Austria, Belgium, Holland and Switzerland I had only a small selection of data. For Russia, in particular, the number of utilized stations is absolutely insufficient.16 For Great Britain I have taken the district means. In reality, then, I have taken into consideration only Scandinavia and Cen- tral Europe, and European Russia simply for the sake of orientation. It would seem that the figures of more than four hundred stations ought to give a very accurate idea of the variations of temperature which occurred during 1900-1909 in Central Europe. This is true only to a certain extent. First of all there are some local complications due to orographical con- ditions To try to discuss these complications would lead me too far afield and would necessitate still more data. I had in view simply to get a general idea of the geographical distribu- tion of the annual departures of temperature and, for that purpose, I had just enough data. The area covered by Scandinavia and Central Europe is absolutely in- sufficient to give the necessary maps for a clear understanding of the climatic variations which take place. Europe is but a fragment of an immense continent: the old world of Asia, Africa and Europe, and the variations of temperature which occur in Central Europe evidently de- pend on those which occur in Asia, in the Arctic regions, on the Atlantic and perhaps also of those which occur in Africa, in Equatorial Africa and the Sahara in particular. Central Europe is probably the least favor- able spot on the earth’s surface to be taken into consideration for the study of climatic variations. There, the variations are far too compli- cated to be understood easily. It would have been a great advantage to me, if I had had the data of all the Russian stations, those of Siberia and Turkestan in particular, and also the Indian data; but then I would have had to face such a number of new problems that it would have been quite impossible to stop the research work in order to write down the results obtained. The ten European departure maps which I publish now (Figs. 12-21) are simply first material for further researches. These maps are most suggestive for many special investigations. In order to advance, how- ever, I will avoid details as much as possible and will pass at once to the main question: the cause of pleionian variations. 1¢J7t is my intention to study more in detail the variations of temperature which oc- curred during the years 1900-1909 in Poland, the Russian Empire and India as soon as circumstances permit. 62 ANNALS NEW -YORK ACADEMY OF SCIENCES Besides the Scandinavian countries, for which I had the complete record of observations, I had at my disposal the monthly means of tem- perature of Bucarest, Kazan, Warsaw, Odessa, Aachen, and Geneva. I made consecutive twelve monthly totals for these stations and also for Bodo, Sydvaranger, Haparanda, and Vestervig. On the following dia- grams (Figs. 10 and 11) I reproduce these totals graphically, together with the curve of Arequipa, which will serve as a type of the direct solar variation in equatorial regions. Bucarest. Odessa. Warsaw. Kazan. Arequi pr: 0g; Fic. 10.—Ourves of the consecutive means at Bucharest, Odessa, Warsaw, Kazan and Arequipa The striking fact which is exhibited on these curves is that the varia- tions of the European regions having a frankly continental climate are radically different from those which lie under the prevalent influence of the Atlantic. The curve of Kazan exhibits tendencies of increase of temperature followed by tendencies of decrease in regular successions, a variation repeating itself independently of the seasons of the year, just as in Arequipa. In Bucarest and in Warsaw we have also the typical pleionian variation. The curve of Aachen, on the contrary, is absolutely different. There we have small ups and downs entirely disfiguring the ARCTOWSKI, CHANGES IN DISTRIBUTION OF TEMPERATURE 63 pleionian curve. The Aachen curve is characterized by brachyechronic variations of small amplitude. In northwestern Europe, then, the brachypleions must have a pre- dominant importance, whereas they do not affect eastern Hurope very greatly. The curves of the Scandinavian stations, on the other hand, belong to a mixed type of pleions and brachypleions. Here we may have Geneva. Aachen. Yestervie. Bodo. 4 aparanda Fie. 11.—Consecutive temperature curves at Geneva, Aachen, Vestervig, Bodo, Haparanda and Sydvaranger a succession of several years during which the changes of temperature will be slow and great, as in Russia, followed by a succession of years of shorter and more irregular changes entirely different from those of Russia. It would seem that there are pleionian and brachypleionian areas and that the border between them may temporarily belong to one area or the other. In reality, however, things are more complicated because the pleionian and brachypleionian variations are coexistent over more or less large areas. The comparison of the curves shows this very plainly. 64 ANNALS NEW YORK ACADEMY OF SCIENCES If we now compare more closely the curves of Kazan, Warsaw, and Bucarest with the Arequipa curve, we notice sufficient similarities to grant that the primary cause of the Russian variations is most probably the same as that which produces the equatorial variations. The main cause of the complications in the geographical distribution of the excess and deficiency of temperature has to be ascribed to the perturbations of atmospheric circulation and transport of water vapor. We have to admit that if the value of the solar radiation changes, the temperature at the earth’s surface must change; but the total atmospheric pressure remains the same. Consequently, a rise or fall of temperature must produce abnormal changes in the distribution of atmospheric pressure. ‘These changes will affect the winds, the rainfall, and also the temperature. The normal, or let us say the Arequipa, variation of temperature must therefore undergo, in different regions, all sorts of modifications due to the local conditions of atmospheric circulation. This fact explains the coexistence, and mutual dependence, of pleions and antipleions and explains also, to a certain extent, the persistence and more or less progressive displacement of the pleions from one region to another. On the other hand, some, at least, of the brachypleions may be con- sidered as peripheric trepidations of the pleions. At present this interpretation is evidently but a simple working hy- pothesis for investigations yet to be made. It will, however, be suffi- cient to compare the curves of Aachen and the Scandinavian stations with those of figures 59 and 60 to arrive immediately at the conclusion that the brachypleionian oscillations are not at all a particularity char- acterizing the purely maritime climate of oceanic islands, as at first one would have been inclined to think. The temperature scale not being indicated on the diagrams (Figs. 10 and 11) I give in the following table (Table III) the values of the highest and lowest consecutive means and their differences. These fig- ures are °C. It would have meant too much work to reduce all the fig- ures utilized to draw the curves into mean temperatures and into °C. The utilized figures are simply totals of twelve monthly means. In the ease of °C. I added fifty to all figures in order to avoid the negative values of the winter months. For totals of °F. the figures of course give an apparently greater amplitude of variation to the curves. The pre- ceding table will serve to make comparisons possible in case anyone would like to examine the amplitudes of individual crests. For my present purpose such comparisons are unnecessary. ARCTOWSKI, CHANGES IN DISTRIBUTION OF TEMPERATURE 65 TABLE ITI.—E.treme values of consecutive means Highest Lowest Drees (STONE) Be clot 0 COO Cn 6 COG IORCRDIDIO Te oar 5 8.4 L9 PACH OTe ca etal oyreactareraialieweusitowenclsic tesco! s/c) ose aha uneuer onsitalie 5 8.2 a) WMESCOIVS ei ctaicics 6 ce) cual anerster ere ice rclie =) s)-sp 3) initer oeieneire tet lerle lolee xe 6.1 Die SOO epee fence ey onr eae ore eueitelle nila Sle oviel.s: wlesle topes erorctarat : 29 2.3 Sydvaranger —3 -o 4.3 EVA PAT ANG (ere t6 0/010) e)= «1= ele) ee) 1s = = = : 5.566 2. —2.3 es SICH EOS Urata eee ha niece eh areas tener hskeriskete: @usi@rersiensyenere .6 9.0 2.6 GES Sar ory cal seal cae av cue faleitevie, ay sh nclovsrceoun (6 9p Bi ae'e ene eibas lorie 8.4 So VESEY? 36.6 di Oo Gc Olo SiDoinItA Gcheae OF alomioralo chen 6.4 229) TSEAANIRY . Se SIERO. CROCE SAIS Ck CREM n ERR Rete 1.4 4.6 Passing now to the description of the maps, we will immediately realize the usefulness of the curves of consecutive means, because these curves eliminate the possibility of hazardous speculations about the dis- placement of the pleions from one year to another. Instead of such superficial considerations, we will find the way to study systematically the progressive transformation of the maps, a task which I cannot under- take at present not having the monthly means for all, or at least a large number of stations. The map giving the distribution of the departures for 1900 (Fig. 12) is practically identical with the map of the same year I have traced, utilizing the departures from the means of 1891-1900." This demon- strates very clearly the fact that annual departures from ten yearly means serve perfectly to indicate the position and shape of pleions and antipleions. On the present map the quasinormal line crosses Denmark, Southern Sweden, Curland, and forms a curve across Russia toward the Azof Sea. North and east of this line the departures are negative, south of it they are positive. The antipleion forms an immense wave with two centers of lowest values, one in Scandinavia, the other in Eastern Russia. In Sweden the greatest negative departure occurs at Quickjock and is — 2.0 C. The Russian data do not permit of locating the eastern center of the antipleion exactly. In Kazan the departure is —1.0. The highest values of the pleion are + 1.5, in Hungary. The pleion is broken up in central Europe by an area of low values. In southern Germany and Bohemia, the departures are below + 0.5 and decrease to 0. along a line extending from the Belgian frontier into Bavaria. The map of 1901 (Fig. 13) shows a radical change in the distribu- tion of temperature. Where we had a negative wave we now have a positive wave. The pleionian departures are + 1.2 in Swedish Lapland and + 1.3 in southern Russia. The quasinormal line goes from Great Britain across Germany and Austria towards Rumania. The lowest departures of the antipleion are — 0.7 in Erfurt and — 0.8 in Gottschee, in Krain. Op: (Ctts, py 121 66 ANNALS NEW YORK ACADEMY OF SCIENCES So 4 \ ——~| AS) Fic. 13.—Temperature departures for the year 1901 . orto ~2e5 oye =e Fic. 15.—Temperature departures for the year 1903 68 ANNALS NEW YORK ACADEMY OF SCIENCES Naturally one asks whether the antipleion of 1900 went down or the pleion went up. The consecutive curves will give some indication about the displacement which took place: The pleionian crest passed Kasan at the consecutive mean: Sept. 1900-Aug. 1901. In Haparanda, Bodo and Sydvaranger the crest occurs at the mean of Noy. 1900—Oct. 1901. In Vestervig and Aachen, Feb. 1901—March 1902. We must admit therefore that, most prdébably, the displacement went from northeast to southwest, but the map shows the existence of an antipleion over the White Sea. Moreover, in 1901, the consecutive curve of Kazan is on the descent. The same is true in northern Scandinavia. If we consider the dates of the occurrence of the minimum, we find: Sydvaranger, Haparanda, Bod6é, November, 1901—October, 1902; Kazan, 1902; Vestervig, Warsaw, Bucarest, February, 1902—January, 1903; Geneva, May, 1902—April, 1903. There is, therefore, a progressive in- vasion of a negative wave coming from the White Sea and spreading out towards the southwest and south. The map of 1902 shows plainly the importance of this antipleion. This characteristic antipleion, with a departure of — 3.1 at its center in Mezen, follows closely the first depression of Arequipa. Therefore, a de- tailed study of the meteorological phenomena of 1902 would be most instructive if one took, besides the European data, those of Asiatic Russia and India. The distribution of the equideparture lines on the map (Fig. 14) shows plainly the dynamical character of the phenomenon. It would not be very difficult to find out how this antipleion invaded Kurope and the reason why could be traced as well, and correlated with the equatorial variation of temperature. The map (Fig. 15) of the departures for 1903 is just as interesting as the map of 1902. There is an important rise of temperature over all the area with the exception of southwestern Europe, Ireland and Scotland. Now there is a pleion centered over Russia, where the departures are + 1.5 in Pernau, Vologda and Vychnyi Volotchek. What became of the antipleion of 1902? Did it go towards the At- lantic and the south, or was there a rise of temperature in situ without any displacement ? The consecutive curve of Geneva (Fig. 11) shows that the antipleion of 1902 certainly did not cross Switzerland to go south; but the curves of Sydvaranger and Haparanda are very steep immediately after 1902, the curve of Warsaw (Fig. 10) shows a regular and progressive ascent from 1902 until 1903, while the curve of Bucarest, on the contrary, shows a slow ascent followed by a very much faster increase of tempera- ture towards the end, and there the values remain high till the mean of AROCTOWSKI, CHANGES IN DISTRIBUTION OF TEMPERATURE 69 1904 Fic. 16.—Temperature departures for the year 1904 Fia. 17.—Temperature departures for the year 1905 70 ANNALS NEW YORK ACADEMY OF SCIENCES Fic. 19.—Temperature departures for the year 1907 ARCTOWSKI, CHANGES IN DISTRIBUTION OF TEMPERATURE Yt September, 1903—August, 1904. Consequently, consecutive maps would show that the formation of the pleion began in the north and extended progressively south or southeast. The map of 1904 (Fig. 16) shows a distribution of the equideparture lines similar, to a certain extent, to that of 1901; but the sign of the departures is reversed, since we have now an antipleion where we had a pleion in 1901. The consecutive curves are again a great help towards the understanding of what happened. lf we consider this depression of temperature as being due to the same cause as that of Arequipa, we may say that in Kazan the reaction is felt first, then in Warsaw and. finally in Bucarest, where it is very much retarded. In Scandinavia the phenomenon appears to be more complicated. There we have two distinct depressions. One is coincident with that of Arequipa, as the curve of Vesterwig shows, and the other is greatly re- tarded. It may be that the second depression of Haparanda and Sydva- ranger is due to a propagation of the antipleion first formed in Russia. The map of 1905 (Fig. 17), if considered from the same point of view, represents the formation of a pleion and that of 1906 shows the same pleion after the maximum of its development. The most important crests on the curves of Kazan and Bucarest occur between 1905 and 1906 and correspond to an Arequipa crest. In War- saw we notice fluctuations; the same in Scandinavia, where they are even more pronounced. This pleion must have been Asiatic. In 1907 (Fig. 19), there are residual pleions over Scandinavia and central Europe and an antipleion over Russia. Between 1907 and 1908 the curves of Warsaw, Bucarest and Geneva show the Arequipa crest. This fact demonstrates that the Russian antipleion did not spread out progressively over central Europe, as a comparison of the maps of 1907 and 1908 seems to indicate. On the contrary, an important interruption occurred, during which a brachypleion (corresponding to the Arequipa crest of 1907-1908) came from the south and invaded southern and west- ern Europe without affecting the Russian antipleion. Finally, this Rus- sian antipleion of 1908 went west in 1909 (Figs. 20 and 21). The con- secutive curves, the curve of Geneva in particular, leave no doubt about this fact. The main result gained by the study of the maps is that, during the years 1900 to 1909, the pleions and antipleions did not move from the Atlantic across Europe towards Asia. On the contrary, the displace- ment was from the northeast towards the southwest, or from the east towards the west. Moreover, these displacements did not cross the areas ANNALS NEW YORK ACADEMY OF SOIENCES Fic. 21.—Temperature departures for the year 1909 ARCTOWSKI, CHANGES IN DISTRIBUTION OF TEMPERATURE 73 of maritime climate.. One may say, therefore, that the big pleionian variations of Europe are a purely continental and, perhaps, Arctic phenomenon. It would be premature to attach any importance to the locations of origin of particular pleions. The areas where they are formed, in situ, are probably not always the same. Besides, the question whether a pleion is of an Asiatic or arctic origin has no importance for the present, simply because it would be absolutely premature to discuss the reasons why, under the influence of a temporary increase of solar radiation, one loca- tion of the polar or temperate regions is more favored than others. It is evident that a temporary increase of energy radiated to the earth’s surface, during, let us say, three months in succession, will not directly influence the temperature of the arctic regions if if occurs during the winter months of the northern hemisphere, whereas the antarctic regions will be greatly influenced. The question of the formation and development of pleions, outside the equatorial regions, must be studied together with the seasonal changes of atmospheric pressure and the temporary alterations of. atmospheric circulation, nebulosity, rainfall, ete. I intend to make such a study for particular pleions and especially for brachypleions. The fact, however, that in Russia there are some striking coincidences between the forma- tion of pleions and the crests of the Arequipa curve is a most convincing proof of a common cosmical cause of these variations. AMERICAN TEMPERATURE Data For 1900-1909. Since the equatorial or Arequipa variations of temperature can be ob- served not only in Russia but also along the Atlantic coast of the United States, in New York in particular,'* it was really fascinating to follow more Closely the changes in the distribution of temperature which oc- curred simultaneously in different regions of the North American con- tinent. Here, it was possible to follow the phenomenon from ocean to ovean, over a much more extensive area than that of central Europe, and, this area being more isolated, it is self-evident that more definite results were obtainable. A research, apparently similar to mine, was made long ago by Helm Clayton.1® It was only after my investigations were nearly completed that I noticed the fact and Clayton’s writings have therefore not at all influenced my work. Clayton studied the monthly departure maps pub- 18H. ARCTOWSKI: “On some climatic changes recorded In New York City.” Bull. Amer. Geogr. Soc., Vol. 45, p. 117. New York, 1913. 1H. HELM CLAYTON: “Weather changes of long period.” Amer. Meteor. Journ., Vol. 2, p. 126. Detroit, 1885. V4 ANNALS NEW YORK ACADEMY OF SCIENCES lished in the Monthly Weather Review of 1884 and 1885, and for the sake of comparison I give his conclusions below.*® The results I have obtained so far are so different from those of Helm Clayton that it is perhaps necessary to insist once more upon the fact that my maps are annual departure maps, whereas the maps utilized by him were monthly departure maps. Even in the case of monthly maps of temperature, however, Clayton’s generalizations must be Considered as simply plausible hypotheses, which may disagree with the observed facts. The departure maps for the months from August to November, 1912 (see Monthly Weather Review), will serve as an example. The method of research I have adopted is evidently the same as that of Clayton** and the pleionian variation, of equatorial regions in par- ticular, is certainly the same phenomenon as the meteorological cycle of twenty-five months’ duration discovered by Clayton,** or the longer cycle, of about three years, advocated by Lockyer and others. Though the method of using consecutive means, and tracing departure maps, has already been used long ago, it has not yet been applied to the scientific study of climatic variations in a sufficiently extensive and careful way to lead to the results of general interest and practical ap- plication which we might expect to obtain. Before entering into the details of the description of the departure maps of the years 1900 to 1909, I will take into consideration the geo- graphical repartition of the range of variation of the annual means of temperature. On Bigelow’s tables** I have formed the differences between the high- 01, There are areas of barometric depression, and elevation, which occupy weeks and months in their movements across the continent from West to Fast. “2. There exist, Independent of the movements of areas of barometric depression, and elevation, numerous seesaw oscillations in. the pressure which have been given the name of surges. “3. In front of and to the south of areas of barometric depression of slow movement and long duration, as in those of rapid movement and short duration, the temperature is above the normal; and below the normal north of them and in their rear which is usually the front of barometric elevations. In front of, and to the south of, areas of barometric elevation of long period, as in those of short period, the temperature is below the normal, and above north of them and in their rear which is usually the front of depressions, (In winter the area of warmth approaches and usually includes the area of lowest pressure, and the area of cold approaches and usually includes the area of highest pressure; in summer, vice versa.) “4, In front of, and within, barometric depressions of long period, as in those of short, the rainfall is above the normal, and below. in their rear. In front of, and within, barometric elevations of long period, as in those of short, the general tendency is towards fair and clear weather with deficient rainfall.” 71 HENRY HELM CLAYTON: ‘“‘A proposed new method of weather forecasting by analysis of atmospheric conditions into waves of different lengths... Monthly Weather Review, 1907, p. 161. See also: HENRY GAWTHROP: Temperature curves (Jbid., p. 576). *2 American Meteorological Journal, Vol. 1, p. 130. 1884. 23'On: cit;, Bull: S: ARCTOWSKI, CHANGES IN DISTRIBUTION OF TEMPERATURE %5 est and lowest annual means. The geographical distribution of these figures is most interesting. The highest differences are those of Bis- marck, Duluth, St. Paul and Marquette. The figures are respectively 9°.8, 9°.7, 9°.5, and 9°.4 F. In North Dakota and the Lake Superior region the range of possible variation of the annual means is therefore above 9°F. From that region, the values diminish progressively towards the east, south and west. The line limiting the region where the differ- ences are above 5° goes from Portland, Ore., towards Salt Lake City, North Platte, Hannibal, Lynchburg and from there northeast, along the Atlantic coast. The difference 7°.2 for Portland, Me., is too high. The series of observations taken in Portland, Me., is evidently not homo- geneous. The values of 5°.5 for Los Angeles, 5°.8 for El Paso and San Antonio are also probably too high, since the line of 4° goes from Eureka southward over Sacramento toward San Diego, then eastward towards Little Rock, Memphis, Atlanta and Wilmington. The lowest value, 3°.1 for San Francisco, and the value 2°.8 for the shorter series of observations of Corpus Christi and Jupiter, are not very much higher than the differences 2°.1 and 2°.6 of the pleionian crests and antipleionian depressions of the consecutive curves of Arequipa and Bulawayo. It follows that if, all over the United States, the varia- tions are primarily due to pleions, having the same cause as the equa- torial pleions, the phenomenon would be four times more pronounced at the center of the North American continent—in Winnipeg, let us say—than under the equator. Of course, in the case of the brachypleions, the difference would probably be very much greater, and if the results obtained from the study of the interdiurnal mean variabilities of temper- ature? are taken into account we must be impressed by the similarity of the results obtained. Evidently, the continentality must have the | same exaggerating effect on climatic variations that it has on the cold and warm waves characterizing the changes of weather. I will pass now to the study of the annual departure maps. The figures utilized have been taken from the Annual Reports of the Weather Bureau and those of the Canadian stations were copied from the Summaries of the Monthly Weather Review. In Table IV, I repro- duce the means of the decade 1900-1909 and the annual departures from these means for all the utilized stations. These figures were inscribed on maps and equideparture lines drawn. The ten maps thus obtained are reproduced as figures 22-31. J. HANN: “Untersuchungen uber die Veriinderlichkeit der Tagestemperatur.”’ Si{tz: Math. Nat. Cl. Acad. Wiss. Wien, Vol. 71, II, p. 571. Wien, 1875. L'ot+ ue cot I Z L°0-— $ CT ed —< Ge — o°T— ZO+ 0° 6F pte ane acho COSA aL CWE OY OOOH ATE ‘os BOIYO L°0-— 20 9~O+ CUR | Yor joe) Ua L o+ oT t°o+ L°Ot+ G°t+ GRP Ce Oe WO i TA CE fo S86 sO LSHSY Sie eehle EES OL AAT ‘auueA0Y,) L°Ot+ tit L°O+ t°O+ eo O'I— 0 I— c:O+ Li- 6°0+ 1°09 Ci CACET TOM n> cers IGT CCH CIOL Cr Hp *Bd00UB}1 BY, ) G07; 0s T205-— 80 E30 SoS 0 S05 6 SoS a0 x0 OOo eee ee "OS eS See 2 Neola Pon Oatee Ole lot> mt Os Osten Gi Ost men (Gin Oa eae) Oa 0 CESS 8 SEAT eec kyo Om ie 60 890sF -Pl0re Oat 6 Ge eee NE ee aoe Ne ae gS OT+ It 8':0—- Got SoG: f opie 0 Ot G'0O— ST GOP cctttttet tte: bitte e eee ees OSSRRY ‘MOISOg SO G0> L0—> $0 C0— Si0> © Silt “ClO 190 Ske MSO nr cemene leo ries neieene eas Cette gta oyBpl ‘ostog ea 20 LIP EI Ott 20-— PS FO 820 820>% SS eT ONC ree are cree eee ere Tan ooye Q TEV 6°I+ gr — tT'o+ 9°0— 2° Tt — 6 o- FO Got tet 8° OF @G).0\'a.® 8 eee) 6.80)8 are, ele, © Beer ac eaoke. (7 (I "'N “‘yOIBUISl = ¢°oOt+ 6°O+ go T— l t+ G0 = oC L°o+ L°O+ 1 o+ 6° T+ G' OF B))8)8 500) 6m eve! S) B48) 6 18)/6)(s) o Bee) 6) ae ere AR “"N ‘uo} Mey sUlY > 8205 G Lat en O°T+ ch0= LiGe “0 GAVE BS O— Os Gitte NOR GGG ie aa eee re aac ee “pW ‘slow [eq eee 0 SOF .6O— $'0— OD ODS S09 NOs: a sen ie eee ene Bag ‘AYO 1oyeg Sp SOs ae EO ae OF FOF 20 20. arOT PES) ONO) RO rc Ome amen eee "to ey “Bsngny S 9:03, S603 (Gt S= (Sette S00 (ONG > PEKOs eon0S eS ee Oleic ammo ene oe Se oe £N ‘4419 onueny e'o+ 8'O+ s'ot a’ O+ F'O— L°0- s°O— S:Ot+ Ge t— O°I+ 6°09 ei ms wetedevee) eo) site she eee e sees Price rl itec et TO 785) ‘BlUuelly > 2°o+ 2°o+ tert 1°O- s'I- 6 o+ L'I-— ec ot 0 L'O+ T'9¢ 6:0) 6) @ -0)1e/\e ,e),0) ele \o¥e1¢)(6| 60. ese =,6 MES <1 6 ‘ol J1eUy =| 1°o+ 6°Oo+ + I— oi It “or S in 0 g°o+ I o— e:tt+ Oz ku ese ove le) .ele a dJeld avalte REO TM OICO OF yi 1p ‘wued| Vy 930—" JA = Sea— Ra G0 8S GeO Saige Oc een Oa sod aus . 8°ot+ €20-- Salas Seilets [DOs ORGies Sin Oe 8°O+ Oise Stee Gad Gia 3 ec nina ies S Se aA oe ie INYILY JLOg Ee fOr Ot SI- OOF -8°0—- FS ROT -BTOSS PF iOs = COti ee cee cee cec ete eae eee he nee punog A118 Q Lor s8:0- Gt= Stir 290 “S96S= —hOP° “Sl O5es 9 Os0sF) 2.0 Ie site orem aie een ee een No[UBIS 710d = F'Ot+ ttt 2 T— t'tt+ L0= ie sot £°o+ ot+ 9°T+ L°GF 9, ©, ©) 6/8) eB) 0 (ue. ©, (@)\p )18) 928) 6. ure Oras Gi eee OF Oe . 0} UO10 7, -°O+ 9°O+ Ce) | i s°ot+ '°T— PA ei ott 8°o+ Oo°tT+ e:tt+ ESP eye) s) (6) 6.8 a) 8 aue))s)\s) (© 8 Jeevan ‘e)'6 oe 018 (p)e1e) 6.9) 1018 aia UOISSULY, R GlO-- Os hele Gale ShO me (ONS eee a0s oe O neon One F°O+ DUGHE Pe ae er ae atime Mab STE NS cn” “"* [Bol UOT | e°ot+ = °O-4- a F'ot+ 9°0— Za e-o+ BOs Soy pae e:ot+ 0°62 een ele}-ai,a)ie1's ual siivj eleiedel avulleltehetelteretanciietelts 77 ** =" dagen? RB P:O+ L0-- T=) «= SO 20 OS COs ORS Ost Om See eee “"julod 19g} By ial e7ot+ 8°o+ e°I-— e°tt+ z'0— I tT O°I— L°Oo+ pit eo G*OF POULIN INOt Oo DO OOOO OOD O.0'O OPO each weyIey) = Leate £°0-+ ONGiae $°O+ Gla SO 820 Sisltste PaGain Sialats aR i een eee COLL OnRLe Ut) x 9°O+ Zot s'I-— g o+ IT S°I- s°o+ s°ot+ S°tt PoIt+ O° SF @0)[eke.e) bys) ese) a) (6 0)6) 01:08 ap (area) is ialatavelle “oo * GjQnoudre x Ett FOr 2:0— 2:0 O'8° S16 BOT OR LeOH eco 8 chy ee ee "*uouRyy puwiy e-o+ 0 ST—- 9'o+ SI—- 8s Ok FTE OSE BT SER Co eet etn tence ene botemign tele XUIUCTI ltt s'ot 6°I- 9°O+ 6°I- 6°I- 6°0- Lit tot 9°o+ G’°GF eve(a)e)efe-a| aim ea eleleleialalatw noises sisi sib isu 6 eeee Soupség 606T S06T L061 9061 COGT FOGT SOG ZOGT LOGI OOGT uvoyy SIIDIS PaplugQ ayy pun vppouvdg a "SUDIWE BSOY] WOdf SIANJADAIP PDNUUD PUD GOGI-OOGL apvoap ayy s0f Cy.) Suvou ainjosodwMat— AT ITAV UG OHO HAI OO HAIG CUE OO rst SH st OID OO CN 1 0 Rss ROR NNWOOCnM On Onn Ae Ke HOnOCOPMRAOnR OM NCD OF ~ wt OO rt SH HCO OD 1D 16 [+++4++ +1)44 +4+4+44+4+14++ “c OnNO-HKot ~~ 10 on Orr MADAMA DANA S © OO 0) 0) 09 19 0 CO b= “ DAI OU H A HD OI OIA SS +4++tt+4+ 1 ++4+4+4+4+4 144+ winrOnrr nt Onwenrne ri ARCTOWSKI, CHANGES IN DISTRIBUTION OF TEMPERATURE oo non 0° Te O78. 0-8" Ora ee Ue UL Ole) One: 6 Pee lee 6 enehe eae apre *IQ9N ‘ajooury G'Fg Ct at CT tea et ae Ch Ce Oe Aa AY ‘Uo0) Sux] 8°GF pein Pia melS Men ages Cy ee ee OL EGY GM ‘leapue'y Pay sursttamranene SIM ‘assoa) By] TSG (ict tee ee eee ences Gua, ‘a[[TAXouy FOL Sra neWexepersrelehats iepereneterete afeltevetere se ints ‘Bla “YSOAL soy a KA 2G para 5 “s'*"BMOT ‘ynyooy LT’ &F OTe a yelve\ecaerelere ise) ee éne) eee eevee le *"VUOTY ‘TTedstjey 8°FS ove 500 teeta he Xetstee ee PE te SC et wl bal Jul “OW “AVID SBesuRy (Naa, Bowen co BrieWe ve) lee ae DI OTe ROT) three Gil (Uae O°SO citi ttt to eeee esses ee oopry QrpAmOsyoRp fo Ae} aa "je Ce Seh\eie¥e fee te ereteie eee e ers yee le ‘pul ‘stjodvueipuy 6 . LG SO kO-e 70ers eee 4s ee elels ie eee be eller 38) ‘souepusdepuy OPP Citi rst ested teers eee eee SrRer “gs SaoanET OF FFP Sie levers es fale jeveie) eke ip 1ele ene 16/6) ene) ene)eltee a i “VUOTW ‘eueley] | a Cee he tee Nee vero erere ie enererene sree enelelerene ba he CO) 1g ‘OIAR YT 0° ZO eres tee vue hu) e) 8s 1808) 8)1aee) Oe) ene) -evele lees) 99) "nN ‘SB.10}}B]] 6°TS SA hase ace a mead $I UND OOF CQ 9 | ‘SINQSLIIe TT ()2CeXe) Be ecrejreste te vce verele:ere: te! eS es enene) boas ane enh OW ‘jeqraueyy a EP a aaa Maine aie ran moe eR OQeOT Ty v\ ‘) or To | | —) | OotnaK © + b= 09 6 HED OO S00 SoC | —) | eed S a ehh oe) | oe ae CD Oot © ab OS 1 St — | 1c OMaien >] | oc Se 100 * ° | be | a ANHDAHRHOHOO rooes | trot | } + a S | S | i) | A 1S | ° | to a) I++4+41 1 14441 144174 C2 6 So ale Yes) — ae — | nN oa = | Yes) S © i | ° bon Nes ) = : ~~ | Sal oo tel a © Shh oe hon =) WHMOOCMOMOMOA TIS HID HO nN | han S | = ~ b ome | as S | oF S | Pot ANDO HAO or for) N ve) 19 <=) re HO AHO TSH NOS N nmOoocoeon — 7 erowtwono +444 144+ fi in 1 1 (= + rah | IS AOACIN ~ = — + TAD AIH NOH OH OS N | INH — + bs io 4) ~~ ° sal DODD ORHNHO~-O-ms baal | = if Rant: ae > LS 5 | o i a re | 420 ++ Yen) _ bo oe) re ) io) — | 09 65 OD rs | 4 + On Sh) P Oost oH | 1 ON 1 + bo Al | + ir = | 1D SHEN LO FANN S On nH OOCONRTNONANRHACOnnA FHAt+H+ AATF HETEFHTHTHHH | AHH H+ 44444444444 | 4444+ i) | Dorma DW nN | re | So | °° | ee rt | rit o31 S me o “O+ "Ose a1) 90 a = ~ S | 7 6) =! | S = 1 1D . S | i | ~ | rr or nel _ rs = —) MDWANO AO HW HIDIOON —) ener PAN HAN ROO A AIS — > i) ao nN | Ko —) + Yes — | 1 5 el | —) | Saal aa poenulyuoj—949 ‘606-0067 apnoap ayy 40f (‘YY .) SupawM ainjoiadwaf— AT ITV L 3201 CD ooo Tall a | To 4K INO H+ patie | onan ~ Or YORK ACADEMY OF SCIENCE IN © O11 00006 + 1 ANNALS NEW r Art OMe OO Heil ooo oo staal Bal | | Lie) —) | VSO rt OU OD 19 1D 4 OO on | | COS HHA AIO Al Hh ooo ono al +444+4+44+44+1+ ‘O+ ro ir ADHOMAAN DION S UHC MmOnnnn Onn NnOn Onn nnOn gp 27120080 SH +L +tttt+tt ttt | +tttt+ SOU HAD HO SD Sh 099 9 a HID SON ONO SO 1 DH HD I AION 89 29 2 a wie 6°0+ T°O— 6°O+ ett c'o— Po 9°0+ 3:0 — Ta ee 9°0O+ 9°O+ co+ 6°0+ G:0— G*T— $:0— Calas c'O+ Go L+ TO+ Leet T°o+ 8°0— L'a eo cco— Okae= 3°0o+ G'0O+ e'o+ G°0-- SiOnc 6°O+ o= rO— OLLs- o+ C0 — Cela (= a0— 0 — (= E30 — c'O+ Let 2 tj = -°0— a0) 3°0= 3) IDM NOOARN OAC HK oo [Se a | Tt — T° 9° ; ir [Bo e— s'o+ o=— (i 8°oO+ O+ 9°0— Pie—= (ESE G30 = Sai Deca -0— GaGa “0 E}())— T°O+ ae N= (2 G°O+ SOma 20} 6°0+ CROs (Bord Te Ont cc0— hag S'o+ c‘o+ 8'O+ T°o— “0 2 i m0) 0 T°0o— T°O+ ay -0— L°O+ 7 0— — Cala 3-0 — Gila La0s a1) F0— Ca0rt BeO= 21 — gs0=- ee DeO= Onis ILO= 6 OnDats mt) 920— (1) es Ze Te €°O+ Lae 7 Cr Oa G2O5e <9? (rar oj — Sil Os 9061 SO6T FOGT penuyu09—930 ‘ 2h oo teit OD ret ret OU SH SH et CU rd OO 3 ON OO ri TH ONO CO Hh He o NO oo ain QHD UH don relia) 00> 10> a Goes S06T HAM RFONMONIMIS sal Pr 0= L° Co — CVD rd ACD de i) | In —) oo aA MOMAMPHANAAAMONK + UH YD ra rt ri OD SH CO CO —) 00 om 1 | OVS OD 41 OO b= SH Ht rc TOD Be OO SO HOV OO 4 2 OO HH HD HO SO AIH OOO AID NID 110 0 ON St SH SH re TO6T NO wo 6) G0 =H VID bh O10 Hh co olny FEFEEEEFEEH LT FEF HEHEHE FETEE FLEET LEFTIST Ht RMMOOMOAMIOMOMOONHOONM 1d CRT OTN ONRNOCOC ROR BRR CORN NONR ROR OH OMNNR ORR ROOR ORO S S f=") re lol SAAMI AM OOAMOMANH MOO HS AIHA c=) ~ ©. a) (00.0) :¢|.0\ a) ©) 0) 0 ere. oe aie ie) sera (ae eboe 3910 ‘SINQesoy <)(a (sles) 6-6) 8) 0) eke) mais) alo ah ele lors sce0buale oNT ‘1ayseqooy 8).¢) 59 8.0.6 a ep id. e)6 aye) aero 2 A) eis ene ae eae Le BA ‘puomyoly Che ONC ec Uwe ee CY OO SO a On ae) ‘Dud pew oj0D ‘olqang 0 048.8 8.6 @) ed Oe 6 816) D6) Se, Bee Ste kee eae een ee BIO ‘puro OOD moar ko omc rites oo i ory ace. od oW ‘purly.10g dis nttie ee ae eae WIN ‘woany 3104 See ee ee USBAN “JuaDSed) 4.107 IB) “WYsvyT sakoy JULIO LEE oS ee Se aA Oe OUBP] ‘011938004 69's (48 S08 a) e.8 6) 6a) 8) 6) eee 10 408 Sopunes Jane ed ‘qsanqgsiid © .b)'6 cee, 0) 8 8)0 46.0. 0.18 618.000) 6) 076/810 08 ls) eee hele ed ats ‘1191 Mie Stich OO ele aay xpuANT 90) 0).600 0e 648 8-90 10/8 ae 61 16) 82 el One 10, td ‘BIqdepeliyd Pe Panne i bth fixcuy oo tien on Bld ‘Blooesua gq CB SLOT UO ares BA ‘AM ‘Banqsrayqey Loe ae nee ee ae xay, ‘ousared Be. .8, 0 ie) 0 (@ 0.18 8 6. (6) 66 806.610 6.8) «(kee eee eure x "N ‘O3S9MSO AN SS Eee Et ee Eee 4qaN ‘eqenO Ge Rtn MnO e: no OOOO OC INO ‘eu0yeIyO «iste Sa ee ee IqaN 21814 WON Ble 8.0 2 @l4 e160. ae) 8 818) 6 01.610) a lepey es aneheenreae: 1A ‘pleuqqI0ON G16).8) 1076/0. 006,619.18 © (4) 0) 6 a6 ee en 6 Oe ele) 6p Olas BA “YIOJION Bib oicMOWd ChCECE CE O10 O DIO CHO 0-6 OOO XN ‘ya0X MON ON OD: Hi OME ONO OOS O00 RYT ‘SUBaTAQ MON ORO CMCC ORO OT FOO OL CAD Oo mor tc auo0y) “UdARTT MAN Bele /e.6).0.'o) olelalelaisie iskelleraaveltule cer scelevels uuaL ‘QTTTAYSBN ©0)/e 6: 0/(o)/e6) 6 (ale) w.e eles) m/ ellsiss I aay ‘ld JJOSUBSBIIBN ©) :ec0-s 10 fejie) eueveyes(ele (¢lfeneis ORI TO OS OOF Hn ‘ayonqueNn siieusiete eles sisebs/(s/anoenesutelietsrs ae) ‘sTed]vuey, ehegated ii alee, a elle) ¢..e)e.n lee eve eele ie sieie alsiess) ele uur ‘pRoqlooyy Bilele\\e' = » 1 --r-- a - a Fic. 31.—Temperature departures for the year 1909 ARCTOWSKI, CHANGES IN DISTRIBUTION OF TEMPERATURE 8Y of the equideparture curves is the same. A seesaw movement would ex- plain the transformation ; but, as will be seen later, such an explanation is not satisfactory. The map of 1908 (Fig. 30) indicates a simple displacement of the pleion towards the north. The shape of the pleion remained practically the same and in the west an antipleionian wave, following the movement, advanced over the continent. ‘The pleion has two crests, one going from Alberta towards Kentucky, with the departures of + 2°.6 in Calgary, + 2°.1 in Medicine Hat, + 1°.9 in Bismarck, Minneapolis and Cincin- nati, and a second wave following a perpendicular direction along the Atlantic coast in the New England States (+ 1°.6 in Boston and New Haven). Curiously enough, farther south, a depression in the pleion is noticeable, the departure being + 0°.5 at Richmond, Raleigh and Char- lotte, and only + 0°.3 in Lynchburg. On the Atlantic there is an anti- pleion, the quasinormal line going from Jupiter, Fla., towards Halifax. In Bermuda, the departure is — 0°.9. Finally, the map of 1909 (Fig. 31) shows the disappearance of the northwestern crest of the pleion under the influence of the advancing antipleion, whose two distinct centers moved from Eureka towards Battle- ford and from Independence toward Valentine. The pleion of 1908 has been reduced to a wave extending from Nova Scotia towards Texas. The most positive departures are + 1°.1 in Sydney, Grand Manan and Char- lottetown, and + 1°.7 in Fort Worth, Tex. To gain a more precise knowledge of the displacements of the pleions and antipleions, which took place during the years 1900-1909, I made consecutive maps and consecutive curves. I will examine separately the results gained by the study of the curves and of the maps. CoNSECUTIVE MApPs The annual departures of 175 stations were utilized to draw the maps I have just described. To obtain similar consecutive maps would have involved a great amount of purely clerical work. I simplified the task by omitting the Canadian data and taking only the means of the twenty- one districts into which the United States are divided in the columns of data published in the “Monthly Weather Review.” I copied the monthly means for the years 1900-1909 for these districts, then calculated the consecutive totals, then the individual means and finally the departures of these means from the normal values. This last operation was some- what arbitrary and I would certainly have done better by taking the de- partures from the ten-yearly means. 88 ANNALS NEW YORK ACADEMY OF SCIENCES At the beginning, I had no intention of doing the work for all the years and so I made the annual departures correspond to those given in the “Monthly Weather Review.” These departures are probably taken from the means of the entire series of observations, and these means increasing or decreasing as the number of years taken into consideration increases, the departures are necessarily not homogeneous. This lack of homo- geneity has no importance, since I adjusted the values, for each year, so they would correspond to the last annual departures (from the normals) given in the annual summaries of the “Monthly Weather Review.” The material which I have at hand consists of ten annual maps, giving the distribution of the annual departures from normal means, and of one hundred eight consecutive maps, showing the progressive changes of the map of each year into that of the following year. First of all, I must say that comparing the ten annual maps, obtained by utilizing district departures, with the ten detailed maps, described previously, one has to admit that the method of grouping the results of different stations to obtain regional averages is most inconsistent and defective. One can imagine how inefficient our daily weather maps would be if instead of utilizing the values given for individual stations we made regional averages. Still, even such smoothed weather maps would give some idea about the position of lows, and highs and cold waves could also be located, though in a very vague and unsatisfactory way. On the consecutive maps I have drawn, the pleions and antipleions are of course badly deformed and most interesting details are lacking, but the more or less progressive displacements taking place can easily be followed and the precise moments when the important changes in the distribution of temperature occur are detected without great difficulty. I will therefore compare the maps, simply in order to reach a better understanding of the transformations of the annual departure maps from one to another. 1900-1901. The consecutive maps show that the antipleion of 1901 came from the south and progressed westward over the States. After some minor oscillations of the quasinormal line,?* the upward movement of the antipleion starts with the map of September, 1900,—August, 1901. The annexed diagram shows the successive positions of this line on the consecutive maps for the periods ending in August, September, October, November and December, 1901 (Fig. 32). The western pleion remained practically unchanged. 26 These quasinormal lines deal with departures from normal values, and, evidently, differ from those of the maps expressing the distribution of departures from ten yearly means. ARCTOWSKI, CHANGES IN DISTRIBUTION OF TEMPERATURE — 89 1901-1902. The western antipleion of 1902 is not the southeastern antipleion of 1901. The consecutive maps show that the center of the antipleion remained nearly stationary over the Atlantic, east of Georgia or South Carolina, and that, at the end of the year, the negative area ex- tending over the States moved eastward. The following diagram (Fig. 53) shows the successive positions of the quasinormal line for the consecutive maps of November, 1901,—October, 1902, until January—December, 1902. The maps show an interesting feature concerning this antipleion. Its disappearance from the map was preceded by a progressive shrinkage followed by an expansion. The shrinkage began with April, 1901,-March, 1902, and continued until June, 1901,—May, 1902, when the quasinormal line lay from Washington over Louisville and Memphis towards Vicksburg; then, the negative area “py Fic. 32.—Successive positions of the Fic. 383.—Progressive displacements of quasinormal line the antipleion of 1901 increased again progressively until November, 1901,—October, 1902, and was followed by the eastward movement shown on the diagram. The consecutive maps do not give a satisfactory account of the forma- tion of the negative wave which, on the map of 1902, extends between the pleions of Canada and Mexico. The regional departure maps do not show the existence of the two pleionian centers of 1901 (Fig. 23) until towards the end of the year. The movement of separation begins with December, 1901,-November, 1902, and corresponds to the rapid drift of the eastern antipleion towards the southeast. The western antipleion of 1902 (Fig. 24) came from the southwest and spread out, progressively, over the entire area of the United States. 1902-1903. Evidently the quasinormal line of the consecutive maps is not the same as the quasinormal line of the departure maps from the means of the decade 1900-1909, but the displacements shown and the transformation of the consecutive maps must be similar to those of the detailed yearly maps. I repeat this statement to avoid misunderstanding. 90 ANNALS NEW YORK ACADEMY OF SCIENCES On the consecutive map of 1902, the western antipleion is very much less developed than on the departure map of 1902 (Fig. 24)- This is evidently due to the fact that the normals adopted in the “Monthly Weather Review” are very different from the means of 1900-1909. The following diagram (Fig. 34) gives the successive positions of the quasinormal line on the maps ending with October, 1902, December, 1902, February, 1903, May, 1903, and August, 1903. Figures 35-37 are the consecutive maps for the years ending with September, October and No- Fig. 34.—Displacements of the quasi- Fic. 36.— Departures of temperature normal line averages for November, 1902,-October, 1808 Ni Bri Sa : AG ‘a Fipenany = TORS S Gennes: ~ ae os Bb Fic. 85.— Departures of temperature Fic. 37.— Departures of temperature averages for October, 1902,-September, averages for December, 1902,-Novem- 1903 ber, 1908 vember, 1903. These diagrams show plainly that we do not have to deal with a simple displacement from west to east or southwest to north- east. We see that the pleion and antipleion are not only dependent upon each other for their displacements but also have a tendency to remain on the continent. The antipleion moving eastward displaces the pleion, first northward then northwestward and finally westward. The pleion and antipleion move en bloc contra-clockwise. ARCTOWSKI, CHANGES IN DISTRIBUTION OF TEMPERATURE 9] 1903-1904. The map of 1904, without a doubt, expresses the con- tinuation of this dynamic phenomenon. The change, however, in the respective positions of the two centers is not progressive. The following four maps (Figs. 38-41) show indeed that for February, 1903,—January, 1904, the temperature conditions were still very similar to those of 1903, whereas for April, 1903,—March, 1904, the distribution of the negative departures already had the character of the map of 1904. The develop- ment of the western pleion was delayed, and it is only on the consecu- tive map ending in November that its maximum development is reached. ~ WM VA At a4 Fic. 38.— Departures of temperature Fie. 40.— Departures of temperature averages for February, 1903,January, averages for April, 1903,-March, 1904 1904 St a ee ed woh ‘ aN Min al | vo . “J ees BY Fic. 39.— Departures of temperature Fic. 41.— Departures of temperature averages for March, 1903,February, averages for May, 1£03,-April, 1904 1904 1904-1905. The consecutive maps showing the transformation of the temperature conditions of 1904 into those of 1905 are most interesting, because they show a slow and continuous movement. The pleion and antipleion remain bound together and both remain on the continent; but the displacement is reversed and now it goes clockwise. The follow- ing two diagrams (Figs- 42-43) will be sufficient to demonstrate inter- mediate stages between 1904 and 1905. The first one is of the twelve- 92 ANNALS NEW YORK ACADEMY OF SCIENCES monthly means ending with February, 1905, and the second gives the distribution of the departures of July, 1904,—June, 1905. 1905-1906. The map of 1905 (Figs. 27 and 44) shows a large pleion in Canada, another on the West Indies and Florida, and an extensive antipleion with two centers, one in ‘Texas and the other on Nova Scotia. The consecutive-map of February, 1905,-January, 1906 (Fig. 45), shows the two pleions joined together, separating the two antipleionian centers. The following maps show the shifting and final disappearance of the northeastern antipleion, but the southern or southwestern antipleion re- mains, undergoing small changes of position or extent. Fic. 42.— Departures of temperature Fic. 44.— Departures of temperature averages for March, 1904,-February, averages for January—December, 1£05 1805 Fig. 48.— Departures of temperature Fig. 45. — Departures of temperature averages for July, 1904,-June, 1£05 averages for February, 1905,-January, 1806 The variation of the departures is of some interest. The highest de- parture in the north is + 1°.1, for 1905, in the North Dakota district + 2°.1, for February, 1905,January, 1906, + 2°.3 for March, 1905,— February, 1906, and then decreases to + 1°.0 and afterwards increases’ again progressively until November, 1905,—October, 1906, when it reaches + 2°.5. At the center of the antipleion, the negative departures for the first three consecutive maps are — 1°.4, —1°.1 and —0°.8. A general ARCTOWSKI, CHANGES IN DISTRIBUTION OF TEMPERATURE 93 increase of temperature takes place all over the States, in the pleionian as well as in the antipleionian area. Then, towards the end of the year, the negativity of the antipleion increases with the increase in positivity of the pleion. _ The contrast between the positive and negative departures decreased and then increased. This particular case shows therefore very well the importance of minor oscillations, taking place independently of the dis- placement of pleions and which, at the same time, may interest larger areas than those covered by pleions and antipleions. I reproduce only Fic. 46.— Departures of temperature Fic. 48.— Departures of temperature averages for January—December, 1906 averages for April, 1906,-March, 1907 +c AN ne Fic. 47.— Departures of temperature Fic. 49.—- Departures of temperature averages for February, 1906,January, averages for June, 1906,-May, 1907 1907 the first two consecutive maps showing the junction of the two pleions (Figs. 44, 45). The other maps simply illustrate a progressive disap- pearance of the southeastern center and the minor oscillations of the southern antipleion. 1906-1907. The maps of 1906 and 1907 (Figs. 28, 29), by their pre- cisely opposite character, seem to indicate a simple seesaw oscillation in the distribution of temperature. The consecutive maps contradict this supposition. The following dia- 94. ANNALS NEW YORK ACADEMY OF SCIENCES grams, representing the conditions for 1906, February, 1906,—January, 1907, April, 1906,-March, 1907, and June, 1906,—May, 1907 (Figs. 46- 49), show that the transformation began with a shght rotary movement, followed by a displacement of the pleionian center from the north to- wards the south, and then by a displacement of the pleion towards the west and of the antipleion towards the east. 1907-1908. The pleion of 1908 (Fig. 30) is so similar to that of 1907 that one would think that nothing extraordinary happened during the year and that there was simply a shifting of the pleion towards the northeast. In reality, the consecutive maps show that the pleion and antipleion moved first around, with the hands of a clock, so that for February, 1907,—January, 1908, the quasinormal line had already ex- tended from North Dakota towards Tennessee and Virginia, as is shown on the following diagram (Fig. 50), then the antipleion extended farther south (April, 1907,—March, 1908), and from then on it was driven away progressively in a northeasterly direction. The lines of the dia- gram show the successive posi- tions of the quasinormal line. The 1908 western antipleion is, in June, 1907,-May, 1908, al- ready on the plateaux and from then on its negativity increases progressively. es C j r ac Fig. 50.—Displacements of the quasi- 1908-1909. Finally, the last normal line on the consecutive de- twelve consecutive maps show parture maps of February, 1907,—Jan- ¢ ane ] = PC wary, 1908, till July, 1907,—June, 1908 that the axis of the pleion of 1908 first turned slightly to a north- south direction, then moved eastward, then back again to the west. For November, 1908,—October, 1909, there were two pleionian centers, one in North Dakota, the other in Texas. The advance of the antipleion towards the northeast is seen only on the last consecutive map. So far, the consecutive maps have served only to explain the transfor- mation of the departure map of one year into that of the following year. One might be satisfied with the results obtained, the consecutive maps having served their purpose in a satisfactory way. The principal conclusion gained is that the method used could be applied to seasonal forecasting ; but it is evident that, for such a purpose, it would be necessary to draw the consecutive maps as correctly as possi- ble, by calculating the means for all the individual stations. It is also evident that the same would have to be done for the rainfall ARCTOWSKI, CHANGES IN DISTRIBUTION OF TEMPERATURE = 95 data, the atmospheric pressure and the velocities and directions of the observed winds. In the present state of our knowledge of the phenomena governing the climatical variations, however, it would be most unscientific to try to make seasonal forecastings, since, at present, we see only the possibility of arriving at practical results by pursuing patiently the research work in a well-established direction. With the immense amount of work which yet remains to be done before a clear understanding of the climatical variations will be reached, any test of the method employed, in order to show its practical value, would be completely out of place. To show how each step forward leads to new questions to be solved and new research work, having apparently nothing in common with the pursued purpose, I will note a few problems arising from a closer ex- amination of the consecutive maps. First of all, during the years 1900-1909, the pleions as well as the antipleions displayed a tendency of persistence. No seesaw movement, between a pleion and an antipleion, leading to the gradual disappearance of both and then to the formation in situ of a pleion on the place for- merly occupied by the antipleion and vice versa, could be traced. Minor seesaw oscillations took place, but they served simply to increase or de- crease the contrast between the pleions and antipleions without destroying them. Together with a tendency to persist goes a tendency of displace- ment. These displacements are generally gradual and continuous, but sometimes they may be very rapid and in striking contrast to the nearly stationary conditions which preceded or followed the rapid change of position. The problem is, then, to know what makes a pleion remain on the map during several years and what makes the pleion move from one region to another. Another fact is the tendency of the pleions and antipleions to remain on the continent. In other words, the phenomenon of the variation in the distribution of the anomalies of yearly temperatures in North Amer- ica is to a great extent a purely North American phenomenon. This leads naturally to another question of some importance. The pleions and antipleions seem to be correlated or bound together. One depends on the other, and if one moves the other moves. The area of the North American continent seems not to be wide enough for the simul- taneous presence of many pleions and antipleions. In order to remain on the continent, the motion of a pleion involves a displacement of the antipleion in an opposite direction. A rotary movement is the conse- quence. It is a pendulation. The following diagram shows in a schematic way the pendulations of 96 ANNALS NEW YORK ACADEMY OF SCIENCES the pleionian center (Fig. 51) and expresses simply the tendency of the displacements, during the years 1900-1909, and may serve as an illustra- tion facilitating the comprehension of the problem. For precision it would have been necessary to have detailed consecutive maps. sae The principal problem is of ae PEE eer ALYY course, what keeps the pendula- eae ware ee ee ar Th \ at; ie tion coing ? Without some exte- rior impulse, the movement would die out or could not even origi- nate. It seems to me highly im- probable that a mechanical work- ing as is exhibited on the diagram could be due to variations of the Atlantic ice conditions. Without doubt, it is the cause of the for- mation of pleions which, repeating itself more or less periodically, gives the impulse to the clockwork. The Russian pleions have shown some correlations with the equatorial variation of temperature, illustrated by the consecutive curve of Arequipa. The consecutive curve of New York also belongs to the Arequipa type. We see now how the tendency of the pleions to maintain their existence complicates the problem of their mode of formation or origin. The following diagram (Fig. 52), which expresses graphically the last oof heb a 7 fot AY TA Wwe ce Fic. 51.—Pleionian pendulations Fic. 52.—Pleionian amplitudes and the Arequipa curve problem I have to mention 1m connection with the study of consecutive maps, shows plainly that, independently of the pendulations, the Are- ARCTOWSKI, CHANGES IN DISTRIBUTION OF TEMPERATURE 97 quipa variation affects the entire system of North American pleions and antipleions. I have taken the differences between the highest and lowest departures for each map, in other words, the total amplitude between pleions and antipleions. The curve is reversed and I have drawn the Arequipa curve underneath to make the comparison easier. From 1900 until 1906 these two curves are similar. Between 1906 and 1907 an interesting anomaly is noticed. The Arequipa temperatures decreased from 1900 to 1909. In the United States, the differences between pleions and antipleions were also decreasing during that period of years. CONSECUTIVE TEMPERATURE CURVES FOR SEVERAL STATIONS IN THE UNITED STATES In order to simplify the work, it was necessary to take into considera- tion the consecutive means of district data, and it is evident that the overlapping maps obtained from these figures are but a first approxima- tion. I will now complicate the problem anew by showing how.the pass- ing pleions and antipleions affected the succession of means observed at individual stations. TABLE V.—Ha«rtreme values of consecutive means Highest Lowest Sith on IMEVA RWIS Uy avetcrens rsie ccsicle Boe ersten Meiayeanbatey 77.6 74.3 Baa IL fs3 EDV ATYU TO rey carats beter steyie ter catactte aire fain cast oc arcbay cue a 73.0 69.6 2.4 19 ENON T) Sache cd BIO Cae ie OO OID Oe eres 68.1 64.7 a4! 1.9 RUA GIS rotor ote ete areas aeabara te aocue cohen evarenatems Gens 58.0 3.8 Drei Wann OM rae prenatal neko, ue ore roere eel nan eietans 56.6 Boia! 4.5 2.5 ING Wg VOD a poretete Oi med Seer Mcn veins AT ari awh 19) 4) 50.2 aye & 3.2 BOTTA sarees recto the uskorene eo) hoe cus oleae 46.7 AD ei: 4.6 2.6 ASH OT te seacket herrea eae cote seenaliete adaeshes 43.6 38.9 4.7 26 Sa mliteStehMamie see sno. c\ 5 Sec ctyace oie aw ogee pea | aX5}603 5.8 ieee 1S) UUM Gece eet rays cestcs cere te done DNs Ich ieee cA 41.5 349)5 (0) 6.5 3.6 Uy OME teptere ret meant coro chic ues oes te ado 44.9 37.6 33 4.1 WYUNNEES): COMTEASES ots caer fs eee es a ee a ae eee cee eens 48.6 43.5 En al 2.8 INOEUDBB Tater cats aie seater erAveccune: wakene 51.8 47.9 3.9 ey \ubaio) ou Lesa Pe eerie tay are Ace ORME ope ry eee Bee ee 57.9 54.4 33,5) 19 SHECVEVOL Ee warmth ere cons ore neiebeecho nn 67.6 63.8 Bye) Pe Al Newi Orleans tics care acts Elects are lnicus Tle@ 67.7 3.9 Qe LBROIREN IEE ahs chetevenr cuca tr yet EIR OL RRL ee eet a RBS 7 50.4 see bets: MEOSWAN SOLER rete uras ic) setacshncce ciuvhotevetaustels lans.‘e 64.6 61.0 3.6 2.0 USS Obra ey acetone at resto eraiere Chie ie 65.1 6272 2.9 iG Corpusy CHristimasctactie cee ee cele eke 72.4 69.0 34 1.9 TACO RU Oe eee perch e, OtaIn ete ee CEE eRe aie Se ae ee 58.8 56.3 PASH) 1.4 Not taking into account, at present, the actual monthly means, I will examine the consecutive annual means, or rather the curves expressing graphically the variations of these figures. : The stations for which T made the necessary calculations are Eastport, Portland, Me., New York, Washington, Raleigh, Savannah, Tampa, Key 98 ANNALS NEW YORK ACADEMY OF SCIENCES West, Sault St. Marie, Duluth, Havre, Miles City, North Platte, Wichita, \ TO +} Tow *l ac . ni “alee . x \ Shreveport, New Orleans, Eureka, Los Angeles, El Paso and Corpus Christi. In figures 53-56, I reproduce the diagrams obtained and in “Yormpa Sovannoly Raleigh Washing on | New York. (09 19/19. Fic. 53.—Oonsecutive temperature curves of Atlantic coast stations Table V, I give the values of the highest and lowest consecutive means and their differences in °F. and °C. ARCTOWSKI, CHANGES IN DISTRIBUTION OF TEMPERATURE 99 The curves of the stations along the Atlantic coast have already been published in my paper on the climatical changes recorded in New York City. In that paper I said: “The Arequipa curve descends; all the others, on the contrary, are ascending curves (Fig. 53). This contrast leads to the conclusion that the pleionian crests are independent of the long-range variations. Annual departures from ten-yearly means may therefore lead to very erroneous conclusions. At Are- quipa, for example, the annual means for 1901 and 1902 are higher than the average of the decade taken into consideration; they form positive departures, elthough belonging to a depression of the curve. “This is a strong argument against using such departures without consider- ing at the same time the trend of the curves. Now, the Arequipa curve has four crests and four depressions. So has the curve of New York. The most important difference between the two curves is that the maxima and minima of the curve of New York occur a few months later than those of Arequipa. One may say about three months later. “All the other curves are identical with the curve of New York in some particulars. For example, the depression of 1904 appears on all the curves. It occurs sooner in Tampa and Savannah than in Raleigh, Washington and New York. In Portland, and even more so in Eastport, this depression is very much retarded. The first crest in the Eastport curve, furthermore, re- appears, although greatly diminished, in the first depression of New York, which later corresponds to that of Arequipa. One can follow the gradual attenuation of this feature going south. For 1902 we have a positive de- parture in Eastport belonging to a pleion. This pleion (1901-1902) has evi- dently nothing in common with the equatorial variation of Arequipa and the other tropical stations. It is another wave having another origin and whose occurrence is marked all along the Atlantic coast in the midst of the anti- pleionian deficiency of temperature. About 1905 the curves of the southern stations differ greatly from the Arequipa curve. The curves of the northern stations, on the contrary, are similar to the curves of Arequipa and New York, except at the end. In Eastport we indeed notice a crest between 1909 and 1910 which is not a retarded crest, and going south, we observe the same attenuation of this phenomenon as between 1901 and 1902.” I will compare the other curves in a similar way. In the following diagram (Fig. 54), I reproduce the curves of Miles City, Duluth, Sault St. Marie and Eastport. These curves are very much alike. The only striking difference is that the variation of Miles City is more or less in advance of the others. The curves of Fig. 55 are those of the line of stations between Havre, Mont,, and New Orleans. The variation of New Orleans is to a certain extent opposite to that of Havre, and so it is most interesting to compare the diagrams one by one and see how the features of one curve gradually disappear in favor of those of other curves. The curve of Wichita, in particular, has a most unsettled appearance, since it participates in the 100 ANNALS NEW YORK ACADEMY OF SCIENCES variations of the northern plateaux as well as of those of the Gulf. A certain similarity with the curve of Aachen (Fig. 11) is undeniable. Some other localities of the middle states would give perhaps even a better example of an unsettled variation. Figure 56, on which the consecutive curves of Hureka, Los Angeles, El Paso and Corpus Christi are reproduced, shows that on the Pacific, as Miles Gly. Duluth. Sault S¥. Marie. Eastport. Fic. 54.—Temperature variation of Montana (Miles City) compared with that of Maine (Fastport) well as on the Atlantic, the pleionian variations are very pronounced, presenting a striking difference with northwestern Europe. Moreover, the curve of Eureka is similar to the Arequipa curve, and even shows exceedingly interesting small details of the Arequipa curve, details which cannot be ascribed to chance circumstances. It is also worthy of note that the Eureka variation is in advance of that of Arequipa. The pleionian crest of 1904-1905 is evidently out of the question. This pleion appears ARCTOWSKI, CHANGES IN DISTRIBUTION OF TEMPERATURE 101 even sooner in Los Angeles, where it corresponds in time to the principal antipleion of the curves of New York and Arequipa; but brachypleionian particularities, such as the V-shaped depression of March, 1909,—Febru- Miles City, Wi chika. Fic. 55.—Temperature variations in Montana, Nebraska, Kansas and Louisiana ary, 1910, appear later in Arequipa than in Eureka. The El Paso curve shows a variation opposite to that of Arequipa and retarded. The Corpus Christi curve is more complicated, opposition and similarity of variation being combined.” 102 ANNALS NEW YORK ACADEMY OF SCIENCES ~ We may say, therefore, that the consecutive curves of temperature for the United States, if compared with the Arequipa curve, may belong to a direct type similar to the Arequipa variation (considered as a standard of the equatorial or direct variation), or to an inverse type.** Some curves, it may be added, must be called indifferent, since, to a certain extent, they belong to both types of variation at the same time. Finally, there is the independent type. 1 Arequipa ; Eureka. Los Angeles. E\ Paso. Fic. 56.—Comparison of variations in California and Texas with the Arequipa curve Most curves may belong, temporarily, to one type or the other, but this is not a complication, because, if the results gained from the comparison of the consecutive maps are kept in mind, it is plain that it could not be 27 Compare the maps published by FrANK H. BiGELow: “Studies on the circulation of the atmospheres. . . .’’ Monthly Weather Review, Vol. 31, p. 515. Washington, 1903. Also: Sir NorMAN Lockypr and Wm. J. S. Lockyrr: “The behaviour of the short-period atmospheric pressure variation over the earth’s surface.’ (Proc. Roy. Soc. London, Vol. 73, p. 457. London, 1904.) AROTOWSKI, CHANGES IN DISTRIBUTION OF TEMPERATURE 103 otherwise. Since the pleions displace themselves, the crests of the curves cannot occur simultaneously everywhere, and since it has not been possi- ble to detect important and persistent seesaw centers, it seems a prtort very improbable that the direct and inverse types of variation could be characteristic for certain given locations. Therefore, there is no fixed location for the inverse or compensating type. An inspection of the consecutive curves shows, however, that in New Mexico, Arizona and Southern California the existence of a center, where the variation displays a striking preference to belong to the inverse type, may be suspected, and that, on the contrary, in Pennsylvania and Oregon, the direct type must be predominant. Are these locations centers of origin of pleionian variations? Not necessarily. Besides, the records of only ten years of observations are insufficient to give definite results. Even such as it is, the result gained leads to further investigations in regard to the question of the mode of formation of pleions in situ. Leaving this question as an unsolved problem, I will pass to another most puzzling subject. . Since, for certain parts of the United States, the consecutive tempera- ture curves belong to the direct type,—that is to say, are similar and coincide more or less in time with the equatorial curves,—the impulse producing these variations must be the same as that which produces the tropical variations. This impulse is evidently extra-terrestrial. There- fore, where the variation is direct, the departures of temperature will not be due to abnormal conditions of atmospheric circulation but will, on the contrary, produce such changes of atmospheric pressure, wind direction and velocity, etc., as may be characteristic for pleions or antipleions. On the maps the pleions do not disappear: they move away. Now the question is how—in a direct type of variation—the pleion corresponding to the second crest of the consecutive curve is renewed. Is it the same pleion coming back from the region it was pushed away from by the formation in situ of the direct antipleion, or is it a new pleion, and if so what became of the first one? Let us call the pleionian crests of the Arequipa curve A, B, C and D (Fig. 1). The consecutive maps show that the crest B of New York went northwest over Canada and then southwest towards California. The pleion came back nearly the same way during 1904-1906. The crest C of New York is therefore the same as B; but, if we try to follow this pendulation on the consecutive curves of individual stations, we do not succeed very well. This is because, as has been shown in Fig. 52, the amplitude of the departures changes independently of the pendulation. 104 ANNALS NEW YORK ACADEMY OF SCIENCES The pleions pendulate and surge at the same time. An old pleion may be reintensified. In the case of the pleions B and C the surging is noth- ing but the superposition of a new pleion upon an old one, so that C is the residual of B, plus a new impulse produced in situ under the influence of the direct solar action. In this way, it is conceivable why the pleionian variations may be more important on the North American continent than the identical variations in tropical regions. « ABpout TEMPERATURE VARIATIONS AND THE CHANGES OF THE ARCTIC Icke CONDITIONS In order to connect the European annual departure maps with those of the North American continent, I utilized the results of the observa- tions made at St. Johns, N. F., Upernivik, Jacobshavn and Ivigtut, on the west coast, and Angmagsalik, on the east coast of Greenland, those of the Icelandic stations Stykkisholm, Vestmanné, Grimsey and Berufjord and, finally, the observations made in Thorshavn of the Faroé Islands. For information about the variations which occurred on the Atlantic Ocean, I took the results of the observations made in San Juan (Porto Rico), Hamilton (Bermuda), Angra do Heroismo and Ponta Delgada (Azores), Funchal (Madeira), St. Vincente de Cabo Verde and, finally, those of St. Helena. The St. Helena observations were extracted from a report of J. S. Dines.** The portuguese data were kindly sent to me by the Director of the Observatory of Lisboa, in manuscript for the years 1906-1909 and in printed form for the previous years.*® The data for Bermuda were copied from the “Monthly Weather Review” and the reports of the Meteoro- logical Service of Canada. Those for Porto Rico were sent to me by Section Director Oliver L. Fassig. The results of the observations made at Danish stations were sent to me by the Director of the Meteorological Institute of Copenhagen.*® Finally, the data for Newfoundland were copied from the Canadian Reports. Unfortunately, for the period of years taken into consideration, only the records of St. Johns are complete. It would have been very desirable to have some records for Labrador and the Hudson Bay region, to obtain a closer connection between Green- land and Canada, but the distances between Ivigtut and St. Johns, Father Point and Quebec are so much smaller than the ordinary pleionian dimen- °8 JOHN SOMERS DINES: Climatological tables for St. Helena, with a report on (Meteor. Office Publ. No. 203: The trade winds of the Atlantic Ocean, London, 1910.) * Anaes do Observatorio do Infanta d. Luis. Obsarvacoes dos postos meteorologicos. %° Meteorologisk Aarbog . . . Anden del: Faeréerne, Island, Grénland og St. Croix. Udgivet af det danske meteorologiske Institut, ARCTOWSKI, CHANGES IN DISTRIBUTION OF TEMPERATURE 105 sions, that it is perfectly safe to say that the records of Greenland and Iceland are entirely sufficient for the purpose of the questions I had to solve. In Table VI, I give the annual departures from the ten-yearly means of temperature for the stations now taken into consideration, the dia- Upernwik. Jace \shavn. vighwr. Anqmagiaki. Stykkighelm. Grimsey. Berufierd. Vestmanwa pa wiee me Thorshavn- ° oa a Fic. 57.—Temperature variations in Greenland, Iceland and Faroe grams (Figs. 57-60) express graphically the succession of consecutive means, and in Table VII, I give the highest and lowest of these values as well as their differences. . ANNALS NEW YORK ACADEMY OF SCIENCES ey Srmmooocooooooo IAM ae ris aia as i) $H4t4444 141144 I tb e900 9 eg A wits BD HN HGS NAT OD St OD 4 2 > or) Leni @ S on) nl 106 Gmr Ces “1 0°6S RH SIO a steerer ies aa : F ; “Busf9H 4S aa ae 6° 3s Toy obser apl9A OF PP 2UTA “S rn orate BLT ERT oe eee wees en “BALOPVIY ‘Teyoung ET ' 8°91 A) sal0zy ‘epes[aq BIU0T 9°T a c’orL best Lie ae eel BeLOZy, “OUIS|O.1OH op BISUY eT ‘ g:¢ Ove wee errs ° 2018] ‘WARYSIONLL O° ee OQ ciettst sts e sees ‘pupjeoy, ‘ouUBUNISoA SZ OUD 'z rr ‘-pupjeo] ‘mpOUsTyyNS ee z'0— SQ sess ee ees ‘= pupjaoy ‘Aasutay Pee . I oF G . $ © (6. eu0he) siera Boee © ae aseye. ‘puvylso] ‘paofJnasg LZ 5 Fe — DOr te puv[uediy “gf ‘yITesseuUsUy o's . 8°6— Qa at oases pue[usedy “AN ‘ypayuaed Fe are Oi ot eC purldeary “A ‘UABYSqooR ls aes eee 5a), 6°0-— $3) 8°tTt+ Gel ete. e Sey @ 8.6.8) 6 lone ‘puvluseds “AA “qn sya] BG bP “A L°8é HE ats anion ok pe ae ; dN ‘suyor YS nae el eas Oe en Pree Bic ---upnuieg, ‘Wo} [Ue] 8°0 $1 “a G° Dh) Oi (aot, Someta cibir **OdLY OJIOgG ‘UBnE UBS a se ° ° sdUe1IDIG JSOMO'T segs SUDIUL BQVIZNIISUOI {oO §S9NPDA BWIA Y— TTA ATV 0) “0 9°0-— $'O-— b'O-— 2°O+ Pot FLT COC ODO OTN CeO CeCe Ft CWO IO) OMFG **BpB3[eq B100g 1) aT) L:0— e'o- e'0— 2:O+ eo OO; “MOO EOIO OO OOMO CH ODS NTO O80 0G -) gasuw 2°ot+ “oO 2°0O- ms) e'o+ e'ot+ s'o+ (DOCK ed “ee ee Cit TRCMICIA Wa Ye yey i DICK CP psexelis) oy eae) SC OVUGO NA 4s m8) 210) T'o- Tot F'Oot+ z°Oot+ e'ot+ O°ST oe Kevieiieli> ye) of ene hesey/ee)(s)/a/\9), 0) eieisse kat aieieiiisi seas aes 8 [Byouny e'o+ 3°0— 9°'0— e209 — e-o+ g'o+ ‘0 6 °F aileneliehie] ele ela nas SOD GO AOC OUT OCOD CACC tap TKS Tt Z°O+ a e'0— eot+ 6'0o+ e:o+ POT oO CBIOIONIO HOI. O.5'5 6 oe eee aes ee ieee! Ana STAT -°T— s°o+ G‘oO— T'ot+ Zot ec'ot+ o°tT+ C0 Ye Se ar NE 2 IOI ee nee ua’ eles ef uABYySsqoovs L°0— 9°o+ OI e'o+ G:o+ z°O+ weet. AU EEO ORO OOK OOD QUOC OOD.O OOD Boost MPATUIOd 9°0— c°o— 9°0— e0= G°o— O° T+ eye 12 en O DO O00. S500 0 TODO O00 11 YI[esseusu0y c°‘o— e°o+ e'ot+ $°0=— 8°0o— 9°0+ T'ot+ pie @ 0..6).6..2) 610 6, @ (a) 0.8. 6 6 6) 6) 8) a8 .0) 8588 65060 8 ere) ere paolynasg c*‘o— Tot F'O+ = io |e c‘ot+ F'Oct Shalt © )eelae| ae) @ 616) 6.8) 8-8 By 6) 6 (0466.6 (O08). 8 608 Ne Sere Laaees AVSUITIY) e°0— eng T20= L:0— G'o— eee e'o+ Ze mon sine § i sinh eeicEagntehn Shee een eae WMOUSIYANS g"0o— TO TelOste 2°O— TOs ‘0 T'o-— FC FID 0 BIRO SOO OID ONS DIGIC OOD OO OTNES-O No yen tapW sey \ c‘o— 20+ e'o+ Z°0— Z°0— 5°o-+- 2°0o+ e'9 sjehedenousilolekelsasielelelels)s)(shsleisleeenens (0) loner mT AGT STOUT; 9061 CO6T FOGL s06T GO6T TO6T 006T Do SPUDIST 0144UD]2P 242 pudD puUd)uUaaLy) ‘punjaoq 40 DUDp aunyn1adwmat— TA q1avL 2 ARCTOWSKI, CHANGES IN DISTRIBUTION OF TEMPERATURE 10% The following questions may be taken into consideration : 1). Is there any trace of the Arequipa variation on the consecutive curves of Arctic stations ? : 2). How far do the pleionian variations of the North American conti- nent extend over oceanic areas ? 3). Is there a system of Atlantic variations independent of those of Europe and North America ? 4). What are the connections between the continental variations and those observed in the Arctic regions ? 5). Do the Arctic ice conditions influence the variations of annual temperatures observed on continental areas ? Sy. Valves $ Fic. 58.—Temperature variations in Newfoundland and Maine The curves on Fig. 57 give an answer to the first of these questions. Two types are distinguishable: the west Greenland type, best represented by the curve of Jacobshavn, and the Icelandic, or let us say Grimsey type. Angmagsalik, on the east coast of Greenland, belongs to the Icelandic type and the curve of Ivigtut is transitional, since it is similar to the curve of Jacobshavn until 1905 and very much more like the curve of Angmagsalik after 1905 or 1906. If we compare the curves of Jacobshayn and Grimsey with the Are- quipa curve we must admit some similarities which are too well pro- nounced to be ascribed to a simple chance circumstance. With the ex- ception of the part comprising the consecutive means of February, 1903,— 108 ANNALS NEW YORK ACADEMY OF SCIENCES January, 1904, to October, 1904,-September, 1905, the Grimsey curve shows all the crests and depressions of the Arequipa curve slightly re- tarded. Jacobshavn, on the contrary, is in advance of Arequipa. Grimsey is a small island situated on the Polar Circle, north of Ice- land, and the latitude of Jacobshavn is 69° 13’ N. Both of these stations, therefore, are under the influence of polar currents and polar ice. If the consecutive temperature curves of these stations display simi- larities with the Arequipa curve, and, in consequence also with thoge of Bulawayo and Mauritius, we must admit that it is absolutely out of the question to search for an explanation of these variations in the changes of polar ice conditions. The second question could not be answered by a simple inspection of Funchel. Fic. 59.—Temperature variations in the Azores, Madeira and Cape Verde Isl. the departures given in Table VI and the curves of consecutive means of Figs. 59 and 60. Therefore, I have drawn maps showing the probable connections between the European and American pleions. The map for 1909, which I reproduce here (Fig. 61), may serve as an example. On this map there is a continuous chain of pleionian centers extending from northeastern Russia over Greenland and Labrador towards Texas and probably Mexico. The quasinormal line goes between Iceland and the Faroes, the departures being + 0°.2 in Berufjord and Vestmanno and —0°.4C. in Thorshavn. In the latitude of Newfoundland, the quasinormal line must go between the continents and run in a southern direction towards the Azores, where Angra do Heroismo belongs to the American pleions and Ponta Delgada to the western European antipleion. The other maps also show, in a convincing way, that there can be no question of a special Atlantic variation. Parts of the Atlantic area be- long to one pleion or the other or are covered by an antipleion separating pleions. ARCTOWSKI, CHANGES IN DISTRIBUTION OF TEMPERATURE 109 The fourth question, “What are the connections between the conti- nental variations and those observed in the Arctic regions?” can best be answered—with the maps on hand—by saying that the north polar annual temperature changes form, probably, an intrinsic part of those occurring in Europe, Siberia and North America. The ice-covered Arctic Ocean connects Nowaya Zembla and the Siberian shores with Arctic America into one immense continental area. On this area, pleions and antipleions are formed and conjugate into one system. The question, “Do the Arctic ice conditions influence the annual tem- peratures observed in Europe and North America?” ought therefore be reversed into the question, “How do the pleionian—or let us say the Fic. 60.—Temperature curves at St. Helena, Arequipa, Porto Rico and Bermuda Arequipa variations—influence the ice conditions?” ‘This is a very wide subject, about which much might be said. Air temperature is only one of the factors influencing the drift of polar ice. Ocean currents, and especially the winds, are more important factors than temperature. It is very well known that in the Arctic, as well as in the Antarctic, the ice conditions of certain regions may vary considerably one year from another and, from the knowledge gained in the North American archipelago—the Northwest Passage in particular—we must infer the existence of long-range or even secular variations. The quantity of icebergs drifting down into the path of the transat- lantic ocean steamers also varies considerably. The same may be said about the Antarctic. 110 ANNALS NEW YORK ACADEMY OF SCIENCES Speaking of ice conditions we must make a distinction between ice- bergs and sea ice, the conditions under which these two kinds of ice are produced and drift being absolutely different. Icebergs are anchored deep in the water and are much less influenced by the direction of the winds than by the ocean currents. They originate at the glaciers. The quantity of icebergs carried down through Davis Strait, for example, and along the Newfoundland Banks, will depend mostly on the factors which acted upon the flow of the glaciers. Sup- posing normal conditions of the glaciers (for example, a regular advance), a succession of cold years followed by a warm year and, in particular, an abnormally warm summer, will favor considerably the production of icebergs. In 1909, for example, much ice was noticed in the Atlantic.** The consecutive temperature curves of Upernivik and Jacobshayn give an explanation to this fact. During 1906 and 1907 we notice a remark- able depression in the curves (Fig. 57), followed by a steep ascent, culmi- nating, in 1908 and 1909, farther south, in Ivigtut. The drift of polar sea ice, on the other hand, is a most complicated phenomenon. In the Antarctic, the conditions are very much simpler than in the Arctic, and, even there, the drift is far from being a simple function of the velocity and direction of the wind. For the north polar basin, the distribution of the surrounding lands and islands, and the existence of well-pronounced ocean currents, compli- cate the ice-drift to such an extent that the possibilities of a successful study of the correlations between the anomalies of the meteorological con- ditions and the abnormal changes of the ice conditions is evidently most problematical. The observations collected by the Danish Meteorological Institute and printed every year concern the ice conditions during the navigable season only and are naturally restricted to the peripheral areas of the frozen sea. In some waters, we have to deal with winter ice, which must melt away during the summer; in other waters it is old drifting polar ice which hinders navigation. To correlate these variable ice conditions, of the navigable season of some arctic seas, with atmospheric temperature data of distant stations is a task which can lead only to very uncertain results. “In the special reprint of the Nautical-meteorological Annual of the Danish Meteoro- logical Institute, “‘Isforholdene i de arktiske Have, 1909,” it is said: “Off New Foundland and on the transatlantic steamer routes uncommonly much ice was observed, floes as well as icebergs. As early as February there was much ice, an@ from March to July the conditions were more unfavorable, than they have been for many years. The icebergs held out uncommonly late, the navigation being much ham- pered at Cape Race as late as August. It was not till September that the ice com- menced to decrease, but still many icebergs were to be met in Belle-Isle Strait and far off-shore off the Strait.” ARCTOWSKI, CHANGES IN DISTRIBUTION OF TEMPERATURE 111 Nevertheless, comparing the yearly temperature departures of the Ice- landic stations and Angmagsalik, on the east coast of Greenland, with the state of the ice in Denmark Strait and north of Iceland, one finds that a correlation is undeniable. Now, why is the variation of Grimsey (Fig. 57) so much more accen- tuated than that of Berufjord or Vestmanné? It seems evident that the proximity of the ice must be the cause. ~ Consequently, the Icelandic consecutive temperature curves could be taken as an example proving the influence of the ice on the variation of temperature. The ice conditions of Denmark Strait must greatly influ- ence the temperatures of Grimsey, especially some of the monthly means, by accentuating them one way or the other. It is not, however, the ice which causes the observed variations of temperature producing the for- mation of pleions and antipleions. The departure maps I have drawn show this plainly. The changes of ice conditions are more or less local phenomena re- stricted to small areas; the formation and development of the pleions and antipleions, on the contrary, are a universal phenomenon. CONCLUSIONS In the case of the annual departures of temperature for the years 1891 to 1900, which I utilized in my previous investigations, I dealt with the results of observations made all over the world and gained therefore some precise knowledge of the distribution and extent of the pleions and anti- pleions, and found that the years 1893 and 1900 were particularly inter- esting, the first being a year of predominant antipleions and the second being a typical pleionian year.*? In 1900, the pleions were not only very accentuated, with exceptionally high departures at their centers, but the areas they covered were fused together in such a way that the antipleions appeared as isolated patches on a pleionian background. The year 1900 was exceptionally warm, the temperature of our atmosphere being above the average, the negative areas being insufficient to compensate the excess of temperature of the positive areas. In some cases,—the map for 1909 (Fig. 61) shows it plainly,—conju- gated pleions form bands of very extensive dimensions. In other cases, there are intercrossing pleionian bands forming a real network with antipleionian patches between. = Op. cit., p. 123. 112 ANNALS NEW YORK ACADEMY OF SCIENCES Therefore, since, as has been demonstrated in this memoir, the Are- quipa variation is not exclusively an equatorial phenomenon, but appears, more or less modified, in North America, Europe and even in the arctic regions, the question is whether the years of conjoined pleions do not correspond to crests of the Arequipa curve and whether the depressions of this curve do not correspond to years: of isolated pleions in a net of conjoined antipleions. In fact, the maps for 1900-1909 show that the years 1904 and 1907, closely following the Arequipa depressions, are years of conjoint antipleions, and that the years 1900, 1908 and to a certain extent 1905, are pleionian years with isolated antipleions. Fic. 61.—Pleionian connections The existence of macropleionian variations, the close correlation of the pleionian phenomenon with the Arequipa variation, the compensating antipleions, and, finally, the dynamic character of these climatic changes, eliminate, it seems to me, the hypothesis attributing such changes exclu- sively to the presence of variable quantities of voleanic dust in the higher layers of our atmosphere. Variations of the solar radiation must be the real and most important cause producing the changes of our climates and keeping them in a dynamie state. The elaborate investigations pursued at the Smithsonian Astrophysical Observatory, and the Mount Wilson Observations in particular, give striking support to this conclusion. In fact, considering the means of the solar constants, observed at Mount Wilson during the summer months of 1905, 1906, 1908, 1909 and 1910, and comparing the differences of these mean values with the correspond- ARCTOWSKI, CHANGES IN DISTRIBUTION OF TEMPERATURE 113 ing differences of temperature in Arequipa, one arrives at the conclusion that a difference of 1° F. corresponds to a change of 0.01 of the solar constant. It would be premature to conclude that the brachypleionian, the pleionian and macropleionian variations are simply due to corresponding variations of the solar constant. Other factors may indeed complicate the phenomenon. Further research into the variations of the climates of Asia, South Africa and Australia, in particular, must be made before any definite conclusions can be reached. It is also evident that the sea- sonal changes of temperature, atmospheric pressure and rainfall have to be taken into consideration. I intend to do this. In this paper, I have shown that in far distant regions of the globe, simultaneously with the appearance of the Arequipa crests, pleions are formed ; that these pleions have a tendency to persist; that, in order to persist, one must displace another. Pleions and antipleions are corre- lated: if one moves, the other moves. In North America the displace- ments seem to be confined to the North American continent. In conse- quence, the pleions must pendulate from one side to the other. More- over, the differences between the pleionian crests and the antipleionian depressions of temperature change. These changes of amplitude seem to be in immediate correlation with the equatorial changes of temperature. The Arequipa curve may, therefore, be considered as a very convenient standard for the study of all the complicated phenomena of climatic variations, and of those observed in North America in particular. It appears now perfectly evident that a more detailed study of the Arequipa variations will advance very greatly the problem of correlations between solar and terrestrial phenomena. 7 o . ats ce emquey gern ay =r nile - ’ on ; AG oe er 4, i z* +; oe ety ; t ist ' P * x / - 4 ‘ ’ qt tie kro ee i tk - Ea aieg . i ‘ . / f 2 7 .% | les ; 7 i," “ } ° d = it ae ie : ; : : v is A ry rs a ioe F SiPeit. ’ ¥ . ? 49 + ‘ A % ; ' ee 7 + ! i he 7 q A « 7 * ' : = . se 7 7 =f ‘ ; F 5 Pian ae , ¢ i , ., | nee ae teint if ot, See FP Saat) a mi i uae oa _ s pith nity . : 5 é iy 2 Wainer sk lists ye hie i Rees) Lat at ‘s wife : s Taek re) Pitas #eeyiy OR gh, Te ay tier e hy) ae aed 7 ry ey ica Rou Yin nae Mei —o , PUBLICATIONS OF THE NEW YORK ACADEMY OF SCIENCES (Lycrum oF NaturaL History, 1817-1876) The publications of the Academy consist of two series, viz.: (1) The Annals (octavo series), established in 1823, contain the scientific contributions and reports of researches, together with the rec- ords of meetings and similar matter. A. volume of the Annals coincides in general with the calendar year and is sold at the uniform price of three dollars per volume. The articles composing the volume are printed separately, each in its own cover, and are distributed in bundles on an average of three per year. The price of the separate articles depends upon their length and the number of illus- trations, and may be learned upon application to the Librarian of the - Academy. The author receives his separates as soon as his paper has been printed, the date of issue appearing above the title of each paper. (2) The Memoirs (quarto series), established in 1895, are issued at irregular intervals. It is intended that each volume shall be devoted to monographs relating to some particular department of Science. Volume I is devoted to Astronomical Memoirs, Volume II to Zodlogical Memoirs, ete. The price is one dollar per part as issued. All publications are sent free to Fellows and Active Members. The Annals are sent to Honorary and Corresponding Members desiring them. Subscriptions and inquiries concerning current and back numbers of _ any of the publications of the Academy should be addressed to THE LIBRARIAN, New York Academy of Sciences, care of American Museum of Natural History. New York, N. Y. __-. PRESS OF JUDD & DETWEILER, INC., WASHINGTON, D. C. ue ; < ANNALS OF THE NEW YORK ACADEMY OF SCIENCES Vol. XXIV, pp. 115-170, pll. VII-XXI Editor, Epmunp OtT1s Hovey THE GENESIS OF CERTAIN PALEOZOIC INTERBEDDED IRON ORE DEPOSITS BY RAYMOND BarTLeTT EARLE NEW YORK PUBLISHED BY THE ACADEMY 4 Auaust, 1914 THE NEW YORK ACADEMY OF SCIENCES (Lycrum oF Naturat History, 1817-1876) OFrricers, 1914 President—Grorck FREDERICK Kunz, 601 West 110th Street Vice-Presidents—CHARLES P. BerKeyY, RAyMOND C. OspurN, CHARLES BASKERVILLE, CLARK WISSLER Corresponding Secretary—HeEnry EH. Crampton, American Museum Recording Secretary—EpMunpD OtTIs Hovey, American Museum Treasurer—Hrnry L. Donerty, 60 Wall Street TIhbrarian—Rawtreu W. Tower, American Museum Editor—Epmunp Otis HovEy, American Museum SECTION OF GEOLOGY AND MINERALOGY Chairman—Cuar._es P. Berkey, Columbia University Secretary—A. B. Pacini, 147 Varick Street: SECTION OF BIOLOGY Chairman—RayMonpd C. Ospurn, 557 West 124th Street Secretary—Witi1AM K. Grecory, American Museum SECTION OF ASTRONOMY, PHYSICS AND CHEMISTRY Chairman—CuHakLES BasKERVILLE, College of the City of New York Secretary—ErneEst HE. SMitH, 50 Hast 41st Street SECTION OF ANTHROPOLOGY AND PSYCHOLOGY Chairman—Cuarxk WIssLER, American Museum Secretary—Rosert H. Low1z, American Museum The sessions of the Academy are held on Monday evenings at 8:15 o’clock from October to May, inclusive, at the American Museum of Natural History, 77th Street and Central Park, West. [Annas N.Y. Acap. Scr., Vol. XXIV, pp. 115-170, pll. VII-XXI. 4 August, 1914} THE GENESIS OF CERTAIN PALEOZOIC INTERBEDDED . IRON ORE DEPOSITS? By RAYMOND BARTLETT EARLE (Presented in abstract before the Academy, 7 April, 1913) CONTENTS Page DATGRO GCE OMe earcesea ie eeree tiare ap otek e eyes SoS al ao a cla a erento oe obs Shera alte. ensues 116 ENCINO WiLEC TIM CTNES era ater anenerars ielici/sevessiee lis: a/.eie sei raelees axe teal eb ellelbpetal allows, @ a heleGuslien cues 8 HT Bart ee As Study ang Criticism Of previous) UHEOLIES)...+ +... 5 = - 118 ASR SUE ATI META Ty neue cpac a teatcy ore are nUete eve slate sie terials 6 sales, ete araunare elese, s seks 118 Iron ore deposits of the Clinton formation................ Paertrorchs ere. 120 GOES Oa ORE sc. ote cies sess) occ yovellere suscero AiSlavs, sie eidveevee"s wlene needa pfeianeleys eve tate 120 PEXGETU Ot CP OSUES ogo leyh eter score recovereve Saye S srevels Delenel eaehelsie oa Sialaleie) ciexelevais 121 OGWELILACOZOVCs CEPOSTUES wis ners oeroinne lore ois choles cheloies opens omen fanaiteh ie, chavaiete eles 123 Wabhana deposit. INewroundland|2. secs cere sete re sere FOO OLE. 23 Morbrook and Nictaux deposits, NOva Seo0tia= ca. c-.- 260+ 65 55 124 Miraavallevadeposit.. Cape Breton: Na Shas sce cis orlee cree slelereelee 124 SUIMMArVaOt et MECOMeESSO/SCNESIS: s16< (oars cle oe. o so areca oom cities nisl seve 124 EAGT eC SP OS ULI OM ate reccy eee eps hese oa ievevacevcten stele way Stele chet cageaie eusvelene 125 SHO GRAY WET ae od poole OU ROO a OC DIC ODO Eos De EID eISnne 129 Residualgenrichment theory soso ciecee ec cle ece +c aah CaS 129 Replacenlentatheonycrrce oye ott ooo mael oes atl o nike acca eae 136 Application recent investi GatlONSs.s. cme. oc ocie oe esses eas cas sisice 131 MheoLyAOtsoriein ale sedimMentaAciOnh a-.c sek cieile c eiele elec soll cileiea ee ciate 131 SeqimentanyeaASpeCh Of OLnenWe da -musvicie crore cnele ce tecehe sis) slele ote 132 Consequences of sedimentary theory.............0c.cesee-- 132 EUG val LOT CO esas « fareterayeaiale oiratebarae traces tee ree oe ta er o iekere eee 133 Inadequacy of sedimentation in open sea......... Heractstnisteom. 13 Sedimentationkinylacoonsiysone sewer eee nee. ieee Syn ees .. 134 Theories of replacement and secondary enrichment....,.......... 135 IBEESISLEN CE TOLIOLE SCAINISE. in tervals alee ee ker ER CL cee sie eisinciee 135 lWnnrenlacedehinrestoneny recieves eietrtorie cee ete screen clon enced eo are 135 Secondarya enrichment GheOLiessee entices case canes occbee eee. 136 SumplemeplacementsteOrys ance cee. sce selec su cee HOPROCC His Or 136 Patil Artesian replacementstheory. «ccs tecncecdtns cecceweeessbecee 137 ANTIETAM “OGIVONTTLOTOSS iciche b ao! PARRA CoRR © ORE CRORE eR a EO eae 137 COMEROUIM CM PACLOLS paretrary cy ete close wre © alaxeveieier she anc ot cave arasimierd okaicse 137 CirematonzOLsarctesiaMmyyaLelscsat<)e5e acs cis os cies co alei oe cro os 138 Dy PiCMESeChlOuS Of ClinbOMmmOre sis naa clack ool oe oe se Selec, ewcene een 139 JUL TSISTONL Ie! 6 Cen Bits Ces eke Bie ec hk CRN CHE ER OLE ee hn se) So ASTUTE OO one cic i ROOT TOR ae Ee See me a ee 139 Ses AOE Le PR ASEM a ek eel ARTS cyt) as ck © woe See. oe Hhdieichs Shc cl gedlsicis' Baveibs 148 ENGI CSS COS eters chet tatarcherne evdi-c era seveXs ie lensa ies atevale. Sicha absban s oihoue aeteke em ersi nate 144 1 Manuscript received by the Editor, 15 December, 1913. (115) 116 ANNALS NEW YORK ACADEMY OF SCIENCES Page WAT OUMEA 0. os fe ha ae 5 6 Zs. oro. Gyn nceieve Blue eset e ee Ruete eee tn eee Caen ie te eee 144 IWiGSE: Vir cL Bip 5 io ce Sloe ccustis ee iol aleve sete doase eher Rel el Te otel c Tete emene onen neta: Seema 145 PENNS YL VAD are oie eiousverokopstene (etateror n/sfebstcrorboyene ete talel ave retcnae Metohonetot eetete 145 IN QW OLS a aaive vas ors cycisint iatote ev entiyeweme, execs a et chante totene eo oRe le heres Nesenelte ene apemea 146 INOW aS COLLARS ic ie Sia\e beled! ereueyeicnstennlavene cleyeueie te ier s ParereneTe exe oeese ano oie 147 Artesian conditions inthe Clinton deposits ss. 1. ciieiere ce icles cierto ie 147 POLOUS TAY OLS oi cose ovee ieee aiote ale MiATCRY, cite ye ataucheceneratorsrston elena 148 TMpErviOus VAVELS/j.lelece cic\ siete le el oicla ele ote 2. ag, ao sk Oe Sk 148 Marine artesian (SLOPCS! 2s 5% ciel. ore, ofexereucle'ai chavars loelenetele ots lots Mokeliote onetonee ane 149 Infiltrationy Of meteor GswabeLs yt teteleereiclenersietetere oe eter meoteretatenenoretere 149 Ore conditions resulting from artesian slopes...............--.-ee-- 150 Depth" Of -GEPOSiESS 0 sé eeyers, em jeue.o-d eteiouc one alskoushe /oascene loaieke teeter etoteehsae tena 150 Hxtent OLGeposit GOwNeGip es cee acc cverenei, victotel ueiel anenetemeb sient cites taliotale ils! Widexdistributionsofe Gepositstrcsy-icrsrrere ce eisietciclel ois cteaeletcacioieneretonsiets 151 Occurrence of odlitic and fossiliferous ore............----++--- 151 Variations im ichemical compositionins. sc siete cee monn erase aay b Ve PIAGIONS) IN IEOREUTC sic.eis, acs, cue veysucieeeretere exe ehoue one Chevelete tenes ent hetereicntetons 152 Variations ANeGICHNESS ier cpesereperereieuel ocieneyeteeveietelera homers hetsues caeto or crepertete 152 Artesian replacement theory as applied to other horizons............ 152 DepositsofmwWwabana, NewLounadlan Glee se see ateroeie etel-tel eee = ene sie 153 Deposits of Mira Valley, ‘Cape Breton, Ne Scpae 2 eerste ie sie eo 153 GCeneralisummanysandeConcluslonSeerersisueiteletereieitoncie teeter ole usier ieliel eyeteietet ) PIG PLUME MS IDUMO STAD Wyevale tape rets)seaeie%el a chonener keke venckoious c: srorsey col euch ckemokeuetonsnettetenele y INTRODUCTION The great economic importance of interbedded iron ore deposits has been recognized for many years. ‘l'wo types of these ores may be cited as notable examples: first, and possibly most important, are the Clinton hematites extending along the Appalachian system of mountains for nearly two thousand miles and having their greatest development in the Birmingham district, Alabama; second, the Wabana hematite deposits, on and near Belle Isle, Newfoundland. The great extent, commercial importance and scientific interest of these Paleozoic interbedded iron ore deposits has long made them a sub- ject of painstaking exploration and study; the extensive development of them that has taken place during the last fifty years, and more especially within the last ten years, has made it possible to gather together a large amount of information of a detailed and accurate nature that under other conditions would have been impossible. Underground mining has largely taken the place of earlier open-cut methods. At Wabana, for example, the ore is now being mined under the sea at a distance of two miles from the original open-cut workings. EARLE, INTERBEDDED IRON ORE DEPOSITS 117 In Birmingham and Attalla, Alabama; in Clinton, New York, and in many other places, the same conditions prevail, and hard ore under- ground mining has largely superseded the soft ore open-cuts. Deep borings, in some cases nearly 2,000 feet in depth and several miles from the present workings, have brought to light conditions un- known in former times. Genetic theories, advanced originally from a study of outcrops and open-cut mines, undoubtedly would have been different, had the present data been available at the time. Yet with all the data accessible at the present time and much literature within reach, the question of the genesis of interbedded Paleozoic iron ores, and more particularly the Clinton ores, is still in controversy. Men of established reputation and wide experience adhere to theories sharply at variance with one another. At least two and possibly three such theories are still advanced and have the active support of men prominent in the mining and geological world. The practical importance of an accurate knowledge of the origin of these ores must not be overlooked. Upon such knowledge depends the interpretation of conditions underground below the present zone of mining, the depth to which the ore may be expected to go and the con- tinuance or change of richness of the ores. It has been the purpose of the writer to include in this paper a brief review of former theories and to discuss these theories in relation to re- cent field and laboratory developments, showing where they fail to meet certain conditions now known to exist and therefore must to some extent be altered or superseded. Further, the paper proposes and elaborates a theory of genesis which the writer has worked out in the field and laboratory and which, in order to separate from other theories and to avoid confusion, he calls the “Arte- sian Replacement Theory.” In brief, the writer attempts to maintain the following thesis: that the Paleozoic interbedded iron ores of eastern North America, and more par- ticularly the so-called Clinton ores, are replacements of porous strata of ordinary mechanical sediments by iron oxide, the agency being ferrugi- nous waters acting under artesian conditions. ACKNOWLEDGMENTS The writer wishes to acknowledge the criticisms, suggestions and as- sistance given by Prof. J. Edmund Woodman, Dr. Albert B. Pacini, Dr. Augustus Pacini, Prof. C. H. Smyth, Jr., S. W. McCallie, E. C. Eckel, E. F. Burchard, Chas. A. Borst, Dr. E. A. Smith, Dr. E. O. 118 ANNALS NEW YORK ACADEMY OF SCIENCES Hovey, Prof. A. W. Grabau, W. M. Bowron, W. C. Phalen, Prof. J. J. Stevenson, Prof. C. P. Berkey, J. Bewley and H. V. Maxwell. The writer made use of slides loaned by Prof. C. H. Smyth, Jr., in- cluding those used as a basis for the sedimentary theory ; also slides loaned by S. W. McCallie, State Geologist of Georgia, and a set of slides from E. F. Burchard, of the United States Geological Survey. The field work was provided for by a grant from the John Strong Newberry Fund of the New York Academy of Sciences. PART I. A STUDY AND CRITICISM OF PREVIOUS THEORIES THE SILURIAN ERA Whether the events that took place during the Silurian era were such as could give rise to immense deposits of iron ore precipitated as original material and thus forming contemporary sedimentary iron ore beds, or whether fossiliferous limestones and sandstones of a nature favorable for the later penetration of iron-bearing solutions were formed during a por- tion of this era, cannot probably be determined without detailed study of probable land and sea conditions at that time; but if any extended study of these deposits is to be made, a knowledge of land and sea areas and. a study of conditions that affected different kinds of deposits at the time the inclosing strata were laid down seem not only desirable but important. According to Dana, the North American continent at this time con- sisted of a broad land area extending from Alabama northward east of the present Appalachian system through Georgia, Tennessee, the Vir- ginias, Pennsylvania and into New York, there turning westward, skirt- ing the Great Lake region, then turning in a more or less southerly direc- tion through a portion of Wisconsin, then abruptly turning northward and running far into Canada. The southern and eastern portions of this continent must have extended far out into the Atlantic ocean. On the west, was a fast closing sea area shut off from the cold waters of the arctic regions by rather extensive areas of the Archean shield. This left an immense interior sea and a gulf or bay nearly 800 miles long, favor- able to the presence of warm currents which undoubtedly greatly influ- enced life in the waters and temperature conditions on the land. That these conditions were responsible to a large extent for the texture, compo- sition and distribution of the Silurian strata, we cannot doubt. Quoting from S. W. McCallie (216. p. 162) : “The region, however, of this tumultuous sea, with its rapid, moving cur- rents, appeared to have been rarely of long duration. Scarcely had the deposits EARLE, INTERBEDDED IRON ORE DEPOSITS 119 of sand and pebbles been fairly inaugurated, when the work was brought to a sudden close by a deepening of the waters, or a change of sediments. The enfeebled currents were now enabled to carry only the finest sands and clays, the materials of which formed the innumerable layers of shale. The later condition seems to have predominated throughout the entire Clinton epoch, but the frequent occurrence of sandstone marks intervals of time and less often repeated, the conditions favorable for the deposition of sands and clays gaye place to conditions favorable to the deposition of limestone. The last named conditions point to the clearing of the waters, and probably also the deepening of the sea. The thinness of the individual limestone beds would indicate that the conditions favorable for the deposition were of short duration. Yet, at the same time, the extended area over which they often occur shows that the conditions were widespread, and not confined to certain special localities. In the entire series of Clinton rocks, from the beginning to the end, we here see no evidence of sudden upheavals or tilting of strata. The character of the sediment points only to general oscillations of the sea floor, many times re- peated.” At the close of the Clinton, no marked changes took place till the close or near the close of the Niagaran period; then came a gentle uplift, as shown by the erosional unconformities. This emergence continued well into the Middle Silurian. Before the close of the Silurian period, two more periods of elevation have been recorded. The layers of sediment that formed a portion of the epicontinental shelf at the close of the Clinton era thus emerged from their original position and at the close of the Niagaran became a part of the coastal plain, the layers of the Clinton formation outcropping and thus placing their catch basins in a position favorable for the easy penetration of me- teoric waters charged with iron-bearing solutions. The following correlation table has been taken from Professor Grabau’s work (128, p. 87): Siluric: Upper or Monroan........... Hiatus and disconformity Akron dolomite Bertie waterlime Hiatus and disconformity Middle or Salinan..... ....-. Hiatus and disconformity Salina.... Camillus Syracuse Vernon Pittsford 120 ANNALS NEW YORK ACADEMY OF SCIENCES Lower or Niagaran......... Hiatus and disconformity Guelph Lockport Itochester Clinton (Disconformity refers to an erosional unconformity. ) J The important points with reference to iron ore deposits to be deduced from the above described conditions appear to be as follows: first, a thick series of sedimentary beds of rapidly changing texture and composition, clays and shales, sandstones and longer continued but less frequent lime- stone deposits; second, an irregular shore outline with consequent effects upon shape and size of sedimentary beds; third, a large area open to the constantly changing waters of moving and probably rapidly moving ocean currents from the south; fourth, to the east, land areas of the older crystalline rocks of both acid and basic types, from which the iron salts were dissolved in the course of weathering; fifth, a long period of slight oscillations of the sea floor. After the deposits had been formed, they were slowly elevated, and land areas were increased by a gradual emer- gence of the sedimentary beds. No violent or pronounced folding took place until the beginning of the Appalachian revolution, but mild dis- conformities followed the slight uplifts accompanied by slight tilting. Erosion of the strata on this coastal plain gave access on the part of the ferruginous waters to the porous beds, in which the water soon devel- oped artesion conditions. Iron OrE Deposits OF THE CLINTON FORMATION The Clinton iron ore was first discovered in or near the town of Clin- ton, New York. It was found later to be very common throughout the Upper Silurian, and the same type occurs also in other geological hori- zons. It has been traced extensively throughout much of eastern North America along the folds of the Appalachian system. The Clinton ores are usually found interbedded with slates, shales, sandstones and impure limestones. TYPES OF ORE The Clinton iron ores are known by various names according to text- ure, composition and location. The fossil ores (Plates VIII, IX, XVI) are composed of large masses of marine fossils, such as bryozoans. corals, crinoid stems, brachiopods and many other characteristic fossils of that age. These fossils are usually EARLE, INTERBEDDED IRON ORE DEPOSITS 121 found in a very fragmental condition and have been wholly or partly re- placed and coated and cemented with iron, or a mixture of iron oxide and calcite cement. The odlitic ores (Plates X, XX, X XI) are composed of odlites of hema- tite cemented ‘together like the fossil ores, but containing the peculiar odlitic structures which have often been referred to as resembling fish roe. The odlites are made up of a nucleus of quartz or other mineral, coated in concentric rings or spheres with iron, usually in the form of red oxide. The flaxseed or “seed” ore is really odlitic ore and has the same pecu- liar structure. In some cases, however, the term refers only to a smaller grained odlite, such as that in Ohio and Wisconsin. Dyestone ore is the name given to the Clinton red hematite of Ten- nessee and some other regions because of the fact that it has long been used as a paint or dye; and even in very early times it was mined in a small way by near-by inhabitants for such purposes. The red flux beds refer to a lean seam or seams of fossil ore in northern New York, the highest iron-bearing seam of the New York Clinton formation, and found typically at Clinton, New York. Soft and hard ore are names used particularly for southern deposits, to denote the difference between unaltered ore and that which, as a result of weathering and a loss of soluble materials like calcite, has left a resi- due rich in iron oxide and high in silica; the resulting soft ore being rather more porous than the hard, and found at or within a few hundred feet of the surface. The hard ore, on the other hand, is called hard, not because of its actual hardness, but because it runs high in lime and has not been affected by surface changes due to weathering. If we were to dissolve the lime from hard ore, the result would be a residue rich in iron and silica and precisely like the soft ores. Because of the value of lime in fluxing iron ores, the hard ores of lower grade are much sought and extensively mined. The soft ores, originally supposed to be the only valuable ores, have been for the most part worked out; and the old open- cut mines have given way to the underground hard ore mines. EXTENT OF DEPOSITS One of the most interesting problems to be solved in a study of the origin of these iron ores is how any single theory of origin can account for a line of deposits extending over such an enormous area. Following the shoreline of the old Silurian land areas bordering the epicontinental sea, we find deposits intermittently all along the line (Plate VII). In Wisconsin, the ore occurs in Dodge County, attaining its greatest #22 ANNALS NEW YORK ACADEMY OF SCIENCES importance at a place called Iron Ridge. This Wisconsin deposit is iso- lated and unlike the Clinton deposits found elsewhere, both in the shape of the body which appears to have resulted from filling of a local basin, and also in respect to the size of the odlites, which are smaller than the oodlites found elsewhere and are peculiarly flattened. Professor Grabau of Columbia University (128, p. 77) has suggested that they may be wind-blown deposits in the nature of sand dunes, which have been buried and later impregnated with iron ore. The ore next appears in New York, beginning near Rochester, where it can be seen in the Genesee Gorge in Monroe County, and then extend- ing eastwardly through Wayne, Oneida, Madison and Herkimer Counties. In this state, the ore is being steadily mined at Ontario and Clinton. The most extensive development is that of the Borst properties, in what appears to be a small bay of the old Clinton sea, covering possibly 2,000 acres near Clinton. Southward, we find the Clinton ores again outcrop- ping in Pennsylvania. Here the deposits extend in a broken belt across the central and southern parts of the state, including Montour, Snyder, Juniata, Blair, Bedford, Mifflen, Centre, Fulton and Huntingdon Coun- ties. These occurrences form a belt of several parallel ridges which run ‘southward into Maryland, where they appear in two beds in Allegany County. The ore then passes into West Virginia, where it appears in Mercer, Monroe, Greenbrier, Pendleton, Hardy and Grant Counties. "These West Virginia deposits have been but little used or explored. ‘The ores are then found in a few isolated sections along a line passing through the western part of Virginia and including Wythe, Giles, Bland, Tazewell, Russell, Scott, Lee and Wise Counties. These deposits have been worked tto a limited extent but are for the most part in a very unde- veloped condition. The ore belt then passes through Tennessee, extending across the entire width of the state, a distance of over 150 miles, into Georgia. The Ten- nessee ores occur in Claiborne, Campbell, Rhea, Roan, Sequatchie, Ham- blin and Marion Counties. Here, owing to the much-folded condition of the strata, many seams in parallel ridges appear running in a southerly direction into Georgia. In Georgia, the ore is found in Dade, Walker, Chattooga, Whitfield and Catoosa Counties. The ores of Dirtseller, Gaylors, Taylors Ridges and Lookout Mountain all are of considerable interest. ; From Georgia we find the ore extending into Alabama, where it reaches its greatest development in the Birmingham district. It is found in Shelby, Tuscaloosa, Bibbs, Jefferson, Dekalb, Etowah and St. Clair Counties. EARLE, INTERBEDDED IRON ORE D#HPOSITS 123 Two other separate occurrences of this Clinton ore have been noticed. One of them is a short spur extending from Kentucky into the southern portion of Ohio. In Kentucky, it appears in Bath, Montgomery and Fleming Counties. In Ohio, it is found in Clinton, Adams, Highland and Muskingum Counties. The other occurrence, which is of no present importance, is that of Missouri, where in Holt County a deep borehole revealed a seam of fossil ore in what appeared to be Clinton strata. From this outline, it is clear that these deposits are of great extent, and that this feature must be reckoned with in accounting for their origin. OrHER Pateozotc DEpositTs WABANA DEPOSIT, NEWFOUNDLAND The Wabana deposit, one of the largest deposits of interbedded iron ore in the world, is located on Belle Isle, in part, and extends for a con- siderable distance under the waters of Conception Bay, Newfoundland. The properties have been estimated to contain 3,635,000,000 tons of hematite iron ore (170, pp. 745-752). Conception Bay on the east is bounded by a shore of Huronian strata, which overly an anticline of Laurentian age, but are not conformable with it. The western shore of the bay is regarded as pre-Cambrian. Between these shores is a great synclinal trough of sandstones, slates and shales of Cambrian age. The upper strata forms Belle Isle and contains six beds of hematite, two of which are being worked at present. They show the following section (Symons; 328, p. 1009) : Feet Inches SLUG URES PES HA ce oad, CS RIC CROCE HELENE ORE TCE CEES mR eR roi 56 LMI AAO Meee eaters’ Sloss wire sheleh crate biata lettre eVeleurieuars ease, 2 1 SAMOStOM Garr mete G oheceore ole « Ciaeieta ee ee nitions waar e ekis 5 eye 15 Levene Ceseare Seto Polos CI IgCa cP Cece IGieRe chet Ais orcs aceon it 3 SAIL SUOM Cemeteries suciattteie: sickeyain eenekaral cia eieretmrecela wibees 2 Be 18 WebOeFeR oil ets Fy peek RCe MORE RCE Ric PEARARDL Er CRCAAY ick nc Rene re ence eae Pa 3 SAMASTOMES 22s ctace stools susie. otitl eins evan dicncyeieie..2 svetenete.it oe a: 1 8 ex OC ES EN eM ote co cxencar cd. Sect EL ROMER ICC CLT ER CPCI CEE SRE ee 3 SAM AISEOME mas Aeaysteeets see cleus olete ae Bioee chit Orel eibe ec austere 2 6 MOMMA GOR ee parcactevataetoi aici; « slenel Sis oro e cris eis ae ieee il B Sandstone and shale........ Menetrsitiste ses eherecefete 3 : Hematiter.. jis. He SOC Oe Oe edi sja weieiene. ae Sandstone ....... aera seit seateesisaciate siete eteie! sya erereeie 46 ae Hematite 2.0.5.0... eihereistee Siosts ake oh aycieucheie.o sit exe 9 A SANASCONELANGESH WLOx.1 s crekersiecveier trereieiesere laine oon LOO “fe Hema titerer cor teaie tete ce cae loool dec ddladecuecdlee 12 6 124 ANNALS NEW YORK ACADEMY OF SCIENCES The ore consists of a dense red hematite, high specific gravity, with bright metallic luster and a highly developed cleavage. The ore is of the non-bessemer type. These deposits are now being mined two miles from the outcrop, under the sea. The quality of the ore has slightly improved with depth. TORBROOK AND NICTAUX DEPOSITS, NOVA SCOTIA Several beds of hematite of a similar type to the regular Clinton ores have been located at Torbrook and Nictaux, Nova Scotia. The two principal beds are the so-called “Shell” vein and the Leckie vein. Sec- tions of this deposit may be found on page 147. MIRA VALLEY DEPOSIT, CAPE BRETON, N. S. The ore deposits of Mira Valley, Cape Breton, N. 8., are interbedded iron ores much like the Clinton deposits. Professor Woodman in his report (356, II-12—(3) ) says in reference to them: “The iron-bearing minerals are magnetite, red compact hematite, specular hematite, and possibly siderite in small amount. They occur in bands, inter- stratified with the sediments. In places they grade imperceptably into the latter across the bedding, and often downward or on the dip and longitudinally or on the strike. . . . These ores furnish perhaps the best examples known of partial and interrupted replacement.” Section on McKinnon farm: Backes Slate wm waillivcgs sre tee,s (sega eke sotete fore reroxthouekovels Mista peiae sreiele hisveais REG SH EMALICC See sista 17 apotste be rojcuaia ar aie ate ce, o onete lava fe eee) cieie uaoualla fale 3.0 Sates An. iii s acceso sctarecotereis oles © atc larsdetNote eigio sve halal abies 1.5 QUMATETALE Riad Secale orate cous tesainvel nt hetaians leona elke ehectke tale elehereister ate ib IMR MOTUS Es, arcane ieoreieateern Vere teabehe is far ciaimielsininne 'arbie wieterseanee 1.0 SST Gee eyes atctisy ze sestaoucsieh Pane ee au ares Cal erator fo mialrane alate leloiarevalsiotercts 0.5 SIDGOM EE eens cclevohie sil sete recclotele aiaichastansin’D cade once char crone ne ote OREN raat 14 er AI NRE Pe RENAE ESATO ARTS EB tone aD hrs oe ecg hac 1.0 (STEN EUCAILRE Aetna ORO DE DIG COG COD OORT IC oO GrO Cod Maer or 2.0 SIH ESAAS oo. BO aCe RG Rule Eon tC ODICLOIE OCr Sta Oo On Dodc 1.0 TOF eyes eho eran a ei oick clolekorcvct si wet ® cuscteaus, 6 eabtavetovaun eievetere ererenete 10335) STATON re eicteketenhs sie cha ature aPelay a sreravatcue ty tote oraiare a tionaeiees SUMMARY OF THEORIES OF GENESIS The origin of the Clinton ores has long been the subject of dispute, and even to-day, with the large amount of data available, we find con- siderable diversity of opinion. Mining men, owners of mines and those EARLE, INTERBEDDED IRON ORE DEPOSITS 125 interested in the financial success of the various properties naturally lean for the most part to the theory that will best support their conten- tion that the richness of the ore body will not diminish with depth. The views maintaining this contention seem to be well sustained in certain places by some of the facts developed in the course of actual mining operations. In other localities, however, data appear to contradict this view to some extent. A theory which would apply to all of the Clinton deposits must ac- count for the odlitic character of some ores and the fossiliferous charac- ter of others, for the occurrence of the ores in conformable strata, for the difference in lime and silica content in the various ore beds and within the same bed, for the presence of waterworn pebbles and grains of sand, for the soft and hard ores so common in the southern localities and for the compact and cleavable ores of some of the northern deposits. To find a theory that will reconcile all of these variables has thus far been impossible, and most writers have admitted that to a limited degree other theories besides their own may have some value. Three principal hypotheses have been advanced to account for these Clinton ores: (1) the theory of original deposition, which has been referred to as the sedimentary theory; (2) the residual enrichment theory; (3) the replacement theory. Of these theories, the first has the greatest number of advocates. It has the merit that depth would not affect the value of the deposits. PRIMARY DEPOSITION The advocates of the primary deposition theory believe that the Clin- ton ores were deposited contemporaneously with the inclosing rocks, i» the form of chemical precipitation at the bottom of the sea. Some claim that the iron was originally dissolved from the ancient crystalline and metamorphic rocks of Appalachia Land. 'The ferruginous waters were carried into inclosed or partly inclosed shallow seas or basins. The iron salts were slowly oxidized and precipitated gradually, forming con- centric layers of iron about particles of sediment on the sea bottom. As the sediments varied in kind and texture in different places and at dil- ferent times, the nuclei about which the iron concretions were formed differed. In some layers, the odlitic structure surrounded grains of well- rounded quartz sand; in other layers, broken fragments of fossils such as crinoid stems, bryozoans, brachiopods or corals were inclosed in iron concretions. It is claimed by some that where calcareous fossils were present, some replacement occurred, but only while the Sri: of original sedimentation was in progress. 126 ANNALS NEW YORK ACADEMY OF SCIENCES The advocates of this theory are numerous. In order to set forth the arguments given in favor of this theory, the writer quotes from several] of its leading advocates. Quoting E. C. Eckel (37, pp. 32-33) : “The principal facts supporting the theory of sedimentary origin may be briefly summarized as follows: bah “1. In mining from slopes running down on the dip of the ore bed, when once the limit of surface weathering is passed—and this may be at any point from 1 to 100 feet below the outcrop—no further important change in the ore is found with increasing depth; though a number of mine workings are now close to 2,000 feet from the outcrop. “2. A number of borings in Alabama have struck the ore at points from one- half to one mile back from the outcrop, and at depths of 400 to 800 feet below the surface. The ore encountered in these borings was hard ore of the usual quality, and not merely a ‘ferruginous limestone. Several borings in New York have struck Clinton ore at distances of from 10 to 15 miles back from the outcrop. These borings showed good ore at depths of 664 to 995 feet below the surface. “3. The physical character of the odlitic ore cannot readily be explained on any replacement theory, while the formation at the present day of original oolitic materials is a matter of common knowledge. “4. The occurrence of fragments of the ore in overlying beds of limestone in the Clinton formation as described by Smyth, points to the fact that the ore had been formed prior to the deposition of this limestone. “5. If the replacement theory were accepted, one would expect that the ore beds would show a greater vertical range; that is, that they would at places occur in rocks of other than Clinton age. Throughout their entire extent, the Clinton beds are closely associated with Silurian and Devonian limestones and shales, some of which offer excellent receptacles for the replacement deposits, but the characteristic red ores are confined to the Clinton itself.” The author goes on to say that primary replacements did not exist to any great extent, but, although no definite proof has been found, it is probable that some secondary replacement has since taken place. Leach- ing, of course, is noted. Eckel has done a very considerable amount of work on these ores while engaged in the economic work of the U. 5. Geological Survey, and, although his principal investigations were upon the deposits of the southern states, he included a wide extent of Clinton deposits in his special study, and his views must be recognized as founded upon accurate information and a broad knowledge of the literature available. According to Professor C. H. Smyth, Jr. (317), the replacement theory for the origin of the Clinton ore was not substantiated by the facts in the field. The caleareous rocks would certainly have caused the iron to be precipitated while it was passing through them, yet the ore is EARLE, INTERBEDDED IRON ORE DEPOSITS 107 found in places directly underlying limestones and shales. Concretions in the lean ore were found to be as ferruginous as those found in the richer ores; accordingly, the substitution took place before the fossil fragments were consolidated into a bed by the cementing material. If the ore is formed by a process of replacement, it should contain some ferrous carbonate, yet this has never been found. The iron did not come from above, for the Clinton ore beds are often horizontal, with no chance for the action of downward-circulating waters. There is no doubt that the ore was laid down in the form of an original precipitate at the same general time that the inclosing sediments were deposited. It is likely that there has been some enrichment of the deposits by the removal of calcium carbonate. The iron is secondary only in respect to the organic fragments, but primary with respect to the ore deposits as a whole. Weathering has contributed to the present condition of the formation as we find it to-day in some localities. Iron oxide and silica were deposited together from solution in meteoric waters. Organic material caused the retention of the iron in such waters. There is a connection between silicic acid, iron and organic acids in the soils, and a deposition of iron and silica together. Odltes were not originally calcareous. Smyth has probably been quoted more than any other writer in sup- port of the sedimentary theory of origin. It is certain that the careful microscopic work done by him is well worthy of careful consideration before coming to any final conclusion in regard to the origin of these deposits. Most of Smyth’s work was founded on a study of the Clinton ores of New York. Quoting D. H. Newland and C. A. Hartnagel (234, p. 50): “The evidence in support of both views has been traversed very thoroughly by C. H. Smyth, Jr., in a paper which represents as well the results of long experience and close study of the Clinton ores both in northern and southern districts. There can be no doubt after an impartial perusal of Professor Smyth’s paper that the theory of sedimentary origin is fully substantiated for most of the occurrences. For the ores under present consideration. this is the only explanation at all compatible with the conditions. “The stratigraphic features presented by the New York section of the Clin- ten do not lend themselves to the conception of vertical circulations of ground water such as would be required to dissolve and carry iron from the overlying strata. The ore beds everywhere lie nearly horizontal; their dip is universally toward the south, at an angle no greater probably in many places than that given by the contour of the original sea bottom on which they were deposited. At no time in their subsequent history have they been steeply inclined. More- over, they are overlain by thick shales not readily permeable to water. Under- ground flowage must necessarily be limited and be dependent for the most part on the cropping out of the more porous strata-like limestone and sand- stone layers. Thus it is directed rather along the bedding planes than across 128 ANNALS NEW YORK ACADEMY OF SCIENCES them. Below the ore, there is also more or less shale intervening before the top of the sandstone and conglomerate basement is reached.” G. P. Grimsley (130, p. 74): “The rocks of the Clinton series in this state are shales, clays, sandstones and an absence of limestone. If there was originally a bed of limestone now replaced by ore, the stratum was a very irregular one, varying in thickness from 6 inches to 5 feet. It expanded and contracted from place to place in a most irregular manner; a relation very unusual for limestone, but often present in sandstones and other shallow water rocks. By the theory of original sea deposition of this iron ore, it would be formed in the Clinton sea in the same manner as sandstones and shales. The iron was precipitated and mixed with sand and clay in which fossils were preserved. The odlitic structure would imply a concretionary deposit, the iron ore being precipitated around sand grains in concentric form. In some portions of the sea, as in the Keyser area, there was only a slight precipitation of iron in the sand. “The difficult factor to account for in this theory is the quantity of iron available for this deposit in the Clinton sea, apparently not duplicated at any other time before or since. There must have been at this time an exceptional quantity of iron present; its source is difficult to explain. There are thus encountered in both theories factors almost impossible to account for; but it seems to the writer that the theory of original deposition offers a more satis- factory explanation of the origin of these West Virginia Clinton ores than that of replacement.” One of the most complete publications on the Clinton ores has been made by S. W. McCalhe (216) for the Georgia Geological Survey. He agrees with Eckel and Smyth in placing the origin as original sedi- mentary deposition, but differs from them as to the source of the iron. He maintains that it came originally from large deposits of glauconite marl. Both J. S. Newberry (282) and T. C. Chamberlin (53) conclude that the Clinton ores of their states were formed by original deposition of their iron content, similar to the Swedish lake ores. H. D. Rogers (290, p. 729): “The regular ores of the Surgent (Clinton) series are to be regarded as among the permanent constituent strata of the formation, and as having origi- pated with other sedimentary materials in the form of very extended, but thin, sheets of ferruginous matter, covering at successive epochs the wide floor of the quiet Appalachian sea.” He goes on to say that the source of the oxides has not been deter- mined. He acknowledges that much secondary enrichment has taken place by enormous quantities of ferruginous matter diffused in marls, slates and shales in contact with the ore bodies, being dissolved in the form of sulphate and then redepositing the iron in the ore beds, reaction EARLE, INTERBEDDED IRON ORE DEPOSITS 129 with the lime of fossils converting it to peroxide. This secondary en- richment is plausible because where the outcrop, the slope of the ground, the thickness of the overlying strata and other conditions are favorable to considerable infiltration of surface water, the ore carries a higher amount of iron than at less favorable places. The fossils often form one-half of the total weight of the ore; it is obvious that if part of the lime thus contained is dissolved out, the remaining peroxide of iron will form a much larger percentage of the total bulk of remaining material. SECONDARY ORIGIN Residual Enrichment Theory The residual enrichment theory starts with the supposition that the ore beds were originally limestones rich in iron, that by a process of leaching the hme carbonate was partly or wholly removed, and that the iron, together with the insoluble material, was left in a much more con- ccentrated form. Similar effects are known in tropical countries; in Cuba, for instance, where silica by a process of weathering known as laterization has been removed from iron-bearing rocks, leaving the iron and insoluble portions rich enough to be classed as an ore and mined profitably. I. C. Russell (292, pp. 22-23) : “Portions of the Silurian rocks of Alabama, readily recognized as limestones when unweathered, are easily mistaken for sandstones and shales when only their weathered outcrops can be seen. The Clinton ore, or fossil ore, inter- bedded with strata of shale and sandstone forms one of the most character- istic beds in the Upper Silurian rocks of Tennessee and Alabama. In the mines of Gadsden and Attalla, Ala., where Clinton ore is worked, the strata are highly inclined (a dip of 70 to 80 degrees to the southeast prevailing) and well exposed for study. “The outcrops of the beds are soft, porous, highly fossiliferous ore, which has a deep brownish red color, and is easily worked and easily smelted. The ore at Attalla retains this character to the depth of about 250 feet, measured down the slope, and then changes to a hard, compact, ferruginous limestone, rich in fossils. The marked difference in the character of the ore in the upper portions of the mines as compared with that of the lower portions is due entirely to weathering. This is shown by its chemical composition. Two typical samples of the ore, selected by me—one from near the surface, repre- senting the ordinary character of the soft ore, and the other from a depth of ‘250 feet, representing the hard ore, but not the most calcareous variety—gave on analysis the following percentages of iron, lime and carbonic acid, after ‘drying at 105 degrees Centigrade: At surface 250 feet ESS CSS S DDC ORES UREOS Ces SCT ee 57.52% 7.75% (GEIO) excel aig Grd SPOS BIG OIC EROS PRICE RES 1.38% 47.64% rare e tiers Marana te, shevavorn el axe \jo,0 6)9/ eel oc) 0: .80% * 34.90%” 130 ANNALS NEW YORK ACADEMY OF SCIENCES Eekel (37, p. 33) claims that Russell was mistaken and that the At- talla ores do not vary with depth. Replacement Theory The replacement theory holds that the iron content of the Clinton rocks in the form found at present has resulted from a replacement of lime carbonate by iron, long after the rocks had been deposited. The iron was introduced by descending waters charged with iron which they had dissolved out of overlying ferruginous rocks. J.J. Rutledge (293, pp. 254-255) : “The conclusion that the iron content of the Clinton iron ore beds of Stone Valley, Penn., is due mostly to replacement by removal and enrichment, seems unavoidable, when it is considered that but a portion of the fossiliferous lime- stone or of the hard ore is found to contain iron oxide when examined in thin sections under a microscope. Calcite cement makes up by far the greater portion of the section. “An analysis of the limestone shows that it contains but 2.12 per cent of FeO and 2.35 per cent of Fe,O,. These seem much too small an iron content to yield as rich an ore as the soft ore, simply by the removal of the calcium carbonate. Field conditions such as the occurrence of weathered shales, dull colored clays and iron-stained sandstones, prove that the action of replacement is still going on. “The iron came originally from the overlying shales and was transferred later to the beds of fossiliferous limestone.” The following reasons are advanced by Rutledge for the statement that the addition of iron was not due merely to the removal of the lime- stone, as would be the case under the enrichment theory : “(a) The character of iron ore concretions where associated with silica. “(b) The invariable association of the soft ores (rich) with the leached, decolorized shales and the hard (lean) ores with unweathered, bright red shales. “(c) The relations of the ores to the shattered sandstones and to the topo- graphic situation of the ores. “(d) The fact that analogous replacements are now going on in the Medina formation. “(e) The observed progressive steps in the transformation of the limestone to an ore, which may be followed in the field, in the sections under the micro- scope and in chemical analyses. “(f) The absence of conditions such as local crumpling, including a shrink- ing of the strata, pointing to a relative rather than an absolute enrichment of the ores.” I. C. White (198, pp. 135-137) : “The iron has evidently been filtered into the bed as the lime has been: EARLE, INTERBEDDED IRON ORE DEPOSITS 131 filtered out of the bed, otherwise the percentage of iron in the bed would not diminish below drainage level. “Wherever the ore is valuable the inclosing rocks are very much weathered ; the lime rocks are changed into clay and the shales overlying the ore are bleached almost white; their iron having presumably been transferred to the ore bed. And this is a reasonable way of explaining the fact that the ore bed does not always keep the same place in the series, for any bed can become an ore bed, provided it is so situated as to be a water-bearer and recipient of the iron-leachings.” A. F. Foerste (110, pp. 28-29) : “As a rule the iron has replaced the substance of the bryozoan itself; all the stages between partial and complete replacement may be noticed, the most complete stages being of course found in the purer ores. Usually, correspond- ing changes are observed in the cement which binds the o@litic grains together into a solid mass. It is evident in these cases that the origin of the odlitic structure is not due to a concretionary segregation of iron particles, but finds its explanation in the gradual replacement of the lime of the fragmental fossil bryozoans, particle by particle, by the iron ore.” N. S. Shaler (303, p. 163) : “The ores were not included in the present iron-ore beds at the time of their deposition, as conditions varied so much at different points that this would have been impossible. The ore-occurrences are due to replacement of limestone beds by iron-bearing solutions derived from overlying shales. The iron could not have been deposited as far from the shore as the limestones were.” J. P. Kimball (190, p. 355) : “Parts of thin fossiliferous limestones of the Clinton group of strata are often replaced by red and brown ferric oxides from extraneous sources. “This replacement has been wrought especially in steep dips by infiltrations from drainage of adjacent ferruginous strata, partially of an inferior series outcropping topographically higher in the flanks of these parallel ridges.’ APPLICATION OF RECENT INVESTIGATIONS THEORY OF ORIGINAL SEDIMENTATION Even though it may be demonstrated that the odlitic hematite can be successfully synthesized in a chemical laboratory in open agitated water ; even though odlitic formations such as the sand of the Great Salt Lake in Utah and the brown iron odlites of the Swedish lakes are being formed in open water to-day; yet no matter how plausible the theory may be in most respects, if a single factor prevails that would be impossible under conditions necessary for original deposition of the iron ore beds, it is 132 ANNALS NEW YORK ACADEMY OF SCIENCES enough to discredit the correctness of the sedimentary theory and to force us to look elsewhere for an explanation of the origin of these forma- tions. Sedimentary Aspect of the Ore Beds That the general appearance of the ore beds would give the impression that they must be regular sedimentary iron-ore beds laid down as the advocates of sedimentation suggest, cannot be doubted. Clean contacts, lens-shaped deposits, widespread occurrence, separated individual iron- coated odlites, non-ferruginous sandstone and limestone beds overlying some of the ore beds and underlying others, all would tend to give weight to the sedimentary hypothesis. Before the theory can be considered as proved, however, it must account for certain conditions that appear to the writer irreconcilable with any theory based upon original deposition. Consequences of Sedimentary Theory In the first place, let us apply the sedimentary theory to the odlitic hematites that are so common in New York, Virginia, West Virginia, Kentucky, Ohio and Wisconsin. Under the sedimentary theory, it is assumed that the Clinton Sea was heavily charged with iron salts in solu- tion, and that, as sediments were being laid down along the shallow and gently inclined shore slopes, myriads of sand grains under the influence of considerably agitated waters were coated with layer after layer of iron oxide, which in many cases alternated with silica. These iron-coated grains finally accumulated into beds in the same manner as any sand stratum would accumulate and were then cemented by more iron and calcite into solid beds or layers of sedimentary rock like any other sedi- mentary deposit. For the sake of argument, let us assume that such conditions did exist. If the sea-water contained enough iron in solution successfully to coat grains of sand until they formed a bed several feet in thickness, would it not be reasonable to suppose that all sediments laid down simultaneously would be coated, impregnated or at least stained with iron? Would not all lenses of clay and shale and limestone be completely saturated with the same iron-bearing sea-water that coated the mass of odlites? Would it be possible for any portion of the shore deposits along the entire length of the Clinton Sea to have escaped without leaving permanent evidence of the presence of such large quantities of iron in solution in the sea- water—iron sufficient to cause deposits within a comparatively short period extending for nearly two thousand miles along the shore, many miles out to sea and in some cases many feet in thickness? EARLE, INTERBEDDED IRON ORE DEPOSITS 13 Co Field Evidence In a small seam of odlitic hematite about eight inches in thickness located at Big Stone Gap, Virginia, the writer discovered what at first appeared to be a small bowlder (Plate XI, fig. 1) entirely surrounded by iron odlites. Upon removing the stone and breaking it, he found that instead of being an ordinary bowlder, it was an original formation, such as is often present in beds of loose sand that have been penetrated grad- ually by mineral-bearing solutions and consolidated by the well known process of cavernous consolidation, leaving loose sand-filled cavities (Plate XI, fig. 1). The “bowlder” was well filled with practically pure loose quartz sand, with no iron-coated or hardly even iron-stained grains (Plate XI, fig. 2) and was merely the first of several also similarly sand filled. Here we have a local accumulation of sand made up of quartz grains entirely surrounded by iron-coated odlites, yet completely ignored by the iron-charged solutions and not even consolidated. A little farther on in the same seam,:another small mass of yellowish white sand was found, which was partly consolidated but had not been penetrated by the solu- ’ tions that coated the surrounding odlites (Plates XII, XIIT). If we still believe that the iron-bearing solutions were a part of the sea-water, how can we explain two sets of different consolidations, one with iron the other without iron, from the same source and at the same time as the surrounding conditions seem to indicate in this case? Let it be emphasized that these occurrences are situated not at margins of ore seams, but well within a distinct stratum of hematite. In Clinton, New York, in a single hand specimen, the writer found odlites coated with iron oxide and other odlites coated with a green min- eral in concentric layers (probably an iron silicate, greenolite) and still a third type of odlite composed of a quartz nucleus, then a ring of iron oxide, then a thicker ring of the green mineral and finally another ring of iron oxide. In a seam in Birmingham, Alabama, a single slide of the Clinton ore shows a calcite-coated odlite and an iron-coated odlite side by side. How would it be possible for the iron-charged sea-water to distin- guish between different grains of the same mineral and coat one with one substance and its neighbor with an entirely different mineral ? At Niagara gorge, we find the Clinton series of limestones, shales and sandstones, but no evidence of iron-ore seams. This indicates that the iron-bearing marine solutions had suspended operations at this point, and yet had been active as far west as Ohio, Wisconsin and Holt County, Missouri. Finally, in applying the sedimentary theory to the fossil-ore beds, we 134 ANNALS NEW YORK ACADEMY OF SCIENCES find conditions similar to those prevailing in the odlitic strata; limestones untouched by iron-bearing solutions, but with overlying and underlying iron-ore seams; lenses from a few inches to many feet in length composed of limestone with only a slight marginal penetration of the iron, and yet entirely surrounded by iron ore (Plate XIV, fig. 1). In places, fractures in the limestone beds have been penetratéd, by iron and the walls of the breaks lined with ore. In other occurrences, the iron has followed seams or laminations in the limestones and has replaced the limestone along these planes of weakness (Plate XV). A generally prevailing condition of apparent replacement in all stages is found in both fossiliferous and odlitic beds, in which both calcite and quartz show corrosion and replacement (see photomicrographs, Plates ROWE KORY KOR) Inadequacy of Sedimentation in Open Sea The conclusion seems justified that, whereas widespread similarity of conditions should be expected, with unbroken evidence of the presence of marine iron-bearing solutions in the sedimentary beds (whether of sand- stone, limestone, shale or clay), yet the reverse conditions actually exist, for (1) small lenses of loose sand untouched by iron-bearing solutions are found, which, under the conditions imposed by any sedimentary theory, could not have been free from the iron; (2) two unlike consoli- dations, one without iron, the other surrounding the first and completely charged with iron—a circumstance that would have been impossible under the sedimentary conditions pointed out by the various advocates of marine deposition of iron ores; (3) differently coated odlites in the same immediate locality would hardly seem possible under such a theory ; (4) the penetration of iron into seams, lamination planes, weakened strata, mudcracks and fissures running off from the main ore beds could hardly be explained under theories depending upon a primary origin of the ores. | Sedimentation in Lagoons In regard to a growth within inclosed lagoons or basins of shallow water, the field evidence in some places may bear this out. The writer’s observations, however, have seemed to show much active wave erosion and considerable rather violent agitation of the sea-water. This is illustrated by what appear to be two well-formed stacks which could only have been shaped by wave erosion; one at Clinton, New York, and the other at Red Mountain, Birmingham, Alabama. Further evidence might be suggested, as, for example, the existence of several kinds of varieties of coral which EARLE, INTERBEDDED IRON ORE DEPOSITS 135 do not thrive in the still waters of inclosed basins but require agitated waters in the open sea. In places, also, the sediments contain a large amount of water-worn material, fossils badly broken and coarse-textured conglomerates. So, although we do see in places such testimony as Pro- fessor Smyth (317) has suggested, yet we find also much evidence of quite different conditions; and, therefore, it seems to the writer that httle importance can be attached to the supposed basins as an aid to the determination of the origin of these ores. * THEORIES OF REPLACEMENT AND SECONDARY ENRICHMENT Persistence of Ore Seams The question of depth and distance from the outcrop to which the ore is known to extend is an interesting one, and the facts are inconsistent with the theories of secondary enrichmefit and replacement, where such theories depend upon leaching of slates and shales or vertical descent of ground waters. One boring has shown good ores, 1,902 feet deep and two and one-half miles from the outcrop; another over 800 feet deep and more than ten miles from outcrop and with a very low dip. The writer is inclined to agree with Professor Smyth (loc. cit.) in part in regard to such data; but as to using this great depth as an argument in favor of original deposition, he cannot convince himself that it applies. Some of the deposits in that event must have extended into waters of very con- siderable depth and distance from shore; and as depth increased, the amount of iron necessary for keeping up the same degree of richness as nearer to the shore must have been great indeed. It would seem highly improbable that these iron-bearing marine waters could circulate over and through the sediments without becoming diluted in the great expanse of water, as currents carried them far out to sea. The present writer would, on the other hand, lay claim to the argument of great depth in support of his ideas of origin, which differ widely from those of the advo- cates of original deposition. Unreplaced Limestone Finally, as to the argument that some overlying beds of limestone would be excellent for replacement of lime by iron and yet remain prac- tically untouched with clean-cut contacts although in close proximity to iron-bearing seams. Here again the writer agrees with Professor Smyth that the facts are against the replacement theories as ordinarily ad- vanced, especially since many layers of impervious rock lying in a more or less horizontal position intervene between the ore seams and the sur- 136 ANNALS NEW YORK ACADEMY OF SCIENCES face. Yet the writers own conception of the origin of these ores is much strengthened by these same conditions, which are indeed neces- sary, according to his ideas, for the existence of the iron-bearing seams. Because of these conclusions, it seems unnecessary to the writer to spend further time on a theory that in the hght of such conditions as have been pointed out seems not only inadequate but impossible. Secondary Enrichment Theories Under the head of secondary enrichment may be classed both enrich- ment due to replacements and that due to residual enrichment. Of the theories of secondary origin, some depend upon leaching of soluble mat- ter and a consequent enrichment of the iron-bearing deposits because of relative insolubility of the iron, and others upon a combination of re- placement with enrichment by a process of leaching of iron from over- lying ferruginous shales and slates and replacement of the lime in the underlying limestone by iron thus obtained. The most that can be said regarding these possibilities is that undoubtedly these alterations have been made, but to a very limited extent, and such methods are wholly inadequate to explain the distribution of the ore, as shown by recent borings and extensive underground mining, which have proved that hard ore does not change materially with depth, and that the above theories only account for the very superficial facies called soft ores. Absence of extensive exploration gave these theories plausibility and caused much favorable comment upon them for a time, but more recent underground mining has caused them to be more or less generally discredited. Simple Replacement Theory We still have one well recognized theory to discuss before advancing the theory of the writer, and that is simple replacement. ‘The advocates of this theory have seen extensive evidence of replacement of the calcite by iron in the fossiliferous beds and have noted the replacement of the lime cement by iron. In respect to the evidence advanced by these au- thors, the writer is inclined to believe that to a large extent it is correct, but in a few particulars he finds himself obliged to disagree with their deductions. The first is the attempt to account for the iron as a leached product from overlying shales. It seems incredible that such immense quanti- ties of iron can have been derived from so limited a source. It also appears, as Professor Smyth has well stated, that the intervening layers of limestone which are comparatively free from iron would have offered EARLE, INTERBEDDED IRON ORE DEPOSITS 137 an excellent field for progressive replacement, whereas we find more or less clean-cut contacts and underlying rich iron ore beds. A second point of disagreement is in relation to the direction of movement of the iron-bearing solutions under these theories, which, in the various papers examined, seems to be by vertical descent of ground water, often limited in depth to a few hundred feet. Thus I. C. Russell (292) refers to a case in Attalla, Alabama, in which the ore changed to ferruginous lime- stone within a few hundred feet of the surface. PART II. ARTESIAN REPLACEMENT THEORY ARTESIAN CONDITIONS CONTROLLING FACTORS Artesian conditions result from a natural arrangement of strata in such manner that they act as a retaining basin or catch basin in porous strata in which water is or may be confined under hydrostatic pressure sufficient to cause the water to rise when the reservoir is tapped. The conditions requisite for the existence of artesian wells, as set forth by Chamberlin (54a), are the following: 1) A porous stratum for the penetration of water ; 2) An impervious underlying layer to prevent the downward escape of water ; 3) An impervious overlying layer to retain the water under pressure ; 4) An inclination of the layers, at least in part, so that the point of entrance is higher than any other portion of the retaining layers; 5) A reasonably large exposure of the porous layer, in order that free entrance may be provided for the penetrating waters ; 6) Sufficient rainfall for water supply ; 7) Absence of any place of escape for the retained water. This summary of the usually quoted factors may be taken to indicate ideal artesian conditions, but many variations may exist and still allow artesian flow, although these requisites or adequate substitutes for them must be present. The pervious medium may be any crystalline or sedimentary non- crystalline rock or stratum which contains enough pore space to permit a circulation of the penetrating water. Sandstones (particularly of coarse texture), fossiliferous limestones and even in some cases coarsely crystalline limestone may serve as a carrier and saturation medium for artesian waters. In some instances, even bedding planes, laminations or fracture systems may be adequate. 138 ; ANNALS NEW YORK ACADEMY OF SCIENCES The impervious floor is not absolutely necessary. In some cases, arte- sian flow might be expected if the underlying layer, although penetrable, was less porous than the middle layer, so that the incoming water would accumulate with greater rapidity than it could escape downward. This would occur in the case of the two sandstones of marked difference in texture, the upper coarser and the lower finer. The impervious layer above is more essential, but even here extreme differences of texture may give some results, even though the overlying layer is somewhat porous. It must be remembered, moreover, that poros- ity is relative and that absolutely impervious strata are unknown. Another very important consideration with reference to these porous and impervious layers is the possibility of accumulations by precipitation or by the mechanical fillmg of voids at or near the line of contact be- tween the strata. If, for example, two sandstones, one coarser than the other, are in contact, precipitation would first occur near the contact in the pore spaces of the finer textured rock, as there the penetrating solu- tions would move with less freedom and rapidity. Mechanical sediments as well as mineral crusts might be expected to play some part toward establishing more complete artesian conditions. Circulating ground waters, following lines of least resistance, tend to establish more or less definite channels, and if these channels are fairly well retained, nature itself will attempt to improve conditions by steadily increasing the density of the carrying medium through cementation and other processes of filling the minute channels of escape. Inclination of artesian beds is necessary only for the purpose of allow- ing gravity to establish hydrostatic pressure. If the water head is suffi- ciently high to develop enough pressure to overcome the friction and other causes of retarding a free flow of water, it is fair to assume that artesian waterways may exist in practically horizontal layers, the move- ment of the water depending upon the amount of pressure exerted by the water column. The absence of an avenue of escape may be accounted for in many ways; for example, a marine slope may be terminated at the lower end by a change of texture from coarse to very fine, as sand to mud; or pre- cipitation may take place at the lower end of a runway, thus filling the voids in the previously porous layers. CIRCULATION OF ARTESIAN WATERS It has been suggested that artesian water is stagnant until tapped; but the writer is inclined to believe that absolutely stagnant artesian waters would be unusual if not impossible. EARLE, INTERBEDDED IRON ORE DEPOSITS 139 In the first place, any leakage, even though very slow, would promote circulation in the inclosed waters ; artesian conditions depend upon rela- tive, rather than absolute, imperviousness. In the second place, differences in temperature would cause the estab- lishment of currents involving a more or less constant circulation. That such differences of temperature do exist between the surface waters at the outcrop of the porous layers and the waters confined at considerable depths is hardly open to question. ‘The density of cooler water is greater than of water at higher temperatures, and therefore such water would be acted upon by gravity, causing the denser water to sink while the less dense would rise. In the third place, dissolved salts would add to density of the waters, and if solutions heavily charged with iron salts were admitted to the artesian runways, they would tend to sink until by precipitation they lost a part or most of their load, after which they would tend to rise and give place to other charged water from above. Fourth, waters charged with insoluble mechanical sediment would tend to sink and set up circulation within the runway. Fifth, oscillating movements of the water would occur because of tidal variations in load, accession of fresh surface water, crustal movements and other minor causes effecting changes of pressure transmitted in various ways, such as through the overlying impervious layer or through the water in the reservoir. Slowly moving waters thus act as carriers of iron salts and other min- erals and gases and precipitate much more readily because of relative confinement and slow method of circulation. TYPICAL SECTIONS OF CLINTON ORES MISSOURI Holt County. Drill hole. Red odlitic hematite of Clinton age, 1,885 feet below surface, showed the following section (Crane, 66) : Feet Inches Rurple shales) GimperviouS)inccseues se eece oes 21 3 Oohtiewhematiten(POrous) esos soe olen oe 3 8 Earthy argillaceous hematite................... es 5 Light green sandy shale (impervious).......... 2, 6 Bluish green shale (impervious)............... 64 9 ALABAMA In the Birmingham District, out of about 80 sections extending over forty-two properties, seventy-two included ore seams either fossil or 140 ANNALS NEW YORK ACADEMY OF SCIENCES odlitic with shale or slate directly overlying and underlying the ore beds. These seams included the Ida, Big and Irondale, besides many too small to be of practical importance. (For sections, see Eckel, 37, pp. 74-78.) Section East No. 2 mine, Red Mountain (Phillips, 266, p. 64) : Feet Inches GClaty amdSoOile vse «cre ete sst svete larseevekerens me tatoystottevenetehs 6 SAMAUSTOME) A airsysotererole cueterete sseketondistercte eiac S etchevsieve’ss 3 CU BI ss reid foie: SR syste garenend © otiekatete: stalel ovoneieye sa: o eke etekoreneh mye i SANGUSTOMEINSS ai srcpons s sisnet olors, shshelerenero etereseilorsiere iol ete feuehere 1 as Cy so oie Sees i otenstotaneredo te lavercteve mucherste Coleus cele isiees eiciews =e 2 ORG aah sco) Rae enter Cee aie eT eh teres Cee ere ate 6 CUA Reta ecoeeccs tale creretens wllorchena s/overmionetcie tienen eke fetee ss 2 OO’ cicieis. 5's. acces ve eek oe ie eile ore) ow ane-akel sraysraneceiere ehene sles 3.0 Gaye tee eiaetorcpereearecsrsbocinite ie orersvenoerte retort ee ctor al OC ere esa ciara ee als Goteue tens We He pee ane: Siisuale ish yee vatere lie el susie 4 Gay) Hearne HRA otek avons to otons athe Lele tioaloretereraearstel 4 ORCL A Sai oe elves Sees ore een aie NO siolate moe eare lie rsleate 4 Qa Ge aia lays tarcicst oer acl oiateteterai ss ies re tele ores a fove teers : 0.5 OTe reve rairccnbostare saiolicte dors lo, clei ose tevokaensh el ove, wile ct hagetehanevele site fe 1 il Gaye asc cctussa a cio tivousieucichensy overs ouave Tekeke mieuer sue tetoueveleue, ose 2 CCE e canetests Gieiins Aer oye wnsh ote ies; eee sretensreusuars Gnsvs eyene siete se 10 DVS iajars aps avcre asco let oper ores ates cakes] custodian el ere Geiss 1 ONG es scat ayaiays arava tore Mceyoielonst sastenerooretereierettaeie Ge orabavere ors 2.0 CLAY pe eraie Sita oteral tie atetrelietevelace uel amet eleienere Tooter sexes 0.5 (Oey AIO OM LCOS RCUEAMEICRERG: Chan ice OG COFOI CREO TGR OR 0.5 Clay Ss BO Getiei Meeleole Sele oisioras ernie Gielen eats eeleles 1 (ON cee IE. CS OORT CI CROC MEIC RRC COLOR FCIO. GO OR ee 2 Ay ate are vie oye ans ictoreilspevolsyet oreveclovevercueme rol verietemeaieeWereteus 0.5 ORS SINE (SLANE yor. hievere eal chelots levetonctione oho evacenehsleteboke 2 (ONT GercaIee Ca IneIO OOOO Cc apBa namo tiacoocoO oes Z Ore sine ram ede eyepeieve sterols overonorersl sleleieserserentelecoreve 1 4 SUE OMS IatS RIE CO TORIES DOE ORG Cea Ors cin: ero Icio Ceioa: ace 0.5 Ones ofine PoraIN EA iene. earths eels) eionstlatelscletcnsterenche 5 GT Biv aieerera cic cisisis seas se eheveteetotelavchernteneler she cyotehcienereres> i (Oney wins Camiiivenlos pa sno ome odgapcoedeuGRGUDOUUoUS a IGT GBOtho oi to ator aCIGlOnS IIOICTa CIGD IE ORrcre nic cinta locate 1 GEO MTIME TET AIN EO: fishers roveycrere + elavore 175 Shales, ore, the ore very sandy and in thin seams TIN SNA Cs jcrone yee reiese hensiere ioiete shevedoye We eyenel siete eveiareexere.te 10 Loam, sandy, red, with loose shales, about...... 80 Loam, ore, the ore sandy and in loose pieces in NEC SAN Give LOAme a Ollitiere rie levlencuererenoueie lene terol ceneievere 10 Pelham) (Prenton) limestones: cs cicrsicieivieicte cle ele = oie Broughton Bridge Gap, N. E. 14 of N. E. 1 Vf /4) k Inches ‘ se Cc %, { Jefferson County. West Red Mountain. Section in Pit in 8. KE. 4 of N. W. 14, Sec. 20, T. 15, R. 1 W. (266, p. 52): Feet DEDTIS;: SOUS a ousie che sersnsis wis « aheteials Gisuseueieio es Olle : Ore, Sandy, aMlare ere vans) rer elciete sieves etenerele D Shale, yellowish, only in places............... 56 OR URI OnG.ccI LGC GDC ROE LOO A .C.cre OTS 6 ShoailessyelllowalShiiivs sreletvorstere ic ciel sii ptolors che hglentese ue (ORG: Ce rerio ctateha ton eis te ces oa nletlsnelaherecsereteme tes cses 1 Shale, ore, the shale is yellowish and carries the ore only in places, in streaks........... aie OGRE rere contre falls stor ere tenehere tors Leet ntelonsnelal shetesers il Shale spy ellowaShincicuretierstrerencrctorexne vices eceies L214 Ore, Soft ascanlet Coloma ce ieuctelstete oierenetetenonetone al Ore, shale, in alternate streaks, the ore is SIO her Game BSG OOo SOMOS os tahstons He eh orate ee 1 @Orewshalen the Orel 1S Softee -icte uokevee svete tee as 7 Section of Big Seam, opposite Oxmoor (266, p. 61) : Feet OSH erie ealcus ie tetas luceicve, ste Rildie he fenedel storeheheke Jers verei eve AG 1 SHAVES MED OLESaiacc. ccleteieleloinlate «wheels ates ate teirare toes Trace OTOH es ere err eerela tc aa Wows rshoha ans (e aistedetaredevers S atts 8 STHILGS Ir raers etnercreretetcde olde ove toranelasdtolerore spaielmerecene sas as ONS AE eieiios stots vol ehetes cs yaiey syoyerahelsioig savatelteueveve, 6 siece 2 SNS Warreraarcr ea tanetatetelensticls, chen sistevoloy en siervesnetenelel siromelerrs He Oressne: siacperedealetedste ts a ORO Dore SONOS COO Ge 8 SHAMLe Wc ateerorsheder slays aleuchevene: suas Biever eS ieiiste Slayereie evexee Inches bo Whore - rc ior) , tO & bo EARLE, INTERBEDDED IRON ORE DEPOSITS 143 Feet Inches STV AIG Be, Petre retrenlel at ciot oars ofr a rebee ot an evie siesta tarioitallor eitelty on a 2 GOTO oy avel = tera eh shal al sh orec a iat sualel ave Shops otseive: « atecerealeratells S3 11 OTT GP acyekerakshatoncr cvercheiat ci suevehatepavat svaverctabelehesctatgneveless 7 ee AO) TEC yaratrexatte Satted slike! i eiier ol aie) «'ahac aavehall ae) Sieravanal cretelatotetaters 1 3 GEORGIA In Georgia, the same arrangement of strata prevails as that found in Alabama. The porous layers of sandstone or limestone are joined top ‘and bottom with impervious layers of slates or shales. In a few cases, fine-textured sandstone joined the much coarser layers of ore. 'Thirty- seven sections taken in this state showed without exception similar favor- able artesian conditions. In nearly all cases, the ore seams consisted of fossil ore and always much coarser than surrounding layers. Walker County, Lookout Mountain (216, p. 96), Edmund Evitt’s property. Lot 220, 12th district: , Feet Inches SIVAN Paya oateratinseperet sane whey Ser aete eek naa eho pita ere teberstels COC aye star cr ctoretete lonley a ele tae ar ars ates spellinielace biel ape le niatage eles. Ye - Solna Oday ere verepete rei seetabek aves cleriaputet tuarade Have ronilen aturey aulsueqec eee 4 OIE RSIS ta to CPO ORCL Oe ORE NCHS IORE REE PCN es PS Seta : 10 HOB KEY: 6 Ber stGawatic Oram Toes CIRCE DIOR BOLE Cre ODER COIR Ero Chattooga County. Lookout Mountain (216, p. 123). W. T. Henry’s property. Lot 171, 13th district: Feet Inches SVQRW G48. o Sets CIC SIE SORE OOO ICICISII Ore ICT Cee Rs OTLB pe eter oat aie shore tore: overdo oesuaisloneta alerane eis ere locas eeiele 1 2 NO NEMONeytest cera telote a cilehetena love ote) ciate char oiglousiefeleiatere.ecets 1 10 Ore rsa fos ee Bie Siete soka, oecl ehouenetabe mye ew swt oh elie ekeetove 5 4 Sly all Ginn: sddapey teres cher ace ws tad patat ee sie iol ats atelier ete aia See 6 Oregeie. DONOR CICRES DIG EEE OLS COO ICRC eT Rae 2 ST werreteuarecerstsiotere Sates, deine aleustetaarele erelew alelstersieiate Dirtseller Mountain. Z (216, p. 130). Lot 150, 14th district: Feet Inches (CURE Soi DO iC OC GEOG SIOIO Ie i TOROS SIRT an ene Ur mae Hae (e) Ler | o OHKRAODEED PA. (QR Sirol CASS atsnelatateheve la a eicvetss lavecs| eveisteve his 14-4. ANNALS NEW YORK ACADEMY OF SCIENCES Taylor’s Ridge (216, p. 133) : Feet Inches Heavy-bedded ‘Sandstone =. --.iscewicis « etereretoieioieis els (OL ey ety oro eR ERIC AOR RIS en AGG Occ otadaes ots 5 S01 1 ce er eee Rot etc mets cGeehenS. Gio ics OR CRON se aL QO cGare wis Biche vsilsice:efo, ashe ovate one corakouehenor onotelive eter ete etey ee al il OTIC TD 5 lays eve ferede asda ho wrest BSc oteke ckerlone o esenel telotersioie te Mr. Maddox’s property (216, p. 135). Lot 160: Feet Inches SANASEONE vac kove sleeve lolekovere, sheseilens ie loveieroterey oareeeieteenere ce ae acs OLE saiehein by avetts taker eyaves of te lois ohare teliartarelovereveleteielel efekeiete ler le 56 7.5 SHA CMe er cccioieucis cite toe rovetnic, eteiGue aincle ere Mra creat fe 3 OTC reo teiver cece svonstovaseuete iota) she tone Weiss ete Jonata tetouetete tereraneiehiwetee te i a bts SAME accrrerctavtaciate eisreuets des fo ete owe iskere ee neraven na eon 7 Dade County. Lookout Creek (216, p. 59). Lot 83, 18th district: Feet Inches Ore (overlain Dy Shale) micccciesiete ae ciclo isacio es Sie 10 SAVOY eoies-sepesouckcieveke ates cisyaie sors iolelin eveieleretelsleois: obeForevere ae 8 (DCs one vs Taifareie ie iofoto eG foie feloiey ase coleie ls cave polevereieis eislaseerere 1 ae SINR Rais otto tehete bee are rern lays fosloue. velo Pavers. Whale elec lord wfoiete 1 ORISA ROR ROS LOND OIG ORT RE CEOS eet rr ee Ste a 6 SIAC Gis thee etocleratelo ae efovelotoelo e Bvsteleos ie ntotowlere tote olor 3 ORC Meo efoecs Biot usters Santis « ike Siewe woth Oho ore iforsiele 6 SAG ercteyratetesspeynlste avers, ote S ove) eaene eho evore ersten ole eneretel shave Be il OC Fer ucuede retteveh oteuecy cle tsieiei@ lavehevere ioiee ete talloreteter tie erences 1 STAC metare.cteystaetshoha eters bie) Sialtehelais lem retslne cam lepeitor aie 2 Herrurcinous) fOSSil Limestone. /reis cl eters ole) ole als ae 9 SH al Oop gee rarctayeee re ates fe ret bieiel old atekors ers tekoetslepeltenstetenerts 6.5 HOSSTIFELOUSs LIMESTONE S scr uss-atereteceuehelotener susie strche ous 4 Shale with thin layers of sandstone and lime- SCOM OM ere caerelavs ous ieis, PA ayeveso elaiis ol eipiecetetenevensielers, Suelo 40 TENNESSEE Safford (295, p. 304): Feet Inches Greenish shale........... Weis rele ie bretedeuetone. doavetereceevels 22 % Orevwithsparcine Of Shales..-v. cco ctciirele cleleroiseeleTe i 6 SANASTONE aN CuSl aimed syste cicie loveietebeiche Sleketoteleleleiete 6 Greenish shale with occasional iron seams...... 67 OolitiChorew calcareous jose vee s celcls we Se nictetelelos 4 a GPECNISHESH RLS che ort sis wcois ct bie c she eer Miolore eieorei Rare loiots 21 3 VIRGINIA Low Moor, Horse Mountain (92, p. 188) : Feet Inches Slates andsthinisaAndSstome vee cls isle icles vic sls, « ie 3 10 Fossil ore, brown and porous.............. 5 7 Hossilsorenredsand tainly; SOOM sec. ete ieee eel = 1 2 CHER Ye Cl ayacde rere sercistoreitoteleiens isos sic leveushe siete) Crelisyeileyoxs 6 RSH ET VEST ang ye earataass Ans RCE AEE CRC OREN Cin, Oc RRO ICA ACER Dy EARLE, INTERBEDDED IRON ORE DEPOSITS 145 Slope 2: Feet Inches WOH CUWENNWANIEG BK5ooobOobbUGODODnDOOCOUuDOUOCS 30 a SMITE Sieyeeereeraia cress related cdenah axed af earteychcbercne) ester ehohehotehshe lis pe FRCAGROSSTINOLGretsr einer eves sch) doyevchoneiel clalciietorsacrelsioneieie 1 4 Ochery shales and thin sandstone....,......... il ShalestanGesamadStOnme acces ctelslersvele ovetetatstoneloictelehs 2, Big Stone Gap, Va.: = ae ee Slate spac yanvea tiered sac cis cterelerclelelcierclereiels 1X PLATE XXIV, a VoLuU ME Scr. ANNALS N. Y. ACAD. ——~s. a hae a £40? Us am ANNALS N. Y. ACAD. SCI. VOLUME XXIV, PLATE X FIGURE 1 FIGURE 2 ‘ . ’ 1 \ ® \ =a . F « . a i | ' ’ ’ «¥ PLATE XI 1 XXIV, VOLUME Sct. ACAD. ANNALS N. Y. 1 FIGURE ANNALS N. Y. ACAD. SCI. VOLUME XXIV, PuatTHy XII FIGURE 1 FIGURE 2 tS * = 4A _ — > ba uw = Ss ~*~ ANNALS N. Y. ACAD. SCI. VOLUME XXIV, PLATE XITI FiGgureE 1 ~. id , , "s ee , a —% 4 a ¥ Tod or > , ee ae woe. te wy” * ed ¢ . ai 4 SPL FIGURE 2 ’ 4 ' ee . ; ’ i ~ si ' oa ie 1 ' . s > o i) ‘ © > F i A 1 7 | ’ , ; ‘ a « - ‘ F . ¥ ‘ ‘ ™ ' = 4 ' - P i 6 = ai ~ ¥ ‘ * « - rt . oe a ANNALS N. Y. ACAD. Scr VOLUME XXIV, PLATE XIV FIGURE 1 FIGURE 2 th yu VOLUME XXIV, PLATE XV ANNALS N. Y. Acap. SCI. y st > note, FIGURE 1 2 FIGURE a 2 2 ANNALS N. Y. ACAD. SCI. VOLUME XXIV, PLATE XVI FIGURE 1 FIGURE 2 ‘» r * z - i a _ 7 7 » 4 > a + < - _ nd f ‘ a = , fy , « ' 1 , ni a vv ~ . - - - 4 ye © ANNALS N. Y. ACAD. SCI. VOLUME XXIV, PLATE XVII FIGURE 1 FIGURE 2 dy. XXIV, Puatr XVITI VoLUME Scr. ACAD. Ni ANNALS N. FIGURE 2 XXIV, Prarn XEX VOLUME ACAD. SCI. Y:. ANNALS N. 1 FIGURE » “= FIGURE ANNALS N. Y. ACAD. SCI. VOLUME XXIV, PLATE XX . i te * oa : x ‘ - = i i « a ° _—_ 7 7 P . . a | . . 2 * ae ‘ ‘ -“ ae © ei : 7 ( ANNALS N. Y. ACAD. SCI. VOLUMD XXIV, PLATE XXI PUBLICATIONS OF THE- NEW YORK ACADEMY OF SCIENCES (Lyceum oF Naturat History, 1817-1876) The publications of the Academy consist of two series, viz.: (1) The Annals (octavo series), established in 1823, contain the scientific contributions and reports of researches, together with the rec- ords of meetings and similar matter. A volume of the Annals coincides in general with the calendar year and is sold at the uniform price of three dollars per volume. The articles composing the volume are printed separately, each in its own cover, and are distributed in bundles on an average of three per year. The price of the separate articles depends upon their length and the number of illus- trations, and may be learned upon application to the Librarian of the Academy. The author receives his separates as soon as his paper has been printed, the date of issue appearing-above the title of each paper. (2), The Memoirs (quarto series), established in 1895, are issudd at irregular intervals. It is intended that each volume shall be devoted to monographs relating to some particular department of Science. Volume I is devoted to Astronomical Memoirs, Volume II to Zodlogical Memoirs, etc. The price is one dollar per part as issued. All publications are sent free to Fellows and Active Members. The Annals are sent to Honorary and Corresponding Members desiring them. Subscriptions and inquiries concerning current and back numbers of any of the publications of the Academy should be addressed to THE LIBRARIAN, New York Academy of Sciences, care of American Museum of Natural History, New York, N. Y. PRESS OF JUDD & DETWEILER, INC., WASHINGTON; D. C. je ae oS, i oe, ANNALS OF THE NEW YORK ACADEMY OF SCIENCES Vol. XXIV, pp. 171-318 Hditor, EpMuND OTIs Hovey CLIMATE AND EVOLUTION BY W. D. MatTtHEW NEW YORK PUBLISHED BY THE ACADEMY 18 Fepruary, 1915 THE NEW YORK ACADEMY OF SCIENCES (Lyozum or Naturat History, 1817-1876) OFFIOERS, 1914 President—@eorcr FrepericK Kunz, 601 West 110th Street Vice-Presidents—CuHaArRLES P, BrerKey, RayMonpD C. OsBuURN,_ CHARLES BASKERVILLE, CLARK WISSLER Corresponding Secretary—Hrnry HE. Crampron, American Museum — Recording Secretary—KpmMuND Otis Hovey, American Museum Treasurer—Hernry L. Donerty, 60 Wall Street Librarian—RawtrexH W. Tower, American Museum Editor—EpmMunD Ot1s Hovey, American Museum SECTION OF GHOLOGY AND MINERALOGY Chairman—Cuar.eEs P, Berkey, Columbia University Secretary—A. B. Pactn1, 147 Varick Street SHCTION OF BIOLOGY Chairman—RayMonpd C. Ospurn, 557 West 124th Street Secretary—Wi.Lu1aM K. Grecory, American Museum SECTION OF ASTRONOMY, PHYSICS AND CHEMISTRY Chairman—CHartrs BASKERVILLE, College of the City of New York . Secretary—Ernest 1. Situ, 50° East 41st Street — SECPION OF ANTHROPOLOGY AND PSYCHOLOGY in Chairman—Crark WissLER, American Museum _Secretary—Roserrt H. Low1z, American Museum ‘ The sessions of the Academy are held on Monday evenings at 8:15 - o’clock from October to May, inclusive, at the American Museum of Natural History, 77th Street and Central Park, West. ‘i 1 ANNALS N. Y. Acap. Sci., Vol. XXIV, pp. 171-318. 18 February, 1915] CLIMATE AND EVOLUTION! By W. D. MatTtTHEW (Presented in abstract before the Academy, 13 February, 1911) CONTENTS Page DTU CLS EMM ETI e Gaye boats 2 ace dates Me nerap etal Stepetea nm del Sials ete wieweite ce 18@) Ore Poe oak cree 172 IFAT eT CRC tal Oneness cto, te otireceroiel avon aller ech eter cred etetotanteyeireltal anal et ele (oraseliohaie @erene si elerer 173 Alternations of elevation and climate during geological time....... a6 Alri Rewmine paseo sien GIECRIN RSIS a aqdae eo bododa soo Ob coo tNaoomadcdc 174 Distribution of land and water, present and past..................- 175 Effects of alternations of elevation and climate upon evolution of COLES EMEA le PAM el Srerere siete oe oiiete- BRE SRO OE OO HO AEIOD pic dino Oc 176 Comparison with the paleontological record........ SA LS chee 178 Interpretation of supposed exceptions................+-00:ss-0--0-- 179 PTINGI PLES sof MISPELSA lla, as sve hevs ie siete sas ceeypoepateyeiewsvnds a)e,8 ERS OM OER otose 180 Review of the evolution of vertebrate life............-....0.-02.0.. 181 Imperfection of the geological record.............-.....++- 189 VeGSTONA CORTE) ALOMrgers vets aieienavaie eleletetsiel cle sellecefeoroi ERE RR RCSA ol CR eear a ts 191 Synchronism and homotaxis........... Sedes sate SPRUE, owt stat oaewetepare Sauaks ts 192 ; TAATAAy CORRARIHOIN Tol Souiln wlNeinKNes oa oobnoe donc s00o 005 Soo ou sO oS 195 GCENCLELSHOsGISPERSAlevere serie ete oles |= la ere raters eve 1s (ots tele eheliotelatevoles sie WO Sraictalehee 200 Oceanicrandaconrinentaluis!andSryen cael seicios cteis o averei ole eyelet ale) eller aie\elote el ereres 202 Faunal differences between oceanic and continental islands.......... 202 Natural rafts and the probabilities of over-sea migration thereby.... 206 Considerations affecting probabilities of over-sea migration in special GASES? ciajsie shaves ons Seaensreues ER hapa ath chenevitie evs tata a ataits vance eyicr otepe aisle swale ateponse SOS ID ISGELSANO te eiey TT AN Aeterna alee eels) elena) feline aterchalaiielajaite).cl stele Bt spelehs ey-vste 209 WB Vis 5 ey ose See NAMM a tic satens raratarey er cheverateiaNenet cliellsiehs essreste leyoue: afousel's Paiste Sas 209 ESPIMAATCS yey cletateiiekechetheleveune ee Bot kekebadeenats revels eC Nraratist sie COVA a aMebeuallaras @isy shane eit 214 (GRIT SHON ERI. ORs een Ales tier 6 EIR RICE CHOY ONC RERORCR) ERE ca ee nCROT on te PRE cc Beek Ped clten! Ors 217 COE INIG Baineena teecicmen ett Nee eS BOA tena atte toe ai cealetes PPR yah LAS RON ed PR eae Pa 218 TO CY OMe efeueyelere sie avy eles eels coe ele Arete eM beleyaie witthensintetese elected 220 IVIISEGIT CEE taeveleerciatsvereter siciere tetetslene ee lers AB AUR etereteralstavernie wileletako crenata ows 221 UOTEEHOR DAIS OSE OF Gis Widid ides Se Gore td DIOP RCo Ce Ie eee ECRECICIcrO taro ac 221 WEI GT ET ASO tea yereievetara sues ce! aleurone eles haere nehereiets elatadhine withoeh atoints asics 222 15 Wiss) Pa ae Gio eels Ns Te eaCTCIAENG adecal etches cial ot cuntlavehancus ie isdel exe (6: 8s, 66F sus)» 223 TOYS Ee Snes eel HAR eee. cI ONEe CRI CHG REMI Aahaicusvsvarsts ESP APNA eat eae ererere sae 223 TPMT AYER Gp aad PROCS Oi he ov PONS CG TS CS RAC NCTE THERE ARNON EE MICRO IGN SNE RCV he 27-34 PMSECULVOLE 2) 6:2 a: eheiei ais ane PieToreisl shetevars saceieie isis einiele\ere! sre Fe SS OP ALONG Oe 224 Chiroptera....... Se OORT, COTO Ie shevlaba tabs ye ate: re deve ata a9 el eae eee 227 LOGO Sg o SES ee dics Ors ORE TG ERC CRO Cie ern CRT aeRO PS AO Ps ctr Cb 228 1 Manuscript received by the Editor, 20 October, 1914. 172 ANNALS NEW YORK ACADEMY OF SCIENCES , Page POCLISSOGEACEVIA «a: odes hisysteusieveve eile ieaclavttevohelevstartetencla is fate MAIS oie Dake Gee Oe UL Se revere rctete ouetetaciereve A COMIC COO OOS See pono Biso bob ofa. - 235 MAMI. Slayers elere ererel« le Be I OC EME OER enc a cA sine Oi 238 RHINOCCLOCIEH she cs Se cease lok elote steer tio ee oC eens aie ee eee 240 AT GLO MACE Aras eie wast tain oroiets le a ole oletoiors oe cuel epitope eel Ieee ea oe 241 Pigs Jana! WCCCALIES isle taisic exe cheseve oasis + cloles ole hele o hein e inion eee 241 FUUIITIATUGS eae tee ie ctepeh oteyei sued tsietst steie« Fe cee dyh ao loko SR ee 242 PrODOSCIMED . Sie ars eis, SF reefs) Posie nS te oy eicete eee tains Gh rete ere te eee eee 254 SUT OTT AN 67 ja, oe whe cereal S-oje ys foueco char aelle \e inftes apeFe feheks aeons teweiore bolted mevete ee shouetot ale toner oeetene 256 Condylarthra and, specialized \suceeSSors......2+-500 o20-+ssene es soe een eOm HOM atayyer.) he chevete tobe teves ose love: csi Mews! eteve owniloletspere eso. ehayeerevavelione ie odetelevenene toleae vent 259 Marsupialia....... a ee OCR REC Ee OS COO DO Ono Cr 262 MON OEREMM ACA. os cys le tere allele ais ogote Sis cae ede Ghchsiete s eutin es estiee AG Uie Et ove E Te ene 270 Summary of the evidence from dispersal of land mammals.......... 270 Interpretation of negative evidence in fossil mammal faune......... 273 Dispersal of reptilia....... 5 Ae te ibre.tocies ere weeks we oseyalic reve lake awed tesegai eet TaiGh eee 274 Dinosauria...... sssduaile seuss eta ateberai’suabote cua ol shelekols wi eiejsbaeneue: ethaleiduelane suoWateustoumeee 275 (OL LOOUE heee airee RNA CIARA Ae Robina ieee eM TCT OR aE mee Ee cans care 6 280 CrOCOMTMART Ki. derok red. setete cicys e.os Clae je Sun steleliowje sietars else jyctale 'oreccle ete, aseporetees 284 TOS COT GUT A area lsratere etore ee cle, bree eve heac a fos oneicerauayenere beta obs, ote sete tells neds Lat velar Rotate 288 PISHEESAIU OF, MOUS eye bate toysites's: eulers' oles uci ew cella we te else ictie: eens Yon shelope borer etter: oxeiciedevastoeee, as) SePemeC ea ee WISMELSAl OL sAIMP MUP ere .= epoca voters @) vie, ai cv recoie eieusle oe eheher shepelernslemey oer sper apcNorarenetete 294 PMispersal vot mresh=wabter f1SheSeyee tacts clove ciereoricle nie Serer tare te etere ote 297 General considerations on the distribution of invertebrates and plants.... 299 interpretation -of distribution data, of ierayiish; .).2 6 ©) ne ene 301 Distribution ot elias MOrtensis. .\. saosin oh oe clowns <>) da .cleeeeemee 303 DIS GED UGLOM TO te ee CICLO evens cotsicves svoncie)oieucliolerc¥ eters totete rots teier ~/eveuorstsicattel a eoeraaet eae 304 CricgicisMmeot, SOMETOPPOSLNE MhivIp OLNESES yei-weveyevere ciel aireeite seve) =e cues oleiene ou-patloueiiosellers 305 On vain speculations......... BHC R TR Oe OUEST EEE oman io ae oc 306 Summary of evidence....... ais echs Peta tin ar noner least ah omens rahe cue) SRE Poe CLAN cue eee Peete 308 FAD PEM GE |S ers Natoma le itctlete, eves ALORS SAGA RAE OR SA ic TEE BOERS enti Ct & 311 THESIS 1. Secular climatic change has been an important factor in the evolu- tion of land vertebrates and the principal known cause of their present distribution. 2. The principal lines of migration in later geological epochs have been radial from Holarctice centers of dispersal. 3. The geographic changes required to explain the present distribution of land vertebrates are not extensive and for the most part do not affect the permanence of the oceans as defined by the continental shelf. 4. The theories of alternations of moist and uniform with arid and zonal climates, as elaborated by Chamberlin, are in exact accord with the course of evolution of land vertebrates, when interpreted with due allowance for the probable gaps in the record. MATTHEW, CLIMATE AND EVOLUTION 73 5. The numerous hypothetical land bridges in temperate tropical and southern regions, connecting continents now separated by deep oceans, which have been advocated by various authors, are improbable and un- necessary to explain geographic distribution. On the contrary, the known facts point distinctly to a general permanency of continental out- lines during the later epochs of geologic time, provided that due allow- ance be made for the known or probable gaps in our knowledge. INTRODUCTION ALTERNATIONS OF ELEVATION AND CLIMATE DURING GEOLOGICAL TIME Several years ago,? I had the honor to give a talk upon “Climate and Evolution” before the Linnean Society. The subject was then new to me—it was an application to vertebrate paleontology of theories in regard to geological history which had been brought forward by Cham- berlin a year or two previously. I have had these concepts more or less in mind ever since, and though I must admit that I am far from having the evidence in shape for final presentation, I desire to submit for gen- eral consideration the conclusions thus far reached. Chamberlin’s theories are to-day well known and are year by year gaining a wider acceptance. So far as they pertain to the present sub- ject, they differ from the older prevailing concept of geological climatic conditions chiefly in that they involve an alternation of climates through the course of geologic time from extremes of warm, moist tropical and uniform, to extremes of cold, arid zonal climates. The former are the results of prolonged base-level erosion and the overflow of large conti- nental areas by shallow seas. The latter are the results of the re-adjust- ments needed to bring the continents once more into isostatic balance, involving the general lifting of the continents, especially of their borders, the expansion of the continental areas to their utmost limits and the renewal of rapid erosion. These alternations of conditions are marked by alternations of the prevalent type of formation in the geological series. The uniform base- leveling corresponds to widespread deposits of limestones and in its waning stages with coal formations. The periods of uplift are marked by thick barren formations, often red in color, by indications of arid conditions in salt and gypsum beds and they finally culminate in great extension of glaciers from boreal and high mountain areas. 2 Jan. 14, 1902. ’T. C. CHAMBERLIN: Jour. Geol., vols. v-viii. 1897-1901. 17a: ANNALS NEW YORK ACADEMY OF SCIENCES Chamberlin’s text book of geology may be consulted, for the more exact and extended exposition of these theories. The present purpose is to in- dicate their application to the evolution of land vertebrates. PERMANENCY OF THE OCEAN BASINS In the first place, we may note that they depend as a fundamental basis on the general permanency of the great ocean basins. The conti- Fig. 1. The areas within the continental shelf (100-fathom line) are left unshaded. This map represents the true relations of land and water in the northern hemisphere far more correctly than does the usual Mercator projection. The unity of Arctogzea and the direct relation is obvious between the various degrees of isolation of the southern continents and of peculiarity of their faune. Zoological regions on north polar projection nents have been alternately partly overflowed, separated and insular, or raised to their greatest extent and united largely into a single mass. The great ocean basins have in the main been permanent. This principle is MATTHEW, CLIMATE AND EVOLUTION 175 dependent upon the known facts in regard to isostasy. The rocks under- lying the oceans are heavier than those underlying the continents, as is proved by the deficiency of gravity measurements in the continents as compared with those in oceanic areas, the deficiency being most marked in certain, mostly high-lying parts of the continents. The conclusion appears unavoidable that in a broad way the present distribution of land and shallow water on the one hand, of deep water on the other, has been substantially unchanged.* Changes in past geography have been of two kinds : 1) The continents have been alternately partly overflowed and then have emerged to the limits of the continental shelf. 2) Certain lines of unstable conditions have been subject to folding and crumpling, accompanied with great changes of level. DISTRIBUTION OF LAND AND WATER, PRESENT AND PAST The present distribution of land and water shows the great land masses located mostly in the northern hemisphere.® The land areas, extended to the borders of the continental shelf, form a single great irregular mass with three great projections, South America, Africa and Australasia, radiating out from it into the southern hemisphere. A rise of 600 feet would unite all the land into a single mass.° Only New Zealand, Mada- gascar, the Antilles and numerous small oceanic islands would remain separate. The Kast Indian islands would be part of the main land. A lowering of 600 feet would isolate North America, South America, Asia, Africa and Australia as separate insular continents. Europe would form a complex of islands and peninsulas much like the East Indies of to-day. : According to the present theory, we have recently passed through an epoch of maximum continental extension and zonal climate culminating in the Glacial age, marked by great aridity in the equatorial zones, by cold and glaciation towards the poles and in high mountain regions. A much earlier extreme of aridity and glaciation is seen in the Permian,’ and less marked extremes at the end of the Trias and at the beginning and end of the Cretaceous. The alternate extremes of warm moist and ‘In this connection, however, the suggestion of Bailey Willis that the present isostatic compensation may be unusually complete must be borne in mind. ° It should be observed that the Antarctic continent, according to the latest data avail- able, equals or exceeds any of the other continents in bulk of emerged land: but it is sur- rounded by deep oceans of vast extent. ® Australia forms a doubtful exception. The soundings in the Indo-Australasian region are insufficient to determine with certainty whether or not there is any continu- ous bridge within the 100-fathom line. “The earlier Paleozoic extremes of aridity—-Cambrian and Devonian—do not come within the scope of this discussion. onditions of com- © > J rTXN Wave ACADEMY OF SCIENCES ANNALS NEW YORK Now the base-leveling and overflow con- msion and growth of marine c The « arrow border of the continental shelf will be unfavorable 2.—The southern continents, south polar projection WiG. 1000-2000, and over 2000 fathoms indicated by progressive ’ The steep margins of the continental shelf The isolation of the southern continents is in contrast to at emergence of the continents will tend to OF TERRESTRIAL FAUNAS uniform climates are seen in the early Carboniferous, in the Jurassic, mid-Cretaceous and Hocene. and will tend to what Chamberlin calls restrictive evolution of faunas. plete emergence of the continents and restriction of the littoral life to ditions are obviously favorable to the exp life, especially of the littoral and shallow seas. the steep and n 176 x TYAN S OX XY RAY sa a RATE 970A Pr, REARS ea ar av ROOK Associated with the isolated continents, we have moist tropical uni- form conditions of climate, and to this the provincial land faune of Ocean depths of 100-1000 shading. Less than 100 fathoms unshaded. are indicated by hachures. EFFECTS OF ALTERNATIONS OF ELEVATION AND CLIMATE UPON EVOLUTION flow and isolation will tend to the restriction of land migration and the expansional evolution and cosmopolitan faunas, while their partial over- development of provincial faune. the unity of the northern land areas. Conversely on land, the gre MATTHEW, CLIMATE AND EVOLUTION wey these periods will be especially adapted. The periods of continental emergence were periods of arid and markedly zonal climate, and the faune must adapt themselves to these conditions. Such conditions, while favoring the spread and wide dis- tribution of races, would be unfavora- ble to abundance of life and the ease with which animals could obtain a liv- ing. The animals subjected to them must maintain themselves against the inclemency of nature, the scarcity of food, the variations of temperature, as well as against the competition of rivals and the attacks of enemies. In the moist tropical climatic phase, animals would find food abundant and tempera- ture relatively constant; but the larger percentage of carbonic acid and prob- ably smaller percentage of oxygen in the atmosphere during those phases would tend to sluggishness. We should expect, therefore, to find in the land life adapted to the arid cli- matic phase a greater activity and higher development of life, special adaptations to resist violent changes in temperature and specializations fitting them to the open grassy plains and des- ert life. Jn the moist tropical phase of land life, we should expect to find adaptations to abundant food, to rela- tively sluggish life and to the great ex- panse of swamp and forest vegetation that should characterize such a phase of climate. The oncoming cold and arid condi- tions should appear first at the poles and spread towards the temperate and tropical regions. Owing to the distri- bution of the great land masses, this would involve a general tendency for the great migrations resulting from the EAST INDIES Central Asia Asia Minor Three pn Cross-section of continental platforms and ocean basins at the equator Vertical scale exaggerated about 170 times. Mid-Atlantic Ridge Vertical scale greatly exaggerated. America Oross-section of continental platforms and ocean basins at 45° north latitude North Wie. 4, Via. 3. 178 ANNALS NEW YORK ACADEMY OF SCIENCES emergence of the continents to be outward from the two great northerly masses, and especially from Asia. The tropical and southern continents would be the refuge of the less adaptable and progressive types. This phase of climate should, therefore, favor a higher development and greater activity of land life, while the geographic conditions favor cosmopolitan faunz. When the climatic pendulum began to reverse its ' swing, the continents became isolated and fheir faunw developed inde- pendently ; but the dominant animals of these faune when first isolated would be those previously developed during the arid phase, and these would readapt themselves to the new conditions of moist and uniform climate, of prevalent forest and swamp and of abundant food. COMPARISON WITH THE PALEONTOLOGICAL RECORD How far do these a priori deductions correspond with the facts, as obtained from the geological record? In the first place, we should keep in mind that our record of the land life of the emergence phases is very defective. 'The sediments of this phase, where deposited along the con- tinental margins, are limited in area, thick and very barren, the condi- tions of their deposition being generally unfavorable to the preservation of fossils. The sediments of the interior of the continents, river and floodplain deposits of the Cenozoic era are more widespread and furnish an extensive record of Tertiary and Quaternary land life; but those of the preceding periods of aridity have been re-eroded and carried down to the marginal and littoral areas during the period that has elapsed since they were first deposited. Of the pre-Tertiary epicontinental de- posits, only the coast margin, littoral and marine deposits are extensively preserved. That means that the record of Mesozoic and Paleozoic land life as preserved to us is chiefly the record of the coast-swamp and low- land regions and that we know nothing of the life of the upland, except by a rare accidental preservation. In considering the evidences of cli- matic adaptation during the Mesozoic, this must be kept clearly in mind. The great mass of evidence in favor of adaptation to progressively arid climate and of dispersal from the northern land regions is derived from the recorded history of the Mammalia during the Tertiary and Quaternary and from comparison of their former and present geographi- cal distribution. It has long been recognized that the present distribu- tion of mammals is due chiefly to migration from the great northern land mass, and the connection of this southward march with progressive refrigeration in the polar regions was made more than a century ago MATTHEW, CLIMATE AND EVOLUTION 179 (1778) by Buffon.* With a clearer perspective of geologic time and far more exact records, it is clear that most of this deployment and dispersal of the mammalian races has taken place since the Hocene epoch of the Tertiary, although remnants of an older dispersal on the same lines are probably traceable in the present habitat of monotremes, marsupials and primitive insectivores. INTERPRETATION OF SUPPOSED EXCEPTIONS There has been a disposition in recent years among students of geo- graphical distribution to lay weight upon certain apparent exceptions to this general rule, where the geological record has not yet afforded evi- dence to support the northerly origin of certain groups now limited to the southern continents or to the tropics and to infer various equatorial or southern continental connections during or previous to the Tertiary, in order to account for these exceptions.® To these hypotheses, there are several objections : 1) The evidence for the general permanence of the great ocean basins and their maintenance formerly. as now, by isostatic balance is very strong and direct, and before allowing any exceptions, we should be very sure that no other explanation will serve. 2) The instances adduced in favor of former equatorial or southern con- nections are distinctly exceptional cases in the faunz, which may, in all the cases I have examined, be accounted for by appealing to the imperfection of the geologic record, by parallelism or by the rare accidents of over-sea trans- portation. 3) The existence of such land bridges would present the opportunity for migration of other parts or of the whole of certain faunz, which bas evidently not occurred. I can see no good reason why the only animals which availed themselves of such continental bridges should be the ones which might be accounted for in other ways, while those which would furnish conelusive proof are invariably absent. SSee K. y. Zittel, History of Geology and Paleontology, p. 43, for a brief summary of . Buffon’s views on this subject. The theory has been more fully presented by many sub- sequent writers. In recent years, it has been very ably set forth in its relations to Tertiary mammalia by Dr. J. L. Wortman (Amer. Jour. Sci., 1903). A very readable little pamphlet by G. Hilton Scribner, entitled ‘‘Where Did Life Begin’, 1884, while totally deficient in geological perspective, sets forth very clearly the diverse effect upon migration of the general trend of the great mountain system, north and south in the New World, east and west in the Old. Alfred Russell Wallace is, I believe, usually regarded as the foremost exponent of this theory on the distributional side: but it is scarcely necessary to catalogue the principal exponents of a view so long and so gen- erally held. ®The distinguished Argentine paleontologist, Florentino Ameghino, has for twenty years past advocated a theory the direct opposite to that currently held, and he would derive practically all groups of mammalia from a South American center of dispersal. The evidence for and objections to this theory will be discussed in the sequel. 180 ANNALS NEW YORK ACADEMY OF SCIENCES 4) Many students of geographic distribution proceed on what appear to me to be wholly false premises. They assume that the habitat of the most primi- tive living member of a race is the original habitat of the race, the most ad- vanced forms inhabiting the limit of its migration. It seems to me that we should assume directly the reverse of this. PRINCIPLES OF DISPERSAL Whatever agencies may be assigned as the cause of evolution of a race, it should be at first most progressive at its point of original dispersal, and it will continue this progress at that point in response to whatever stimulus originally caused it and spread out in successive waves of migration, each wave a stage higher than the previous one. At any one time, therefore, the most advanced stages should be nearest the center of dispersal, the most conservative stages farthest from it. It is not in Australia that we should look for the ancestry of man, but in Asia. In the same way, in considering the evidence from extinct species as to the center of dispersal of a race, it has frequently been assumed that the region where the most primitive member of a race has been found should be regarded as the source of the race, although in some instances more advanced species of the same race were living at the same time in other regions. The discovery of very primitive sirenians in Egypt while at the same time much more advanced sirenians were living in Europe has been regarded as evidence that Africa was the center of dispersal of this order. It is to my mind good evidence that it was not. It is very common to see references to the African facies of the Miocene or Pliocene mammals of Europe; but it is much more correct to say that the modern African fauna is of Tertiary aspect and is in large part the late Tertiary fauna of the northern world, driven southward by chmatie change and the competition of higher types. The chief arguments advanced in support of the method here criticized appear to be that the modification of a race is due to the changes in its environment and that the primitive species are altered more and more as they spread out or migrate into a new environment; but, assuming that a species is the product of its environment, the conclusions drawn would only hold true if the environment remained constant. This is assuredly not the case, and if it were there would be no cause left for the species to change its range. In fact, it is the environment itself, biotic as well as physical, that migrates, and the primitive species are | those which have followed it, while those which remained have had to adapt themselves to a new environment and become altered thereby. Probably, it is never the case that the environment of the marginal MATTHEW, CLIMATE AND EVOLUTION 181 species is an absolute replica of the older environment of the race. In many cases, it must be profoundly modified by its invasion of new regions, and there are many features in the evolution of a race which appear to be only partly, if at all, dependent on environmental change. But to assume that the present habitat of the most generalized members of a group, or the region where it is now most abundant, is the center from which its migrations took place in former times appears to me wholly illogical and, if applied to the higher animals as it has been to fishes and invertebrates, it would lead to results absolutely at variance with the known facts of the geologic record. REVIEW OF THE EVOLUTION OF VERTEBRATE LIFE To my mind, this hypothesis of the evolution of land life in adaptation to recurrent periods of aridity supplies a satisfactory background of cause for the whole evolution of the higher vertebrates. We may set aside earlier periods of aridity and continental extension signalized by the development of invertebrate land types, whose early terrestrial adaptation is wholly hypothetical, since the known portion of their history is so small and so remote from their origin that we cannot project it backwards with any sort of exactness. As Barrell has pointed out, the arid period of the late Devonian coincides with the probable time of the first adaptation of vertebrates to terrestrial life. In the arid period of the Permian, we see the conditions more clearly prevalent which favored a much more extensive development of land life, and this period marks the rise and early differentiation of the Reptiha. That reptiles first differentiated from amphibia as a dry-land adaptation seems to be obvious; that the period of their rise corresponded with the greatest ex- treme of aridity, continental emergence and glaciation between Cambrian and Quaternary would, I think, be also generally admitted. The domi- nant order of land reptiles up to the close of the Mesozoic was the dino- saurs, preéminently a dry-land adaptation in their inception, since their most marked characteristic lies in their long limbs, bipedal progression and general parallelism in proportions and structure to the large ground- birds of modern times, which are to-day peculiarly inhabitants of arid regions. The relationship and origin of the more specialized, mostly gigantic, dinosaurs of the later Mesozoic can be best explained by regard- ing them as a succession of derivatives from smaller and more lightly constructed upland dinosaurs, mostly unknown to us, the larger and more specialized types being re-adapted to a swamp life and inhabiting the coast marshes whose sediments are still preserved, while the more direct 32 ANNALS NEW YORK ACADEMY OF SCIENCES line of dinosaurian evolution inhabited the uplands, where the sediments, if such were deposited, have long since been removed by erosion, and the fauna is consequently unknown to us, except by inference. It is quite impossible to trace the evolution of the dinosaurian phyla through the same nearly direct series of known forms as can be done in the phyla of Tertiary mammals. But I may observe that if our knowledge of the Ter- tiary sediments were limited to the coastal ‘swamp deposits,—if in this country, for instance, we knew only the Tertiary of the Atlantic and Gulf coasts,—we would be equally at a loss for any direct ancestral series illustrating the evolution of the Mammalia. The same explanation, namely, that the geological record in the Meso- zoie is defective where its evidence would be most direct as to the evolu- tion of land vertebrates, applies both to birds and to mammals, but espe- cially to the former. The exceeding scantiness of fossil birds and mam- mals during the Mesozoic and their apparently sudden appearance in the record, already well deployed, is often explained by supposing them to have evolved mainly in some continent not yet investigated. It appears to me that a simpler and more probable explanation lies in the fact that the formations of the interior of the Mesozoic continents have in general not been preserved and that this facies of the Mesozoic faunz is conse- quently unknown to us. It may be objected that remains of dry-land animals would be brought down by rivers and deposited in their deltas and thus preserved to our day. This may, of course, occur in exceptional cases. How rare is the exception, we may judge from the exceeding rarity of remains of land animals in true marine deposits, where the chances for their preservation should be almost equally great. In marked contrast with the evolutionary record among dinosaurs, stands the record of development of the non-marine crocodiles and che- lonians, whose normal habitat was the swamp regions and whose more direct evolution is in consequence recorded since the Mesozoic. Remain- ing in a constant environment, they evolved but little, though their abun- dance and geographical distribution varied. Throughout all the evolutionary history of the vertebrates, we see numerous examples of races which, having become adapted to a higher plane of life, have re-invaded a lower plane. In each instance, the higher organization and greater activity acquired in the higher plane have caused them to become dominant, increase rapidly in size and spread widely in the absence of efficient competition. Thus we find various groups of ma- rine reptiles appearing with apparent suddenness in the Mesozoic, becom- ing very abundant and of gigantic size, spreading very widely and then MATTHEW, CLIMATE AND HVOLUTION 183 being replaced by new invasions from the land instead of evolving further in their new habitat. The ichthyosaurs, plesiosaurs, mosasaurs, sea- crocodiles, sea-turtles, are examples of this sort among reptiles; the ceta- ceans and seals among mammals. ‘These invasions from a higher to a lower plane of active hfe have been very frequent, so that their recogni- tion is necessary in tracing evolutionary series. The converse movement from a lower to a higher plane, as from aquatic to amphibious, from amphibious to terrestrial, from terrestrial to arboreal or aérial, have been slow, difficult and for the most part have occurred but once or twice in the geological history of vertebrate hfe. The higher field once occupied, the lower adaptation was handicapped in its attempts to rise. IMPERFECTION OF THE GEOLOGICAL RECORD Everyone is familiar with Darwin’s classic illustration of the imper- fection of the geological record ;'° but I doubt whether the majority of paleontologists realize how very imperfect our record is, even to-day. We know more about fossil mammals in’ proportion to their modern numbers than about any other of the larger groups of land animals; yet the num- ber of species of which we have any adequate knowledge is but a minute fraction of the number which must have lived since the class first came into existence. Were it not so, the fossil species would vastly outnumber the living forms; as it is, they form a small minority. Moreover, the greater number of recorded fossil species are hardly more than nomina nuda, each known from a single fragmentary jaw, a tooth, a scale, a broken bone, indicating indeed that an animal otherwise unknown lived at a certain time in a certain locality but giving very little information as to its entire structure, its habits, its geographical and geological range. The relationships of these imperfectly known species, provisionally stated by the describers and adopted without the query by subsequent writers. are one of the most fertile sources of error in paleontological theories. Mammals undoubtedly existed during the entire Mesozoic, an era about three times as long as the Cenozoic. Two thirds of their evolution must have taken place during that time; and by the end of it, the principal modern orders were already defined. But we have not a skeleton, or even a skull of a single Mesozoic mammal." Two jaws and a few teeth from the Triassic, a number of more or less fragmentary jaws from the upper Jurassic and various teeth and fragments of jaws from the uppermost Cretaceous represent the sum total of our real knowledge of the first two In the Origin of Species, at the end of Chapter X. U Setting aside Tritylodon as of doubtful affinities. 184 ANNALS NEW YORK ACADEMY OF SCIENCES thirds of the evolutionary history of the Mammalia. The rest is theory and hypothesis. Assuredly, we have no right to assume that the few species which have been founded upon these fossil remains represent at all adequately the number and variety of mammals that lived during the Mesozoic; nor can we even suppose that they fairly represent them. Only two’’ of the numerous phyla of early Tertiary mammals can be at all directly derived from known Mesozoic ancestors. The rest are descended from unknown forms. We may suppose, from the evidence at hand, that the known Jurassic and Cretaceous mammals were arboreal swamp-dwellers and that the chief reason why we know so little of the Mesozoic mammals is that the deposits of the upland regions where they chiefly lived have not been conserved to our day, or at all events have not been recognized and suffi- ciently explored for fossils. In the Tertiary, mammals suddenly spring into (apparent) promi- nence, mainly, it may be assumed, because the fluviatile and eolian forma- tions of the Cenozoic still exist in many localities, although they are being rapidly eroded and carried down to the coastal swamp and sea margin areas of deposition. Epicontinental deposits of Eocene age are rare and scattered, and our knowledge of Eocene mammals is obtained from only a few localities and largely from fragmentary specimens. Through the following Tertiary epochs, these deposits become progressively more ex- tensive and abundant, and our knowledge of fossil mammals is corre- spondingly greater. Finally, in the Quaternary, they form a mantle over most of the earth’s surface, and the fossil mammals are so well known and so many specimens from so many localities have been found that we can get a fairly accurate idea as to the range of many species, not merely as discovered in one or another continent, but as to what parts of that continent they inhabited. ' If our knowledge of fossil mammals is incomplete, that of fossil birds is very much more fragmentary. They probably came into existence at about the same time as mammals, but the early stages of their evolution are even more obscure, and comparison of the living members of the class affords less evidence than with mammals as to their source and course of progress. They are even rarer than mammals in the Mesozoic. ‘Two skeletons and a feather from the Jurassic of Bavaria, a number of skele- tons and fragments from the late Cretaceous of Kansas and a few frag- ments of the skeleton from Cretaceous formations in New Jersey and Europe,—these are all we know of a class which was probably very large 12 Plagiaulacide and Didelphyide. MATTHEW, CLIMATE AND EVOLUTION 185 and varied during the Mesozoic. Our knowledge of Tertiary and Quater- nary birds is much more extensive, but it bears no comparison to our ac- quaintance with Tertiary mammals, and the materials on which it is based are for the most part very fragmentary, their identification often questionable. We may say, however, that Mesozoic birds are more com- pletely known than Mesozoic mammals; that is to say, we know the entire skeleton of two or three, and in consequence can estimate their affinities more certainly and exactly. On the contrary, the fraginentary remains of Cenozoic birds make our estimates of their affinities proportionately uncertain and inexact. The Reptilia are a more ancient class than either birds or mammals and include the ancestral types of both. Our knowledge of fossil reptiles, in comparison with their probable numbers and variety, past and present, is much less than with mammals, more than with birds. We cannot, as with Tertiary mammals, reconstruct approximate evolutionary phyla of the several races from known fossil forms; yet the evidence is sufficient to give a reasonable basis for inferential phyla of some degree of exactitude among many of the Mesozoic and Tertiary reptiles. But the origin of the Reptilia, lke that of the Mammalia, is wrapped in obscurity, and the interrelationship of the more ancient groups is a puzzle not yet solved. We have a fairly extensive acquaintance with the Reptilia of certain habitats at certain epochs; but there were evidently long intervening periods and important faunal facies of which we know nothing or next to nothing. The Amphibia are not a very important group at present and are al- most unknown as fossils, except for the so-called armored amphibians or Stegocephalia, whose relations to the modern frogs, toads and salamanders are still far from clear. This ancient group was abundant and varied in Carboniferous, Permian and Triassic times and is supposed to have given rise to the Reptilia; but the relationship has not been satisfactorily dem- onstrated by fossils, nor is there direct evidence of the interrelationship of the several groups of stegocephalians. A wide gap separates the oldest four-footed vertebrates from any known fishes, living or extinct. ZOOLOGICAL REGIONS, Past AND PRESENT The zoological divisions of the land surface of the earth are given by Lydekker?® as follows: 18 RICHARD LYDEKKER: A Geographical History of Mammals. 1896. This is a modifi- cation of the regions proposed by Sclater in 1858 (Jour. Proc. Linn. Soe., vol. ii, pp. 130- 146) and adopted by Wallace in 1876 (Geographical Distribution of Animals). 186 ANNALS NEW YORK ACADEMY OF SCIENCES 1. Australian region ) | Polynesian fo ; ’ ms \ Notogewic Realm Hawaiian Austro-malayan “ ) i) ia Neotropical “| Neogieic ~~ Malagasy . Ethiopian 7 Oriental cs Arctogeeic Holarctic 2 Sonoran Ce “cc “ oO 9 “ DEN The Polynesian and Hawaiian regions have played no material part in the evolution of mammalian faunas and do not call for any special con- sideration here. The lmits of the remaining regions are shown on the accompanying map. The eight principal “regions” are by no means equally distinct, and their combination into three “realms” does not re- move this defect. Of the five included in Arctogeea, the Sonoran is closest, the Malagasy and Ethiopian farthest removed from the central Holarctic region, if we take into account both the recent and extinct faune. The true relations of the several regions might perhaps be better represented thus: Boreal Subregion Nearctic Palearctic ai Ei eel Sonoran Holarctic Region Oriental Region Subregion | | | Mediterranean Austro-malayan Subregion Subregion Neotropical Ethiopian Region Region Malagasy Australian Subregion Region The Holarctic region in its broader sense, including the Sonoran and Mediterranean subregions, is bounded by the tropic of Cancer, except where (as in Asia) the dispersal of the fauna from a northern center has been hindered by east and west mountain systems, or (as in America) facilitated by north and south mountain systems. The Sonoran sub- MATTHEW, CLIMATE AND EVOLUTION 187 region includes most of the United States and northern Mexico; the corresponding subregion in the western half of the Old World is the Mediterranean, including Europe south of the Alps and Pyrenees, part of southwestern Asia and Africa north of the Sahara desert. The Oriental region corresponds in the eastern part of the Old World to the Mediterranean and Sonoran subregions, but, partly because it in- cludes the great East Indian islands and partly because of the barrier interposed by the Himalayan ranges, it is more clearly differentiated from the Holarctic and may best be regarded as a region of itself. eo NEOTROPICAL HOLARCTIC AUSTRALIAN olarchie 7 Gene Mar cosmopolitan Anlelopes, Elphants Anlelopes, Horses, Herbivoraus and MODERN dorms cay A few High er Fuminants Cattle, D Phi (BPA THR FT C Ma ial Survivors of the Myomorph Frodents 2kle , Deer, Khinaceros Elephants, Dee arnivorous, 7Suplals ulochthohous fauna| “Dogs, Cals, Bears (Fauna of Llivcene and Pleistocene Holarctica) A few Placentals Giant Edenlates.| Man. Modern genera of) Elephants ,Cattle , Frit genera Flerbivorous and PLEISTOCENE |2crauchenta a7 oi fiacental Mammals | PRhinoceroses , 7 Carnevorous Toxadox. Holaretic | Mastodons, Elephants, \ Horses, Camels. of African TATHTNGLS Wr Z Carnivor res 8 Ungulales| Horses , Riinocerases: arsupla Ss, MIOCENE OLIGOCENE EOCENE PALEOCENE CRETACIC Pic. 5.—Characteristic features of the mammal faune in different zoéloyical regions at swecessive epochs of the Cenozoic ORIENTAL ETHIOPIAN Autochthonous Mostly modern genera faur ra dointnant poy sisal Mammals. Earliest invasion of Elephant: & Mastodons 5 J 7 Holarelic fauna FA iota Monkeys Maslodons, E lephanis Fhinocerases, Gira, Beliss Ruminants ed Horses eve Peculiar lypes of \ Modern farilies of Ungulales -E£aentates| Placental Mammals. Marsuptal Carntvores) Mastodors. Monkeys ? Mastodons Rhinoceroses Primitive Fuminanis Masiodons ele. Evolution of Evolution of modern » Pec ue ‘ypes of families of placental Peculiar Cypes of Ungulales . Lnvasio -Ungulates ele. nam mals Primitive Placental\| Modern orders of Ungulates. Placental Mammals arsupial Carnivores| (: Ongulales, Carnivora, Primates Primitive Placental Carniveres and Ongulates Marsupials domimant No Plaeentals posilt vely Arzaw 72. Austromalaya is the debatable ground between the Oriental and the very distinct Australian region; but the consensus of opinion classes it by preference with the Australian. It includes Celebes, the Moluccas, Timor and smaller islands and is separated from the Oriental region by “Wallace’s Line.” The Australian region includes Australia, New Guinea and Tasmania and is the most remote and archaic of all the great (continental) regions of the globe. New Zealand is included in the Polynesian (island) region. The Ethiopian region is connected with the Holarctic by the Mediter- ranean subregion. It is perhaps more distinct than the Oriental, cer- 158 ANNALS NEW YORK ACADEMY OF SCIENCES tainly less so than the Neotropical region. The Malagasy subregion is related not to the modern but to the Tertiary Ethiopian region; its sup- posed Oriental affinities will be considered later. The Neotropical region is connected with the typical Holarctic through the Sonoran, as the Ethiopian is through the Mediterranean interme- diates; but the relationship is more remote. During the Tertiary, the region was much more distinct than it is now. In considering the records of past faune of one or another of these regions as a guide to the dispersal of different groups, it is very necessary to remember that our records are often chiefly or wholly from a small part of the region, often far from typical. Our knowledge of Palearctic faunze in the early Tertiary is wholly from western Europe, an outlying, marginal part, more or less submerged and archipelagic. Its relations to the main body of Palearctic land life were probably much like those of the East Indian archipelago to the continental portion of the Oriental region. In the later Tertiary and Quaternary, we obtain a broader outlook on the Palearctic fauna, but even then it is incomplete. In the Oriental region, we know nothing of the land life of the early Tertiary, and in the later Tertiary we know only the life of its northern borders, close to the Palearctic region and doubtless more nearly approxi- mating the Palearctic fauna then than now, as the Himalayan barrier was less complete. The result of these two facts will apparently be that the early Tertiary Palearctic fauna will appear by the record to be less progressive than it really was and that the Tertiary Oriental fauna will appear to be more progressive than it really was. In the Nearctic Tertiary, the record is chiefly confined to the Western plains; we know little of the Canadian Nearctic—presumably more progressive. In the forested regions of the Kast and South, where we might expect to find primitive survivals, or on the Pacific coast, where we might expect to see stronger Palearctic influ- ence, our knowledge is very imperfect, although the few available data are in conformity with a priori deductions. In the Neotropical region, our chief dependence is upon the Argentine faune which should be both the most progressive and least influenced by Northern immigration. In the Ethiopian region, we have but a single glimpse of the Tertiary land fauna, and that is derived from Egypt, where we might expect to find a transitional fauna, combining true Ethiopian autochthones with immigrants from Palearctic or northwestern Oriental faune. But, since the water barriers to the north of Egypt were more extensive and the MATTHEW, CLIMATE AND EVOLUTION 189 desert barrier to the south less developed in the early Tertiary than they are to-day, we should expect that the autochthonic element would be dominant and that Tertiary Egypt belonged to the Ethiopian zodlogical region, although modern Egypt does not. These may serve as instances of the caution with which the geological record must be used in attempting to estimate the position and source of regional faunz. The regions here adopted are based primarily upon the present and past distribution of mammals. Birds, reptiles, amphibians, fresh-water fishes and the various groups of terrestrial invertebrates are not wholly in accord with this arrangement so far as their present distribution is concerned. This is partly because the means and limitations of their dispersal differ, chiefly, as I shall attempt to show, because so little is known of their former distribution. ForMER BARRIERS AND BRIDGES The general principle of dispersal on the lnes of the present continents is open to an obvious objection. The outlines and connections of the continents were different in former times. The relations of land and water were not the same. In fact, if one depends upon a text-book knowl- edge of geology he may find authority for an assured behef that they were fundamentally and altogether different in different geologic periods. It is necessary therefore to point out that the stratigraphic no less than the life record is a defective one, and that the really proven changes in the distribution of land and water are limited to those summarized on page 175. The geotectonic hypotheses so ably and brilliantly elaborated by Suess,’* Haug’® and other writers, are not facts but theories, and I must confess to a decidedly skeptical attitude towards some of their con- clusions. There are too many gaps in the chain of their arguments; too many known facts with which their conclusions appear to be inconsistent. The permanency of the continental platforms is indicated by the ab- sence of abyssal deposits in their sedimentary succession wherever this has been adequately studied. The platforms have been extensively over- flowed by shallow seas, but such submergences were temporary, and inter- vening periods of uplift are indicated by gaps in the marine succession. Where the geologic records are fragmentary, widely scattered and imper- fectly correlated, there often is a tendency to exaggerate the extent and permanency of such overflows, as also to assume extensive unknown con- tinents to account for the existence of clastic sediments which were more 44E. Suess: Antlitz der Erde. 1888-1901. 6H. Hauc: Traité de Geologie. 1912. 190 ANNALS NEW YORK ACADEMY OF SCIENCES probably derived from unsubmerged adjoining portions of the existing continents. We are apt to assume that great displacements of strata in- volve correspondingly great changes of level. They do not necessarily ; more probably, in most instances, the erosion has kept pace more or less closely with the displacement. Even where great changes of level have occurred, they often have been, and more often may have been, of re- stricted extent and compensated by opposite changes in regions imme- diately adjoining, and most of them have had but little extensive or permanent effect on the general configurations and relations of the conti- nental platforms. The relative permanency of the North American continent is very clearly brought forward in Schuchert’s maps.'® “Yet even here, if one may venture a criticism on so thorough and conservative a study, there is a certain loss of conservatism where the outlines run into territory where the evidence is inadequate, as in the Antilles and the Arctic seas. The imperfect data available for the South American continent appear to in- dicate general conditions very similar to those of its northern neighbor ; nor does it appear that Africa and Australia were any less permanent land platforms. Northern Eurasia appears to have been similarly perma- nent, but across Central Europe and extending southeastwardly to the East Indies lies a broad strip of disturbance where great changes have occurred during later geologic time. But the extent and permanency of the great central sea which is so frequently depicted as interposing a broad ocean between the Holarctic and the Ethiopian and Oriental land masses is by no means certain, especially as regards its eastward exten- sion. I cannot find in the recorded facts proof that it afforded any more continuously effective bar to dispersal along the lines of the present con- tinental relations than did the middle Cretaceous overflow in North America or the early Tertiary one in South America. Perhaps the most widely accepted departure from the permanency of the ocean basins is the supposed Gondwana Land, invented to account for certain similarities in southern Paleozoic floras, and since used to account for almost all cases of similarity among southern flore and faune which were not demonstrably due to dispersal from the northern conti- nent. This theory has in its original form gone so long uncontested that it is very generally regarded as incontestable. New discoveries have been interpreted in terms of it, the weakness of the original evidence, the pos- sibility that it might be otherwise interpreted, has been forgotten, and like the Nebular Hypothesis, it has become almost impossible to dislodge it from its place in the affections of the average geologist. 16 CHARLES SCHUCHERT: Bull. Geol. Soc. Amer., vol. 20, pp. 427-606, pll. xlvi-ci. 1910. MATTHEW, CLIMATE AND EVOLUTION 191 If the distribution of animals be interpreted along the lines here advo- cated, there is no occasion for a Gondwana Land even in the Paleozoic. But it is chiefly as affecting Mesozoic or Cenozoic dispersal that we are here concerned with it. One may summarize the arguments for it by saying that a considerable number of groups of animals and plants which are absent in the northern world, either living or fossil, are found in the southern continents and some of them in certain oceanic islands as well. Most of the groups are unknown or almost unknown as fossils; those which have any considerable fossil record are steadily being eliminated from the list by the progress of discovery, showing that they or their an- cestors did formerly inhabit the northern world. The remaining groups agree with those southern faunal groups which have admittedly come from the north, in being of primitive and archaic type and in that their representatives in the different southern regions are but distantly related, the remoteness being in a very direct proportion to the present isolation of the region. There are a few instances of exclusively southern types closely related (e. g., Galaxias) ; but, although they have been cited in corroboration of the evidence from the groups above mentioned, they are in fact, if thus interpreted, directly contradictory. For the distant relations of the one series is interpreted to mean a very ancient connection, but isolation since; while the other series would indicate a very recent connection and earlier isolation. The explanation here lies not in a northern ancestry, but that the ocean does not form an impassable barrier to their dispersal. This has been proven in the case of Galazias; it is probably the explana- tion of all similar distributions. The relations of the Glossopteris flora are a different and far more com- plex problem of distribution. The clue to its interpretation lies perhaps in its association with Permian glaciation; but it is outside the limits of the present essay and will not be discussed here. REGIONAL CORRELATION The geological correlation of widely distant formations is so intimately bound up with problems of geographical dispersal and migration that the two series of problems must needs be studied and solved together. We cannot arrive at a correct understanding of the history and causes of the geographical distribution of animals, present and past, without correct correlation of the geological succession in different regions. Nor have we, up to the present time, any reliable methods of exact correlation in widely distant regions except the comparison of fauna and a considera- 192 ANNALS NEW YORK ACADEMY OF SCIENCES tion of their source and the history of their migration and dispersal. Absolute standards, as of world-wide changes in physical or climatic con- ditions, may serve in the future to give us broad lines of correlation inde- pendent of paleontology ; but at present their universality is hypothetical, the exact train of physical phenomena which they entail and the indices by which they may be recognized in the stratigraphic succession are im- perfectly known. Paleontology is for the present our sole recourse in correlation. Probably it will always be our chief dependence, at least in exact and detailed comparison. SYNCHRONISM AND HOMOTAXIS The ordinary methods of paleontologic correlation can be applied with accuracy and certainty only over limited areas of the earth’s surface. When applied to far-distant regions, we meet first with the difficulty that there is little identity of faune, only an equivalence more or less exact. Nor can we be sure that equivalent or even identical forms were contem- poraneous in all parts of the earth. They certainly are not so to-day. The modern land fauna of Australia, as Huxley long ago insisted, is in its broad lines a Mesozoic fauna. Examined in detail, it shows indeed the marks of a long period of independent evolution and specialization. Yet the degree and amount of specialization. is far less than that which the faune of the northern continents have undergone during the Ceno- zoic. The modern fauna of the East Indies or of Central Africa has a great deal in common with the later Tertiary faunz of Europe and north- ern Asia. Central America and tropical South America bear similar relations to North America. While Huxley’s dictum that an older fauna in one region may be homotaxial with a later fauna in another does not apply to the extent of involving identity of all or most of the species, yet it very clearly does apply in a broad way to the land faunz and probably to a less extent to the marine faunz as well. The rate at which evolution and differentiation progress varies as between the faunz of different re- gions. It varies as between the different constituents of a fauna. Neither the partial identity nor the general equivalence of two faune is sufficient to prove them synchronous, except under certain conditions to be con- sidered later. Another method very generally used in correlation of faunz which contain little or nothing in common consists in an estimate of their rela- tive antiquity as indicated by the proportion of extinct to surviving spe- cies or genera. This also involves the assumption that the rate of progress wT. H. Houxupy: Q. I. G. S., vol. xviii, pp. xl-liv. 1862. MATTHEW, CLIMATE AND EVOLUTION 193 of evolutionary change is constant in all parts of the earth, at least for members of the same group. But if the rate varies in different regions for the fauna as a whole, we have no reason to believe that it would be constant for common or similar groups. The practical application of this method is very unsatisfactory. In illustration of this, I may instance the widely divergent views entertained by different authorities as to the age of the later geological formations of Argentina in comparison with European standards. Able and authori- tative discussions of this problem have appeared within the last few years by Ameghino,’* Roth,"® Gaudry,?° Scott,?* Hatcher,” Ortmann,”* Stan- ton, von Ihering, Wilckens, Cossmann, Wiman and others, dealing with the vertebrate and invertebrate fossils and stratigraphic relations of the formations. The field work has been extensive, the collections large, the faune are large and varied and in large part well known; but the results are widely discordant. The amount of discordance is indicated by the correlation of the four principal terrestrial formations, as given by Ameghino, Roth, Gaudry and Schlosser. The correlation of widely distant formations is so intimately bound up with problems of geographic distribution and migration that the two series of problems must be studied and solved together. The methods relied upon by Roth and Ameghino are substantially the same as those generally used by northern authors. Why then do they lead to such dis- cordant results? It is because the data on which they rest prove not con- temporaneity but homotaxis. Granting that two faune in widely remote regions contain the same proportion of extinct species, granting that they represent equivalent stages of evolutionary progress, they are not thereby shown to be contemporaneous, unless they are at the same dis- tance (measured not in miles but in difficulty of advance) from the main center of dispersal of the fauna which they contain. Very obviously, if iW, AMEGHINO: “L’Age des Formations Sedimentaires de Patagonie,’’ Anal. Soc. Cient. Argent., tom. L, LIV; pp. 1-281 of separata. 1903. ‘“‘f’ormations Sedimentaires du Cretacé Superieur et du Tertiaire de Patagonie,”’ Anal. Mus. Nac. Buenos Aires, tom. Xv, pp. 1-568. 1907. 19 SANTIAGO RoTH: “Beitrag zur Gliederung der Sedimentablagerungen in Vatagonien und der Pampasregion.’’ Neues Jahrb., Beil.-Bd. xxvi, s. 92-150, taf, xi-xvii. 1908. 20 A, GaupDRY: ‘‘Fossiles de Patagonie, etc.’”” Ann. de Paléont. I. 1906. 2 W. B. Scorr: Mammalia of the Santa Cruz Beds in Rep. Prine. Univ. Exp. Pata- gonia, vol. v. 1903. Int. Cong. Zool., Berne, C.-R., pp. 241-247. 1905. A History of the Land Mammals of the Western Hemisphere. 1913. 27. B. Harcuer: “On the Geology of Southern Patagonia.’ Amer. Jour. Sci., vol. iv, pp. 827-354. 1897. ‘Sedimentary Rocks of Southern Patagonia,” ibid., vol. ix, pp- 89-108. Jbid., vol. xv, pp. 483-486. 1903. 23 A. ORTMANN: Tertiary Invertebrates in Report Prine. Univ. Exp. Patagonia, vol. iv, pp. 45-332, pll. xi-xxxix. 1902. See for further references the bibliography in Ameghino, 1907, supra, pp. 3-18. 194 ANNALS NEW YORK ACADEMY OF SCIENCES TasBLeE I.—Correlation of the Four Principal Terrestrial Formations Ameghino a || x 1906 4 Roth, 1908 Gaudry, 1906 | Schlosser, 1912 Pleistocene ) Pampean Pampean I | > Pampean (= Pliocene | J > Pampean ____________—|- : | Santa Cruz Miocene Santa Cruz : eee J Santa Cr | Pyrotherium Oligocene Santa Cruz | il Eocene Santa Cruz | Pyrotherium | Pyrotherium | Notostylops Paleocene Notostylops Upper ee ne Cretaceous Pyrotherium | Notostylops ee Notostylops 4 c i] the principal center of dispersal of Mammalia was in the Holarctic re- gion, the fossil mammals in southern regions invaded by that northern fauna will appear in their homotaxial relations to be more ancient than they really are. The modern fauna of South America, of Africa, of the Oriental regions, will be in the same stage of evolution as the late Ter- tiary and Quaternary faune of Holarctica. Its species will be more nearly related or equivalent to Pliocene and Pleistocene species of Europe and North America than to their modern fauna. The late Tertiary mammals of the southern continents will approximate in homotaxis the middle or early Tertiary mammals of Holarctica; and the middle Ter- tiary southern faune will approximate the early Tertiary or late Cre- taceous faune of the north. _ On the other hand, if we believe, as does Dr. Ameghino, that the prin- cipal theater of evolution of the mammals lay in the temperate regions of South America, and that the mammal population of the North was derived by migration from that center (by way of Africa across a tropical land bridge not now existing), it will be equally obvious that the southern formations will be more ancient than their homotaxis, impartially con- sidered, would lead us to believe. The result will be to assign to the MATTHEW, CLIMATE AND EVOLUTION 195 Cretaceous period those southern faunze which are homotaxial with the early Eocene of the North; to the Kocene those faune which are homo- taxial with the Middle Tertiary of the North, and so on. To a certain extent, the intercalation of marine formations may pro- vide a check on this relationship, but it must be remembered that the same theories of dispersal may also apply to marine faunz, wholly or in part. Homotaxial marine faune may be far from contemporaneous. The chief center of dispersal of marine faunze may be assumed to be either the equatorial oceans and coasts, the northern, or the southern seas, or both north and south equally. Only when the movements of dis- persal are in opposite directions on land and in the seas will the marine faunz furnish an adequate check on the homotaxis of land faunze; and in that case the true synchronism must be arrived at by balancing con- flicting evidence derived from terrestrial and marine faunal comparisons. It is true that if we eliminate the idea of faunal dispersal altogether and regard each race of animals as evolving and dispersing independ- ently, governed by its own conditions and causes of change, we may in the present imperfect state of our knowledge lay out various and inde- pendent centers of dispersal for different races, whose successive appear- ance in one or another continent will furnish data for a true correlation. There has been a strong tendency in the last half century to work on this theory, but in the present writer’s opinion at least, the supposed evidence in favor of this view is due chiefly to the imperfection of the geologic record, and its very wide acceptance to a lack of appreciation of the underlying causes of evolutionary progress and dispersal. I do not understand how anyone can reconcile the theory that each race of animals evolves and disperses independently and that the common biotic and physical environment is not a controlling factor, with the plain fact that regional faunz do exist to-day. The conditions that control the dispersal of one race are largely identical or correlated with those that control the dispersal of others, and every change in these conditions will affect not one race only, but a large part or the whole of a fauna, in a manner and to a degree largely identical, causing similar changes in the range of the fauna. ‘ TERTIARY CORRELATION IN SOUTH AMERICA Before setting forth the evidence as to the dispersal of the mammals, it is necessary to attack a problem which has caused much acrid contro- versy, namely, the age of the later formations of the Argentine Republic. The difference of opinion among authorities has already been indicated, 196 ANNALS NEW YORK ACADEMY OF SCIENCES as also the fact that the true correlation is so intimately related to the direction of migration that the two problems must be settled together. In view of the great and well merited reputation of Dr. Ameghino and the immense array of data which he has marshalled in support of his theories of correlation and phylogeny, it is not surprising that they should find a very considerable acceptance, not in South America alone but else- where. Few scientists indeed are disposed to accept his derivation of the horse family from carly South American ancestors or of the various families of Carnivora from the same source, for in these and other cases the evidence for northern ancestry is almost universally accepted as con- vineing; but many writers are willing to accept Ameghino’s determina- tion of the age of the Argentine formations, although more critical as to his phylogenetic views. The two, however, must stand or fall together; and it is precisely be- cause the Equidwe, Procyonide, ete., if their generally accepted phylog- enies be admitted, afford incontrovertible evidence against the validity of Ameghino’s correlations of the formations of the Argentine, that he has been compelled to devise different phylogenies for these cases. Few scientists will be willing to believe Ameghino’s assertion that Merychippus and its successors in the equine phylum have nothing to do with the Anchitheriine which they so closely resemble in teeth, in skull, in feet, in all details of the skeleton, but must be derived from the South Amer- ican Notohippide on the strength of a much more distant resemblance in the second upper molar, unsupported by any near resemblance whatsoever in the remaining teeth or in any points of construction of skull or of skeleton. It is not my intention to present here any detailed refutation of Dr. Ameghino’s argument, but to point out that if the northern origin of the Equide be accepted, the age of the Pampean and related forma- tions must be far later than that he has assigned to them. The first ap- pearance of true equines in South America is in the Pampean. The three best-known genera are Lquus, [ippidion and Onohippidion. The first might be regarded as of Palearctic origin; the second and third have no Old World predecessors, but may be directly derived from the North American Pliohippus. They are, however, much larger and more pro- gressive than Pliohippus, and. in size, reduction of the lateral digits, etc., are equivalent to Hyuus. We can hardly doubt that they came to South America from North America, nor can I see any practical alternative to believing that Hquus arrived by the same route. Now, the first appear- ance of Hquus in North America is at the base of the Pleistocene. In Argentina, it first appears in the middle Pampean. The middle Pam- pean cannot therefore be older and is presumably younger than Lower MATTHEW, CLIMATE AND EVOLUTION 197 Pleistocene. Hippidion and Onohippidion are found (fide Roth) in somewhat older levels; but as they are much advanced over anything in our Middle Pliocene (Blanco), it would seem that their first occurrence in the Pampean must be placed at the top of the Pliocene or preferably in the lower Pleistocene. I conclude that the Pampean formation ap- proximately represents the Pleistocene epoch. Beneath the Pampean of Ameghino, but included in it by Roth, are fossiliferous beds in which certain Procyonide and Urside are found. If we admit the North American source of these carnivora, they would indicate Pliocene age for the beds containing them. Dr. Ameghino, who regards them as Oligocene and Miocene, is compelled, therefore, to set aside the North American ancestors of the Procyonide and to regard them as of South American origin and the Urside as either autoch- thonous or arriving in South America from the Old World via Africa. As with the Equid, the only shadow of plausibility for such phylogenies lies in the incompleteness and careful limitation of the evidence that is adduced in their behalf. Phlaocyon of the North American Miocene, which is intermediate between Cynodictis and the Procyonide in almost every detail of the perfectly preserved dentition, skull and skeleton is merely** “un vrai Canide sans relations avec les Procyonidés,” while the South American genera are derived through hypothetical ancestors from the carnivorous marsupials of the Santa Cruz. Here again, Dr. Ameghino is compelled, in defense of his theories of correlation, to adopt these im- possible phylogenies, because if the Procyonide are of North American origin the Argentine formations are demonstrably of later date than those which he assigns to them. Phlaocyon is a far more primitive procyonid than any of the South American genera. Leptarctus of the Upper Miocene may be their equivalent, but it is very imperfectly known.” If these Argentine genera are derived from the Oligocene Cynodictis and related genera of Holarctica, Phlaocyon being about half way between the two groups, then their age is indicated as Pliocene, not as Oligocene or Miocene. Also with the Urside; to admit them as arriy- *4 PL, AMEGHINO: Ann. Mus. Nac. Buenos Aires, tom. xv, p. 396. 1906. Dr. von Ihering has since attempted to prove what Ameghino merely asserted. His argument rests upon an untenable interpretation of a single feature in the dentition, ignoring all other char- acters of teeth, skull and skeleton, and, if true, would involve not only that Bassariscus has nothing to do with the Procyonide (which he asserts), but also that the Procyonide have nothing to do with the carnivora but are of wholly diverse ancestry. See H. vy. IHERING, Systematik, Verbreitung und Geschichte der sudamerikanischen Raiibthiere. Archiv f. Naturg., 76 Jahrg. I. Bd., s. 113-179. 1910. °>'The type of Leptarctus is an upper premolar of doubtful affinities. Wortman re- ferred to it in 1894 a lower jaw from the Upper Miocene, which is unquestionably procyonid and hardly distinguishable from Procyon. Ameghino and von Ihering ignore this record. 198 ANNALS NEW YORK ACADEMY OF SCIENCES ing via North America would compel Ameghino to conclude that their first occurrence in South America in these same sub-Pampean beds must be materially later than the evolution of the phylum in the Palearctic region (Miocene) and that the genus Arctotherium of the true Pampean in South America, unknown in North America until the Pleistocene, indicates, like Hquus, that the Pampean is a Pleistocene formation. The distribution of Smilodon in North and South America is in exact accord with that of Arctotherwm. The relations of the South American Proboscidea to those of North America correspond to those of the Equide. The Camelide, Cervide, Canide, ete., also support the Pleistocene age of the true Pampean. The Edentata, whose migration appears to have been in the reverse direction, will be discussed later. In the Santa Cruz fauna, we have not the direct evidence that the Pampean faune afford for correlation by means of groups of admittedly northern origin. The evidence has been very fully discussed by Hatcher, Ortmann, Scott and others, and so far as it is based upon the relations and age of associated marine formations, I am not competent to criticize it. The criterion used by Ameghino and Roth, of proportions of extinct to living genera, I regard as untrustworthy, partly for the genera] reasons already given (p. 192) and partly because of the personal equation that must always affect the number of genera and species described as new, as compared with those referred to known genera and species. Unless the standards of diversity for genera and species were approximately the same, and in this instance they are certainly very wide apart,”° the com- parison of the proportions of extinct to surviving genera and species in Argentine formations with those of Europe or North America would be misleading. Perhaps the most important correlation is that of the Notostylops fauna, Lower Cretaceous according to Ameghino, Upper Cretaceous ac- cording to Roth, Paleocene according to Gaudry, Upper Eocene in Schlos- ser’s view. Here there is an apparently strong point for Cretaceous age in the presence of dinosaurs in association with the fossil mam- mals. Dinosaurs disappeared from the Northern world at the end of the Cretaceous." They are entirely unknown in any Tertiary formation. Nevertheless, the possibility of their survival into the early Tertiary in South America must be considered.?8 The mammalian fauna with which °6The European fossil rodents are, for the most part, referred in accordance with the old conservative standards of genera and species, while Ameghino is much inclined to hairsplitting in generic and specific distinctions. Scott in his revision is more conserva- tive, but not so as to equalize the standards in question. *7 The latest dinosaur formations of North America are, however, regarded as Paleo- cene by Knowlton, Lee, Peale and other authorities. * The same arguments apply to the occurrence of a Mesozoic type of Crocodile, Votosuchus, in the Notostylops fauna. MATTHEW, CLIMATE AND EVOLUTION 199 they are associated is in part closely related to the Paleocene fauna of Europe and North America and for this reason has been regarded as equivalent. But these genera of Northern affinities are associated with a large number of larger and more progressive genera, structurally de- rivable, according to the canons of evolutionary development universally accepted by paleontologists, from the more primitive types which are common to the Notostylops beds and the Paleocene of the North, and leading apparently into the various specialized groups peculiar to the later South American Tertiaries. These more progressive types are un- known to any northern Tertiary fauna; they appear to be derived from the more primitive group whose affinities are so close to the Puerco, Torrejon and Cernaysian mammals; and they point to the conclusion that the Notostylops fauna is in reality decidedly later than the Paleo- cene, the more primitive group of its fauna being little altered survivals,”® corresponding to the primitive survivals (Condylarthra, ete.) which are found in the Wasatch and Wind River faunz of North America. Taking the Notostylops fauna as a whole, it appears to me to represent an Hocene stage of development, conditioned by an isolation which began in the Paleocene and hence prevented the incoming of any Perissodactyla, Artiodactyla or Carnivora from North America.*° This same isolation will satisfactorily account for a later survival of the dinosaurs, of Meso- zoic Crocodilia and some other primitive elements, if they were in fact contemporary with the Notostylops fauna. The age of the Pyrothertwm beds is much less definitely determinable. Dr. Roth, indeed, doubts the existence of this fauna as distinct. If accepted, it would presumably be intermediate between the Notostylops and Santa Cruz faune and provisionally referable to the Oligocene. The sequence of the Argentine faune will then be Pampean (s.s.) — Pleistocene Monte Hermoso etc. — Pliocene Santa Cruz — Miocene Pyrotherium — ? Oligocene Notostylops — Eocene. So far as the correlation of the Pampean and Santa Cruz is concerned, their fossils agree wholly in preservation and degree of petrifaction with those preserved in similar Pleistocene and late Miocene formations, re- °® Little altered, that is to say, so far as the parts known to us are concerned; their adaptation, whatever it was, not involving radical changes in dentition from the primary type. *® Schlosser (in Zittel’s Grundziige d. Pal.. Rey. Ed. 1912) regards the Notostylops fauna as Upper Eocene. Scott (History of Mammals of West. Hem.) places it as Eocene. 2()() ANNALS NEW YORK ACADEMY OF SCIENCES spectively, in the western Plains, and the degree of consolidation of the matrix is the same. We have in the West two fossiliferous formations, the Bridger (Eocene) and John Day (Oligocene), which are, like the Santa Cruz, composed of an andesitic volcanic ash, and similar ash strata are found in different levels of our Western Miocene formations. Now, the Santa Cruz matrix and fossils are very much less consolidated or thoroughly petrified than the Bridger and“decidedly less so than the John Day, while they agree very well with the volcanic ash beds in the middle and upper Miocene. As there is no reason to suppose that the rock-making processes work at a different rate in different continents, this evidence is entitled to some consideration. On similar grounds, the Pampean fossils would be referred to middle Pleistocene, and the few fossils that I have seen from Monte Hermoso agree best with Pliocene fossil mammals from North America. I should place no weight on this kind of evidence except when, as in the present instance, the climatic conditions and the origin and method of deposition of the formations are substantially similar. The foregoing digression is somewhat outside the limits of this dis- cussion. It appears, however, to be necessary to show briefly the reasons on which the age assigned to the South American mammalian faunz are based. It might, indeed, be logically objected that these correlations are based on the northern origin and migration of certain phyla and cannot, therefore, be used in support of the theories here advocated. But the phyla on which the demonstration rests are so universally admitted to have arisen in the north, and the evidence that they did so is so com- plete and conclusive, that there is no reasonable alternate to accepting them as such. And if so, the correlations of South American faune must be approximately as here stated, a conclusion supported by the wholly independent evidence of the degree of consolidation of the forma- tion and of petrifaction of the fossils contained. CENTERS OF DISPERSAL Whether the evolution of a race be regarded as conditioned wholly by the external environment or as partly or chiefly dependent upon (un- known) intrinsic factors, it is admitted by everyone that it did not appear and progress simultaneously and @quo pede over the whole sur- face of the earth, or even over the whole area of a great continent. The successive steps in the progress must appear first in some comparatively limited region, and from that region the new forms must spread out, displacing the old and driving them before them into more distant MATTHEW, CLIMATE AND EVOLUTION 2()1 regions. Whatever be the causes of evolution, we must expect them to act with maximum force in some one region; and so Jong as the evolution is progressing steadily in one direction, we should expect them to con- tinue to act with maximum force in that region. This point then will be the center of dispersal of the race. At any given period, the most advanced and progressive species of the race will be those inhabiting that region; the most primitive and unprogressive species will be those remote from this center. The remoteness is, of course, not a matter of geographic distance but of inaccessibility to invasion, conditioned by the habitat and facilities for migration and dispersal. If the environmental conditions in the center of dispersal pass the point of maximum advantage for the race-type that 1s being developed and become unfavorable to its progress, we should find its highest types arranged in a circle around a central region, which was the former point of dispersal, and the more primitive types arranged in concentric ex- ternal circles. The central region will be unoccupied, or inhabited by specialized but not higher adaptations. It would appear obvious that the present geographic distribution of a race must be interpreted in some such way as this by anyone who accepts the modern doctrine of evolution. Yet there are many high authorities on geographic distribution who proceed apparently upon a precisely op- posite theory. According to these authors, the distribution center of a race is determined by the habitat of its most primitive species, and the highest and most specialized members of the race are most remote from its center of dispersal. This principle may be true enough so far as concerns the first appearance of a given race, 1. e., provided the most primitive species are also the oldest geologically; but it appears to me to be the direct reverse of fact as regards the present distribution, or the distribution at any one epoch of the past. The only ground on which it could be defended would be that the progress of the race is due to its migration, and those members which did not migrate did not progress. But this involves the view that its progressiveness up to the time that its geographical environment changed was due to staying at home, and the same progress after its environment changed was due to not staying at home. It seems to me that the prevalence of this view must be due to some fallacious notions about migration, unconsciously retained, in- volving a concept of it as analogous to travel in the individual. The successful business man, no doubt, may pack up his baggage and take to traveling, leaving home and going elsewhere and profiting much thereby. Nations have done the same thing, likewise to their advantage. But there is very little analogy here to the zodgeographic migration of spe- 202 ANNALS NEW YORK ACADEMY OF SCIENCES cles rectly of transference of habitat, although this may be the final result. It seems obvious that the conditions which brought about the early progressiveness of the race in a particular locality would, so far as they which is a question of expansion or contraction of range, not di- were external, cause the continued progressiveness of those individuals which remained in that region; so far as they were intrinsic, they would affect the main bulk of the race, the center of its range, more than any outlying parts of it. The present writer is very thoroughly convinced that the whole of evolutionary progress may be interpreted as a response to external stimuli; and intends here to point out what he regards as the most important of these stimuli. It is therefore necessary to point out that these postulates regarding centers of dispersal and migration are not dependent upon the theories to be proved—we are not reasoning in a circle. OCEANIC AND CoNTINENTAL ISLANDS FAUNAL DIFFERENCES BETWEEN OCEANIC AND CONTINENTAL ISLANDS One of the strongest arguments for the relative permanency of the deep oceans, especially during Cenozoic time, is afforded by the marked and striking contrast between the faune of those large islands which are, and those which are not, included within the continental shelf. The continental islands have the fauna of the continents to which they belong, large as well as small, differing only in the absence of types of recent evolution or of unsuitable adaptation and in the survival of primitive types which have disappeared from the mainland. But no question could be raised as to their former union with the mainland, no other possible solution would explain their fauna. We are compelled to assume the former connection of the British Isles with Europe, of Ceylon with India, of Japan with Korea or Siberia, of Sumatra, Java and Borneo with the Malayan mainland, of the Philippines with Borneo, of New Guinea and Tasmania with Australia, of Newfoundland and Cape Breton with Lab- rador and Nova Scotia. In each and all of these cases, the evidence is overwhelming, and, with the exceptions cited, the faunal identity is complete. . On the other hand, with all those islands which are separated by deep ocean from the mainland, we find that just that evidence is lacking which would afford convincing proof of former union with the mainland. Their faune are widely different from those of the adjoining mainland; they lack just those animals which could not possibly have reached there except by land bridges; they point often to long periods of independent evolution and expansion, and the primary elements of the faune of every MATTHEW, CLIMATE AND EVOLUTION 203 one of them are such as might possibly at least have reached the island without continental union, whether by accidental transportation, by swimming or by other means. Take for example the mammals of Sumatra, Java and Borneo. We cannot reasonably suppose that the rhinoceroses, tapirs, deer, wild dogs, felids and numerous other large animals common to them and the ad- joining continents reached these islands except by land. They are too large for transportation on “rafts” of vegetation such as occasionally drift to sea from the mouths of tropical rivers. They are dry-land ani- mals not given to swimming long distances. And we would not invoke the agency of man to account for a whole fauna. But most important is the fact that all the animals that we might fairly expect to find there in view of a former land connection are really present. Contrast with this the fauna of Madagascar.* There are no ungulate mammals there, except for the bush-pig, possibly introduced by man (in accord with known customs of the Malays) and a pigmy hippopotamus (now extinct) which might have reached the island by swimming, as hippopotami are known to travel considerable distances by sea from one river mouth to another. The great majority of the unguiculate groups of the mainland are also absent. The only representatives are a few very peculiar carnivores of the family Viverride, a peculiar group of insectivores (Centetide) and a peculiar group of Cricetine rodents, each apparently evolved on the island from a single type introduced Jong ago, a species of shrew (Crocidura) of more recent introduction and a variety of bats. There are numerous lemurs and no monkeys there; and the lemurs appear to have radiated out from a single group*? into a number of peculiar types, two of which, now extinct, paralleled the ungulates and the higher apes in several significant features. ‘The fauna of the island does not resemble the present fauna of Africa, nor can it be derived from any one past fauna, known or inferential, of that conti- nent. The attempt to derive it from the present or from any known or inferential past fauna of India involves still greater difficulties. On the contrary, the Malagasy mammals point to a number of colonizations of the island by single species of animals at different times and by several methods. Of these colonizations, the Centetide are the earliest, perhaps pre-Tertiary ; the lemurs, rodents and viverrines are derivable from one or more middle Tertiary colonizations: and in both cases the “raft” 31A,. R. Watuace: Island Life, pp. 381-412. 1881. See also Trouessart Catalogus Mammalium and Suppl. Quing.; Lydekker, Geog. Hit. Mam., pp. 211-226. 1896. Lydek- ker’s arguments for continental union are mostly invalidated by more recent discoveries. 32 See W. K. Gregory's studies upon the affinities of the Lemuroidea, forthcoming in Amer. Mus. Bulletin. 204 ANNALS NEW YORK ACADEMY OF SCIENCES hypothesis may reasonably be invoked.** The hippopotami may have arrived by swimming and the bush-pig and the shrew may have been introduced by man, while the bats may readily have arrived by flight. The extinct ground birds are easily derived from flying birds. Dr. Arldt,** in his discussion of the Malagasy fauna, poimts out its composite character, derived from severak successive invasions. This, I think, is clear enough; but it seems equally clear that these were not faunal invasions due to land connection but sporadic colonizations by a few species all at different times. The characters of the mammalian fauna, both negative and positive, practically exclude the theory of land connections during the Tertiary. The West Indian islands afford another marked instance. In spite of its nearness to Florida, there are no North American mammals in Cuba, except the manatee,—analogous with the hippopotamus in Mada- gascar. Nor are the other islands richer in fauna. As also in Mada- gascar, we have a peculiar and very primitive insectivore Solenodon (Cuba and Hayti), a number of peculiar extinct ground-sloths, of which Megalocnus is the best known, and which although Pleistocene in age are derivable not from the Pliocene or Pleistocene ground-sloths of North or South America but from the Miocene ground-sloths of Patagonia, and evidently differentiated through a long-continued period of isolated evolution, and a couple of chinchillas—the hutias of the larger islands, the (extinct) Amblyrhiza in Anguilla. The Solenodon may be referred to a more ancient colonization, the ground-sloths probably arrived during the Miocene, the chinchillas more recently; and the direction of the prevalent ocean currents points out the reason why these are of South American derivation. Those who, hke Dr. J. W. Spencer,*® believe in gigantic elevation movements connecting the Antilles with the mainland in Pliocene and Pleistocene would account for the absence of the conti- nental fauna by invoking a subsequent subsidence which drowned out everything else. The improbabilities involved in this hypothesis on strati- graphic and faunal grounds have been pointed out by W. H. Dall, R. 'T. Hill’® and others. *3'The moist tropical conditions of early Tertiary times would fayor the formation of such rafts, the small size and arboreal habits of the animals concerned would increase the chances of their being caught on such rafts and the uniform climate and conse- quently more placid seas would increase the distance over which the raft might be trans- ported before it broke up. *' THEODORE ARLDT: Entwicklung der Kontinente und ihrer Lebewelt, pp. 119-142. 1907. ° J. W. Spencer: “Reconstruction of the Antillean Continent,” Bull. Geol. Soc. Amer., vol. vi, pp. 103-140. 1895. 38 W. H. DALL: “Geological Results of the Study of the Tertiary Fauna of Florida,” Trans. Wagn. Inst., vol. iii, pt. vi. 1903. R. T. Witu: “Geological History of the Isthmus of Panama and Portions of Costa Rica,” Bull. Mus. Comp. Zo@l., vol. xxviii, pp. 151-285. 1898. MATTHEW, CLIMATE AND EVOLUTION 205 Cuba, while near in actual distance to the North American continent, has been comparatively inaccessible to sporadic colonization from that source, on account of the direction of the ocean currents; but coloniza- tions from South (or possibly Central) America have reached it. New Zealand is more remote and inaccessible, and, during the whole Mesozoic and Cenozoic eras, we have evidence of but two colonizations by land vertebrates, neither implying any necessary continental connection. The rock-lizard (Sphenodon) may, for aught we know to the contrary, be derived from a marine form; all its early Mesozoic relatives were aquatic, some apparently marine. The few other reptilia may be best accounted for by sporadic colonizations of later date. The moas are probably de- rivatives from flying birds. When we come to the smaller oceanic islands, their poverty of fauna is still more conspicuous. If their fauna is due to sporadic colonization, this should be expected, as the chances are reduced directly in proportion to the smaller length of coastline on which an immigrant might land, as well as by their effective distance from the mainland. The colonization of a group of islands one from another may be due to former land con- nection and subsequent isolation, or to the same method of accidental transport, subject to the same laws of chance. It is quite possible that in certain instances the small size and unfa- vorable environment of islands formerly connected with the continent may account for non-survival of the continental fauna. The Falkland Islands are a case in point; but even here, we find the survivors closely allied to the continental fauna and including types which afford the con- clusive proof of continental connection which is uniformly lacking in oceanic islands.** The characteristics of continental and oceanic island faunz have been very fully and ably elucidated by Wallace (Island Life), and it is in- tended here merely to assert that the progressive increase of our knowl- edge of the past life of the world tends only to emphasize the distinctions in the source of their faunze which he has so clearly demonstrated and, so far as my acquaintance with the subject goes, to reduce still further the number of continental connections which he regarded as permissible. To the argument so often advanced that the transportation of a species across a wide stretch of sea and its survival and success in colonizing a new country in this way is an exceedingly improbable accident, it may be answered that, if we multiply the almost infinitesimal chance of this *7 Introduction of Canis antarcticus by human agency in prehistoric times is, however, a possible explanation of its occurrence. It is the only alternate to a Pleistocene land connection. 206 ANNALS NEW YORK ACADEMY OF SCIENCES occurrence during the few centuries of scientific record by the almost infinite duration of geological epochs and periods, we obtain a finite and quite probable chance, which it is perfectly fair to invoke, where the evidence against land invasion is so strong. Furthermore, the fact that continents have not in general been peopled in this way one from another is well accounted for by the fact that speties already existed there which filled the place in the environment and by their competition prevented the new form from obtaining a foothold, or greatly reduced the chances thereof. In oceanic islands, however, the favorable environment existed without the animal to fill it. Very often, on account of this lack, some other type was evolved to fill its place; birds being widely distributed on account of their powers of flight have in many oceanic islands developed large terrestrial adaptations to take the place of the absent or scanty mammals. NATURAL RAFTS AND THE PROBABILITIES OF OVER-SEA MIGRATION THEREBY The following series of facts and assumptions may serve to give some idea of the degree of probability that attaches to the hypothesis of over- sea transportation to account for the population of oceanic islands. 1) Natural rafts have been several times reported as seen over a hundred miles off the mouths of the great tropical rivers such as the Ganges, Amazon, Congo and Orinoco.* For one such raft observed, a hundred have probably drifted out that far unseen or unrecorded before breaking up. 2) The time of such observations covers about three centuries (1 set aside the period of rare and occasional exploring voyages). The duration of Ceno- zoic time may be assumed at three million years (Walcott’s estimate). 3) Living mammals have been occasionally observed in such records of nat- ural rafts. Assume the chance of their occurrence (much greater than of their presence being noticed) at one in a hundred. 4) Three hundred miles drift would readily reach any of the larger oceanic islands except New Zealand. Assume as one in ten the probability that the raft drifted in such a direction as to reach dry land within three hundred miles. 5) In case such animals reached the island shores and the environment afforded them a favorable opening, the propagation of the race would require either two individuals of different sex or a gravid female. Assume the proba- bility of any of the passengers surviving the dangers of landing as one in three (by being drawn in at the mouth of some tidal river or protected inlet), of landing at a point where ‘the environment was sufliciently favorable as one in ten, the chances of two individuals of different sexes being together ss A recent number of the Popular Science Monthly (Sept., 1911, vol. lxxix, pp. 303-307) gives the recorded observations of the drift of a natural raft of. this sort, covering over a thousand miles of travel. MATTHEW, CLIMATE AND EVOLUTION 207 might be assumed as one in ten, the alternate of a gravid female as one in five. The chance of one of the two happening would be 1/10 + 1/5=3/10. The chance of the species obtaining a foothold would then be 3/10 « 1/5 x 1/10 — one in a hundred. If then we allow that ten such cases of natural rafts far out at sea have been reported, we may concede that 1000 have probably occurred in three centuries and 30,000,000 during the Cenozoic. Of these rafts, only 3,000,000 will have had living mammals*® upon them, of these only 30,000 will have reached land, and in only 300 of these cases will the species have established a foothold. This is quite sufficient to cover the dozen or two cases of Mammalia on the larger oceanic islands. Few of these assumptions can be statistically verified. Yet I think that, on the whole, they do not overstate the probabilities in each case. They are intended only as a rough index of the degree of probability that attaches to the method, and to show that the populating of the oceanic islands through over-sea transportation, especially upon natural rafts, is not an explanation to be set aside as too unlikely for consideration. I have considered the case only in relation to small mammals. With reptiles and invertebrates, the probabilities in the case vary widely in different groups, but in almost every instance they would be consider- ably greater than with mammals. The chance of transportation and sur- vival would be larger and the geologic time limit in many instances much longer. Wind, birds, small floating drift and other methods of acci- dental transportation may have played a more important part with in- vertebrates, although they cannot be invoked to account for the distribu- tion of vertebrates. The much larger variety and wider distribution of infra-mammalian life in oceanic islands is thus quite to be expected. And the extent and limits of such distribution are in obviously direct accord with the opportunities for over-sea transportation in different groups. On the other hand, the transportation of very large animals in this way may fairly be regarded as a physical impossibility, which could not be multiplied into a probability by any duration of time. The only methods of accounting for such animals would be by evolution in loco from small ancestors, by swimming, by introduction through the agency of man and by actual continental union. The first hypothesis would involve evolution in an isolated and more or less altered environment and would result in wide structural differ- ences from any continental relatives. The second applies with greater probability to large than to small animals, but, except for animals of * Small reptiles and invertebrates would only rarely be observed, if present. 208 ANNALS NEW YORK ACADEMY OF SCIENCES more or less aquatic habits and within certain limits of distance, it is an apparent physical impossibility. The third may be either intentional or accidental and should be considered in connection with the known custom among Malays and other races, of taming various captured animals and taking them along on sea-voyages. Its application is, of course, limited to distributional anomalies of late Pleistocene or modern origin. The last hypothesis, where it traverses the doctrine of the permanence of ocean basins, appears to me unnecessary, as I have failed to find a single im- stance of distribution which cannot reasonably be otherwise explained. CONSIDERATIONS AFFECTING PROBABILITIES OF OVER-SEA MIGRATION IN SPECIAL CASES The probabilities of over-sea transportation to an oceanic island will obviously be much greater if the island is large, and correspondingly re- duced if it be of small size. The distance from the mainland will greatly reduce the chances of such rafts making a landing, for two reasons: first, the chances of survival of the animals are reduced proportionately to the length of their journey (or rather, in a varying relation, which for con- venience we may consider as a direct proportion) ; second, most rafts will be carried out from one or more points along the coast, but not from all points equally (that is to say, from the mouths of one or more great rivers, where the conditions are favorable, seldom from any of the small rivers). If we disregard prevalent winds and currents and consider the rafts as drifting out in all directions the probability of their landing on a given island will be directly proportioned to its length opposite the main- land, inversely to the distance. The probabilities of survival of animals, so far as it depends on the raft holding together, will also be inversely as the number of days exposure to the sea, hence as the distance. Compar- ing Saint Helena, 1100 miles from Africa and 10 miles diameter, with Madagascar, 200 miles from Africa and 1000 miles in length, we see that the probabilities of effecting a colonization would be 100 K 314 X 514, or 3025 times greater in the case of Madagascar. New Zealand, 800 miles long and 1200 miles from the Australian coast, will receive 8/10 & 1/6 < 1/6, or 1/45 as many colonizations as Madagascar, but 80 11/12 x 11/12 or 67 times as many as Saint Helena. I beheve that it is to their small size rather than to unfavorable con- ditions for survival that the poverty of fauna, especially of higher verte- brates, in the smaller oceanic islands is due. The oceanic currents and prevalent winds do, of course, profoundly modify the above generalities in each individual instance. They have MATTHEW, CLIMATE AND EVOLUTION 209 prevented the populating of Cuba from North America, while facilitating invasions from South and Central America. The present set of currents reduces the probability of mammals reaching Madagascar from the Afri- can mainland, while increasing the chances of Oriental animals reaching it. It reduces materially the opportunities for Australian fauna to reach New Zealand. We have no adequate data on which to base theories as to the former set of oceanic currents. A worldwide uniformity of climate would prob- ably reduce the north and south movement of the waters; the east and west element of their motions is conditioned by the rotation of the earth, and its velocity would be reduced proportionately to the north and south movements; so that a more uniform climate would bring about a reduc- tion of velocity rather than change in direction. The third principal conditioning element is the conformation of the continents, and doubtless the flooding of great areas and the opening up of broad though shallow passageways between seas now separated would profoundly modify the surface currents in many regions. The opening of a broad passage be- tween North and South America would allow the Caribbean current to pass into the Pacific instead of being deflected northward and eastward along the shores of the Gulf of Mexico to find an outlet between Cuba and Florida. The absence of this initial part of the Gulf Stream would obviously be unfavorable to North or Central American animals reaching western Cuba. The great equatorial current would sweep across from Africa along the northern coast of South America, and uninterruptedly into the Pacific; transportation from Africa to South America or from South or Central America to the Galapagos Isiands would thus be facili- tated. DISPERSAL OF MAMMALIA MANKIND We may with advantage begin our review of the special evidence in support of our theory with the migration history of man. This is the most recent great migration; it has profoundly affected zodgeographic conditions; it is the one where our data are most complete and accurate ; we can perceive its causes and conditions most clearly, and we have a great deal of corroborative evidence in history and tradition. All authorities are to-day agreed in placing the center of dispersal of the human race in Asia. Its more exact location may be differently in- terpreted, but the consensus of modern opinion would place it probably in or about the great plateau of central Asia. In this region, now barren . 210 ANNALS NEW YORK ACADEMY OF SCIENCES and sparsely inhabited, are the remains of civilizations perhaps more ancient than any of which we have record. Immediately around its bor- ders lie the regions of the earliest recorded civilizations,—of Chaldea, Asia Minor and Egypt to the westward, of India to the south, of China to the east. From this region came the successive invasions which overflowed Europe in prehistoric, classical and medixval times, each tribe pressing on the borders of those beyond it and in its turn being pressed on from Fic. 6.—Dispersal and distribution of the principal races of man No attempt is made to indicate anything beyond the broader lines of dispersal. behind. The whole history of India is similar,—of successive invasions pouring down from the north. In the Chinese Empire, the invasions come from the west. In North America, the course of migration was from Alaska, spreading fan-wise to the south and southeast and continu- ing down along the flanks of the Cordilleras to the farthest extremity of South America. Owing to the facilities for southward migration af- forded by the great Cordilleran ranges, the most remote parts of the New World are the forests of Brazil and of northeast South America. In the northern continent, Florida is the most distant from the source of mi- gration. MATTHEW, CLIMATE AND EVOLUTION D111 In Africa, the region north of the Sahara has been overrun by succes- sively higher types from the east. The great desert was a barrier to southward migration, being pierced only by the narrow strip of the Nile valley, from whose head spread out the successive populations of central and southern Africa. The main trend of migration followed the eastern highlands, the valleys of the Niger and Congo being more remote. In the East Indies, the succession of great islands to the southeast, perhaps more connected formerly than now, formed stepping stones of migration to the distant continent of Australia. The lowest and most primitive races of men are to be found in Aus- tralasia, in the remoter districts of southern India and Ceylon, in the Andaman Islands, in southwest and west central Africa and, as far as the New World is concerned, in northern Brazil. These are the regions most remote, so far as practicable travel-routes are concerned, from Central Asia. A century ago, the present habitat of primitive races was taken to be approximately the primeval home of man. With our present under- standing of the conditions and causes of migration, a theory more in ac- cord with tradition and history is generally accepted, and the dispersal center of man is regarded as situated in central or southern Asia. The influence of the old opinion is perhaps seen in the tendency to place this region south of the great Himalayan ridge and in tropical or semi-trop- ical climate. This last assumption—that man is primarily adapted to a tropical cli- mate—is, I think, only partly true at best. Its general acceptance is perhaps due, among other reasons, to the supposed relation between loss of hair on the body and the wearing of clothes, the first being regarded as an earlier specialization in an environment of tropical forests, the second as a secondary adaptation resulting from migration to a cold climate. But here, it seems to me, we are putting the cart before the horse. We may more reasonably regard the loss of hair in the human species as a result of wearing clothes and conditioned by this habit, rather than attrib- ute it to any climatic conditions. This view is supported by several points in which the loss of hair in man is differentiated from the partial or com- plete loss of hair common in tropical animals, the following two being most clearly significant. 1) It is accompanied by an exceptional and progressive delicacy of skin, quite unsuited to travel in tropical forests. I do not know of any thin-haired or hairless tropical animal whose skin is not more or less thickened for pro- tection against chafing, the attacks of insects, ete. 2) The loss is most complete on the back and abdomen. The arms and the legs and, in the male, the chest, retain hair much more persistently. This is 212 ANNALS NEW YORK ACADEMY OF SCIENCES just what would naturally happen if the loss of hair were due tu the wearing of clothes,—at first and for a long time, a skin thrown over the shoulders and tied around the waist. But if the loss of hair were conditioned by climate it should, as it invariably does among animals, disappear first on the under side of the body and the limbs and be retained longest on the back and shoulders. It will not be questioned that the higher races of man are adapted to a cool-temperate climate, and to an environment rather of open grassy plains than of dense moist forests. In such conditions they reach their highest physical, mental and social attainments. In the tropical and especially in the moist tropical environment, the physique is poor, the death rate is high, it is difficult to work vigorously or continuously, and especial and unusual precautions are necessary for protection from dis- eases and enemies against which no natural immunity exists and which are absent from the colder and drier environment. This lack of adaptation to tropical climate is also true, although to a less degree, of the lower races of man. Aithough from prolonged resi- dence in tropical climate they have acquired a partial immunity from the environment so unfavorable to the newcomer, yet it is by no means com- plete. The most thoroughly acclimatized race—the negro—reaches his highest physical development not in the great equatorial forests but in the drier and cooler highlands of eastern Africa; and when transported to the temperate United States, the West Coast negro yet finds the environ- ment a more favorable one than that to which his ancestors have been endeavoring for thousands of years to accustom themselves. In tropical South America, the Indians, as Bates long ago remarked, seem very im- perfectly acclimatized and suffer severely from the hot moist weather ; much more than the negroes, whose adaptation to tropical climate has been a much longer one. : In view of the data obtainable from historical record, from tradition, from the present geographical distribution of higher and lower races ot men, from the physical and physiological adaptation of all and especially of the higher races, it seems fair to conclude that the center of dispersal of mankind in prehistoric times was central Asia north of the great Hima- layan ranges, and that when by progressive aridity that region became desert it was transferred to the regions bordering it to the east, south and west. We may further assume that the environment in which man pri- marily evolved was not a moist or tropical climate, but a temperate and more or less arid one, progressively cold and dry during the course of his evolution. In this region and under these conditions, the race first at- tained a dominance which enabled it to spread out in successive waves of MATTHEW, CLIMATE AND EVOLUTION 91: ~~ migration to the most remote parts of the earth. The great mountain ranges to the south impeded migration in this direction, while to east and northeast, west and northwest, migration was easy and rapid. Reaching the New World by way of the Alaskan bridge, the long uninterrupted chain of the Cordilleras facilitated migration along their flanks to the farthest limits of South America. There is little evidence if any, in the New World, of any migrations of inferior races long preceding those of the Amerind tribes, which would seem to have branched off at a moderately high stage in the evolution of mankind. Per contra, we find in South Africa, in Australia, in penin- sular India and elsewhere, remnants of what must have been an early cycle of migrations. Each group of this early cycle, derived primarily from a different part of the central region of dispersal, has specialized further in proportion to its isolation and yet retains a predominance of the common primitive characters representing the stage of development attained when it left the dispersal center. The populating of Africa by the negroes may be regarded as the latest phase of this early cycle of dis- persal, or should perhaps be considered independently. The later development of the race is conditioned by its splitting in the region of dispersal into an eastern or Mongolian and a western or Cau- easian stock. This split was presumably conditioned by the east-west elongation of the dispersal center caused by the facility of expansion in these directions and the mountain barriers to the south. All the east- ward migrations from this time on bear a distinctly Mongolian stamp. An early phase of this stage is represented by the population of the New World and the variously mixed Malayan peoples. A later phase appears in the more typical Mongolian races. All the westward migrations, on the other hand, are of Caucasian affinities, this stock splitting, as the re- gion of favorable environment widened out westward, into northeastern or Nordic, southwestern or Mediterranean groups. The peoples of north- ern Europe are derived from the successive migration waves of the first, those of southern Europe and northeast Africa from the second; the in- termediate Alpine stock of central Europe is considered to represent a somewhat older migration allied to the Slavic peoples, who are to-day the principal population of eastern Europe, the latest cycle of Caucasian dispersal. I have gone into this brief recital of the migration and dispersal his- tory of mankind, not to present anything novel or authoritative, but be- cause we have more evidence, direct and indirect, and more insight into the conditions and causes which controlled its course, than with any other 214 ANNALS NEW YORK ACADEMY OF SCIENCES race of mammals. I believe that these controlling causes have been sub- stantially the same in the lower animals as in man and their methods and routes of dispersal largely identical.*° PRIMATES We have seen that the dispersal center of man is in central Asia; that, in the present distribution, the survivors of the earliest cycle are found DISTRIBUTION OF PRIMATES == Movern AntTHRopoiDed (MONKEYS, APES , BABOONS ) Wu ” LemuroipEa (LEMURS ,LORIS, TARSIER) E, Eocene (Ano OligoceneyL EMUROIDS 0, Oricocene ANTHROPOIDS M, Miocene y P, Puiocene os) | ore RS uth til oo” ™ il =i Fic. 7.—Dispersal of the Primates The marginal position of the modern lemurs, the progressive disappearance of the order from the more central regions which it formerly inhabited are clearly shown. in Africa, peninsular India, the East Indies and Australia; that the populating of the New World belongs to a later cycle of distribution, and we have no good evidence that the earlier cycle ever reached it; that the dominant migration in the Old World has been east and west, prog- ress to the south being hindered by the transverse mountain system to the south of which more primitive types long survived, while in the New World the dominant line of migration has been to the southward from Alaska, and eastward migration has been slower. 40 One notes, too, the same fallacy in interpreting the data; some authors are disposed to place the center of dispersal of European races or languages in western Europe or in northern Africa because they find there the most primitive surviving races or languages. MATTHEW, CLIMATE AND EVOLUTION 915 In the hving Primates we have survivors of pre-human stages in the evolution of man, specialized to a varying extent in different directions from him, so that they have not come into direct rivalry with him, and have hence survived. The latest infra-human cycle is represented by the anthropoid apes, surviving to-day in the forests of West Africa and of the East Indies. We may suppose that these are remnants of a cycle of dispersal from a central Asiatic source, but we have no sufficient data to define its extent or time, except as late Tertiary and probably limited to Arctogea. Nearest to man in intelligence and habits, this cycle has been swept out of existence, except for the few members which were or became adapted, LIVING AND EXTINCT GROUPS OF PRIMATES LEMUROIDEA ANTHROPOIDEA NSECT- IVORA Progressive EVOLUTION oF THE HIGHER Groups oF PRIMATES DURING THE TERTIARY PERIOD Fic. 8.—Phylogenetic relations of the living and extinct groups of Primates The circles indicate the size and known geological range of the several groups, the dotted lines their most probable derivation. Their supposed relations to certain Insectiy- ora and intermediate extinct groups are also indicated. as our own ancestors were not, to tropical forest hfe. The arboreal habitat may be interpreted as a partial reversion. The doubtful and fragmentary remains of anthropoid apes in the Pliocene of Europe and of northern India are about all that the geological record has to state in regard to the former distribution of this cycle. The next lower cycle is that of the monkeys and baboons of the Old World, and as a very doubtful early phase, the New World monkeys. The Old World monkeys inhabit most of Africa, India and the East Indies. To the northeast they extend to southern Japan. Closely re- lated forms are found in the late Miocene of central and southern Eu- 216 ANNALS NEW YORK ACADEMY OF SCIENCES rope, in the Pliocene. of India, in the so-called Pliocene (which may be Miocene) of China. These may all be referred to a central Asiatic source. The dispersal of this cycle must date back at least to the be- ginning of the Oligocene, for it had reached as far as Egypt at the date of the Faytim fauna as shown by Schlosser’s recent discoveries.** With the New World monkeys, the evidence seems rather to point to inde- pendent evolution in South America from éarly Tertiary Primates of an Eocene cycle of dispersal. For no remains of Primates have been dis- covered in any Oligocene or later formation in the United States, while the later Tertiary formations of the Argentine have yielded remains of a number of Primates apparently intermediate between Eocene lemurs ~and South American monkeys. The oldest cycle of primate dispersal is that represented by the lemurs. These are now most abundant in Madagascar; a few exist in west and central Africa, peninsular India and the East Indies. Lemuroid pri- mates, lacking certain specialized characters of modern lemurs but other- wise closely related, and equivalent in stage of development, are found abundantly in the Eocene of Europe and the United States. They are very doubtfully represented in the early Tertiary formations of the Argentine. We know too little of the Tertiary of other parts of the world to make any inference as to the extent of their distribution at that time, or the course of its subsequent changes. They disappear in Europe and North America at the end of the Eocene; in South America, they may have evolved into New World monkeys, while in the Old World they must have given rise to the higher primates. It is reasonably cer- tain that the theater of their evolution was not Europe, and although they are not known in the Oligocene Faytim fauna of Egypt, we may doubtfully suppose that they had reached that continent at some time during the Eocene. Madagascar most probably received its lemurs from Africa, but it is reasonable to suppose that only a single type, allied to the Eocene Adapide, reached the island, and in the favorable environ- ment radiated out into a number of diverse adaptations taking the place of various mammal groups not present in the island fauna. From the fact that the European and North American lemurs are in an equivalent stage of development, although not very closely related, we may fairly infer that they were derived very early in the Tertiary from an intermediate center of dispersal, presumably Asia north of the Himalayas. 41 Max ScHLOSSER: “Beitriige zur Kenntniss der Oligoziinen Landsiiugethieren aus dem Fayum Aegypten,”’ Beit. zur Pal. u. Geol. Oest-Ung., B’d xxiv, s. 52. 1912. MATTHEW, CLIMATE AND HVOLUTION 917 CARNIVORA The modern land Carnivora are spread over all the great continents except Australia, where a single species of wild dog, probably introduced by man, is their only representative. They are found equally in all the continental islands (7. e., those included within the continental shelf border), and a few have reached Madagascar and other large oceanic islands. Miocene relatives of Cyon, Lelicyon Ly caon & Dingey nope = Dinocynops (Pleislocen e) Ce dingo, / probably intro : ‘diseeld in latel§ Pleislocene Eocene lo Recent = Fic. 9.—Distribution of the modern Canide The jackals (Ethiopian and Oriental) are slightly more primitive than the true wolves and foxes; the Neotropical ‘Dog-foxes’’ more distinctly so. Cyon, Icticyon and Lycaon appear to be dispersed remnants of an aberrant group formerly Holarctic; the ancestry of the more typical Canidz is also found in Holaretica. The order is unquestionably of Holarctic origin. Primitive Carnivora (Creodonta) are abundant in all the earlier Tertiary formations of Eu- rope and North America, one group (Miacidw) ancestral to the higher Carnivora (Fissipedia), others which became extinct during the Oligo- cene. True fissipede Carnivora first appear in the Upper Kocene in Europe and North America and differentiate into the diverse modern types through the remainder of the Tertiary. They did not reach South 218 ANNALS NEW YORK ACADEMY OF SCIENCES America until the Pliocene, their place being supplied up to that time by carnivorous marsupials. In Australia, their place is still taken by carnivorous marsupials. In Africa, primitive Carnivora (creodonta) appear in the Oligocene, represented only by the extinct family of hyaeno- donts, all of them derivable from Eocene hyznodonts of the Holarctic region; but the contemporary Holarctic Fissipedia had not yet reached Africa. , The modern Jand Carnivora are divided into seven families, each rep- resenting one or more broad phyla. ‘The various divergent adaptations of the phyla and secondary re-adaptations of subphyla have brought about an amount of convergence and parallelism which makes it difficult to disentangle or to state accurately the true genetic relationship in any terms of classification. Some of the phyla are Holarctic, others Pale- arctic or Nearctic. In all of them, we find the most primitive modern survivors in the tropical regions, the most advanced types in the Holarctic. Canide.—The Canidex are the most cosmopolitan family of the order. It is also the most progressive family in its adaptation to the open plains and arid climate of the great modern continents. The gradual adapta- tion of the race to these conditions from primitive arboreal forest-living ancestors can be traced through successive stages in the Tertiary forma- tions of Europe and North America, but most completely in the latter country. The lengthening of the limbs and their adaptation for swift running, the reduction of the long balancing tail to a short comparatively unimportant organ, the perfection of the shearing and crushing teeth and, especially, the steady increase of brain capacity are the chief lines of progress. While most of the surviving Canide conform pretty closely to a single type, we find a tendency among their Tertiary ancestors to branch off on the one hand into more predaceous, on the other into more omnivorous types. Most of these have disappeared, but in the Oriental Ethiopian and Neotropical regions we find in the genera Cyon, Icticyon ~ and Lycaon survivors of a more predaceous group which is known from the Oligocene and Miocene of the Holarctic region. This group has disappeared from Holarctica by the end of the Tertiary; two or three representatives are found in the Pleistocene of South America. Among the more typical modern dogs, the wolves and foxes are the most pro- gressive types, the jackals slightly less so, the African fennec retains most nearly the primitive long tail, the South African Otocyon, while anomalous in possessing an extra molar tooth, is likewise normally primi- tive in several characters and the Neotropical “dog-foxes” show a marked resemblance in many details to the late Tertiary Canide of North Amer- ica. The fact that the Canide are preéminently adapted to open country MATTHEW, CLIMATE AND EVOLUTION 219 and more or less arid climate is of primary importance in explaining their present dominance and cosmopolitanism, their close association with man, their absence from Madagascar and other oceanic islands; and it makes it most probable that the introduction of the dingo to Australia was through human agency although undoubtedly as early as the late Pleistocene. In their adaptation and distribution this family of Car- nivora largely parallels the Equide among Perissodactyla. TABLE I1.—Distribution of the Canide Neotropical Holarctic | Ethiopian Oriental | Australian . Canis AA Canis ae Canis Cues Recent Teticyon Canis Cae Cyon Canis Canis ‘, Pleistocene | Icticyon Canis Canis als Canis Dinocynops Y . ; Canis (Record in- | Canis Pliocene (?) Amphicyon Ga: ete. | * de quate) | Vulpes (No record) Tephrocyon Miocene None ae Nae yonine No record | Amphicyon | (No record) Cyon, ete. Cephalogale : ‘ Cynodictis a2 Amphicyon Oligocene None Daphznus, ete. None Cephalogale (No record) ‘‘Amphicyon’’ Cynodictis, ete. Eocene None Cynoid (No record) Miacidee 42Fayfm fauna, Egypt. Although this locality is not to-day within the Ethiopian province, its fossil mammals are generally regarded as representing the Hthiopian and not the Mediterranean fauna of the Oligocene. My own impression with regard to it is that it is transitional, as the Egyptian fauna is to-day, but dominantly Ethiopian instead of dominantly Mediterranean. 220 ANNALS NEW YORK ACADEMY OF SCIENCES Procyonide.—The family Procyonide includes a member of omnivor- ous specializations from the central phylum now represented by the Canide. All of them are arboreal, partly retaining and partly reverting to the primitive mode of life in this respect. They are mainly Neotrop- “ WH, Procyonidde wee Procyon = NT 7 EN Neae Ter liarpanceslops =: , =~ of Viverrdé No Procyonidae fs unl Pliocene Fic. 10.—Distribution of the Procyonide and Viverrida, formerly Nearctic and Palearctic, but now surviving chiefly in the peripheral regions The geographical position of Aelurws is anomalous for a member of the Procyonide, to which family it is usually referred. Its true affinities, however, are doubtful. ical, but the raccoon, the most dog-like of the family, survives as far north as the Sonoran region. The panda of the Himalayas is usually placed with Procyonide, but its true affinity is not very clear. MATTHEW, CLIMATE AND EVOLUTION TABLE III.—Distribution of the Procyonide 221 Holarctic Neotropical = Sonoran Palearctic Procyon Nasua Alurus Recent Cercoleptes Procyon (Affinities question- Bassaricyon able) Bassariscus Pleistocene (Not recorded) Procyon Amphinasua Parailurus Pliocene Pachynasua Probassariscus (Affinities question- Cyonasua able) ; Leptarctus Miocene None Phiaocyon C | dicti. Hae ynodictis — Oligocene NORE (Probably ancestral in part) Eocene — None Miacidee Mustelide.—Primarily the Mustelide represent a more predaceous adaptation than the Canide. Their development through the Tertiary in the Holarctic region can be traced almost as completely as that of the dogs. Like the Canide (though not as early), they perfected during the later Tertiary a differentiation of the back teeth into shearing and crushing types, and they are equally progressive in brain development but much less so in running powers, retaining to a great extent their primitive forest-living habitat. They are to-day chiefly holarctic, the most progressive typical mustelids being the martens, weasels, ferrets and wolverenes. Karly in the Tertiary there appear divergent side branches, specialized descendants of which survive to-day in the badgers, skunks and otters of the northern world, the intermediate forms being now extinct or confined to India and Africa. Urside.—The bears are regarded by many paleontologists as an off- shoot from the Canide, but, on structural evidence, they appear to be related rather to the Mustelidew. Their distribution indicates derivation 299 ANNALS NEW YORK ACADEMY OF SCIENCES from a Palearctic source. The most primitive bears first appear in the Miocene of Europe; in the New World, they first appear in the Pleisto- cene. They are to-day chiefly Holarctic; the single South American species is distinctly primitive; the Oriental sun-bear and sloth-bear are partly aberrant, partly primitive. The Thibetan 2/uropus is aberrant and specialized ; its relation to the typical Wrsid@ is not very close. Vwerride.—The Viverride are now almost exclusively Oriental and Kthiopian and have conserved the primitive type more than any other ——_ No Ur sidae unt) Pleislocene 1 = a 7r Vy Se Ancestry D Lae == Of Ursidae fe =" tn later Tertiary Fig. 11.—Distribution of the Urside, Pleistocene and Recent The group appears to have dispersed from a Palearctic center, its Tertiary ancestral series being found in Europe and in the Pliocene of India and China. Carnivora, except some of the Procyonide which have a somewhat corre- sponding geographic position in the New World. The three most pro- gessive genera, /erpestes, Genetta and Viverra, survive to-day along the southern borders of the Palearctic region; the remainder are Ethiopian or Oriental, the most primitive living genera being west African and East Indian. Herpestes and Viverra occur in the Oligocene and Mio- cene of Germany and France, and more primitive extinct genera in the Upper Eocene of Europe. MATTHEW, CLIMATE AND EVOLUTION 993 The primitive character of the viverrines is especially seen in their imperfect differentiation of shearing and crushing back teeth, their rather short limbs, long bodies, long tails and relatively small brain capacity. Hyenide.—tThe family Hyenide is generally regarded as a specialized offshoot from the Viverride and is apparently connected with the Kuro- pean Miocene viverrids by a series of intermediate forms. The latest development of the race, the genus Hyena, inhabited Europe and Cen- tral Asia and China in the Pliocene and Pleistocene but is now found only in India, Africa and southwestern Asia. Felida.—The Felide are almost as cosmopolitan as the dogs and are even more uniform in type, the cheetah being the only marked living variant. A notably different specialization is shown in the extinet macherodonts or sabre-tooth tigers, and in the Tertiary sequence in Europe and America we find approximate genetic series, parallel in the two countries, by which the true cats and macherodonts converge towards a common primitive type, in which the upper canines are moderately elongated. According to this phylogeny, the clouded tiger of Sumatra and Java is the most primitive living felid, while the double series in Europe on one hand and North America on the other, would indicate northern Asia as the center of dispersal of the race. The range of some of the modern species is very great. The puma extends in the New World from Alaska to Patagonia, the tiger in the Old World from Man- churia to Java. We may note, however, that the tiger is regarded by Blanford as a recent immigrant into southern India; while, on the other hand, it is known that the northern range of the lion has been pro- gressively restricted during prehistoric and historic times from northern Europe to its present limits of southwestern Asia and Africa. PINNIPEDIA When dealing with littoral and marine mammals we must expect to find the conditions of their evolution somewhat different. If the hypoth- esis be valid that the progressive refrigeration of the polar regions was the dominant cause of evolutionary progress and geographic dispersal, an examination of the map will show that the Arctic-North Atlantic basin affords the most favorable region. The Arctic basin centers around the pole, and a broad shelf of shallow water encircles it, extending as far south as latitude 45°. The North Pacific basin was closed to the north- ward by the Alaskan land-bridge during a large part if not all of the Tertiary, and its shores plunge suddenly to great depths, margined by 294. ANNALS NEW YORK ACADEMY OF SCIENCES high mountains, affording less opportunity for expansional evolution of the littoral fauna. The Antarctic continent appears equally unfavorable, and dispersal from that center would also be hindered by the broad stretches of ocean. We may expect, therefore, to find the littoral fauna of the North At- lantic most progressive, that of the North Pacific less so, the tropical faune containing many relict types of discontinuous distribution, and the Antarctic faune partly composed of types from the north which had crossed the barrier of warm water when the climatic zones were less differentiated than they now are; partly of groups developed in the south. Whether these groups were closely allied on the different southern con- tinental shores would depend on their ability to cross the great barriers of deep ocean that separate them. The distribution of the pinnipeds accords with these principles. The most specialized family is the Phocide, originating apparently in the Atlantic-Arctic basin, where Phoca, the most progressive genus, is found in the North Atlantic and Arctic seas and has penetrated into the North Pacific as far as California and Japan. Southward in the Atlantic it is succeeded by the less progressive Monachus in the Mediterranean and Antillean region. The Antarctic Phocide are also primitive and archaic, related more or less nearly to Monachus. In the Pliocene of Belgium are found extinct genera closely related to Phoca and others more primi- tive allied to Monachus. The walruses, also Arctic and North Atlantic, have penetrated into the North Pacific only as far as Bering Sea; they are likewise recorded from the Pliocene of Northern Europe and along the North Atlantic in the Pleistocene as far south as Virginia. The third family, the Otariidz, is decidedly more primitive in struc- ture, being less specialized for marine life. They are found in all the southern seas and on the North Pacific coasts. They are unknown to the North Atlantic and Arctic shores and have never been found fossil in either Europe or eastern North America. Desmatophoca and Ponto- leon of the Miocene of Oregon are perhaps ancestral types, but more evidence is necessary before its North Pacific origin can be regarded as satisfactorily indicated. INSECTIVORA Among the Insectivora we deal with a number of very ancient races, whose relationship is much more distant than in many other mammalian orders. They are small, and most of the surviving members are scarce and little known, while they are still less known as fossils. So far as MATTHEW, CLIMATE AND EVOLUTION 225 we have any satisfactory evidence, the different races or most of them appear to have originated in the Holarctic region and spread to the southward. ‘The most primitive division, the zalambdodonta, includes four families, the Centetide of Madagascar, Chrysochloride of South Africa, Potamogalide of West Africa and Solenodontidz of Cuba. Fossil zalambdodonts are found in the late Miocene in South America, in the early Oligocene (and recently in the Basal Eocene) in North America. These indications are in conformity with the derivation of the group Soricidae and Talpidde since Oligocene Fic. 12.—Distribution of Soricide (right to left shading) and of Talpide (left to right shading) The less specialized Soricide are more widely dispersed, the highly specialized Talpide limited to Holarctica. Ancestral types of both are found in the Tertiary of Europe and North America, but the evidence as to their phylogeny is very scanty. from very ancient Holarctic ancestors, the modern zalambdodonts being the last surviving remnants of a dispersal from the north in early Ter- tiary or possibly pre-Tertiary times. But the evidence is too slight to be conclusive. The hedgehogs are more clearly of Palearctic origin, the most primi- tive survivors being the East Indian Gymnura and Hylomys, while the most progressive genus, Hrinaceus, is Palearctic and is preceded in the 226 ANNALS NEW YORK ACADEMY OF SCIENCES Oligocene aud Miocene of Europe by more primitive ancestral forms. A relatively primitive genus, Proterix, occurs in the Oligocene of South Dakota, as a contemporary with more progressive genera in the Oligocene of Europe. The family is otherwise unknown in the New World. The moles and shrews are also evidently of northern origin. Of the two families, the Soricide are more primitive in structure and have spread more widely; the more specialized Talpide are still limited to Holarctica, and in the extreme north their exclusion from the areas of permanently frozen subsoil has split their range into two disconnected areas. The most progressive and abundant shrews are Holarctic, while the Oriental and African species (Crocidurine) retain some primitive characters. Fossil moles and shrews in the middle Tertiary of Europe and America indicate that the divergence between the two families was not then so great as now. ‘The modern genera are reported to occur (but on inadequate evidence) as early as the Miocene in Europe and America. Jaws of several minute talpoid genera are known from the Middle and Lower Eocene of North America. They are unknown in the extra-Holarctic Tertiary, but this negative evidence is of no weight in view of their minute size and rarity. The Tupaiide of the East Indies and Macroscelidide of Africa occupy a somewhat anomalous position, since they are of higher type in brain development than other Insectivora and in many respects are nearer to the higher placental mammals.** Their distribution so remote from the great northern dispersal center may perhaps best be accounted for by considering the fact that their specializations, adaptations and habits of life are of a less unusual kind than in most of the lower insectivores and would bring them more directly into rivalry with certain groups of rodents, with which they were unable to contend successfully and were compelled to retreat southward in consequence. No fossil remains cer- tainly referable to these families are known, although quite a number of early Tertiary genera of Europe and North America have been or might be provisionally referred to them.** There are a large number of primitive Insectivora in the Eocene of North America and a few in Europe, which do not seem to be nearly ancestral to any modern group but rather indicate that the order once #8 This anomaly in distribution is now removed by the studies of Gregory and Elliott Smith, which show that the true relations of Tupaia and presumably of Macroscelides, are with the Primates, rather than with the Insectivora. Their geographic distribution is quite normal on this view of their affinities. 44 Hntomolestes of the Middle Eocene of North America is regarded by Dr. Gregory as probably related to Tupaia, and a number of other small mammals from the Bridger and Wasatch may be related to this group of Insectivora. MATTHEW, CLIMATE AND EVOLUTION 997 took a much more important place in the mammalian faune of the world than it does now. This should be kept in mind in considering the rela- tions of the Insectivora. CHIROPTERA I am not sufficiently acquainted with modern Chiroptera to venture an opinion as to whether or not their geographical distribution indicates their place of origin, but I should not expect to find much satisfactory evidence, as they are known to be of very ancient specialization and to have greater facilities for wide distribution than terrestrial animals. Dr. Andersen,*® in his recent Catalogue of the Chiroptera in the British Museum, remarks: “The evidence afforded by the geographical distribution of Bats has generally been considered of doubtful value; hence they have either been entirely excluded from the material worked out by zodgeographers, or at least treated with pronounced suspicion as likely to be more or less unreliable documents of evidence. This un- willingness or hesitation to place Bats on an equal zodgeographic footing with non-flying Mammalia would seem to be due partly to the precon- ceived idea that owing to their power of flight Bats must evidently have been able easily to spread across barriers which in ordinary circumstances are insuperable for wingless Mammalia; partly to the fact that hitherto very often whole series of distinct forms have been concealed under one technical name. . . .” [the author cites a series of instances of this kind which] “tend to show that the present distribution of the Mega- chiroptera has not been influenced to any great, and as a rule not to any appreciable extent by their power of flight; if it had the Fruit-bat fauna of islands could not so commonly as is actually the case differ from that of a neighoring group or continent, and the tendency to dif- ferentiation of insular species or forms would have been neutralized by the free intercourse between neighboring faunas.” The belief that bats are more easily able to cross ocean barriers than non-flying mammals is probably based, not on the preconceived idea that they could, but upon the plain fact that they have done so far more frequently. Birds and bats are found upon numerous oceanic islands where no non-flying mammals, and very few non-flying animals at all, exist. That they have wings and occasionally use them for so long a journey, whether voluntarily or involuntarily, is a natural explanation. I cannot see any other reasonable interpretation of the fact that they are present and the terrestrial mammals absent in so many remote oceanic 4K. ANDERSEN: Catalogue of the Chiroptera in the British Museum, Vol. I, Megachi- roptera, p. lxxvi. 1912. yore} ANNALS NEW YORK ACADEMY OF SCIENCES islands. With bats, as with most birds, the intervening ocean acts as a hindrance, but their wider distribution shows that it is less of a hindrance than with terrestrial mammals. RODENTIA The abundant and dominant order of: Rodentia lends, in general, strong support to the theories here advocated; but there are certain serious difficulties which can be reconciled only by appealing to the im- perfection of the geological record. The rabbits and picas form a group apart, the former Nearctic, the latter Palearctic since the Oligocene, and both Holarctic since the Pleistocene, the rabbits having extended their range over most of the Oriental region and a large part of the Ethiopian and Neotropical. A single specimen is recorded as from the Pleistocene of South America; their introduction to Australia is known to have been by civilized man. Of the remaining rodents, the myomorph families are evidently of Holarctice origin, as they first appear in Europe and North America in the Oligocene and the highest and most progressive modern types (e. g., Arvicoline) are now Holarctic, while in the southern continents they are unknown until the Pleistocene and various primitive survivals are found still living in Oriental, Ethiopian and Neotropical regions. We may note, however, that the very abundant and typical group of Cricetine has its most primitive living representatives in tropical regions, that as we go south in South America, the genera approximate more toward the more specialized arvicoline type, in the same way that they do as we go northward in the northern continents.*® ‘Since there is no doubt that the cricetines are of northern origin, appearing first in South America in the Pliocene or Pleistocene, while they are common in the Holarctie regions from the Oligocene to the present day, we must suppose that the higher development of the Antarctic genera, to which Oldfield Thomas has called attention, is a case of parallelism with that of the Arctic genera and that the colder climate of the far south is the stimulus which reversed the usual conditions of geographical distribution. A review of the fauna of the Argentine as compared with that of tropical South America tends to show, I think, that this condition is general throughout, and that the fauna is more progressive and more nearly equivalent in development to those of the northern world than is that of the intervening tropical zone. This is equally true of autochthonic races and of those which are demonstrably of northern origin. Compare distribution of the 46 OLDFIDLD THOMAS, MATTHEW, CLIMATE AND EVOLUTION 229 genera of Procyonide, Canidee, Cervide, Tatuide and Dasypodide among mammals. Among the sciuromorphs, the squirrels are of early appearance (Oligo- cene) in the northern world but are now most abundant in the Hast Indies. The more specialized and later appearing marmots are chiefly Holarctic. The highly specialized beavers and pocket-gophers are Hol- arctic and Nearctic respectively, from their first appearance. A marked exception to the rule is seen in the survival in the western Sonoran sub- iG : x Wj Geomyidae Oligocen 10 CORE SS NX deteromyidae Anteslors of. — Pedetidae /Geomyidie ana Stele: Van MM Aromadirae eleromylade {?Anomaluroidea — ia hnevido os Fic. 13.—Distribution of Geomyoidea, Anomaluride and Pedetide The Geomyoidea are of Nearctic origin, but the more primitive Heteromyide have spread into part of South America. The Anomaluride are thought to be the nearest living relatives of the early Tertiary Theridomyide. The Pedetide are an aberrant specialization, derived perhaps from the same group. region of Aplodontia, the most primitive living sciuromorph in several respects. I have no explanation to offer of this anomaly, save that we have not yet balanced properly the essential qualities of progressiveness among Rodentia. Among the hystricomorphs, we find serious difficulties in the distribu- tion. The most primitive living group is certainly the Anomaluride of West Africa; but, like the Pedetide of South Africa, they offer a 230 ANNALS NEW YORK ACADEMY OF SCIENCES puzzling admixture of characters, which makes it doubtful whether they should be reckoned as pertaining to the same stock as the other hystrico- morphs. The remaining families, while chiefly South American, are also partly represented in the Ethiopian, Oriental and Holarctic regions. It may be possible, in view of the facts that the European Theridomyide antedate geologically any specialized hystricomorphs, are apparently di- rectly intermediate between the primitive rodent type (Paramys and its allies) and the hystricomorphs and show the early stages of differentia- tion of several hystricomorph families, that the Hystricomorpha are a u | hes | ami Lerliary anceslo No Erelaizontidde oY Hy slricidde anil Tertiary ances lors of Brethizon lidae Fic. 14.—Distribution of the true porcupines (Hystricide) and New World porcupines (Frethizontide) The Hystricide appear to be of Palearctic dispersal, the Erethizontide are apparently of Neotropical origin. group of Holarctic origin which has spread into all the southern conti- nents and specialized independently on parallel lines. But their entire absence from the recorded North American ‘l'ertiary is then explainable only by the defective record, and our knowledge of North American Ter- tiary rodents is so extensive that I should hardly regard this assumption as justifiable. The fact that the highest and most specialized types are South American necessarily involves the idea that that continent has been the most important center of their later development and dispersal, MATTHEW, CLIMATE AND EVOLUTION 234 and the alliance of the African to the South American genera and of the New and Old World porcupines must be regarded as more remote than it appears. Dispersal from South America by help of Antarctic or transatlantic land-bridges will not solve the problems of their distri- bution much better. The most specialized porcupines in most respects are the hystricids of the Old World—late Tertiary in Nurope, now chiefly Oriental and African. The Nearctic porcupines (rethizon) are more advanced in several features than the Neotropical (Synetheres). Yet the ancestors of the New World porcupines at least occur in the late Tertiary of South America and are absent or unrecorded from the Ter- tiary of North America. The distribution of the Octodontide in Africa and South America would possibly admit of being interpreted by parallel development from theridomyid ancestors; but the parallelism must have been singularly close, and the absence or non-recognition of Therido- myide from the North American Tertiaries appears surprising. I have been unable to frame any hypothesis which will fit all the facts of distri- bution in this group,*? except by assuming that the South American Hystricomorpha, which as Scott has shown are all clearly derived from a single stock, reached South America from Africa in the Oligocene by over-sea raft transportation. This involves so long a voyage that I hesi- tate to accept it as a reasonable probability, even though the winds and currents would obviously favor transportation in this direction. The Hystricide may fairly be assumed as of Old World origin, and probably Palearctic, since they are represented in the later Tertiary of Europe and are unknown in the New World. The Erethizontidz must apparently be derived from South America, since they are unknown in the Old World, and unknown in the North American Tertiary, while Steiromys of the Patagonian Miocene appears to be ancestral. The primary type of the simplicidentate rodents, as I have elsewhere shown,** must be regarded as being represented by the Ischyromyide of the American and European Kocene, in particular by Paramys and Sciu- ravus. All other rodents may be derived from this group by divergent, parallel and in some respects convergent evolution. Modern rodents rep- resent a great number of independent derivations from this primary stock, their association into sections and families being to a considerable extent artificial. 47 The hypothesis of migration to or from South America across a land-bridge from Africa to Brazil is equally unsatisfactory as an attempt to explain the relations of the hystricomorph families and is entirely at variance with the evolution and distribution of other mammalian orders, besides being highly improbable on isostatic grounds. The supposed evidence in its fayor from lower vertebrates and invertebrates is due, so far as I have been able to examine it, to a lack of appreciation of the principles of dispersal of races and of parallelism and of the imperfection of the geological record. 48 “Osteology and Relationship of Paramys and Affinities of the Ischyromyide,”’ Bull. Am. Mus. Nat. Hist., vol. xxviii, p. 43-71. 1910. 232 ANNALS NEW YORK ACADEMY OF SCIENCES There are no rodents in the Notostylops Beds of South America (Ko- cene) ; presumably therefore none in preceding epochs. There are none in the Paleocene of Europe and North America; presumably therefore their sudden appearance in the true Eocene of these regions was due to migration from some other region, equidistant from either, as their de- velopment is almost equivalent in the two,—therefore probably Asia. The few Theridomyide of the Oligocene of Africa are rather primitive forms, certainly not more progressive than their contemporary relatives Vy Oclodontdae Be : = SN aly AE lyslricoimorp ‘S Ss Dasyproclidae ae (Theridomyiade) y —— Chinchillidae / é7 pper Hocene | | | Cavitdde ve and Oligocene JVo ie erlia ry Hystlrico- morphs Miocene ancestors of Octodonts , Chinchillids and Caviids Fic. 15.—Distribution of the Neotropical families of Hystricomorphs The Octodontids are also found in Africa, and the Theridomyide of the early Tertiary of Europe are apparently ancestral to these families of the Hystricomorpha. No hy- pothesis satisfactorily explains the accepted relationship and distribution, in Europe, affording thus a slight indication that they were Palearctic immigrants. In Australia the evolution of Marsupial analogues of the more abundant rodent types of Arctogwa affords strong evidence that the true rodents were absent from Notogwa until the end of the Tertiary; a view confirmed by the limited amount of adaptive radiation which the invading Muride have undergone in that continent up to the present day. The Australian Muride can only be accounted for by over-sea transpor- tation, for the family appeared and evolved during the middle and later Tertiary, and the peculiarities of the Australian fauna are explained by all writers as due to isolation extending through the Tertiary period. MATTHEW, CLIMATE AND EVOLUTION 233 TaBLe 1V.—Distribution of the Rodents S. America | N. America Asia Europe Africa Australia Myomor- M yomor- Muride and| “pha | PP oy. | Mygmor | Myomor Hystrico- | Sciuromor- | ~~) ; é he é Recent morpha pha ape ox eat ae Sees Muridee Lagomor- | Lagomor- | ~* ee g ee nh ee pha pha ia r st See eo ea Ee eHiioon Corie) ze | pha morpha Hystrico- ee Soka eee E morpha Sciuromor- | Sciuromor- | Sciuromor- Pleisto- and ha. na ahha 5 9 gene Muridz Te somor- Te omor- a omor- ' Lagomor- ah A an ‘ Pa pia Erethizon Hystricidee | Hystricidze 1,3 Myomor- Sciuromor- : Pili Hystrico- vee wee See te Ree eae 2 Tene | emmornpha ‘ cha Leporidze? | pha erat : Hystricide | Ochotonide Leporidee y Sc I Hystricidee Sciuromor- Sciuromor- meateae ys pha pha Miocene ree Myomor- ? Myomor- i P pha pha Leporidee Ochotonidee Sciuromor- Sciuromor- pha pha Myomor- Myomor- Oligocene Pepalo: pha 2 pha ’ Leporidee Ochotonide | Therid - Ischyromy- Therido- Ade idee myide |) mye Therido- Bocce None Ischyro- 2 myidee 2 myidee Ischyro- myidee Paleocene None t None Cretace- a Rae None ? None 42 A hystricomorph, recorded by Ameghino from the Pyrotherium beds. ANNALS NEW YORK ACADEMY OF SCIENCES raw) Oo PERISSODACTYLA The order Perissodactyla is represented to-day by three widely sepa- rated families—the rhinoceroses, Ethiopian and Oriental; the tapirs, Neotropical and Oriental, and the horses, Asiatic and Ethiopian. The last group is the most progressive and modernized, but the whole order must be regarded as having seen its best days and as passing towards ex- tinction in competition with the better organized and more adaptable =— DISPERSAL OF THE PERISSODACTYLA (HORSES , RHINOCEROSES ano TAPIRS) i J Distripution of Existing Horses | ZA 5 : RniNoceroses ——~+ KY ROY ” » TAPIRS @) HORSE . Surroseo Centres or Dispersal \ x RHINOCEROS \— @ OURING TERTIARY Perioo TAPIR fetes / ea RENE SS AS ae Fic. 16.—Dispersal and distribution of the Perissodactyla The tapirs are on the whole the most primitive and their present distribution widely discontinuous. The rhinoceroses are less widely dispersed and the horses the most cen- tral in their present distribution. All were inhabitants of Tertiary Holarctica, but their dispersal centers appear to have been Palearctic, as indicated. Artiodactyla. The geological record affords abundant evidence of the Holarctic origin of all the Perissodactyla. The ancestry of each race can be traced back in the Tertiary faunz of Europe and the United States, in a series of approximately ancestral stages, sometimes closer in one re- gion, sometimes in the other, to a group of closely allied primitive peris- sodactyls in the early Eocene of both countries. In South America, the order is unknown until the late Pliocene and Pleistocene. In other re- MATTHEW, CLIMATE AND EVOLUTION 935 gions we know too little of the early Tertiary faune to say when the perissodactyls first appeared, but they are absent from the Oligocene fauna of Egypt, from the Pleistocene and modern faune of Australia and of all oceanic islands. This accords with the natural inference from their size, proportions and habits that they would be strictly limited by land connection in their geographic distribution. Besides the surviving groups, the early perissodactyls gave rise to sev- eral extinct families, the lophiodonts, paleotheres, titanotheres and chali- Tertiary anceslors of Horses Eguidae . 7_-f oe Pes a (nm (ale Lerliary Wo Eguidae until Pleislocene. Fic. 17.—Distribution of Equide, living (solid black) and Pleistocene (shaded) Early Tertiary ancestors are found in Nearetic and Palearctic regions. The American series is more direct than that of Europe until the late Tertiary. This and other con- siderations indicate the center of dispersal as in northeastern Asia or northwestern North America. cotheres, none of which are known to have invaded the southern conti- nents. Equide—The best known phylum of the order, that of the horses, is certainly not a direct genetic succession, as regards known species, but approximately so as regards the known genera. The successive genera are progressively more specialized in accordance with their geological 936 ANNALS NEW YORK ACADEMY OF SCIENCES sequence. ‘They are identical or closely allied in the European and North American sequence. In North America, the series is more complete, the approximation to a direct genetic sequence is closer and the successive stages appear earlier in time. This is reasonably interpreted by suppos- ing that the center of dispersal was intermediate between Europe and the western United States but nearer to the latter. That is to say, it was either in boreal North America or in as vrs Asia. The absence of |_| Sinald 4-Tieet Horsey Synall 5Toed | Larger 5-Toed | Large (Toca Horses _| ae TERTIARY PERIOD (QUATERNARY EOCENE EPOCH OLIGOCENE MIOCENE PLIOCENE |PLEISTOCENE] RECENT —— - Gees - - - - eee and North Americ ) __-!_ 7A qe = _(North America) | | = ma (NortAl America) | A. | | | me EE | North América) | eee | ee ae —— | <——=> (North Abnerica) | eee America)t — — ~ te | | | | | | | | | qe (Biropene t--- C — Hypohippus Ve ce America, Asia and Europe) — — _ (North America) a Se ee ao (North America) a (North America)_ eR aeties Be ees Pee (North America)_\— — ¥ — 22 REE Ee Gh Ne (South Amkrica)-: ee —_— oe fee a kee yee (South ee aed eoy i POET (ee Hipparion (North America , Asia , Europe and North Nafrica) ‘orth and South America, Asia , Europeyand Africa)_ = Sl Fic. 18.—G@eologic range and phylogenetic relations of fossil Bquida | | | | | | | | | | | | | | | The overlap in geologic range of the genera, and the sudden appearance of each new stage, indicate that our record is not derived from the center of dispersal of the race; although the American series is sufficiently direct to indicate that it was not very remote. primitive survivors of the race in the Kast Indies is natural ; as the horses were very early adapted to open plains, unfitted for mountain or forest habitat, the great transverse Himalayan chain would form an almost im- passable barrier and the heavily forested regions of the East Indies would have no attractions to tempt the ancestral horses to pass around its east- ern end. MATTHEW, CLIMATE AND EVOLUTION 937 TasLeE V.—Distribution of the Equide Neotropical Nearctic Palearctic Oriental |Ethiopian Recent Equus Equus Equus Hquus : Hippidion Equus Equus Pe es Pleistocene Onoh ippi- 2 Hipparion Equus (India) Fiquus dion, ete. Hipparion Equus ? Equus Pliohippus (Siwalik ) Pliocene — Mo eis Hipparion Hipparion Parahippus Hypolippus Hypohippus ( Hipparion Pliohippus Protohippus Hipparion Hipparion | Merychippus (L’rSiwalik) | Hypohippus Miocene a \ Merychippus Hypohippus Anchitherium Parahippus Parahippus Anchitherium Miohip; Oligocene fein: ps Mesohippus ae Anchilophus Es Lophiotherium ne Orohippus Pachynolophus Eohippus Hyracotherium 238 ANNALS NEW YORK ACADEMY OF SCIENCES Tapiridea.—The tapirs are the most primitive living perissodactyle, retaining the primitive number of digits in fore and hind feet and the primitive short-crowned grinding teeth. They are to-day limited to the Jast Indies and tropical America. In the Pleistocene, they inhabited the Sonoran region and continental India and the marginal parts of the Palearctic region. Their Tertiary ancestry has been traced back in Eu- rope and in North America to the Oligocene Protapirus, which is pre- ceded by a less direct ancestral series in the Eocene of North America; te lerkary ane lerkiry ances 0 a oe Tapirs * Vo Tapyrs untul Pletslocene Fic. 19.—Distribution of the Tapirs living (solid black) and Pleistocene (shaded) Ancestral types are found in the Tertiary formations of Europe and North America. The relations of the two series and the Pleistocene and modern distribution indicate a dispersal center in eastern Asia. but ancestral tapirs have not been identified in the Eocene of Europe. The data are insufficient to determine the center of dispersal except as probably in the Palearctic region. Tapirs are unknown in South Amer- ica until the Pampean (Pleistocene) ; they do not appear to have reached Africa at all. The arid climate of the Afro-Asiatic connection and the heavily forested path of migration to the East Indies would seem to be the features that determined the dispersal of the horses into Africa, the tapirs into Malaysia. MATTHEW, CLIMATE AND EVOLUTION 939 TABLE VI.—Distribution of the Tapiride Neotropical Nearctic Paleearctic Oriental % Tapirus 8. s. Tapirus Recent Tapirella None None ( Rhinocherus) lei Teoria (al Tee SSE gees 1 Tapirus Pleistocene apirus (s. 1.) apirus (s. 1.) | Tapirus (s. 1.) ( Rhinocherus) Pliocene None u Tupirus (s. 1.) 2 Cp ee Lapis: (8.15) 9 Miocene None Tapiravus Paratapirus d Oligocene None Protapirus Protapirus ? Tsectolophus i Eocene None Helaletes*® Lophiodontidx*! ? Systemodon * True affinities of these genera require revision. 51 Affinities between tapirs and rhinoceroses. 7. o / S PL Ra a Ar ly Tertiary ancerrees Of Rhinuoceroses y RAIROCerOseS| | 12. Midale o la TerTeé rlia ru Hi Pic. 20.—Distribution of the Rhinoceroses, living (solid black) and Pleistocene (shaded) Primitive rhinoceroses are found in the Palearctic and Nearctic Tertiaries and late Tertiary of India and Africa. Comparison of the Palearctic and-Nearctie series indi- cates that the center of dispersal was in west-central or southwestern Asia. 240 ANNALS NEW YORK ACADEMY OF SCIENCES Rhinocerotida.—The rhinoceroses are intermediate between horses and tapirs in adaptation. ‘The Tertiary history of the group is much the same, approximate series being found in Europe and North America as far back as the Oligocene or Eocene, but the phyla are less direct and complete, and there is a greater diversity of type among them. The Palearctic series appear to be more direct, and this, in connection with the fact that the race never reached South America, may be taken to indi- cate that the center of dispersal was Palearctic rather than Nearctic, less northerly than that of the horses, less easterly than that of the tapirs. At all events, the relations of the later Tertiary rhinoceroses indicate that North America was much more remote from the center of dispersal than Europe, while southwestern Asia was very close to it. TasLteE VII.—Distribution of Rhinoceroses Nearctic Palearctic | Oriental Ethiopian Rhinoceros Ceelodonta Non None : aa8 eeu Soe Ceratorhinus Opsiceros Elasmotherium : Celodonta Rhinoceros ‘ a) Ui “1 2 = stocen None . : psiceros Pleistocene None Opsiceros Opsiceros Opsiceros Ceratorhinus aisd | Opsiceros Rhinoceros Pliocene Teleoceras Ceratorhinus Teleoceras ? Teleoceras Aceratherium | Teleoceras Teleoceras Miocene Aphelops _ Aceratherium Rhinoceroses | Rhinoceroses Diceratherium | Diceratherium | I-- Ceenopus _ Aceratherium* Dy, Wand s02 a) » T aac : aus nO or th, ac- Oligocene Prigonias Canopus | Diceratherium | Nd Perissodac S Metamynodon Prohyracodon Teleoceras tyla Hyracodon | Cadurcotherium Y Amynodon | focene Triplopus ; ? Lophio|dontidee® 52 Includes a number of subgenera recently defined by Abel. 53'This family may be regarded as ancestral to both rhinoceroses and tapirs, but the more exact derivation is doubtful. 54 Gaj fauna, upper Aquitanian auct. Pilgrim. It should perhaps be regarded as Lower Miocene. MATTHEW, CLIMATE AND EVOLUTION 941. ARTIODACTYLA The great and diverse order of artiodactyla can fairly be regarded as of Holarctic origin as a whole. Its distribution can most readily be con- sidered group by group. Pigs and Peccaries.—These two groups are characteristic of the Old and New World respectively. The pigs are now chiefly Ethiopian and Oriental, the peccaries Neotropical in distribution. The peccaries first reached South America in the Pleistocene and ranged throughout the 6a uced by early me Terliary anceslors Of true Pgs Tertiary ancesfors of peccarces No peccaries terlil Lletstocene Fic. 21.—Distribution of pigs and peccaries In Old World, broken shading Sus only; full shading, other genera. In New World, full shading Dicotyles. The dispersal center of Dicotylide was Nearctic, of Suide Pale- arctic. The living South American genus is more primitive than the Pleistocene genera of North America, Platygonus and Mylohyus (the Pleistocene North American species referred to Dicotyles are all Mylohyus). United States from the Oligocene to as late as the Pleistocene. Pigs were common in the Oligocene and later Tertiary in Europe and were present in India in the Miocene, probably earlier. The Tertiary ancestry of the pigs in Europe can be traced back to a common ancestral group in the Eocene, and the same is true of the peccaries in the western United States. 949 ANNALS NEW YORK ACADEMY OF SCIENCES TaBLeE VITI.—Distribution of the Pigs and Peccaries Neotropical | Nearctic | Palzearctic | Ethiopian | Oriental | ees Potamoche- | Sus . 1 ; rus | ec 20LY LES us . | ° Recent Dicotytes Sus Phacoche- | Babirussa rus ? Dicotyles Pleistocene | Dicotyles Mylohyus Sus Platygonus ? Platygonus | Sus, Hippohyus Pliocene ? Prosthen- | Platygonus | Sus | Samitherium nOps | Potamocherus | Listriodon Prosthen- Sus ? Sus, Hippo- . nops : LYUSs Miocene None v (No record) hy ; Potamocherus Desmathyus | Listriodon Hyotherium ete. Palzocherus Paleocherus Oligocene None Percherus | Hyotherium | None®* Palxocheerus® etc. ?Helohyus | Cebocheru - Eocene None ue oe ne s (No record) Ruminants.—Under this term, we may conveniently include all the selenodont artiodactyls,—the camels and tragulines, deer, antelopes, sheep and cattle, besides various extinct groups. They are admittedly of Northern origin. In South America, they do not appear until the end of the Tertiary (Wicrotraqulus, Monte Her- moso) ; their representatives in the Oligocene of North Africa are much more primitive than the contemporary artiodactyls of Europe; the high- est and most progressive types are found to-day in Asia, and the most antique and primitive survivals in the East Indies, West Africa and trop- ical America. The several groups indicate in their present distribution, and in what is known of their past history, that their centers of dispersal 55 Schlosser has shown that Geniohyus is a Hyracoid, not an Artiodactyl. °° Gaj fauna, regarded by Pilgrim as upper Aquitanian. MATTHEW, CLIMATE AND EVOLUTION 942 were in different parts of the northern world, as we have seen among the Perissodactyl groups. The camels appear to have been of American origin. An ancestral series is found in the Tertiary of the western United States, going as far back as the Upper Eocene.** In the Old World, they first appear in the Phocene; in South America, in the Pleistocene (Pampean) ; and the Lerliary Anceslr of the Cametltaae { until Pleislocene Fie. 22.—The dispersal center of the Camelide was in North America They reached the Old World in the Pliocene, South America in the Pleistocene. They survive on the margins of their range but became extinct in North America early in the Pleistocene. North American Pleistocene camels were more advanced than the living types of the marginal areas. camels of the Pleistocene in North America were about as specialized on the whole as the living llamas of South America or the camels of Africa and Asia. In North America, the race is now extinct. The center of dispersal would appear to have been in this continent,—how far to the north we have no means of estimating; but the exceptional directness of the phylogenetic series as represented by our western fossils indicates, in my opinion, that these fossils lived in or close to the racial dispersal center. 57 It forms a singularly direct and complete phylum, so supercharged with intermediate and connecting forms that it is very difficult to classify and arrange the fossils into species and genera, while every gradation of structural evolution is abundantly illus- trated. 244 ANNALS NEW YORK ACADEMY OF SCIENCES TaBLE 1X.—Distribution of the Camels Neotropical Nearctic Paliearetic | Ethiopian Oriental Recent Auchenia None Camelus None None Bschatius Camelus Pleistocene | Auchenia Camelops ? Procam- te Camelus Camelus elus °° Miaucheni 9 Parac c Pliccane None Pliauchenia Lee anacam ? Camelus ® ete. elus —_—__—_— (No record) = Procamelus a is Protolabis = = Miocene None Miolabis ete. None None Oxrydactylus Protomeryx : - ete. = r r one F i . None No recoré Oligocene None | Poébrotherium None None (No record) | Hotylopus Hocene None Protylopus None (No record) The tragulines, recent and extinct, are a heterogeneous assemblage of primitive ruminants, whose real affinities have been much disputed. In the present writer’s opinion, the living East Indian cheyrotains should be associated with Hypertragulus of the North American Oligocene and perhaps Microtragulus of the South American Pliocene, and the center of distribution of this group hypothetically placed along the northeastern coast region of Asia (cf. tapirs). The living water-chevrotain (Hye- moschus) and most of the so-called tragulines of the European Oligocene and Upper Eocene are to be regarded as primitive stages of true Pecora. Leptomeryx, Protoceras and Heteromeryx are related forms from the North American Oligocene. Among these primitive forms, some (Lep- fomeryx) display affinities to the deer, others (Protoceras, Heteromeryx) to giraffes and antelopes. 580, sivalensis of the Siwalik beds is doubtfully congenerie with the modern species and, along with the so-called Procamelus described by Mme. Payvlow from the Pleistocene of Russia, appears to be an intermediate stage between Procamelus and Camelus. 59 A doubtful Camelid, based on a single upper molar from the Pliocene (or Miocene) of China. MATTHEW, CLIMATE AND EVOLUTION eo i SDT In the later Oligocene of Europe and the Miocene of the United States appear more definitely deer-like types (Dremotherium, Blastomeryz), and in the succeeding formations we find progressively higher types of deer in Europe and North America, but always appearing earlier in the Old World. The deer—-excepting the isolated primitive survival repre- sented by the “water-chevrotain,” closely related to Dorcatherium, a Miocene genus in Europe—have not reached the Ethiopian region, but ic Leplomeryx in Dorcatheriumy = mid-Lerliary ; Cervi mid-Terliary; Cervidae~ tn laler Tertiary A tn later Terhiary Re 1) No Cervidae = By 9 moschus = until Plecs locene cf Dorcalh?m) == Rangifer, Alces, Cervus canadensis and allies YS Other Cervidae , primilive lypes mostly in Tropes Fic. 23.—Distribution of Cervide and pro-Cervid Tragulide The highest and latest appearing types are still confined to the circum-Arctic regions ; the genera of the more peripheral regions are more primitive. The earliest and most direct ancestral series is found in Europe and Asia; the parallel series in North America is less direct and more retarded. A primitive survival is found in West Africa, pro- tected by the desert from competition of higher types. were easily able to reach North America in the Pleistocene. I take it, therefore, that their center of dispersal was well to the east and north in Asia (cf. horses). Their migration into the Ethiopian region was checked after the Miocene by the progressive aridity of the desert region between, 246 ANNALS NEW YORK ACADEMY OF SCIENCES which served as a barrier to these forest-living ruminants, although not to the plains-living antelopes. TABLE X.—Distribution of Tragulide proper Recent Pleistocene Pliocene Miocene Oligocene Eocene Neotropical Nearctic Palearctic | Ethiopian Oriental None None None None Tragulus None None None | None Tragulus ?? Microtrag- E AS None None Tragulus None None ? None None [Hypertrag-| 57,1, None ulus] ? None None None None [Primitive Artiodactyla ]| MATTHEW, CLIMATE AND EVOLUTION 947 TABLE XI.—Distribution of Cervide and Pro-Cervid Tragulines Neotropical | Nearctic | Paleearctic | Ethiopian Oriental Cervus, Cervus, Hyemos- R t Odocoileus Alces Alces chus Cervus eee! Mazama Rangifer Rangifer, (W. Afri- (sensu lato) Odocoileus Dama eae Cervus, eee : Odocoileus Alces vets Tene Pleistocene Tigenan Rangifer UN : Cervus (s. 1.) Odocoileus Megaceros Cervus (s. 1.) ‘ Cervus Cervus Lae Moschus sinveene None (Sale) (Sil) mete: (No record) Doreathe- rium Bee £ eS N 1) Dorcathe- i J 0 recorc “ium & Btiocene pone Blasto- Dremothe- ( aes meryxr rium Lepto- ioe a Prodremothe- Oligocene None merys | eer 5 ie None rium ete. GeOCus Gelocus ete. ete. Eocene None Primitive Artiodactyla No record The antelopes, on the other hand, while also appearing fairly early in the European geologic record and abundant and well advanced in south- west and southern Asia as early as that record is revealed to our eyes, are imperfectly represented in North America—first appearing in the Plio- cene and not widely varied even to-day, while they have not reached South America at all. They are to-day most abundant and varied in Africa. From these facts, I infer that their center of dispersal was well to the 6 Family is structurally ancestral to American Cervide. % This group is referred generally to the Tragulide, but the common characters are persistent primitive features, and I regard it as a little altered survivor of the ancestors of the Cervide. Tragulide as here limited are a distinct phylum, primitive in many~ features but aberrant in others. ® Family Tragulide as usually referred, but affinities are with Hyemoschus, not with Tragulus; the group may fairly be regarded as ancestral to the Cervidwe, while the traguline group certainly is not. Hypertragulide, but Leptomeryr 248 INNALS NEW YORK ACADEMY OF SCIENCES C. Other Antelopes (Antelopine etc.) D. Cattle (Bovine) Fig. 24.—Distribution of the Bovide, existing (solid black) and extinct (shaded) The sheep and goats are regarded as the highest group; the muskoxen represent a specialized Arctic adaptation (cf. Eskimo among mankind). ‘The cattle are a somewhat southerly type; their formerly wide northern distributfon has been greatly restricted, and for the theory that they are of Oriental origin there does not appear to be any real evidence. The remaining Bovid subfamilies, usually grouped under the term “‘antelopes,”’ are to a varying extent primitive and aberrant. The Holarctic groups are nearer to the sheep and goats and the more primitive groups are limited to the Ethiopian region and the East Indies. MATTHEW, CLIMATE AND EVOLUTION 249 west and south in Asia (cf. rhinoceroses). The sheep and goats are a comparatively recent development of the highest antelopes and must be assigned a center of dispersal somewhat more to the north. The cattle’ are of comparatively recent appearance in Europe, as also in America. Judging from their present distribution, one would say that their center of dispersal was in southeastern Asia, the southward slopes of the Himalayas. Fic. 25.—Distribution of the Giraffes, existing (solid black) and extinct (shaded) On present evidence their dispersal center would appear to have been in south central Asia. But the affinities of the Tertiary Giraffide to other contemporary ruminants need careful and judicial reconsideration. antelopes. This inference from their modern distribution conforms with the geological record. They appear suddenly in the upper Miocene of Europe, but an ancestral series is found in India as far back as the upper Oligocene.** The giraffes have approximately the same center of dispersal as the 83 See G. E. PILGRIM: Rec. Geol. Sur. Ind., vol. xliii, p. 301. 1913. 250 ANNALS NEW YORK ACADEMY OF SCIENCES TABLE XII.~—Distribution of Bovide and Antilocapride | Neotropical | Nearctic Paleearctic Ethiopian Oriental Ovis Sheep and Cattle oy Spann Oreamnus Goats !! e Cattle !! Recent None Bison Cattle * ae git | Antelopes! | Antilocapra | Antelopes Rein ' Sheep and Bison aa 7 ae Goats Cattle Cattle Pleistocene | NOE etka Cattle! Antelopes | Antelopes B Antelopes! Cattle! ! ae eer [Meryco- Cattle : Pliocene oe dus |" Antelopes !! ce lopes!! 2 —_—_——__—— (No record) | pak Antelopes!! Cattle!! Miocene None Ease aye % (late Mio- Ante- | a cene) lopes!! | [Ancestral Oligocene | None None Primitive None None Ruminants | 6 Merycodus is a distant relative, combining characters of Bovide and Cervide. MATTHEW, CLIMATE AND EVOLUTION 251 TABLE XIII.—Distribution of Girafide Nearctic Palearctic Ethiopian Oriental = Giraffa | Recent None None Ocapia Pleistocene None None Sivatherium i None (unless Hydaspithe- 9 “ Pliocene ? None in China) (No record) AP Giraffa ete. Felladotherium ? Giraffa Miocene ? None Samotherium (Norecord) | Progiraffa etc. ete. [Syndyoceras | Ancestral ree miley Oligocene and Proto- Primitive None : phe @inat ceras © Ruminants fide) i 6& Remote and archaic collateral relatives, family Protoceratide. eertain that Dromomeryx and other undescribed It is by no means genera from the North American Miocene provisionally referred to the Cervide and Brachyodont Bovide are not related to the Giraffide ; but on present evidence the dispersal center of the family appears to be India, and their range confined to Palwogea. 252 ANNALS NEW YORK ACADEMY OF SCIENCES Besides these surviving groups of ruminants, there are several groups which have not survived. The anthracotheres are one of the earliest of these specialized races; I have elsewhere®® detailed the data upon which may be predicated a North Asiatic center of dispersal for this group. The living hippopotami show a modicum of resemblances to this type, which may mean that they are derived from some early members of it. Their present habitat is Ethiopian; but inthe Pliocene and Pleistocene their range was far to the northward—even as far as England on one OX Anthracotheres We a Z- Terliary LEG flippopolame / wa Jiocene ana Ea ay Modern Hpppopolamé Fic. 26.—Distribution of the Anthracotheres and Hippopotami The Anthracotheres were a large and widely dispersed group in the Oligocene and Miocene, especially in the Old World, but found also in the Oligocene of North America. The Hippopotami appear to be specialized survivors from the same stock; they are con- fined to the Old World and their range has been greatly restricted since Pliocene and Pleistocene. hand and northern India on the other. While the present distribution of the large hippopotamus is Central Africa, smaller and more primitive precursors have been stranded on the one side in West Africa, on the other (now extinct) in Madagascar and also found refuge in the Mediter- ranean islands until the Pleistocene. (The aquatic habits of the hippo- 6 Bull. A. M. N. H., vol. xxvi, pp. 1-7. 1909. MATTHEW, CLIMATE AND EVOLUTION 253 potamus have enabled it to reach these island retreats more easily than terrestrial competitors. ) TaBLE XIV.—Distribution of Anthracotheres and Hippotami Nearctic Palearctic | Ethiopian Oriental Malagasy Hippopota- Recent None None Mus None None Cheropsis Hippopota- Hippopota- Pleistocene | None Choro - Hippopota- Miss wie OP mus (dwarf ( Gupcns) species) Hippopotamus * Hippopota- : Hexaprotodon Pliocene None aie (No record) | 4, erycopota- MUS (No record) Anthraco- therium Miocene Arretothe- | Brachyo- | Anthraco- | pomimerye rium dus theres Sivamerysx ete. Merycops Anthraco- ie ob Aes 8 Anthraco- therium “Brachyo- Oligocene therium | Ancodus Ancodus dus”? (No record) Ancodus “Brachyo- Anthraco- dus” therium Eocene None Ancodus (No record) The remaining groups of ruminants are not of especial interest in this The entelodonts are Holarctic; the oreodonts Nearctic; anoplotheres and cenotheres Palearctic; there is no evidence that they originated elsewhere or that they reached any other zodlogical region. Entelodon (sensu lato) appears simultaneously in Europe and the United discussion. 87H. minutus is (fide Bate) congeneric with the Liberian species. The rules of pri- ority call for the application of Hyopotamus Kaup to this genus, instead of Chwropsis. 254 ANNALS NEW YORK ACADEMY OF SCIENCES States in the beginning of the Oligocene, without direct ancestry in either continent, and is regarded by Peterson®* as probably from an Asiatic source. NON- RUMINANT PECCARIE 0 CSS PLEISTOCENE PLIOCENE . Ni, é ara ratatatat i wy ry AA MIOCENE OLIGOCENE YD Ws, 1% Wei 0.9 ,y, iX? ay x mY Uff LLL X\ NW Wy (5 Mi ¢ OF THE EOCENE - HIGHER GROUPS OF RUMINANTS PALAARCTIC ETHIOPIANKORIENTAL Fig. 27.—Phylogeny and distribution of the Artiodactyla YR “ Most of the families appear to have originated in the Nearetiec or Palearctic region and spread thence outwardly to the more peripheral regions. The higher types are of more recent origin and are still dominant in the Holarctica. PROBOSCIDEA The later Tertiary and Quaternary history of the mastodons and ele- phants agrees with the various groups that we have been considering in indicating Asia as the center of distribution of the race. Elephants are now limited to the Ethiopian and Oriental regions, but in the Pleisto- cene their range was over the whole of Europe, Asia and North America, as well as Africa. The northern species, although of smaller size, are more progressive than the southern species in the specialization of the teeth, proportionate length of tusks, shortening of skull with concomitant elongation of trunk. The more primitive mastodons first appear in India in the Oligocene, in Europe in the lower Miocene, in North America in the middle Miocene. The intermediate stages leading to the mammoths and elephants are best shown in the Pliocene and Pleistocene of India; a less exact series may be found in North America. The mastodons reached South America in the Pleistocene; the mammoths and elephants never reached that continent. The earlier stages in the phylogeny of the Pro- 6 QO, A. PETERSON: Mem. Carn. Mus., vol. iv, pp. 145-148. 1909. MATTHEW, CLIMATE AND EVOLUTION 955 boscidea have not, however, been found either in Europe or North Amer- ica but have been recognized in the Oligocene of Egypt. From this fact, it has been generally concluded that the Proboscidea first evolved in the Ethiopian region. But it should be remembered that northern Egypt is not strictly within the Ethiopian region but belongs with all of northern Africa to the Mediterranean subregion of Holarctica. Owing to its prox- imity to the Ethiopian region, it contains Ethiopian elements in its mod- ern fauna and may have contained more in the past. But it is not clear that the Oligocene Proboscidea must be numbered among these. There is no evidence that their center of dispersal was not Asiatic in early as in later Tertiary ;°° but it must have been too far to the south to admit of their reaching Europe or North America, until after their spread into northeast Africa. We must therefore conclude, apparently, that the dis- persal center was transferred to the north and east during the course of the Tertiary—a quite exceptional feature, beside which the question of its original location, whether in southern Asia or in Africa, appears much less important. TABLE XV.—Distribution of the Proboscidea Neotropical | Nearctic Palearctic | Ethiopian Oriental Recent None None None Loxodon Elephas Hlephas : : Elephas Hlephas : Pleistocene | Dibelodon jeetianien, | aepetracien ? ene 2 Hlephas Pliocene None Dibelodon a Wate (No record) | Stegodon lophodon Trilopho- is . : Tetralophodon Miocene None nee ho D ie oe D mone Trilophodon = Dinotherium rium Paleomas- | Heniimasto- F todon don Oligocene | None Neue Merithe- | Dinotherium rium ? Meritherium Merithe- Eocene None None None Pi (No record) ® Certainly the Proboscidea of the Oligocene Gaj fauna of India are far more ad- vanced than the Egyptian Fayfim genera, if Pilgrim’s correlation of the Gaj beds is cor- rect. This, by our methods of interpretation, would indicate that India was much nearer than Egypt to the dispersal center. 256 ANNALS NEW YORK ACADEMY OF SCIENCES SIRENTA The most primitive sirenians are found in the late Eocene of Egypt. As these were apparently contemporary with more progressive types in the Middle and Upper Eocene and Oligocene of Europe, they indicate, if anything, that the Mediterranean shores hald a less progressive fauna than the North Atlantic. The Oligocene and Miocene types are approxi- mately ancestral to both the modern groups, manatees and dugongs. Ap- parently the manatees became characteristic of the North Atlantic, the dugongs of the Indian Ocean shores. The progressive cold of the later Tertiary and Pleistocene has driven the manatees out of the Arctic and northerly Atlantic shores and their northern limit is now Florida on the western, and the African coast on the eastern side. They have not been found fossil north of 40° N. lat. on the American coast,’® for the excel- lent reason that there are practically no Tertiary littoral deposits north of that latitude. The occurrence of Manatus in West Africa and in the West Indian and South American coasts is among the arguments used in support of a transatlantic bridge; but there is no evidence at all that the ancestors of Manatus did not inhabit the whole of the North Atlantic and Arctic basin during the Tertiary. It is certain that they did inhabit parts of the intervening European and American littoral, and the negative evi- dence elsewhere is obviously worthless, because there are no formations known in which they might be found. CONDYLARTHRA AND SPECIALIZED SUCCESSORS We may here consider the distribution of a number of extinct groups of Tertiary ungulates or semi-ungulates, whose rise and culmination took place at an earlier epoch and under different conditions from those which we have discussed. ‘The Condylarthra are an extremely primitive group of hoofed mammals, fulfilling nearly the theoretical requirements for the common ancestral type of all placental ungulates. The earliest known artiodactyls and perissodactyls are, however, too much specialized to be immediately derived from the known Condylarthra. Condylarths first appear in the Paleocene of North America and Europe and in South America in the Notostylops fauna, here regarded as Eocene. In North America, they develop through the Taligrada into the Amblypoda, culmi- 7 Yor distribution of manatees during Tertiary vide Hay, Bibl. Foss. Vert. N. A., U. 8. G. S. Bull. 179, p. 583-4, 1902; of Old World Sirenians, Abel, 1904. Abh. Geol. Reich- sanst., xix Bd., s. 214; 1906, Neues Jahrb. Bd. ii, s. 50-60; 1912, Palwontographica, lix Bd., s. 292. MATTHEW, CLIMATE AND EVOLUTION 257 nating in the highly specialized Dinocerata. In South America, they apparently develop during the Tertiary in absence of Artiodactyla and Perissodactyla into a great variety of hoofed mammals, the Toxodontia and Typotheria, Litopterna, Astrapotheria, Pyrotheria. The Arsinoi- theria of the Oligocene of Africa, perhaps also the Hyracoidea and Probo- scidea, may also be regarded as evolved from primitive Condylarthra, in absence of the higher ungulates of the Asiatic center of dispersal. We have therefore direct or inferential evidence that at the beginning of the Eocene the Condylarthra inhabited the Palzarctic, Nearctic, Neotropical and Ethiopian regions. There is no reason to suppose that they were ueeat NOTOUNGULATE Bae x z (| Crelacéc BU NOTHE RIA RQ Palaearciic. GG, Wearelic. — Meotropicel|I| [III] Ethiopian and Oriental Fic. 28.—Relationship of the Condylarthra to the Notoungulate and Subungulate groups of hoofed mammals In indicating the distribution, Egypt, Syria ete. have been included with Ethiopia, as the essential facts in this case could thus best be represented. ‘“‘Bunotheria”’ are the common ancestral stock (hypothetical) of the Creodonta-Carnivora-Condylarthra-Ambly- poda group. absent from the Oriental region, but they evidently did not reach Aus- tralia or Madagascar. The worldwide dispersal of the condylarths at the opening of the Ter- tiary (partly hypothetical and exclusive of Australasia and Madagascar) may be regarded as due to the epoch of elevation and disturbance which closed the Cretaceous. The subsequent development of peculiar and highly specialized ungulates during the Eocene in the several great con- tinents is attributable partly to the isolation of these continents during that period due to submergence of the low lying connecting regions, 258 ANNALS NEW YORK ACADEMY OF SCIENCES partly to the prevalence of more uniform climatic conditions all over the world and the consequent lack of environmental pressure tending to force a change in habitat. Towards the end of the Eocene began a period of progressively intensified elevation and disturbance, with re- frigeration of climate beginning at the poles; this culminated in the Glacial epoch. ‘The northern fauna suecessively invaded the tropical and southern continents and swept before it ‘nearly all their autochthonic faune. In Africa, we see this invasion in progress in the Oligocene; the anthracotheres, forerunners of the great ruminant invasion have already appeared; to these may yet have to be added Palwomastodon as a fore- runner of proboscidean invaders (although on the present record the Proboscidea may appear an autochthonic group); while the hyracoids, with Meritherium, Arsinottherium, Barytherium and some less known types are apparently autochthonic since Paleocene. Unfortunately, our view stops here; we know little of the progress of this invasion until the late Pliocene, when these invaders had themselves disappeared before a succession of later invasions or become modified into new types. In South America, the isolation lasted much longer, and owing to the great southward extension of the continent, a highly progressive inde- pendent center of dispersal was set up in Argentina. Whatever criti- cisms may be made of the phyletic theories of Dr. Ameghino, so far as they affect the evolution of the mammalian races of the northern world, I think that there can be no question that he has brought out a remark- ably complete series of phyla in the autochthonic races of South America. The closeness of these series, and the large amount of progressive evolu- tion which they involve, on lines analogous to those of the northern mammals, are fair indices that the controlling forces were similar and that the southern end of the continent was the chief center of dispersal. The various types of structure which were developed in northern mam- mals during the Tertiary, in adaptation to the progressive change of environment, are almost all paralleled, occasionally exceeded in degree by these southern races; but they are very generally seen in different combinations, as Professor Gaudry has so clearly shown.™4 Had the Condylarthra reached Australia, we should expect to find there a group of placental ungulate orders peculiar to the region, like those of Tertiary South America, persisting to the present day. But we find, instead, that the marsupials evolved into the herbivorous fauna. In Madagascar the lemurs may be regarded as filling the place which T ALBERT GAUDRY: Annales de Paléont., t. ill, pp. 41-60. 1908. MATTHEW, CLIMATE AND EVOLUTION 259 primitive ungulates would have taken, if they had reached the island; but the case is not so clear. EDENTATA The edentate orders afford among the unguiculates a broad parallel in their distribution and history to the Condylarthra and their successors among the ungulates. Their extinction has been somewhat less complete ; a few highly specialized survivors remain in the Neotropical, Ethiopian and Oriental regions. Gly plodonis and Ground-sloths a Tee [Eas ——— Ss Pholitola andd Lubulidentata Dis pe Xenarthra throughout Ter lary Wi, Xenarthra INS Photidola {II | Tubulidentata Fic. 29.—Distribution of the Edentate orders The New World edentates or Xenarthra may have originated in Cretaceous North America, but their Tertiary dispersal centers were South American, apparently in or near to Patagonia. The dispersal centers of the Pholidota and Tubulidentata would appear to have been Palearctic, but very little is known of their fossil record. The super-order Edentata is an artificial assemblage including the three surviving orders Xenarthra, Pholidota and Tubulidenta and the extinct order Tzeniodonta (= Ganodonta). The Teniodonta of the Ho- cene of North America may perhaps be regarded in a broad way as rep- resenting the primary type of the Xenarthra, but even this is doubtful. 260 ANNALS NEW YORK ACADEMY OF SCIENCES They are far more primitive and nearer to the generalized eutherian type; but they show certain unique Xenarthran peculiarities in foot- construction and in the pelvis, and the dentition in the two known phyla progressively evolves on lines leading towards, although not into, the PHYLOGENY OF he: aD ENT Ais ARMADILLOS ANTEATERS _TREE-SLOTHS i Wn ae RR We ‘hd (mad BN a GLYPTO nts kone LOTHS \ ‘S see \ FER KES, \ ‘y D \\ CaN x y PLY Nearclic North American) , \\\\ ) Weolropical (South American) CRETACIC Unknowte Fie. 30.—Distribution and phylogeny of Xenarthra and Teniodonta The aberrant North American groups appear to be relicts indicating a northern origin of the Xenarthra, but the evidence is not conclusive. MATTHEW, CLIMATE AND EVOLUTION 961 specialized edentate types. The xeniodonta range from Paleocene to Upper Eocene in North America and are doubtfully recorded in the early Eocene of Europe. They may be hypothetically regarded as a Cretaceous-Eocene ancestral group in the northern world, from whose early members budded off the ancestral Xenarthra in the Nearctic, pos- sibly also the Pholidota and Tubulidentata in the Palearctic, the whole group being driven southward at the beginning of the Tertiary, except for a few lingering remnants, rare and little known. Of these lingerers, we may instance in the (Bridger) mid-Eocene of Wyoming Metacheir- omys, whose affinities are distinctly armadilloid and an unnamed but more primitive genus in the Lower Eocene of Wyoming approximately ancestral to it; “Lutra” franconica of the Oligocene of Germany, shown by Schlosser to be related to the Aardvark, Palwomanis and Orycteropus of the Miocene of Samos, and more doubtfully Palworycteropus and Necrodasypus (in part) of the Oligocene of France. Whether the rare ground-sloth remains from the (?) Middle Miocene’ and Lower Pliocene of the western United States are to be regarded as surviving Northern edentates or as immigrants from the south is not certain, but the latter explanation is more probable. The Old World edentate groups, although still surviving in Ethiopia (Manis, Orycteropus) and the East Indies (Manis), are not known to have undergone any considerable expansion during the isolation-period of the early Tertiary.7** The Xenarthra, on the other hand, are first represented in the early Tertiary of South America by armadilloid forms and they blossomed out in the isolated continental conditions that pre- vailed during the Tertiary in that continent into a wide range and diversity of type, just as the Condylarthra appear to have done under the same conditions there and the marsupials in Australia. Of the five principal groups—tree-sloths, ground-sloths, anteaters, armadillos and glyptodonts—only the second, fourth and fifth are known as fossils, and only the first, third and fourth have survived. The fossil groups reached their maximum of size and specialization in the Pleistocene, and rein- vaded North America in the Pliocene and Pleistocene (possibly earlier, “= There is some question as to the true horizon of the ground-sloth claw found by Sinclair in the Mascall formation (Middle Miocene) of Oregon. The specimen may have washed down from the overlying Rattlesnake Beds, Lower Pliocene [oral communication from J. C. Merriam]. 73 But this may be due only to the imperfection of the geologic record. We know nothing of the early Tertiary faune of the Ethiopian and Oriental regions, save for the Oligocene of Egypt. The Eocene faune of South Africa, India and the Bast Indies may have included a considerable expansion of pholidate or tubulidentate mammals, corre- sponding to the xenarthral expansion of the New World, but earlier extinguished because of the earlier invasion of those regions from the north. 962 ANNALS NEW YORK ACADEMY OF SCIENCES vide supra), but only the armadillos have maintained any foothold in the northern world until modern times and these only in the southwest corner of the Sonoran region. ‘The anteaters and tree-sloths might be expected to have originated in Patagonia and to have been driven north- ward to tropical South America in accord with the theory of climate and evolution here advocated. The geological record, however, has failed to show any certain evidence of this, and, as the Patagonian record is a com- paratively full one, this fact should be counted as evidence that climatic change is not the only causal factor of evolution. We must suppose, if the record be adequate, that these groups originated and evolved in tropi- cal South America. The armadillos are an extremely persistent group, and the record gives no really convincing evidence of a Patagonian dis- persal center, although it might be so interpreted. Glyptodonts and ground-sloths appear in the Pliocene and Pleistocene of North America. The Pleistocene genera except Megalonyzx are closely allied to the genera of the Pampean formation, in part identical there- with (Brachyostracon, ? Glyptodon, Chlamydotherium, Megatherium, Megalonyx, Nothrothervum, Mylodon). These, or allied genera equiva- lent in specialization, inhabited South America from Ecuador to Pata- gonia in the late Pliocene and Pleistocene. The only genera found in the Pliocene of North America are Megalonyx and Glyptotherium, decidedly more primitive and are best interpreted as earlier forerunners of the main invasion which appeared at the beginning of the Pleistocene. Mylodon has been recorded from the Blanco beds of Texas, but this is an error. MARSUPIALIA Marsupials are at present almost limited to the Australian and Austro- malayan region, where, in the absence of placental mammals, they have diversified into a wide variety of size, habits and adaptation, paralleling the adaptive radiation of the higher mammals in the northern continents. A single unspecialized group, the opossums, representing quite nearly the primitive type from which all marsupials are derivable, survives in the Neotropical region, one or two of its species ranging northward into the Sonoran subregion of Holarctica. Another primitive survivor in the Neo- tropical region is the rare little Cenolestes, formerly regarded as a primi- tive member of the diprotodont marsupials, but now considered to be of polyprotodont affinities, its diprotodont resemblances being due to paral- lelism. What we know of the paleontology of the order is in complete accord with the theory of their being primarily of northern origin, their dispersal preceding that of the early placentals. MATTHEW, CLIMATE AND EVOLUTION 263 The fragmentary and little known mammals from the Mesozoic for- mations of Europe and North America were in large part marsupials, so far as we can judge from what is known of them. The most distinctive group among them were Multituberculata or Allotheria. Gidley’ has recently (1909) brought forward strong evi- dence for the view that these animals were an archaic, early specialized branch of the marsupials paralleling the later diprotodonts.7* They occur (doubtfully) in the Rheetic of Germany, certainly in the Upper lerliary Opossums probabl: Opossums Since Crelgced S3 : Borhyaenizs unrecerded /72 Pha (29 gers: Caenolestias lalér Ter lary only Opossums i'n early Terliar Thylactnus MAustralia (7 =. Pleistocene :) —. —— — 7 — = =. a tify Qpossiiis, = Caenolestes SSN Dasyures;Diprotodontia Fic. 31.—Distribution of Marsupials This is probably to be regarded as due to a very ancient dispersal from the north, fol- lowed by differentiation and dispersal during the Tertiary of specialized adaptations parallel in the Neotropical (Borhyenids and Cenolestids) and Australian regions (Thy- lacine-Dasyures and Diprotodonta). ‘The Phalangers of the Austromalayan islands are regarded as marginal types from an Australian dispersal center. 74 J. W. GIDLEY: Proc. U. S. Nat. Mus., vol. xxxvi, pp. 611-626. 1909. 7 Recent discoveries, made since these lines were written, indicate that the relationship was not as close as had appeared. Dr. Broom has even maintained that these animals were nearer to monotremes than to marsupials, but in my judgment he has failed to adduce any really valid evidence for this view. But while they are in the Metatherian stage of evolution I do not think they can be included in the order Marsupialia on the data now available. See forthcoming article by Walter Granger in Bulletin Am. Mus. Nat. Hist. 964. ANNALS NEW YORK ACADEMY OF SCIENCES Jurassic and Lower Cretaceous of England (Plagiaulaz™*) and Wy- oming (Ctenacodon). They again appear in the uppermost Cretaceous and Paleocene of North America (Lance formation of Wyoming, Fort Union of Montana, Puerco and Torrejon of New Mexico) and Europe (Cernaysian) in the genera Ptilodus, Neoplagiaulax, Polymastodon and Meniscoéssus. They are questionably recorded in the Eocene Notostylops beds of Patagonia, in the genera Propolymastodon, Polydolops ete. (which more probably belong to the same group as Cwnolestes). They are not known elsewhere except for part of a jaw from the middle Cretaceous (Belly River) of Canada and a jaw (Karroomys) from the Jurassic (Karroo beds) of South Africa. The front half of a skull long ago found in the Karroo beds and described as T’ritylodon is probably to be referred to this group, although its mammalian nature has been questioned.** In addition to the Multituberculata, there are in the Jurassic and basal Cretaceous of England and Wyoming a number of mammals with simpler and more numerous teeth whose affinities are very uncertain. Whether they are ancestral to marsupials, to placentals, to both or to neither is, in the writer’s opinion, an unsettled question. Its definite solution must probably await the discovery of more complete material. In the uppermost Cretaceous (Lance formation) of Wyoming are found in addition to teeth and jaw fragments of Multituberculata, a variety of tritubercular teeth, some associated with fragments of the jaw. These appear to be more definitely referable to the polyprotodont marsupials ; some of them may be quite near to the opossum. I have seen no evi- dence among them of placental mammals, although most of them are too fragmentary to exclude the possibility of the presence of Eutheria. The Paleocene fauna of New Mexico, Montana and France contains numerous placentals and a few Multituberculata, but no polyprotodont or true diprodont marsupials have yet been positively recognized in it. It is evidently not derived (except for certain of the Multituberculata) from the fauna of the Lance formation. Yet it is almost, perhaps quite, con- temporaneous with it and must be supposed to represent a distinct facies of the fauna, differing in habitat from that of the Lance formation (the Fort Union is partly intermediate). Polyprotodont marsupials certainly persisted in North America and Europe, for we find the remains of species nearly related to the existing opossums in the Lower and Middle 7% Bolodon is a synonym of Plagiaulaa, fide Gidley. 7 Broom has recently made a careful restudy of the affinities of Tritylodon, and con- cludes that it is a mammal, but not closely related to the marsupials, and represents an archaic specialization with many primitive characters inherited from the cynodont reptiles. R. Broom: Trans. S. Af. Phil. Soc., vol. xvi, pp. 73-77. 1905. Proc. Zodl. Soc. London, 1910, pp. 760-768. 1910. Bull. Am. Mus. Nat. Hist., vol. xxxiii, pp. 115- 134. 1914. MATTHEW, CLIMATE AND EVOLUTION 265 Eocene of Wyoming, in the Oligocene of Colorado and in the Upper Eocene to lower Miocene of France and Germany. They are not known from any later formation in any of the northern continents. In the Southern continents, they assumed a much more important position. In South America, in the absence of placental carnivora, the polyprotodont marsupials developed into a number of large and small predaceous mammals (Borhyenide), so closely paralleling some of the predaceous marsupials of Australia that they have been referred to the same family (Thylacinide). Pseudo-diprotodont marsupials were also fairly common, taking the place in the fauna held by Insectivora in the North, this group of placentals (except for a single type) not having reached South America. The marsupials of South America did not de- velop into groups taking the place of northern ungulates, rodents or primates, since primitive placentals of these groups (Condylarthra, ? Hystricomorpha, ? Lemuroidea) had penetrated into South America before it was separated from the Northern world, and there developed along lines sub-parallel to the development of the higher placental groups in the North, but distinct and less progressive. In Australia, the marsupials assumed a still more important position, as the only mammals of that continent. The placental mammals of the northern Tertiary did not reach Australia, except for a few strays—bats and mice and the dingo—which were too few in numbers and of too re- cent introduction to affect seriously the course of mammalian evolution on that continent. In the absence of placentals, the marsupials developed into a wide variety in size, form and habits of life, partially paralleling the higher mammals. . The near resemblance between the modern Australian Thylacinus and the Borhyzenide of Tertiary South America has been used as an argu- ment for an Antarctic connection between the two. Such a hypothesis will not bear close examination. The resemblance is not closer than between parallel adaptations in distinct families of true Carnivora, whose genealogy has been more or less completely traced back through inde- pendent lines of descent from unspecialized common ancestors. It is not closer, for instance, than that between the Oligocene Felide and the modern Cryptoprocta of Madagascar, whose common descent from an unspecialized placental carnivore (Viverrid or Miacid), analogous to the marsupial didelphyids, is generally admitted. The common char- acters distinguishing thylacinids and borhyzenids from the didelphyids are, without exception, such as would naturally be assumed independently in adaptation to predaceous terrestrial life and have been so assumed in numerous independent parallel adaptations of the same sort among 266 ANNALS NEW YORK ACADEMY OF SCIENCES placental Carnivora. On the other hand, Thylacinus has retained cer- tain didelphyid characters which are already lost by the most primitive of the Borhyenide (palatal vacuities, posterior position of the orbits,’® an external lachrymal duct, double perforation of the basisphenoid), while in other features (brain development, cursorial specialization, etc. ) it is more progressive. ‘he Borhysnidz are more progressive in the reduction of the last molar, in the differentiation of enamel from dentine, less so in the cursorial adaptation of the limbs and feet. Descent from a common ancestral type is undoubtedly shown, but some at least of the above differences point back to Didelphyide as this common type. The characters which Sinclair uses to separate the thylacines are the reduced number of incisors, the carnassial specialization of the molars and especially the loss of the metaconid. Every one of these features, besides numerous other common characters which he does not specify, may be paralleled in two or more distinct lines of Carnivora whose com- mon ancestors are not more predaceously specialized than Didelphys. The loss of the metaconid oceurs in Cyon, Ischyrocyon, Simocyon and Enhydrocyon among the Canidee, in all the post-Oligocene Felide, in Gulo, Megalictis, Mustela, etc., among the Mustelide, in the later Hye- nide, in Hyanodon and Pterodon among the Hyenodontide, in Patrio- felis among the Oxyzenide, in all the later Mesonychide. Each one of these genera is independently descended from genera in which the metaconid is well developed. In every case, it is simply a stage in predaceous adaptation of the molars, nor can it be assigned any other significance in the marsupial carnivores. There is, in short, no evidence for assuming a closer affinity between thylacines and borhyenids than common descent from didelphyid ancestors, and there is strong evidence against such an assumption. But if this be true, these animals afford no evidence for Antarctic connections between the southern continents ; for we have seen that Didelphyid marsupials were certainly present in the Mesozoic and early Tertiary of Holarctica and of South America, and we have no reason to believe that they would have had greater diffi- culty in reaching Australia in the Mesozoic or early Tertiary than the murine rodents found at a later date. The supposed presence of Diprotodont marsupials in the South Ameri- can Tertiary and in modern Australia has also been used in support of Antarctic connections between the two continents. The recent mor- phologic studies of Dederer*® and Broom®® have shown that Cwnolestes *8 Interpreted by Sinclair as a progressive character in 7hylacinus, but certainly the reverse in analogous placental adaptations. 7 PAULINE H. DepereR: Amer. Nat., vol. xliii, p. 614. 1909. sR. Broom: Proc. Linn. Soc. N. S. W., vol. xxxvi, p. 315. 1911. MATTHEW, CLIMATE AND EVOLUTION 267 is not a true diprotodont, but in fact belongs to the polyprotodont divi- sion of the Marsupialia, and with this genus must be associated all of the Epanorthids and probably all of the so-called Paucituberculata of the South American Tertiaries. If then the Diprotodonta, so dominant and so widely varied in Australia, were wholly absent from South America, while parallel adaptations were developed there from the Polyprotodonta, the distribution of these marsupials affords a valid argument against instead of for any Antarctic connection during the Tertiary. In view of the great amount of adaptive divergence seen in the various Pleistocene and modern genera of Australian Diprotodonta, the origin of the suborder in Australasia or its earliest invasion of that zoological region, must be dated far back in the Tertiary. On our present evidence it may well be regarded as wholly autochthonic, derived from early Ter- tiary or possibly from late Mesozoic polyprotodonts. Nevertheless, in view of the defectiveness of the Mesozoic record, where we should chiefly expect to find this group, if anywhere in the North, and the presumable ‘rarity of Tertiary survivors, there is nothing unlikely in the view that they originated primarily in the North lke their polyprotodont and allotherian relatives and were driven southward with the former group and somewhat more thoroughly extinguished in the north, while in Aus- tralia they blossomed out into a great adaptive expansion paralleling the absent ungulate mammals. It is probable that the opossums survived in North America throughout the Tertiary, although there is no clear record of them in our Miocene and Pliocene.*t But we know only a small part of our Pliocene fauna as yet, and the Miocene, although better known, represents chiefly the animals of the open plains, the forest fauna being very incompletely rep- resented. On the other hand, it seems probable that the apparent dis- appearance of marsupials from Western Europe after the Lower Miocene was real, and it is probable that they had disappeared even earlier from Asia. They have not been found in the later Tertiaries of India or China, so that they must have been rare if not absent at that time. The Eocene Tertiary of Asia, where they might be expected to be common, is altogether unknown.*? 81 A very badly preserved skull from the Colorado Miocene and a jaw fragment from the South Dakota Miocene in the American Museum collections are perhaps marsupials; but I have never been able to see in either specimen satisfactory proof that they were so, and have consequently never recorded them. ©The earliest Asiatic Tertiary fauna is that of the Bugti beds of India, lower Bur- digalian or upper Aquitanian according to Pilgrim, Rec. Geol. Sur. India, vol. SH ie) pp. 264-326. It is therefore either late Oligocene or early Miocene. 268 ANNALS NEW YORK ACADEMY OF SCIENCES TABLE XVI.—Distribution of Polyprotodont Marsupials | Neotropical Nearctic | Paleearctic | Ethiopian | Oriental | Australian Didel phyidee 9 . + 7 NI. as | Lhylacinidse Recent Ganolestee Didelphys None ‘ None None Dasyuride Pleistocene | Didelphyidee | Didelphys | None None None peor Payiee se rab G Dasyuridse Bore yee = None (Record in- Pliocene Didelphyidze | ~ ees None eaicnaate None Epanorthidee say eon ate) ; Borhyzenidee None after Miocene Didelphyidee ace lower Mi- (No record) | None Epanorthidee cene 5 is 3 v ; z Oligocene Dilelobyide Peratherium| Peratherium| None = a iS) ‘BS = QO ~ o ao mA a Polydolops, ete aS >, as Borhyzenidie & 2 eS Eocene Didelphyidee | Peratherium| Peratherium 3 2 ‘5 Caroloame- = ra S) aie o = ghinia = a a eo y e = S 3 S N Cimolestidee} , . . a & = Cretaceous Thleodon |(N© record) S S ie g Proteodidel- 5 S) Comanche phys * (No record ) 3s 2 = fo fo} ie) : : ; Zi Z Jurassic Triconodontidze ~ 2 53 Fragmentary remains, referred to Hy:enodontide by Dr. Ameghino. & Jurassic, fide Ameghino. 8 Doubtful fragments of jaws which may be Didelphyid. 86 Hxcept on borders of Australian region. - MATTHEW, CLIMATE AND EVOLUTION 269 TasLe XVII.—Distribution of Diprotodont and Pseudodiprotodont Marsupials and Allotheria (Multituberculata) Peer: Anlothera Di prowsounls Neotropical | Nearctic Palearctic | Ethiopian Australian Macropodidee Recent Cenolestes*" | None None None Phaseolomy- idee ete. Diprotodon- tidee Pleistocene None None Macropodidee Thylacole- onide ete. Pliocene Spar None None thidee® Mi Epanor- N N iocene thides®® one one Oligocene as None None None Wynyardia®? Eocene P Peep : ? None Paleocene Ptilodus Neoplagiau- Polymasto- lax don ete. Cretaceous Ptilodus Meniscoes- sus ete. Comanche Jurassic Ctenacodon | Plagiaulax | 7. ate ae Tritylodon Triassic Microlestes | Karroomys 87 This genns is a pseudo-diprotodont, as its real affinities are with Polyprotodontia, as shown by Dederer and Broom, I. c. 88 Affinities probably with Cenolestes. 89 Combines Polyprotodont and Diprotodont characters. 270 ANNALS NEW YORK ACADEMY OF SCIENCES MONOTREMATA The monotremes are the lowest group of mammals, far removed struc- turally from any others. Their connection with the main stock must date back to the end of the Paleozoic era. Nothing is known of their evolutionary history. The Multituberculata of the Mesozoic and Basal Eocene are regarded by Broom as ancestral to them, but this view is not supported by additional evidence since obtained. Xenotherium®® of the North American Oligocene, referred by its describer to the monotremes, is an Insectivore related to the Chrysochloride ; Scotwops** of the South American Tertiary is an Armadillo,®? and other genera referred by Ameghino to the Monotremes probably also pertain to other groups. We find them to-day limited to the Australian region, and surviving even there only by virtue of unusual specializations of habit; Hchidna protected by its coat of spines, Ornithorhynchus by its amphibious habitat, both genera burrowing and nocturnal. Presumably, these genera repre- sent the last relic of the early Mesozoic dispersal movements of the Mammalia. SUMMARY OF THE EVIDENCE FROM DISPERSAL OF LAND MAMMALS The foregoing review of the several groups of land mammals shows that the more recently evolved and dominant races of Mammalia are to- day mainly Holarctic, and many of them have not yet reached the more peripheral regions; that the ancestry of all these dominant races has been found in the Holarctic Tertiary formations, sometimes in Europe, some- times North America, more generally a series in each country of equiva- lent approximately ancestral stages. Where the geological record is ade- quate, these races are shown to be newcomers in the peripheral continents which they have invaded, and any ancestral series is absent. Their repre- sentatives in the peripheral continents are to a varying degree primitive and allied to earlier stages in the evolution of the race as represented in the Tertiary record of Holarctica, but they have specialized more or less along parallel or divergent lines from the direct line of descent of the northern representatives. When the parallel series in Europe and North America are sufficiently complete they are seen to be not parallel phyla of independent local evo- lution, but periodically recruited by more progressive new stages, appar- 89 HaARL DouGLass, 1905. (The name is preoccupied by Xenotherium Ameghino, 1904, a genus of typotberes.) *1L, AMEGHINO, 1887. WwW. B. Scorr: Rep. Prin. Exp. Patag., vol. 5, p. 12. 1903. MATTHEW, CLIMATE AND EVOLUTION yal ently from a common center of dispersal. The relations are like those of one side and the other of a branching tree whose trunk region is unknown to us. The more ancient and primitive groups of the Mammalia have mostly disappeared, or are in process of disappearance, from Holarctica. In the peripheral continents, they have undergone in many cases a notable local adaptive radiation and expansion, extensive in proportion to the isolation of these continents from the northern realm, more complete during the early and middle Tertiary than now. When the reunion to Holarctica permitted the northern fauna to invade the peripheral continents, these autochthonous groups were in general unable to maintain themselves against the competition of the more progressive northern races, and have either wholly disappeared or left a few scattered survivors, mostly aber- rant specializations which did not come directly into competition with the invading races. The survival of the major part of the marsupial radia- tion in Australia is attributable to its continued isolation. The apparent fact that Neotropical races of Edentata were able to invade North Amer- ica during the Pliocene and Pleistocene may be ascribed to two factors: 1) No Nearctic groups of closely analogous specialization existed at that time. 2) Owing to the far southerly extension of South America, the evolu- tion of mammals in that region was, so far as controlled by climatic change, more progressive and more nearly equivalent to the Holarctic evolution than in Australia or Africa. Its products therefore were better able to maintain themselves against their northern competitors. If we regard the Proboscidea as of Ethiopian origin, we must suppose that they too constitute an exception to the general rule that the races evolved in the peripheral regions have been unable to invade Holarctica. But the recent discoveries of Pilgrim and Cooper in the Oligocene of India tend strongly to show that the Proboscidea were from the first, as they certainly were in the later Tertiary, a group of Asiatic, not African, dispersal. The dominant influence of climate in controlling the range of modern mammals has been emphasized by C. H. Merriam. The mammals adapted to north temperate or even boreal climate are the most specialized and last evolved members of their respective races. The most primitive sur- vivors of northern races, and surviving members of races formerly abun- dant in the north, are met with chiefly in tropical regions. Similar rela- tions are seen in the faunz of the antarctic as compared with the southern tropical regions, although less obvious. This is especially seen in South 979 ANNALS NEW YORK ACADEMY OF SCIENCES America. It is displayed there quite as clearly in races, such as the crice- tine rodents, cervide, etc., which are admittedly of Northern origin, as it is in any autochthonous groups. Hence, it cannot be attributed to a gen- eral Antarctic dispersal center, but must be explained as a parallel evolu- tion under similar climatic stimulus. The general distribution of Mammalia 6n,these lines is almost univer- sally accepted; but many writers have pointed out certain supposed ex- ceptions and found it necessary to account for them by various hypo- thetical continental bridges. A careful consideration of these supposed exceptions shows that, if due allowance be made for parallelism and for the imperfection of the record, each one can be more satisfactorily inter- preted in accordance with the general law. And the acceptance of any such continental bridges would entail migrations of other groups which assuredly have not occurred. The hystricomorph rodents of South Amer- ica afford a single exceptional instance, in which over-sea transportation from Africa appears to be the only reasonable interpretation of the evi- dence at hand. I place much greater weight on the evidence from mammalian distri- bution than on that of any other terrestrial group for several reasons, as follows: 1) Their past history, the time, place and method of evolution of the vari- ous races, is better known than in any other group of land animals or plants. 2) The complexity of structure in the hard parts which are preserved as fossils is greater, affording a larger amount of evidence by which we may dis- tinguish parallel or analogous races and determine the closeness of their real affinities. As Stehlin® has recently observed, a single tooth of a mammal affords as much structural evidence whereby to determine its relationships as the entire skeleton of most invertebrates. Where our evidence is thus lim- ited (to a single tooth, for example), we may, and frequently do, find difficulty in deciding the exact aflinities of a fossil mammal. But where we have the skull or the skeleton or even the entire dentition, the results are correspond- ingly sure and precise as the data are more extensive. 3) Owing to their nearness to ourselves, their large size and other causes, we are better able to understand their adaptation and observe and appreciate the factors which may affect their evolution and migration. In dealing with the evidence furnished by the lower vertebrates and invertebrates, we are hampered by the wider limits of time within which the migration may have taken place, by the relative simplicity of the _ structure of the hard parts, which makes it less easy to distinguish paral- ®3“ijber die Siiugethiere der Schweizerische Bohnerzformation.” Verh. Schw. Naturf. Gesell., 98 Jahresvers. 1910, Basel. P. 11 of separate. MATTHEW, CLIMATE AND EVOLUTION 273: lelism from immediate affinity,°* by the relative scarcity of fossils as com- pared with living species (among land animals), and by our less certain knowledge of the causes which may control their evolution, their means of migration, and their true evolutionary history and affinities. INTERPRETATION OF NEGATIVE EVIDENCE IN FOSSIL MAMMAL FAUNA In considering a Tertiary mammal fauna, we must keep in mind the facts that there may be large facies of it that are represented imperfectly, if at all, in our records, and that there may be important parts of it which have left little or no record, owing to their habitat, small size or other circumstances. We may, with some reserve, conclude that the en- tire absence from the record of a group which is abundant in other faune indicates its real absence from the fauna. But we are not justified in so concluding in the case of rare or inconspicuous races. It is fair to as- sume that the absence of. Perissodactyla from the Oligocene fauna of Egypt or the Miocene fauna of Patagonia was real, and not a matter of defective record. The same assumption would be unjustified in the case of didelphid marsupials and dilambdodont Insectivora respectively. But the most conclusive evidence of the absence of a certain group from a given fauna is that while it is not found fossil, another group is found to have become adapted on parallel lines, taking its place in the fauna. The absence of Perissodactyla and Artiodactyla from the Miocene of South America is confirmed by our finding Litopterna, Toxodontia and Astra- potheria, which parallel in adaptation the horses, rhinoceroses, tapirs, camels, ete., of the North; the absence of Carnivora by the parallel adapta- tion of marsupials to take their place. The evolution of lemuroid pri- mates in Madagascar into large quadrupedal forms apparently paralleling certain groups of Ungulates,*? affords some evidence that the Tertiary hoofed mammals were unable to invade Madagascar. The absence of fissiped Carnivora from the recorded Oligocene fauna of Egypt would not be conclusive in itself; but, coupled with the excep- tional variety and abundance of the more archaic creodonts of the family Tt may be noted in illustration of this point that a natural cast of the entire carcass of a mammal would afford far less secure information as to its real affinities than would a fossil skull, and less even than a lower jaw with reasonably perfect teeth. The parallel adaptations so frequently recognized among mammals lead to superficial resemblance of distantly related types whose true affinities are readily recognized by the internal struc- ture. If, as among most invertebrates, we had only an external skeleton to guide us, the real affinities would not be so securely recognized. % The skull and the short limbs of Megaladapis are very suggestive of such types as Promerycocherus. The feet do not, however, indicate a terrestrial habitat, nor are the teeth efficient in grinding. The resemblance in teeth and skull of Archeolemur to the Anthropoidea is very marked. a4. ANNALS NEW YORK ACADEMY OF SCIENCES Hyzenodontide, it is very strong evidence that fissiped Carnivora had not y , af g yet invaded the Ethiopian region, at least in any considerable numbers. DISPERSAL OF REPTILIA The essential adaptive feature which distinguishes mammals and birds from the reptiles out of which they arose lies in the non-conducting cov- ering to the skin,—of hair or fur among mammals, of feathers among birds. The assumption of this covering enabled the body to be kept at a uniformly high temperature, thus favoring the maximum of bodily activ- ity, and making it practicable to develop the circulation and the entire organization to a much higher standard. It also made these classes of animals independent of the temperature of their environment. It ena- bled them to withstand cold or variable climate and to take full advantage of the conditions of the colder regions, which appear to favor a higher development than can be attained in moist tropical countries. The initial development of mammals and birds took place, so far as we are able to judge, during the great arid period of the Permian-Triassic. They appear to have been derived from unknown groups allied respec- tively to the theromorphous reptiles and to the ornithischian dinosaurs. We know almost nothing of their Mesozoic evolution, because the upland epicontinental formations of the Mesozoic, in which this record should be chiefly preserved, have been totally swept away, or if any remnants re- main, they have not been recognized and sufficiently explored to recover it. The formations of the swamps and coastal marshes, river-deltas, lit- toral regions and shallow seas of the Mesozoic are extensively preserved and their inhabitants well known to us. But of the upland fauna, we get only an occasional glimpse in such deposits as those of Solenhofen, where a few remnants of the fauna of the adjoining uplands have been pre- served in great perfection. We have, indeed, indirect evidence as to the nature of the upland fauna of the Mesozoic, for the successive groups of swamp dinosaurs, the marine birds and pterodactyls of the later Mesozoic and the abundant and varied mammalian fauna which appears at the beginning of the Tertiary are not derivable, any of them, from their predecessors in the swamp or marine faunz, but must be traced back to ancestors distinctly adapted to dry-land life, which reinvaded the coast- swamp, littoral or marine provinces. ‘This will appear more in detail in the discussion of the several orders. The point here to be emphasized is that the dry-land vertebrate fauna has been throughout the dominant facies and has repeatedly reinvaded the swamp and sea-coast provinces, the higher activity and better organization acquired on land giving its MATTHEW, CLIMATE AND EVOLUTION © -~2 Or members, when readapted to the marsh or littoral conditions, an advan- tage which enabled them to supersede the autochthonous dwellers in those conditions. Per contra, there have not been a succession of invasions of the dry land by the vertebrate inhabitants of swamp and sea-coast. Once established on dry land, the primary groups of dry-land reptiles held their own and evolved and expanded into higher types and greater variety, but they were not recruited, so far as the evidence shows, by new invasions from the swamp and aquatic fauna. DINOSAURIA The dinosaurs appear to be primarily a dry-land adaptation (properly speaking, two distinct but parallel adaptations) of the primitive rep- tiles.°° Their most obvious adaptive characters lhe in the long limbs and swift-running gait and the general parallelism to the ratite birds. As such, the conditions of life would tend to greater activity and higher de- velopment and enable them, when they reinvaded the swamps during the epochs of great swamp-extension, to reach greater size and dominance. It is these readaptions that are chiefly known to us and are apt to give the idea that the dinosaurs were distinguished by gigantic size and mass- ive proportions. In fact, these are no more typical of the order as such than the whale, hippopotamus and elephant are fairly typical of the mammals as such. There must have been multitudes of small dinosaurs, mostly inhabiting the upland, a smaller number living among the swamps and marshes, but we know comparatively little about them. Some notion of their numbers and variety in the Triassic is gained from the innumer- able footprints spread over the Triassic shore-deposits of the Connecticut River. But of all this multitude, we have actual remains of only two or three types. The Compsognathus skeleton of Solenhofen is, perhaps, an example of the small lght-limbed upland dinosaurs of the Jurassic ; Hallopus and Podokesaurus are perhaps fairly representative of their Triassic ancestors. The Jurassic sauropods, while highly specialized for aquatic hfe and river-bottom wading, yet retain a few features indicative of former land life. One of these is the long limbs, which it would seem must have been acquired on land. Another is the fact that the knee bends forward as it does in all other dinosaurs, while in reptiles primarily am- phibious the knee bends outward and the limbs are short. The elbow of the Sauropoda, on the other hand, bends outward, as in reptiles generally, not backward, as it does in primarily quadrupedal land animals, and this 6H. VON HUENE: Geol. u. Pal. Abh., N. F., Bd. xiii, s. 22-38. 1914; Neues Jahrb., Beil. Bd. xxxvii, s. 577-587. 1914. : 276 ANNALS NEW YORK ACADEMY OF SCIENCES I take to be an indication that their quadrupedal gait is partly secondary and that they are derivable from long-limbed, partly bipedal ancestry. The shortening of the feet and pillar-like construction of the limbs is an obvious parallelism with the specialization of these parts seen in all large land mammals and is an adaptation to their great size. No near parallel can be found to this group among living animals; the hippopotamus affords some suggestions, but diverges widely in many respects.** I have already referred to the primary adaptation of the dinosaurs as a dry-land adaptation of the Reptilia. ‘To a limited extent, the mod- ern lizards represent a corresponding adaptation but not carried so far or occupying so important a place in the fauna. The lizards have to compete with the large and varied dry-land fauna of mammals, and rela- tively to these, they occupy but an unimportant niche in the terrestrial life. They suggest, however, the sort of animal which in the absence of a higher competing type evolved into the dinosaurs, and their more specialized types (e. g., Chlamydosaurus) mimic them in proportions in a most instructive manner. Dinosaurs are first recorded from the Triassic; those which we actually know®® are of moderate to large size, slender and long limbed as com- pared with other reptiles, not highly specialized in dentition, unarmored and some but not all bipedal in gait. Indirect evidence in the multitudes See W. D. MattHrew: “The Pose of the Sauropodous Dinosaurs,” Amer. Nat., vol. xliv, pp. 547-560. 1910. ®§The principal references on ‘Triassic dinosaurs are the following : R. Broom: “On the South African Dinosaur Hortalotarsus,” Trans. S. Afr. Phil. Soce., vol. xvi, pp. 201-204. 1906. E. Fraas: “Die neuesten Dinosaurierfunde in der schwabischen Trias,’ Die Natur- wissenschaften, Bd. I, Heft 45, pp. 1097-1100. 1913. F. voN HuENE: “Die Dinosaurier der europiischen Triasformation.” Geol. u. Pal. Abh., Supplem. Bd. I. 1908. ———: “Hin primitiver Dinosaurier aus Elgin,’ Geol. u. Pal. Abh., Bd. xiv (N. S., Bd. x) Heft. I. 1910. : “Beitriige zur Geschichte der Archosaurier,’ ibid., Bd. xvii (N. S., Bd. xiii) vente) eb OWa: : “Ueber die Zweist’ammigkeit der Dinosaurier,”” Neues Jahrb. Beil., Bd. xxxvii, s. 577-589. 1914. Bh. voN HUENE und R. 8. LuLL: “Neubeschreibung des Originals von Nunosaurus agilis Marsh,’ Neues Jahrb., Bd. I, s. 184-144. 1908. ——— ——: “On the Triassic Reptile Hallopus victor Marsh,’’ Amer. Jour. Sci., vol. xxv, pp. 113-118. 1908. O. JAEKEL: “Ueber die Wirbethierfunde in der Oberen Trias von Halberstadt,’ Pale- ont. Zeitsch., Bd. I, s. 155. 1913. R. S. LuLu: “Fossil Footprints of the Jura-Trias of North America,’ Mem.. Boston Soc. Nat. Hist., vol. v, pp. 461-557. 1904. ———: “Dinosaurian Distribution,” Am. Jour. Sci., vol. xxix, pp. 1-39. 1910. ——: “The Life of the Connecticut Trias,” ibid., vol. xxxiii, pp. 397-422. 1912. O. C. MarsH: ‘Notes on Triassic Dinosauria,” ibid., vol. xliii, pp. 543-546. 1892. ———: “Restoration of Anchisaurus,” ibid., vol. xlv, pp. 169-170. 1893. : “Dinosaurs of North America,” U. 8. Geol. Sur., 16th Annual Report, pp. 143- 244, pill. 1896. MATTHEW, CLIMATE AND EVOLUTION ony of footprints of the Connecticut Valley sandstones shows that there must have been also a great number and variety of small bipedal three-toed forms all presumably dinosaurs, and other reptiles with shorter feet and more numerous toes which may also have been dinosaurs, although not generally so referred. Liull** states in regard to the latter: “These forms seem to represent survivors of the ancient stem from which the dinosaurs arose; they may, however, represent primitive quadrupedal dinosaurs which had not yet acquired the erect gait.” He calls attention to their possible relationship to Protorosaurus and Kadaliosaurus. From these and other fragments of evidence, we may reconstruct a concept of the dinosaurs as a land adaptation developed during the arid Permo-Triassic climatic phase, corresponding to the later deployments of the mammals along the same lines of adaptation and under a similar impelling cause of progressive aridity and continental expansion. Dur- ing the base-leveling and submergence and moist tropical climate of the Jura, these dry-land adaptations reinvaded the swamps and _ coast- marshes, the least specialized types (cf. Protorosaurus), more quad- rupedal and some of them long-necked, reverting farthest towards an aquatic life and specializing into the peculiar Sauropoda, while the higher bipedal types retained more of their terrestrial habitat but evolved into huge, massive armored and bizarre creatures, to be paralleled in habit and type at a later date by the bizarre specializations of the Eocene Mam- malia. These are the familiar dinosaur fauna of the Upper Jura and basal Cretaceous. The drier uplands of that time must have been ten- anted by lighter, smaller dinosaurs, but of these, in my opinion, we have little direct evidence. But that they continued to exist and carry for- ward their primary lines of adaptation is shown by the subsequent history of the order.1”° In the Lower Cretaceous occurred a swing towards emergence and arid conditions, not extreme, but sufficient to wipe out the sauropod dinosaurs in the northern world. They survived, however, in the southern conti- nents until, in the middle and later Cretaceous, the pendulum swung back to a marked extreme of submergence and moist-tropical climate, and their remains are found in late Cretaceous beds in South America, East Africa, Madagascar and Australia. The correlation of these beds is in need of revision, however; they may be Comanchean. In the Northern CONES Se bovis Ber ose jo, Zhe | aityay oR, S. LuLy (“Dinosaurian Distribution,” Amer. Jour. Sci., vol. xxix, pp. 1-39, 1910) has admirably summed up the data regarding the geological occurrence of dinosaurs. While not agreeing in all respects with his interpretation, I take pleasure in noting the accuracy and clear presentation of the evidence as worthy of the high regard in which its author is held by his confréres. : 978 ANNALS NEW YORK ACADEMY OF SCIENCES world, at all events, they did not reappear after the early Comanchean. A dinosaur fauna largely similar to that of the Jurassic in habits and adaptation in other respects, developed during the late Cretaceous in the North. It contains no Sauropoda, but it includes amphibious types (Trachodontide) with marked aquatic adaptation, gigantic terrestrial swamp and forest dwellers, like the ceratopsians, tyrannosaurs and anky- losaurs, and many smaller more agile forms. These Cretaceous giants, however, appear to have evolved, not from amphibious or aquatic dino- saurs of the Jura, but, in part at least, from small and little known forms, of more upland adaptation, which had been much more highly specialized for dry-land life than any of the Jurassic swamp dwellers, and had re- adapted themselves to the forest and swamp environment of the later Cretaceous. The trachodonts and ceratopsians, for instance, while re- lated to the earlier iguanodonts, cannot be directly derived from them but must be traced back to some unknown contemporary which was highly progressive in developing efficient grinding dentition, compact feet with flattened hoofs, ete.—characters which in a survey of mammalian adapta- tion we find to be especially associated with upland habitat. The evi- dences of former dry-land adaptation are not so clearly shown in the other swamp-giants of the late Cretaceous, but they may perhaps be shown by further study." In sum, we may find in the hypothesis of recurrent climatic change, and in the primary adaptation of the dinosaurs as a dry-land adaptation of Reptilia and their secondary readaptations to forest and swamp life, a fairly satisfactory solution of their distribution and phylogeny. Lull, in his able discussion of the subject (1910), explains their adaptation along these lines. But at present our data, both of correlation and identification, are too uncertain to allow of positive and detailed con- clusions in regard to the centers of dispersal and course of migration of the dinosaurs. That the sauropods survived in the southern continents long after their extinction in the north appears proven, if we accept the stated geological correlations of the southern formations where they are found and set aside as an erroneous identification the reported occurrence of a sauropod in the Danian of France.’°? That the Theropoda survived into the Eocene in South America and Theropoda and Predentata into the Paleocene in North America is not improbable on a priori grounds, 10. J,, DOLLO (Bull. Soc. Belg. Géol., xix, p. 441. 1905) has shown that the quadrupedal gait of many of the Predentate dinosaurs is a secondary adaptation from bipedal ances- try. I believe this to be true, to a less extent, of the Sauropoda as well. 102, Nopsca (Rep. Geol. Mag., vol. vii, p. 261. 1910) states that the femur on which this recorded occurrence is based is not a sauropod but a trachodont dinosaur, allied to or identical with Telmatosaurus of the Gosau beds of Austria. MATTHEW, CLIMATE AND EVOLUTION 279 but the evidence that they actually did so survive is open to serious question. So far as they go, the facts accord with the dispersal of the dinosaurs from the northern land mass. And so far as I have been able to review the data, the migrations of the order could be made to conform with the present distribution of continental and abyssal areas (Mada- gascar excepted'™) about as well as with the different distribution upon which they are plotted by Dr. Lull. It is significant in this connection to note that young individuals are very rarely found in the dinosaur formations. ‘Thousands of individuals are found together in some of the great quarries, pertaining to a great number and variety of genera and with a wide range in size, but it is very rare to find young individuals among them. This fact is well known to collectors, but has not, as far as I know, been commented upon in print. It is true that young individuals are less clearly distinguished from adult among reptiles than among mammals, the chief difference being the imperfect ossification of the bone structure, and that such im- perfectly ossified bones are likely to be poorly preserved and might often be rejected by collectors on this account. But making all reasonable allowance for these considerations, there remains a very notable contrast with fossil mammal quarries and fossiliferous formations, in which young individuals are always to be found among any considerable number of adult specimens and often are more numerous than mature individuals. This may be interpreted in conformity with the above theories as to the habitat of dinosaurs, by supposing that the young dinosaurs were more dry land or upland animals, retaining the ancestral habitat, and coming down into the swamps only when they reached maturity and their larger size made an amphibious or aquatic habitat more suitable. The young animals would rarely or never visit the swamps and deltas, whose formations have alone been preserved, and their fossil remains would be correspondingly scarce. Young crocodiles, so far as I can gather from various descriptions, are somewhat more terrestrial in habit than the full-grown animal, but the difference is evidently not considerable. Analogous cases among fish, marine types breeding in fresh water and vice versa, are well known. The migration of birds has also some analogy, if, as may often have been the case, the swamp dinosaurs resorted to dry land for breeding and egg- laying purposes. In either case the breeding or egg-laying place would be presumptively the ancestral habitat of the race. 103 The Cretaceous sauropoda of Madagascar may have reached that island in the same Manner as the hippopotamus did at a later period, namely by swimming. 280 ANNALS NEW YORK ACADEMY OF SCIENCES CHELONIA The publication of Dr. Hay’s splendid monograph’** upon the extinct Chelonia of North America has added a great deal to the available data for explaining the distribution of this group. So far as the Tertiary and modern distribution goes, it conforms to the same lines of dispersal as do the various orders of mammals. The pre-Tertiary history of the order is mostly too fragmentary to afford any important data bearing, pro or con, upon the theories here presented. The whole order is in general conservative and persistent to a high degree, like the Crocodilia. The occurrence of giant tortoises (Testudo) on several oceanic islands and in Australia and Patagonia (Metolania) has been adduced as evi- dence for continental connection of these islands and for an Antarctic connection of the two southern continents. Here, as in the case of the carnivorous marsupials cited on page 265, the evidence will not bear close examination. In the first place, we know that large tortoises of the genera Testudo and Stylemys are among the most abundant fossils in the Middle and later Tertiary of the Nearctic, Palearctic, Oriental and Ethiopian regions. So far as we can judge, they were cosmopolitan, except Australia and Patagonia. They occur in the Pleistocene of Cuba and Madagascar and survive to the present day in certain islands in the Indian Ocean and in the Galapagos Islands. So far as these oceanic islands are concerned, if we assume that their presence in one involves continental union, it must do so in all. If such continental union oc- eurred, it is hardly conceivable that, in each instance, tortoises alone would have made their way to the islands. We must infer for each and every one of them a vertebrate and invertebrate land fauna. Where is that land fauna, and why has it perished? The idea of selective drown- ing might possibly be entertained if we had to do with only a single instance, but is too absurd for serious consideration, when we deal with several instances of the survival of the same race. The only reasonable method of accounting for the presence of Testudo on these islands is that its facilities for oceanic distribution are somewhat better than those of mammals and that it arrived by over-sea transportation. The most recent argument for land connection of the Galapagos Islands is by Dr. Hay.?°° He advocates a connection with Central Amer- ica, via a submerged ridge which is shown in the reports of the Blake Expedition to extend southwest from Costa Rica towards the islands. 14 QO. P. Hay: “Fossil Turtles of North America,’ Carnegie Institution Publ. No. 75. 1908. 1061 OSE. EUAN, 1G. MATTHEW, CLIMATE AND EVOLUTION 281 The depth of this ridge Dr. Hay omits to state, but the soundings indi- cate it as being upwards of ten thousand feet, so that it does not mate- rially affect the improbability of an elevation to this extent. The Gala- pagos Islands are purely volcanic in origin and stand upon a platform less than a thousand feet in depth, similar on a smaller scale to that which surrounds the continents and presumably open to similar inter- pretation. If so, the islands have, probably, been more or less completely united at periods of continental emergence and completely isolated at periods of continental submergence (if any such haye occurred since they were first upbuilt from the ocean floor by voleanic ejectamenta) but never connected with the mainland. As the island platform is less extensive than Madagascar or Cuba, farther from the mainland and without intervening island stepping-stones, the opportunities for success- ful colonization through rafts or other means of transport have been fewer, and have not succeeded in introducing any mammals or amphib- ians and but few reptiles and invertebrates. The most favorable oppor- tunity for such colonization would be when the islands were at their maximum elevation—towards the end of the Tertiary, if this corre- sponded with the elevation of the mainland—as at that time the extent of coast and consequent probability of making a landing would be much greater. The subsequent isolation of the islands by submergence ac- counts for the presence of distinct although related species on different islands. Thus the series of “miracles of transportation,” which Dr. Hay finds it so difficult to accept, dwindles down to a single “miracle” and to one which he must invoke to account for the populating of the more remote Pacific islands, and which, when considered in relation to the time involved, does not really involve any serious improbability. On the other hand, if a miracle be an exceptional occurrence in apparent contravention of all probabilities, and without assignable causes in nat- ural law, I think the processes of selective drowning, or of selective migration of sporadic elements of a fauna, involved in the alternate hypothesis, in addition to the elevation during the late Tertiary of abyssal depths to the surface, unwarranted by any valid evidence, does involve a series of miracles, almost as unworthy of belief on the evidence offered, as the special creation of the species of the Galapagos Islands appeared to Darwin. The present distribution of species of Testudo on the islands of the Indian Ocean has been partly changed by man, so that there is some uncertainty about its details. Lydekker states it as follows: “Madagascar, probably the Comoros, North and South Aldabra—small islands lying to the northwest of the northern point of Madagascar—the Mascarenes YR ANNALS NEW YORK ACADEMY OF SCIENCES HOM or Mascarenhas, situated to the east of Madagascar and including Réunion, Mauritius and Rodriguez and lastly the Amirantes and the Seychelles, which are the most northern of the whole assemblage and only about four degrees south of the equator.” °° Jach of these groups of islands, except the Mascarenes, stands upon a shallow platform, and is surrounded by abyssal ocean, upwards of 5000 feet between the Comoros and Africa, elsewhere upwards of ten thousand feet. The three Mascarene islands rise separately from abyssal depths. Madagascar is about 180 miles from the African coast; the other islands are 400 to 600 miles from Madagascar; the present normal set of current is unfavorable to transportation from Madagascar. It is very frequently asserted that a bank of shallow soundings con- nects India with Madagascar through the Amirante Seychelles group, and that this indicates a former continental bridge of which these islands are remnants. The facts are as above stated; the so-called bank is very little above the general level of the floor of the Indian Ocean and is not differentiated from it in any features of relief that would suggest its former continental character. The transportation of natural rafts five hundred miles against the normal set of current—or five times that distance if from the East Indies—is the most improbable element in this explanation. There is no valid reason to suppose that the general direction of winds and cur- rents differed materially in the later Tertiary from the present day conditions. I do not think it necessary to assume with Dr. Lydekker that the tortoises were of gigantic size when they reached the islands or to ignore, as he does, the elements of parallelism in considering their affinities to continental species. Nor does it appear that the difficulties which he admits in accounting, on the hypothesis of former continental union, for the absence of the rest of the fauna, should be ‘‘set aside for future consideration.” They add so greatly to the improbability of the hypothesis, that in conjunction with the physiographic difficulties it appears wholly out of range of reasonable probability. On the other hand, an investigation of the very variable direction of the winds and currents in the Indian Ocean would probably yield data to reduce the improbabilities in the hypothesis of over-sea transportation as above stated. The third possible hypothesis is that the present distribution is due in part to human agency, not necessarily limited to the historic period. If this factor may account for a species of Canis in Australia distinct from the living species of Arctogea, it may perhaps help to account for peculiar species of tortoises as well. 106 Science Progress, October, 1910, p. 303. MATTHEW, CLIMATE AND EVOLUTION 983 As for Miolania, it occurs in the Notostylops Beds of Patagonia and in the Pleistocene of Australia. The Notostylops Beds are Kocene, as here advocated. The persistence of so highly specialized a genus for so long a period appears surprising; if they are Lower Cretaceous, as Ameghino asserts, it is quite unprecedented. My acquaintance with chelonian anat- omy is not adequate to warrant my venturing an opinion as to how far parallel evolution from less specialized Pleurodira might account for this anomaly. But we certainly do not know to what extent this genus or a less specialized pleurodiran ancestor may have been aquatic or even ma- rine in its habits. And unless we suppose that it had some such semi- marine adaptations which would enable it to cross a marine barrier im- possible for terrestrial mammals, I do not see how to account for its reach- ing Australia without any of the Notostylops mammalian fauna accom- panying it. We cannot believe that a placental fauna ever reached Aus- tralia, for if it had we should not see the development of a marsupial fauna on analogous adaptive lines to take its place. Miolania, then, could cross some barrier, presumably an ocean barrier, which land mammals could not; and it becomes merely a question of how wide a barrier this extinct chelonian of unknown habits could cross. The present lines of the continents within the continental shelf would not present materially greater difficulties in its reaching Australia via Antarctica than Testudo has managed to surmount in reaching Mauritius and the Seychelles, and I think we are justified in saying that the occurrence of Miolania has no weight as evidence of former Antarctic connections of the Southern conti- nents and, in fact, is opposed to any actual land connection. The following notes on the distribution of the land Chelonia are sum- marized from Dr. Hay’s monograph: Cryptodira are the dominant group of turtles ae compare with the pla- centals among mammals. All continents except Australia. Chelydride.—Central America, eastern North America and New Guinea. Apparently a relict-distribution, but the family is unknown fossil. Dermatemydide.—Part of Central America. Found in abundance in North America in the Upper Cretaceous and in reduced numbers during the Tertiary. Emydide—Chiefly Holarctiec and Oriental. A few have reached South America, none in Ethiopia, Madagascar or Australia. First known in Holarctic Lower Eocene. Testudinide—Very abundant in Tertiary Holarctica but now mostly re- stricted to its southern margin. Abundant now in Ethiopia and a few species in Neotropical and Oriental regions; also in oceanic islands. Present in Su- matra, absent in Java, present in Celebes but absent in Borneo. These and other features are very suggestive of man’s having had much to do with the local extinction of Tortoises. For obvious reasons this family would be pecu- liarly subject to his ravages. 284 ANNALS NEW YORK ACADEMY OF SCIENCES Pleurodira.—Now limited to the three southern continents, Holarctic in the later Mesozoic and early Tertiary, and the extinct Amphichelydia from which they are descended were likewise a Holarctic group. The occurrence of closely related genera in South America and Madagascar is used in support of a Brazilian-Ethiopian-Malagasy land connection. It would be interpreted in conformity with the views here advocated, as due to common descent or to parallel evolution from Tertiary Pleurodira of Holarctica. Trionychide.—The distribution of this group is exceptional in that it is en- tirely absent from the Neotropical region and the Pacific coast of North America, while common to eastern North America, the Ethiopian, Oriental and southeastern Palearctic regions and New Guinea. Ameghino records Triony7r from the Notostylops Beds of Patagonia,” indicating if the identification be correct that the group was formerly present in South America. It is found abundantly in the Cretaceous and Tertiary of North America and in the older Tertiary of Europe; absent from Australia and Madagascar. Presumably this is a relict-distribution of an ancient group, whose facilities for transportation were relatively limited. It should be noted that the hy- pothesis of over-sea transportation on rafts would be less applicable to aquatic animals than to their terrestrial relatives, as they would be less likely to be earried out to sea on floating vegetation, on account of their ability to leave it at will for the shore. But the absence of the group from the Neotropical and Western Nearctic, and its presence in New Guinea, are anomalous features. CROCODILIA The crocodiles are usually regarded as the most conservative of the reptilian orders. This is true enough, so far as adaptive specialization from the primitive amphibious environment into the higher plane of ter- restrial habitat is concerned. Their expansional tendencies have been in the other direction, towards invasion of the marine province. The present geographic distribution of the group is as follows: Narrow | Gavialis, India. snouted | Tomistoma, East Indies. Alligator, Southern United States, China. Broad Crocodilus, Africa, southwest Asia, Oriental and northern Austra- lian regions, tropical America and West Indies. snouted Caiman, Tropical America. Osteolemus, West Africa. This is very clearly a remnant-distribution and is explained, at least in part, by the occurrence of crocodiles in the Tertiary. Fossil Crocodilia are abundant in the early Tertiaries of Europe and North America. The Kuropean species, according to Zittel,'°’ belong partly to Crocodilus, 17 WL. AMEGHINO: “Age des Formations Sedimentaires de Patagonie,’ Anal. Soe. Cient. Argent., tom. L, Liv, p. 52 of separata. 1903. 108 K, A. VON ZITTEL: Grundziige der Palwontologie, 2e Aufl., ii Abteil., s. 272. 1911. MATTHEW, CLIMATE AND EVOLUTION 285 partly to the extinct genus Diplocynodon, in which the proportions of the skull are like Alligator, but with a notch for the lower canine, like Croco- dilus, while the armor, especially the belly armor, is like that of Caiman. North American Tertiary Crocodilia are all with one exception referred to Crocodilus, but the armor is incompletely known, and they may prove also to include Diplocynodon. Gavialis is recorded from the late Tertiary of India; Tomistoma and Crocodilus occur in the Oligocene of Egypt and Tomistoma in the Miocene of southern Europe. The common Egyp- “fy Alligators WY Crocodiles Eee a ILI] Calmans VA Crocodiles tn late Crelaceous and early Tertiary __ Osleolaemu Fic. 32.—Distribution of the Crocodilide Originating probably in Cretaceous Holarctica, they have been restricted to the pe- ripheral continents by inability to become adapted to cold climates. Note discontinuous occurrence of crocodiles and of alligators, the last the most specialized, as Caiman and Osteolemus are the most primitive of the living genera. tian Oligocene species of Tomistoma is intermediate between this genus and Gavialis. The Upper Cretaceous crocodiles are nearly allied to those of the early Tertiary. The Jurassic and Comanchiec crocodiles include also long-snouted gavial-like forms, more or less marine in habitat, and broader-snouted crocodile or alligator-like forms of more strictly fresh-water habitat. All 286 ANNALS NEW YORK ACADEMY OF SCIENCES have amphiccelous vertebrae, whereas all Cenozoic and late Cretacie croco- diles have proceelous vertebra. It is commonly believed that certain of the narrow-snouted types (Teleosaurus) led into the gavials, the broad- snouted (Goniopholids) into crocodiles and alligators: and that the broad-snouted types first appearing in the Upper Jura are derived from the teleosaurs which first appear in thesMiddle Jura. Neither of these propositions seems to me to be probable. 'The narrow-snouted crocodiles are characteristic of marine or semi-marine formations, the broad-snouted kinds of fresh-water formations; the known formations of the Middle Jurassic are chiefly marine, those of the Upper Jura chiefly fresh-water. If we turn back to the Trias, we find that in the allied Parasuchia there were also long-snouted (Mystriosuchus and Rutiodon) and broader- snouted (Belodon) types—both of fresh-water habitat, but apparently less aquatic than Crocodilia; in the alhed Pseudosuchia the snout was short, and the adaptation to amphibious or fresh-water life; while the more distantly related dinosaurs were terrestrial and_short-snouted. Upon these data, it appears to me more reasonable to suppose that the Triassic Mystriosuchus and Rutiodon, the Jurassic Geosauride, Teleo- sauride and Metriorhynchide and the Tertiary Gavialide are all inde- pendent successive adaptations to a fish-eating diet and a more or less marine habitat and that the Jurassic Goniopholide are the source of all the modern Crocodilia. This will also relieve us from the necessity of supposing that proccelous vertebra and a number of other identical char- acters were independently and simultaneously acquired in two phyla of diverging adaptation. 'The accepted view involves the anomaly of asso- ciating divergent adaptation with convergent structural evolution. However this may be, we are justified in assuming certain characters as primitive among the modern Crocodila, since they are common to all the older types. These are the following: 1) More complete and consolidated ventral armature. Common to all the Mesozoic genera, retained in Diplocynodon of the European Tertiary and the modern Caiman and Osteolemus. 2) A notch instead of a pit in the upper jaw for reception of the lower canine. Common to all the short-snouted crocodiles of the Mesozoic and Ter- tiary, retained in the modern Crocodilus. 3) Amphiccelous vertebre. Common to all Crocodilia and related groups up to the middle Cretaceous, lost in most Upper Cretaceous and all Tertiary’” and modern genera. 4) Large supratemporal and small lateral temporal fenestrie. The upper temporal fenestra is large in all Mesozoic Crocodilia, considerably smaller in the gavials, quite small in Crocodilus, Alligator and Caiman. 109 Except Notosuchus of the Patagonian Eocene. MATTHEW, CLIMATE AND EVOLUTION 287 5) Posterior nares more anterior in position. In the Mesozoic crocodiles, the choanz are situated at the posterior end of the palatines in the long- snouted groups, while in the short-snouted Goniopholide, they have moved further backward, between the palatines and pterygoids. In the modern Crocodilia, they are still farther backward, entirely enclosed within the ptery- goids. This is an adaptation to lying submerged with the nostrils only pro- jecting above the surface of the water and enables the animal to breathe comfortably in this position. It would naturally develop in the slow, omniy- orous broad-snouted crocodiles and not in the swift-moving fish-catching, long- snouted types; hence its greater development in Goniopholide than in teleo- saurs, ete. The fact that it is fully as much developed in gavials as in croco- diles is another reason for deriving both from Goniopholid ancestry. According to the above criteria, Alligator is the most progressive mod- ern genus.’!° Caiman is primitive in (1); Osteolemus in (1) and (2) ; Crocodilus in (2) and to some extent in (4); Gavialis and Tomistoma are primitive in (4), divergent in adaptation in other respects, so that comparisons would be unprofitable. We may conclude, therefore, that so far as they go, the Crocodilia accord with the general lines of distribu- tion of other groups. They ranged much farther north during the Ter- tiary than they do now; the most progressive modern genus, Alligator, has the most northerly range, and the Neotropical Caiman, the West African Osteolemus and the cosmopolitan tropical genus Crocodilus are primitive in one or another respect. The gavials also had a wider and more northerly distribution during the Tertiary. That the present limits of range are conditioned chiefly by tempera- ture and climate, and that the much wider range in the early Tertiary was due to a warmer climate towards the poles, will hardly be questioned. Of previous limitations and expansions of range in the order, due to previous secular alternations of climate, there is no adequate evidence. The distribution of the more primitive modern genera in widely sepa- rated parts of the tropics; the occurrence of the most progressive genus on the northern borders of the range of the order in two widely separated regions, and, finally, the survival in the Eocene of Patagonia of a croco- dile, Notosuchus, of the Mesozoic type which had disappeared from the Northern world by the Middle Cretaceous,—these facts point to a north- ern rather than a tropical or southern center of dispersal for the order; but the evidence is slight and far from conclusive. 10 R, L. Ditmars, of the New York Zoédlogical Park, has observed that crocodiles are decidedly more active and ferocious animals than alligators. I would not interpret this, however, as meaning that they are more progressive, in the sense here used, since the adaptation of the typical Crocodilia is not towards an active life. 288 ANNALS NEW YORK ACADEMY OF SCIENCES LACERTILIA Lizards are the largest group of the Reptilia, comprising over 1800 species, mostly of small size. Most of them are active animals, and a large proportion are adapted to rocky and desert habitat and arid climate. They are more dependent on external warmth than mammals and birds, and consequently are excluded from the colder regions; their means of dispersal are perhaps less limited than with mammals, if we may judge from their wider distribution, for they do not appear to be of more ancient origin. Unfortunately, the rarity and fragmentary nature of their fossil remains stands in marked contrast with those of mammals, and our evidence as to their evolution and dispersal is chiefly indirect, based upon the modern distribution, and is neither conclusive nor con- vineing. Such as it is, it compares fairly well with corresponding dis- tribution features among the smaller Mammalia and points to the same conclusions. But it emphasizes the importance of occasional over-sea transportation as a factor in distribution. Gadow observes? in regard to the Geckos, the most cosmopolitan of all lizards: “Although not at all aquatic, they are particularly fit to be transported acci- dentally on or in the trunks of floating trees, to which they cling firmly, and they can exist without food for months.” Other groups are somewhat less easily transported in this way, and to quote the same authority: “It is a most suggestive fact that most of those families of Reptiles, and even of other vertebrates which have a wide distribution and are apparently debarred from transgressing Wallace’s line, are also absent from Madagascar.” The iguanas are chiefly Neotropical, but they occur also in Madagascar, in the Fiji and Friendly Islands and in the West Indies and Galapagos Islands, as well as on the American continent. Fossil iguanas are re- corded from the Upper Eocene and Oligocene of Europe and from the Upper Cretaceous and Middle Kocene of the western States. If these determinations be correct, they must formerly have been more cosmo- politan. Their presence in Madagascar is most reasonably explained by their former presence in Africa, which is rendered probable by the fact that they occur in the early Tertiary of Holarctica, along with various mammalian groups which certainly did reach Africa. Their disappear- ance from the mainland of Africa may be coupled with the invasion of other later developed groups, Zonuride, Varanide, Lacertide, which 11 HANS Gapow: Cambridge Natural History, vol. viii, Amphibia and Reptiles. 1901. The distribution data for lizards and amphibians are mostly based upon this authority. Nid gt MATTHEW, CLIMATE AND EVOLUTION 289 were prevented from reaching the New World by the absence of any land bridge or land approximation within their temperature limits. One: genus of Zonuride has likewise reached Madagascar. Bearing in mind the progressive limitation of northerly range of the Lacertilia by the secular refrigeration of the polar regions during the’ Cenozoic, we can see that, if the distribution of land and water has not greatly changed except within the 600 feet limit, any families arising; during the middle or later Tertiary would be limited to the old or to the; new world. While the distribution of various lizards in oceanic islands compels us to admit that they can cross considerable bodies of water and obtain a foothold on an imperfectly populated island area, yet the proba- bilities of their crossing the whole width of a broad ocean and maintain- ing themselves against competitors trained in the broad arena of a great continent appear to be very much less and almost negligible. Conversely then, we may assume that a distribution, such as that of the Scincide,. Iguanide, Geckonide, Anguide and Amphisbenide, involves the evolu- tion and cosmopolitan distribution of these families as early as the Ko-. cene. The Agamide, Varanide, Lacertide, Zonuride, Chameleontide: are Old World families, and none are known from the New World. The Zonuride may well be regarded as of Ethiopian evolution; if not, they must be a remnant of a very ancient stock. The same may be said of the Chameleons, except that if Ethiopian they reached as far as India. The Lacertide, the highest, or at least most typical family of lizards, are evidently the most recent development; they have not yet reached Mada- gascar or Australia, and their northern limit is higher than in any other lizards. The Varanide and Agamide have not reached Madagascar but have spread widely through Australia. The evidence from extinct lizards is very slight, the remains are scanty and mostly too fragmentary for positive family identification. Of the several genera from the Hocene and Oligocene of North America, two are positively referable to the worm- like Amphisbeenide, whose present distribution in tropical America, the West Indies and Africa is thus partly explained as a remnant of a former wider northerly range and presumably Holarctic. Of the remaining North American Tertiary genera, Peltosaurus and Glyptosaurus are re- ferred to the Anguide ;* the remaining genera are too fragmentary for reference or have not been studied.1? 12 WARL DouGLASS: Ann. Carn. Mus., vol. 4, p. 278. 1908. 18 The recorded presence of Iguanidse (Jguanavus) in the Cretaceous and Eocene, while not provable, is not unlikely ; that of Chameleon (C. pristinus) in the Upper Cretaceous is improbable and based upon insufficient evidence; the reference of Thinosaurus (Middle Eocene) to the Varanide appears to be merely a matter of bibliographic convenience ; the specimens are probably definitely referable, but the only expressed opinion as to their affinities is by Boulenger (1891), who suggests their relationship to the Telide. 290 ANNALS NEW YORK ACADEMY OF SCIENCES In general, so far as I can judge, the Lacertilia lend no support to the theories of transoceanic bridges. Their widespread insular distri- bution must in some cases, and may in most others, be explained by over-sea transportation. ‘Chey lend some support to late Tertiary eleva- tions to the continental shelf line so as to include the continental islands and to a line of separation in the East Indies which some, but not all, were able to cross; those which did succeed in crossing it spread widely through Australia, indicating more continental conditions, and also indi- cating in these familes a capacity for crossing marine barriers which enabled some of them to reach Madagascar, New Zealand and various Pacific islands. The ratio of their abundance in regional faune is apt to be inversely to the full development of mammalian life. Where mammals are scanty, as in oceanic islands, lizards partly take their place; and this is true of some continental regions as well as of oceanic islands. In the typical continental fauna, the lizards are largely restricted to desert or rocky habitat and are of small size. Yet these last are the most typical mem- bers of the order.. They show what its primary adaptation was. Various readaptations appear, to fossorial, to aquatic, to arboreal or to terrestrial forest life, repeated again and again in different families and causing frequent parallel divergencies from the primary type. This primary type, I regard as an adaptation to a Mesozoic arid period. The moist uniform climatic phase of the early Tertiary would tend to develop large forest living and aquatic forms and restrict and provincialize the more typical lizards. During the middle and later Tertiary, the typical lizards would expand and multiply in numbers and variety, but, on account of their lack of adaptability to cold climate, their evolution was not so much a successive series of dispersals from a Holarctic center, as a provincial evolution from the arid centers of the great continents. Such a priort hypotheses are of little value, however, except as confirmed, modified or refuted by detailed study of the affinities and geographic distribution of the genera of each family, checked by a wider knowledge and more thorough study of the fossil forms. Until the fossil Lacertilia have been thoroughly studied and their affinities authoritatively esti- mated, any conclusion whatsoever as to the evolution and distribution of the order remains highly hypothetical. Dr. Gadow’s recent study" of the distribution of Cnemidophorus and its interpretation is an excellent example both of the value of such de- tailed studies and the need of carefully distinguishing between what the u4H. Gapow: “A Contribution to the Study of Evolution based upon the Mexican Species of Onemidophorus,” Proc. Zool. Soc. London, vol. 1, pp. 277-375. 1906. MATTHEW, CLIMATE AND EVOLUTION 291 data themselves indicate and what is assumed as true from other evidence. He concludes,— 1) That the species are the product of their environmental conditions; 2) That their dispersal center was in western Mexico, whence they have spread northeast as far as Texas and Florida, southwardly into South Amer- ica, northwestwardly into Lower California ; 3) That the primitive type was nearest the Texas and Florida species. He assumes—evidently on some other grounds— 1) That a great land area stretched out from Mexico far into the Pacific during the Tertiary all the way between Lower California and Central America ; 2) That the central tableland of Mexico was a vast fresh-water lake during most of the Tertiary ; 3) That Cuba was connected with the American mainland during the Oligo- cene (this assumption underlies the statement that, since the Floridian Cnemi- dophorus did not reach Cuba, its migration must have occurred as late as Miocene). Ortmann,’ reviewing this paper, takes, as proven by Gadow’s studies, not merely the points actually indicated but also the assumptions which are entirely unnecessary to explain the data but which Dr. Gadow evi- dently feels obliged to take for granted. In fact, these assumptions interfere with a reasonable interpretation rather than help it, and all of them are questionable, to say the least. The great Tertiary lake is, I suspect, on all fours with the vast interior “lakes” of the Plains region of the United States, which the progress of physiographic and paleon- tologic studies have relegated to the domain of myth. The connection of Cuba with the mainland of either North or South America inyolves the same difficulties as the connection of Madagascar with Africa. The recent discoveries by Dr. de La Torre of a Pleistocene vertebrate fauna in Cuba strongly confirm this analogy between the Cuban and Malagasy faune. ‘The existence of extensive land west of the present Pacific coast line is an equally unnecessary and improbable hypothesis. On the other hand, Dr. Gadow fails to take into account the barrier between North and South America which prevented or hindered intercommunication of land faune during a large part of the Tertiary, while it permitted inter- communication of marine faune during the Eocene. I am not here concerned with its nature but may venture to point out that its bearing on the differentiation of species would be important. For, once across that barrier, an invading species would find itself in unfamiliar environ- ment on account of differences in the autochthonic fauna and flora, even m5 A, BH, ORTMANN: Geog. Jahrb., vol. xxxi, p. 262. 1908. 992 ANNALS NEW YORK ACADEMY OF SCIENCES though the physical environment were similar. If the rising of the Mexican tableland conditioned the dispersal of the genus from that center, we can see in this different biotic environment the reason why the marginal species in North America should be primitive, while the marginal species in South America are highly specialized. In general, it would be true that the species of the dispersal center (or those nearest to it, where, as in this case, it has become ill adapted for the habitat of the race) will be the most progressive and those of the marginal areas nearest the primitive stock. But where the scattering primitive forms, in fol- lowing the primitive climatic conditions, are brought into a new floral and faunal environment, this may profoundly modify them and cause a rapid divergence and specialization. DISPERSAL OF BrrDs As a class, birds are extremely difficult in their taxonomy. They are held closely to type in comparison with mammals, and the differences between them are mostly directly and obviously due to adaptation. Adaptive parallelism obscures the true affinities to such an extent that even at the present day the major classification is somewhat uncertain. This difficulty is the greater on account of their rarity as fossils. There is no reason to interpret this rarity as indicating any lack of abundance of birds in the faune of Tertiary and later Mesozoic time; it is presum- ably to be accounted for by their generally upland habit, small size and the lightness and fragility of the skeleton. The small minority of fossil birds which are known from anything more than a few fragments are, with two or three notable exceptions, aberrant types—ground-birds, marine or lacustrine types, whose habitat facilitated their preservation as fossils. By far the most notable and instructive of these exceptions is Archaeopteryx. It has been customary to class the greater number of the ground-birds (Ratite) as a more primitive sub-class. On @ priori grounds, this may be correct enough, since it would appear theoretically that feathers must have preceded flight, the ability to fly being conditioned by high organ- ization plus small size, and this would involve a rapid circulation and high temperature, which could hardly be attained without a nonconduct- ing coating over the body. But it appears certain that most, and possible that all of the existing ground-birds are readaptations to terrestrial habitat from flying ancestors, and their resemblances are due almost wholly to adaptive parallelism. Owing to their powers of flight, the dispersal of birds is much less MATTHEW, CLIMATE AND EVOLUTION 293 limited and conditioned by distribution of land and water or by moun- tain or desert barriers than is that of mammals. Climate and environ- ment are much more important factors. Their dispersal is accordingly much wider, and this is especially true of the more migratory and strong- flying types. The general course of their dispersal from the northern land masses is in some respects much more obvious than with the Mam- malia, provided we allow for the extreme imperfection of their geological record; but on this account, it is not supported by the mass of direct evidence which we have among mammals. The most primitive living birds, the penguins, are Antarctic in their distribution, and as fossils are known only from the Antarctic Tertiaries, where they include gigantic terrestrial adaptations. It is of interest to note that the only actually known land vertebrates of the Antarctic con- tinental area are penguins. If this continent had been united during the late Mesozoic and early Tertiary to Australia and South America, we should expect to find a fossil mammal fauna, probably highly pro- gressive and specialized before the spreading ice swept it out of existence. We might, indeed, -hope to find a few marine adaptations from this mam- malian fauna still haunting the edges of the Antarctic pack. But in fact, the three items which to my mind have a bearing upon early Ter- tiary conditions in Antarctica all point towards continued isolation and obviously parallel the fauna of oceanic islands. These are,— 1) Gigantic land-penguins in the ? Eocene deposits of Seymour Island (also in Patagonia). Compare with the gigantic land birds of various oceanic islands, correlated with paucity or absence of land mammals. 2) The living marine penguins are not readily interpreted as a pri- marily marine adaptation, but they are very easy to understand as modi- fied survivors of a group formerly of terrestrial habits, altered to meet the present conditions under which alone could life be maintained on the Antarctic shores. 3) The occurrence of Miolania, as interpreted on page 283, is sug- gestive of the former presence of giant land-turtles in Antarctica, al- though not explainable as evidence of former land connections with South America and Australia. There may be other indirect evidence in the distribution of marine Vertebrata and Invertebrata, which, if conservatively interpreted, would confirm or disprove these indications. So far as they go, they suggest that ground-birds and land-turtles were the large land vertebrates of Tertiary Antarctica as in oceanic island faune of to-day. The distribution of modern land birds is universally interpreted in 294. ANNALS NEW YORK ACADEMY OF SCIENCES terms of Northern derivation. Oceanic, desert or mountain barriers have been much less efficient in limiting their range, and the efficiency of the climatic factor is much more obvious than with mammals. Their dispersal from a Holarctic center in successive waves of migration is indicated by the dominantly Holarctic habitat of the highest and latest developed groups, by the generally tropical habitat of archaic groups often highly specialized, whose ancestors or relatives are in many cases known from the Holarctic Tertiary, and by the fact that the southern continents are peopled, not by a series of dominant groups corresponding to the Holarctic groups, evolved in a common Antarctic center, but chiefly by groups of more or less tropical affinities and by a few northern groups which have crossed the tropic barrier. There are many groups of birds living to-day in the widely separated tropical regions whose ancestors have not thus far been discovered in the Holarctic Tertiary. But they correspond, both in distribution and in relative position in the classifica- tion, with other groups which the geologic record proves to have origi- nated by dispersal from Holarctica, and there is no valid reason for assuming any other origin. The geologic record of Tertiary birds is far more fragmentary than that of Tertiary mammals and especially in the Nearctic region. It should further be observed that the perching birds represent the primary adaptation from which the various specializations—terrestrial, wading, marine, etc.—have diverged, and that, in consequence, these divergently specialized forms retain various archaic features which have been lost by the central group. The relations, dispersal and present distribution of birds are thus wholly in accord with the principles here set forth. The detailed appli- cation of these principles is beyond the limits of the present discussion. DISPERSAL OF AMPHIBIA The modern Amphibia include a few small and for the most part highly specialized survivors of a group whose period of dominance dates back to the Paleozoic. Of their Mesozoic and Tertiary ancestry almost nothing is known. The Stegocephalia, the dominant Amphibia of the Permian, were far less aberrant and much nearer to the contemporary primitive Amphibia; their interrelationships are still far from being pre- cisely definable, and, until these are better understood, it is futile to dis- cuss the evidence which they may furnish as to former geographic con- nections. The distribution of the modern Amphibia is often notably discontinu- ous, and in the absence of evidence from extinct types as to the real MATTHEW, CLIMATE AND EVOLUTION 295 origin of these discontinuous distributions they are interpreted by many authors as affording evidence for various transoceanic bridges. But they are not essentially different from various instances of discontinuous dis- tribution among Mammalia, except that they are probably in some cases of more ancient origin, and are less restricted by ocean barriers. The urodele Amphibia are Holarctic, save for one family, Plethodon- tide, which has spread into northern South America and has also reached Hayti. Although thus limited in dispersal, they would seem to be an ancient group represented as far back as the Wealden by Hylwobatrachus, said to be related to the modern Cryptobranchus.*® Their distribution within Holarctica is more or less of a relict type, broken up by the unfa- vorable environment of so large a part of this region, especially of the central portion. The cecilians are tropical but have not reached Aus- tralasia. The frogs and toads have a wide dispersal, and so far as a superficial view may show, the most primitive or archaic families are limited to the peripheral continents and oceanic islands, while the more progressive groups are more cosmopolitan, but have not yet reached all of the outly- ing regions. Some of the families, at least, would appear to be of ancient origin; Paleobatrachus, allied according to Gadow'** to the Aglossa of the Ethiopian and Neotropical regions, is recorded from the Jurassic of Spain, and is said to be common in the older Tertiary of Europe. Among the modern families the Cystignathide are chiefly Australasian and Neo- tropical, but a few are still found in North America. This distribution parallels that of the polyprotodont marsupials, except that the latter have not reached New Zealand or the Antilles, or entirely disappeared from the East Indian islands. The Discoglosside inhabit the East Indies and North America but have disappeared from the intervening portion of Holarctica; Discoglossus and other genera are found in the Middle Ter- tiary of Germany. The Pelobatidz stretch across Europe and Asia and northwestern North America. These three families represent evidently three successive dispersals. The other families are more cosmopolitan. The genus Bufo has failed to reach Australasia, Madagascar or New Zealand, but is replaced in Australia by a (more primitive?) member of the family. The Hylide are to-day chiefly South American and Australian, but a few members still inhabit North America. They are not found in Africa or the Orien- tal region, where it seems reasonable to suppose that they have been dis- placed by the true frogs (Ranid), peculiarly varied and abundant in 16. Broiwi, in Zittel’s Grundziige der Paleont., Vertebrata, s. 176. 1911. iH. (GApow: 1. cs pil45. L900. 296 ANNALS NEW YORK ACADEMY OF SCIENCES these regions. The Ranide, lke the Bufonide, represent a less ancient dispersal, probably from a southern Palearctic or Oriental center, since they have reached northern Australia on one side and northwestern South America on the other, and, while they have reached Madagascar and the Solomon Islands, they have failed to reach the Antilles. These suggested lines of dispersal are based upon the present distribu- tion interpreted in accord with the principles outlined in previous pages of this article. While the past history of the Amphibia is too little known Cyslignalkidde (scogloss(dae Pelobatidae Fic. 38.—Distribution of three families of Anura These may be interpreted as due to three successive dispersals from the north. The other families of frogs and toads are more widely spread, and their regional abundance has conditioned certain peculiarities in the distributions here shown. to confirm them by adequate direct evidence, I believe that good infer- ential evidence might be obtained from a comparison of the progressive or archaic characters of the skeleton in the different families. The fossil Amphibia afford sufficient evidence to determine the broader lines of their evolution and differentiation, although they tell very little about their past distribution. The same conditions hold true with regard to the fresh-water fishes. MATTHEW, CLIMATE AND EVOLUTION 297 DISPERSAL OF FRESH-WATER FISHES The fresh-water fishes afford many striking illustrations of isolated primitive survivals in the southern continents and especially in their tropical parts. With marine fishes, the distribution is wider, as we should expect, and the dominant types are generally world-wide in their distri- bution. Yet, even with marine fishes, a superficial survey seems to show the majority of primitive survivals along the southern coasts. Fishes are, it is to be remembered, dominantly marine. The wider field and more varied opportunities for development afforded by the ocean waters, in contrast with the limited and isolated fields and uncer- tain tenure afforded by fresh-water rivers and lakes, have conditioned this. The fresh-water habitat for aquatic groups of animals stands in somewhat the same relative position to the marine habitat as does the insular to the continental habitat for land animals. It is the refuge for survivors of primitive faune. And, as in the insular land faune, we are constantly confronted there with the occurrence in widely remote regions of archaic types apparently nearly related, whose similarity is partly due to independent adaptation to a similar environment, partly to persistent primitivism. Lepidosiren in tropical South America, Protopterus in tropical Africa, Ceratodus in tropical Australia are perhaps the most prominent examples of extremely ancient survivals. These are survivors of early Mesozoic or even Paleozoic marine and estuarine fishes of world-wide distribution, and they have endured, in their tropical refuge, the several successive periods of zonal climate which affected the environment of temperate and tropical regions. More pertinent to the problem in hand are the relationships of early Tertiary fishes of the northern continents to the modern South American, African (and Australian?) fishes. Here, again, I am compelled to dis- sent from the interpretations and conclusions of so distinguished an au- thority as Dr. Eigenmann,"'* who, as it seems to me ignores certain very important parts of the evidence. There is a marked similarity between certain parts of the fresh-water fish faune of South America and of Africa. Higenmann and others would explain this by a former continental union, but it is certain that some, at least, of these now tropical types existed in the northern conti- nents during the early Tertiary. Higenmann’’® asserts, indeed, that no 188 See especially C. H. EIGENMANN: ‘‘Fresh-water Fishes of Patagonia,’’ Reports Prine. Univ. Exped. Patagonia, vol. iii, parts iii-iv. 1909-10. 19 C, H. EIGENMANN: Popular Science Monthly, 1906, p. 523. 298 ANNALS NEW YORK ACADEMY OF SCIENCES part of the modern South American fresh-water fish fauna is derived from North America; but how he reconciles this with the recorded pres- ence of several of the most typical genera in the Green River Eocene of Wyoming, I do not see. A few cases in point may be noted, as follows: Lepidosteus, now Central American and southern Sonoran. Abundant in all the Eocene formations of the northwestern States, as also in Europe. Phractocephalus, Arius, etc., now South American, nearly related to Rhineastes of the Bridger and Amyzon beds of the western States. Osteoglossus of Brazil, Borneo and New Zealand, Vastres and Hetero- tis, also southern types, closely related to Dapedoglossus of the Green River shales (Kocene). The characins, which form so important an element of the modern South American fauna, are, as Eigenmann holds, largely a local expan- sive radiation conditioned by the immense ramifying river-systems of that continent. But, considered in their more general relations, they are a primitive group, the northern cyprinids being a higher and later de- velopment. The catfish, which in the North have the characteristics of a disappear- ing group, are numerous and dominant in South America. Eigenmann calls attention to the paucity of the Patagonian fauna and its apparent relations to that of New Zealand and Australia (Galaxiide and Ap- lochitonide). He does not, however, attach any great weight to this as evidence for a former Antarctic connection, regarding it as “highly theo- retical and precarious” so far as the fresh-water fish are concerned—but “The evidence from other sources of a former land connection has be- come conclusive.” I might observe here that many students in other eroups are equally doubtful of the conclusiveness of the evidence for Antarctic connections in the groups with which they are familiar, while equally ready to accept as conclusive the evidence in groups with which they are not familiar. As regards a connection of tropical Africa with tropical South Amer- ica, Kigenmann is much more positive, basing it mainly upon the chara- cins and cichlids, common to both continents. There is no species or genus common to the two continents. Both families are relatively primi- tive, as compared with northern related groups. As regards their former presence in the northern world (which EKigenmann does not allude to) or their parallel adaptation from marine forms of Cretaceous or early Ter- tiary time, there is little satisfactory evidence. Nevertheless, the fact that they represent an adaptive divergence from an intermediate and more primitive type ancestral to carp and catfish is a suggestive one. MATTHEW, CLIMATE AND EVOLUTION 299 If now we compare the general relations of tropical fresh-water fishes with those of the North, it will appear very clearly that the highest and latest in appearance of the several groups are still limited to the northern world, and that, in the tropics, more primitive groups exist, many of them known to be former residents of the northern world, others much nearer to known or inferred ancestral groups than are any members of the present northern fish fauna. Where the environment favors, some of these groups have branched out into an immense variety and number, far exceeding what is‘known in the colder north. But they are distinctly less progressive. In the southern continents, we meet with some remark- able parallelisms to the dominant types of the North, very suggestive at first of Antarctic connections, but probably explainable (as in Galaaias) in other ways. ‘These groups impress one as highly progressive, although less so than the northern groups; but they do not appear to have con- tributed materially to the tropical faune. In some respects the fresh-water fishes present nearer analogies to the birds than to mammals in their distribution; and this is no doubt con- ditioned by their less strict limitation to land connections for their mi- gration, and to the greater antiquity of the class. GENERAL CONSIDERATIONS ON THE DISTRIBUTION OF INVERTEBRATES AND PLANTS It would be unwise to attempt any survey of the paleogeographic data afforded by invertebrates and plants. Lacking both the special knowl- edge necessary for a critical consideration of the data, and the time neces- sary to make even an adequate compilation, it would add nothing to the argument. While, for reasons already given (page 272), placing most weight on the evidence obtainable from mammals, I fully recognize the importance and variety of evidence outside the Vertebrata, and the force which attaches to cumulative evidence from several independent sources. At the same time I must express a strong conviction that the sources are not really independent, and that concordant results in several groups which flatly contradict the results obtained by a study of mammals, can only indicate one of two things. Either the interpretation of the evi- dence among the Vertebrata is incorrect or there are factors of error common to the interpretation of the several other groups which accord in their disagreement. What these factors may be, I have already indi- cated and have attempted to show that they account for discordant results based upon the distribution of the lower vertebrates and interpreted as involving radical changes between continental and abyssal regions which 300 ANNALS NEW YORK ACADEMY OF SCIENCES are highly improbable, to say the least, from a geological point of view, and which are not merely unnecessary but apparently impossible when we attempt to explain the distribution of the higher vertebrates in accord- ance with them. It is true that the evidence against such changes in pre-Tertiary times is less weighty, and that it diminishes further in the older periods of geologic time. And the antiquity of many groups of invertebrates, especially of land invertebrates, makes it impossible to limit the hypo- thetical land bridges which their distribution is supposed to require, to the Tertiary or even the Mesozoic. ‘The permanency of the ocean basins in the older geologic epochs is beyond the limits of this discussion. So far as a superficial acquaintance shows, the general distributional relations of most land invertebrata and of plants appear to me to accord with those of the mammalia. Primitive and archaic’*® types abound chiefly in the tropics. The most progressive and dominant types are Holarctic. The southern continents show common groups suggestive of an Antarctic radiation, but which may, like the marsupials or the chryso- chloroid insectivores, be remnants of formerly cosmopolitan groups whose resemblance is due rather to persistence or to parallel evolution under similar climatic stimulus than to such close affinity as would involve Antarctic continental connections. Where, as in the earthworms, we have no knowledge at all of their past distribution, it is impossible to test this interpretation of their present distribution ; nor in such a group does it seem possible to estimate how much and in what manner slow progressive climatic change might affect their structural evolution, although climatic conditions are evidently important in controlling their range. The point that I desire to emphasize is that, if such an interpretation as I have suggested be possible, it should be accepted in preference to one which would involve such unexplainable difficulties in the distribu- tion of the higher animals and such improbable physiographic changes. No hypothesis can be finally accepted that does not conform to the facts of distribution in all groups of animals and plants. It is not a matter of preponderant evidence. Every anomaly must be explained, every dis- tributional fact must be interpreted in accord with the rest, before we can consider theories of paleogeography as conclusively proven. It is not sufficient that the evidence in one group or in ten groups has been interpreted on concordant lines, so long as there remains an eleventh group which cannot be so interpreted. But, pending a final agreement .° Archaic is used in the sense of divergently specialized but little progressive. MATTHEW, CLIMATE AND EVOLUTION 301 in our deductions from the evidence afforded by the various classes, it appears to me that we should hold to conservative views rather than adopt hypotheses of continental relations so much at variance with gen- erally accepted geological principles and inferences. To illustrate the point that these discrepancies are a matter rather of interpretation. than data I may venture to discuss one or two instances among invertebrates prominently used in paleogeography. INTERPRETATION OF DISTRIBUTION DATA OF CRAYFISH I am indebted for my data on this interesting group to Dr. Ortmann’s valuable discussion of the geographical distribution of fresh-water Decapoda.**t The interpretation, however, which I would place upon the facts differs widely from his. As Professor Huxley has observed, the real difficulty in explaining the distribution of the crayfish is in their occurrence in the north and south temperate zones, separated by a wide tropical belt in which none now occur or are known to have occurred in the past. Two explanations offer themselves : 1) Independent adaptation from marine types in the northern and southern hemispheres. This would involve either former Antarctic con- nections or independent adaptation also of the several southern groups from marine types. 2) Former cosmopolitan distribution of crayfish, with subsequent dis- appearance from the tropical belt and differentiation of the isolated south- ern groups and of the more progressive northern groups. The latter view is generally accepted, and seems to me more consonant with the facts of distribution, e. g., presence of crayfish in Madagascar, while they are absent from South Africa. I am unable to agree with Dr. Ortmann that crayfish on oceanic islands necessarily involve a former land connection, since such land connections as he finds it necessary to postulate would apparently involve the presence on these islands of con- tinental faunz which are not now present, and whose absence cannot be reasonably accounted for. For the reasons already presented I see no difficulty in supposing that the crayfish of Cuba, Madagascar, New Zealand or Fiji have reached those islands by accidental transport of natural “rafts” through the agency of ocean currents, or by other acci- dental means. The Australian and South American crayfish I should regard as derived from the north, by way of the existing or slightly sub- 1A. E. ORTMANN: “Geographical Distribution of Fresh-water Decapods and its Bear- ing upon Ancient Geography,” Proc. Amer, Phil. Soc., vol. xli, pp. 267-400. 1902. 302 ANNALS NEW YORK ACADEMY OF SCIENCES merged land bridges,.at a time when the northern crayfish were much more primitive than now, and when, for reasons which I do not venture to suggest, the tropics were a more favorable environment than now. The northern crayfish have since evolved into Potamobius and Cambarus, the southern specialized into the more divergent Parastacus of South America, Cheraps and Hugeus and Astacopsis of Australia and Tasmania, Paranephrops of New Zealand and ? Fiji arid Astacoides of Madagascar, Of these southern genera, Astacoides is the nearest to the northern types. This is to be expected, if the southern genera are remnants of a cosmopohtan distribution derived by dispersal from the north; for the Malagasy genus would be a derivative from Ethiopian crayfish, which would be Jess remote from the north, and would be correspondingly more advanced than in South America or Australia. As far as the more special distribution of the northern crayfish is concerned, Dr. Ortmann’s paper affords data for the following interpretations. Two genera are concerned, Cambarus of the eastern Sonoran region, and Potamobius (Astacus of most authors) of the Old World and western Sonoran region. In his discussion of the genus Cambarus Ortmann states that the more primitive forms of the first, second and fifth groups belong chiefly to the south towards Mexico, and interprets this as meaning that the genus came from Mexico. But, according to the principles here adopted, this should mean that the center of dispersal is to the north and east; and the discontinuity in range to the south and west is exactly what we should expect, Dr. Ortmann’s attempt to find an explanation for it on the opposite theory of migration being curiously complex and unconvine- ing. The most primitive species occur in such widely divergent points as Mexico and Cuba. The more primitive genus Potamobius has a more discontinuous range, in Europe, part of Eastern Asia and Western North America, the Asiatic species being nearest to Cambarus (i. e., highest in development) but parallel, not truly closely related. This, I take it, is correctly interpreted by Ortmann as indicating an Asiatic center of dispersal for this genus. But in place of supposing with Ortmann that Cambarus originated from species of Potamobius pushing down southward into Mexico and thence northward again (as Cambarus) into the United States, it seems to me that the rational explanation would be to suppose that both genera are the disconnected remnants of a formerly Arctic center of dispersal. This would be first split in two by a progressively unfavorable environment, one division passing down into America east of the Cordilleras, and developing into Cambarus, the other part in Asia progressing more MATTHEW, CLIMATE AND EVOLUTION 303 slowly into Potamobius and spreading east and west from that center, as the American group spread southward. DISTRIBUTION OF HELIX HORTENSIS Dr. Scharff!?? regards the distribution of Helia hortensis as an im- portant part of the evidence in favor of a late Cenozoic bridge connecting Europe with eastern North America. The species is well known in Europe and has always been regarded as indigenous there. It occurs along the North Atlantic coast, and in Labrador, Greenland, Iceland and the Shetland and Faroé Islands. It was formerly considered as intro- duced on this side of the Atlantic by human agency; but it has been found in old Indian shell-mounds and more recently in undoubtedly Pleistocene deposits in Maine. It is unknown in Asia or western North America. Hence, Dr. Scharff concludes that it must have migrated from _ Europe to America across a land bridge via Iceland and Greenland in Pliocene or Pleistocene times. The early opinion that Helix hortensis is an introduced species in this country was founded, so far as I recall, mainly upon the peculiar local range and habitat of the species, very different from the truly indigenous New England land-snails, and my early experiences in land-snail collect- ing in southern New Brunswick were quite in accord with this evidence. It is quite possible that Helix hortensis, like the genus Hquus, is both introduced and indigenous. Granting that it is at least partially indigenous, what evidence is there that the present distribution is not the remnant of a Tertiary circumpolar distribution? The fact that it is not recorded in the Tertiary of Asia? But what proportion of the presumably abundant Miocene or Pliocene land-snails of Asia is known to us? It can only be a minute fraction at the best—less than one per cent. So the chances are a hundred to one that if Helix hortensis or an ancestral form of the species existed in the Tertiary of North Asia, we should have no record of its existence at present. We do, however, have a good deal of indirect evidence that an environment favorable to the present habits of the species existed during the later Tertiary in the region intervening between its present discon- tinuous distribution areas, and that the environment became unfavorable in that intervening region at the close of the Tertiary. I can see no need for assuming a transatlantic land bridge to account for the distribu- tion of this species. And the explanation here suggested is in harmony with the known course of distribution of those members of the northern 22 R, F. ScHarrr: Proc. Roy. Irish Acad., vol. xxviii, p. 19. 1909. 304 ANNALS NEW YORK ACADEMY OF SCIENCES land faunz whose past history is preserved to us in the geologic record. It involves only those minor changes of continental level (a few hundred feet) of whose occurrence during the Pleistocene we have ample evidence. On the other hand, if we assume such a Transatlantic land bridge during the late Tertiary we must suppose an elevation of upwards of five thousand feet, a huge disturbance of the isostatic balance of whose possibility we have no real evidence; for the submerged channels so often cited in support of these immense uplifts have been shown by Chamberlin to be much more probably due to “continental creep,” to the slipping” down, so to speak, of marginal sediments to a lower level.'** In any ease, there could be no evidence as to the period at which these old channels were last above water. They may have been submerged since the Permian, for aught we know to the contrary. Furthermore, we have to explain the non-migration of a multitude of forms which got just so far as conservative land elevations could carry them, but no farther. DISTRIBUTION OF PERCIDA Another instance upon which Dr. Scharff lays great stress is the dis- tribution of the perches. Here, the false impression produced by the use of a Mercator’s projection map in plotting the distribution of north- ern forms, seems to me to be very obvious. ‘This map does not give the northern regions in their true proportions or relations. Transferring the distribution of this family as plotted by Tate Regan, to a north polar projection map we get the real relations and proportions with approxi- mate correctness. It then becomes obvious that the perches are centered around the drainage basin of the Arctic ocean. In North America they have extended down the Atlantic coast drainage area and into that of the Gulf of Mexico as far as the Rio Grande. In Asia they have been ad- mitted by the old Hyrcanean Sea into the present Caspian and Aral basins; and a glance at the late Tertiary geography of Europe will show how they have reached the drainage basin of the northern Mediterranean. They are not now found in the Arctic drainage area of western North America, Greenland or Iceland, where the environment, now or in the Pleistocene, is amply sufficient to account for their extinction. What need of a transatlantic land bridge to account for this distribution. 23There is another possible explanation. The progressive building out seaward of barrier reefs around a number of separate centers until they joined into a platform would naturally leave deep intervening channels, especially off the mouths of great rivers where the influx of mud and fresh water hindered the growth of the coral organisms. The submarine contours around the West Indian islands especially suggest this explana- tion, which I offer tentatively for the consideration of my better-versed confréres, MATTHEW, CLIMATE AND EVOLUTION 305 A fourth instance cited by Dr. Scharff is the distribution of the river- mussel Margaritana, and as he well observes, numerous other instances would probably show similar discontinuous distribution. But, so far as I have been able to find such instances, the same reasoning and the same explanation apply to them all. Criticism oF Some Opposing HypoTHEsEs It is not practicable to take account here of the flood of paleogeo- graphic discussions of recent years which have advocated all sorts of consistent or inconsistent changes in continental outlines. They agree for the most part in failing to take into account certain considerations which to my mind are essential elements in any problem of distribution. Among the geological considerations are the following: 1) Evidence that the present distribution of the deep ocean basins is in the main due to isostatic balance. This affords a strong presumption in favor of its permanence. 2) Absence of abysmal deposits in the geological formations of any continental region. Chalk deposits are not an exception, as it has been shown that they were deposited in shallow epicontinental seas rather than in deep oceanic basins. 3) Abrupt ending of an elevated line of disturbance and its continua- tion as a submerged line of disturbance does not necessarily indicate that the submerged portion was formerly elevated, although it does reduce the improbability of its former elevation by indicating a line of dis- turbance and hence of possible elevation. 4) The presence of marine formations of Cretaceous or Tertiary age over large portions of the interior of the great continents does not indi- cate that these continents first came into existence as such during the Cretaceous or Tertiary. In the better known portions of the earth’s surface we know well enough that these marine formations were due to ‘periodic temporary submergence, interrupted by periods of more or less complete emergence. It is but reasonable to apply the same explanation to the less known regions. I see no more reason to suppose, as do Von Thering, Scharff and others, that South America first came into existence as a united continent in the Tertiary, than to conclude on similar evi- dence that North America was but a group of isolated land masses until the end of the Cretaceous. In this country, we have positive proof of its antiquity; but the evidence for recent origin of the South American would apply just as well to the North American continent. A similar presumption of antiquity applies to Australia, Asia and Africa. 306 ANNALS NEW YORK ACADEMY OF SCIENCES Among zodlogical considerations we may mention the following: 1) The discontinuous distribution of modern species is again and again taken as proof that the regions now inhabited must have been con- nected across deep oceanic basins, without considering the possibility that it is a remnant of a wider past distribution, or that it is due to parallel evolution from a more primitive type of intermediate distribution, now extinct. Yet so many instances are known where the geological record has furnished proof that one or other of these explanations applies to cases of discontinuous distribution, that it would seem that these ought to be the first solutions of the problem to be considered, and that in view of the known imperfection of the geologic record, mere negative evidence is not sufficient to cause them to be set aside. 2) No account is taken of faunal interchanges often much more ex- tensive, which would presumably have taken place if the land bridges assumed had existed, but which have not taken place. It may here be urged that this too is negative evidence. But the negative evidence de- rived from an appeal to the geological record is weak, not per se, but because of the demonstrated imperfection of this record. On the other hand, there are many instances where a land bridge is well proven, and in these cases it is not a few scattered exceptions but an entire fauna that has migrated, subject only to the restrictions imposed by climatic or topographic barriers of other kinds. I may venture upon a discussion of a few instances in order to show the type of objections which appear to me to apply to much of the evi- dence cited in favor of most of these transoceanic land bridges. On VAIN SPECULATIONS According to some distinguished paleontologists,1** progress is to be made only by ignoring the possibility that races have originated in or migrated from regions of whose former life we have substantially no record, and assuming that they must have evolved in one or another re- gion where the record is more or less known, and that the actual record must be the sole basis for any conclusions. They refuse to consider the arguments for origin elsewhere, on the ground that such hypotheses are “vain speculations” and “serve merely to conceal our ignorance.” To this I may answer that a fair and full consideration of the data at hand shows that such hypotheses, of one kind or another, are absolutely necessary, unless we are to abandon all belief in the actuality of evolution and are to treat it as merely a convenient arrangement of successive spe- 124 Depéret, Thévenin and others. MATTHEW, CLIMATE AND EVOLUTION 307 cies and faunas independently created. Such a view was held by Agassiz and most of his predecessors, but it is unnecessary to consider it in the present state of scientific belief. If, on the other hand, we accept the belief that the successive species of each phylum are genetically related, how are we to explain the fact that these phyla are usually approximate and not direct, and that where the evidence is most complete, the fact that they are not in a direct lne of structural evolution stands out most clearly. Take for example the ancestry of the horse, the most striking, easily recognizable, widely known and thoroughly studied illustration of mammalian evolution. It was possible, when the “documents” were few and imperfect, to trace a sup- posedly direct line of ancestry through European predecessors. Later, when the fossil fields of the western United States were first explored, a much more direct line of ancestry was found in this country, and the European series was recognized as not being the direct line. But the further progress of exploration in America, and the discovery of complete skeletons of the supposed ancestral stages known at first only from frag- mentary specimens, has demonstrated that this line too is an indirect and approximate series so far as the succession of the known species is con- cerned. This has been recognized in recent years by American students, and variously phrased or interpreted. The most probable explanation of the facts is to suppose that the known phylum is approximate, not direct ; that the direct line of descent leads through unknown or imperfectly known species, and that those known to us are offshoots of varying close- ness. The direct line is, then, admittedly through hypothetical species, and the only question is whether the habitat of these species was in the regions where we have searched vainly for their remains, or in the much greater intervening region where we have not searched. Horses are found throughout the Tertiary in central and western Europe on the one hand, on the Western Plains of America on the other. There is every reason to believe that they inhabited all or parts of the intervening region and we have no right at all, in weighing the evidence, to refuse to take this re- gion into consideration, on the plea that it has furnished no “documents” as yet. To place such limitations on our theories would hardly tend to solving our problems, however much it might seem to simplify them. It is merely to prefer a conclusion that we know to be false to a conclusion that we cannot prove by direct evidence to be true. What I have stated in regard to the fossil ancestry of the horse applies to most mammalian phyla, in greater or less degree according to the per- fection and number of our “documents.” Where these are few and frag- mentary, it is still possible to build up phyla which cannot be proven to 308 ANNALS NEW YORK ACADEMY OF SCIENCES be inexact. But, as knowledge increases and becomes more exact, these phyla are more and more broken and complex, and direct genetic series become more limited in extent. This is to be expected, for the regions which up to date have been at all thoroughly explored are but a small fraction of the area which the group concerned must have inhabited. And on @ priori grounds, the chances are greatly against the particular species which was to become dominant inhabiting the particular regions which we have explored. Professor H. F. Osborn has very well expressed the conditions of evo- lutionary progress by stating that each group is highly polyphyletic, con- sisting of numerous subphyla evolving along more or less parallel lines. But we are here concerned less with the disentaglement of the subphyla of a group than with its dominant center of dispersal as a whole. And from this point of view it seems to me misleading and erroneous to as- sume that it must have migrated only from one to others of the regions where its remains have actually been found, instead of attempting to locate from the indirect evidence available the true center of dispersal. In contrast with the views here criticized, I may venture to quote from an address in which Dr. Stehlin?®® has recently summarized the phylo- genetic results of his monumental studies upon the Eocene fauna of Eger- kingen, a work of extraordinary thoroughness and ability which, as a recent reviewer observed, has involved a revision of the entire Kocene mammal fauna of Europe: “Where then dwelt these yet unknown herds of mammals evolved during the Eocene, whose existence is recorded through their influence upon Europe and North America the more clearly as we analyze more closely the data obtained in these continents? We can scarcely be wrong if we look to the huge continental mass of Asia, still almost unexplored by the paleontologist. The future, and, it may be hoped, the near future will show how far our present anticipations are correct.” SUMMARY OF EVIDENCE The geologic evidence for the general permanency of the abyssal oceans is overwhelmingly strong. The continental and oceanic areas are now maintained at their different levels chiefly through isostatic balance, and it is difficult to believe that they could formerly have been reversed to any extensive degree. The floor of the ocean differs notably in its relief from the surfaces of the continents, and only in a few limited areas is the relief suggestive of former elevation above sea-level. The continental shelf is 2H. G. STEHLIN: Verh. Schw. Naturf. Gesell., 93 Jahresversammlung, Sept. 1910. P. 29 of separata. MATTHEW, CLIMATE AND EVOLUTION 309 so marked, obvious and universal a feature of the earth’s surface that it affords the strongest kind of evidence of the antiquity of the ocean basins and the limits beyond which the continents have not extended. The supposed evidence for greater elevation in the erosion channels across its margin have been shown to be better interpreted as due to “continental creep.” The marine formations now found in continental areas have all been deposited in shallow seas. No abyssal deposits have ever been cer- tainly recognized among the geologic formations of the continental plat- form. Leaving out of consideration speculative hypotheses as to a formerly smaller amount of water on the surface of the globe, shallower ocean basins in Paleozoic times and different land and water distribution in the older geological periods, it is sufficient for the purposes of this discussion to emphasize the great weight of geological and physiographic evidence for the permanency of the continental masses as outlined by the conti- nental shelf, during the later geological periods, and especially during the Tertiary. The present distribution of continents and oceans on the surface of the globe (as outlined by the continental shelf) consists of a great irregular northern mass including Europe, Asia and North America, with three great partly isolated projections into equatorial and southern latitudes, South America, Africa and Australasia, and a smaller Antarctic land mass wholly isolated. The three peripheral continents are isolated from each other and from the Antarctic land by broad and deep oceans, but with the doubtful exception of Australasia, are united to the central mass by shallow water or restricted land connections. A rise of 100 fathoms would unite all the continents and continental islands, except perhaps Australia, into a single mass, but would leave Antarctica, New Zealand, Madagascar, Cuba and many smaller islands separate. A further elevation of five times this amount would not alter materially the boundaries of land and sea. A submergence of 100 fath- oms would isolate the three southern continents, and cause shallow seas to spread widely over the interior of all the continental masses, reducing some of them to isolated fragments or archipelagoes. Such cyclic alternations of emergence and overflow are recognized by many geologists as the dominant feature of the earth’s history, corre- sponding to the succession of periods into which geologic time is divided. The greater disturbances resulting in folding, faulting and mountain making, while involving much greater changes of level, affect more lim- ited areas, adjacent to lines of unstable equilibrium, especially along the borders of the continental platforms. 310 ANNALS NEW YORK ACADEMY OF SOIENCES Associated with these great cycles of elevation and submergence are climatic cycles from extremes of cold or arid zonar climates culminating in glacial epochs, to the extremes of warm humid uniform climates which accompany or follow the extremes of submergence. The effect upon terrestrial life of progressive elevation of the land areas, accompanied by a progressively cold climate at the poles and arid climate in the interior of continents, would be to adapt the terrestrial life to cold, arid and highly variable climatic conditions. The environ- ment favorable to this adaptation will appear first near the poles, and the northern and southern faune will be more progressive and will tend to disperse towards the equatorial regions. The wider area of emerged con- tinents will tend to expansive evolution of the land faunz, and their union into a single land mass will facilitate cosmopolitan distribution. Owing to the conformation of the continents the dispersal will be chiefly from the Holarctie region, the Antarctic and southern lands being unfa- vorably situated for the evolution and dispersal of dominant races and contributing but little to the cosmopolitan faune of the emergent phase. These conditions are also favorable to the development of higher, more active and more adaptable types of terrestrial life, which tend to supplant even in moist tropical regions the less adaptable remnants of the tropical faune which find there their last refuge. During the opposite phase of the cycle, the faunz become progressively readapted to the moist tropical climatic environment. But owing to the higher evolutionary stage acquired during the arid phase, the higher and dominant types of the new fauna are evolved chiefly by readaptation of the dominant types of the arid phase and only subordinately by expansive evolution of the tropical fauna surviving through that phase. The paleontologic record appears to be in exact harmony with these principles, provided due allowance be made for its imperfections. The geographic distribution of animals and plants affords far more complete data, but their true significance has in my opinion been misinterpreted by many zodgeographers. When interpreted in harmony with the prin- ciples of dispersal shown to be true among mammals, they yield fully concordant results. The geologic record is to-day far more incomplete than is generally admitted, and will always be incomplete. Negative evi- dence, while sometimes of high value, is more often worthless and should never be admitted without a careful canvass of the situation in each instance. The population of oceanic islands is notably incomplete and cannot be interpreted as due to continental connection. The difficulties in the way of over-sea transportation are best explained by the hypothesis of natural MATTHEW, CLIMATE AND EVOLUTION 311 rafts; the degree of probability that attaches to this hypothesis is esti- mated. The dispersal of mammals is then considered at some length, order by order, and it is shown to accord fully and in detail with the principles here set forth, and to be impossible of explanation except upon the theory of permanence of the ocean basins during the Cenozoic era. While the prominence of the Holarctic region as a center of dispersal is ascribed to its central position and greater area, some evidence is given to show that climate is also a factor in the greater progressiveness of the northern, since it is also noticeable in the southern as compared with tropical faune. The distribution of the Reptilia appears to be in conformity with the principles here outlined, and extends their application to the Mesozoic era. The distribution of birds and fishes and of invertebrates and plants is probably in accord with the same general principles, modified by differ- ences in methods of dispersal. The opposing conclusions that have been drawn from the distribution of these groups are believed to be due to an incorrect interpretation of the evidence. A few instances, which have been prominently used to support opposing conclusions, are analyzed and shown to conform to the conclusions above set forth, if interpreted upon similar lines as the data for mammalian distribution. APPENDIX Since this paper was written two very readable and instructive books on geographic distribution have appeared, “The Wanderings of Animals” by Professor Gadow,?*° and “Distribution and Origin of Life in America” by Professor Scharff.1°*7 Both writers, and especially Doctor Scharff, be- long to what may be called the bridge-building school of paleogeography, and the general criticisms expressed in the earlier part of this article apply largely to their interpretations. It is with no intent to depreciate their value that I observe that there are numerous errors of fact in those portions of the evidence with which I am best acquainted, for in a subject of so wide a scope most of the evidence is necessarily compiled and not very well understood, and errors more or less essential will slip in. It is for that reason that I have avoided detailed discussion of the parts of the evidence on the present subject with which I am not well acquainted ; and, in spite of a good deal of checking and revision, I have no doubt that the foregoing discussion contains various inaccuracies. 26 HANS Gapow: “The Wanderings of Animals.’’ Cambridge Manuals of Science and Literature, No. 64. 1913. VR, F. Scuarrr: Distribution and Origin of Life in America. Macmillan Co., pub- lishers, New York. 1912. 312 ANNALS NEW YORK ACADEMY OF SCIENCES A more serious criticism is the illegitimate and often partisan use made of negative evidence. This is doubtless due to the same cause, a mere book knowledge of the fossil record, and failure to examine and weigh its evidence. But it is very obviously affected by a readiness to rely on negative evidence that favors their theories and to ignore a vastly greater amount of negative evidence that does not. \” forsthe theory of Holaretic dis- Dr. Gadow considers it “awkware¢ persal of the marsupials in the Cretaceous that no survivors have been recorded in the Tertiary of Asia. He prefers to believe that the Aus- tralian marsupials arrived via Antarctica from South America. If it is “awkward” for the one theory, that although survivors are found in the early Tertiary of both Europe and North America, none have been found in Asia, then it must be equally “awkward” for the theory that Dr. Gadow supports that none have been found in Antarctica. For we know even less about the early Tertiary of Asia than we do about the Antarctic Tertiary. If the absence of zalambdodont insectivores in the Eocene of Europe is to be assigned any weight, then equal weight should be assigned to their absence from the Oligocene and Eocene of South America and from the Pleistocene of Cuba, of Madagascar and South Africa. We know as much about the one fauna as we do about the others. The negative evidence has no weight in any of these instances; per contra, the fact that zalambdodonts are known to have lived in the early Tertiary of North America (Paleocene to Oligocene) affords a presumption of their presence in the nearly allied early Tertiary faunas of Europe, just as their presence in the recent faunas of Madagascar and South Africa and in the Miocene of South America affords a presumption of their presence in the nearly allied faunas which immediately preceded them. Equally, the presence of marsupials in the early Tertiary of Europe on one side of Holarctica and of North America on the other side raises a strong presumption of their presence in the intervening region of Asia from which no fossils are known. ‘They are not found in the later Ter- tiary of Europe and America, so that we should not expect to find them in the later Tertiary of Asia. On the contrary, the small fragment of evidence that we have as to the Tertiary fauna of Antarctica affords a slight presumption against the presence of mammals on that continent. Doctor Gadow’s statement that the Chiroptera did not reach America until the Pleistocene is another curious instance of the misuse of the fossil record, which no one familiar with the character of our Tertiary formations and the necessary limits of the fossil faunas would be likely to make; nor would anyone acquainted with the variety and specializa- tion of the New World genera be inclined to believe that it was all the MATTHEW, CLIMATE AND EVOLUTION 313 result of post-Pliocene immigration and differentiation. Most of the ereodonts, he informs us, “died out with the Eocene or rather they were modernized into the typical Carnivora in various parts of the world. Some, however, kept on to almost recent times as highly specialized creodonts, e. g. the sabre-toothed tigers: Nimravus in North American Oligocene; Machewrodus from Miocene to Pleistocene in Europe and Asia, whence in the Pleistocene it appeared as Smilodon in America. : 8 Tt is perhaps unnecessary to point out that the macherodonts were not creodonts but typical Carnivora of the family Felidae, and that their evolutionary series is fully as complete and progressive in the Nearctie as in the Palearctic record. I may also note that “small swine” (meaning I suppose the primitive bunodont artiodactyles from which both pigs and peccaries are derived) appeared in North America quite as early as in Europe; that the genera Procamelus and Pliauchenia do not mark the splitting of the Camelide into camels proper and llamas; that Dorcatherium is not identical with “Hyomoschus”’ (Hyemoschus) and is an older name; that Arsinoitheriwm is not a pair-horned dicera- there but is a representative of a distinct order of mammals; that the precise relations of the American Eocene tapirs have yet to be deter- mined; that Protapirus does not first appear in the Lower Oligocene of Europe but in the Mid-Oligocene of Europe and North America; that there is no reason to believe that the European Paratapirus is more di- rectly in line of descent of the later tapirs than is the so-called Tapiravus of the American Miocene, and that the very fragmentary and inadequately studied record of the evolution of the Tapiride is quite inadequate for the positive and exact statements which Gadow makes as to their “wan- derings.” The statements as to the evolution of the horse show a surprising amount of inaccuracy, considering that this is so widely known a story. Apparently, it is in part the result of an attempt to criticize and modify the conclusions of American writers on the basis of a hasty survey of the incomplete materials available in European museums. The Eocene ancestors are disregarded, because they “are still so very generalized that they lead to horses, rhinos and tapirs as well as to other distinct groups.” While this is not far from the fact as regards the Lower Eocene Hohippus, it certainly is not true of Orohippus and Epihippus of the Middle and Upper Eocene. The relations of Miohippus to Mesohippus are hardly to be dismissed with a “perhaps.” Desmatippus is not an ancestor of Parahippus but is identical; Hypohippus is not intermediate between Para- and Merychippus but is an aberrant type descended from Miohippus 128 Op. cit. 314. ANNALS NEW YORK ACADEMY OF SCIENCES through Anchitherium; the American Miocene series does not come to an end with Merychippus, but this genus gives rise through numerous intermediate species to Protohippus, Phiohippus and Hipparion. Hip- pidion is not a descendant of Hipparion but of Pliohippus. There is, it is true, a considerable gap between Hipparion gracile and Hquus, this species being too specialized in tooth pattern and its lateral digits ex- ceptionally heavy; but most of the Ameritap hipparions are simpler and less aberrant in tooth pattern and the shafts of their lateral digits reduced often to mere threads. The proximal splints in these forms are very nearly as much reduced as they are in Hquus; the gap which Doctor Gadow declares has been “slurred over” lies simply in the fact that no specimens have yet been found in which the shafts of the lateral digits are discontinuous but the distal rudiments preserved. Anyone familiar with the difficulty of securing proof of this condition in a fossil species, and with the imperfection of our record of the Pliocene Equid, will hardly consider this as a serious gap. Certainly, it is trifling in com- parison with the gaps in any of the other mammalian phyla which Doctor Gadow accepts without difficulty. As for the derivation of Hquus from primitive species of Hipparion rather than of Protohippus, my opinion to that effect rests upon intensive studies of Miocene Equide undertaken for Professor Osborn’s monograph of the Evolution of the Horse (in preparation) and I do not think it fitting to publish the evidence in its support at present. The sirenians, Dr. Gadow tells us, afford strong support of the theory of a transatlantic bridge, the earliest being known from the Kocene of Jamaica and Egypt, ete. They would, undoubtedly, if there were suf- ficient reason to believe that they were absent from the more northerly parts of the North-Atlantic-Arctic shores during the early Tertiary. But there is none whatsoever; the North Atlantic coasts either extended dur- ing the Tertiary beyond their present limits to or towards the continental shelf, or else their marine and littoral deposits have been destroyed by glaciation; at all events none remain above water worth mentioning from New Jersey on one side around to the British Isles on the other. That no littoral vertebrates should be known where there are no littoral deposits is not surprising; yet it is upon this worthless negative evidence that the “strong support” rests. I have limited myself in the foregoing criticism to noting a few points in regard to fossil mammals. Dr. Scharff’s book is far too extensive for any detailed criticism here, even within these limits. I can note only that, while highly instructive as well as entertaining, it is far from being either accurate or fair in its treatment of the geological aspects of the MATTHEW, CLIMATE AND EVOLUTION 315 subject or the fossil record. The view-that during the Glacial Epoch the glaciers were confined in this country to the higher mountain ranges’”® is one that even a biologist is hardly excusable for upholding. Nor does it seem that anyone discussing the Tertiary geography of North America should be so little informed as to suppose*®® that the eastern and western portions of the continent were separated during the EHocene by an ocean barrier. In his argument against the permanency of the ocean basins, Scharff is, on the other hand, able to quote high authority. But the weakness of the argument is nevertheless apparent. That there have been great changes of level along certain lines of disturbance has never been questioned. But the conclusion that the continental platforms have never been submerged to abyssal depths, based upon the entire lack of abyssal deposits in their geological succession, is not disproved but rather confirmed by the recognition of abyssal deposits on an oceanic island lying along a line of high disturbance. For that merely proves that abyssal deposits are recognizable as such when they do occur, absence from the continental platforms remains untouched. Nor does the oc- currence of ancient sedimentary and metamorphic rocks on some, espe- cially of the larger, oceanic islands afford any evidence that they are remnants of former continents. The same processes of sedimentation, regional metamorphism and orogenic upheaval must of necessity occur in any oceanic island of considerable size and antiquity, and produce similar results both stratigraphic and petrographic. Moreover, if such islands lie in a line of disturbance which is continued under the ocean to an adjacent continent the same earth-movements may well affect both areas without raising the intervening region above the abyssal depths in which it now hes. Dr. Scharff adopts Ameghino’s correlations of Argentine formations, and Von Ihering’s assertion that the continent of South America did not exist as a single land mass until late in the Tertiary. I may note by the way that Ortmann'*! not long ago, in reviewing Pfeffer’s**? essay on the zodgeographical relations of South America, rebuked him severely for not being aware of this “undoubted fact,” which he declared was not a theory at all. The real facts are that marine and fresh-water forma- tions of Jurassic, Cretaceous and early Tertiary age occur extensively in the interior of South America, indicating that the broad low-lying in- terior of that continent was periodically flooded by shallow seas. The conditions parallel those of the North American continent very closely, 122 Op. cit., pp. 46 ff. 190 Tbid., p. 357. 131 A, KE. ORTMANN: Amer. Nat., vol. xxxix, pp. 413-416. 1905. 182 G, PFEFFER: Zoél. Jahrb., Suppl. 8, pp. 407-442. 1905. 316 ANNALS NEW YORK ACADEMY OF SCIENCES a so far as they are known. ‘The North American continent we know existed as such throughout geological time, although extensively flooded at times by shallow seas, especially during the Middle Cretaceous. The same is presumably true of South America. Like Doctor Gadow, Doctor Scharff makes a wholly unjustifiable use of negative evidence where it may serve to support his views. He is a much more reckless bridge-builder, and appears to be quite unconscious of any difference in probability between such a bridge as the Alaska- Siberia connection and the various trans-Atlantic and trans-Pacific bridges which he invokes. Yet the Alaskan bridge is in existence to-day, only a few yards of its planking removed, if one may so speak, the sub- structure intact, and the marks of the missing planks still showing on the undamaged portion, while the huge bridges which he “prefers” to believe in are, except for the Icelandic ridge, scarcely indicated by so much as a sandbank on the flat abyssal floor of the vast intervening oceans. That he can claim support of a kind from so high an authority as Suess may be true, but scientific problems should be settled by ex- amination of the evidence, not by citations of opinion from selected authorities. “et Doctor Scharff does not at all believe in accidental transportation by floating vegetation or other natural means. Why, he demands, do not the advocates of such views cite instances of such transportation in mod- ern times, and why is it only the more ancient animals that are so trans- ported? The argument is curiously parallel to the favorite anti-evolu- tionist demand. Why, if man has evolved from a monkey, do not the scientists take a monkey and turn him into aman? Of course, the proof demanded is an impossibility. If any instances of such transportation were noted during the last few centuries, they would be ascribed to human agency; but the probabilities within that time are slight except in islands near the coast, such as Krakataua; for more distant islands they are made probable only by the vast length of geological periods, and it is a matter of course that the more ancient the type, the longer time and consequently better chance there has been for its transportation by accidental agencies. Like all authors who advocate union of the Galapagos islands with the mainland, Dr. Scharff does not distinguish between a union of the islands with each other, which is geologically probable and is an almost unavoid- able conclusion from a study of the fauna, and their union with the mainland, which is highly improbable on geological and physiographical grounds, and is not merely unnecessary to explain the fauna but im- possible to reconcile with its peculiarities by any reasonable theories which take into account all of the consequences of such union. MATTHEW, CLIMATE AND EVOLUTION 319 While Doctor Scharff’s interpretation of the data is based upon funda- mentally different principles above noted, and his statements as to fossil distribution are often inaccurate or incomplete, yet the numerous dis- tributional data which he presents of modern invertebrates are of great interest, and, if interpreted along the lines which I have used, they fall completely into line with the vertebrate evidence. We cannot usually indeed check the conclusions drawn from modern distributional relation- ships by the fossil record. Many groups are altogether unknown, and the record in others is very scanty, but the same general relations clearly apply. The survival in Western Europe on one side, in southeastern North America on the other side, of a somewhat primitive cycle of Hol- arctic distribution ; the survival in the Mediterranean region on one side, in Central America and the Antilles on the other, of a more primitive cycle; of a still more primitive cycle in Africa and South America; and the progressively greater amount of divergent or parallel specialization in the survivors of the earlier cycles; the antique and fragmentary char- acter of the faunx of the oceanic islands, progressively more so in pro- portion to their smallness and isolation—all these conform to the verte- brate distribution. And with invertebrates as with vertebrates, every year adds to the number of the types which, while now limited to the peripheral continents and oceanic islands and highly discontinuous in their range, are shown to have inhabited formerly the central Holarctic region. It appears that many, one might perhaps say most, invertebrates are more readily transported across ocean barriers than vertebrates, espe- cially mammals, even making due allowance for their greater antiquity. This also we should expect. I do not think it necessary to catalogue the errors or inaccuracies in presenting the evidence afforded by fossil vertebrates. Such errors are unavoidable in a subject of so broad a scope, and excusable enough, if they do not lean too much to one side. I shall cite but one instance, and this in justice to my distinguished confrére Professor Depéret. Doctor Scharff concludes his summary of the North American records of the Evolution of the Horse with the following remarks :1%* “And yet not a single transition from one genus to the other seems to be known. No wonder that one of our foremost paleontologists exclaims, “The sup- posed pedigree of the horse is a deceitful delusion, which simply gives us the general process by which the tridactyl foot of an ungulate can be transformed in various groups into a monodactyl foot in view of an adaptation for speed, but this in no way enlightens us on the paleontolog- ical origin of the horse.’” Such a statement, coming from so excellent 183 Op. cit., p. 147. 318 ANNALS NEW YORK ACADEMY OF SCIENCES an authority, seems startling until one verifies the quotation and finds that it refers, not to the American records, but to the ancestry of the horse as presented in Gaudry’s'** Enchainements, to the European series Paleotherium, Anchitherium, Hipparion gracile and Equus. Depéret takes care to premise that he is speaking only of this European series, and while I think the criticism goes too far—it should at least be modi- fied by changing “ungulate” to “perissodactyl” in view of what we know about the Litopterna—yet the criticism is largely justified in its proper context. As applied to the American series it is altogether unwarranted. 1% A, GAUDRY: Enchainements du Monde Animal, vol. iii, Mammiferes tertiaires. 1878. PUBLICATIONS OF THE NEW YORK ACADEMY OF SCIENCES (Lyceum or Natura History, 1817-1876) The publications of the Academy consist of two series, viz. : (1) The Annals (octavo series), established in 1823, contain the sci- entific contributions and reports of researches, together with the records of meetings and similar matter. A volume of the Annals coincides in general with the calendar year and is sold at the uniform price of three dollars per volume. The articles composing the volume are printed separately, each in its own cover, and _are distributed in bundles on an average of three per year. The price of the separate articles depends upon their length and the number of illus- trations, and may be learned upon application to the Librarian of the Academy. The author receives his separates as soon as his paper has been printed, the date of issue appearing above the title of each paper. (2) The Memoirs (quarto series), established in 1895, are issued at irregular intervals. It is intended that each volume shall be devoted to monographs relating to some particular department of Science. Volume I is devoted to Astronomical Memoirs, Volume If to Zodlogical Memoirs, ete. The price is one dollar per part as issued. All publications are sent free to Fellows and Active Members. ‘The Annals are sent to Honorary and Corresponding Members desiring them. Subscriptions and inquiries concerning current and back numbers of any of the publications of the Academy should be addressed to THe LIBRARIAN, New York Academy of Sciences, i ; | care of . American Museum of Natural History, New York, N. ¥, PRESS OF JUDD & DETWEILER, INC., WASHINGTON, D. C. ~ ANNALS OF THE NEW YORK ACADEMY OF SCIENCES Vol. XXIV, pp. 319-346, pll. XXII-XLI Editor, EpmMuND OTIs Hovey DEVELOPMENT OF THE NEURAXIS IN THE DOMESTIE:-CAT. [QO THE STAGE OF TWENTY-ONE SOMITES BY H. von W. ScHULTE AND FREDERICK TILNEY NEW YORK PUBLISHED BY THE ACADEMY 31 Marcu, 1915 - acai ape THE NEW YORK ACADEMY OF SCIENCES (Lyozum or Natural Hisrory, 1817-1876) OFFIcErs, 1914 President—GrorGE FREDERICK Kunz, 601 West 110th Street Vice-Presidents—CuarLrEs P. Berkey, RayMonp C. Ospurn, CHARLES BASKERVILLE, CLARK WISSLER Corresponding Secretary—Hrnry HK. Crampron, American Museum Recording Secretary—EpmuND Otis Hovey, American Museum Treasurer—HeEnryY L. Douerty, 60 Wall Street Inbrarian—Ratru W. Tower, American Museum | Editor—Epmunp Oris Hovey, American Museum SECTION OF GEOLOGY AND MINERALOGY Chairman—Cuar.eEs P. Berkey, Columbia University Secretary—A. B. Pacint, 147 Varick Street , SECTION OF BIOLOGY Chairman—RaymonpD C. Ospurn, 557 West 124th Street _ Secretary—Witiiam K. Grecory, American Museum SECTION OF ASTRONOMY, PHYSICS AND CHEMISTRY _Chairman—CuaRLES BASKERVILLE, College of the City of New York Secretary—ErneEst E. Smitu, 50 East 41st Street SECTION OF ANTHROPOLOGY AND PSYCHOLOGY Chairman—Cuark WIsSLER, American Museum _ Secretary—Rosert H. Lowiz, American Museum y _. The sessions of the Academy are held on Monday evenings at 8:15 o'clock from October to May, inclusive, at the American Museum of © Natural History, 77th Street and Central Park, West. fAnnats N. Y. Acap. Scr., Vol. XXIV, pp. 319-346, Pll. XXIJI-XLI. 31 March, 1915] DEVELOPMENT OF THE NEURAXIS IN THE DOMESTIC CAT TO THE STAGE OF TWENTY-ONE SOMITES* By H. von W. ScHULTE AND FREDERICK TILNEY CONTENTS Page VR AROYETO AKON Hida equi oe o Go oroo Ola nro Gin OO bine Groidio DIG sia aia tke ici exInO Coun OIC O 319 MER CMGLON TOL CTO OSs re cise eae eee oe icine Hee e aloie wale ine aiiehel ores tage taged sae tries Embryos prior to the appearance of intersomitic clefts. .....-..-+-+-- 321 HAD RVGSTOL ONE ISONMEE s onielelel elec uiceie 6.8 alaictisis ese nia oleate eleiern imines eee Pena pees Ei EV OSM ORME O ISON ILES raeielete aro Sse tira: = iesvearo ni ale afte facm is sayetbe torr in euro ever shee es PO PEVOROE TENT CET SOMMCCS reraie acs cite 4 sob tate eats oe cise aisle veensiaelereelele mn Vooe HMI OMO te LOM A SOMME ya -terats eve cle seya ei aici alelist el ct cvelole ois) sino lai-ieialallstekos ken: 322 PMIPGLVGSTOL SEVETIE SOUMEES a Ae er ae c.« heriercace eae er eone io atid. c Soantatoder hs shah cn eomeen DTT OETOS Gwe Ethane Sonmnisococcoagoo mes odusodhsoboogmandccaMocgooS Gre BI DAVO. Ol: ahs Conse Jane donor ooeodubon somo roMbociwie mono ar comcocs 324 PRIN EYVOSEOE sce SOMILCS Hsia sir leo cue sions sie b oitisrte ee mee cleo sO HWMbEVOsOLitwelverand thirteenusomitesieerisc se elecieictreites sisi erar mes EMPL VOSwO fe LOURLECIE SOMMEES iu cts eters ails siateicssvare e siaieress wheres vereielei never 327 EMD EVOROLSESFCENe SOMES rare ec conerce sin ever lo rohaceravacteca suv simysieleleleie «tiene 329 HimbEyOLOrnsevenbeenusOmitesincctvasese loot ish oon cata o sustske cae ee Oe HM MpEyorot nineteen SOMILEES <2) ores) sve. ouaie ce chore « ve iclere eia@ lls) ee rote le] Susie ere eueoo HIMpPEVOLOLAGWeNty-ONEsSOMIULCS yan even ae. s el cctst a ms eracinioe see nee csi oo RAN CTT OMI CECLES tere cis rections iavnretehcl eval tia’s feveusindare @ wickeiontalelts bre Showin tise bwere SORA EOE: CHOSTIES. IE THIS INUIT NO La Berend come co OR OiereIdioinae arn aa Sale eae 336 EFOSCHCE al ON ere eee PNetere eevee ie es aie ovens bi alana eta pele iStelouhis lo Se by oe eibo ate te ee OS MESON CED HALO er micyatavercrote: ote caketenetescvalorehonarete tobananerctanera ace Oiles slaee Seid seats igh ale ete 342 RUHOM Pence pHalomerrre,cvereiwrevozs tele ve coher aio erates ic eve Sie eer ae ome escetecueiearelorererers 344 INTRODUCTION The present paper is a study of the morphogenesis of the neuraxis in its early stages, with especial reference to an attempted interpretation of the forebrain in terms of the longitudinal zones of the neural tube, viz., the basal and alar laminew, and the ganglionic crest. The theoretical problem can be stated briefly. Since the ganglionic elements are in- cluded in the wall of the neural tube at the time of its separation from the ectoderm—Neumayer has shown this for reptiles, and it is also true 1 Presented in abstract by the senior author at the meeting of 11 May, 1914, under the title “Early Stages in the Development of the Brain in the Domestic Cat.” Manuscript received by the Editor, 17 September, 1914. : (319) 4 7 At > 7 = Doe. 4 Sr eal _ ist tie ANNALS N. Y. Acap. Scr. VoLUME XXIV, PLatE XXVII ITIAXX GLVYId AIXNX GINATIOA ‘IOS “VOW ‘X(N SIVNNY XIXX GLvV1d ‘AIXX OFNAIOA ‘IOS ‘GVOV ‘AX ‘N SIVNNY XXX GIVId “AIXX FWNATIOA IOS ‘dVOY ‘A ‘N SIVNNY PLATE XXXI VoLUME XXTV, ANNALS N. Y. AcaD. Scr. ee cae ae lines al ANNALS N. Y. Acab. Scr. VOLUME XXIV, PLATE XXXII Fie. 2 ANNALS N. Y. AcabD. SCI. VoLUME XXIV, PLATE XXXIII AIXXX GLVId ‘AIXX FNATION ‘10S “dVOV ‘XK ‘N SIVNNY AXXX WLV1d ‘AIXX INAI0A : ‘10g “VOW “X 'N STVANYV m IIAXXX GilVId ‘AIXX @NOAIOA ‘IOS “GVOV “XA ‘'N SIVNNYW IIIAXXX @LYIgd ‘AIXX GDNAI1I0OA ‘IOS ‘dVOV “A ‘'N SIVNNY G Dla T ‘Oli XIXXX FLlVId ‘AIXX @FNAIOA ‘IOg ‘dVOV ‘A ‘'N SIVNNY Tx QLVId ‘AIXX GNOIOA I! ‘DL ‘1OS ‘avoy ‘XK CN SIVNNY ITX GLI “AIXX GNATIOA 109 “GVOY ‘X 'N SIVNNY PUBLICATIONS OF THE NEW YORK ACADEMY OF SCIENCES (Lyceum or Natura History, 1817-1876) The publications of the Academy consist of two series, viz. : (1) The Annals (octavo series), established in 1823, contain the sci- entific contributions and reports of researches, together with the records of meetings and similar matter. A volume of the Annals coincides in general with the calendar year and is sold at the uniform price of three dollars per volume. The articles composing the volume are printed separately, each in its own cover, and are distributed in bundles on an average of three per year. The price of the separate articles depends upon their length and the number of illus- trations, and may be learned upon application to the Librarian of the Academy. The author receives his separates as soon as his paper has been printed, the date of issue appearing above the title of each paper. (2) The Memoirs (quarto series), established in 1895, are issued at irregular intervals. It is-intended that each volume shall be devoted to monographs relating to some particular department of Science. Volume I is devoted to Astronomical Memoirs, Volume II to Zodlogical Memoirs, ete. The price is one dollar per part as issued. All publications are sent free to Fellows and Active Members. The Annals are sent to Honorary and Corresponding Members desiring them. Subscriptions and inquiries concerning current and back numbers of any of the publications of the Academy should be addressed to Tue LIBRARIAN, New York Academy of Sciences, care of American Museum of Natural History, New York, N. Y. PRESS OF JUDD & DETWEILER, INC., WASHINGTON, D. 0. nA : ae at eis & ANNALS OF THE NEW YORK ACADEMY OF SCIENCES Vol. XXIV, pp. 347-443 Editor, EpmMunp Otis Hovey RECORDS OF MEETINGS CHARTER, CONSTITUTION AND MEMBER- SHIP IN 1914 OF THE NEW YORK ACADEMY OF SCIENCES WITH INDEX TO VOLUME XXIV NEW YORK PUBLISHED BY THR ACADEMY 14 May, 1915 . THE NEW YORK ACADEMY OF SCIENCES (Lyceum or Natura Histrory, 1817-1876) OFFICERS, 1914 President—Grorch Frepertck Kunz, 601 West 110th Street Vice-Presidents—CHARLES P. Berkey, RayMonpD C. OsBuRN, CHARLES BASKERVILLE, CLARK WISSLER Corresponding Secretary—HENRY KE. CRAMPTON, American Museum Recording Secretary—EpmuND Otis Hovey, American Museum Treasurer—HENryY L. Douerty, 60 Wall Street Inbrarian—Rawtru W. Tower, American Museum Editor—EvmunpD Otis Hovey, American Museum SECTION OF GEOLOGY AND. MINERALOGY Chairman—Cuar.eEs P. BerKxey, Columbia University Secretary—A. B. Pacin1, 147 Varick Street SECTION OF BIOLOGY Chairman—Raymonp C. Ossurn, 557 West 124th Street Secretary—Wituiam K. Grecory, American Museum SECTION OF ASTRONOMY, PHYSICS AND CHEMISTRY Chairman—CHaARLES BASKERVILLE, College of the City of New York Secretary—ErNEsT E. SmitH, 50 East 41st Street SECTION OF ANTHROPOLOGY AND PSYCHOLOGY _Chairman—C.ark WISSLER, American Museum Secretary—Roxsert H. Lowir, American Museum The sessions of the Academy are held on Monday evenings at 8:15 o'clock from October to May, inclusive, at the American Museum of Natural History, 77th Street and Central Park, West. . be pw ° i la ns son [ANNALS N. Y. Acap. Sci., Vol. XXIV, pp. 347-448. 14 May, 1915] RECORDS OF MEETINGS OF THE NEW YORK ACADEMY OF SCIENCES January to December, 1914. : LIBRAR Y By Epmunp Otis Hovey, Recording Secretary NEW Yor« BOTANICA: GARE? BUSINESS MEETING 5. JANUARY, 1914 The Academy met at 8:22 P. M. at the American Museum of Natural History, President George F. Kunz presiding. The minutes of the last business meeting were read and approved. The following candidates for membership in the Academy, recom- mended by Council, were duly elected: ACTIVE MEMBERSHIP F. H. Pike, Columbia University, R. G. Eccles, 681 Tenth Street, Brooklyn, Hermann von W. Schulte, College of Physicians and Surgeons. The Recording Secretary then reported from the Council the recom- mendation that Mr. Emerson McMillin be elected a patron in recognition of his direct gifts of more than $1,000 to the active work of the Academy. On motion, the Academy unanimously adopted the recommendation, and Mr. McMillin was declared a patron of the Academy. The Academy then adjourned. Epmunp Otts Hovey, Recording Secretary. SECTION OF GEOLOGY AND MINERALOGY 5 JANUARY, 1914 Section was called to order at 8:15 Pp. M., Vice-President Charles P. Berkey presiding and forty-three members and guests being present. (347 ) 348 ANNALS NEW YORK ACADEMY OF SCIENCES The minutes of the previous meeting of the Section were read and approved. The following scientific programme was then offered: Miss Marjorie O’Connell, A Revision or tHE GENUS Zaphrentis. Chester A. Reeds, THE OGLITES OF THE CHIMNEYHILL ForMaA- TION OF OKLAHOMA. Charles P. Berkey, THE Origin or SOME OF THE COMPLEX STRUCTURES OF THE ANCIENT GNEISSES or New York. SUMMARY OF PAPERS Miss O’Connell’s paper gave an account of a proposed reclassification of the genus Zaphrentis which will eliminate the inconsistencies at pres- ent extant. This paper has been published as pages 177-192 of Volume XXIII of the “Annals.” Dr. Reeds’s paper was illustrated with lantern slides showing thin sections of odlites, in which he pointed out typical structures. Hand specimens were also shown. This paper was discussed by Professor Grabau, Dr. Berckhemmer and Dr. Berkey. Dr. Berkey presented a continuation of his paper on this subject begun late last year, and this section of it was illustrated with lantern slides. Discussion was postponed until the next meeting owing to the lack of time. The Section then adjourned. A. B. PAcINI, Secretary. SECTION OF BIOLOGY 12 JANUARY, 1914 Under the auspices of the Section of Biology, a general meeting of the Academy and its Affiliated Societies was held in the main lecture hall at the American Museum of Natural History at 8:15 Pp. M. President Kunz introduced the Chairman of the Section of Biology, Professor Raymond C. Osburn, who presided. The reading of the minutes of the last meeting was dispensed with and the following programme was then offered : RECORDS OF MEETINGS 249 CONFERENCE ON THE PILTDOWN SKULL AND THE ORIGIN OF MAN Henry Fairfield Osborn, GrotocicaL AGE AND SuccESSION OF FARLY Human TYPES. J. Leon Williams, ON THE PILTDOWN AND OTHER PREHISTORIC SKULLS. R. Broom, CRITIQUE OF KEITH’s AND SMITH WoopD- WARD’s RESTORATIONS OF THE PILTDOWN SKULL. William K. Gregory, Tre Base oF THE CRANIUM IN ANTHROPOIDS AND MAN. SUMMARY OF PAPERS The substance of Professor Osborn’s paper will appear in Volume XXVI of the “Annals.” Dr. Williams gave a careful statement of the essential facts regarding the discovery of the Piltdown remains and the principal points of the reconstructions attempted by Drs. Keith and Smith Woodward. Professor Broom defended Smith Woodward’s reconstruction, which he held to be far better than Professor Keith’s. Dr. Gregory spoke in substance as follows: Some years ago a work by the Dutch anatomist Van Kampen directed my attention to the importance of the detailed characters of the base of the brain-case as indicating the relationships of various groups of mam- mals. The special characteristics of the bony portions of the organs of hearing are highly significant, in revealing descent from common an- cestors among widely different animals. I therefore propose to pass rapidly in review before you the basal view of the skull in many families of Primates and to point out the significance of the resemblances and differences in the auditory region. Lemuride. In this family the auditory bulla or bony resonating cham- ber of the middle ear is swollen up in a more or less hemispherical or ovoid form. It completely incloses and hides from view the delicate ring of bone upon which the tympanic membrane is stretched and which is known as the tympanic annulus or tympanic bone. ‘The existing lemurs have evolved into widely diverse forms: here we have a more or less in- -sectivorous form, and here a large sloth-like, leaf-eating form, and yet the formation of the auditory region is essentially identical in all. This formation is one of the characteristics which these now very diverse lemurs have probably inherited in common from remote and extinct ancestors, such as have been found in the Eocene formations of Wyoming. 250 ANNALS NEW YORK ACADEMY OF SCIENCES ; Indriside. The Indriside include certain highly specialized lemurs from Madagascar, such as the Sifakas and Indris. As compared with the lower lemurs great advances have been made in the structure of the teeth and in the size of the brain-case, but the formation of the auditory bulla remains the same and this is one of the characters which reveals relationship with the typical lemurs. The Nycticebide include certain curiously modified lemuroids of Africa and the oriental region, such as the slow loris and the galagos. These lemuroids have undergone considerable modification in the basi- cranial region. It is much wider, the mastoid region in the back of the skull is swollen up, the tympanic annulus is no longer concealed by the bulla but lies as a short rim at the external border of the bulla. The South American monkeys of the family Cebidee differ markedly from the lemurs as follows: the tympanic annulus is not concealed by the auditory bulla but lies external to it and is closely joined with it, the suture disappearing in the young animals. ‘The base of the brain-case has been greatly widened and the bulla itself is further in toward the mid-line than it was in the lemurs. ‘The tympanic thus forms a short napkin-ring-like spout, called the bony auditory meatus. The monkeys of the Old World or catarrhine Primates. In these the auditory bulla is not so much expanded as in preceding families. It is pierced by a large canal or foramen for the carotid artery. The tym- panic bone now form a greatly elongated spout leading to the outer ear. In this slide we see the wide range in structure among the existing macaques and baboons. Beginning with a short-faced more round-headed macaque we pass by almost imperceptible gradations to these very highly specialized baboons with enormous elongated faces and massive jaws and teeth. Amid all this diversity in form the structure in the auditory region remains constant, as we see by comparing the most specialized form, the mandrill, with the primitive form figured at: the left. We pass now to a much higher group of the Old World apes, the Sem- nopithecine, which includes the langurs, the guerezas, the long-nosed monkey and others. Were it not for their high vegetarian specializa- tions these monkeys, so far as the skull is concerned, might almost be regarded as ancestral to man. Again we have a wide variation in form from short-faced to long-faced types, but the formation and arrangement in the auditory region is the same as in all other Old World apes, namely, the bulla is not greatly inflated and is pierced by the carotid foramen or canal while the tympanic forms a long tapering spout. The anthropoid apes all agree again in the structure of the auditory region. Here is the auditory bulla, pierced by the carotid canal, and RECORDS OF MEETINGS ool here is the long spout-like tympanic. Passing on the right to the human skull, notwithstanding its marked differences in the proportion of vari- ous parts, we see a fundamental agreement with the anthropoids and with all the Old World monkeys in the auditory region. Here again is the auditory bulla pierced by the carotid canal, now greatly enlarged, and here again is the long spout-like tympanic somewhat altered in ex- ternal contour. If this agreement stood by itself it might be ascribed to convergent evolution, but taken in connection with hundreds, even thousands, of other agreements it can only mean common ancestry with the anthro- poids and the Old World monkeys. The Piltdown skull fortunately preserves a portion of the auditory region together with the articulation for the lower jaw. It conforms, as do all other human skulls, to the type common to the Old World mon- keys, the anthropoids and man. In the form of the articulation for the lower jaw it is very man-like, but in the form of the lower jaw itself it is more orang-like. It is thus a synthetic type combining in a way not hitherto known the characters of man and of apes. Conclusion. If any of you may think that I have overestimated the significance of this fundamental agreement in the structure of the audi- tory region in man and the Old World monkeys and anthropoid apes, I ean only reply that the study of evolutionary relationships is a science in itself, that by long experience the facts force themselves upon us and compel us to place certain values upon them. The conclusion that mankind is related by common origin with the Old World monkeys and anthropoids is irresistibly forced upon paleon- tologists, who are familiar with scores of other well-established evolu- tionary series. But in view of the incredulous attitude of many it is the duty of those who are familiar with the facts to place them before the public. | The conference was followed by a collation which was served in the Eskimo Hall. The Section then adjourned. WiLiiaAmM K. Grecory, Secretary. SECTION OF ANTHROPOLOGY AND PSYCHOLOGY 26 JANUARY, 1914 Section met in conjunction with the American Ethnological Society at 8:15 p. M., Professor Franz Boas presiding. Or raw) ANNALS NEW YORK ACADEMY OF SCIENCES The meeting was devoted to a public lecture, as follows: Fay Cooper Cole, Hr Witp Tribes or MINDANAO. SUMMARY OF PAPER Mr. Cole first described the Island of Mindanao and its history, then discussed in more detail the life of two of the pagan tribes—the Bukid- non and the Bagobo. The Bukidnon, who inhabit the north-central portion of the island, have for centuries been harassed by the wild Manobo warriors on the east and by the slave-hunting Moro on the west. The many conflicts with these enemies caused them to develop a unique culture, one phase of which is shown in the tree dwellings found in part of their territory. The presence of three well marked physical types in the population is another point of interest brought out by this paper. The natives’ views concerning the spirit world and some of the ceremonies made to pro- pitiate the superior beings were described and illustrated in the talk. Going to the Bagobo, on Dayao Gulf, a glimpse was given into their traditions, laws and customs, particularly those which led up to and ex- plained the custom of human sacrifice, and the organization known as Magani—the members of which gain the right to the title and a dis- tinetive type of dress by slaying a certain number of enemies. The lecture was illustrated by about seventy shdes showing the coun- try, the people and their homes and several native crafts. The Section then adjourned. R. H. Lowi, Secretary. BUSINESS MEETING 2 FEBRUARY, 1914. The Academy met at 8:17 py. M. at the American Museum of Natural History, President George F. Kunz presiding. The minutes of the last business meeting were read and approved. The following candidates for membership in the Academy, recom- mended by Council, were duly elected : AcTIVE MEMBERSHIP W. J. Matheson, 182 Front Street, A. A. Goldenweiser, Columbia University, Marguerite T. Lee, 66 West 95th Street. RECORDS OF MEETINGS Or Oo ASSOCIATE MEMBERSHIP Miss Ruth Raeder, Barnard College. The Recording Secretary then reported the following deaths: Henry W. Boettger, Active Member since 1905, died 13 January, 1914, Dwight A. Jones, Active Member since 1905, died 7 December, 1913, David L. Pettigrew, Active Member since 1896, died 19 January, 1914, | George 'l'aylor, Active Member since 1907, died 20 September, 1913. The Academy then adjourned. Epmunp Otis Hovey, Recording Secretary. SECTION OF GEOLOGY AND MINERALOGY 2 Fesruary, 1914 Section was called to order at 8:15 p. M., Vice-President Charles P. Berkey presiding. The minutes of the last meeting of the Section were read and approved. The following scientific programme was then offered : Henryk Arctowski, A Srupy oF THE CHANGES IN THE DISTRIBUTION oF TEMPERATURE IN EUROPE AND NortH AMER- 1cA DurInG THE YEARS 1900-1909. Charles P. Berkey, OrtciIn oF SoME CoMPLEX STRUCTURES OF THE ANCIENT GNEISS OF NEw York: IGNEouS Con- TACTS AND TRANSITIONS. SUMMARY OF PAPERS Dr. Arctowski’s paper has been published as pages 59-113 of Volume XXIV of the “Annals.” Dr. Berkey: The following items cover the chief modifications and re- statements discussed : The Sedimentary Series: 1). The fundamental formations, including the Fordham gneiss, were originally a simple sedimentary series. 2). There is nowhere in the region any evidence of an older basement. 3). The relation between the Manhattan schist and the Inwood limestone is essentially a conformable transition showing considerable oscil- lation and local variation. 304 ANNALS NEW YORK ACADEMY OF SCIENCES 4). There is also either a conformable or a simple overlap relation be- tween the Inwood and Fordham formations with some interbed- ding of gneiss within the limestone. 5). The streaked character of the Manhattan schist is due chiefly to igneous impregnation of an already strongly foliated metamorphic rock. Its other characters are of primary sources. 6). The chief impurities of the Inwood, exclusive of the pegmatite dikes, are recrystallized primary matters. 7). The strongly banded structure of the Fordham is of complex origin— its fundamental cause is primary sedimentary difference and strong bedding structure. The rock has been injected along these weak- ness lines with igneous matters. 8). Distortions are for the most part of regional dynamic origin, but in some instances, in the larger intrusives, it is in part of flowage origin. 9). The prevalent granitic composition of the Fordham is in part of pri- mary (arkose) origin and in part simple injection and in part an impregnation of granitizing solutions. 10). There are variations in the Fordham indicative of original inter- bedded limestones similar to the Inwood, and shales similar to the Manhattan, besides the more abundant sandstones and arkoses. 11). There is no direct evidence as to the exact geologic age and no per- fectly satisfactory correlation. The Igneous Series: 12). The large intrusive masses are represented by: (a) “Yonkers gneiss.” (0) “Ravenswood granodiorite.” (c) “Staten Island serpentine.” 18). The smaller igneous representatives include: (ad) “Pegmatite dikes.” (e) “Anthophyllite rock.” (f) “Hornblende schist strips.” (g) Lenticular and irregular masses and streaks of pegma- titic matter in the schists. (h) Some of the bands of the Fordham gneiss. (i) Much impregnating granitic matter now intimately mixed with materials of other origin. , 14). The principal structure of the Yonkers and the Ravenswood is essen- tially primary and of two types: (1) A superimposed structure derived from and in part pre- serving the structure of partially absorbed older masses. (2) An induced structure due to movement in the magmatic mass during crystallization. 15). The structure of the hornblende schist, “Anthophyllite rock’ and serpentine is chiefly secondary (metamorphic). 16). The igneous representatives vary in age. RECORDS OF MEETINGS Dy | | 17). There is very great difference in extent of igneous effect in the differ- ent fundamental formations. I judge that there is a strongly selective influence exerted by the formations themselves. General: 18). As a result, all of the formations are complex in composition—in part primary, in part metamorphic, in part introduced, and 19). All of the formations are also complex in structure—in part of pri- mary sedimentary control, in part induced by metamorphism (re- erystallization), in part of primary igneous habit, in part a pri- mary structure emphasized by its control over igneous injection, and in part purely secondary dynamic modification. The Section then adjourned. AY Bs PACING; Secretary. SECTION OF BIOLOGY 9 FrBruary, 1914 Section met at 8:15 p. m., Professor Raymond C. Osburn presiding. The minutes of the last meeting of the Section were read and approved. The following programme was then offered: W.D. Matthew, SoME REMARKABLE Extinct ANIMALS OF SoutH AMERICA. Robert Cushman Murphy, Hasirs, ANAroMy AND RELATIONSHIPS OF THE SEA KueeHant (Mirounga leonina). SUMMARY OF PAPERS Dr. Matthew: The American Museum collections of extinct South American mammals includes a series of eight mounted skeletons repre- senting the Edentates, hoofed animals and Carnivores that flourished during the Pampzean and Santa Cruzian epochs. The chief characteris- tics of these animals were outlined. Mr. Murphy had studied the Sea Elephants at South Georgia, a small island in the Antarctic Ocean, where he had secured a series of specimens of them and of other animals for the American Museum of Natural His- tory and for the Brooklyn Institute of Arts and Sciences. He exhibited a remarkable series of photographs of hving Sea Elephants; also a series of skulls representing the principal genera of the Phocide, arranged ac- cording to their structural affinities, the extremely long-skulled Sea Leopards being at the left, Phoca, of intermediate structure, near the center, and Monachus and Mirounga, with widened skulls, at the right. 306 ANNALS NEW YORK ACADEMY OF SCIENCES After discussion of the papers the Section adjourned. WILLIAM K. Grecory, Secretary. SECTION OF ASTRONOMY, PHYSICS AND CHEMISTRY 16 Fepruary, 1914 Section met at 8:15 p. m., Vice-President Charles Baskerville pre- siding. The evening was devoted to the following lecture: H. T. Barnes, THe Puystcan Errects Propucep By ICEBERGS IN THE NortH ATLANTIC. SUMMARY OF PAPER Professor Barnes said in abstract: The formation, drift and melting of icebergs form for the physicist subjects of inquiry of great practical value. The whole matter has been, until recently, neglected, but popular opinion now demands careful investigation of it. Although the need for such inquiry is no more acute than it has always been, the lecture now offered to the Academy deals with our present knowledge and points out the great value which would result from a careful investigation of the Labrador current, for which a large appropriation by the various govern- ments concerned would be necessary. The lecture was followed by a collation which was served in the Eskimo Hall. mMuND OTIs Hovey. ° Councilors (to serve 3 years), BAsHrorD Dean and C. STuartT GAGER. Finance Committee, Freperic S. Ler, Joun Tartock and W. J. MATHESON. At the close of the elections, Dr. George F. Kunz gave his address as retiring President, entitled RECORDS OF MEETINGS 401 “THE New York ACADEMY OF SCIENCES, ITS PAST, ITS PRESENT AND ms HULTUREs after which the Academy and guests listened to an address by Mr. Raymond L. Ditmars, Curator of Reptiles at the New York Zoological Park, on the “Book OF NATURE,” illustrated with some of his latest motion-picture reels showing various phases of most interesting animal life. The Academy then adjourned. Epmunp Otis Hovey, Recording Secretary. REPORT OF THE CORRESPONDING SECRETARY We have lost by death during the past year the following Honorary Members : Sir John Murray, elected 1912, died 16 March, 1914, Heinrich Rosenbusch, elected 1887, died 20 January, 1914, August Weissmann, elected 1909, died 6 November, 1914, and the following Corresponding Members : Theodore Nicholas Gill, elected 1858, died 25 September, 1914, Seth E. Meek, elected 1888, died 7 July, 1914, Charles Sedgwick Minot, elected 1878, died 19 November, 1914, J. Selden Spencer, elected 1890, died 3 December, 1914, N. H. Winchell, elected 1898, died 1 May, 1914. Two Corresponding Members have been elected Honorary Members. There are at present upon our rolls 47 Honorary Members and 115 Corresponding Members. Respectfully submitted, Henry EH. CRAMPTON, Corresponding Secretary. REPORT OF THE RECORDING SECRETARY During the year 1914, the Academy held 9 business meetings and 27 sectional meetings, at which 72 stated papers were presented as follows: Section of Geology and Mineralogy, 18 papers; Section of Biology. 22 papers; Section of Astronomy, Physics and Chemistry, 7 papers; Section of Anthropology and Psychology, 25 papers. Four of the sectional meetings were of general character and of par- 402 ANNALS NEW YORK ACADEMY OF SCIENCES ticular interest and were followed by a social hour, with refreshments, in one of the exhibition halls of the Museum. The first was held under the auspices of the Section of Biology on the evening of 12 January, when a “Conference on the Piltdown Skull and the Origin of Man” was participated in by Professor Henry Fairfield Osborn, Dr. J. Leon Williams, Professor R. Broom and Dr. W. K. Greg- ory. The second was held on 16 February, under the auspices of the Section of Astronomy, Physics and Chemistry, when Professor H. T. Barnes, of McGill University, lectured upon “The Physical Effects Pro- duced by Icebergs in the North Atlantic.” The third meeting was held on the evening of 23 March, under the auspices of the Section of An- thropology and Psychology; Professor Hiram Bingham, of Yale Uni- versity, delivered a lecture on “Recent Exploration in the Land of the Ineas.” The fourth, held on 2 November, was under the auspices of the Section of Geology and Mineralogy, and Professor Reginald R. Daly lectured upon “Problems of Volcanic Action.” In addition to these general meetings of the Academy, one public lec- ture was given to the members of the Academy and the Affiliated Socie- ties and their friends on 30 April by Dr. L. A. Bauer, Director of the Department of Terrestrial Magnetism at Washington. The’ title was “Following the Compass.” At the present time the membership of the Academy is 497, which includes 478 Active Members (of whom 19 are Associate Members, 124 Fellows, 98 Life Members and 10 Patrons) and 19 Non-resident Mem- bers. There have been 11 deaths during the year, 22 resignations have become effective and three names have been dropped from the roll. One member has been discontinued temporarily at her own request. Fifty- four new members have been elected during the year, one of whom failed to qualify and five of whom commuted their annual dues by a single pay- ment of $100 each. One patron has been elected. ‘Two names have been transferred to the life membership list on account of twenty-five years’ payment of annual dues. Four associate members have taken up active membership. As the membership of the Academy a year ago was 481, there has been a net gain of 16 during the year 1914. Record is made with regret of the loss by death of the following active and associate members : Albert S. Bickmore, Active Member since 1873. Samuel H. Bishop, Active Member since 1907. Henry W. Boettger, Active Member since 1905. 1 Including 38 members-elect who have not yet paid their first annual dues, RECORDS OF MEETINGS 403 F. F. Hahn, Associate Member since 1912. Karl Hutter, Active Member since 1910. Dwight A. Jones, Active Member since 1905. J. Langeloth, Active Member since 1905. Mrs. Charles Tyler Olmsted, Active Member since 1907. Charles J. Perry, Active Member since 1905. David L. Pettigrew, Active Member since 1896. George Taylor, Active Member since 1907. Respectfully submitted, EpmMunpD Otts Hovey, Recording Secretary. REPORT OF THE LIBRARIAN During the current year the Library of the New York Academy of Sciences has received by exchange and donation two hundred eight vol- umes and one thousand five hundred twenty-two numbers. The Natur- historisches Verein in Augsburg has very graciously supplied the Acad- emy Library with volumes III (1850), V-XIII (1852-1860), which ‘ were lacking in the files and for which special acknowledgments are herewith extended. Through the system of inter-library loans the scientific books have been made more accessible to students and investigators, and it is there- fore a pleasure to report that the use of the library has much increased. Respectfully submitted, RatpeH W. Tower, Inbrarian. REPORT OF THE EDITOR The parts of the Annals which have been published this vear are the following : ; VOLUME XXIII Pages A. C. Hawkins—Lockatong Formation of the Triassic of New Jersey ATAU SVs crevevcierayelte eel er ayelsle a) cheyeiel oisi cnebeheieleliels 145-176 Marjorie O’Connell—Revision of the Genus Zaphrentis.............. 177-192 Charles R. Fettke—The Manhattan Schist of Southeastern New York State and Its Associated Igneous Rocks....... 193-260 ¥. O. Hovey—Records of Meetings of the Academy................. 261-316 Charter and Organizaticn of the Academy............ 317-322 Constitutions ands By-laws)... 226s siseie © cieieeicie einie awre ele « OLBSOOO Membership, of the Academy..................+-+02... odol—-342 GE Ee Ss ad's oa waster tet sy a. SRB ie at 343-353 AQ4 ANNALS NEW YORK ACADEMY OF SCIENCES VOLUME XXIV Elvira Wood—The Use of Crinoid Arms in Studies of Phylogeny... 1-17 ©. ©. Mook—Notes on Camarasaurus Cope......c.cccc ccc cece or sce 19-22 Alexis A. Julien—The Genesis of Antigorite and Talc............... 23-38 Henryk Arctowski—A Study of the Changes in the Distribution of Temperature in Europe and North America dur- incr themnears: 1900-1900 a eee eee eee 39-113 Raymond Bartlett Earle—The Genesis of Certain Paleozoic Interbed- ded “iron IOre Deposits a. sc ricre tele tele siete 115-170 There is likewise in press a paper by W. D. Matthew entitled “Climate and Evolution” and one by H. von W. Schulte and Frederick Tilney en- titled “Development of the Neuraxis in the Domestic Cat to the Stage of Twenty-one Somites.” The first portion of Miss Laura E. W. Benedict’s paper on “Bagobo Ceremonial, Magic and Myth” is in press. This is to form the first paper of Volume XXV of the Annals which is to be de- voted exclusively to anthropological papers. The Publication Committee has accepted a paper by W. K. Gregory entitled “Present Status of the Problem of the Origin of the Tetrapoda,” for publication in Volume XXVI of the Annals. Respectfully submitted, EpmunpD Oris Hovey, Editor. REPORT OF ‘THE TREASURER MEMBERSHIP Paid up, Active Members (2 of these were elected after 1 May and paid fay tops Ie) Sis be Aa eaig od Scid do 8 clots OO Ope rece EerO Ice aioercib oo d.ocn oc. ec 212 IEG ih, LAO IM eNO. os caeod doo comp boa adeoceoCoomooeDUdoooddD ile Delinquent Active and Associate Members................-se-eeeeeee 51 Ibrity ravVerer) ares Avel VERON bo doc dbs ue OdeUOmOOe doa OG cD osOOn do adonsGoC 104 444? RECEIPTS DECEMBER 1, 1913—NovEeMBER 30, 1914 Cashvonnhand: Mecember! ds mOlS weiss ee cis_-) 1 Sele siclelersiehelels)oletbie $2,821.67 Mitewmembhership cee cike cle citer lelakelelala aXe clots vel steel ebelal-eleante 100.00: Income from investments: Interest on mortgages on New York City real estate.. $771.26 Interest on railroad and other bonds.............-. 1,325 .00 2,096.26 1 Including four deceased members whose dues haye been paid to the end of the year. RECORDS OF MEETINGS 105 iM KSSE Grol lophalic [MEMS 4 ac Oop Dodo ano sbeObbaudoLGOpeUODOOOr 40.55 AGHA: 1 1n) aeres onlyay Colbert AAS | caso d oc ouaceace auuOOdude 30.00 ff os FE AQUI era rondee cus ye fovict oueieveraahatetehecerese ts 115.00 fr es SSO LOD ee. saetel eve ust aiale caren etshoheers 2,710.00 ; 2.855 .00 ASSOGIALG MEMDELSHIP) GUS LOM aerate cata ale eel elslcrela ele! 3.00 i; “ So Mem noereee) svekevaveve euenctckeie vers vere 3.00 se ef SSN el QUAY vedatt ata tor sieiaeers ofeceust sielerevever 51.00 57.00 SAEeSHOfePUDITCATLOMSE re are cuotete cutie shelter sish ee of oa isi sche ersustele vuslevcl sel leterelerereys 195.85 ContributLiony toncosteol publication tickets ieicleiietels eiclctaciclel el eleleveleler che 250.00 SUNDSERIPELONS «CO marmialle cCaimm er GLOWS) cpesrecrcters ciate cierets cle leleie loreal elle 188.00 Esther Herrman Research Fund (return of grant)............... 150.00 Part payment on Deane-Brennan mortgage.............0.6.-.+-<. iGsaos Sale of Lawyers’ Mortgage Company’s bond.............:.......+ 1,000.00 ORO [iGo Swieveyy (Guiloseenoavon)) ogcarossaoscoscodoodccngcceuopcor 1,000.00 Porto Rico Government (refund of advances made on account of ALO LAGE XTVEIM SOS!) pera ata eicay arrester Mee be ov Wercr ay oiler os nace tie sal evaieral shebakore rele rerotavere 1,990.82 Cashmore ote umMlyaiike tars oemcrs eteceee ove: eneiskowcoveiers ereres sels eusreu gio ler seekers 3,000.00 Loan from American Museum of Natural History................. 250.00 MOYES Gene oe Oris EE oo LG creo ean DO Bei cite ae ae $17,158.48 DISBURSEMENTS DECEMBER 1, 1913—30 NoveMBER, 1914 Publicavionsronyaccount Of eATIMAIS tc siete calor ciel ielelelele eielbalalaeie ciate $2,165 .96 Bublicationm ote ule lyse ior, viets cowilosayoleielenoleva, slaves, #0/e0e evevele 406 elnieiels 607.86 VE COLGINey SCCretAMVIS MEX PENSE is cioteisis isis. slexcisies sie oe dievelaese.e thats clave ers 417 .82 Recording Secretary’s and Hditor’s allowances................c00. 1,500.00 WE CEUE WE OMIMICE CC aeercrs toner ane oe ehatoh oral ous 1s. soy Bis ololecy 19:84 fave ava evenayersiteatepatere 98.40 GeneralM@exqcnsSestrrrepewrcievaysccrcic, anes ec aee steteusteoni eve die eelatele aie ahevensuererels 149.60 HStherserrmanehescarem Hund) (C2amts)) cicero © cle ciel eels eels ec) ele ee 990.00 OMNES LEON Se NEG werIsy, eHiUI Gi (SLATES ie cejersie esl @ al oleic sie elele clepeliele/acleys 75.00 Jenni meen pve! hie CIE ego doeosacaeckooopooueaeoooede 226.30 ATE Ghoes, gles (GRAVE!) Gon ocoado bop Benoa nes OOO GUDOOOCOCUUOG 10.00 CG EMeralMeMneecin LS rarars stevale re ele slave oe se ais) steiciz evel) 9} Shelaraietslayslle w Giehere qelelie rene 861.20 VERA QUATEerS ee COMM|IETCE Stee seciciets aicheseleses =) =o: 6) olay © el oieile) 5, s/o [eisisyele elels 324.49 Purchase of mortgage from Lawyers’ Mortgage Company.......... 2,000.00 Porto Rico Survey (advances for field expenses)..........+-++-+- 3,000.00 Sections of Geology ands MineralOgiyien. mice «)> « «1s ciele oi< = e101) olelele © eles 20.43 SOGHOD Or TOMAR oe oéadeac oso O OS SOOO OPO OME O OUD OUGdCU So Od Dol 39.28 Special Membership, Committe@s. a5. mice sce. sco cee wcrc vencs 100.00 American Museum of Natural History (repayment of loan)........ 250.00 AVA ents OsMOterl mel ai Keaps mreyols les) eleteieiolstahs)