m PROCEEDINGS OF THE AMERICAN ACADEMY OF ARTS AND SCIENCES. Vol. LI. FROM MAY 1915, TO MAY 1916. BOSTON: PUBLISHED BY THE ACADEMY 1916 Ubc Cosmos press EDW. "W. WHEELER CAMBRIDGE, MASS. II I 3X. CONTENTS. Page. I. New Indo-Malayan Laboulbeniales. By Roland Thaxter . . 1 II. Polymorphic Transformations of Solids under Pressure. By P. W. Bridgman 53 III. A Quantitative Study of Certain Perthitic Feldspars. By Charles H. Warren 125 IV. The Glacial-Control Theory of Coral Reefs. By Reginald A. Daly 155 V. The Australian Honey-Ants of the Genv^ Leptomyrm^x Mayr. By William M. Wheeler 253 VI. Mitosis and Multiple Fission in Trichomonad Flagellates. By Charles A. KoFom and Olive Swezy 287 VII. Expansion Problems with Irregular Boundary Conditions. By Dunham Jackson 381 VIII. The Mechanics of Telephone-Receiver Diaphragms, as Derived from, their Motional-Impedance Circles. By A. E. Kennelly and H. A. Affel 419 IX. On the Development of the Coral Agarida fragilis Dana. ByJ. W. Mavor 483 X. (I) Compositae new and transferred, chiefly Mexican. By S. F. Blake. (II) New, reclassified, or otherwise noteworthy Sperma- tophytes. By B. L. Robinson. (Ill) Certain Borraginaceae new or transferred. By J. Francis Macbride 513 XI. On the Life-History of Ceratomyxa acadiensis, a new species of Myxosporidia from the eastern coast of Canada. By James W. Mavor 549 XII. Polymorphic Changes under Pressure of the Univalent Nitrates. By p. W. BRffiGMAN 579 IV CONTENTS. Page. XIII. The Pathological Effects of Radiant Energy upon the Eye. By F. H. Verhoeff, Louis Bell and C. B. Walker ..... 627 XIV. Records of Meetings 819 Biographical Notices 843 Officers and Committees for 1916-17 933 List of Fellows and Foreign Honorary Members 935 Statutes and Standing Votes 951 RuMFORD Premium 965 Index 967 Proceedings of the American Academy of Arts and Sciences. Vol. 51. Xo. 1. — Augu.st, 1915. CONTRIBUTIONS FROM THE CRYPTOGA.MIC LABORATORIES OF HARVARD UNIVERSITY. No. LXXVII. NEW IXDO-MALAYAX LABOULBEXIALES. By Roland Thaxter. CONTRIBUTIONS FROM THE CRYPTOGAMIC LABORATORIES OF HARVARD UNIVERSITY. No. LXXVII. NEW INDO-MALAYAN LABOULBENIALES. By Roland Thaxter, Received May 24, 1915. Presented May 12, 1915. The new forms which are described in the present paper have been obtained from several sources. A portion of the Javan material was found on insects purchased from the late M. Rouyer, and a small number were also obtained from certain hosts kindly collected for me by Dr. W. P. Thompson. All the forms from Samarang, which include a great majority of the Javan species, were partly found on miscellaneous insects which ]Mr. E. Jacobson was so kind as to have his assistant collect for me, while the infested crickets were discovered by Mr. Jacobson himself, who is thus the first to have observed that hosts of this nature were bearers of Laboulbeniales, and to whom I am especially indebted for this noteworthy addition to our knowledge of the group. Through the very great kindness of Mr. T. Fetch of the Peradeniya Gardens, to whom I am under deep obligations for his trouble in personally collecting miscellaneous insects which he has sent me for examination, our knowledge of the Ceylon forms has been very materially increased. A few specimens have also been obtained from the Museum of Comparative Zodlogy in Cambridge. I am further indebted for determinations of Staphilinidae to Dr. Max Bernhauer; of the forficulids to Dr. Malcolm Burr; of the other Orthoptera to Mr. A. N. Caudell, and for certain generic determina- tions to Mr. Schwartz and Mr. Barber of the National Museum. To all these gentlemen I desire to express my obligations for the trouble which they have taken in my behalf. The new and very striking forms on Gri/Uus and Gri/Uofalpa collected by Mr. Jacobson, have a special interest since not only are these new types of hosts for the Laboulbeniales, but the species which infest them are unusually interesting and peculiar. The two forms on GnjUus, which I have placed in the genus Laboulbenia, are noteworthy in that they combine the characters of this genus and of Ccraiomyces in such a way as to definitely confirm my previously expressed opinion 4 THAXTER. that the two genera might have to be united (see These Proceedings, L, p. 45). The genus Ceraiomyces was originally based on the diptero- philous C. Dahlii from New Guinea, in which not only are cells III-V replaced by a single cell but the appendage has the appearance of being single, that is of arising from one basal cell. In this species, also, a well de\eloped penetrating rhizoid is developed. What appears to be the basal cell in this species should probably be regarded, however, as the insertion-cell; the basal cells of the two branches which arise from it, representing the basal cells of the typical outer and inner appendage in Laboulbenia. A sufficient numlier of forms ha\"e now accumulated on Diptera and Coleoptera which, taken in connection with the two forms on Gryllus described below, make it evident that Ceraiomyces had best be discarded. The forms which ma}' be included in this general section represent a tendency toward the development of a more simple receptacle, just as the aquatic species, and some others, show a tendency toward greater complication than is present in the type-forms. In one of the two closely allied species described below on Gryllus, the black insertion-cell bears but a single basal appendage- cell, while cell \^ is present and is proliferous in a fashion resembling that seen in Laboulbenia proliferans. In the second species, which is so closely allied that it might almost be regarded as a variety only, the structure is exactly that of a typical Laboulbenia, although in both the host is penetrated by a well developed rhizoid. The species here- tofore described under Ceraiomyces may therefo're best be transferred to Laboulbenia which will therefore include the following forms: Laboulbenia Dahlii, L. Selenae, L. Epitricis, L. obesa, L. mi- niscula, L. dislocata, L. Trinidadensis, L. Chaetocnemae, and L. Nisotrae. Dimeromyces falcatus nov. sp. Male indiridual. Pale straw-yellow or nearly hyaline, consisting of usually five superposed cells (four to eight) terminating in a small unicellular appendage bluntly pointed, slightly tapering, about three times longer than broad; the basal cell running into the large long somewhat irregular foot, somewhat bent, about twice as long as ])road, narrower above the foot, with a broad contrasting distal black- ish brown band; the cells above usually successively somewhat smaller or subequal, all or most bearing single antheridia more or less definitely superposed, usually three to five in number. Antheridia long slender colorless; the stalk-cell sometimes even longer than the venter and NEW IN DO-MALAYAN LABOULBENIALES O neck; the venter without basal cells and consisting of two antheridial cells which discharge by rather long canals into the cavity of the neck, which is straight or curved and about as long as the venter. Total length, including terminal appendage-cell (12-15 /i) and foot (25 /x), SO-9-i XSfx. Antheridia, including stalk-cell, about 40 X 5 ix. Female it^dividual. Amber-brown, falcate and often slightly sig- moid, the cells thick-walled. Receptacle not distinguished from the primary appendage with which it is continuous; the axis consisting of usually twenty-five superposed cells, six of which belong to the receptacle proper; a basal cell considerably longer than broad, hya- line below a broad partly diagonal Ijlackish band; the other five cells very thick-walled with small rounded lumen, especially the lower four, the lowest of which produces a short simple rather closely septate pale sterile appendage which, bending beside the base of the perithe- cium above it, is more or less appressed against the receptacle and extends just above the base of the primary appendage, its origin be- coming obscured by displacement, so that it appears to arise from the basal cell of the receptacle ; three of the cells above it usually producing perithecia. Primary appendage simple, paler distally and tapering to a blunt termination, the cells squarish or slightly longer than broad. Perithecia two to normally three, arising from successive cells in a slightly oblique series, almost sessile, erect; the very short stalk- portion bent al>ruptly upward; the main body asymmetrical, slightly broader distally, the inner margin convex, the outer concave, tapering rather abruptly to the tip which is subtended by a slight elevation on the inner side; the rounded apex of its slightly curved distal portion bent outward. Perithecia 85-100 X 20-24 m- Spores, female, IS X 3 iJL. Total length to tip of primary appendage 350-390 jj.. Re- ceptacle 80 X 20 fx, face view X 30 /x. Secondary appendage 45-70 /x. On the antennae of Gryllus mitratus Burm. Samarang, Java. The many peculiarities of this species, which grows appressed on the antennae of its host, need hardly be pointed out. The male individual is of special interest, since its antheridium is of a type more simple than is usual in this genus; the antheridial cells, of which there appear to be never more than two, cutting off no basal cell, so that there are present only a long stalk cell, two antheridial cells and the neck for discharge of sperm-cells. The sixth cell of the receptacle in the female individual appears to be the basal cell of the primary appendage, which in other species is often separated from the append- age by a darkened septum. The single secondary appendage is so small and inconspicuous that it may be readily overlooked. THAXTER. Dimeromyces brachiatus nov. sp. Male individual. Receptacle hyaline, subtriangiilar above its stalk-like base, the lumen of which is nearly or completely obliterated; consisting of from fi^'e to seven cells obliquely superposed, those between the basal and terminal cells much flattened and each asso- ciated with a corresponding simple secondary appendage : the terminal cell larger, somewhat rounded distally, bearing the primary appendage distally and the single antheridium laterally. Primary appendage simple, rather short, hardly tapering, with blunt rounded apex, and consisting of a small almost square or slightly flattened basal cell, followed by one to two flattened yellowish browTi cells distinguished by dark septa; the rest of the appendage a single elongate cell. wSec- ondary appendages superposed in a single series, somewhat divergent and mostly in contact throughout; similar to the primary appendage, the basal cell large, the flattened suffused cells one to three in number. Antheridium solitary, arising just above the base of the uppermost secondary appendage, its basal or stalk-cell somewhat larger than that of the latter, subtriangular; the venter and neck relatively short, somewhat curved; the antheridial cells about eight in number, the basal cells clearly defined, the neck rather abruptly distinguished, but rather short and strongly bent. Receptacle 46 X 12 ix. Foot 18 /x. Appendages 28 X 4 ju. Antheridivmi 27 X 9 /x. Female individual. Hyaline or faintly yellowish. Receptacle similar to that of the male, consisting of eight to nine cells; the termi- nal one small, triangular, bearing the primary appendage terminally; the remaining cells above the basal cell bearing either secondary appendages, perithecia, or secondary appendiculate axes: the terminal cell always bearing a secondary axis laterally, the subterminal giving rise to the first perithecium and the cell next below to a secondary appendage, while the remaining cells may produce either of these structures without regularity. Primary and secondary appendages similar to those of the male, but smaller. Secondary axes greatly elongated, suberect, slightly flexed, of about the same diameter throughout, consisting of a single series of a hundred cells more or less very thick walled, arranged in vertical pairs, the successive pairs slightly displaced from right to left; the upper cell of each pair asso- ciated distally and externally with a snuill simple, closely appressed appendage similar in all respects to the secondary appendages of the male, and alternating right and left from successive pairs. Perithecia NEW INDO-MALAYAN LABOULBENIALES. / originating from a basal cell like that which subtends the secondary axes; the first always arising from the sjibterminal cell of the recepta- cle, others rarely arising from cells lower down; short-clavate, sub- symmetrical, broader distally, tapering to the rather small but abruptly broader apex; the lips more or less distinctly and broadly papillate. Perithecia 90-120 X 15-20 ix, its basal cell 12 X 8 m- Receptacle 50-00 X 20 m; foot 20 m- Primary and secondary ap- pendages 18-22 X 4 ^i. Secondary axes up to 580 X 12 n, their appendages 20-28 X 4 /x. On the elytra of Hetcrophaga sp. nov. No. 2107, Peradeniya, Ceylon. This remarkable species is clearly separated from all other known forms by its sterile and peculiarly appendiculate secondary axes. The host, which has been kindly examined for me by Mr. Champion, is said ])y him to be a new species near H. niiiduJa Motsch. Dimeromyces Petchi nov. sp. Male individual. Receptacle hyaline, siibtriangular, or sometimes more elongate, consisting of from seven to ten superposed cells; the basal cell short and triangular, or sometimes somewhat elongate; the rest obliquely superposed, flattened, each giA'ing rise to an antheri- dium ; the series terminated by the basal cell of the primary appendage which is subtriangular and not otherwise distinguished from the cells below. Primary appendage simple, its subbasal cell small, more or less suffused, and abruptly distinguished from the basal cell by a blackened septum, the cell above it slightly larger, but little suffused, and followed by the more elongate two-celled portion, the walls of which are distally swollen and may become more or less disorganized. Secondary appendages absent. Antheridia arising in a single row, more or less displaced by crowding, so that the series may appear double: short and stout, the stalk-cell usually slightly longer than broad, and protruding somewhat upward between the two antheridial cells from which no basal cells are separated; the venter not clearly distinguished from the stalk; the neck deeply blackened, contrasting; the efferent tube short, broad, truncate, slightly tapering, quite hyaline, diverging subterminally from the black neck at an angle of from 45° to 90°. Receptacle, inclufling the small foot and the basal cell of the primary appendage, 45-80 X 15-20 fx. Primary appendage 45 X 5.5 n near base. Antheridia about 28 X 9 ^t. THAXTER. Female individual. Receptacle hyaline, usually turned so that it is viewed edgewise, somewhat broader opposite the lower secondary appendage, consisting of a longer basal cell and three obliquely super- posed flattened cells terminated liy the undifferentiated basal cell of the primary appendage ; which is small and not otherwise distinguished from the cells below it; the cell above it is distinguished by a black septum, but is not deeply suffused, and is in general like that of the male, though sometimes smaller. A single secondary appendage arising from the subbasal cell of the receptacle, similar to the primary, and separated from its basal cell by a contrasting black septum. Perithecium usually arising from the subterminal cell, of the receptacle, rarely from the terminal, above the secondary appendage; erect or l)ent sidewise at the base, sessile or the stalk very short, the ascus- apparatus filling the whole cavity; asymmetrical, transparent, pale smoky brown, distally rounded outward on one side below the rather abruptly distinguished tip, the base of which is more deeply suffused with blackish ; forming a more or less definite transverse black hand, above which it is quite hyaline except for a deep blackish suffusion just below the hyaline apex; which is rounded, slightly asymmetrical and bent. Spores 30 X 4 /x. Perithecia 75 X 20-24 ^t. Receptacle 55-65 X 25 fx. Appendages about 40-45 X 7-S m- Total length to tip of perithecium 120-135 fx. On the right inferior posterior surface of the prothorax of a small carabid allied to Tachjs, Peradeniya, Ceylon. No. 2093b, and Samarang, Java, No. 20Slc. Known also from Borneo and the Philippines. This species is not nearly allied to any known form. The antheridia are unusually numerous and quite unique in form and appearance owing to the deeply blackened neck and abruptly bent hyaline dis- charge tube. I have taken the liberty of naming this very distinct form for Mr. T. Petch, whose admirable work on the fungi of Ceylon is well known to all mycologists and to whom I am indebted for all the Ceylon species herewith described. Dimeromyces appressus nov. sp. Male individual. Receptacle lying nearly flat on the substratmn, the antheridium toward the female; two-celled, the basal cell suffused, broader than long, extending outward beneath the antheridium, the base of which rests on it; subl)asal cell concolorous, subtriangular, NEW INDO-MALAYAN LABOULBENIALES. 9 much narrower, lateral in relation to the base of the antheridiuni. Basal cell of the appendage hyaline, separated from the rest of the appendage by a thin dark septum, the rest of the appendage about the same diameter throughout, two to three celled, the basal cells shorter, the tip bluntly rounded. Antheridium solitary, arising from the subbasal cell of the receptacle, its base resting almost wholly on the basal cell; its stalk-cell short and broad, distally slightly pointed; the antheridial cells four with clearly distinguished basal cells; the venter short and rather stout; the neck stout, purplish, not abruptly distinguished, relatively long and stout, curved and slightly tapering throughout. Receptacle, exclusive of foot, 10 X 9 m- Antheridium 18X6)U. Appendage 20-24 m- Total length to tip of appendage about 30 jj.. Female individual. Receptacle prostrate, its three successively larger cells flattened, somewhat obliciuely superposed, not distin- guished from the basal cell of the appendage; the whole, including the foot, subtriangular; the lower (outer) margin straight, the upper divergent from the foot to the upper angle of the basal cells from which it conAerges to the thin dark terminal septum which separates the basal cell from the rest of the primary appendage; which is two to three celled, rather stout, the lower cells short, the tip bluntly rounded. Secondary appendage solitary, arising from the subbasal cell of the receptacle; when growing on the abdomen bent abruptly upward at right angles to its axis, erect, its long basal cell sometimes rather deeply suffused, and followed by two nearly equal thin-walled purplish cells, slightly longer than broad; the rest of the appendage mostly hyaline, one to two-celled; when growing on the foi'ceps smaller, shorter, and lying nearly parallel to the axis of the receptacle. Perithecia solitary, arising from the thirtl cell of the receptacle, curved abruptly upward at its abruptly broadened base, erect; the stalk elongate, broader below, distally merging gradually with the faintly purplish body of the perithecium ; which is hardly distinguished from it, slightly asymmetrical, its inner margin nearly straight, the outer convex; the tip not distinguished, except near the apex which is subtended externally by a more or less well defined hump or promi- nence; the lip-cells Aariably somewhat asymmetrically prominent, slightly oblique inward, that is away from the secondary appendage. Body of perithecium 45 X 12^ (ascigerous portion), the stalk 70- 80 X 10 M, at base. Spores about 28 X 2. .5 m- Receptacle, including basal cell of appendage and foot, 20 X 10 m- Primary appendage, exclusive of basal cell, 12-16 X 5 yu. Secondary appendage 35-45 X 5.5 iJL. 10 THAXTER. On the inferior surface of the abdomen near the tip and on the forceps of Labia pilicornis Motsch. Peradenyia, Ceylon, No. 2112. This species is allied to D. Labiae and D. rninutii'tsimus from both of which it is distinguished by its pedicillate perithecium, as well as by other points of difference. The specimens obtained from the al)domen differ distinctly from those which grow on the forceps, the secondary appendage of the latter lying parallel to the receptacle and primary appendage, instead of projecting at right angles as in the material from the abdomen, in which this appendage is also longer, darker and more conspicuous. The two seem otherwise practically identical. Rickia rostrata nov. sp. Axis hyaline indeterminate broad, somewhat narrow below, elongate, its margins somewhat uneven, the cells relatively large and thin-walled. Foot small, basal cell of the receptacle broader than long, the sul)l)asal cell flattened, slightly intruded between two small nearly equal cells which lie right and left above it and form the bases of the two lateral rows of the axis which consist of three cell-rows: the cells of the axial row somewhat larger than the others, seldom o\er twenty in number, rarely twenty-seven but often less, the lower more or less elongated vertically, those below the perithecium often shorter, the distal por- tion of about six small cells ending below the base of the primary appendage and in contact with the venter of the perithecium, the wall-cells of which become obliterated in this region: the cells of the anterior and posterior rows similar, narrower, a majority of them appendiculate, the anterior row of seldom more than twenty cells, often less, rarely twenty-five : the posterior row of twenty-two cells or less, rarely thirty, ending at the base of the rostrate tip of the perithecium, where it bears the primary appendage. Primary appendage smaller than the secondary ones, short, hyaline, evanescent ; the base relatively small, truncate-conical, projecting free at an angle of 45° or less to the axis, distally suffused. Secondary appendages hyaline, variably elongate, mostly evanescent, subtended by a small cell separated from the axis-cell which is prolonged to form a slender dark purplish brown base simulating an appendage, free and projecting horizontally out- ward, or more often obliquely upward. Antheridia similarly borne, their irregular cells becoming more or less free in a mucus group. Venter of the perithecium hyaline, or faintly tinged with yellowish brown, broad, rounded outward, the ascogenic cell placed laterally NEW INDO-MALAYAN LABOULBENIALES. 11 on the inner side, the distal portion contracted abruptly to form a characteristic rostrate termination almost twice as long as the venter, I^urplish brown, lighter and slightly constricted above the middle; the apex blunt, often slightly bent. Spores 20 X 7.5 m- Venter of perithecium, exclusive of marginal cells of axis, 42-48 X 25-28 ^u ; the rostrate termination 63-72 X 10-12 ix. Secondary appendages 24 X 3.5 fjL, not including their projecting basal cells which are 7-12 X 2.5 M- Total length to tip of perithecium 240-450 X 20-35 /jl. On the right elytron of Taniignathus ruficoUis Kr., Java, and Sara- wak, Borneo. This species of which the specimens from Borneo are taken as the types, the Ja^'an material being immature, is aberrant in several re- spects; more especially in its rostrate perithecium with inflated venter and laterally placed ascogenic cells. The secondary appendages sug- gest those of Monoicomyces Lcptochiri in the general appearance of their dark projecting basal cells. Although a large species, it is not readily seen, since it lies perfectly flat on the elytron. Rickia Tomari nov. sp. Form rather short and stout, foot small, structure determinate or subdeterminate, nearly hyaline, except the deep brown hyaline-tipped perithecium. Basal cell somewhat longer than broad, slightly in- truded between, or broadly rounded and slightly overlapping a pair of subbasal, nearly equal, somewhat irregular cells above which three cell-rows are distinguished: an axial row of three larger and somewhat dissimilar cells, followed by three or two successively smaller cells which lie in contact with the base of the perithecium on the posterior side : a posterior row of usually six more regular and somewhat rounded cells, the upper two or three small and extending to or nearly to, the end of the axial series, terminating in the flattened basal cell of the primary appendage; all its cells except the uppermost, cutting oft' three or four small cells which lie side by side somewhat irregularly and horizontally, each giving rise to well developed antheridia of the normal type, or to short appendages, smaller than the antheridia and much more numerous, both distinguished by dark brown cup-like suffusions at the hardly constricted subtending septa: the anterior row consisting of usually six marginal cells, the two terminal ones lying beside and united to the base of the perithecium, and extending higher up than the corresponding cells of the axial row on the opposite 12 THAXTER. side; a seventh cell lying immediately below the base of the perithe- cium, between the third cell of the axial and the fourth cell of the anterior series, the members of which, except the small terminal one, cut off small cells like those of the posterior series; the lowest, how- ever, like the posterior subbasal cell, Ijearing only a single appendicu- late cell. Perithecium nearly symmetrical, erect, rich black-ljrown except at the base where it is yellowish, and at the contrasting hyaline tip; the venter slightly inflated below, tapering slightly above; the upper half of its suffused portion broad, with often nearly parallel margins, and slightly enlarged below the abruptly distinguished, contrasting, hyaline, nearly symmetrical, blunt-conical tip. Peri- thecia 80-90 X 22-24 /x- Antheridia 10 X 3.5 m- Total length to tip of perithecium 135-150 X 30-36 fx. On the elytra of Tomarus sp.. No. 2095: Peradenij'a, Ceylon. This species is not nearly allied to any other that is known to me, and is easily distingviished by its white-tipped, deep l)rown perithe- cium, the body of which is more or less clearly distinguished into an inflated ventral and broad neck-portion, above which the short tip is abruptly differentiated. Rickia marginata nov. sp. Hyaline or faintly yellowish, with thick white walls, the appendages only l:)ecoming slightly brownish, broad and flattened, very variable in size; the cells of the receptacle multinucleate. Basal cell longer than broad, broader distally, variably intruded between two paired elongated subbasal cells, the posterior somewhat longer, distally separated by a long single axial cell which extends above the base of the perithecium and, together with a small flat cell lying above it close against the venter and terminating in an appendage, constitutes the axial series: the anterior series consisting of the anterior subbasal cell and a normally single and elongate cell, which may occasionally be irregularly divided, and extends from its apex to the base of the perithecium; the posterior series also consisting normally of the posterior subbasal cell and a single greatly elongated cell, rarely divided transversely, followed by a smaller cell usually somewhat longer than broad, often lying nearly horizontally and ending in the basal cell of the primary appendage, around which it cuts off small, often numerous, appendiculate cells all of which bear single appendages not distinguishable from the primary appendage; the NEW INDO-MALAYAN LABOULBENIALES. 13 whole forming a dense terminal group, the members usually much more elongate than those which arise lower down: the two subbasal and the lateral cells also giving rise to numerous secondary appendages, subtended by small lateral basal cells, either closely associated to form a continuous margin, or variable and irregularly distributed singly or in groups; the lumen of these cells small, the walls very thick, the single cells or groups slightly prominent beyond the margin, each separated from the appendage by a large cup-like brownish l^lack suffusion at the constricted septum: the appendages variably elon- gated, unicellular, cylindrical, or irregularly somewhat inflated, or often slightly tapering. Antheridia apparently few in numl)er, becoming more or less free and irregularly developed. Perithecium concolorous, one half free, and convex on the inner side; wholly free and straight, or slightly concave externally, usually projecting out- ward at a slight angle to the axis; relatively small in well developed individuals; the walls thick; the tip not distinguished, truncate conical or bluntly rounded. Perithecia 75-80 X 35-38 /i. Spores 46 X 4.5 M- Receptacle 150-625 X 46-85 fj.. Lateral appendages 30-80 X 5 m; terminal up to 300 X 8 /x. In various positions on Heterophaga jmncfulata Motsch. (det. Champion); No. 2108 Peradeniya, Ceylon. As will be seen from the above description, this species is clearly distinguished by the unusual character of its perithecium, the greatly elongated cells of which separate often very numerous appendiculate cells seriately arranged along their margins. The occasional occur- rence of septation in these long cells appears to be a secondary phe- nomenon. The antheridia have not been very satisfactorily made out, but, as in some other instances, appear to become more or less free in irregular groups. Although far more highly developed this species remotely resembles Rickia Lispini in the character of the cell-series which form the receptacle. Rickia Coptengalis nov. sp. Rather long and broad with blackish brown and yellowish brown suffusions; the receptacle of about the same width throughout, except where it becomes narrowed at the base; the basal cell narrower than the somewhat rounded foot, hyaline and contrasting with the two suffused, paired, and nearly triangular cells above it, between which its bluntly rounded distal end is intruded. Above this pair of cells the 14 THAXTER. receptacle is triseriate; the lower halves of the basal cells of the two lateral series in contact, the upper separated through the intrusion of the basal cell of the middle series. Cells of the middle series slightly longer than broad, for the most part, thirty-two to thirty-four in number, their walls becoming suffused with yellow brown from below upward; the distal eight or nine small, flattened, with rounded con- tour, partaking of the yellow brown suffusion of the perithecium to which they are united; the series ending a short distance below the base of the deeply suffused tip, and opposite and within the basal cell of the primary appendage. The anterior row consisting of from forty-two to forty-four cells, the lower two or three deeply suffused, the suffusion decreasing upward; these, and a few cells above them, decreasingly ol)lique in relation to one another; three or four of the distal cells of the series, which ends abruptly at the persistent insertion of the trichogyne which subtends the deeply suffused tip of the peri- thecium, being without appendages; while of the remaining cells, about eight or ten at both the upper and lower ends of the series, cut off one, or two, appendiculate cells from the upper outer angle; while all the rest cut off three such cells, the successive groups of threes forming a continuous and more or less symmetrically disposed lateral series, the successive groups partly separated by the external pointed prolongation which characterizes each of the cells of the main anterior axis in this region; the appendiculate cells bearing, subtended by the usual cup-shaped suffused septa, either short hyaline inconspicuous appendages, or longer curved flask-shaped purplish brown antheridia, the venters of which are at first quite hyaline, two antheridia more commonly associated with one appendage in these sets of three. The posterior row ending in the two-celled base of the primary append- age, the subbasal cell of which is small and free, while its basal cell appears to end the distal row, from the members of which it is in no way distinguished, and which comprises from thirty-five to forty cells similar in general to those of the anterior series, and producing an- theridia and appendages in a similar fashion, and subsymmetrically placed in relation to them. Perithecium rather long, slightly and subsymmetrically inflated, erect, or turned slightly to one side, deeply suffused with brown, translucent; almost wholly enclosed, except the broad short nearly symmetrical abruptly distinguished opaque tip, and a very small portion of its anterior margin; the apex hyaline, contrasting, abruptly distinguished, the distal margin slightly uneven from the minute projecting lip-cells. Perithecia 100-112 X 28-30 ^u, or including the marginal cells of receptacle, X 46-55 /x. Total NEW INDO-MALAYAN LABOULBENIALES. 15 length to tip of peritheciiim 450-500 n, greatest width 40 fx. Antheri- dia about 12 /x. Appendages about 8 ju. On Copfengis SJupardi Pasc, Island of Djilolo. No. 1779, M. C. Z.; the type on the margin of the right elytron ; the variety on the inferior surface of the thorax and prothorax. The variety of this species above referred to, was found on the same host, and differs in its smaller size, having a maximum length of 175 fjL, the members of its three cell-rows being half as numerous in each series; none of the marginal cells separating more than two appendiculate cells; the antheridia longer and somewhat more slender, the appendages longer and more conspicuous, the tip of the perithe- cium bent abruptly sidewise so that the apex appears to be evenly rounded and suffused, although it is in reality hyaline and lateral in position. Although the differences just mentioned appear to be constant, the two forms resemble one another so closely that it has not seemed desirable to separate them. The species is a very striking and beautiful one, peculiar from the almost symmetrical lace-like pattern of its cell-structure, and the very large numbers of conspicu- ous purple-necked antheridia which it produces. Rickia Onthophagi, nov. sp. Rather long, hyaline to pale straw-yellow, structure subdeterminate, foot relatively large. Basal cell longer than broad, slightly narrower and faintly suffused below; its distal septum horizontal separating it from two cells above which may be symmetrically paired or obliquely related, the posterior cell placed higher and pointed below; the recep- tacle above consisting of three cell-series; the basal cell of the axial series either lying wholly above the basal cells of the two lateral series, or intruded between them for half their length: the axial series con- sisting of thirteen or fourteen cells of unequal length, but longer than broad, the upper six or seven cells much smaller and nearly isodia- metric, lying in close contact with the inner margin of the lower half of the perithecium : the anterior row consisting of nine or ten cells extend- ing to the base of the perithecium, each cell cutting off a vertical series of subtriangular small cells, three or four from the upper to one from the lowest members, all of which give rise to antheridia, or rarely to very small appendages: the posterior series similar to the anterior, consisting of about twelve to fourteen cells, and terminating in the two-celled, somewhat prominent, divergent base of the primary 16 THAXTER. appendage. Perithecium one half free on the inner side, wholly free externally, concolorous, the l)ody long elliptical or broader below, the tip abruptly distinguished, short, broad, purplish Ijrown, except the bluntly rounded subhyaline apex. Autheridia relatively long, curved, sharp and beak-like. Appendages minute, hardly distinguishable and very few in number. Perithecium G5-100 X 28-35 /jl. Spores 50 X 7 ju (in perithecium). Antheridia 15 X 5 /i. Appendages 4 X 3.5 M- Total length to tip of perithecium 260-340 X 40-44 m- On the inferior tip of the aI)domen of Onthophagus sp. No. 2094, Peradeniya, Ceylon. This species which is the first member of the family reported on Scarabeidae, presents no very striking peculiarities. It is distin- guished by its very numerous antheridia and scanty minute appen- dages. Rickia compressa, nov. sp. Straight or slightly curved, rather stout, subsymmetrical, wholly hyaline. Basal cell very small, subtriangular, the subbasal broader, flattened, sometimes obliquely divided, followed by two superposed pairs of cells above which the receptacle is triseriate; the middle row consisting of from thirteen to fifteen cells, the eight lower squarish, the upper ones extending along the posterior margin of the perithecium against which they are flattened, the two to three distal ones free externally above the insertion of the primary appendage; the free divergent basal cell of which terminates the posterior row of nine to twelve cells; the anterior row^ consisting of eight or nine cells, its external margin evenly continuous with that of the perithecium, to the base of which it extends: a variable but small number of the cells of both the anterior and posterior rows separate a small cell distally and externally, which form the basal cells of the hyaline slightly tapering secondary appendages. Perithecium externally free, straight, erect, rather stout, but slightly inflated, the tip abruptly distinguished, com- pressed, symmetrical, ending in a blunt nearly symmetrical apex. Perithecium, 00-75 X 20-22 /x, not including marginal cells of the median series. Receptacle, to tip of primary appendage-cell, 140- 155 /x. Secondary appendages 20-40 X 4-5 /x- Total length to tip of perithecium 175-200 X 32-35 yu- On the antennae and prothorax of Lrpfochirus sp., vel aff. No. 1417, Java, (Rouyer). This species is described from four specimens in good condition but NEW INDO-MALAYAN LABOULBENIALES. 17 may show more variability than is indicated by the diagnosis when abundant material is available. It is most readily distinguished by the abruptly differentiated compressed tip of the perithecium. It differs from R. Lcptochiri in having a triseriate receptacle. Rickia Uropodae nov. .sp. Form rather stout, habit more or less crest-like, axis indeterminate. Basal cell abruptly distinguished, more than twice as long as broad, slightly curved, with thick brownish yellow walls, and followed by two small paired cells symmetrically placed; above which the cells of the axis are triseriate, hyaline: the axial row united to the perithecium laterally nearly to the apex, three or four of its distal cells extending free beyond the insertion of the primary appendage, some of them bearing secondary appendages, its cells below the base of the peri- thecium small and squarish: of the two lateral axis-rows one, the anterior, is more nearly straight or not strongly curved, usually consisting of from eight to twelve cells, terminating at the base of the perithecium; one to four of its slightly prominent and radially elon- gated cells bearing at irregular intervals single, relatively large, com- pound antheridia which are subtended by single small triangular cells: the posterior lateral series, which is strongly curved, consists of from twenty-five to thirty or more cells which are somewhat more elongated radially (broader), and all of which without exception give rise to large pear-shaped bladder-like appendages, hyaline, becoming brown, short, slightly curved outward, subtended by a single small cell to which they are attached by a narrow stalk, and distinguished by a dark septum; the series ending in the large highly differentiated primary appendage, the terminal (appendage-) cell of which is evanescent; the two-celled sessile base large, cylindrical or slightly narrower dis- tally in the region of its upper and much smaller cell; the whole brownish yellow, free, projecting outward at a small angle from the axis below it, its origin about three quarters of the distance from the base to the apex of the perithecium. Perithecium dark rich brown, wholly free externally, turned to an almost horizontal position by the curvature of the axis; the tip broad, not very clearly distinguished, about half free on the inner side, the apex blunt, broad, somewhat asymmetrical. Perithecium, exclusive of marginal cells, 60 X 18 ju. Basal part of primary appendage 16 X 8 ju. Secondary appendages 18 X 9 M- Total length 100-120 X 30 ju. 18 THAXTER. On various parts of a species of Uropoda parasitic on large Passali, Java (Thompson). A very distinct and beautiful species, occurring rarely on various parts of its host, more commonly dorsally, sometimes in company with R. Bcrh'siana. The antheridia are very well developed, and their compound character is more clearly distinguishable than is usual in species of this genus. They do not persist, however, and usually appear like more or less broken appendages after the perithecium has begun to mature. It is more nearly allied to R. cristata than to other described species, but is clearly distinguished by its pear-shaped secondary appendages, as well as by numerous other details of struc- ture. Rickia uncinata nov. sp. Axis slender, elongate, hyaline, simple, or occasionally branched, consisting of three parallel rows above the basal cell. Appendages two or three to ten, scattered irregularly, short, nearly cylinch'ical with broadly rounded apex, the basal cell minute, subtriangular, hardly projecting and distinguished by a black septum; an antheri- dium usually subtending the perithecium on its outer side; the primary appendage three-celled, the terminal cell shorter than the secondary appendages and more inflated, distinguished by a blackened septum from its two-celled base which is subcylindrical and projects free at an angle of about 45°, terminating the outer axis cell-row opposite the middle of the perithecium. Perithecium terminal, hyaline, or be- coming slightly suffused with brownish, slightly curved, the median axis-row extending along its convex margin for about three fifths to two thirds of its length, and ending just below the persistent blackened base of the trichogyne; the tip contrasting brown, the apex partly hyaline, the whole abruptly recurved outward. Perithecia 38-42 X 12-16 /z, including the marginal cells. Primary appendage, including base, 24 ij. long; secondary appendages 12X4/^. Total length 150-500 X 10 M- On large Passali from Java. (Thompson). This species resembles R. nutans in general habit, but is at once distinguished by its three-ranked axis, and the absence of apical appendages on the perithecium. It rarely branches, except when injured. A single specimen found grow'ing on the leg of a species of Macrochcles parasitic on the same host, appears also to belong to this species. This individual, however, is relatively short and stout, its NEW INDO-MALAYAN L.\BOULBENIALES. 19 receptacle even shorter than the perithecium, the marginal portion beside the perithecium being more prominently developed, almost all the cells here bearing appendages; the primary appendage projecting forward against the tip of the perithecium, which is not othenyise exactly like that of the type. Rickia nutans nov. sp. Receptacle colorless, slender, elongate, often branched, the axis consisting of two rows of cells above the basal cell; the three-celled primary appendage remaining above the third cell from the foot, the double axis extending indeterminately beyond it, its distal cell small and distinguished by a finally blackened septum, the other two cells forming a stout basal portion which is wholly free, distinguished by a constriction from the cell which bears it and broader toward its base. The two cell-series, whether primaiy or secondary, otherwise without appendages, except one to three which always subtend the perithecium below its convex side: the axis terminations curved and slightly broader as the perithecia are reached, about five flattened cells of one of the cell-series forming a hyaline narrow contrasting margin, extend- ing to the apex of the perithecium on its convex side; one to three of the terminal cells of the other series larger, obliquely elongated laterally and upward, the uppermost in oblique contact with the perithecium for a short distance above its base. Perithecium continuing the curvature of the axis which bears it, rich brown, contrasting, the tip much darker, but not otherwise differentiated; two of the lip-cells forming relatively long, stout, tapering, divergent, blunt appendages, which are directed vertically or obliquely do^\^lward. Perithecia 58-66 X 20 /u, the apical appendages 15-16 ju. Total length 750 n or less, the diameter about 12 jjl, just below the perithecium about 16 n. x\t the tip of the abdomen of large Passali. No. 2114a, Peradeniya, Ceylon. This species is very clearly distinguished by its brown appendiculate perithecium and nodding habit, which suggests the head and neck of a flamingo. The third row of axis-cells is not developed as in more typical species, and is represented by the single cell which subtends the primary appendage near the base. 20 THAXTER. RicKiA Berlesiana (Bacc.) Paoli. This species is not uncommon on large Passali in Java, Ceylon and Australia, as well as on various genera of mites which infest them. Among the latter it has been found on Canestrinia sp., Urojjoda sp., Cclacnopsis sp., and Macrochdus sp. which were collected for me in Java by Dr. W. B. Thompson. On Passali it grows, as a rule, much more luxuriantly, sometimes reaching a length of 800 ^i. It is a grace- ful and very striking species, and is easily distinguished by the dark brown contrasting color of its perithecium and axial row of cells. Unlike most other species having a similar long and slender hal)it, it does not appear to produce secondary branches of the axis, even when the primary perithecium has been destroyed, although nearly sessile secondary perithecia are occasionally met with. RiCKiA DiscoPOMAE Thaxter. Since this species was descril)ed on the mite Diseoporna, it has been found in far better condition growing on the passaline beetle on which the latter is a parasite. It may attain a length of a millimeter, and often branches several times irregularly, as many as six perithecia being sometimes developed at the tips of a corresponding ninnber of axes. No individuals have been noticed on Javan material. Tettigomyces nov. gen. Receptacle consisting of an indeterminate series of cells superposed in a single row, or the distal ones longitudinally divided; foot large, black, without penetrating rhizoids. Appendage clearly distin- guished, or a mere continuation of the receptacle: in the type consist- ing of a short series of cells each of which gives rise on its inner side to two opposed series of usually paired antheridial cells, the two series of paired cells arching over a central cavity into which the sperm-cells are discharged, and which opens by a subterminal pore; the cushion- like compound antheridium thus formed compact and clearly defined in the type, while in other species the antheridial cells may be indis- tinguishable, or more or less irregularly associated in rows with the bases of sterile branches which may arise from the appendage. Peri- NEW INDO-MALAYAN LABOULBENIALES. 21 thecia somewhat indeterminate, the wall-cells numerous in each of the four rows; solitary, or several developed at intervals from an elon- gate receptacle. Trichogyne branched, more or less persistent at the base of the perithecium on the inner side. Asci eight-spored. Spores 1-septate; the basal (upper) segment twice as long as the terminal and with lateral cushions at the tip. Ascogenic cells more than two. Owing to the unusual variations which are exhibited by the diiferent species of this genus, it is very difficult to define it satisfactorily. In T. Gryllotalpae, which is taken as the type, the antheridium may be as clearly defined as it is in Eucantharomyces. In fact these organs are very similar in general appearance in the two genera. On the other hand there are some species in which I have been unable to discover any signs of antheridial cells; and in others the latter are associated with the bases of certain sterile branches, arising near the base of the appendage, which closely resembles that of some species of Ceratomyces and its allies, a resemblance which is further accentuated l)y the char- acters of the multicellular receptacle and perithecium, and the position of the trichogyne which is left behind at the base of the perithecium. In fact a species like T. brcvis would be placed in Ceratomyces with- out hesitation, were it not for the presence of these peculiar groups of antheridial cells which, from analogy with the type, must be con- sidered compound antheridia. The genus must therefore find its place among the Peyritschiellaceae. There is a certain resemblance between some species of this genus and Spegazzini's Cochliomyces which, although its characters are not at all clearly indicated by the published figures and descriptions, appears to differ in possessing appendages on both sides of the perithecium. There is, moreover, a superficial resemblance to Edeinomyces which, however, differs in the character of its antheridia and determinate perithecia. The trichogyne is usually more or less persistent and might readily he mistaken for a branch of the appendage. The antheridium in the type is terminated by the peculiar spine found in Eucantharomyces and various other genera, which appears to correspond to the persistent apex of the spore. In the present in- stance this spine lies just beside the subterminal pore through which the sperm cells are discharged from the common antheridial cavity. The asci are certainly S-spored in some cases although I have not in every instance been able actually to count this number. It is somewhat remarkable that a single host should in the same locality be parasitized by so many distinct species, but although I have endeavored to reduce the mmiber which may be distinguished 22 THAXTER. as far as possible, I have been forced to the conclusion that at least seven must be recognized on the Javan host, which, although some of them may be mingled on the same substratum, retain their indiAidu- ality, without essential departure from their type form. I have even been uncertain whether the type itself might not properly be sub- divided. In addition to the eight species here described three others are known to me; two from Africa and one from South America. Tettigomyces Gryllotalpae nov. sp. Colorless to pale straw-colored, sometimes faintly suffused with brownish. Receptacle variably elongated, slender or rather stout, nearly uniform throughout, or somewhat broader distally; consisting of from twelve to forty or more superposed flattened cells which be- come squarish; the distal one divided by a longitudinal septum into two cells, one of which subtends the perithecium, while the other forms the base of the appendage. Perithecium solitary, straight, or somewhat curved outward, tapering more or less continuously from the base to the tip, often somewhat inflated below; the tip well, often abruptly distinguished, narrow and nearly cylindrical, straight or slightly bent, the apex symmetrically rounded or slightly oblique; the outer wall-cells 18-21 in each row, the inner 15-19; usually becom- ing more or less prominent at maturity, so that the outline may be conspicuously corrugated. Appendage erect, or diverging at right angles from the perithecium; variably and often abnormally devel- oped; consisting of from four or five to a dozen, usually somewhat flattened and obliquely superposed cells, slightly prominent externally, all of which, except the basal and terminal ones, may bear, on the inner side, paired double rows of antheridial cells; the terminal cell some- times bearing also a short sterile branch beside the terminal minute spine-like process. Perithecia 135-310 X 35-62 /x. Spores 45-50 X 4.5 fji. Appendage 40-60 X 25-40 m- Receptacle 550-1400 X 25- 65 IX. Total length to tip of perithecium 235-1560 /x. On the inferior surface of the abdomen and on anal appendages of Gryllotalpa Africana Palis. Samarang, Java. This species varies very greatly not only in size, as will be seen l)y the measurements above given, but in the character of its perithecium and appendage. The latter in the type form is more or less erect and stout, with the rows of antheridial cells lying subhorizontally or obliquely inward and upward, and this character is usually associated NEW INDO-MALAYAN LABOULBENIALES. 23 with a stouter receptacle, the outUne of the perithecium being more conspicuously corrugaterl and the tip broader, somewhat asymmetrical and slightly bent distally. The more common type on the abdomen differs at maturity in possessing a usually more slender receptacle, which may be very greatly elongated, the appendage projecting at right angles to the main axis and roughly triangular in outline; the rows of antheridial cells vertical, or even oblique outward; the sterile cells decreasing rapidly in size to the tip: the perithecium having a less prominently corrugated outline, and the tip longer, more abruptly distinguished, nearly isodiametric, slender, with a symmetrically rounded apex. A third variation which occurs on the inferior surface of the thorax and adjacent parts of the abdomen and legs, is usually smaller and stouter; the appendage becoming very soon disorganized and remaining as a yellowish mass beside the base of the perithecivmi. The latter is less characteristic in form; the tip not well distinguished and stout, the outline hardly corrugated. Although the association of characters in these three \'ariations is more or less constant, and the extremes might readily be separated as distinct species, intermediate conditions are sufficiently numerous to make the series more or less continuous. This species is by far the most alnuidant of those described, and from its large size and conspicuous black foot, is readily seen even with the naked eye. A species very closely allied, and perhaps identical, has also been examined from African Gryllotalpac. Tettigomyces pterophilus nov. sp. Nearly colorless. Receptacle slender, variably elongated, more or less uniform above the basal region which may be slightly narrower, straight, curved, or somewhat sinuous, the cells slightly prominent and separated by more or less distinct constrictions at the septa. Appendage, which is assumed to arise opposite the base of the upper- most perithecium, not distinguished from the receptacle, slender, elongate, curved outward, tapering to a sterile termination; bearing at irregular intervals from its inner side, single scattered branches of variable number, simple or usually not more than once branched, and resembling the termination of the appendage. Perithecia one to four, or even five, superposed and arising at variable intervals from the receptacle, the axial trichogyne more or less persistent and associated with a small appendage just above it, the l>asal cell of which appears 24 THAXTER. to produce a small number of antheridial cells. The perithecium straight, or somewhat curved outward, tapering continuously from its hardly inflated base to the tip, which is but slightly distinguished, nearly conical, the apex bluntly pointed, often surmounted by a short apiculus; the inner wall-cells twenty-four or less, the outer usually twenty-six, those below the tip slightly prominent. Perithecia 125- 190 X 25-50 ^t. Spores 45 X 2.5 m- Total length to tip of appendage 400-1000 X 20-30 M- On the wing-tips of GnjUofalpa Africana Palis. Samarang, Java. This species was found in considerable numbers on a single individual of its host, all the others examined being quite free from it. Although the perithecia are similar to those of T. brcvis, the two species illustrate the extremes of development in the receptacle of this genus. There is great variability in the development of different individuals, smaller forms with siilgle perithecia being nearly as abundant as those in which several are produced. Tettigomyces Indicus nov. sp. Receptacle variably elongated about the same diameter throughout or somewhat broader below the perithecia, consisting of sixty or fewer cells superposed in a simple series; the cells much flattened and irregular. Appendage continuous with the receptacle, not distin- guished from it, greatly elongated, simple, divergent, distally flexed inward, tapering; consisting of about eighty or less superposed cells. Perithecia one to several arising at intervals from the receptacle; usually solitary, rather stout and short, the body hardly narrower distally, but very abruptly distinguished from the tip, to the base of which its outline bends almost at right angles; the tip short and more or less pointed ; the outer and inner rows of wall-cells having seventeen and fifteen cells respectively, those immediately below the base of the tip becoming prominently rounded outward, the lower cells of the inner row not at all prominent and having half, or less than half the transverse diameter of the corresponding cells of the outer row. Perithecia 120 X 44 m- Receptacle 100-230 X 20-28 /jl. Appendage 200-400 M- On bristles from various parts of Gri/Uotalpa sp., M. C. Z. Scudder Collection; No. 2678; North India. Although numerous specimens of this species were obtained but few are fully matured and from these the perithecial characters ha^•e been NEW INDO-MALAYAN LABOULBENIALES. 25 taken. No signs of antheridial cells have been seen, and the append- age appears to be simple, without sterile branches. The general habit is somewhat like that of T. pfvrophUu.s, although it is at once dis- tinguished by the form of its mature perithecium. In only one of the specimens examined are two perithecia matured. Tettigomyces chaetophilus nov. sp. Quite hyaline. Receptacle stout somewhat broader distally, con- sisting of from four to eight single superposed cells, which may be followed by from one to three cells once divided longitudinally; the cells separated on the perithecial side much smaller than those which are continuous with the appendage. Appendage erect, its axis coinci- dent with that of the receptacle, sometimes slightly curved outward distally, consisting of from seven to seventeen flattened superposed cells, two or three of the distal ones bearing sterile branches succes- sively or irregularly from the inner side; the terminal cell usually bearing two such branches; the latter simple or once branched, com- paratively slender, short, and tapering. Perithecium divergent and strongly curved outward, especially distally, the main body tapering only just below the tip ; which is abruptly distinguished, rather short and stout, the distal and basal halves well distinguished, the distal tapering more rapidly to the bluntly pointed apex which may be slightly apiculate; the outer and inner rows of wall-cells containing usually twenty and eighteen cells respectively; the cells of the inner row, except about three just below the tip which are larger, having about one quarter to one third the transverse diameter of those in the outer row. Perithecia 100-120 X 2S-3o m- Receptacle 40-75 X 30-40 IX. Appendage 50-100 /x. Total length to tip of perithecium 150-200 IX, including the foot (40 m) which is sharply pointed below. On bristles of the 'abdominal antennae' of Grylloialpa Africnna Palis. Samarang, Java. This species is more nearly allied to T. galeata from which it is at once distinguished by its smaller size, and the entirely different con- formation at the tip of the perithecium. The extreme dift'erence between the transverse diameter of the cells of the outer row of wall- cells and those of the inner distinguish it from all other known forms, with the exception of T. Indicus. 26 THAXTER. Tettigomyces galeatus nov. sp. Quite hyaline. Receptacle broader distally, often much narrower toward the base, both margins usually even, consisting of from six to ten superposed flattened cells, followed by from three to six tiers of two cells each; those on the perithecial side smaller, and becoming divided radially and longitudinally, so that there are actually three cells in each tier. Appendage consisting of from six to nine cells, much flattened, and forming an erect, or usually but slightly divergent series; the terminal cell bearing two simple branches, two or three of the cells immediately below also bearing single simple stout tapering branches, usually absent from the remaining lower cells, from which double or single series of antheridial cells are separated on the inner side so as to form a more or less well defined antheridial cushion. Peri- thecium relatively large and stout, distinctly inflated below, when mature; tapering at first abruptly, then hardly perceptibly to the broad short tip which is well defined by an abrupt indentation of the margin on the outer, and a slight indentation followed by a slight elevation, on the inner side; the inner lip-cell forming a broad bluntly rounded terminal projection with small lumen, extending some distance beyond the pore which is lateral and external; the lateral lip-cells forming small papillae symmetrically placed on either side, and the outer lip-cell forming a similar, often less distinct subtending papilla, the whole tip having thus a somewhat galeate hal)it. Perithecia 190-225 X 55-75 /x. Receptacle 100-140 X 60-SO m- Total length to tip of perithecium 250-400 /x. On the inferior surface of abdomen of Gryllotalpa Africana Palis. Samarang, Java. This species occurred somewhat rarely in the material examined. It is easily distinguished from its nearest ally T. confusus by its galeate tip and sul^erect appendage, which also differs in the absence of branches near its base. Tettigomyces confusus nov. sp. Receptacle slightly broader distally, or nearly the same diameter throughout ; consisting of from six to ten cells superposed in a single series, followed by from two to five tiers of two cells each; those on the perithecial side smaller, and usually divided in a radial vertical plane NEW INDO-MALAYAX LABOULBENIALES. 27 SO that each tier becomes three-celled. Appendage not distinguished from the receptacle, consisting of usually six successively smaller cells; the series somewhat curved outward, and bearing irregularly paired simple, stout, tapering branches from the inner side; the basal cells of which, above the base of the perithecium, produce more or less con- spicuous groups of antheridial cells. Perithecium relatively short and stout, sometimes shorter than the receptacle, more or less evenly inflated below at maturity, tapering evenly to the tip, which is short, stout, and barely distinguished: the inner lip-cell prolonged to form a short blunt appendage; two, which are outer or lateral, ending in rounded prominences somewhat variably placed, forming two more or less distinct papillae lying side by side; the fourth forming a pointed apex which subtends the prolongation of the fu'st; wall-cells variably, usually conspicuously prominent, those in the outer rows twenty to twenty-four, in the inner eighteen to twenty in number. Perithecia 120-200 X 40-60 m- Appendage 40 m; its branches 100 X 20 ^ at base. Receptacle 80-175 X 32-50 m- Total length to tip of perithe- cium 230-390 M- On abdominal ' antennae ' of Gryllotalpa Africana Palis. Samarang, Java. This species is very closely allied to T. galcafus, but is at once dis- tinguished by the conformation of the tip of the perithecium which does not vary essentially in any of the numerous individuals examined. Tettigomyces brevis nov. sp. Nearly colorless. Receptacle short and broad, the foot large and broad; consisting of from three to seven single superposed much flattened cells, followed by from one to three cells once longitudinally divided, this biseriate portion continuous with the appendage on one side and the base of the perithecium on the other. Appendage not distinguished from the receptacle, consisting of a series of from five to eight much flattened successively smaller cells which curves out- ward, or may be even recurved somewhat; all the cells bearing rather stout erect or somewhat curved septate tapering branches of variable length, which arise more or less in pairs from basal cells associated with numerous antheridial cells which are more or less distinctly \isible as a rule. Perithecia variably elongated, the base slightly broader, tapering, usually very slightly, to the short stout tip; which is moder- ately well distinguished; the apex broad, slightly asymmetrical. 28 THAXTER. flattened, or irregularly rounded; the wall-cells more or less con- spicuously prominent, except distally ; the outer rows containing from twenty-seven to fifty cells, the inner from twenty-three to forty-one. Spores slender, 50 X 3 /i. Perithecia 200-450 X 40-60 m- Appendage about 40-60 /j., the branches about 120-150 /z. Receptacle 25-75 X 40-60 M- Total length 230-550 fx. On the inferior margin of the abdomen of GnjUofalpa Africana Palis., Samarang, Java. This species is very clearly distinguished by its short receptacle and more or less indefinitely elongated perithecium, the basal cells of which are more clearly defined than in some other species of the genus, and form an irregular cell-group at the junction of the appendage and receptacle. The appendage, though it bears numerous erect branches in a tuft, often shows a copious formation of antheridial cells, the arrangement of which is very irregular as compared with that seen in the type species. It is most nearly related to T. acuminatus. Tettigomyces acuminatus nov. sp. Nearly colorless. Receptacle stout, consisting of six to eleven flat- tened superposed cells often paired, followed by three or four which are once or twice longitudinally divided, the distal tier always of three cells, two of which are smaller, isodiametric and subtend the base of the perithecium ; while the third, which is nearly twice as large, sub- tends the appendage; the cells on the perithecial side, especially just above the foot, more or less prominent singly or in pairs, the opposite margin usually even. Appendage consisting of eight or more flattened cells from which a cell is separated on the inner side bearing a furcate branch; the antheridial cells inconspicuous or lacking. Perithecia rather stout, curved slightly outward, considerably and somewhat asynnnetrically inflated, tapering rather rapidly to the tip, which is abruptly distinguished above the rounded termination of the rest of the perithecium; tapering, relatively long, one or all four of the lip- cells prolonged to form a slender sharp process; wall-cells hardly prominent, except sometimes near the base on the inner side; the outer rows consisting of from twenty-four to twenty-six cells, the inner of twenty-two to twenty-four. Perithecia 150-200 X 40-60 fx. Spores (in perithecium) 40-45 X 3.5 m- Receptacle SO X 40 ix. Total length to tip of perithecium 250-310 /x. NEW INDO-MALAYAN LABOULBENIALES. 29 On the right margin of the inferior surface of the abdomen of Gryllo- talpa Africana Palis. Samarang, Java. This species is most nearly related to T. brcvis, from which it is most readily distinguished by the form of its perithecium, especially at the tip, as Avell as Ijy its somewhat more highly developed receptacle. The appendages in the six types are all in bad condition and it is not possible to determine the exact character of the secondary branches which it bears. The latter appear to I)e similar, however, to those of T. brcvis and T. cladophorus. Dichomyces gracilis nov. sp. Basal cell of the receptacle hyaline, broader than long; middle cell of the lower tier barely translucent, the lateral cells opaque, wholly united to the middle tier, and extending upward above the base of the terminal tier, no portion being free ; the middle tier pale dirty yellowish brown, concolorous with the terminal tier and the perithecia, and consisting of series of from six to fifteen cells on either side of the median cell; the two series bent upward and usually strongly inward, so as to overlap the distal tier nearly to the insertion of the primary perithecia; the terminal tier consisting of series of from six to fifteen cells, on either side of the median cell, which bend upward rather abruptly. Perithecia two to four, usually two or four, rarely more, very slender and elongate, with a venter, neck and tip, and even a stalk-portion often more or less distinctly differentiated at maturity; the ^•enter slightly inflated just below the mid-region, and narrower toward the base, faintly purplish, tapering slightly to the neck-part which is perhaps a third as long and of equal diameter throughout; the tip slightly darker, moderately well distinguished, tapering to a broad flattish apex without appendages. Appendages relatively short, hyaline, the basal suffusion unusually prominent. Antheridia well de- veloped, purplish brown. Perithecia 200-300 X 22-30 /jl. Receptacle, to base of primary' perithecia, 125-160 /x, to tips of lateral series 150- 235 /i. Appendages 20-25 /z. Greatest width distally 60-110 m- On the anal appendages of an undetermined Staphylinid near Phi- lonthus. No. 1416, Java (Rouyer). This species is most nearly allied to D. Argentinus Speg. from which it appears to differ in its peculiar incurved middle tier, the form of its perithecia, which occur only on the terminal tier, and in its much shorter appendages. 30 THAXTER. Monoicomyces Leptotrachelae nov. sp. Basal cell small, the lower half suffused with blackish; the subhasul cell hardly distinguishable as such, owing to the densely crowded branches and branchlets bearing perithecia and antheridia which arise from it on either side of the main appendage, the basal cell of which is about as broad as long, the subbasal abruptly narrower^ edged externally with blackish, the suffusion extending up along the margin of a short outcurved branch which arises from it distally and externally; one or two hyaline stouter branches also arising from it distally. Perithecia usually six in number, the stalk-cells moderate, broad distally, the basal cells large and hardly distinguished from the somewhat inflated venter, above which the body is subconical, tapering gradually and subsymmetrically to the blunt apex; the tip distinguished only by a slight irregularity of outline. Anther- idial appendages usually eight, the stalk-cells arising close to that of the perithecium, about twice as long as broad, the pair next above slightly longer than broad and producing no antheridial cells, only the lower of the two pairs above it, associated with antheridia; their cells subequal and about as long as broad ; the whole appendage of nearly equal diameter throughout, the terminal cells producing four to eight stout branches of variable length, which are usually curved outward, or even slightly recurved. Perithecia 75-110 X 20-27 fj.. Body of antheridium 40 X 12 /x, its branches 40 X 5 /x. Total length to tip of perithecium, largest, ISo n. On the abdomen of Lcptotrarhda Jumna Bernhauer. Samarang, Java. The primary appendage of this species recalls that of M. Echidno- glossae, while its closely crowded perithecia and antheridia give it a general habit not unlike that of M. Alcocharae. It is more closely allied to the first of these species, but differs from the fact that the subbasal paired cells in the antheridial appendages are not.associated with antheridia which are produced only from the small pair next above. Monoicomyces Stenusae nov. sp. Hyaline. Basal and subbasal cells subequal, small, but clearly distinguishable. The primary appendage consisting of a small more or less dome-shaped basal cell, separated from the small cell above by a deeply blackened septum continuous with the outer blackened margin NEW INDO-MALAYAN LABOULBENIALES. 31 of the subbasal cell, which is also continuous with a similarly blackened branch, the whole forming a black outcurved or even recurved process from the convex side of which one or two hyaline erect branches or branchlets arise. Secondary receptacles variously developed, typi- cally two, growing in opposite directions, rather slender and usually curved; their basal cells short, and giving rise at once to the first perithecium and antheridial appendage; the second much longer, bearing the second perithecium and antheridium distally; the third still longer and usually terminated by two antheridia. In less well developed specimens only one branch may develop, with a single perithecium and antheridium; or both, if developed, may be much shorter and the habit more compact. Perithecia somewhat variable, typically with a well developed stalk-cell, which is narrower in the middle and broader distally than the small compact basal cell-region of the perithecium; which is abruptly distinguished, also, from the suddenly inflated venter above ; the venter short, the rest of the body long, tapering evenly to the blunt point, the junctions of the wall- cells indicated by two successive elevations by which a neck and tip are distinguished. Antheridia relatively long and slender, the basal cell-pair and the subbasal about equal; the cells of the two antheri- dial pairs bulging distally so that this portion of the antheridium is marked by successive elevations and depressions; the distal cells giving rise to from two to four usually curved rather short branches. Perithecia; longest, including basal cells 135X34;u; stalk-cell 40 X 12 ;u. Body of antheridium 58 X 9 )U, but very variable. Total length to tip of perithecium, longest 200 ijl. On various parts of Stcnusa Ccyhnica Kr. No. 2085, Samarang, Java. The short compact forms occur on the legs, the more highly de- veloped specimens on the abdomen and elsewhere. Although not departing widely from the usual type this species does not appear to be nearly allied to any other form, when well developed. Monoicomyces Amauroderae nov. sp. Hyaline, except for the brownish yellow perithecia. Basal and subbasal cells about equal in size, hardly longer than broad; the primary appendage simple, consisting of five or six superposed cells; the basal cell bent abruptly upward from the receptacle, and distin- guished by a small !)lack septum. Fertile branches typically two 32 THAXTER. growing in opposite directions, sometimes more; usually consisting of a single cell somewhat longer than broad, distally pointed, and bearing the antheridium and perithecium which diverge right and left at nearly a right angle. Stalk-cell of perithecium very elongate, abruptly distinguished by a slight distal enlargement from the basal cell-region, which is also relatively very large and long, and slightly broader distally than the venter of the relatively small perithecium which is hardly inflated and not distinguished from the usually curved rather elongate distal portion which tapers slightly to the l)lunt apex; the tip hardly or not at all distinguished. Stalk-cells of the antheri- dium relatively very long, (each al)out eight or nine times as long as broad), the basal pair broader distally and, abruptly distinguished from the narrower base of the second pair which is shorter and also distinctly broader distally; abruptly distinguished from the two pairs above, the cells of the upper smaller and separated from the lower by a distinct constriction, bearing directly two terminal appenflages, one usually longer than the other, sometimes only one, without evident basal cells in some instances. Perithecium: stalk-cell 156-275 X 20 /i; basal cell-region 55-65 X 20-28 /x; main body 118 X 20-25 m- Spores 30 X 4 M- Antheridia 62-82 X 18-20 m- Receptacle 32 m- Primary ap- pendage 125-175 X 10-15 /i. Antheridial appendages 40-150X8^1. On the inferior abdomen and thorax of Amaurodvra Kraepclini Fauv. No. 2078, Samarang, Java. A species well distinguished by its peculiar long-stalked perithecium, with highly developed basal cells, and its simple primary appendages, as well as by its unusually elongate antheridial appendages. Monoicomyces denticulatus nov. sp. Basal cell hyaline slightly narrower distally, somewhat broader than long; subbasal cell hyaline much smaller, bearing distally the some- what divergent primary appendage which is distinguished by a black- ish brown septum and consists of a single cell, nearly twice as long as broad, stout, distally rounded and externally rather deeply suffused with blackish brown. Fertile branches two; consisting of single more or less deeply suffused cells, which arise on either side of the sub- basal cell, and bear single antheridia and perithecia. Antheridial appendages symmetrically paired, slightly smoky, their stalk-cells hardly distinguishable from the basal cells, deeply suffused externally, nearly twice as long as broad, the two pairs above much smaller, both NEW INDO-MALAYAN LABOULBENIALES. 33 producing antheridia; the cells of the lower decidedly larger than those of the upper; the terminal pair closely associated to form an evenly rounded termination, in the types without appendages of any kind. Perithecia short and stout, pale straw-colored with a slight smoky tinge; the stalk-cell short, broader distally, somewhat longer than broad ; the body asymmetrical, one margin straight and ending in a tooth-like, brown, erect and slightly cur^■ed appendage at the tip; the other slightly and broadly convex, and abruptly indented below the small hyaline broadly rounded apex formed by the other lip-cells, which are slightly exceeded by the tooth-like appendage. Perithecia 125-134 X 40 IX, "the stalk-cell 40-45 m- The tooth-like process IS X 10 fx at base. Antheridia 65 X 24 ix. Primary appendage 24 X 12 ix. Receptacle, exclusi^■e of foot, about 24 n. Total length to tip of peri- thecium 215 ix, including foot. On the tip of al^domen of Homalota nigrcsccns Fauv. Samarang, Java. But three specimens of this small species have been examined, one of which is immature, while the others are well developed and in good condition. Each bears a single perithecium only, but it is probable that more may be formed. The form is clearly distinguished from all others known by the tooth-like appendage at the tip of its perithe- cium. It is possible that the antheridia may in some cases be appen- diculate, but there is no indication in the types that such is the case. Herpomyces Panesthiae nov. sp. Male indicidual. Receptacle consisting of four superposed cells, the basal much larger and longer and distally inflated; the rest subequal; all, or only the two terminal ones bearing antheridia directly, or single cells from which antheridia arise distally: the antheridia twelve or more, of the usual type ; the group broad below, the necks more or less appressed. Receptacle SO /x, its basal cell 40 X 20 fx. Antheridia 60 fx. Total length to tip of antheridia about 150 /x. FcmaJc indiindual. Hyaline. Primary receptacle minute, the basal cell more elongate, narrow below, the rest subequal, the terminal one nearly circular in outline, except its base, and terminated by a bhuit apiculus; the subbasal cell gi^'ing rise to two secondary receptacles each of which produces a single perithecium on either side of the pri- mary receptacle. Secondary receptacle rather strongly curved, distally broader, the cells vertically elongated, one of them, the largest, extend- 34 THAXTER. ing from the base to the lower basal cell of the perithecium, which is also subtended by a short cell, below which a third cell extends down to the substratum and from which three or four distally pointed suc- cessively shorter cells are separated laterally, externally, and somewhat obliquely, all of them extending to the substratum. The lowest basal cell of the perithecium associated with a general constriction which it wholly occupies, flattened and connected by a more or less narrow isthmus with a broader portion which lies immediately below the ascig- erous cavity and forms, together with the three remaining basal cells which are more or less rounded and sul)equal, a short nearly symmetri- cal rather al^ruptly inflated base, broader than that of the ascigerous venter; the latter is relatively rather short, slightly inflated, the junc- tions of the wall-cells barely indicated, the outline subeven, the struc- ture of the distal portion similar in general to that of H. tricusindatus; the third wall-cell of the anterior row slightly concave, not at all prominent, its margin continuous with that of the cell below, subtend- ing an erect incurved spinous process, a similar process arising from the fourth wall cell of one of the lateral rows, these two processes extending distinctly beyond the erect spine which subtends the blunt short in- curved tip of the perithecium. Perithecium, from basal cells to tip of upper spine, about 92 ju; to tip of lower spines about 102 ^t, greatest width 30 /x. Primary receptacle 26 X 7/^; secondary receptacle about 60 X 28 IX. Total length to tip of longest spine 180 ^i. On the antennae of Pancstkia lobipennis Brunn. Near Peradeniya, Ceylon. This species occurs rarely on the above mentioned host, usually singly, and always produces a single pair of perithecia only. It differs in this respect from H. Paranensis and H. fricuspidafus to which it is nearly allied, as well as in minor details of its secondary receptacle and perithecium. The secondary receptacle is considerably twisted, so that it is almost impossible to see its broad face, as in the two species just mentioned. Synandromyces Javanus nov. sp. Pale straw-yellow throughout. Basal cell of the receptacle erect, narrow, almost completely surrounded by the subbasal cell and the stalk-cell of the appendage, which are nearly ecjual and lie almost symmetrically on either side of it, extending nearly to its base, meeting for a short distance above it. Body of the antheridiiun consisting of NEW INDO-MALAYAN LABOULBENIALES. 35 two superposed pairs of rounded cells separated by indentations; the upper smaller, each bearing a single antheridium distally. Stalk- cell of the perithecium short and very slender, the very broad base of the perithecium arising from it laterally and asymmetrically, the body short and stout, much inflated, asymmetrical when viewed sidewise, tapering abruptly distally; the tip itself not well distinguished, broad; the apex rounded, slightly sulcate. Perithecium 120-135 X 40-50 fx. Receptacle, including stalk-cell, of appendage, 40 X 32 /x, free portion of antheridial appendage, to tips of antheridial cells, 35^0 X 22-25 (jl. Total length to tip of perithecium 155-175 /x. At the base of the posterior legs of a cryptophagid beetle, belonging to the section Telmatophilini. No. 2147, Samarang, Java. This species differs from others known to me in having only two antheridia. A second and very distinct species was found on the same host, having a much reduced appendage and a much longer black perithecial stalk-cell. The single specimen, however, is too broken for description. Arthrorhynchus Nycteribeae (Peyr). Specimens of Pcnicillidia Jenynsi Wstw. collected at Peradeniya, Ceylon, and infested by this species have been very kindly communi- cated to me by Dr. Hugh Scott. The individuals appear to corre- spond in all essential respects to European material and differ only in their somewhat larger size, some of them having a total length of about a millimeter, and paler color. Stigmatomyces Stilici nov. sp. Basal cell of the receptacle about twice as long as the subbasal, narrower below and more or less deeply suffused with blackish brown; distally hyaline and broader than the base of the subbasal cell from which it is thus abruptly distinguished; subbasal cell faintly suffused with yellowish brown, about as broad as long, its basal portion thin- walled and distinctly differentiated, five sided, distally pointed and obliquely separated from the stalk-cell of the appendage on one side and that of the perithecium on the other. Stalk-cell of the appendage subtriangular, externally convex, the basal septum oblique and practi- cally continuous with the short inner margin, which is in contact with 36 THAXTER. the base of the perithecial stalk-cell; basal cell somewhat darker, separated by a slight constriction, distally slightly protruding between an inner and an outer larger cell, each bearing two antheridia; the outer externally convex, and followed by a similar external cell, also bearing two antheridia, while above it three antheridia follow in a vertical series, the uppermost subtended by the usual spinous process; all the antheridia relatively long, with long appressed necks directed obliquely upward in a coherent group. Stalk-cell of the perithecium well developed, hyaline, broader distally, its distal septum horizontal below the flattened group of basal cells; the body of the perithecium becoming more or less well distinguished into a broader somewhat inflated venter, a neck-portion subtendetl l)y a slight elevation, and a somewhat shorter tip; the whole usually straight, tapering gradually, and ending in a bluntly rounded apex subtended on either side by minute papillae. Perithecia 125-195 X 30-40 ix stalk-cell 40-60 X 20-23 ju. Appendage proper, to tip of antheridia, 45-50 /x; its stalk- cell 15-20 /x. Total length to tip of perithecium 200-310 /i. Spores about 35 X 4 ju. On the abdomen and elytra of Stilicus Ceylonensis Kr., No. 2098, Peradeniya, Ceylon; No. 1826 (Types) Borneo. This species belongs to the type separated under the name Zcugan- dromyccs in my paper on Argentine Laboulbeniales. A comparison of abundant material in good condition leads me to believe, however, that the latter name should not be retained, and the Argentine species on Scopacus lacvis should be changed to Stigmatomyces australis nov. comb. Cryptandromyces Javanus nov. sp. Hyaline. Basal and subbasal cell of the receptacle somewhat obliquely and asymmetrically associated, subequal, nearly twice as long as broad. Stalk-cell of the appendage lying parallel to and closely united with the subbasal cell of the receptacle, the distal end of which is also in contact with it; the basal and subbasal cells of the appendage subequal; or the former larger, bearing distally two erect series of three or four antheridial cells separated by slightly oblique septa, at first somewhat coherent and evanescent after separation. Stalk- cell of the perithecium terminal in relation to the subbasal cell of the receptacle, its base of the same width anrl somewhat compressed or irregular, its distal end broader and obliquely related to the outer basal cell of the perithecium above, the basal cells of which are well NEW INDO-MALAYAN LABOULBENIALES. 37 developed and clearly defined; the region somewhat narrower than the base of the venter from which it is well distinguished ; the body of the perithecium rather short and stout, slightly and subsymmetrically inflated; the tip hardly distinguished, short, blunt-pointed, slightly asymmetrical. Perithecia 60-70 X 23-25 m, the stalk 20^0 n. Appendage including stalk-cell 30-38 /x. x\ntheridial series about 28 /i. Receptacle 35-40 ii. Total length to tip of perithecium 100- 120 m- On a small mahogany brown scydmaenid, No. 1419: Java (Rouyer). The antheridia in this species are somewhat more persistent than in the type, although they seem to disappear as soon as they have functioned. In the present instance two groups are present in at least one of the specimens, which seem to be coherent and when sepa- rated each series closely resembles the antheridial branchlet of the Rhadiuomyccs-type of the genus Corethroinyccs; although it seems doubtful that there is any close relationship between the two. The eflPerent necks are short or obsolete. Cryptandromyces subgaleatus nov. sp. Hyaline. Basal and subbasal cells not differing greatly in size, somewhat variably related, the latter united to the stalk-cell of the appendage, which is similar. Appendage consisting of a rather long and slender axis curved toward the perithecium, near the base of which may arise an antheridial branchlet of three seriate antheridial cells; a similar branchlet sometimes terminating the main axis, or arising from the stalk-cell of the appendage. Stalk-cell of the appendage relatively long and slender, al)ruptly distinguished from the base of the perithecium, the cells of which are clearly defined and subequal; the region abruptly broader than the stalk-cell, and somewhat nar- rower than the venter above; which is slightly inflated, and tapers distally to the relatively very broad tip, which is of about the same diameter throughout; the apex slightly broader and subgaleate, slightly oblique inwardly, with a small projection. Perithecium 86 X 20 Mj including basal cell region (10 /x), the stalk-cell 25 X 8 m- Ap- pendage to 215 X 8 M, the antheridial branchlet 30 /x. Receptacle about 20 X 10 /x- Total length to tip of perithecium 125 /x. On the elytra of a small beetle near Scydmacnus. No. 2145, Samarang, Java. Only one of the four si^ecimens examined is mature enough to have 38 THAXTER. profhiced spores. The species seems well distinguished by its broad siibgaleate apex, slender and well developed perithecial stalk-cell, and its long appendage. Corethromyces Medonis nov. sp. Receptacle geniculate between the basal and subbasal cells, the former squarish or slightly broader than long, quite hyaline; the latter opaque, broader distally, sometimes twice as long as broad, indistin- guishable from the main appendage with which it is continuous. The primary appendage hyaline along the inner margin, otherwise nearly opaque, the subdistal and much smaller distal cells giving rise to a group of hyaline branches which are rather short, once or twice branched, directed inward; some of the branchlets consisting of seriate antheridia of the normal type. Perithecium hyaline, becom- ing yellowish; the hyaline stalk-cell well developed, narrower below, the outline from its base to the apex of the perithecium almost sym- metrically fusiform; body of the perithecium slightly inflated, the outline of the margins somewhat undulate; the tip not distinguished, often slightly bent outward, the apex bluntly rounded. Perithecium 78-100 X 20-28 m; the stalk-cell 20-39 X 12-16 /x. Primary appen- dage 27-31 X 9 /x, its longest branches 50 /x. Total length to tip of perithecium 125-175 /x. On the inferior abdomen of Mcdo7i curtus}\.r.; No. 2074, Samararig, Java. On Mcdon Binnanus Fauv., No. 2368, Borneo. A small species more nearly allied to C. purpurasccns than to other described species, and belonging to the Cryptohium-mhdihitmg group. Corethromyces decipiens nov. sp. Receptacle becoming almost wholly opaque, a small hyaline point just above the foot, the basal cell small, geniculate, prolonged poste- riorly and distally to form an opaque, free, somewhat divergent, spur-like upgrowth, spoon-shaped or of nearly equal diameter through- out, blunt-tipped; its base united throughout to the long flat distally hyaline otherwise deeply suffused subbasal cell; its remaining free portion lying beside the hyaline appendage, the branches of which reach some distance beyond its extremity. Appendage arising in the angle between the base of the perithecium and the black spur-like NEW INDO-MALAYAN LABOULBENIALES. 39 process from the receptacle, quite hyaline; consisting of a basal cell and two or three smaller terminal cells from which a group of branches arises reaching to about the middle of the perithecium, once or t .vice branched, the lower branchlets antheridial, others sterile, rather slen- der, with curved or slightly recurved tips. Perithecium straight or rather strongly curved, rather slender and long, the stalk-cell about as large as the basal cell of the appendage beside it, the basal cells rela- ti\'ely large, the body somewhat crooked and ending in a snout-like tip, the apex broad, rounded, the outer lip-cell somewhat more promi- nent. Perithecia 75-100 X 16 fi, the stalk-cell, 12 X 8 ju. Recep- tacle including foot 25-30 yu, its spur-like process 35-40 ^t. x\ppendage including branches 55-65 /jl. Total length to tip of perithecium 120- 150 M- On Mcdon hirmanus Fauv., No. 2119, Borneo (Type) on bristles near tip of abdomen on upper side. On M. ochraceus Boisd., Borneo, No. 2369. On M. curt us Kr., No. 2087, Samarang, Java. The specimens growing on hairs of the host are somewhat more slender and more strongly curved than those which grow on the body (No. 2369), and the spur-like process is more slender. The latter is similar to those which are developed on many of the species parasitic on Stilici to which the present form is most nearly allied. Corethromyces Thinocharinus nov. sp. Basal cell deeply suffused above and along its posterior margin, where the suffusion is continuous with that of the foot, above which it forms elsewhere a narrow hyaline contrasting arc, the cell extending upward and outward on the posterior side to form a free blackened \arialjly developed spur-like somewhat divergent prolongation, its inner margin hyaline; sometimes rather short and straight, or longer and subsigmoid, extending beyond the longest branches of the append- age. Subbasal cell usually more or less suffused, translucent, small and angular, extending down beside the basal cell nearly to the foot, from which it is separated by the hyaline area of the former : the short stalk-cell of the perithecium rising from it distally and anteriorly; the appendage subterminally on the opposite side. Appendage hya- line, consisting of usually two or three larger superposed cells, from the upper of which arise several branches, some of their branchlets producing seriate antheridia. Perithecia relatively large, hyaline or faintly yellowish, the basal cells distinct and about as large as the 40 THAXTER. short stalk-cell; the main body nearly straight or slightly curved, sometimes subsigmoid, relatively rather long, tapering slightly; the tip rather broad, variably modified, often snout-like and irregularly bent, the posterior lip-cell usually prominent or forming a short but well defined concolorous projection; the apex flat or bluntly rounded. Perithecium 55-72 X 12-13 ^t. Appendage with branches 35-50 ^u. Receptacle 9 X 7 ju, the spine-like process 18-55 X 3 /x. Total length to tip of perithecium 75-95 /U. On the inferior surface of the abdomen, near the tip of Thinocharis pygmaea Kr. Samarang, Java, No. 2084. This species corresponds closely to the numerous forms which com- prise the section of the genus parasitic on Stilici. The basal cell is hardly distinguishable, from its small size and the suft'usion which obscures it, while its relations are further confused from the displace- ment downward of the subbasal cell. It is most nearly related to C. decipiens. Corethromyces orientalis nov. sp. Basal cell of the receptacle relatively large, wholly blackened and not differentiated from the foot; continued upward to form a blackened prolongation, which is closely united to the subl)asal cells of the receptacle and of the appendage, bending abruptly inward, being free and more slender above the latter; the free portion lying parallel to the appendage, the main axis of which it may equal in length, subbasal cell of the receptacle hyaline, its outer margin concave, below obliquely, or almost \ertically, separated from the l)asal cell and its extension, as well as aliove from the stalk-cell of the perithecium which arises from it externally. ]Main axis of the appendage consist- ing of a basal cell which becomes displaced so that its base seems to lie against the blackened prolongation of the basal cell of the receptacle, the appendage thus becoming turned so as to lie across the stalk-cell of the perithecium, and of two to three smaller terminal cells all of which may bear one or more very elongate attenuated branches or branchlets, the whole quite hyaline. Peritheciiun and its stalk-cell hyaline, or the former faintly reddish purple, the whole subfalcate or even subsigmoid; the stalk-cell well developed, several times longer than broad, its diameter about the same throughout, as broad as or slightly broader than the base of the perithecium, which is usually strongly curved, hardly inflated l)elow, tapering slightly and gradually distally; the tip not at all distinguished; the -apex broad, bluntly NEW INDO-MALAYAN LABOULBENIALES. 41 rounded. Perithecium 66 X 16 y., the stalk-cell 40^5 X 13 m- Pri- mary appendage 25 /z; the longer branches 250-275 m- Receptacle about 28 fx, the free part of its prolongation 24 X 4 /i. Total length to tip of perithecium 135-145 ji. On the abdomen and prothorax of Stilicus Ceylonensis Kr. Buiten- zorg, Thompson (Type) No. 2488; No. 2077, Samarang. Allied to C. Stilici and differing especially in its appendage and re- ceptacle. Corethromyces appendiculatus nov. sp. Receptacle consisting of two very small cells, the basal hyaline, or but slightly suffused, the subbasal more or less deeply involved by a suffusion which is continuous with that of the main axis of the append- age. The latter dark blackish olivaceous, deeper externally, the dark septa clearly visible, the cells four or five in number, the three lower larger, the basal closely united to the stalk-cell of the perithecium, the subbasal often strongly convex, the distal bearing a terminal and one or two subterminal branches, the latter from the inner side; the branches sometimes sparingly branched, short, hyaline, except the basal cells of the terminal one which is a continuation of the main axis. Perithecia relatively short and stout, the stalk-cell but slightly larger than the basal cell of the appendage, to which it is almost wholly united; often separated from the perithecium which may diverge from it at an angle, crossing the axis of the appendage, by a distinct external constriction; the basal cell-region hyaline, not distinguished from the portion above it which becomes gradually broader and more deeply suffused with olive brown from below upward; the middle third broadest, the outer margin rather strongly convex; the tip short and broad, not at all distinguished, its inner margin deeply suffused, the suffusion extending upward into a rather slender curved appendage formed from an outgrowth of one of the lip-cells; another of the lip- cells prolonged to form a suffused projection about half as long, straight, with broad base and rounded tip, the other two lip-cells ending in dissimilar, small and unequal, asymmetrically placed hyaline protrusions. Perithecia 50-60 X 16 yu; the stalk-cell 8-10 X 5 ijl; the free part of the longest terminal appendage about 6 X 2 /x. Appendage, main axis 28-45 X 7 /.i, the hyaline branches about 28 /i. Total length to tip of perithecium 70-85 /x. Receptacle 10 X 5 /U- On the elytra of a silphid l)eetle near Anaspis. No. 2079, Samarang, Java. 42 THAXTER. This peculiar species is most nearly ailed to C. obtutms which it resembles in general habit and coloring. It is clearly distinguished by the peculiar conformation at the apex of its peritheciura. Stichomyces Pterogenii nov. sp. Foot relatively large; receptacle sulKlecvunbent, broadest at the base when seen sidewise; colorless or faintly yellowish, continuous with the primary appendage from which it is in no way distinguished, the whole consisting of about six superposed cells; the four lower somewhat larger, rather irregular in size, somewhat broader than long; the second, third and fourth giving rise laterally either to single perithecia, or to single short branches; the subbasal cell more often producing a short antheridial branch, and the cell above it an up- curved perithecium; the terminal cells of the axis giving rise on one or both sides to rather slender, elongate, tapering branches, which are simple or once l)ranched near the base, turned upward parallel to the perithecium. Perithecium rather slender and long, borne laterally on a short stalk-cell which is bent abruptly upward; the venter somewhat broader, tapering slightly to the hardly differentiated tip, which is short, often slightly bent, tapering to a blunt point. Peri- thecia 40-50 X 9-12 jjL. Spores 18-20 X 2 /x in perithecium. Main axis 30-35 X 10- 12 jj. at base. Longest branches 175 ijl. On the elytra of Pterogenius Nietneri; No. 2109, Peradeniya, Ceylon. This rather nondescript form is somewhat doubtfully referred ta the genus SticJiomyccs, of which it may be an aberrant and reduced type. The tendency to produce branches and perithecia on either side of the main axis, and the common occurrence of a branch below the perithecium from the subbasal cell would make it difficult to in- clude it in Corethromyces which it resembles in some respects, nor does it seem possible to include it in the genus Chactomyces, although the position of this last genus is somewhat uncertain. The perithecium,. which is normally single, is more often produced by the third cell. The axis lies nearly parallel to the surface of the host, and the peri- thecium and long slender branches, at the bases of which short antheri- dial branchlets may arise, are abruptly bent upward. Stichomyces Cybocephali nov. sp. Very constantly coherent in pairs; straight, nearly colorless, the foot relatively very large, the axis consisting of from six to eight NEW INDO-MALAYAN LABOULBENIALES. ' 43 superposed cells, which are broader than long and rather irregular in size; the two or three distal ones more often without branches. Perithecia or appendages arising from the second, third or fourth cells, laterally and sometimes from opposite sides even of the same cell. The branches scanty, often lacking or only one or two, curved upward, usually simple, few-celled, with a terminal antheridium. Perithecia suberect, or sometimes curved outward, borne on a short stalk-cell, concolorous, relatively long and slender, the body nearly symmetrical; the tip not clearly distinguished, except by a minute external papilla, short, tapering slightly to a blunt or truncate apex. Perithecium 30-35 X 7-8 n. Spores 25 X 2 /i. Axis 20-27 X 7 m- Branches 15-18 n. Foot 12 X 5.5 m- On the elytra of Cyhocephalus sp. No. 2106, Peradeniya, Ceylon. This form like the preceding species is provisionally referred to Stichomyces, but is of uncertain position. What appear to be antheri- dia have been seen in several specimens borne singly and terminally, usually on the branch which subtends the perithecium. Laboulbenia helicophora nov. sp. Basal and subbasal cells of the receptacle uniform dirty yellow brown, the subbasal somewhat longer and broader ; the distal portion swollen and becoming deeply suffused with blackish in the region near cell III which, like cell four, is relatively very broad, the two forming a prominent bulging in this region. Insertion-cell immediately over the septum between cells IV and V, separated from the perithecium by a portion of the upper surface of the latter; the group of appendages compact, more or less uniformly dull yellowish brown, the outer usually' simple, rigid, slightly divergent, somewhat tapering, rather short; the basal cell of the inner bearing once furcate branches on either side similar to the outer appendage. Perithecium more than half free, broader at the base, and bent toward the appendages; al- though the inner margin is nearly straight up to the tip which, by a sudden constriction abruptly distinguishes the peculiar lip-cells, the inner of Avhich forms an appendage deeply blackened, except along its inner side, which, projecting outward, curls upward and inward in a short helix; the outer lip-cell also forming a conspicuous black, slightly shorter, stouter appendage, its bluntly rounded tip slightly upturned; the pore hyaline and associated with a small black process from the base of the outer appendage, which curves over it. Perithe- 44 THAXTEK. cia 175 X 40-45 m; the two lip-appendages spreading 64 ^i. Ap- pendages, maximum length, 150 /x. Receptacle 275-315 X 78-85 m- Total length to tip of perithecium 390-425 /x. At the base of the anterior legs of PericaUus sp., No. 1408, Java (Rouyer). A very striking species, most nearly allied to L. Javana, from which it differs in the general form of its receptacle, which recalls that of L. Texana, by its curled apical appendage, and much greater size, as well as in other points. Laboulbenia manubriolata nov. sp. Variable, dirty straw-yellow throughout, or with brown shades; the subbasal, or both the basal and subbasal cells becoming suffused with blackish brown, in the type, and coarsely tuberculate. Basal portion of the receptacle rather slender, curved, the curvature including both the base of the subbasal cell and the whole basal cell, the latter one half to one third as long as the former; cells III and VI usually somewhat elongated, parallel, of about the same width; cell II longer, usually slightly prominent distally, separated by a horizontal septum from cell IV, which is often rather long and externally concave, bulging slightly below the insertion-cell. Appendages hyaline or becoming brownish, the basal cell of the outer relatively large and bearing a distal outcurved prolongation, the rest of the appendage once or twice branched; the branches somewhat tapering, rather slender, soon broken; basal cell of the inner appendage smaller producing a short branch on either side, usually furcate. Perithecium more than one half to one third free, becoming tinged with brownish, rather long and narrow, the tip relatively broad and slightly outcurved, blackened below the hyaline pore. Perithecia 80-135 X 20-26 ijl. Appendages unbroken, longest 150 pt.. Receptacle, to insertion-cell, 120-300 X 30^0 fjL. Total length to tip of perithecium 160-400 m- On various parts of a small Carabid allied to Tachys. No. 2081d (Type) on elytra, Samarang, Java. No. 2093, Peradeniya, Ceylon. This species is most nearly allied to L. Tachyis, and L. maritima and is distinguished by the handle-like protrusion of the basal cell of the outer appendage. The species varies very greatly in size and color. The form having the basal and subbasal cells blackened and coarsely tuberculate, is taken as the type. The last mentioned character is evidently not a result of age, since it appears in a few very young speci- mens in which the perithecia have not begun to form. NEW INDO-MALAYAN LABOULBENIALES. 45 Laboulbenia Grylli nov. sp. Basal cell of the receptacle hyaline without a blackened foot, twice as long as broad, its base abruptly rounded and penetrating the host by a simple rhizoidal filament, the subbasal cell abruptly slightly broader above it, greatly elongated, hyaline, slightly broader distally, sometimes narrower in the mid-region, the terminal portion above it slightly and abruptly narrower; cells III-V yellowish, granular-punc- tate, cell II somewhat larger than the other two combined, cell V small, triangular, hardly reaching to cell III, and usually proliferous to form a small branched appendage similar to the branches of the main appendage. Main appendage arising from a single small basal cell seated on a well defined black insertion, its cells mostly rather short and stout, more or less constricted at the septa, the basal cell bearing two branches, erect, several times successively branched, the branch- lets divergent often at almost a right angle; both the primary and secondary appendages hyaline or pale yellowish, bearing here and there a few short antheridia of the usual type. Basal cells of the perithe- cium small but clearly defined above cell VI, which extends downward in a point beside the end of cell II. Perithecium very large, almost wholly free, paler straw-colored, with a brownish tinge below; a symmetrically inflated venter clearly distinguished from a greatly elongated, but slightly tapering, nearly hyaline neck-portion, above which a long slender tip is abruptly distinguished, which is suffused with pale reddish brown, becoming blackened distally, except the tip of the outer lip-cell, which subtends the external pore. Perithecia 500-750 X 100-120 /x; the tip 70-75 X 18 m; the neck 50 /x at base and 32 fx at distal end. Spores 45 X 7 /x. Receptacle, basal cell about 160 X 70 m; the subbasal 550-780 fx by 60 /x; 70-80 /x at distal end. Total length to tip of perithecium 1300-1700 fi. On the inferior surface of the abdomen, near the tip, of Gri/llus alhifrons Sauss. Samarang, Java. This fine species is very conspicuous from its great size, being one of the largest of all the Laboulbeniales. It bears a very remote resem- blance to L. {Ccraiomyces) Dahlii, which also penetrates the host by means of a well defined rhizoid which, in the present species, appears to be always simple. Two infected hosts were sent me by Mr. Jacobson and abundant material has been examined. In every individual the cell-arrangement is exactly that of LabouUienia, except that the main appendage arises from a single basal cell. This cell. 46 THAXTER. however, the basal wall of which is blackened in the usual fashion, should probably be regarded as the equivalent of the ordinary inser- tion-cell of other species, and always produces distally an inner and outer branch, which would thus be homologized with the inner and outer appendages of the usual type, such as is present in the very closely allied species described below. That this is also the condition in the type of Ccraiomyccs, {C. Dahlii), above referred to, seems also evident, and the peculiar black, bottle-shaped cell in this species, which bears terminally an inner and an outer branch, should be regarded as a modified insertion-cell. The secondary appendage, which arises from the proliferation of cell V, and may sometimes be partly lateral in position, corresponds exactly to the similarly developed appendage seen in L. proUfcraufi. Laboulbenia subulata nov. sp. Receptacle shorter than the perithecium, the basal cell somewhat shorter than the subbasal, curved and narrower below, rather abruptly enlarged just before penetrating the host by means of a usually simple rhizoid, which is abruptly bulbous just below the integument; the sub- basal cell stout, broader distally, hyaline or pale straw colored, concol- orous with the basal cell; cells III-IV small, nearly equal, separated externally by an indentation, deeper yellow, granular-punctate; cell V triangular, reaching to the apex of cell III. Insertion-cell normal, reddish brown, rather broad, giving rise to an outer and inner append- age in the normal fashion, the basal cells of which are similar; both appendages branching two or three times successively, hyaline, rather short; the ultimate branchlets tapering to blunt extremities; the cells not separated by conspicuous constrictions. Perithecium subulate, elongate, the venter relati^'ely small with a more or less clearly defined narrower basal portion, slightly inflatefl, but hardly distinguished distally from the broad tapering neck-portion; the tip distinctly bent, not distinguished from the latter, except for the clearly visible terminations of the wall-cells, and a slightly purplish-brown suffusion, deeper at the apex, which is slightly asymmetrical, the inner lip-cells more prominent, with conspicuous round lip-valves, the pore lateral and external. Perithecia 500-700 /x; the venter 120 X 58 /x; the tip 48 X 20 fjL. Spores 28 X 4 /x. Receptacle 275-400 X 58 m- On the inferior surface of the abdomen of GnjUus Burdigalcnsis- Cerisyi Ser\'. Immature specimens of apparently the same species on GryJlu.s mitraius Burm. NEW INDO-MALAYAN LABOULBENIALES. 47 Although numerous immature specimens of this form are available for examination, few mature individuals have been seen, and it is not improbable that the measurements will vary considerably from those given when abundant material is available. The type corresponds to that of Laboulhenia in every detail, and though closely allied to L. Grylli seems abundantly distinct. Misgomyces ornatus nov. sp. Relatively short and stout, pale amber-brown, the posterior margin more deeply suffused. Receptacle not distinguished from the primary appendage; the former consisting of four superposed cells, the basal longer and narrower, distally geniculate, the rest broader than long, the uppermost larger, five sided, bearing the stalk-cell of the peri- thecium on one side and continuous with the primary appendage on the other. Primary appendage consisting of normally three super- posed cells, the basal larger and associated with a small cell which is separated from it on the inner side and lies between it and the stalk- cell of the perithecium to which it is firmly united ; the two upper cells flattened, sub-hemispherical, bearing on the inner side branches more or less copiously branched to form a usually short dense tuft; the ulti- mate branchlets usually terminated by paired antheridia, which are relatively long, without well distinguished necks; the distal cell of the appendage also bearing at first a short outcurved evanescent append- age, continvious with the deeply suffused posterior marginal wall. Perithecia borne on a short, but well defined, stalk-cell; its basal cells clearly developed; the lowest, which lies at the right, much larger and somewhat broader than the portion of the base above it; the body rather short and stout, the outer margin straight, the inner convex, distally broad and abruptly distinguished from the tip, which is subtended on the inner side by a conspicuous short brown blunt tooth-like process; the pore minute and lateral on the inner side, a subterminal outer wall-cell assuming a terminal position and extend- ing upward and outward to form a blunt-tipped somewhat divergent concolorous appendage, which becomes relatively long and slightly sigmoid. Perithecia, including base, without terminal appendage, SO-90 X 23-30 Ai; its terminal appendage 25-75 X 4.5 ^i; the stalk cell 14 X 11 yu. Spores 35-40 X 2.5 m- Total length to tip of pri- mary appendage 65 X 25 /j., the tuft of branches 35-45 /x. Antheridia 32 X 2 M. 48 THAXTER. Near the middle of the outer margin of the right elytron of a small carabid allied to Taclu/.s. Peradeniya, Ceylon, No. 2093f : Samarang, Java, No. 2081 f. This species is referred to Misgomyccs not without hesitation, since its receptacle and perithecium appear to be generally determi- nate. Although its perithecium is \ery peculiarly modified, and quite unlike that of any species of Misgomyccs known at present, its receptacle and appendage correspond sufficiently well to justify this provisional reference. The same form is known from Borneo and the Philippines and is always found growing appressed in the position above indicated. The apex of the perithecium with its peculiar appendage, which varies considerably in length according to the age of the individual, suggests a pointed monk's hood. Misgomyces Lispini nov. sp. Wholly hyaline, usually straight and erect. Receptacle indetermi- nate, Ijecoming broader distally; consisting normally of a series of about ten to fifteen cells, much flattened, except the subtriangidar basal cell; the uppermost bearing the perithecium on one side and the primary appendage portion on the other; the latter consisting of two, or usually three superposed cells closely united to the base of the perithecium, the two upper, or only the uppermost, separated from it by a long narrow marginal cell which extends nearly to its tip, and appears to be originally derived from the lowest member of the series; the terminal cell bearing one, rarely two 'appendages' from its nearly horizontal or outwardly ol^lique upper surface, usually distinguished by a constricted dark septum, and often more or less copiously branched. Perithecium, including the marginal cell which assumes the position of its inner wall, often nearly symmetrical, four fifths to two thirds free on its inner side, subconical, tapering rather rapidly and evenly from a broad base to a small truncate apex, the tip but slightly distinguished and only on the inner side just above the marginal cell. Perithecia 58-62 X 26-^32 fx. Total length to tip of primary appendage portion 78-105 X 27-43 ^l. The branched ap- pendage 50-75 iJL. Total length to tip of perithecium 110-140 ix. On various parts of Lispinus imprcssicoUis iNIotsch., No. 2076, Samarang, Java. No. 2101, Peradeniya, Ceylon. Also known from Borneo. This species occurs usually in small numbers on its host mid vai'ies NEW INDO-MALAYAN LABOULBENIALES. 49 from a very stout to a somewhat slender form. The antheridia are sometimes produced in considerable numbers, especially when the perithecium has been destroyed or has failed to develop. In such cases they arise in tufts, and are usually paired at the ends of branch- lets as in M. ornatus, and as in this instance, are relatively long, with- out a clearly defined neck-portion. Misgomyces Clivinae nov. sp. Long and slender, nearly hyaline, or faintly tinged Avith brownish yellow. Receptacle consisting of about ten superposed cells, the basal considerably elongated, the next two or three above it squarish, the remainder more elongate, all filled with coarse evenly granular proto- plasm. The axis of the perithecium coincident with that of the receptacle, the primary appendage-portion erect, mostly free, consist- ing of two or three superposed cells somewhat elongated, the basal one connected with the perithecium by a small cell which is closely united to it, the terminal one bearing several evanescent branchlets. Peri- thecium straight, symmetrical, slightly inflated, the tip tapering to the- blunt symmetrical apex. Pei'ithecia 112X42//. Total length to- tip of perithecium .580 /u. Greatest width of receptacle 32 ji. Pri- mary appendage portion 70 X 12 ju. On the margin of the elytra of Clicina sp., No. 1415, Java (RouyerJ. Six specimens of this species have been examined of which three, only, are matured. In all of these, the branches of the main append- age have disappeared. The species is, hoAvever, very clearly dis- tinguished by its long and slender habit, and could not be confused with any known form. Rhachomyces orientalis nov. sp. Receptacle determinate, consisting of four superposed cells, followed by two cells more or less obliquely associated side by side; the basal cell narrow, rather short, distally broader and suffused with blackish brown: the second, subbasal, more deeply suffused, broader than long, subtriangular, distally pointed or obliquely separated from the third cell, and bearing a single simple straight appressed short usually four-celled primary appendage; the basal cell of which is broader, deeply suffused and distinguished above by a broad V>lack slightly 50 THAXTER. constricted septum, the rest externally blackened, their septa some- what oblique: the third cell obliquely separated from the fourth, hyaline or with a basal blackish suffusion, much larger, bearing just al)0\'e the l)ase of the primary appendage a large blackened obliquely separated cell, the inner face of which is almost wholly in contact with the fourth cell of the receptacle, and which bears distally two rigid erect appressed branches, one of which is usually longer, often reaching nearly to the tip of the perithecium ; while the other is much shorter, and bears a terminal antheridium; the basal cells of both nearly opaque, the rest deeply suffused externally, hyaline along the inner margin, the septa oblique: the fourth cell similar to the third, somewhat larger, distally pointed and ol)liquely separated from two cells lying right and left immediately above it; one of which is deeply suffused and gives rise distally on its inner side to the short almost wholly inclosed stalk-cell of the perithecium, and externally to a small subhyaline cell from which arise two appendages similar to that borne by the third cell, each of which is characterized by a deeply suffused basal cell bearing paired erect stiff branches; the second of the two paired cells above cell four, bearing a simple short appendage with a blackened base which usually terminates in an antheridium, and also bears laterally a secondary appendage similar to that arising from cell III, its basal cell deeply suffused and l)earing two erect branches. Perithecium only partly surrounded by the l)ranches of the append- ages, which may extend nearly to its tip; uniform pale straw-colored, subsessile, erect, straight on one side, and somewhat convex on the other, the tip more or less clearly distinguished, usually slightly bent, broad, short, the apex truncate, often oblique and slightly sulcate. Perithecium 100-116 X 25 fi. Receptacle (50-80 X 22 ju- Longer ap- pendages 110 M- Total length to tip of perithecium 160-200 ju. On the trochanters of a small Carabid near Tachys, No. 2081, Samarang, Java. I have taken as the type of this species a stouter, darker, more compact form which inhabits the trochanter of this host and has also been received from the Philippines. An immature specimen which seems to be the same is also among the Ceylon material. In addition to this type-form, one occurs on the elytra in both the Javan and Philippine material, which is somewhat more elongate, the receptacle and the subtending cells of the appendages nearly hyaline, and the perithecium furnished with a short tooth like projection near the mid- dle, which has not been noticed in any of the specimens inhabiting the trochanter. Still another variety has been obtained from Borneo NEW INDO-MALAYAN LABOULBENIALES. 51 and the Philippines, in which the base and neck cell of the perithecium are greatly elongated, the perithecium angular, and the pale receptacle longer. The cell-arrangement in all these forms seems, however, to be identical; although that last mentioned is slightly twisted so that the appendages do not at first appear to arise in the same order, that is three sets superposed on one side and one placed at the base of the perithecium on the other. Although quite different in general appear- ance I am of opinion that they should not ba separated specifically. Proceedings of the American Academy of Arts and Sciences. Vol. 51. No. 2. — September, 1915. CONTRIBUTIONS FROM THE JEFFERSON PHYSICAL LABORATORY, HARVARD UNIVERSITY. POLYMORPHIC TRANSFORMATIONS OF SOLIDS UNDER PRESSURE . By p. W, Bridgman. Investigations on Light and Heat made and published with aid from the RuMFORD Fund. CONTRIBUTIONS FROM THE JEFFERSON PHYSICAL LABORATORY, HARVARD UNIVERSITY. POLYMORPHIC TRANSFORMATIONS OF SOLIDS UNDER PRESSURE. By P. W. Bridgman. Received, May 1, 1915. CONTENTS. Page. Introduction 55 Experimental Procedure 57 The Effect of Impurities on the Transition Lines 59 The Equation of the Transition Lines 63 Data 68 Potassium Sulfocyanide 68 Ammonium Sulfocyanide 72 Potassium Sulfide 76 Potassium Chlorate 78 Potassium Nitrite 81 Carbon Trichloride 84 Carbon Tetrabromide 90 Silver Iodide 97 Mercuric Iodide 104 Phenol Ill Urethan 118 Introduction . In this paper data are given sufficient for a thermodynamic specifi- cation of a number of polymorphic transitions between soHds. Besides the transition quantities themselves (change of volume, latent heat, slope of the transition line) it has been possible in a good many cases to give at least a rough approximation to the difference of compressi- bility, thermal expansion, and specific heat of the reacting phases. The field is a comparatively unworked one. Hitherto investigations on polymorphic changes have been concerned for the most part with the temperature of transition at atmospheric pressure. There are not many measurements of latent heat or change of volume even at atmospheric pressure. Still fewer measurements of the effect of pres- sure have been made; many of these are due to Tammann, whose pressure range was considerably lower than that of this paper. The 56 BRIDGMAN. range of this paper is from 0° to 200°, and from 1 to 12000 kgm. The measurements to be given here must be regarded as only the beginning of an attack on an immense field. The most important immediate task would seem to be the collection of a large mass of data, so that we may become familiar with the general types of phenomena. About the only discussion that can be attempted is a thermodynamic one, and even from this narrow point of view the measurements are not sufficient to completely determine the behavior of the several phases. We can go only a little way toward the solution of the general problem, which is to predict from the properties of any one phase when to expect new polymorphic phases, and what their properties will be. We may perhaps expect more definite results when methods such as are used here are taken in conjunction with recent X-ray methods for determin- ing crystal structure. Thermodynamically, the transition between two solids is charac- terized by the same elements as determine the melting and the vapori- zation curves, and from this point of view the discussion of previous papers ^ is applicable. As a matter of fact, however, the character of the solid-solid curves may vary much more widely than that of the melting or vaporization curves. For instance, all vaporization or melting curves, absolutely without exception, either rise or fall over their entire length, while solid-solid curves may have either a maxi- mum temperature or pressure. In the previous discussion it was not necessary, therefore, to consider certain special relations between the thermodynamic constants at the maximum points, but now these relations become of importance. A discussion of these is given in the following. Another matter that needs reconsideration is that of the effect of impurities at a point of maximum pressure. Evidently the usual statement of the effect of impurity as causing a displaced temperature of equilibrium will not serve here. In the following are given the slight modifications necessary in the usual discussion to find the pressure shift of the equilibrium line. It has been possible to give a much more thorough investigation of the difference of compressibility, thermal expansion, and specific heat between the several phases than was possible in the case of melting. The reason for this is the much greater relial)ility of the experimental measurements of the difference of compressibility between the separate phases, because, except in those cases where the impurity forms mixed crystals, there is absolutely no rounding of the corners of the volume- 1 P. W. Bridgman, Phys. Rev. 3, 126-203 (1914), and 6, 1-33 (1915). POLYMORPHIC TRANSFORMATIONS OF SOLIDS. 57 pressure isothermals preliminary to a reaction. Furthermore, in the ease of those substances whose phase diagrams contain triple points, it is possible to calculate completely the three differences from the thermodynamic data of the transition curves alone. Among the data which do not enter into the thermodynamics of phase change, but are nevertheless of significance for the mechanism of the transition, may be mentioned the amount of superheating or subcooling a phase will support, the velocity of transition, and the width of the band of inrUfference. To fletermine these data so as to have absolute significance is a matter of great experimental difficulty, since slight impurities and the character of the apparatus produce great changes in the measured effects. No such measurements have been attemptefl in this work. Nevertheless a number of qualitative and comparative measurements have been made incidentally in the course of the measurements, which are of orienting value. Since these measurements are so different in their character from the ther- modynamic measurements which form the bulk of this paper, it has seemed best to reserve most of them for anotlier paper. This paper is the first of several which are to deal with solid transi- tions under pressure. In this first paper T will give the preliminary thermodynamic discussion which will be necessary to all the work on solid transitions, and also data for a number of such transitions. Discussion of the data may be profitably left for later papers, after a greater mass of data has been presented. Experimental Procedure. The procedure was, except for some minor details, like that of two previous papers.^ The most important difference is in the receptacle for holding the substance under investigation. Most of the sul)stances were solid throughout the range of investigation. If the substance was one not dissolved by kerosene it was usually rammed dry into a thin shell made of a piece of steel tubing, open at both ends, the tubing just fitting the inside of the pressure cylinder. The shell was held in a heavy steel form during the filling, so that it should not be bulged by the ramming into place of the material, which was accom- plished with a steel piston and a heavy hammer. Frequently after the shell had been filled, numerous holes were drilled laterally through the walls, to procure as ready access as possible by the kerosene to all 58 BRIDOaUN. ^— A parts of the mass. It might at first thought seem that the steel shell would produce irregularities by tending to hinder the change of volume accompanying the reaction, but no irregularities of this nature were ever found. Sometimes, when the substance could be melted without danger of decomposition, it was melted, instead of hammered, into the shell. Under these conditions the shell was, of course, closed at the bottom end. If the substance were one soluble in kerosene, two slightly different methods were used. By the first method, the salt was formed into a cylinder of suitable dimensions, either by ram- ming into a split mound or by melting into a form, and then sub- merged under mercury in a steel shell, being prevented from rising by a clip at the upper end. Or it might be hammered or melted into a steel cylinder, which was then inverted below the surface of mercury as shown in Figure 1. This method is applicable if the substance melts somewhere within the range of the experi- ment, and was used in all such cases. The air, of course, was exhausted in this case. The chief trouble with this method of the inverted cup is that the cup is very likely to be ruptured by the change of volume during either melting or the change from one solid to another. Substances are strikingly different in their rigidity and the readi- ness with which they rupture the cylinder. It is significant that the shell was never ruptured while determining transitions in the low pressure appa- ratus; this is another bit of evidence showing the greatly increased rigidity, or better, viscosity, pro- duced by high pressures. The shells finally had to be made of hardened chrome nickel steel of the dimensions shown, and even then they were some- times broken. The dimensions of the free space allowed for the mercury are of importance, since the kerosene must not come in contact with the substance in consequence of the changes of volume brought about by the transitions or the compres- sibility. The proportions shown sufficed for all sub- stances investigated here up to 12000 kgm. In other respects the apparatus was of the same design as that previously used, but accidents necessitated the renewal of various parts. Three lower cylinders were ruptured, two by amal- gamation, and one by a violent explosion. This explosion also ruined Figure 1. The modified form of container for those substances which melt within the range of the exper- iments. POLYMORPHIC TRANSFORMATIONS OF SOLIDS. 59 the upper cylinder and the connecting tube. If any one should ever have occasion to do work similar to this again, he should use every pre- caution when working with mercury salts. These decompose slightly in contact with the steel, and deposit free mercury, which gradually works its way through the cylinder, even when it is 5 inches in diam- eter, as was that used here. The flaws which resulted from this action of the mercury were not traced to their true cause until two cylinders had been ruptured. The coils with which pressure was measured were also replaced a number of times. These were fre- quently destroyed when the substance under investigation decomposed at the higher temperatures and the decomposition products found their way to all parts of the apparatus. Fortunately a number of well seasoned coils were in hand, so that new ones could be substituted and measurements immediately begun without waiting for seasoning. The new coils were always calibrated in the way already descril)ed by determining the melting point of mercury at 0°. The materials were obtained in as pure a state as possible from various chemical houses, and used in most cases without further puri- fication except careful drying. However, those organic substances which had a convenient melting point were further purified by cr^'stal- lization in a thermostat at constant temperature. Slight impurity is of very much less importance in the case of solid transitions than it is for melting, because here the transition coordinates are not affected except in those rare cases where mixed crystals are formed. At least one such case was noticed, however. The Effect of Impurities on the Transition Lines. In an earlier paper ^ a deduction was given of the temperature depression of a transition point which was applicable at any pressure, since it did not involve the assumption of a vapor phase, as such deductions usually do. To find how a whole transition line is affected by impurity, one has only to apply the formula given there to every point in order to obtain the equation of the modified line in terms of a displacement at every point parallel to the temperature axis. But this process evidently fails if the transition line itself is parallel to the temperature axis. Such transition lines do not occur for melting, but they are comparatively common for solid transitions, and it becomes important to find another expression for the shift of the equilibrium line. 60 BRIDGMAN. The formula is readily derived by the thermodynamic potential. We first deri\e again l)y this method the ordinary expression for the temperature depression, since it is very easily done, and may perhaps make matters a little clearer. Let us suppose that pure (1) and (2) are in equilibrium at p, t on the transition line. (See Fig. 2.) Figure 2. Showing the displacement of pressure and temperature pro- duced by impurities. »■ Then we know that at this point the thermodynamic potentials are equal, or Zi = Z2. Let us now subject the phase (2) to a tempera- ture increment (At) only, but the phase (1) to a simultaneous tempera- ture increment (At) and pressure increment {^p'). We demand that At and Ap' be so chosen that (1) and (2) are in equilibrium under the new conditions, that is, AZj = AZo. We have in general Qt Jp ' \dp = — i^ At + V Ap Now if A;;' is chosen as the negative of the "osmotic pressure" of impure (1), it is obvious that impure (1) at p, t + At will be in equili- brium with pure (1) at p -\- Ap', t + At, and that hence the impure (1) will be in equilibrium with pure (2), where now pressure and temperature are the same on both impure (1) and pure (2), as they actually are. We have, therefore, — S]_ At + Vi Ap' = — So At ViAp' TVi At= — = - ^^Ap si — So AH where Ap (= — A^^') is the osmotic pressure of the impurity. This is the formula derived in the previous paper, except for sign. Ap is POLYMORPHIC TRANSFORMATIONS OF SOLIDS. 61 always positive; hence Ar is negative, and the transition line is de- pressed. The procedure is easily modified so as to give the pressure shift of the equilibrium line. Let now the pressures only be altered. The pressure on (2) is increased by Ai>y, and that on (1) by Api. (See Fig. 3.) Under these conditions pure (1) at p + Api, t is to be in Figure 3. Showing the displacement of pressure produced by impurities. equilibrium .with pure (2) at p + Api, r. If now we choose Ap-2 — Api, equal to the osmotic pressure of the impurity dissolved in (1), then we have impure (1) at p + Ap2, r in equilibrium with pure (1) at p + Api, T, and hence in equilibrium with pure (2) at p + Ap2, r. That is, Api gives the pressure shift of the transition line. We have or also Whence AZi = AZ.2 ro Ap2 = ("1 Api Ap2 — Api = Ap (Ap = osmotic pressure) Ap2 l\ Ap This gives the pressure shift (= Ap2) of the line due to impurity. It might haxe been obtained from the value given above for the tempera- .. . r/r ture shift 1)\" putting Ar = -— Ap, but the method of deduction there given did not make it evident that this substitution would be allowable (It . . . when -T- = oo , as it is in those cases to whicli we wish to applv the dp 62 BRIDGMAN. formula. Conversely, of course, the formula for the pressure shift does not apply at a horizontal part of the transition line, where Vl — t'2 = 0. These formulas have been subjected to no restriction in the deriva- tion except that the phase (1) is that in which the impurity is dissolved. This might, if we liked, be the phase stable at the higher pressure or the lower temperature, instead of as we have shown it. The formulas show that in all cases the effect of impurity is to shift the transition line into the region of the pure phase. For a given concentration of the impurity, that is, for a given osmotic pressure, the displacement is greatest for those substances with a small latent heat and a small change of volume. These are much smaller for the solids investigated here than for liquids, so that one would expect in general the displace- ment of the transition lines to be greatest for the substances investi- gated here. But as has been remarked, very few of these substances contain dissolved impurities (form mixed crystals), so that most of the transition lines are unaffected by what impurities there may be. If the impurity is soluble in both phases, we get for the pressure shift ^ ViApi — V2Ap2 • . Vi — ^2 and for the temperature shift ^^=-AH Ti Api — V2 Api This shows that if the total amount of impurity is slight, and if it is so distributed between the two phases that Vi Ajh = % Ap2, then there is no shift of the transition line. The phenomena in the neighborhood of a triple point offer no particular difficulty. It may be shown directly by substitution that the displaced transition lines must pass through a triple point as well as the original lines, no matter what the relative amounts of impurity dissolved in the three separate phases. This of course is what we know must be the case from other considerations. It should be noticed that although these formulas are entirely valid when the impurity is dissolved in more than one phase, nevertheless the conditions under which they are derived are not always close to the conditions of practise. We have assumed a knowledge of Api and A;^. This demands that we know the way in which the impurity is divided between the two phases. In practise this problem of the distribution of the impurity must usually be solved first, since the POLYIVIORPHIC TRANSFORMATIONS OF SOLIDS. 63 practical conditions usually give us the total amount of impurity present in the two phases together. The distribution problem is not touched above; to solve it would require a knowledge of the heat and volume effects of solution. The Equation of the Transition Lines. This equation has already been developed in a previous paper. ^ The equation is (p - po) \Av + A^ (r - To) -i Aa (p - po)] - {Aso - ACp) (r - tq) — ACp tIocj— = 0 To For the slope of the line we find cIt Ai'o + A|3 (r — To) — Aa (p — po) dp and for the second derivative Aso + ACp log A^{p — po) To "^^"^^Y-2A/3^ + Aa' .dpj dp rfV _ _ ^^ dr dp"^ Av dp The notation is as follows; A v is the difference of volume of the two phases at the point po, tq of the transition line {Av = Vi — vo) ■ A|3 is the difference of the thermal expansions. Aa is the difference of the compressibiHties considered as positive. ''dvi\ fdi\ ^dpjr \dp ACp is the difference of the specific heats (ACp = Cp^ — Cp), and A^o is the difference of entropy between the two phases (A^o = AH/tq). The equation presupposes that throughout the range of application Aa, A/3, and ACp remain constant. In the preceding paper very little discussion was given of the shape of the curve determined by the above equation, because the shape of the melting curve is determined essentially by the variations with pressure and temperature of Aa, A/3, and ACp. In particular, Aa 64 BRIDGMAN. varies greatly over the pressure range of 12000 kgm. But in the case of the solids to be studied here, it seems reasonable to assume a con- siderably better approach to constancy of the three differences. For instance, the compressibility of a liquid such as water has decreased to I its initial value at 12000 kgm., whereas the compressibility of steel does not vary more than a fraction of one per cent over the same range. Steel is of course an extreme case, and the difference of compressibility may change much more than the compressil)ility, but in any event we would seem to be justified in presuming that Aoc for two solids is more constant than for a solid and a liquid. A dis- cussion of the above equation is of more \'alue here, therefore, than in the case of the melting curves. It will not pay to completely discuss all possible combinations of the constants. We will, however, show that the curve may take a great variety of shapes according to the relations between the constants, and that some of these shapes do not seem to exist in practise. We infer, therefore, that in practise the con- stants flo not assume all possible relations to each other; it will be our problem in the latter part of this paper to determine these constants,, and find within what relative range they are actually restricted. We notice in the first place that if Aa > 0, that is, if the high tempera- ture phase is more compressible, as it is in many cases, that the curve crosses the pressure axis twice. For we obtain immediately on put- 2 Ac- ting T = To that p — po = "T — • This means that when the pressure has been increased sufRciently to squeeze the high temperature phase into a volimie as much smaller than the low temperature phase as it was originally larger, the two phases can coexist again in equilil)rium. This in itself indicates a rather unusual state of affairs; the only example found so far is //f/Zi- But the equation furthermore indicates that the curve may under proper conditions break up into two curves, the second pressure just found lying on the second curve. This would mean that the same phase may exist in two isolated regions of the phase diagram, the two regions of stability being separated by the region of stability of the second phase. Such cases, if they occur at all in practise, are extremely rare; the only suggestion of such a thing of which I know is concerning the modification of XHiNOs stable below —16°, made by Wallerant.^ This however has not been ^•erified in a more recent very careful investigation by Behn,^ although 2 F. Wallerant, Bull. soc. fr. min. 133-374 (1905). 3 U. Behn, Proc. Roy. Soc. 80, 444-4.57 (1907-08). POLYMORPHIC TRANSFORMATIONS OF SOLIDS. 65 he was not able to prove definitely that such was not the case. It is interesting to bear in mind that this is at least a theoretical possi- bility. For the present, then, we shall assume that Aa does not remain constant over a wide enough range to allow curves with two branches, and we shall consider only the nearer branch of the cur^'e in the following. If the two phases have the same specific heats (ACp = 0) the equa- tion becomes particularly simple, degenerating into a hyperbola. This hyperbola has one vertical asvmptote and one inclined at an angle Aa . / * TTiri^- The hyperbolic form includes a number of types of behavior. - '^P If Aa > 0, the curve may rise to a vertical asymptote, or may rise to a maximum and fall to a vertical asymptote, or rise to a maximum and fall to an inclined asymptote, or fall to a vertical asymptote, or fall to a minimum and rise to a vertical asymptote, or fall to a minimum and rise to an inclined asymptote. And if Aa<0, the curve may rise to a vertical asymptote or rise to an inclined asymptote with con- vexity either up or down, or may fall to a vertical asymptote, or may fall to an inclined asymptote with convexity either up or down. Of these cases, at least those with a vertical asymptote do not appear to occur in practise. The hyperbola degenerates into one important special case. If ACp = 0, and Aa = 2A/3t^ the curve breaks up into two straight lines (we neglect the second line). Many of the transi- tion lines actually found are nearly straight, so this condition is ap- proximated to in practise. If now, in general, ACp 5^ 0, we see that the curve can ne\er have a vertical asymptote. It may, howe^'er, have a vertical tangent, and be double valued with respect to temperature, if ACp<0. This is a case met in practise, benzol and ice for example. In general the equation demands at high pressures either a maximum or a minimum, or a curve with two branches. Since these are not actually of frequent occurrence, the legitimate use of the equation must be restricted to a comparatively narrow pressure range. We may perhaps say that very roughly the usual type of the curve is one concave toward the pressure axis, whether rising or falling. If the curvature is reversed, and there are examples of this, it would imply in general either an exceptionally large positive value of A/3, or a negative value of ACp. We discuss now the relations at a horizontal or vertical tangent. W^e have seen in the previous paper that the transition data give us two relations between Aa, A/3, and ACp. In general one other relation ACp-\-Av-TA^ 66 BRIDGMAN. is necessary to compute these three quantities. But at a horizontal or vertical tangent, the relations simplify in such a way as to determine uniquely two of these three quantities without the necessity for a third relation. We shall assume for the deduction that we know AV and AH along the transition line. These date are, as a matter of fact, determined in the following work. We are at liberty there- fore to use the following relations: dAv dT .- . -r- = —A^-Aa dp dp dAH ^ dr dp dp' dr Now at a horizontal tangent, "i- = 0. We immediately see that at such a point the equations give us the means of determining both Aa and Aj8 from the transition data alone. These values are; dp ^ T dp We cannot find ACp from the equations; the only method is direct experimental determination. These relations are of application to Hgl2. The equations furthermore show that for those substances which we may call normal [Aa>0, A/3>0], the maximum of the AH curve must come before the maximum of the transition curve. dj) At a vertical tangent, 3" ~ 0, and we get These relations are of application to water or benzol. We cannot determine Aa at a vertical tangent from the transition data alone. Two of the relations may be written in a simpler form, involving the curvature of the transition line. These relations are ; Aa = X curvature, at a horizontal tangent [~r — ^ j> fdr ACp = — r Ai) X curvature, at a vertical tangent ( ;/" = °° and POLYMORPHIC TRANSFORMATIONS OF SOLIDS. 67 It is to be noticed that either a horizontal or a vertical tangent demands properties that are in a certain sense abnormal. At a hori- zontal tangent, to the right of the point of tangency, the phase with the smaller volume has the higher compressibility, and immediately below a vertical tangent the phase stable at the higher temperature has the smaller specific heat. We have examples for both these cases. We have seen that it is not possible in general to determine Aa, Aj8, and ACp from the data of the transition curve alone. (By the transi- tion data we mean the curve giving the relation between pressure and temperature, and the values of AV and AH along this curve.) It is important to notice, however, that if the three phases come together to a triple point, we do have enough to determine Aa, A/3, and ACp completely on each of the three transition lines. The reason for this . , dAv , dAH . , ^ P , , , , ,. IS that —1 — and — r — are mdependent ot each other on the three Imes, whereas Aa, A|3, and ACp are not. We have evidently Aq;i3 = Aq!i2 + Aq:23 etc. That is, at a triple point there are only six unknowns to dAv determine, and six equations, one involving —7— and another mvolving — 1 — on each transition line. The six equations are obvious from the above; we need not bother to write them down. This information is most important. In the following the results of the calculations are given for those substances whose phase diagrams contain triple points. In the case of those substances which do not have a triple point it has been possible in a good manj^ cases to obtain some evidence as to the value of Aa from the difference of slopes of the isothermal jj-v lines above and below the transition point, and so, with the equations above, to calculate AjS and ACp. It must be emphasized that the values so found are in many cases very inaccurate. I have, however, thought it worth while to give them, because this is a matter about which absolutely nothing seems to be kno\\Ti, and is obviously one of ex- treme importance for the understanding of polymorphic forms. In fact, even the sign of these quantities does not seem to be known at present; we do not know (except perhaps in a few cases) which of two polymorphic forms has the greater compressibility or thermal expan- sion or specific heat. The values given in the following would fully justify themselves if they should even give the sign correctly. 68 . bridgman. Data. The data for the individual substances follow. Potassium sulfocyanide. This was Kahlbaum's purest, "zur Analyse," used without further purification. Two sets of readings were made; one consisted of five points on the transition curve and two on the melting curve at higher pressures, and the other was of one point on the transition curve at low pressures. The transition from one solid to the other is sharp, so that apparently the solid does not form mixed crystals with whatever impurity is present. There is some impurity present, however, as shown by a rounrling of the corners of the melting curve, but it must be small in amount, because the melting and freezing ran unusually rapidly and approached closely the same values from alcove and below, nearly independently of the amount melted. The impurity is most probably a small amount of moisture. For the determination at low pressures the sul)stance was carefully dried in vacuum for seven hours at 100°. The value for AV found here lies on a smooth curve with the values found at higher pressures, so the effect of the impurity must be very slight. For the determination at high pressures, during which melting was to occur, the KSCN in the form of fine crystals was hammered com- pactly into a steel shell, open only at the top. For the determination at low pressures, without melting, the fine crystals were hammered into a steel shell open at both ends and perforated on the sides with numerous small drill holes. The somewhat greater restraint offered by the steel shell in the first run seemed to have no harmful effect. The quantity used was about 47.5 gm. Pressure was transmitted directly by kerosene. The experimental results are shown in Figures 4 and 5 and the computed values of AH and AE in Figure 6. Table I gives numerical values at even pressure intervals. It is especially to be noticed that the values given for the melting curve are only approximate; since no great effort was made to get these very accurately. Only two points were determined, so the curvature can not be stated. The transformation curve and the AJ' curve for the transition solid-solid are both like the corresponding curves for liquids. The AH curve falls, however, whereas for the majority of liquids it rises. The effect of pressure on the transition and melting does not seem to have been previously determined, so there are no values for compari- POLYMORPHIC TRANSFORMATIONS OF SOLIDS. 69 son at high pressures. The transition point by extrapolation (the lowest point was 240 kgm.) from the present measurements is 140°, against 146° by Gossner.* The melting point extrapolates to 171.2°, against the values 161.2° by Pohl in 1851 (quoted in the French Tables), 173.8° by Wagner and Zerner,^ and 174.2° by Wassiljew.^ TABLE I. PoTASSirM SuLFOCY.\XIDE. Pre.s.sure 1 Temperature AV cm^./gm. Jp Latent Heat kgm. m./gm. i Cliange i of Energy kgm.m./gm. I-II. 1 140°. 0 0.00306 0.00954 1.325 1.325 500 149 .2 274 892 1.297 1.283 1000 157 .9 J L 245 841 1.256 1.231 1500 166 .1 221 801 1.212 1.179 2000 173 .9 200 772 1.158 1.118 2500 181 .5 183 752 1.107 1.061 3000 189 .1 170 743 1.058 1.007 3500 196 .5 159 735 1.016 0.960 4000 208 .3 151 728 0.969 0.909 LiQ uiD — I (Approximate values). 1 171°. 2 .0497 .0214 10.3 10.3 500 181 .9 480 " 10.2 10.0 1000 192 .6 463 u 10.1 9.6 1500 203 .3 446 u 10.0 9.3 It should be noticed that the extrapolated value of the melting point is probably too high, because no account is taken of the certain curva- ture of the melting curve. It is to be remarked that a high transition temperature is not presumably more correct than a low one, as it is in the case of melting. Impurity in the form of mixed crystals must be proved before such a presumption can be recognized, and even then, 4 B. Gossner, ZS. Kryst. 38, 110-168 (1903). 5 K. L. Wagner and E. Zerner, Monatsh. Chem. 31, 833-841 (1910). 6 H. Wassiljew, Chem. Centralbl. 1910, II, 56. 70 BRIDGMAN. there is the possibihty that the impurity is dissolved in the low temper- ature phase, contrary to the case for melting. Gossner gives the density at room temperature as 1.898. There seem to be no other values of AV or AH at atmospheric pressure. No other modifications were found to 12000 kgm. at 20° or 160°. The directly measured values of the difference of compressibility, which were self consistent, indicate an abnormally large difference between the two phases. The difference increases rapidly with in- creasing pressure and temperature, the low temperature phase being more compressible. But the measurements cannot be accurate, the values being much too large, because they lead to impossibly large 200' 4J ^ 180' imiTfHHHtftlT'l 160" 140° i ■ /' i ■ ! '! . i:_^^,J/__:4..,..,4 1 J -^-\-- V - I ■ (■■■ !■:: 1000 2000 3000 4000 Pressure, kgm./cm.^ Potassium Sulfocyanide Figure 4. Potassium Sulfocyanide. The observed equilibrium pressures and temperatures. differences of the specific heats. The direct measurements would indicate that the low temperature phase has the higher expansion and the higher specific heat. In addition to the measurements of the difference of compressibility, the difference of expansion was measured at low pressures (77 kgm.). This measurement is more moderate than the high pressure ones, but also demands that the low temperature form have the higher compressibility, expansion, and specific heat. The values at 77 kgm. are as follows; Aa = -O.OeO A/3 = - O.O3I2 ACp = —6.1 (kgm. cm.) POLYMORPHIC TRANSFORMATIONS OF SOLIDS. /i The reason for the too high difference of compressibility found at high pressures is not clear. At any rate, one would seem justified in assum- ing that KSCN is abnormal in that the low temperature form is more compressible, expansible, and has a higher specific heat than the high temperature form, and that the differences are unusually large. 1000 2000 3000 Pressure, kgm./cm.^ Potassium 5ulfocyanide 4000 Figure 5. Potassium Sulfocyanide. The observed differences of volume ])etween the hquid and solid I, and between solids I and II. Figure 6. Potassium Sulfocyanide. and the changes of internal energy. 1000 2000 3000 4000 Pressure, kgm./cm.^ Potassium Sulfocyanide The computed values of latent heat 72 BRIDGMAN. NH4SCN. This substance was Kahlhaum's purest, "zur Analyse," used without further purification, except heating to 100° for several hours in a vacuum, to remove all traces of moisture. Measurements were made with three different fillings of the apparatus. The first two were at the higher pressures, and for these the dry NH4SCN powder was hammered into a compact mass in a steel shell, open at both ends and with perforations in the sides. Pressure was trans- mitted directly by kerosene. The third filling was for the run at low pressures, and for this the dry powder was hammered into the in- verted nickel steel shell, and pressure transmitted to it by mercury. The difference in the pressure transmitting medium caused no per- ceptible difference in the behavior. The quantities used were about 34 gm. for the first two fillings, and 21 gm. for the third. A transition point was first found at about 2200 kgm. at room temperature. NH4SCX was known to be dimorphous, but the tran- sition at atmospheric pressure is about 90°, so it seemed at first that there was here a third new modification. On determining the transi- tion at other temperatures, however, it appeared that this was not a new form, but that the transition line runs from high to low tempera- tures with increasing pressure; that is, the high temperature form has the smaller volume. This somewhat unusual behavior does not seem to have been noticed ; at any event this fact has not found its way to tables of crystalline properties, nor in Gossner,* who has published the most extensive investigation on this substance. Four points were found with the first filling of the apparatus, from 0° to 67°. The temperature was then raised to 200° in the search for other modifica- tions. At about 4500 kgm., an irreversible change took place, with an increase of volume of about 0.04 cm' gm. No other transition could be found on increasing pressure to 12000 kgm., and then no transition at all on releasing pressure from 12000, although the melting curve ought to have been crossed. On opening the apparatus, it was found that the irreversible transition was really a decomposition. Very little gas was given off by the decomposed products, but the smell was probably the worst that has ever been artificially produced. The solid products of the decomposition were of a bright yellow color, probably due to free sulfur, but no further analysis was attempted. The products of the decomposition found their way through the whole of the apparatus, and deposited themselves on the manganin coil, short circuiting it, so that it had to be thrown away. The secontl filling of the apparatus was made to redetermine one point at about 2200 kgm., the AT' value for which did not fall on a smooth curve I'OLYMORPHIC TRANSFORMATIONS OF SOLIDS. 73 with the other values. For this determination, a new manganin coil was used, which however, had been previously carefully seasoned. Both the values of equilibrium pressure and temperature and AF found with this coil lie on a smooth curve with the other points, bespeaking the accuracy of the manganin coils. Two points were determined with the low pressure apparatus; one by the method of varying temperature at a constant pressure of 77 kgm , and the second by varying pressure at constant temperature, the equilibrium pressure being about 240 kgm. The values found at 77 kgm. did not agree with the other values, and were discarded; the error is apparently an effect of air in the apparatus, which of course becomes vanishingly small at higher pressures. The transition between the two solids runs cleanly and sharply, very nearly the same pressure being reached from above and below. 1000 2000 3000 Pressure, kgm. /cm/ Ammonium 5ulfocyanide Figure 7. Ammonium Sulfocyanide. The observed equilibrium tempera- tures and pressures, and the changes of volume. The change of volume of NH4SCN was large enough so that it was possible to measure the time rate of reaction, that is, the time rate of recovery of pressure. These results will be given in another paper. The retardation effects, either superheating or subcooling, were not large; no careful attempt to measure these was made. It is probably safe to say that the reaction ran invariably if the transition line was crossed as much as 300 kgm. in either direction. The experimental points are shouTi in Figure 7, the computed values of A// and AE in Figure 8, and a summary of the numerical values is given in Table II. The exceptional curvature of the transition line is to be noted. 74 BRIDGMAN. There seem to be no previous determinations of the effect of pres- sure on the transition point, nor of the change of volume or the latent heat at atmospheric pressure. The previously listed value of the transition temperature is 92°, against 87.7° found by extrapolation. Gossner * also states that the transition is very sharp (as was found E 4 . 0 1000 aooo 3000 E Pressure, kgm./cm.^ •^ Ammonium Sulfocyanide Figure 8. Ammonium Sulfocyanide. The computed values of the heats of transition and the changes of internal energy. above), that the melting point is 169°, and the specific gravity 1.305 at room temperature. Several fairly consistent measurements of the difference of com- pressibility of the two phases were made. The results are shown in TABLE II. Ammonium Sulfocyanide. Pressure Temperature AV cms./gm. dt dp Latent Heat kgm.m./gm. Change of Energy kgm.m./gm. 1 87°. 7 0.0409 -0.03336 4.423 4.423 500 71 .5 412 3146 4.508 4.714 1000 56 .2 414 2976 4.581 4.995 1500 41 .7 417 2828 4.636 5.261 2000 27 .9 419 2698 4.675 5.513 2500 14 .7 422 2590 4.683 5.738 3000 2 .0 424 2494 4.677 5.949 POLYMORPHIC TRANSFORMATIONS OF SOLIDS. 75 Table III. The high temperature phase, that is, the phase with the smaller volume is less compressible, more expansible, and has the smaller specific heat. The differences between the two phases become smaller at the higher pressures. The sign of the difference of compres- sibility seems natural, but one might perhaps expect the high tempera- ture phase to have the liigher specific heat. It is remarkable that the behavior of the two phases here is the exact opposite of water and ice I. Water at low pressures is more compressible, less expansible, and has a higher specific heat than ice. No other modifications were found to 12000 kgm. at 0°, or to HOOD at 100°. The decomposition at high temperatures prevented search for other modifications at higher temperatures, and also made it use - TABLE III. Ammonium Sulfocyanide. Pressure Aa A;3 ACp 1 -.O549 + .O3I7 -3.6 1000 -.0642 .O3I6 -3.6 2000 -.0634 .O3I4 -3.5 3000 -.O527 .O3I3 -3.0 less to try for points on the melting curve. It is a little surprising that no other modifications were found. The abnormality of the high temperature modification, since it is formed with decrease of volume, might lead one to expect that at high pressures it would be replaced by another modification giving a phase diagram something like that of Agl. It may be, of course, that it is not the high temperature form that is abnormal, but that instead it is the low temperature form that has an abnormally large volume. In this case, the phase diagram is the natural one. The possible relationship of the several modifications of KSCN and NH4SCN is suggested by the crystalline forms. At first sight it is surprising to find different types of phase diagrams for these two sub- stances, because K and NHi usually replace each other isomorphously. Gossner * has shown, however, by a comparison of the crystalline forms, 76 BRIDGM.VN. that the ordinary modifications of KSCN and NH4SCN are not cor- responding modifications. The low temperature form of NH4SCN is monoclinic, and the high temperature form rhombic, whereas the low temperature form of KSCN is rhombic, and the high temperature form belongs to some other system of which all that can be said is that it is doubly refracting. The rhombic forms of the two substances are isomorphous, forming mixed crystals. If XH4SCN and KSCN are really isotrimorphous we would expect another modification (not yet discovered) 6i KSCN at low temperature. Failure to find an- other form of NH4SCN at high pressures is evidently because the lower transition is of the ice type. The absence of the upper transi- tion line for NH4SCN may be explained by supposing that the rhom- bic modification melts before the temperature can be raised to the transition value. Potassium Sulfide. — This was Kahlbaum's purest, "zur Analyse," not further purified. Two runs were made; the first at higher pres- sures, and the second at 77 kgni. The quantity used was about 47 gm. For both runs the salt was hammered into an open steel shell with lateral perforations, and pressure was transmitted directly to it by kerosene. K2S is exceedingly hygroscopic; it was necessary to work very rapidly in transferring it to the shell, and it was almost impossible to avoid the collection of some slight amount of moisture from the air. It does not seem to have been previously noted that KoS is dimorphic at atmospheric pressure. This is doubtless to be explained by the very slight change of Aolume (accompanied by a very small heat effect), and the inconvenience of making observations at atmospheric pressure because of the extreme hygroscopy. With the apparatus and the quantity used the discontinuity of the piston displacement was only 0.01 inches. Observation of the transition at low pressures is further rendered difficult by the fact that the transition point is fairly near the decomposition point at atmospheric pressure. At 77 kgm. efl^ects were observed only 15° above the transition temperature which are probably to be explained by incipient decomposition. The transition line must be passed by about 200 kgm. in either direction before the reaction runs. The change of volume is so slight that the pressure does not return to the equilibrium value after the reaction starts, but the reaction runs to completion with exhaustion of one phase, before the pressure can return to the equilibrium. This means that the equilibrium pressures are enclosed within considerably wider limits than is usually the case. The p-t points are, therefore,. POLYMORPHIC TRANSFORMATIONS OF SOLIDS. 77 more irregular than usual. The same cause produces an even greater irregularity in the AJ' values; we have here a large percentage error due both to the extreme smallness of the change and the necessity for a considerably wider extrapolation than usual. To reduce this error as much as possible, the measurements of AV were made with both increasing and decreasing pressure. The very small change of the volume and the comparatively large lag of the reaction combine to make impossible measurements of the rate of the reaction. After the reaction had once started, it ran to completion within the time required for the dissipation of the ordinary heat of compression. The experimental results are shown in Figure 9, the computed values for AH and AE in Figure 10, and the numerical values in Table 0 12 3 4 5 Pressure, kgm./cm.' x 10' ' Potassium Sulfide Figure 9. Potassium Sulfide. The observed equilibrium temperatures and pressures and the changes of volume. IV. The beha^•ior of the various quantities is like that of a typical liquid. There are no other values for comparison, since it was not knoA\Ti before that KoS is dimorphic. The direct measurements of the difference of compressibility are irregular and inaccurate, but the probability is that the low tempera- ture form of K2S is the more compressible, the difference of compressi- bility being of the order of O.O5I . This means that the low temperature form is also more expansible, and has the higher specific heat. The difference of expansion is of the order of O.O48, and of the specific heats of the order of 3 kgm. cm. per gm. 78 BRIDGMAN. No other modifications were found to 12300 kgm. at 20° or to 12000 at 200°. Potassium Chlorate. — This substance was Kahlbaum's purest, used without further purification. It was hammered into a compact I -32 ^ .30 Pressure, kgm./cm.' x 10^ Potassium Sulfide Figure 10.""' Potassium Sulfide. The computed values of the heat of transition and the change of internal energy. mass in a thin steel shell, open at both ends, with lateral perforations, and the pressure was transmitted directly to it by kerosene. 5.3 gm. were used. It is remarkable that it did not explode under pressure by spontaneous combination with the kerosene. NaClOs does so TABLE IV. Potassium Sulfide. j Pressure Temperature AV cm'./gm. .It dp Latent Heat kgm.m./gm. Change of Energy kgm.m./gm. 1 146°. 4 .000948 .0120 .331 .331 1000 158 .4 918 u .330 ..321 2000 170 .4 886 a .327 .309 3000 182 .4 855 u .324 .298 4000 194 .4 825 u .321 .288 5000 206 .4 794 .317 .278 POLYMORPHIC TRANSFORMATIONS OF SOLIDS. 79 explode; this will be described in detail later. Since the transition line of KCIO3 does not cross the axis of atmospheric pressure, one filling of the apparatus and one series of runs was sufficient to deter- mine all the data of the transition curve. A run with a second filling, pressure transmitted by mercury, was made to more carefully de- termine the velocity effects at 0°. The reaction is a slow one, and shows considerable lag on both sides of the line. The pressure limits within which the equilibrium values were enclosed are wider than usual, the limits being very much widest at the low temperatures. The wide pressure limits at the low temperatures make the AV values rather irregular at the low end of the curve. The phenomena of reaction velocity are especially inter- esting and easy to follow in the case of this substance, and an especial study was made of them. These will be given in detail in another place. Attention may be called, however, to an experiment at 0° extending over four days. The equilibrium pressure found under these conditions is considerably higher than that to be extrapolated from the high temperature values. The experimental results are shown in Figure 11, the computed value of AH and AE in Figure 12, and the numerical values in Table V. Within the limits of error, the transition line is straight, and the curve for AV is also straight, with a very small slope. The new phase was not known before, so there are no other values J.00> Pressure, kgm./cm.' x 10' Potassium Chlorate for comparison. No other forms were found be- tween atmospheric pressure and 12000 kgm. at room temperature, and none between 3000 and 12000 at 200°. There are, however, rea- sons for suspecting that there may possibly be another modification at 0°, and that the transition line drawn should properly end in a triple point above 0°, with the line Figure 11. Potassium Chlorate. The observed equilibrium tempera- tures and pressures, and the changes of volume. 80 BRIDGMAN. branching below the triple point. The reasons for this suspicion are the high values of the equilibrium pressure found at 0° from the run lasting four days, and also the high value found for AV at 0°. The high value of AV may be due to experimental error, which is larger than usual at 0° because of the great lag, but the high equilibrium E bD u u a be ^Pressure, kgm./cm.^ x 10' Potassium Chlorate Figure 12. Potassium Chlorate. The computed heat of transition and the change of internal energy. pressure can hardly be explained in this way. In any event, the data given above for the lower end of the curve should be used only with the greatest reserve; it seems that probably either the curve should get steeper toward 0°, instead of remaining straight as drawn, or else there is a new modification, and the transition line should branch. TABLE V. Potassium Chlorate. Pressure Temperature AV cm'./gm. di dp Latent Heat kgra.m./gm. Change of Energy kgm.m./gm. 5680 0° 0.02510 0.976 .07026 -1.356 5880 20 2506 n .07525 1.398 6090 40 2501 u .08023 1.443 6290 60 2497 il .08521 1.485 6500 80 2492 ti .09018 1.530 6700 100 2488 u .09513 1.572 6910 120 2484 it .1001 1.617 7110 140 2479 u .1050 1.658 7320 160 2475 tl .1098 1.701 7520 180 2470 u .1147 1.743 7730 200 2466 u .1196 1.786 POLYMORPHIC TRANSFORMATIONS OF SOLIDS. 81 An attempt to find the crystalline form of KCIO3 II will be described in another place. Seven measurements of the difference of compressibility l:)etween the two phases were made. These all point to the same conclusion; that the difference is very small, probably less than 0.062. This enables us to set upper and lower limits for the diflference of expansion and specific heats. Assuming Aa = 0, we find A/3 = — 0.0622, and if Aa = 0.062, A/3 = -O.OtIo. A value for Aa only 10 % higher than O.O52O would give a positive instead of a negative value for A/3. We conclude that the difference of thermal expansion between the two phases is unusually small, and that possibly, although not certainly, the high pressure phase is the more expansible. Aa and A/3 are so small that the values of ACp are not affected by changes within the limits of error. ACp is nearly constant o^■er the entire range at — 0.024 (= —0.00056 gm. cal. per gm.). Here again is a case where the phase stable at the lower temperature has the higher specific heat. PoTASSiuiNi Nitrite. — This substance was Kahlbaum's purest, used without further purification, except drying. After the runs had been made the irregularity of the results led me to take up the question of the purity of the substance with the J. T. Baker Chemical Co., and I learned from them that it is impossible to prepare KXO2 pure commercially. There is always impurity of water and KNO3. The impurity may be fairly large in amount, varying from 10 to 15%. An analyzed sample from the J. T. Baker Chemical Co. contained 89% KNO2. The amount of water was not stated. These results, there- fore, must be taken as merely indicating the order of the effects which may be expected with pure KNO2. It is unfortunate that there was this large amount of impurity, because the results are interesting both because of the direction of curvature of the transition line, and the behavior of the reaction velocity. Three fillings of the apparatus were used. With the first, only one point was determined, at 22°. The salt was not dried for this run, and considerable rounding of the corners led to the suspicion that moisture was present. For the second filling of the apparatus, the salt was dried in vacuum over H2SO4 for two days. Five points were found with this filling. For the third filling, the salt was dried in vacuum at 100° for several hours; three points were found with this filling. The different methods of getting rid of the moisture seemed to make no difference; the three runs were in essential agreement. In all three fillings the KNO2 was hammered into a compact mass; the first two 82 BRIDGMAN. runs were made with the pressure transmitted directly by kerosene, while for the third run, the compacted mass was submerged under mei'cury, by which the pressure was transmitted to it.^ Probably none of the attempts to remove the moisture was entirely successful; the KNO2 was always different in texture from any of the other sub- stances. When hammered, it compacted into a coherent semitrans- lucent mass, in texture like celluloid when cut with a knife. ^ The pressure limits within which the equilibrium was shut varied from 40 to 360 kgm.; in general the limits were wider at the lower temperatures, but the variations were very irregular. The irregularity is partly mixed up with the varying quantity of impurity at different stages of the reaction. The AT isothermal curves showed the effect of impurity by being rounded at the corners. The rounding took place 9 80 5 6 7 8 9 10 Pressure, kgm./cm.^ x 10^ Potassium Nitrite Figure 13. Potassium Nitrite. The observed equilibrium pressures^and temperatures and the changes of volume. at both corners; this is very unusual, and is the only case of it that I have found. It means that both of the phases form mixed crystals with the impurities, and that the impurity which is miscible with one phase is different from that miscible with the other. As one would expect from the rounding of both corners, no lag of the reaction was noticed in either direction from the transition line, at least beyond the region of indifference. Beside the unusual rounding of both corners of the isothermals, KNO2 is unique in that the reaction velocity is greater with rising POL'i'MORPHIC TRANSFORMATIONS OF SOLIDS. 83 pressure than with falling. This was found at all temperatures; it may well be that it is not a genuine effect peculiar to the pure sub- stance, but is connected with the separating out of mixed crystals of impurities. The experimental results are shown in Figure 13, the computed 5 6 7 8 9 10 Pressure, kgm./cm.^ x 10* Potassium Nitrite Figure 14. Potassium Nitrite. The computed values of the heat of tran- sition and the change of internal energy. TABLE VI. Potassium Nitrite. Pressure Temperature AV cm3./gm. di dp Latent Heat kgm.m./gm. Change of Energy kgm.m./gm. 5000 -3°.0 .0312 .0169 4.99 3.43 6000 14 .0 327 171 5.49 3.53 7000 31 .4 341 182 5.71 3.32 8000 50 .8 355 212 5.42 2.58 9000 74 .4 369 274 4.68 1.36 9500 89 .4 376 343 3.97 .40 10000 109 .3 383 472 3.10 -.73 84 BRIDGMAN. values of A// and AE in Figure 14, and the numerical values in Table \l. The marked upward concavity of the transition line is unusual. There are no other results for comparison. The data are too inaccurate to justify any attempts to estimate Aa, AjS, or ACp. No new form was found to 12000 at .30° or 60°. Carbon Trichloride. — -This was Kahlbaum's purest, used with- out further purification. The powder was hammered cold into an inverted steel shell, and the pressure transmitted to it by mercury. Four fillings of the apparatus were made, two for the points at high pressures, and two for the points at low pressures (80 to 400 kgm.) About 21.5 gm. were used for the high pressure determinations, and 33 gm. for the low pressures. C2CI6 is known to have three modifications at atmospheric pressure, and hence two transition lines. These have already been investi- gated by Tammann "^ up to 3000 kgm. The substance promised to be an interesting one, because the transition lines diverge at atmos- pheric pressure, a phenomenon not shown by any other substance of which I know. It was of interest to see whether these lines would tend to come together again at high pressures. It is unfortunate, however, that this question could not be answered, because at the higher pressures on the transition line above 3000 kgm. the substance decomposes. The first run gave four points on the II-III curve to 177°. The transition runs cleanly on this line with no perceptible rounding of the corners. There is, howe\'er, some lag on both sides of the line. This lag, combined with the small change of ^■olume, once or twice resulted in the pressure limits from above and below being as wide as 200 kgm., one phase being all exhausted before pres- sure could be restored to the equilibrium value. The lag at other points was not so great, and the pressure range was only 40 to 60 kgm. wide. The effects were very different on the I-II curve. The first point attempted on this curve was at 177°. Above the proper value of the transition pressure, the pressure l)egan to rise slowly with time, just as it does with premature melting due to impurity. In the sup- position that the effect was really due to impurity, the work was hastened so as to minimize as far as possible the effect. On arriving at the transition point, however, the rise of pressure continued at a nearly constant rate to nearly 1200 kgm. beyond the supposed point. On increasing pressure again, the automatic rise of pressure continued 7 G. Tammann, " Kristallisieren und Schmelzen," 298. POLYMORPHIC TRANSFORMATIONS OF SOLIDS. 85 up to 3000 kgm. beyond the supposed transition point, that is to 6000 kgm. But above 6000 kgm. tiae secondary pressure reaction was a decrease after every increase. 6000 kgm. is above the II-TII line, but no trace of this transition was found on crossing the hne. On relieving pressure to 3500 the rise of pressure continued. On taking the apparatus apart, the CoCle was found to have decomposed, and a good deal of mercury had disappeared. A suffocating gas was ex'olved, probably chlorine. There was, however, no apparent effect on the steel cylinders. This decomposition is an interesting one. It evi- dently was produced by the high temperature and not by the high pressure, for the reaction runs with increase of volume. The effect of pressure would be to prevent the decomposition; this is what occurred at 6000 kgm. Neither did the decomposition occur at the higher pressures of the II-III curve at 177°. The products of the decomposition found their way to all parts of the apparatus, which had to be taken apart and cleaned. The manganin coil had to be carefully cleaned, but recovered its normal resistance. With the second setting up of the apparatus, one point was found on the lower end of the II-III curve, and the other points were found on the I-TI curve, not rising to the temperature of the previous decomposition. The order of points on the I-II curve was 102°, 136°, 86°, 104° and 120°. The first three of these determinations gave consistent values for pres- sure and temperature, and the second and third apparently gave good values for AT. The first point at 102° gave a value for AT impossibly low, as low as the II-III curve. This possibl}^ was due to the transi- tion not having been complete. The fourth and fifth points gave values of the pressures considerably lower than the other points and too high values for AV; evidently the decomposition had begun and was progressing. The decomposition had probably started at 136°, the highest temperature to which the material had been subjected, and then proceeded at an accelerated rate. The points at low pressures were determined six months after those at high pressures. Six determinations were made, both at constant temperature and constant pressure. There were difficulties here also. At 77 kgm. there was evidence of decomposition at 90°. This may have begun earlier; the A F value of this attempt was not good. The other attempts by the method of varying pressure at constant tempera- ture gave good results on the I-II curve, but AV on the II-III curve was lower than to be expected from the high pressure points. This might be due in large measure to the extreme slowness of the reaction from below on the II-III curve. Working at low pressures, the pres- 86 BRIDGMAN. sure could not be released far enough to ensure the completion of the reaction, and hence the AV ^'alues were too low. We see that the results are affected by numerous irregularities, due to unavoidable decomposition at the high temperatures, and the slow- ness of the reaction at the low temperatures and pressures. But on the II-III curve the points all lie regularly, as do also the AV points at high pressures. Satisfactory values for AV at low pressures could not be obtained, and the value listed for atmospheric pressure was obtained by extrapolation of the high pressure values. It is to be noticed that the II-III curve shows increasing curvature at the high pressures. This is unusual ; it may possibly be the effect of impurities, although the regularity of the AV points would seem to preclude this. On the I-II curve, the p-t values at high pressures are irregular because of decomposition, and only one good value of AV was found at high pressure. The values both for _p, t and AF at low pressures are, how- ever, apparently satisfactory for I-II. To obtain better values would require a great multiplication of experiments. It might be w^orth trying Avater as the transmitting medium, as possibly the action of mercury is such as to start the decomposition. In any cAent, it would probably not be possible to greatly exceed the temperature reached here without decomposition, and the interesting question as to whether the lines I-II and II-III ultimately approach each other would remain unanswered. 0 12 3 4 5 6 Pressure, kgm./cm.^ x 10* Carbon Trichloride Figure 15. Carbon Trichloride. The observed equilibrium temperatures and pressures. POLYMORPHIC TRANSFORMATIONS OF SOLIDS. 87 The experimental results are shown in Figures 15 and 16, the com- puted values of A// and AE in Figure 17, and the numerical values in Table VII. All the determinations of the -p-t points are shown in the .030 • .025 E ^.020 a> T, 015 I .010 ^.005 .000^ 0 12 3 4 5 6 Pressure, kgm./cm.^ x 10' Carbon Trichloride Figure 16. Carbon Trichloride. The observed differences of volume be- tween the several phases. 3.0U ^2.0 1.0 bo _:!_._. l_ njiTr-tr-:;]--::::::^;::::;::;!":: 0 12 3 4 5 6 Pressure, kgm./cm.^ x SO^ Carbon Trichloride Figure 17. Carbon Trichloride. The computed values of the heats of transition and the changes of internal energy. figures, but the bad values of AV are omitted, as these are much more subject to error; in fact, one of the AV points falls on the wrong curve. No other modification was found to 12000 kgm. at room tempera- ture. 88 BRIDGMAN. There are a few values for comparison. For the transition point, Schwarz ^ found by an optical method that the II-III transition was between 43.1° and 46.6° and the I-II point at 71.1°. The value found above for the II-III transition was 42.7°, by extrapolation. I have adopted the value of Schwarz for the I-II point, because of the irregu- larity of my points. The transition line was drawn as the best straight TABLE VII. Carbon Trichloride [C>Clf,]. Pressure Temperature Av cm^./gm. dr dp Latent Heat kgm.m./gm. Change of Energy kgm.m./gm. II-III. 1 42°. 7 .0097 .0273 1.122 1.122 1000 69 .8 85 268 1.087 1.002 2000 96 .3 75 260 1.065 .915 3000 121 .9 67 250 1.059 .858 4000 146 .3 61 238 1.073 .829 5000 169 .5 56 225 1.102 .822 6000 191 .3 52 208 1.161 .849 I- II. 1 71M .0280 .0326 2.96 2.96 1000 103 .7 259 (( 2.99 2.73 2000 136 .3 238 u 2.99 2.51 line from Schwarz's point to the high pressure points. If, however, it had not been for Schwarz's value, and if only the lowest points of the I-II curve had been used, as most probably unaffected by impuri- ties, the I-II line would have shown distinct concavity toward the pressure axis, and the extrapolated transition point at atmospheric pressure would have been about 69°. Tammann ^ has determined the effect of pressure on the transition up to about 2000 kgm. He does not mention finding decomposition, but his highest temperature was 8 W. Schwarz, Diss. Gott. (1892); Beibl. 17, 629-639 (1893). POLYMORPHIC TRANSFORMATIONS OF SOLIDS. ^\f only 130°. Both his transition lines are straight. His II-III transi- tion point at atmospheric pressure is 41.8° and his II-III line lies at pressures from 30 to 50 kgm. higher than mine. The agreement in both these cases is better than is usual between his results and mine. On the I-II line, his atmospheric transition point is at 67.3°, and at 130°, his transition line is about 30 kgm. higher than mine. There are apparently no other determinations of AV or the latent heat at atmospheric pressure, and Tammann gives no values at high pressures. Four measurements of the difference of compressibility between II and III were made. There can be no doubt that III is less compressi- ble, but there may be considerable question as to the numerical value of the difference. On account of the decomposition between I and II, TABLE VIII. Carbon Trichloride. Differences between II and III. Pressure Aa A;8 ACp 1 +0.0;24 + .O438 -.073 2000 .O52O4 .O44I + .23 4000 .06174 .O45O .70 6000 .0:446 .O452 1.25 no good values for the difference of compressibility between these forms could be obtained. The values of Aj3 and ACp computed from the experimental values of Aa are shown in Table VIII. The numeri- cal accm'acy is probably great enough so that one may say without question that II is more expansible; it is however, questionable whether the difference of expansion increases with rising pressure as the table shows. On the average, the specific heat of II is also greater than that of III ; one may well doubt whether the difference is negative at atmospheric pressure as the table would show. The crystalline forms of the several modifications has been deter- mined by Lehmann.^ The high temperature form is regular, the 9 O. Lehmann, ZS. Krj^st. 6, 580-589 (1882). 90 BRIDGMAN. intermediate one asymmetric, and the low temperature modification rhombic. Carbon Tetrabromide. — This substance is not carried in stock by any of the large chemical houses, but was made to order by Hoffmann and Kropff, and guaranteed by them to be chemically pure. Two separate lots were made by them, of about 100 gm. each. The second lot was perhaps purified more carefully than the first, being " triply distilled with steam," and showed a somewhat higher transition point. The particular interest of this substance lies in its suspected poly- morphic isomorphism with CCI4. In the first paper of this series two new modifications of CCI4 at high pressures were found, and since the writing of that paper, I have learned that a transition point at at- mospheric pressure, on the prolongation of the transition line at high pressures, has been discovered by Goldschmidt -^^ at —47°. Now it is known that CBr4 is dimorphic at atmospheric pressure, the transition being at about 47°. The suspicion was strongly suggested that the transition point of CBr4 at 47° corresponds to the newly found one of CCI4 at —47°, and that at high pressures another modification of CBr4 would be found corresponding to the third modification of CCI4. As a matter of fact, another modification was found, but it does not have at all the relation to the other two forms that is suggested by CCI4. It is unfortunate that here, as in the case of CiClg, the sub- stance begins to decompose just where things are getting most in- teresting. Measurements were made with five difl^erent fillings of the apparatus. With the first lot, two sets of readings at high pressures were made; one at 200°, which was terminated by decomposition, and the other to only 100°. A third filling with this lot was made for the low pres- sure point. With the second lot, one filling was made for high pres- sure measurements to 123°, and one filling for the low pressure point. The quantity used varied from 35 to 59 gm. In all cases the sub- stance was hammered cold into the inverted steel shell, and pressure transmitted to it by mercury. These several runs- are perhaps worth describing in some detail, both because the new modification belongs to a new type, and because of the decomposition effects, which are interesting in themselves, and which might without detailed description, be thought to make ques- tionable the validity of the new transition. It was of course expected, 10 V. M. Goldschmidt, ZS. Kryst. 51. 26 (1912). POLYMORPHIC TRANSFORMATIONS OF SOLIDS. 91 in analogy with CCI4, that a new modification would be found on increasing pressure at room temperature. None was found, however, up to 12000 kgm. Temperature was then raised to 127°, and pressure relieved from 12000. Two transitions were found, at 2800 and 2310. These transitions both ran sharply, with no rounding of the corners, and were entirely reversible. The change of volume at the lower transition was considerably^ less than that of the higher. Even the approximate location of the melting curve was not known at the time of this run, so that the low pressure transition might have been melting, although this seemed unlikely in view of the complete absence of rounding of the corners. Temperature was then raised to 152°, and the transitions found with decreasing pressure. Two transitions were found, but separated by a much wider interval than before, and there was some preliminary rounding of the corner at the lower point. This lower point might have been melting this time, therefore, but there seemed no connection with the low point at 127°. Temperature was now raised to 176°. The regular transition was found at high pressures, on a line with the other two points, and a lower point was also found. This lower point was too high to fall with either of the other two lower points, and the transition was partly reversible and partly irreversible; that is, on increasing pressure the return of ^'olume to the original value was not complete. At 176°, pressure was now raised to 6000, and the temperature raised to 200°. Pressure was now increased to 12000 kgm. Two notable decreases of volume were found on the way to 12000, at approximately 6800 and 7500. But on decreasing pressure, no transition of any kind could be found down to 4000 kgm., that is, considerably beyond the higher of the two transition lines. The apparatus was now cooled and taken apart; the CBr4 was found completely decomposed into carbon and bromine. The products of the decomposition reached all parts of the apparatus, and there was considerable trouble in cleaning it. In view of this decomposition and of the fact that the higher of the transition points lie on a line which extrapolates with little curvature to the known transition point at atmospheric pressure, this first run was supposed to indicate only one modification, the lower transition points being ascribed to some obscure effect of the decomposition. With this conclusion, the subject was dropped for over a month, when the rest of the first lot of material was used in the low pressure apparatus to fix the transition data at atmospheric pressure. Two runs were made with this; one with increasing temperature at con- stant pressure, and the other with decreasing pressure at constant 92 BRIDGMAN. temperature. The corners of the cur^'es were rounded, an effect not found before at high pressures. Evidently the CBr4 had become impure on standing. This same lot of CBr4 was then used again in the high pressure apparatus to give two points on the transition line below 127°, the lowest temperature previously used. The data were apparently all in; but on working them up, it appeared almost unmis- takable that the two low pressure transition points first found were genuine, really belonging to a new modification, and were not the effect of decomposition. The evidence which made this almost inescapable was the values of AT, which had not been computed before. There was a discontinuity in the values for AV along the transition line, the amount of discontinuity corresponding very closely with the AV found for the questionable points. Furthermore, the points on the lower end of the transition line, below 127°, did not lie smoothly with those above 127°, but there was a pronounced change of direction at about 110°. The perfect reversibility of the two suspected transitions strengthened the probability. The reason why it took me so long to admit that this was a new form, was because thus is a new type of transition, in which a new form appears at high pressures and tempera- tures with a volume intermediate between the two low pressure forms; there seems to be no excuse for the existence of such a form. In order to make certain that there was really a new form, a fresh lot of CBr4 was ordered from Hoffmann and Kropff, with the request that they take special precautions in purifying it. This lot was triply distilled in steam. The new material was first used in the high pressure apparatus. The first point was obtained at 108°, below the sup- posed triple point, and the second at 123°, above the supposed point and as near to it as was convenient to work. Only one transition was found at 108°, but two at 123°. The lower of these two, which is the questioned one, was found to be reversible exactly as before. There could be no doubt of the genuineness of the transition. It is unfortunate that this second lot decomposed at a lower temperature than the first lot; the decomposition was barely perceptible as low as 108°, and at 145° was proceeding so rapidly that even p-t values for the higher pressure could not be obtained. The AT' values at 108° and 123° with this second lot are both too high because of decomposi- tion. The rest of this second lot was used for a determination of the transition point at low pressures by the method of changing pressure at constant temperature: The value found by extrapolation for the transition temperature was 46.2°. This is somewhat higher than the A'alue given by the first lot, and as there was less roimding of the cor- POLYMORPHIC TRANSFORMATIONS OF SOLIDS. 93 ners, this sample is evidently purer, although it was not perfectly pure, because there was perceptible rounding for this also. The agreement of the two samples at high pressures is better than at atmospheric pressure; doubtless considerable of the impurity shown by the first sample at low pressures was due to a gradual decomposition, and did 'not affect the values at high pressures, which were determined nearer to the time of preparation of the sample. The decomposition phenomena are of special interest; apparently there are two kinds of decomposition. The first sample showed both kinds. The first kind of decomposition is one with increasing volume and took place in notable degree for the first sample at 176°. This is evidently an effect of the high temperature, and not of the pressure. The second decomposition is one with decreasing volume, and was shown by the first sample at 200° and about 7000 kgm. Because of the decreasing volume, we can tell that this is probably a decomposition brought about by the high pressure and not by the high temperature. Such decompositions are not usual; this is the first case of it I have found. Apparently the result of this decomposition with decreasing volume is complete resolution into carbon and chlorine. The second sample showed only the first kind of decomposition with increasing volume at 145°; at this temperature the decomposition ran very nearly to completion, only a small trace of the II-III transition being found on decreasing pressure. The apparatus was cooled and taken apart after the decomposition at 146°. There was some free bromine, but apparently no free carbon. The larger part of the mass was a white solid, but several cu. cm. of a colorless liquid had been formed. After this run, the steel shell with the CBr4 was sealed into a glass tube to preserA'e the decomposition products. The colorless liquid was unstable; after several weeks it had entirely disappeared, its place being taken by hexagonal crystalline plates. This is not the crystalline form of either of the forms of CBr4. The chemistry of this decomposition might prove an interesting subject for investi- gation. As regards the accuracy of the results given above, the jj-t values are probably pretty good, but the AT' values are too high on the I-III and the II-III curves. This is the result of the slow decomposition, even at low temperatures. It is known that at atmospheric pressure C'Br4 decomposes at 100°, so that the decomposition found here is not surprising. As a result of the high AF values, the AH values are also at fault. There can be no question of this, because they do not check at the triple point. There seemed to be no way of adjusting the ^■arious 94 BRIDGMAN. quantities easily so that the A// values shoukl check, and the values of the tables are left with this error. This is the only substance for which the values have not been so adjusted as to check at the triple point. In the case of other substances the amount of necessary adjust- ment is so slight and is indicated so unambiguously, that the probable accuracy of the individual curves can only be increased by demanding that the necessary conditions at the triple point be accurately satisfied ; but here the error is great enough so that the precise direction in which the adjustment should be made is not indicated unambiguously and hence the adjusted values would be likely to have as much or more error than the unadjusted ones. The amount of adjustment is not so very large even here; an increase of one value of Ai/ by 3.5 % and a decrease of another by the same amount would satisfy the conditions. The experimental results are shown in Figures 18 and 19, the com- puted values of hH and I^E in Figure 20, and the numerical values are collected in Table IX. In the figures the values of A I' are not given which are obviously in error because of decomposition. In the phase diagram the initial direction of the melting curve is indicated. This is taken from the work of Wahl, who found that because of decomposi- tion, the melting curve could be followed only about 10° above the normal melting point. The initial slope of the melting curve is un- usually high ; in this work no attempt was made to obtain any of the data of the melting curve. Other results for comparison are as follows. For the transition point at atmospheric pressure Schwarz ^ gives 46.1° by an optical method, and Rothmond ^^ gives 46.91° by a thermometric method. Wahl ^^ seems to have been the only previous experimenter on the effects of pressure. His transition curve extrapolates to 47.3° as the transition point at 1 kgm. The extrapolated value found above from the second purer lot was 46.2°. The specimens of Rothmund and Wahl were, therefore, probably purer than those used above; in particular, Wahl's specimen seems to have been purified with very great care. Wahl also followed the transition line I-II up to 1500 kgm. At 1500 his transition temperature is 2.8° higher than mine, whereas it starts 1.1° higher at atmospheric pressure. The difference of slope, which amounts to about 3.5%, may be an effect of impurity. One point investigated with especial care by Wahl was that of the 11 V. Rothmund, ZS. phys. Chem. 24, 705-720 (1897). 12 \V. Wahl, Trans. Roy. Soc. 212, 117-148 (1912). L/ f Hi/ ■/ 180' 160" 140' 5 120' g_100' E ^ 80' 60° "^^0 1 2 3 4 5 Pressure, kgm./cm.^ x 10^ Carbon Tetrabromide Figure 18. Carbon Tetrabromide. The observed equilibrium tempera- tures and pressures. 0 12 3 4 5 Pressure, kgm./cm.^ x 10* Carbon Tetrabromide Figure 19. Carbon Tetrabromide. The observed differences of vohune between the several phases. 0 12 3 4 5 Pressure, kgm./cm.^ x ID* Carbon Tetrabromide Figure 20. Carbon Tetrabromide. The computed values of the heats of transition and the changes of internal energy. 96 BRIDGMAN. breadth of the band of indifference between the two phases. He found it to vary from 30 to 55 kgm., not showing any regular dependence on pressure or temperature. The breadth of the band found by me above TABLE IX. Carbon Tetrabromide. Pressure Temperature aV cm^/gm. dt dp Latent Heat kgm.m./gm. Change of Energy kgm.m./gm. I-II. 1 46°. 2 .0205 .0305 2.15 2.15 500 61 .4 192 " 2.10 2.00 1000 76 .7 179 " 2.05 1.87 1500 91 .9 167 a 2.00 1.75 2000 107 .2 155 u 1.93 1.62 I-III. 2250 119°. 5 .0029 . 1054 .108 .044 2500 145 .8 29 u .115 .043 2750 172 .2 29 u .123 .043 II-III. 2000 108°. 5 .0123 .0238 1.97 1.72 3000 132 .3 110 K 1.87 1.54 4000 156 .1 097 " 1.75 1.36 5000 179 .9 085 (( 1.62 1.19 Triple Point. 2180 112°. 6 varied from 43 to 58 kgm. at the higher pressures, essentially the same as that found by Wahl. At 100 kgm., however, lower than any of the measurements of Wahl, the band was only 11 kgm. wide. Rothmund POLYMORPHIC TRANSFORMATIONS OF SOLIDS. 97 has also remarked on the sharpness of the transition at atmospheric pressure. The essential agreement of Wahl's and my results at high pressures for the width of the band of indifference, by methods entirely different, would seem to indicate some real physical significance in the absolute values found by us, characteristic of the substance, and not a property of the particular form of apparatus. The width of the II-III band was distinctly greater than that of the I-II band, being from 70 to 100 kgm., while that of the I-IIT band was less, 11 to 33 kgm. There seem to be no previous determinations of either AV or AH for the transition. The data do not justify any attempt at estimating Aa, A^, or ACp. Silver Iodide. — This substance was Eimer and Amend's purest, used without further purification. Only one filling of the apparatus was necessary in getting the high pressure points, and one for the low J1234567 Pressure, kgm./cm.^ x 10* Silver Iodide Figure 21. Silver Iodide. The observed equilibrium temperatures and pressures. pressure points. The transitions ran cleanly, and gave no such trouble from lag as was found by Tammann.^^ The dry powder was 13 G. Tammann, "K. and S," p. 302, and ZS. f. Phys. Chem. 75, 733-762 (1911). 98 BRIDGMAN. compressed by ramming into an open steel shell, and pressure trans- mitted directly to it by kerosene. Apparently there was no soluble impurity present, because no rounding of the corners of the isothermals was ever detected. The quantities used were 85 and 103 gm. .025!^-^5"^'T^"^'¥^'«f: B be .020 .000 0 12 3 4 5 6 Pressure, kgm./cm.^ x 10^ Silver Iodide Figure 22. Silver Iodide. The observed differences of volume between the several phases. Id 2.0 012345G7 Pressure, kjjrsi./cm.' x 10^ Silver Iodide Figure 23. Silver Iodide. The computed values of the heats of transition and the changes of internal energy. The width of the band of indifference was not nearly so great for this substance as was expected. It was greatest on the I-II curve, being here nearly constant at 80 kgm., least on the I-III curve, where it varied from 16 to 25, and intermediate on the II-III curve, where it varied from 20 to 60, being greatest at the lowest pressures. The POLTMORPHIC TRANSFORMATIONS OF SOLIDS. 99 TABLE X. Silver Iodide. Pressure Temperature AV cm3./gm. dt dp Latent Heat kgm.m./gm Change of Energy kgm.m./gm. I-II. 1 144°. 6 .00860 -.0146 2.460 2.460 1000 129 .6 914 155 2.374 2.465 2000 113 .5 964 167 2.232 2.425 3000 96 .1 1020 182 2.067 2.371 I-III. 3000 104°. 8 .01390 .02904 1.808 1.391 4000 134 .0 1286 2947 1.776 1.262 5000 163 .9 1130 3060 1.614 1.049 6000 195 .4 0936 3260 1.339 .777 6500 212 .3 0830 34- 1.188 .649 II-III. 3000 30°. 0 .0239 -.415 .175 .892 2951 50 .0 2395 .390 .198 .905 2897 70 .0 240 .347 .237 .932 2835 90 .0 2408 .297 .294 .976 2800 100 .0 2412 .273 .329 1.005 Triple Point. I-II .01010 -.01787 2.105 2.39 2810 99°. 4 I-III .01402 .0290 1.801 1.41 II-III .02412 -.276 .325 1.00 100 BRIDGMAN. reactions on all three curves were fairly rapid. The total time re- quired to shut the pressure within the limits above, both from above and below, averaged about 20 minutes on the I-III curve, one hour on the I-II curve, and 20 minutes on the II-III curve, except at 30°, which required 50 minutes. In no case was much lag observed in going across any transition line in either direction. No careful meas- urements of this were made; at 107° the reaction from III to I ran on crossing the transition line by only 45 kgm. The experimental results are shown in Figures 21 and 22, the com- puted values of A// and AE in Figure 23, and the numerical results in Table X. None of the experimental points have been discarded. It should be noticed that although in the table AT is given as increasing with increasing temperature on the II-III line, the increase is slight, and perhaps within the limits of error. No theoretical deductions should be based on this increase of AV. A notable feature of the phase diagram is the downward convexity of the I-III curve; such curvature is never shown by a melting curve, but there are several other examples of it for a solid transition. There are several previous determinations of the constants of the transition. For the temperature of transition there is 145° by Kohl- rausch ^* by an electrical method, 142° by Rodwell ^^ from the dis- continuity in length, Schwarz ^ gives 146.9° on heating and 145.4° on cooling, by the sudden change of color, Bellati and Romanese ^^ give the transition point as approximately 150°, Mallard and Le Chatelier ^^ as 146°, and Steger ^^ 147°. Tammann ^^ extrapolates to 144.2° from his measurements at higher pressures. The extrapo- lated value from the above work is 144.6°. For the latent heat of transformation Mallard and LeChatelier give 6.8 cal. and Bellati and Romanese 6.25 cal. at 150° against 5.77 cal. (2.46 kgm. cm.) found by calculation from Clapeyron's equation above. There seem to be no direct measurements of the change of volume of the transition at atmospheric pressure except the value 0.0028 cm^ per gm. which I have deduced with rather questionable assumptions from data of Rodwell. The effect of pressure on the transition point was first investigated by Mallard and Le Chatelier, and has later been made the subject 14 W. Kohh-ausoh, Wied. Ann. 17, 642 (1882). 15 G. F. Rodwell, Proc. Roy. See. 25, 280-291 (1876-77), and Trans. Roy. Soc. 173, 1125-1168 (1882). 16 M. Bellati and R. Romanese, R. Inst. Ven. 1, 1043-1069 (1882-83). 17 E. Mallard and H. Le Chatelier, C. R. 97, 102-105 (1883). 18 A. Steger, ZS. phys. Chem. 43, 595-628 (1903). POLYMORPHIC TRANSFORMATIONS OF SOLIDS. 101 of two papers by Tammann. Reference is made to Tanimann's paper for a historical account of our knowledge of the relation of the two phases. It is to be noticed that my notation is different from that of Tammann. I have employed consistently the numeral I for the high temperature form, whereas here Tammann applies it to the low tem- perature form. The special interest in the pressure effects on this substance is because the forms I and II are of the rather unusual ice type, the form at the higher temperature having the smaller volume. Mallard and Le Chatelier found at 20° and 2475 kgm. a large decrease of volume; one would not expect a transition here from the initial trend of the curve at 1 kgm. This led Tammann to take up the question, and he showed in his first paper that the effects were doubt- less due to a third modification. He was not able to obtain any very great amount of information as to this third modification, however, because of the large amount of lag which the reaction showed. He took up the matter again later, however, and was able to get nuK-h better numerical results. The data of the later paper of Tammann cover about the same range as the present paper as far as the I-II and the II-III curves go. Tammann, however, followed the II-III curve carefully to a temperature 10° lower than that used here, and made measurements connected with the subcooliag, etc., down to the tem- perature of liquid air. The only extension of range in this paper is on the I-III curve, which is followed here to 200°, whereas Tammann did not determine accurately any points on it, but showed only that its slope was positive, and calculated its probable value. Tammann in his second paper devoted a good deal of attention to the question of lag. He found that at 90° the equilibrium pressure could be shut between the limits 2956 and 2957 kgm.; at 60° the limits were 2980 and 3040, and at 18°, 2760 and 3026. The limits within which spon- taneous formation of the nuclei took place were considerably wider than this. In these experiments, on the other hand, the effect of lag was found to be much less. As already stated, the greatest width of the band of indifference of the II-III curve is 60 kgm. at 30°. The reason for this difference between our two experiments is not clear; possibly the kerosene had a catalyzing effect on the reaction which the heavy oil used by Tammann did not have, or it may be that this was in some way an effect of the much higher pressure to which the Agl was subjected in these experiments. Before any readings had been taken, the Agl had been subjected to 12000 kgm. at room temperature. It is to be noticed, however, that Tammann in his first set of experi- ments reported in " Kristallisieren und Schmelzen" found a nmch 102 BRIDGMAN. greater lag than he did in his later experiments. He was not able to repeat the early experiments, and was able to explain them only by supposing that the particular specimen of machine oil which he had used to transmit pressure had some peculiar property which the later specimens did not have. A comparison of the numerical values of Tammann's later paper and the values found above gives the following results. Tammann found the change of volume I-II between 1000 kgm. and the triple point to be constant within the limits of error at 0.0095, the probable error being 5%. I found above that AT' increases on the transition line from 0.0086 at atmospheric pressure to 0.00102 at 3000. Tammann's triple point is 2940 kgm. and 100°, against 2810 and 99.4° above. Tammann found the I-II line straight within the limits of error, with a slope of 0.01473; I found the slope to increase in absolute value from 0.0146 at 1 kgm. to 0.0182 at 3000 kgm. Tammann calculates the slope of the I-III line at the triple point to be 0.0195; I find above 0.0140. The slope of the II-III line Tammann gives as —0.75, and I find it to vary from —.415 to —.273. At the triple point Tammann gives for the change of volume (uncorrected) of II-III, 0.0205, and for I-III, 0.0115, against 0.0241 and 0.0140 above (corrected). Tam- mann, however, had to leave an unexplained discrepancy of 7-10% in the AV values for II-III, the change of volume being different for the two directions of reaction. For the values of Aa, A/S, and ACp, two methods of estimation are available. At the triple point the various quantities may be com- puted, and there are also experimental determinations of the difference of compressibility. Neither of these methods gives satisfactory results, however. The calculations at the triple point are uncertain because the variation of AV along the II-III line is so slight that even its sign is uncertain. The experimental values of the difl'erence of compressibility are irregular and uncertain. We can, however, make the following rough statements at the triple point. I is more com- pressible than II, the difference being of the order of 0.065. The difference of expansion between I and II is very small and is uncertain as to sign. The specific heat of I is greater than that of II, and the difference is of the order of 0.3 kgm. cm. per gm. Ill is more compressi- ble than I, and the difference is fairly large, of the order of O.O55, but there is a good deal of uncertainty as to the numerical value of this difference. Ill is less expansible than I, the difference being of the order of O.O44 and III has a smaller specific heat than I, of the order of 0.3. From these values, of course, the differences between II and III may be found immediately by a subtraction. POLYMORPHIC TRANSFORMATIONS OF SOLIDS. 103 In addition to these very rough values at the triple point, there have been several other determinations at atmospheric pressure. It has been shown directly that the relations of I and II are abnormal both in regard to expansion and specific heat. The form II has a negative coefficient of expansion. We have no accurate measurements of it at the transition point. The best are by Fizeau ^^ on cr^-stals. He found at 40° for large crystals the average contraction of 0.0648 and for the precipitated salt b}^ a dilatometric method 0.0674 cm^./gm. (I have taken the density at room temperature as 5.67 from data of Rodwell ^^). Fizeau and Rodwell have both shown, however, that the coefficient increases rapidly with rise of temperature. Rodwell's data would give an average coefficient between 70° and 150° of — O.OaSSS. From data of Rodwell we also find that the mean expansion of I between 150° and 450° is about O.O55. These measurements of the expansion are too uncertain to justify us in making calculations of the difference. The data for specific heat are somewhat more concordant. Two observers agree in finding that I has a smaller specific heat than II. At the transition point Bellati and Romanese ^^ give for the specific heat of I 0.0577 cal. and for that of II 0.0654, making a difterence of 0.0077 cal. Mallard and Le Chatelier ^^ give for the specific heat of I between 154° and 347° 0.055, and for II between 20° and 127°, 0.059. Since however, they estimate the accuracy of the individual measurements as only 3%, we evidently cannot place much reliance on their difference. If we assume the value of Bellati and Romanese for the difference and combine with the values given above for —. — (= —0.0653) and —. — (= 0.005), dp dp we find A/3 = — .00002, whereas we know A/3 to be positive. If A/3 is to be positive with the above values for —, — and , — , the difference dp dp of specific heats must be at least twice as great as it is. We can get a much better value for Aa. If we assume A/3 = 0, and it evidently is very small, we find Aa is of the order of 0.065, the same as found above at the triple point. We may accept this value with some confidence therefore. Combined with the value of Richards and Jones ^° for the compressibility of the ordinary variety, this means that I is about twice as compressible as II, although it has the smaller volume. The 19 Fizeau, Pogg. Ann. 132, 292 (1867). 20 T. W. Richards and G. Jones, Jour. Amer. Chem. Soc. 31, 158-191 (1909). 104 BKIDGMAN. indications are that the difference of specific heat between I and II changes sign along the transition line; at atmospheric pressure I has the smaller specific heat, but at the triple point that of II is the smaller. It must be insisted, however, that all these values are exceedinglv , dAV , clAH rough ; and are not given at all accuratelv bv the above dp dp data. No other modifications were found to 12000 kgm. at either 20° or 200°. The form of I is cubic and that of II hexagonal, but the two forms are very little different. Mercuric Iodide. This substance is the only one found so far which shows a maximum transition temperature between two solids, a result both unexpected and important. A great deal of time was spent on this substance in order to exhaust every other possible ex- 23456789 Pressure, kgm. /cm. ^ x lO"* Mercuric Iodide 10 U Figure 24. Mercuric Iodide. The observed equilibrium temperatures and pressures. It was not possible for this substance to shut the equilibrium conditions within narrow limits, and in the figure the limits of the reaction are shown. The text should be consulted for further particulars. planation of the effects, and to establish this result beyond doubt. It is particularly difficult to get accurate results for this substance, because of the unusually wide region of indifference. The interpreta- tion of the earlier results was also obscured liy occasional leaks through a minute flaw in the cylinder; this flaw was due to amalgamation POLYMORPHIC TRANSFORMATIONS OF SOLIDS. 105 produced by mercury set free by decomposition, in the presence of steel, either of the Hgl2 of this experiment, or of other mercury salts used in previous experiments. ^: The Hglo was obtained from Eimer and Amend, "c. p." and was used without further purification. For the first runs it was hammered cold into a thin steel shell, and pressure transmitted directly to it by kerosene. After this first run small transverse holes were drilled through the sides of the shell to facilitate the reaction, and cut down the width of the band of indifPerence, but no effect from this could be observed. For the later runs, the Hglo was initially melted into the steel shells and then the lateral holes drilled. Xo effect of the greater initial compactness in increasing the reaction velocity was tobe .004 0 -.002 > .004^^ 12345678 Pressure, kgm./cm.^ x 10 Mercuric Iodide 9 10 Figure 2.5. Mercuric Iodide. The observed differences of volume. observed. Three different samples were used, from 86 to 100 gm. in amount, but each of these samples was used for a number of settings up of the apparatus, and one was used for both high and low pressure measurements. The effect of pressure is to compact the Hgl2 into a mineral like mass, without the slightest tendency to dissolve in the kerosene. After long use a slight amount of decomposition with setting free of minute globules of metallic mercury could be detected where the surface of the Hglo comes in contact with the steel. This is to be expected, because iron will deposit mercury from mercury salts, but the decomposition was too slight to produce a perceptible change in the total volume of the reacting ciuantity of Hgl2. In view of the importance of the maximum temperature exhibited by this substance, a detailed description of the experiment will be given. When I started to investigate this substance, all that I knew about it was that it had a transition at atmospheric pressure at 127°, 106 BRIDGMAN. the red form changing to yellow at this temperature, and that presum- ably the yellow had the larger volume, so that the transition line would rise with increasing pressure. It will be easier to follow the descrip- tion of the many runs made with this substance to turn at once to the phase diagram of Figure 24. The first run was made at room tempera- ture, and no transition was found to 12000. Temperature was then raised to 150°, and transition found at about 8000 with very wide limits of indifference. Other points were then found at 108°, 141°, 159°; the pressure limits are indicated in the diagram. Temperature was then raised to 200°, and no trace whatever of the transition could be found between 12000 and 700. Pressure was then raised to 6800, and temperature reduced to 160°, but no transition could be found down 123456789 Pressure, kgm./cm.^ x 10* Mercuric '"dide 10 Figure 26. Mercuric Iodide. The computed values of the heat of transi- tion and the change of internal energy to 700. This was difficult to explain. What had become of the transition found at lower temperature and also of that at atmospheric pressure at 127°? On starting again at room temperature, raising pressure to 8400 and temperature to 162°, the transition was found again where expected, but on raising temperature to 180°, the transi- tion disappeared. Another run at 185°, however, indicated a transi- tion at 9600, but this was later traced to leak. Several other curious effects between 160° and 200°, and 8000 to 12000 kgm. were also later traced to leak, but were at first thought to be due to another transition. At this stage the most plausible explanation seemed to be a rising transition curve from 1 kgm. and 127° to about 7500 and 160°, and here a triple point with the curve found below 160°, and the supposed curve above 160°. To test this, temperature was raised from 126° POLYMORPHIC TRANSFORMATIONS OF SOLIDS. 107 to 153° at 2200 kgm., but no transition line was found. Again the question, what had become of the line starting at 127°? Very shortly after this the flaw in the cylinder grew to visible size, a new cylinder was made, the mysterious points found before at high temperatures and pressures disappeared, and the correct explanation of the effects was found. In brief, the curious effects are due to a transition curve with a maximum, combined with a region of indifference of unusual shape. This region is indicated by the circles in Figure 24, which mark the TABLE XL Mercuric Iodide. Pressiu-e Temperatiire AV cm^./gm. dt dp Latent Heat kgm.m./gm. Change of Energy kgm.m./gm. 1 127°. 0 .00342 .0267 .513 .513 1000 149 .8 217 187 .491 .469 2000 165 .2 127 122 .456 .431 3000 174 .9 065 069 .420 .400 4000 179 .9 024 028 .390 .380 5000 181 .2 -008 -010 .365 .369 6000 178 .2 -045 -059 .345 .372 7000 169 .7 -100 -134 .330 .400 8000 152 .4 -175 -234 .318 .458 9000 122 .3 -270 -361 .296 .539 9500 102 .4 -325 -427 .286 .594 10000 i 79 .4 -390 -495 .278 .668 experimentally determined limits of this region; the transition curve is drawn through the mean points. We see that the region of indiffer- ence is of different shape on the two branches of the curve. If Hgl2 is heated at atmospheric pressure, the red changes to yellow pretty sharply on passing through 127°, but on cooling again, the yellow may persist for several hours, even at room temperature. That is, the red cannot be much superheated, but the yellow can be very considerably subcooled. With increasing pressure and temperature, the shape of the region changes as indicated, until on the descending branch of the transition line, the equilibrium may be overpassed in either direction, but to a very much less extent than at atmospheric pressure. This 108 BRIDGMAN. shape of the indifferent region makes it impossible to obtain equilib- rium values on the ascending branch by shutting the pressure within two limits approached from above and below. The best that could be done here was to use the method of varying temperature at constant volume, varying the temperature by small steps, and relying on the relatively small amount of superheating to get a fairly good value for the transition temperature. The change of volume can be determined from these data by methods already used. Evidently the failure of this method to give any results at 2250 kgm. on the first trial men- tioned above was because the temperature was not raised sufficiently high. The shape of the region of indifference gives rise to curious effects in the neighborhood of the maximum. Thus at room temperature, the pressure was once raised to 6500, the temperature was then raised to 165°, and pressure was raised, with a transition at the expected place, but on releasing pressure no reverse transition was found on crossing either branch of the curve. The location of the indifferent region evidently explains this. On another occasion, temperature was raised at 6600 completely through the indifferent band, the transition running as to be expected on emerging at 180°, but on lowering tem- perature to 175° and reducing pressure, no reverse transition could be foimd. The existence of the maximum was further established by the following runs. First, temperature was raised to 190° at about 2500 kgm., giving the regular transition point at about 165°. At 190°, pressure was increased to 10,000 with no transition; at 10,000 tem- perature Avas lowered to 120° with no transition, and at 120° the tran- sition was found again at the point to be expected on lowering pressure. Again, starting at 120° and 9000 kgm., temperature was raised to 195° with the transition at the expected point; at 195° pressure was lowered to 4600 with no transition; at 4600 kgm. temperature was lowered in steps of 10° to 120°, passing a transition with almost im- perceptibly small change of volume at 134°, which therefore marks the lower limit of the indifferent region, and at 120°, pressure was raised again, the volume returning to its initial value when the initial pres- sure was reached. To make still more certain that the initial con- ditions had been recovered, pressure was increased beyond its initial value at 120°, and the regular transition found both from above and below. That is, we have here the experiment of describing a complete circuit and coming back to the starting point, with practically only one discontinuous change of volume. If the pressure of decreasing POLYMORPHIC TRANSFORMATIONS OF SOLIDS. 109 temperature had been chosen a Kttle higher than 4600, so as to exactly equal the maximum, no trace of a second transition would have been found. The existence of the maximum involves properties that at first sight seem so unnatural, that perhaps this elaborate procedure to establish it has been worth while. A maximum would seem unlikely because at this point the change of volume is zero; at lower pressure the high temperature form with the larger volume is the more com- pressible, which is of course natural, but at pressures higher than the maximum, the high temperature form has the smaller volume and a higher compressibility than the low temperature phase. Now this peculiar relation of the compressibilities was shown by experiment to actually exist. On the falling branch of the curve, the compressi- bility of the phase with the smaller volume was the higher. This could be determined from the difference of the slope of the volume isothermals above and lielow the transition. Some attempt will be made later to find the magnitude of this difl'erence; it cannot be determined accurately, but the fact is absolutely beyond question. This curious relation of the compressibilities is not an absolutely new thing; it is shown by ice and water, but there the explanation is proba- bly to be found in the greater rigidity of the crystalline framework as contrasted with the structureless liquid. Regarding the accuracy of the results, the values given here are probably more irregular than those for any other substance which does not decompose. The explanation, of course, is the great width of the band of indifference, and the sluggishness of the reaction. The irregu- larities in the AV values are considerably greater than in the values for pressure and temperature. The values of AV probably have their greatest error on the rising branch of the curve. The AV curve as shown was so drawn as to nearly pass through the directly determined value at low pressure, and also to pass through zero at the pressure of the maximum temperature. In spite of the abnormally wide band of indifference, the pressure approached a stationary value after the reaction had once started as rapidly for this as for many other substances; the usual time for reaching a stationary pressure being only about 15 minutes. This suggests most strongly that there is no direct connection between the width of the band of indifference and the velocity of reaction; there seems every reason to think that even in infinite time the apparent width of the band of indift'erence would not become materially less. There are no other values at high pressures for comparison. The 110 BRIDGMAN. value 127° was adopted for the transition point at atmospheric pressure. This could be safely done without making a redetermination of it for this special sample, since this value is certainly as accurate as the other values at high pressures. There are a number of measurements at atmospheric pressure. For the transition point we have 127.2° by Reinders,* 127° by Steger/^ 130° by Guinchant,22 and Schwarz^ gives the limits 124.5° to 130.2° by a thermal method, and 126.3° to 129.3° by an optical method. There is one value of the change of volume, 0.00135 cm^ per gm. by Reinders,^^ against 0.00342 found above. It would seem that the larger would be more likely to be correct. For the latent heat there is the value 3.0 cal per gm. by Berthelot, as quoted by Varet,^^ and 1.53 cal. by Guinchant,^^ as against 1.2 cal. calculated above. If a smaller value of the change of volume were used, the calculated value of the latent heat would be smaller. We have already seen that at a maximum point the relations be- tween the compressibilities and thermal expansions of the two phases may be deduced from the curves for Av and All, the specific heat dis- appearing from the relations at the maximum point. These values at the maximum are included in the follow^ing. It has also been stated that direct measurements of the difference of the compressibility showed that above the maximum the phase of smaller volume has the larger compressibility. The numerical value of the difference could not, however, be determined with much accuracy. Within the limits of accuracy, Aa would seem to be constant on the falling branch at O.OeSS cm.^ per gm. per kgm. per cm^. Seven determinations of this quantity were made; the deviations of the several values from 0.0633 are; +35%, -15%, 0%, -25%, 0%, -60%, and 0%. The devia- tions are large, but still this is as accurately as one can hope to get quantities as small as these by a method like the above. This value is also not inconsistent with that deduced from the AV curve at the maximum point, where we have already seen that Ace is determined by the AV curve alone. It was not possible to make any measure- ments of Aa on the rising branch of the curve. If, however, we make the assumption that Aa is constant on both branches of the curve (an assumption which is very doubtful) we may compute A/? and ACp on the two branches. These values are given in Table XII, directly 21 W. Reinders, ZS. phys. Chem. 32, 494-536 (1900). 22 J. Guinchant, C. R. 145, 68-70 (1907). 23 R. Varet, Ann. Chim. Phvs. 8, 79-141 (1896). POLYMORPHIC TRANSFORMATIONS OF SOLIDS. Ill as computed, without any adjustment. These figures would demand that Hgl2 be abnormal in many respects. On the rising branch of the curve the yellow modification is less expansible than the red, but above 6000, on the descending branch, the yellow becomes more expansible. The place at which AjS changes sign need not necessarily coincide with the maximum. In general, however, the phase with the smaller volume, whether at the higher or lower pressure seems to have the higher thermal expansion. AC'p calculated in accordance with the assumption above as to Aa comes out negative, the yellow having the smaller specific heat. But it has already been shown that ACp is particularly sensitive to errors in the assumed value for Aa, and in TABLE XII. Mercuric Iodide. Pressure A/3 ACp 1 -.0438 -1.50 2000 -.0436 -1.73 6000 + .O42O -1.31 8000 + .O423 -0.43 10000 + .O42I -0.18 this case there is a direct determination of the specific heat by Guin- chant, from which it appears that the yellow at atmospheric pressure has the larger heat, the difference being 0.0040 cal. At higher pres- sures, ACp calculated with the assumption as to Aa, decreases rapidly. Combining this with the experimental value at atmospheric pressure, we are probably justified in the conclusion that the difference of speci- fic heats becomes less at high pressures. The sign of A^S found above was verified by direct measurement on the rising branch of the curve, but the direct measurements did not give regular numerical results. Phenol. This was Kahlbaum's purest, further purified by crystal- lization at constant temperature in the thermostat. Two lots were purified; the first was pure white in color, and the second was very 112 BRIDGMAN. slightly pinkish, but it had a melting point 0.02° higher than that of the first lot, and a trifle higher than any previously recorded for phenol. The purity is apparently, therefore, as great as can be obtained by ordinary means. The pinkish specimen was used for the high pressure determinations. This specimen was not perfectly pure, however, because some rounding of the corners could be detected. It was melted into the inverted steel shell and pressure transmitted to it by mercury. Four different fillings of the apparatus were used; one for the point at low pressures and the other three for the points at high pressures. The quantities used varied from 13.5 to 17.5 gm. 23456789 10 1112 Pressure, kgm./cm.^ x 10^ Phenol Figure 27. Phenol. The observed equilibrium temperatures and pressures. The experimental values of pressure and temperature are shown in Figure 27, the experimental values of AV in Figure 28, the computed values of AH and AE in Figure 29, and the numerical values are col- lected in Table XIII. No points have been discarded. ( It has been known for some time that phenol has two modifications. These were discovered by Tammann,^* who has written two papers on 24 G. Tammami, 11). 'K. und S.," 308, and ZS. phys. Chem. 75, 75-80 (1910- POLYMORPHIC TRANSFORAIATIONS OF SOLIDS. 113 ---f' .ts^.^:i:; 23456789 10 Pressure, kgm./cm ^ \ 10^ Phenol 11 12 Figure 28. Phenol. The observed differences of vokime between the several phases. 123456789 10 1112 Pressure, kgm./cm.^ x 10^ Phenol Figure 29. Phenol. The computed values of the heats of transition and the change of internal energy. TABLE XIII. Phenol. Pressure Temperature AV cm3./gm. dt dp Latent Heat kgm m./gm. Change of Energy kgm.m./gm. Liquid — I. 1 40°. 87 .0567 .0140 12.70 12.70 500 47 .5 471 125 12.12 11.88 1000 53 .4 395 111 11.62 11.22 1500 58 .7 335 099 11.15 10.73 2000 63 .3 280 088 10.72 10.16 I-II. 1327 0°.0 .0593 .0843 1.926 1.139 1565 20 .0 580 a 2.021 1.113 1803 40 .0 568 u 2.113 1.089 2039 60 .0 556 a 2.202 1.068 Liquid — II. 2000 62°. 1 .0831 .02218 12.56 10.90 3000 82 .1 767 1868 14.58 12.28 4000 99 .8 714 1684 15.81 12.95 5000 115 .9 669 1553 16.75 13.40 6000 131 .0 630 1461 17.42 13.64 7000 145 .2 595 1389 17.92 13.76 8000 158 .8 564 1334 18.26 13.75 9000 171 .9 538 1289 18.58 13.74 10000 184 .6 513 1254 18.73 13.60 11000 197 .0 489 1228 18.72 13.34 12000 209 .2 468 1211 18.64 13.02 Triple Point. L-I .0270 .00861 10.57 10.01 2085 64°. 0 L-II .0825 .02171 12.79 11.07 I-II .0555 .0843 2.221 1.063 POLYMORPHIC TRANSFORMATIONS OF SOLIDS. 115 it. The pressure range of the results given here is considerably greater than that of Tammann, but Tammann's temperature range on the I-II curve is more extensive. Tammann's second paper is almost exclu- sively occupied with the lag phenomena on the I-II curve. He finds that the band of indifference grows rapidly wider at the lower tempera- tures, becoming so wide at liquid air temperatures that II may be realized at atmospheric pressure. In this work, the same qualitative behavior of the band of indifference was found. At 0° the width of the band was 180 kgm.; at 25°, 80 kgm.; at 35°, 50 kgm.; at 50°, 20 kgm.; and between 57° and 62° no difference could be detected in the pressures reached from above and below. From one of Tam- mann's diagrams I should estimate the width of his band to be about 350 kgm. at -10°, 200 kgm. at 0°, 50 kgm. at 10° and above 30° the scale of his diagram does not allow an estimate. Apparently the width of his band decreases with rising temperature more rapidly than mine. This emphasizes that too little is known at present of the various factors which determine the width of the band to allow us to attach significance to the absolute width of the band when measured under different conditions. Tammann's transition line I-II showed one other remarkable feature in that at the upper end it rises much more rapidly than it does at the lower temperatures. Tammann's sub- stance was somewhat impure, as he himself recognized, so that it was ques- tionable how much of the curvature could be laid to the account of the impurity, but Tammann suggested that it might be found that for the pure substance the transition line would pass through a verti- cal position, and consequently the latent heat through a zero value, before reach- ing the triple point. In the effort to settle this question, I made careful meas- urements of the transition coordinates between 50° and the triple point; these are reproduced on a much enlarged scale in Figure 30. It is seen that no point departs from the straight line (this is the same straight line which passes through the low temperature points) by more than 3 kgm., which is the limit 50 1900 2100 Pressure, kgm. /cm/ Phenol Figure 30. Phenol. Points on the upper end of the tran- sition line I-II on a very much enlarged scale. There is no curvature appreciable. 116 BRIDGMAN. of sensitiveness of the pressure measurements. We conclude, there- fore, that the transition Hne rises to the triple point with no percepti- ble change of direction, and that the curvature found by Tammann must have been an effect of the impurity. Tammann determined the phase diagram with two specimens, one somewhat purer than the other. In the following comparison of his results with those found here only the data of his pure sample will be used. His L-I curve is 0.3° lower than mine at atmospheric pressure, and 1.2° lower at 2000. At 20° his I-II line is 50 kgm. higher than mine, and 4 kgm. lower at 60°; this is an effect of the curA'ature found by Tammann. Above the triple point, his L-II curve runs lower than mine and with a smaller slope; at 2600 it is about 3° lower. He finds as the coordinates of the triple point 2196, kgm. and 62.8° against 2085 and 64.0° of mine. The discrepancies are at least in part due to impurity. Tammann's value for AT at atmospheric pressure, 0.0163, is much lower than mine; he recognizes that his value is too low, due to impurity. His value for AV, I-II, is also very much lower than that found here, 0.0315 against 0.0568 at 40°. Curiously, however, his value for AT, L-I, at the triple point is almost exactly that found here, 0.0273 against 0.0270. Besides Tammann's values at low pressures there are also a number of other values for comparison. Hulett ^^ has measured the effect of pressure on the melting point up to 300 kgm. He finds for the melting point 40.75° against 40.87° above, and for the initial slope 0.0149° per kgm. against 0.0140 above. It should be remarked that the value of Hulett for the melting point is sometimes misquoted as 41.11°, because of a misprint in the original paper by which the melting point under 25 kgm. is given twice, once as that at atmospheric pressure. Other values for the melting point are 40 to 41° by Calvert, quoted by Schiff,^^ and 42.5° to 43° by Behal and Choay ^^ for a specimen of phenol s>Tithesized by a new method. For the change of volume we have 0.054 to 0.051 by Heydweiller ^8 and 0.0532 by Block ^^ against 0.0567 above. For the latent heat, there is 24.93 cal. by Pettersson and Widman,^° and 26.9 cal. by Eykmann,^^ against 29.7 above. 25 G. A. Hulett, ZS. phys. Chem. 28, 629-762 (1899). 26 R. Schiff, Lieb. Ann. 223, 247-268 (1SS4). 27 A. Behal and E. Choay, Bull. Soc. Chim. Paris, (3) 11, 602-603 (1894). 28 A. Heydweiller, Ann. Phys. 61, 527-540 (1897). 29 H. Block, ZS. phys. Chem. 78, 384-426 (1912). 30 O. Pettersson and O. Widnian, Forh. Stock. (3) 36, 75-79 (1879^. 31 J. F. Eykman, ZS. phys. Chem. 4, 497-519 (1889). POLYMORPHIC TRANSFORMATIONS OF SOLIDS. 117 There are three different methods of attack on the values for the difference of compressibility, expansion, and specific heat for this substance. In the first place, a direct measurement of the difference of thermal expansion of the liquid and the solid I was possible at low pressures. The phenol was so pure that the rounding of the corners was very slight, and this value ought to be fairly good. Then direct measurements were made of the difference of compressibility of I and II along the transition line. These values are fairly good, espe- cially at the lower temperature. And finally, at the triple point, we have theoretically the means to calculate all the differences between the two phases. But actually the various derivatives are not known accurately enough to permit this calculation at the triple point. For instance, a rigorous solution of the six equations at this point demands that I is less compressible than II, whereas direct experiment shows that it is more compressible. The values follow. At 77 kgm. the liquid is more compressible, more expansible and has a greater specific heat than I, as follows: Aa = 0.000025 ] A^ = 0.00027 [ ACp = 6.9 [kgm. cm./gm.] J A/? was found by direct experiment and the others were computed. The value found by Block ^^ for A/3 is considerably less than that found here, 0.00017 against 0.00027. At the triple point it was established by direct experiment that I is more compressible than II. The difference of compressibility increases somewhat at the lower temperatures. At the triple point the difference is of the order of 0.000005. The difference of compres- sibility between liquid and I calculated from this is about O.O4I5. This value may probably be counted on with some certainty because it is not much affected by changes in the value of Aan. The differences in expansion between the two phases are, however, very susceptible to errors in the compressibility. One cannot conclude with certainty which of I or II is more expansible. The liquid is certainly more expansible than I and the difference is of the order of 0.0002. More than this cannot be stated. The differences of specific heats between the liquid and the two solids are also very susceptible to slight errors ; small changes in the constants may change the sign. The difference of specific heat between the two solids is not so sensitive, however, and it seems fairly certain that the specific heat of I is lower than that of II, perhaps a surprising result. The difference is small, and of the order of 0.3 kgm. cm. per gm. ii8 BRIDGMAN. No other modifications of the solid were found to 12000 at 25"^ to 12500 at 200°. or Urethan. This was obtained from Eimer and Amend, and further purified by crystalhzation at constant temperature in the thermostat. Beautiful, colorless, transparent crystals were obtained, columnar in form, some of them 2 or 3 inches long. The melting isothermals showed very little rounding of the corners, so the purity must have been fairly high. A remarkable property of the crystals is their great flexibility; they may be bent in the fingers like paraffine. It is possible that this may be due to twinning, as in the case of calcite; 160' 140" 120" |p!|liSr|aruJffipfli]Mxl^^ wm 3450789 10 11 12 Pressure, kgrn./cm.^ x 10* Urethan Figure 31. Urethan. The observed equilibrium temperatures and pressures. it would be an interesting subject for investigation. Two fillings of the apparatus were used. The first gave the low pressure point and nine points at higher pressures. After this, the splitting of the steel shell necessitated the using of another sample, with which twelve points at high pressures were obtained, completing the phase diagram. The quantities used were 17 and 13.5 gm. The urethan was melted into the inverted nickel steel shell and pressure transmitted to it by mercury. Two new solid modifications were found; these are so situated in the phase diagram that all the triple points and equilibrium lines POLYMORPHIC TRANSFORMATIONS OF SOLIDS. 119 possible for a four phase system actually exist; six equilibrium lines and three triple points. The experimental values for pressure and temperature are shown in Figure 31, the experimental values of AF in Figure 32, the computed values of AF and AE in Figure 33, and the 23456789 10 U12 Pressure, kgm./cm.^ x 10^ Urethan Figure 32. Urethan. The observed changes of volume. 0 1 2 3 4 5 6 7 8 9 10 11 12 Pressure, kgm./cm.^ x 10^ Urethan Figure 33. Urethan. The computed values of the heats of transition and the changes of internal energy. numerical values are given in Table XIV. It was necessary to dis- card two points in the immediate neighborhood of the L-I-II triple point, which were determined before it was known that there was a triple point here. The three melting curves are entirely normal. 120 BRIDGMAN. TABLE XIV. Urethan. Pressure Temperature cm'./gm. dt dp Latent Heat kgm.m./gm. Change of Energy kgm.m./gm. Liquid — L 1 47°. 90 .05990 .01105 17.39 17.39 500 53 .0 4620 0940 16.03 15.80 1000 57 .3 3762 0840 15.54 15.17 1500 61 .0 3182 0688 15.46 14.98 2000 64 .2 2774 0592 15.82 15.27 LlQUII ) — n. 2500 67°. 3 .03130 .00738 14.70 13.92 3000 70 .6 2384 603 13.59 12.87 3500 73 .4 2026 519 13.53 12.82 4000 75 .8 1880 461 14.23 13.48 Liquid — m. 4500 80°. 3 .06294 .01263 17.62 14.79 5000 86 .5 6110 1202 17.98 14.97 6000 98 .0 5724 1114 19.06 15.63 7000 108 .7 5354 1053 19.42 15.67 8000 119 .0 5000 1008 19.45 15.45 9000 128 .9 4670 0969 19.28 15.08 10000 138 .4 4350 0938 19.00 14.65 11000 147 .7 4056 0913 18.69 14.23 12000 156 .7 3780 0887 18.32 13.78 I-] n. 3160 0° .0572 .1040 1.504 -.309 3260 10 573 11 1.561 -.307 3350 20 574 1.617 -.306 POLYMORPHIC TRANSFORMATIONS OF SOLIDS. 121 I-II. 3280 30° .00930 -.0392 .719 1.024 3030 40 952 (( .760 1.048 2770 50 976 " .804 1.074 2520 60 999 u .849 1.101 II-III. 3470 30° .04800 .0612 2.376 .711 3640 40 4748 u 2.429 .701 3800 50 4696 a 2.479 .695 3960 60 4644 a 2.527 .687 4130 70 4592 u 2.574 .677 Triple Point, Liquid — I-I] L-I .02530 L-II .00530 16.19 15.59 2350 66°. 2 .03550 I-II .007866 15.31 14.48 .01020 - .0392 .883 .112 Triple Point, Liquid — II- III. L-II .01840 .00438 14.70 13.91 4230 76°. 8 L-III .06396 .01300 17.32 14.61 II-III 1 .04556 .0612 2.61 .68 Triple Point, I-II-III. I-II .00922 -.0392 .702 1.015 II-III 3400 25°. 5 .04820 I-III .0612 2.352 .713 .05742 .1040 1.650 -.303 122 BRIDGMAN. The transition line I-II is somewhat unusual in that A F increases with increasing temperature. The band of indifference shows interesting variations. On each of the three curves, I-III, I-II, and II-III, the band decreases in width with increasing temperature; but besides the general effect of tempera- ture, there is a very marked specific effect of the nature of the reacting phases. Both the bands I-II and II-III are broader at their lower ends than I-III is at its upper end, although the temperature of the lower end of the first two bands is the same as that of the upper end of the I-III band. The actual widths are as follows. I-III is 200 kgm. wide at 0°, and 60 kgm. wide at 25.2°; I-II is 250 kgm. wide at 35.3°, 75 kgm. at 50°, 20 kgm. at 58°; II-III is 60 kgm. wide at 35.3°, 20 kgm. at 50°, and 3 kgm. at 70.2°. Along with this change in the width of the band there goes a parallel change in the reaction velocity. No trouble was ever found in forcing the desired reaction to take place; but no observations were made of the precise amount of subcooling or superheating. It was observed once at 50° that I had to be car- ried 650 kgm. across the line before II appeared. No other transitions were found to 12700 kgm. at either room tem- perature or 153°. The efltect of pressure on the melting point of urethan has also been measured by Tammann,^^ who did not find any other modifications. Tammann's pressure range was to 2905 kgm. At the upper end of the melting curve he must therefore, have had the modification II without knowing it. Tammann remarks that below 1500 kgm. the melting curve is parabolic, but that above 1500 it is nearly linear; this linear- ity is evidently the effect of the overlooked transition. In fact, if one plots Tammann's points, he will find some evidence for the change in direction of the curve above the triple point, although the points are rather irregular. Tammann also looked for other modifications of the solid out to 3000 kgm. at 40°; if he had tried 60° instead, he would probably have found II. The melting point at atmospheric pressure was 48.14° against 47.90° above. The presumption from this alone would be, therefore, that Tammann's specimen was purer than mine. But at high pressures his curve drops below mine, being 1 .8° lower at the upper end, and he also states that the reaction was unusually slow for a melting reaction. Both of these points suggest impurity. In the work above, the melting set in as rapidly as for any normal liquid. It almost seems as if the design of Tammann's apparatus had been 32 G. Tammann, "K. und S.," p. 239. POLYMORPHIC TRANSFORMATIONS OF SOLIDS. 123 such as to permit the comtamination of the Hquid at high pressures. Tammann finds for AV at atmospheric pressure 0.0573, and at 1455 kgm. 0.0378, against the corresponding values 0.0599 and 0.0322. The agreement at atmospheric pressure is comparatively good, but it is unusual that the discrepancy should be in the direction it is at high pressures. Other values for comparison are 47° to 50° for the melting point by McCreath,^^ and 48.5° by Block.^^ Block gives for the change of volume 0.0576. For the latent heat there is 40.8 cal. by a direct method and 41 by a cryoscopic method by Eykmann.^^ The value found above, 40.8, is in unusually good agreement. Urethan has three triple points at each one of which the difference of compressibility etc. between the phases can be calculated. The data are not accurate enough, however, at the two triple points in- volving the liquid to give accurate results. Thus a solution of the equations at the triple point L-I-II demands that I be very much more expansible than the liquid, which seems very unlikely. Probably a large part of the error is occasioned by the abnormally rapid change of slope of the AV curve for L-II. The only information of value is, therefore, that given at the triple point I-II-III. We find on solving the six equations at this point the following values, Aai2 = O.O5II Aa23 = O.O4II Aai3 = O.O4I2 Ai8i2 = - O.O55 • Ai323 = O.O3I35 AjSis = O.O3I3O ACpi2 = - 0.127 > ACp23 = 0.43 r kgm. cm./gm. ACpi3= 0.30 ) There is doubtless a large amount of uncertainty in these numerical values; probably Aoios is too large, and Aj8i2 too small. The signs of the quantities are of interest. The compressibilities follow the order of volumes, as we should expect, and the specific heats the order of temperatures. Thus as we raise temperature at constant pressure we may change III to I, and then I to II. On passing through each of these transitions the specific heat increases. As regards the thermal expansion, we do not know whether to expect the phase at the higher temperature or that with the larger volume to have the greater expan- 33 D. McCreath, Ber. D. Chem. Ges. 8, 383-384 (1875). 124 BRIDGMAN. sion. We see that in this case the order followed is that of temperature. II is stable at a higher temperature than I and has the greater expan- sion, although it has the smaller volume. I is more expansible than III, and II than III. The difference of compressibility of the several solids as given directly by the difference of slopes was inappreciable. It is a pleasure to acknowledge generous assistance from the Bache Fund of the National Academy, and from the Rumford Fund of the American Academy of Arts and Sciences. The Jeffekson Physical Laboratory, Harvard University, Cambridge, Mass. Proceedings of the American Academy of Arts and Sciences. Vol. 51. No. 3. — October, 1915. A QUANTITATIVE STUDY OF CERTAIN PERTHITIC FELDSPARS. By Charles H. Warren. A QUANTITATIVE STUDY OF CERTAIN PERTHITIC FELDSPARS. By Charles H. Warren. Received, April 26, 1915. CONTENTS. Page Introduction 127 Summary Description of the Structural Characteristics of the Feld- spars Studied 129 Quantitative Methods of Study Employed Microscopic 134 Chemical 139 Discussion of Results 141 Discussion of the Origin of i'erithitic Intergrowths 143 Summary 153 Introduction. — It is the chief purpose of this paper to present the results of a quantitative microscopic and chemical study of certain perthitic feldspars from granite pegmatites. The feldspars here considered are all from granite pegmatites from widely separated localities. They show a considerable range in the relative proportions of the two feldspar phases present, and were found upon examination to be satisfactorily free from foreign inclu- sions and from alteration. The number of specimens originally chosen for study was ten. Of these four were unfortunately sent to Europe during the early summer of 1914 for sectioning and have been held there because of the war, so that the results of the investigation can not, at this time, be made quite as comprehensive as was origi- nally planned. However, as it is believed to be unlikely that the results obtained from the study of these four feldspars would have materially modified those obtained from the six remaining prepara- tions, plus the two studied by Makinen, and here included, it has 128 WARREN. been thought best to put on record the results of the investigation as it stands. In recent years J. H. L. Vogt ^ has given us in an elaborate and suggestive paper, an hypothesis regarding the physico-chemical rela- tions of the alkali-feldspars. He has not, so far as the writer is aware, furnished us with a detailed quantitative microscopic and chemical study combined of particular feldspar intergrowths such as is here described. In an admirable paper on the Granite Pegmatites of Tammela in Finland, Eero Makinen ^ has given us a careful chemical and microscopic study of certain perthites from some of the Tammela pegmatites, including a quantitative estimate of the amounts of the two feldspars present in the perthites. Both Vogt's and Makinen's work will be referred to more fully beyond. Aside from Makinen's results there has been, so far as the writer is aware, no quantitative study made of these intergrowths which is at the same time both microscopic and chemical. The importance of perthitic feldspars in many members of the granite and syenite families of rocks, and the lack of precise knowl- edge regarding them, is sufficient to indicate the importance of our having further ciuantitative information regarding them and has led the writer to attempt the present investigation. To anyone familiar with the characteristics of perthites and microperthites it will be obvious that data of a very precise character will be difficult to obtain. The coarser textured perthites of the granite pegmatites clearly ofi'er a better chance of yielding quantitative results of greater precision than the finer microperthites of the granitoid rocks and were, there- fore, made the subjects of a first attempt. The results are offered in the belief that they at least furnish a rather close approximation to the truth so far as the granite pegmatite feldspars are concerned, and the hope is entertained that they may, if applied with caution, throw some further light on the mode of origin of perthitic intergrowths in general. With regard to the geological relationships of the pegmatites from which the perthites here studied have been taken, little can be stated at the present time. With the exception of the feldspar from Perth, Ontario, and perhaps that from Bedford, Ontario, the pegmatites are connected with an ordinary type of granite, one rather siliceous, low in lime, iron and magnesia, and relatively high in the two alkalies, 1 Tschermak's Min. u. Pet. Mitt., 24 (1905). 2 Die Granite Pegmatite von Tammela in Finnland und ihre Minerale, Bull. d. 1. Commission Geologique de Finnlande, 1913. PERTHITIC FELDSPARS. 129 potash predominating. The Bedford feldspar has apparently a simi- lar relationship, though little is known about it; that from Perth is probably connected with some of the soda-rich magmas of Ontario. Summary Description of the Structural Characteristics of the Feldspars Studied. — The microscopic characteristics of the perthitic feldspars are so familiar to mineralogists and petrographers and have been so often and so fully descriljed, that it seems unnecessary to give a detailed description of the samples here studied. A few general observations will be made and a brief summary of some of their individual char- acteristics will be given. The general form of the albite member of the intergrowth is most commonly that of relatively thin lamellae following a direction approx- imately parallel to the crystal plane, 801. Other directions are some- times followed, nor does the albite always form lamellae, but rather more or less flattened and irregular prismoid bodies. The crystal structure of the two members is, however, always substantially parallel. In sections parallel to the base (001) the outcrop of the lamellae is in general roughly parallel to the edge 001-100. Taken in detail there are many minor curvings and changes of direction, and often very marked deviations may be noted. Perhaps the most common one is in a direction that appears to be in a direction parallel to that of a imit prism. Sometimes a single band will curve gradually around in this direction ; again a number of short lamellae will coalesce along the same direction. In a recent paper A. F. Rogers ^ has called atten- tion to intergrowths showing this characteristic, citing particularly a perthite from near Port Henry, N. Y., but referring also to that from Perth, Ont. The present writer has noted perthite with this feature developed in a quite striking manner in specimens from Haddam Neck, Conn., on which it was originally intended to carry out quan- titative studies. There is a great variation in width in different samples and in the same sample. A single lamellae will also gen- erally show much variation in this respect. They rarely exceed 2 mm. in width in any specimen and not very often 1 mm. The usual width is much less. The length of the bands also varies greatly even in the same specimen. They frequently pinch down and swell again, or pinch out entirely; branching and anastomosing are exceed- ingly common. In sections parallel to the bracy-pinacoid the lamellae make an angle 3 Jour. Geo!., 21, 5 (1913). 130 WARREN. which is in the neighborhood of 70° with tlie basal cleavage. As a whole their direction is more uniform than in basal sections although more or less variation is to be seen, particularly in samples in which the albite bands are less numerous. The bands in general are more continuous and are probably less prone to branch and coalesce than in basal sections. The contacts with the niicrocline are seldom smooth; generally they are irregular and sometimes finely serrated. Minute projec- tions extend out from the band into the microcline following the principle cleavage directions, most commonly that of the base. These may in rare instances extend out for some distance or even to the next parallel band. Beside the main lamellae there is without exception a series of exceedingly minute lamellae which run parallel to the g- n- eral direction of intergrowth. In some sections there is perhaps a gradation of these into the larger ones but generally they appear to be of a different order of magnitude. Much less commonly minute lamiellae or flakes of albite may be seen following still other directions than those already mentioned, but these are unimportant. In those perthites which are poorer in albite the lamellae appear in general to be more irregular in direction, distribution, width, and in form, than in those richer in albite. For the perthite from a given locality, however, the general features of the intergrowth taken as a whole are uniform and are quite characteristic. In all the feldspars here studied the potassic feldspar is a microcline. In fact the WTiter has never yet seen a perthite which was not a micro- cline perthite, and doiibts very much if there is such a thing as a pegmatitic perthite which contains orthoclase. Usually the charac- teristic polysynthetic twinning is present. In many cases it may be very faintly developed or missing altogether. This last fact may account for the statement, so frequently made, that the potassic member is orthoclase. It is worth noting in this connection that the potassic feldspar in many granitic microperthites shows the same characteristics, and also sometimes shows the albite twinning without the pericline. The optical properties are always those gWen for microcline so far as the writer's observations have extended. A somewhat general peculiarity of these perthites is that cleavage surfaces are rarely perfectly continuous. Small variations of level may be noted from place to place. This slight deorientation has doubtless been caused by crushing stresses whose effect is also some- times recorded in the slight deorientation of parts of the microcline (less commonly in the albite) in certain feldspars as seen under the PERTHITIC FELDSPARS. 131 microscope in thin section. The albite is usually well twinned, though sometimes so minutely as to be hardly discernible. This feature renders the exact measurement of the extinctions difficult. Remarks on the individual specimens: No. 1, from Perth, Ontario, Canada. — This specimen, from the original locality, is the richest in albite of any of the feldspars studied. It consists of a rather dark reddish-brown microcHne intergrown with about an equal amount of a light red to almost white albite. The red color is due to the pres- ence of exceedingly minute crystal scales of hematite which are chiefly contained to the microcline. They are usually arranged along defi- nite crystallographic directions. Some hematite is found along frac- tures in the albite, or is more irregularly placed. It is estimated that there is about one-third as much of this material in the albite as in the microcline. The albite lamellae seldom exceed 1 mm. in width while a commonly observed width is 0.5 mm. The microcline bands will in general average broader than the albite. The orientation of the bands is fairly uniform parallel to the usual direction, but there is a common tendency to bend off toward the direction of the prism and often many short bands coalesce along this same direction. Pinching and swelling, branching and coalescing, often in a very com- plex fashion, are common particularly in certain areas. The albite forms minutely pointed surfaces of contact with the microcline as a rule. The microcline sometimes shows very distinct polj^synthetic twinning but more often this is faintly developed and is often not visible at all. There is a slight alteration of the microcline to a fine opaque dust but on the whole it is quite fresh. The extinction angles for the two feldspars are as follows : — Microcline — on 001 . . ..16° to 17° n " 010.. .. 5° Albite — " 001 . . . 4.5° (( " 010.. ..19° Sipecinien, No. 2, Westficld, Mass., U. S. A. — Color white; perthitic structure not conspicuous. The microcline is practically devoid of structure as seen under the microscope. The albite lamellae rarely exceed 0.8 mm. in width, the majority running from 0.1 mm. to 0.3 mm. ; they vary much in width along their length, pinching out and frequently forking at the ends. Besides the usual very minute lamellae running parallel to the general direction of intergrowth there is a set which makes an angle of about 12° with the basal clea\age in 132 WARREN. 010 sections. The microeline is quite fresh; the albite contains fre- quently minute vesicles. A few muscovite flakes were noted. The extinction angles are: — Microeline — on 001 .... 18° — " 010.... 4.5° Albite — " 001.... 3.7° — " 010.... 18° Specimen No. 3, from Thirteen Island Mine, Bedford Township, Frontinac Co., Ontario, Canada.— The color is a light flesh red; perthitic structure can be seen only imperfectly although its presence is suggested by a lining seen most clearly on the 010 cleavage. The feldspar is, therefore, essentially a microperthite. The albite lamellae are very small, rarely exceeding 0.1 mm. in width and averaging con- siderably less. There is, however, a fairly sharp distinction in order of magnitude between the larger bands and the series of very minute lamellae which lie between them. The smaller series have the appear- ance of being more strongly segregated midway between the larger ones. The latter are quite irregular in direction and they commonly break up into several veinlets of smaller bands, or even into a series of lenticular bodies, which may coalesce farther on into a single band again. In some areas the structure is extremely complex and great care and patience had to be exercised in order to obtain accurate measurements. The microeline in the proximity of the albite bands frequently contains numerous minute specks or vesicles, some of ferruginous character. (Hence the red color of specimen.) The twinning in the microeline, though very fine, is exceedingly perfect and beautiful; that in the albite is also very fine so that exact extinction measurements are difficult to make. Furthermore the extinction positions in the microeline in 010 sections are variable. This is probably due to a slight deorientation of the feldspar and is doubtless connected with the slight variations in the level of the cleav- age sui'faces, particularly noticeable in this feldspar. The extinction angles are: — ■ Microeline — on 001 .... 17° — " 010.... 5° to 10°. Albite — " 001 .... 4° to 5° (average of several measurements made on broader twin lamellae). — on 010.. ..17.8° PERTHITIC FELDSPARS. 133 S'pccimen No. 4, from Mineral Hill, Delaivarc-Co., Pa., U. S. A. — The color is slightly brownish to cream; only an indistinct lining indicates the perthitic structure so that this feldspar is essentially a microperthite. The albite bands are rarely over 0.2 mm. in width and range commonly from 0.05 mm. to 0.1 mm. They are individu- ally quite irregular as to shape, continuity and direction, though taken as a whole they are uniform for this feldspar. A few deorientated crystals of albite were noted closely associated with the albite bands. Both feldspars are well twinned, particularly the microcline. A few well grown small crystals of muscovite are present; the alteration is very slight. Extinction angles are as follows: — Microcline — on 001 .... 16° to 17° on broader twin bands. — " 010.... 6° Albite — " 001 ... . 3° — " 010. ...19° Speciinen No. 5, from Pikes Peak, Colo., U. S. A. — Color a fresh light green; rather short albite lamellae are easily seen although the greater part of the albite is inconspicuous. The albite lamellae rarely exceed 0.6 mm. in width. They are sometimes of the usual elongate form, but more often they are short, irregular in outline and discon- tinuous, and are usually rather widely separated from one another. They often taper rapidly or may even terminate bluntly. They may follow for a short distance the usual direction of intergrowth and then turn off, apparently toward the direction of the prism. In 010 sections the lamellae appear in general to be longer but are also quite irregular, particularly as to width. Sometimes the position of a lamella will be occupied by a series of rather stout but small shreds. A few independently orientated albite crystals are present. Twin- ning is strongly developed in both members. The extinction angles are: — jNIicrocline - - on 001 . . ..17° (( - " 010. . . . 5° Albite - " 001. . .. 3° iC _ - " 010.. ..19° Specimen No. 6, from Grafton, N. H., U. S. A.^ — White to almost 4 The locality of this feldspar is not absolutely certain. It is from New Hampshire and probably Grafton. 134 WARREN. clear in color; the albite lamellae are very clearly marked. They rarely exceed 0.6 mm. in width. The bands follow fairly well the usual direction although they frequently bend away from it. This is true also in the 010 sections where there is a rather unusual amount of departure from the usual angle of about 70° with the basal clea\'- age. A few independent albite crystals were noted. Both feldspars are strongly twinned. The structure of the microcline as seen in 010 sections is peculiar. Between crossed nicols the extinction is not uniform. There seems to be a sort of a fine lamellar structure devel- oped with varying distinctness in different areas. This is often more distinct against a cleavage line where minute wedge-shaped lamellae, having different positions of extinction, are to be seen. The extinction angles vary from 1° to 7° but are never sharp. The impression is that there has been a slight deorientation of the feldspar following particu- larly along the general direction of the intergrowth which is sub- stantially that of the Murchinsonite parting. It is probably here also connected with the slight variation in level observed in the main cleavage surfaces and has probably been caused by crushing movements in the pegmatite subsequent to its consolidation. The twinning in the albite shows also some irregularities of position. Extinction angles are: — • Microcline — ^ on 001 .... 17° — " 010.... l°to7°. Albite — " 001 ... . 3° — " 010.... 17° Quantitative Methods of Sfudy employed; Microscopic. — The relative amounts of the two members of the intergrowths were determined by the well knowTi micrometric method first described by Rosival. From a carefully chosen clea\'age fragment of each feldspar to be studied, two very large thin sections were cut, one parallel to the basal, and the other parallel to the brachypinacoidal cleavage. These sections were especially prepared for the writer by the firm of Voigt & Hochgesang, of Gottingen, Germany, and no little credit is due to them for the exceptional quality and size of the sections. The measurements were all made with the large mechanical stage of Zeiss, which is superior to any other with which the author is familiar for measurements of this kind. The scales used were all carefully calibrated. In general the lines of measurement were run at distances of a millimeter apart. PERTHITIC FELDSPARS. 135 The magnifications used were naturally varied according to the width of the particular lamellae in the section under measurement, \\itli the basal sections particularly, it was foimd convenient to diminish the intensity of the illumination by lowering the condensing system and slightly raising the objective so as to bring out the contrast be- tween the refractions of the two kinds of feldspar, thus clearly marking their respecti\'e boundaries. Crossing of the nicols was, however, constantly resorted to to make sure of the correct position of the boundaries. In the case of the brachypinacoidal sections it was in general found more convenient and quicker to measure with crossed nicols. Sometimes both the width of the microcline and al!)ite bands were measured, but in general only the albite was measured, the micro- cline being determined by difference. In order to obtain a true estimate of the accuracy of these measure- ments, the deviations of the individual measurements from the mean should have been found and the precision measure computed by the usual methods. Such a procedure involves a great expenditure of labor and is, in the present case, unwarranted because, however pre- cise the results may be, there would still remain the rather consider- able error (estimated to be probably ± 1 per cent) which arises from the fact that there are included within the microcline member many minute lamellae or stringers of albite impossible to measure exactly. There is also the impossibility, as will be pointed out later, of getting at the exact composition of the albite phase itself. It seemed, there- fore, sufficient to choose as units, in judging the precision of the results of the micrometric measurements, the separate lines of measure- ment. Accordingly, in Table No. I are given for inspection, the total length measured for each section, the numl)er of lines, their average length, and the average deviation from the mean of the volume per- centages obtained from each line. The deviation of the mean itself from the true value is V/i times as probable as any single observation, n being the number of separate observations (here the numljer of separate lines of measurement). In the column headed A.D. are given the products -p^ for each feldspar measured. This indicates that, so far as errors of observation are concerned, the means are probably within one per cent of the true ^'alue. An estimate of the probable error (P. P].) by the usual formula, using the single lines of measure- ment as the units of observation, would have given slightly smaller values than those given under A.D., but the computation is hardly worth while in this instance. Inasmuch as the average width of a — V d -v ,d c8 to Cd M • o :s fflOO 1^^: m V ■ < . fi^ 2 ■ o „ o w Orientation of Thin-section Total distance meas- ured in mms. i VI d £ la o d p ® .a be 03 -* < Average deviation of single lines from mean % (a.d.). Probable deviation of mean Percentage (A.D.). Volume Percent of Microcline Bands (Mean). Volume Percent of Plagioclase Bands (Mean). II to 001 504 20 25.2 54.68 45.32 II to 010 217.3 10 21.7 55.36 44.64 Average of 001 and 010. 721.3 30 24.0 2.79 0.55 55.02 44.98 II to 001 446 16 27.9 81.20 18.80 to 010 369 15 24.6 82.13 17.93 Average of 001 and 010. 815 31 26.3 1.00 0.18 81.64 18.36 II to 001 268.8 14 19.2 88.20 11.80 II to 010 299.8 15 20.0 85.60 14.40 Average of 001 and 010. 568.6 29 19.6 1.53 0.28 86.90 13.10 II to 001 196.0 13 15.1 87.42 12.58 1 to 010 212.4 10 21.2 87.73 12.27 Average of 001 and 010 408.4 23 17.7 1.40 0.29 87.58 12.42 II to 001 398.0 20 19.6 89.52 10.48 1 to 010 305 14 21.8 87.50 12..50 Average of 001 and 010. 703.0 34 20.4 2.9 0.57 88.51 11.49 II to 001 362 21 17.2 86.78 13.22 II to 010 516 17 30.5 89.65 10.35 Average of 001 and 010. 878 38 23.0 2.6 0.40 88.22 11.78 Harkasaari, Tanimela, Finland Pakkalanmaki, Tammela, Finland TABLE I. Re.sults of micrometric measurements made to determine the relative amounts of microcline and albite in certain perthites from Granite Pegmatites; also relative amounts of the potassic and sodic feldspar molecules as calculated from the chemical analyses given in Table II. 136 ^ « ted % of f very nu of Plagioc Ss o o •s ^ a " 1 2 o o Estima ume 0 bands ^ ^ 22_>;. ^ ^ >. <^ cS \^ ^ < '^ ^ < O ^ < 2.0 53.02 46.98 52.3 5 47.7 5 2.57 51.9 45.8 1.5 2.0 79.64 20.36 79.2 20.8 2.55 73.3 25.7 1.5 2.0 84.90 15.10 84.6 15.4 2.56 76.9 21.5 1.2 2.0 85.58 14.42 85.3 14.7 2.55 80.4 18.G 1.4 2.0 86.51 13.49 86.2 13.8 2.55 80.6 18.5 1.7 2.0 86.22 13.78 85.8 14.2 2.55 77.4 21.7 1.35 75.6 24.4 67.2 30.8 2.0 72.3 27.7 71.1 27.3 1.6 5 Corrected for FeO. 6 The specific gravity used for the microcline was 2.62 for all; that given in the table is for the albite. 137 138 WARREN. single lamellae in these intergrowths is well under 1 mm., and usually much under this figure, it is at once obvious that the total length of measurement made in each case exceeds by many hundred times the average length of a single lamellae.^ The presence of very small lamellae, often exceedingly minute, in the microcline member, having their general direction parallel to that of the larger ones, has been just referred to. In measuring each sec- tion the practice was followed of omitting temporarily the measure- ment of any lamellae that fell under two divisions of the scale used. After the section had thus measured an attempt to estimate the amount of these very small lamellae was made for each section. This was done by using a Zeiss 3 mm. apochromatic objective with the No. 12 compensating occular and a strong illumination. While only approximate estimates were possible, it is believed that the figure given in the table represents the true value for these lamellae to within =1= 1.0 per cent.^ This estimate, approximate though it is, is certainly better than a guess or than omitting them altogether from considera- tion as was done by Makinen. In computing the weight percentages from those of volume obtained directly from the micrometric measurements, 2.62 was used through- out as the specific gravity of the albite member. This was the best value found for the albite in two of the samples by direct determina- tions with heavy solutions. For the microcline member slightly different values were used in some cases, as shown in the column headed Sp. Gr. In selecting these values we were guided by the speci- fic gravity of the feldspar mixture as a whole. In the case of the feldspar from Perth, Ontario, for example, the specific gravity of the microclme is a little higher than the other because of the included hematite scales. There is, of course, some chance for error here but it cannot be as large as that introduced by the uncertainty regarding the true amounts of the very small lamellae referred to above, for a change of 0.02 in the specific gravity used would only effect a change in the second decimal place of the weight percentages. 7 No general rule for the length of measurements required for a desired accuracy such as was laid down bj' Rosival (200 times the length of the diam- eter of the average grain) can be adopted for measurements by this method in thin-sections or in general. Tlie precision attainable must be estimated for each separate case by the usual methods apphcable to such cases. 8 The writer feels reasonably sure that the error is in all cases a negative one for there are always a number of lamellae too small to be measured. The values for the different sections were so nearly constant, that in view of the error involved, it was deemed best to adopt for all what appeared to be the best value, vis. 2.0%. PERTHITIC FELDSPARS. 139 The figures given for the various feldspars, of course, refer strictly speaking only to the particular fragment from which the thin-sections were cut, as do also the figures given for the chemical composition. No doubt, there exists some variation in the relative amounts of the two feldspars present in difi^erent feldspar fragments taken from the same pegmatite and even between different parts of the same large crystal. However, from the WTiter's observation on other smaller sections of these feldspars, as well as upon perthites in general, he believes that the samples here studied are quite representative of the 1 2 3 4 5 6 7 8 9 Albite 9 Perth, Ont. Canada Westfield, Mass. U. S. A. Bedford, Oatario, Canada. Mineral Hill, Pa. U. S. A. Pikes Peak, Colo. U. S. A. Graftcn, N. H. U. S. A. Tammela , Finland frrm Harka- saari. Pakkal- anmiiki. N. H U. S A. SiOo 66.50 65.00 65.10 65.30 65.00 65.34 65.05 64.96 A1203 18.40 19.50 19.13 18.83 19.21 19.08 19.19 18.91 Fe^Oa 1.05 .35 .32 .15 .20 .18 .28 .43 MnO tr. tr. tr. tr. tr. tr. tr. tr. MgO .07 .05 .00 .05 .03 .06 — — CaO .30 .30 .24 .28 .35 .27 .41 .32 K2O 8.77 12.36 12.98 13.60 13.62 13.05 11.19 12.09 1.78 Na,0 5.40 3.03 2.53 2.20 2.20 2.56 3.57 3.31 9.98 H2O .20 .26 .16 100.52 .30 .12 .20 .15 .18 Total 100.69 100.85 100.71 100.73 100.74 99.84 100.20 Sp. Gr. 2.597 2.568 2.571 2.565 2.562 2.570 2.564 2.62 TABLE II. Chemical analyses of alkali-feldspar intergrowths, Nos. 1-6 by Warren; Nos. 7 and 8, Eero Makinen; No. 9, analysis of albite separated from pertbite by Warren. run of the perthites or microperthites from granite pegmatites in general, and that, therefore, the results here given are equally repre- sentative. Chemical. — The entire fragment remaining after the thin-sections had been cut was first very carefully crushed in a large flat-bottomed, chilled steel mortar, loss by the flying away of particles being guarded against. This material was then slowly ground in a screened, mechani- cally operated agate mortar. From this material the final sample 9 Estimated to contain about 3 to 4 per cent attached raicrocline. 140 WARREN. was taken for chemical analysis. The analyses may, therefore, be expected to represent fairly the same material upon which the micro- scopic measurements were made. The results of the analyses are col- lected in Table II, together with two by Makinen.^° The figures for all of the radicals with the exception of the alkalies are the average of closely agreeing duplicates. The figures for the alkalies are in all cases but one ^^ the average of three separate determinations. In estimating the alkalies the perchlorate method of separating the soda and potash was employed substantially as described by Gooch,^^ and was found to be very satisfactory, in fact, the writer prefers it for several reasons to the usual platinic chloride method. Blank determi- nations for alkalies were carried out under precisely the same condi- tions as the real determinations and furnished a slight correction which was applied. The total amounts of the three feldspar molecules (Mic, Ab., An.) for each sample as computed from the analytical data are given in Table I. In order to ascertain the amount of the potassic feldspar molecule actually present in solid solution in the albite member, a number of attempts were made by heavy solution separations to obtain the albite in a pure condition for analysis. As a matter of fact it was found extremel}^ difficult to get anything like a clean separation of the two feldspars in most cases. After repeated trials it was found that by using acetylene tetrabromide diluted with benzene or xylene, the perthite from Grafton, N. H., could be made to yield an albite fraction that contained only a small amount of attached microcline. Careful microscopic measurements of the amount of microcline present, indi- cated that its amount could not be far from 3 to 4 per cent. As this material appeared to be the best that could be obtained, it was used for duplicate alkali determinations whose results are given in column 9, of Table I. In using these results to determine the amount of potassic feldspar in the albite an allowance was made for the presence of three per cent of the microcline as such. Such a calculation can, of course, be only approximate, nor is it possible to estimate the exact amount of the error. It seems likely, however, that the error is no greater than that introduced into the calculations from other sources, and it may be noted that this error is of about the same order as that intro- lOLoc. cit., 66. . 11 In this one, No. 5, one of the determinations was rejected as bemg obvi- ously considerably too high for soda. 12 Methods of Analysis, F. A. Gooch, John Wiley & Sons, (1913). PERTHITIC FELDSPARS. 141 duced by the error in the estimate of minute albite lamellae contained in the microcline, and there is an even chance, at least, that it may off- set that error. It is necessary to assume that the amount of potassic feldspar dissolved in the albites of the other specimens is the same as in the one analyzed, and it was also assumed that the lime shown by the analyses, is all present in solid solution in the albite as the anor- thite molecule. There appears to be very little likelihood of these two assumptions being far wrong. Making the assumptions noted above, the computation, from the data collected in Tables I and II, of the chemical compositions of the two mixed crystals which enter into the perthitic intergrowths is a simple matter. The results are given, culculatcd to 100%, in Table III, in which are also repeated in an abbreviated form a part of the data of Table II. Taking into consider- ation the various sources of error it is thought that the total error is certainly not greater than 2.0 units in the final figures for the mixed- crystal phases and is probably less. Discussion of Results. — Reference to Table III will show that there is quite a jump from No. 1, the feldspar from Perth, Ontario, with 52.3 per cent of the microcline member to No. 8, the next nearest, with 72.3 per cent. The remaining feldspars show a distribution of values from 72.3 per cent to 86.2 per cent of microcline. Nos. 4, 5 and 6 are, however, quite close together, and admitting that the percentages given may be 1 per cent too high or too low, may be said to be practi- cally identical so far as the relative amounts of the two members are concerned. A preliminary study of several other feldspars from other localities leads the writer to believe that these too would, if measured, fall within the range above given. As the feldspars studied are be- lieved to be representative of granite pegmatite perthites in general, it is perhaps safe to say that such feldspars will most often be found to contain from 70 per cent to 87 per cent, approximately, of the micro- cline phase. The amounts of the orthoclase and albite (plus anorthite) molecules as computed from the chemical analyses also show a considerable difference between No. 1 (Or. 51.9 per cent) and the next nearest, No. 7 (Or. 67.2 per cent). The rest lie nearer together. Omitting No. 1, the range of the remaining is from Or. 67.2 per cent to Or. 80.6 per cent, or 13.4 per cent. It will be seen that the amount of KAlSiaOs in the microcline falls between 93.0 per cent and 88.5 per cent (±2.0). The variation appears to be greater than the estimated error and probably represents real differences of composition. The amount of Ab+An ranges between 142 WARREN. "^1 C3 a -a S o d o d q O 00 >fl lo »o ■^ lO a> o q 0> 00 >^S5 o 1 .-r 1 CO t-; I— 1 OS M :c5 00 (N l^ i-H 00 3 a t- IN t- (M a a Ph d -a a .g 1 t- o "^ C*< o ■2W^ co lO rJH t^ CO •3 °^ 00 t^ CM r^^ P S3Ph-«! cor>;-^q ■* q q lO qH a_-t» '* i^ '^ o <£ c-i i>^ »o d I>H-3^ •ifiP 00 - '- - - o 00 o Cl 00 ■* o r}< CD o 00 o OS t^ o t^ OS T-l OS 00 O CO (M Ol 1—1 ?2 05 00 o 05 CO o O I-H CO C5 t> 6 03 • r- 1 C 1? >o o 00 00 _ c t3 ^ ^6 .^^ il um o^ Is _• |.^ 6 1^ _ ii3|3 + a.il^lilitliisit ojJiQia I'Bonnaqo niojj oiJ^oraojonn oq^ inojj pa^BinoiBO T3 0) I s ■ - s -fe ^ ^ S =3 *?? ^

> bi lyi -Q a; 0) O c3 -*^ -C •« Oj c tD '^ - O r^ 03 '"' 03 tc t3 in >> fi ^ fc- 03 & 4^ ^' n' O O M tr >^ c — Hi ti tq X a C3 o o &0 u ffi 03 C! o3 •^ ^ o •— 1 o o3 -tJ ?^ OJ r/: •3 4) o b a^ -C & t« s o m u CJ -o -t-3 V3 rt o "3 o u n d u o o o3 ^ a> -1 -:3 o ii 53 O PERTHITIC FELDSPARS. 143 7.0 per cent and 11.5 per cent. The albite mixed-crystal shows about the same order of variation in the amounts of NaAlSisOg. The anor- thite variations are proportionally larger, the amounts of this mole- cule varying from about 3.0 per cent to 8.0 per cent. The writer, therefore, concludes that the microcline mixed-crystals of the granite pegmatite perthites will in general contain around 9.0 per cent of the albite phase in solid solution, and that the albite mixed-crystal contains about 8.0 per cent of the KAlSisOg molecule plus a small, but rather variable amount, 3.0 per cent to 8.0 per cent, of anorthite. The ^elati^^e amounts of the orthoclase molecule and of the albite plus anorthite will in general be found to range from about 70-30 per cent to 80-20 per cent, though in exceptional cases much greater variations may occur, as No. 1 with 51.9 per cent of the orthoclase molecule present. Discussion of the origin of Perthitic inter growths. — It was formerly held by several mineralogists ^^ that the perthitic intergrowths were formed by the introduction of albite from without into previously formed microcline crystals. This view is probably now entertained by few. While there are undoubted instances where albite has been introduced from without by later mineralizing action along cleavage or fracture lines, or by more irregular replacement about the margins, there is usually in such cases, independent evidence in the surround- ing rock that such action has been going on. In the case of the great majority of perthitic intergrowths, at any rate, such evidence is lack- ing. It is difficult to see how albite could be introduced into the microcline so uniformly and extensively, with the perfect preservation of the microcline's structure, if the openings, along which the former was introduced, were prepared by a breaking of the original micro- cline. Furthermore, such breaking could hardly be effected by any form of crushing without a general deorientation of the fragments, nor is it likely that the microcline crystal could be ruptured in such fashion by any sudden change of volume incident to a possible passage through an inversion point. Anyway fracturing of the micro- cline would most likely follow along the more important directions of weakness, the principle cleavages, and these directions are not those followed by the albite growths. It also seems improbable that the albite could have worked its way into the microcline along lines of easy solubility by a replacement process. If such were the case we should expect to find the replacement more marked about 13 In this connection see particularly O. Wenglein, Inaug. Diss. Kiel., (1903). 144 WARREN, the marginal parts of the crystals. Instead we fine that a striking feature of such intergrowths is the uniformity of the albite distri- bution taken as a whole throughout the entire crystal. We should also hardly expect that a wholesale replacement such as would be represented by the case of feldspar No. 1 (47.7 per cent Ab.) could be effected without causing a profound dislocation, or even a complete breaking up, of the original microcline crystal structure. It seems necessary to abandon this theory, particularly as the one given below offers an entirely reasonable explanation of the observed facts and is in keeping with the physico-chemical theory relative to such systems. J. H. L. Vogt ^* appears to have been the first to definitely formulate a physico-chemical theory to account for the origin of the perthitic intergrowths. As a result of an extensive statistical study of the chemical data regarding the compositions of the feldspars which were first to crystallize from rock magmas of different composition, he was led to conclude that the alkalic feldspars, orthoclase and albite (plus some anorthite), formed a broken series of mixed-crystals with a eutectic point between them. He estimated that the two mixed- crystal phases had the approximate composition. Or, 72 per cent; Ab + An, 28 per cent, and Ab -f x\n, 88 per cent; Or, 12 per cent (Points i and q on the diagram, Fig. IV). He was also led to conclude, judging, among other things, from their rough constancy of chemical composition and their structure, that the so-called " cryptoperthites " described by Brogger and others,^^ represented a eutectic mixture ^^ — a simultaneous crystallization of the two mixed-crystals — and that the eutectic composition lay in the region Or, 40^4 per cent; Ab + An, 60-56 per cent or approximately Or, 42 per cent; Ab -}- An 58 per cent (Point E, Fig. IV) . He furthermore pointed out that, in all probability with lowering temperatures (and perhaps also only under considerable pressures) a change would occur in the chemical composi- tions of the mixed-crystals, due to diminished solubility of the com- ponents and also, probably, to the occurrence of an inversion point in the potassic feldspar from an a to a |S modification, viz. — from orthoclase to microcline. He estimated the chemical compositions of the finally resulting crystals of microcline to be in the neighborhood 14 Loc. cit. 15 Zoit. fur. Kryst. n. Min., 16, 537. 16 It of course does not follow that the cryptoperthitic feldspars represent the only form which the feldspar eutectic can assume. Separate crystalliza- tions of the two members might form under certain favorable conditions. A fine grained aggregate or an intergrowth would perhaps be the most probable form to result from such a mixture. See later regarding this point. PERTHITIC FELDSPARS. 145 of Mic. 85-90 percent; Ab + An, 15-10 per cent. The albite set free by such a change — unmixing — might be expected to withdraw more or less completely from the microcline, and orientate itself in positions such as those occupied by the albite lamellae in the perthite intergrowths. This unmixing would also doubtless coarsen the tex- ture in the case of the cryptoperthite feldspars. The present writer believes that there is little doubt but that in the compositions of the homogeneous alkali-feldspar phenocrysts of our effusive or intrusive rocks we have evidence that the potassic and sodic feldspars form what amounts to a continuous series of mixed-crystals. He has also pointed out in the case of the alkaline-granites and their associated cover rocks of porph^Titic texture, occurring in the region of the Blue Hills and Quincy, Massachusetts,^^ that, whereas, in the relatively quickly chilled (quenched) porphyries, the phenocrysts are, or were originally, a homogeneous soda-potash-feldspar, in the under- lying and more slowly cooled granites, the feldspars, of substantially the same composition, are a microcline-microperthite. It was also pointed out in the paper alluded to, that the phenocrysts of the porphyries were easily altered and recrystallized, wholly or in part, and appeared to have been generally quite unstable as originally formed. The conclusion was drawn that the alkali-feldspars probably form at relatively high temperatures a complete series of homogeneous mixed-crystals, but at lower temperatures, probably as a result of an inversion, they formed a broken series of mixed-crystals with a eutectic point, as was proposed by Vogt. The phenocrj^sts would, therefore, represent the earlier formed phase caught in the rapidly chilled and consolidated magma and prevented from unmixing by the great inertia to molecular rearrangements characteristic of the solid state — an unstable phase in other words. The granitic feld- spars on the other hand, crystallizing under more fa^'orable condi- tions to the adjustment of equilibria, underwent the unmixing more or less completely, and being crystallographically very similar, finally crystallized in parallel position as a perthitic intergrowth. The writer suggested that the feldspars might first crystallize as a series of mixed-crystals belonging to type I of Rooseboom's classifica- tions of binary systems. Dittler ^^ has offered on the other hand experimental evidence to show that these feldspars belong to type III of Rooseboom's classification, viz. a continuous series of mixed-crystals 17 These Proceedings, 40, No. 5, 817-323 (1913). 18 Dittler, E. Die Schmeltzpunktkurvc von Kalnatronfeldspathen, T. M. P.,B. B. 3, 513-522 (1912). 146 WARREN. having a minimum melting point. He determined the melting inter- vals of some nine soda-potash feldspars ranging in composition from Or., 6.0 per cent to Or., 67.6 per cent, the majority of them lying in the range Or., 40.0 per cent to 50.0 per cent. The melting intervals were determined upon crushed material under the microscope by the method of Doelter. To the writer the data given by Dittler are not conclusive. The number of points determined seem rather too few; and in view of the very considerable errors that are associated with this method of melting point determination, fully pointed out by A. L. Day,^^ it seems doubtful if Dittler's melting intervals have much significance. Makinen,^° following Vogt, has also expressed the opinion that the perthites are the result of an unmixing, and suggests an analogy in the behavior of the system NaCl-KCl studied by H. Brand. ^^ He has also pointed out that the unmixing process must have taken place somewhat above the temperature 575° C.^^ He finds that in the Tammela pegmatites, muscovite has clearly replaced the feldspar after the albite bands were formed, and as this muscovitization took place during the pneumatolitic period, during which only a quartz was formed, the temperature of the albite crystallization is at least fixed above the inversion temperature of the quartz, viz. above 575° C. Further consideration of the problem of the relations of these feld- spars has led the writer to abandon the idea suggested by him in a former paper and above referred to, that the soda-potash feldspars of the porphyries represent a series of mixed-crystals having a definite range of stability at relatively high temperatures. This assimiption leads to serious difficulties when one attempts to apply it to the crystal- lization of the ordinary granites which are characterized by the pres- ence of the two feldspars separately crystallized, difficulties from which the theory put forward by Vogt is in large measure free. In its general features the latter seems to the writer to be well in accord with the observed facts and to lend itself well to the interpretation of the textural relations of the two feldspars as they occur in their rocks. To account for the development of the homogeneous alkalic feld- spars which are intermediate in composition between the two mixed- crystal phases of Vogt, anorthoclase etc., it seems sufficient to the author to consider them as entirely metastable crystallizations. They represent the first crystallizations of feldspar from a magma chilled 19 Fortschrift der Min., Kryst, u. Pet. 4, 134-137 (1914), 20 Op. cit. 75. 21 Neues Jahr. fur Min. Gcol. u, Pal., 32, 638 (1911), 22 Op. cit., 74. PERTHITIC FELDSPARS. 147 with relative rapidity. It seems reasonable to suppose that the two alkalic feldspars, at least, possess, in the neighborhood of their crystal- lization interval, a strong tendency to crystallize together in all pro- portions on account of the close similarity in crystal structure and volume which characterize them. The rate of the process which tends to separate them into two types of mixed crystals, one potassic and the other sodic, is probably a slow one, possessing relatively little energy, much less rapid and energetic than the tendency of the two feldspars to crystallize from the rapidly chilled and probably undercooled magma. Under such conditions the latter tendency dominates and the two feldspars crystallize together either as a homogeneous (so far as can be told) crystal or as an exceedingly fine intergrowth (crypto- perthite). Once formed in this fashion, though in a metastable con- dition, the inertia of the solid state, doubtless especially great in substances like these, indefinitely postpones further change in the direction of the establishment of a true equilibrium, and the crystals often persist as formed. The case is quite different when a longer period of time is allowed for crystallization, as in the case of the magmas which consolidate at relatively deeper levels or even in the case of the groundmass of the porphyries. ^^ Here we would have slower cooling, less undercooling, and probably also more abundant liquid or vapor phases. The separation of the two mixed-crystal phases can then take place with less or greater perfection according to the perfection of the equilibrium adjustment permitted by the rate of cooling and other conditions. The result would be either the formation of a perthitic or microperthitic intergrowth, or, under still more favorable conditions, of separate unorientated crystalli- zations of the two phases. The process of feldspar crystallization may be further considered with the aid of a diagram representing the kind of equilibrium which is believed to hold for these minerals. The diagram. Figure IV, is sub- stantially that used by Vogt. The true position of the curves is of course not known but the various points on the eutectic line and along the base may be considered as located approximately. The positions of the points N and M have been located in accordance with the figures obtained for the compositions of the mixed-crystals in the present investigation, and lie at or somewhat outside Vogt's lower estimate (about 10% of Or.) for the corresponding points. The points i and q 23 See in this connection the"Jdescription of the^groundmass feldspars in the Blue Hill porphyries, Op. cit.;.249 and 322.* 148 WARREN. have been located in accordance with \'ogt's estimate, at 72% and 12% of orthochise respectively, which represents probably a fair approximation to the truth. x\ssuniing that the diagram represents in a general way the conditions of equilibrium for the crystallization of these feldspars, it is evident that if we start with a melt having a composition represented by a point lying well to the right of i, say Or. 95% (composition .r), we should have, as the temperature fell to ordinary temperatures, represented by the base line, assuming of course that equilibrium was established throughout, as a final product, a homogeneous mixed-crystal of the microcline type of the same Or. 1(10-; 9U Ali.Oj 10 composition with which we started. If we start with a mixture say of the composition represented by a feldspar like No. 4 with about 80% Or. (composition ij on the diagram), crystallization would begin when the curve AE was reached and the composition of the first crystalliza- tion would be that of the point t/o. Assuming a continuous adjust- ment of equilibrium as crystallization progressed, the composition of the solid phase would change continuously along the line Ai until a composition yz was reached when the whole mixture would become solid. With still further lowering of temperature, if we assume that at some temperature, say at that represented by the point C, an inver- sion of the orthoclase to microcline occurred, then when the temperature PERTHITIC FELDSPARS. 149 of the solid reached this temperature at yi there would occur a sharp change in the composition of the solid crystals and we should have formed a crystal phase of the composition represented by the point ?/5, a microcline mixed-crystal, and simultaneously a certain amount of an albite mixed-crystal would separate out. The composition of the microcline crystal would then change along the dotted line until N was reached; the albite phase would change along gM. If there was no sharp inversion point then the mixed-crystals, 7/3, would persist until the point ye was reached when an unmixing would begin with the formation of a microcline mixed-crystal whose composition would then change along the line rN until N was eventually reached, and an albite mixed-crystal of a composition represented by corresponding points on the line qM. The composition of the albite would eventually be that of the point M. The albite phase set free by such an unmix- ing, inasmuch as it takes place in the solid state, would not in all probability escape to any extent from the original crystal, but rather would withdraw in some more or less regular manner and crystallize within the microcline. As the two phases are closely similar in crystal structure and volume, the unmixing is perhaps in the nature of a parallel shifting of the units of crystal structure throughout the space of the original crystal, and the two finally become fixed in parallel orientation as the familiar perthitic intergrowth. We can easily imagine conditions, such as rapid cooling, under which the inversion and unmixing process would not take place, or take place imperfectly, and there might then result homogeneous orthoclase crystals contain- ing all of the 20% of albite, or a crystal containing anyway less of the intergrown albite phase than that which would have formed had the equilibrium been completely established throughout. The fair degree of constancy in composition exhibited by the two mixed-crystal phases as deduced in the present investigation on the pegmatitic perthites, indicates that the unmixing approaches in these, at least, a fairly definite end point. If we start with a melt of a composition represented by the line z, crystallization would begin when the temperature had fallen to the curve AE with the formation of a crystal of the composition s^. This would change continuously along the lower curve until the point i was reached when crystals of composition q would also begin to form, and these two mixed-crystals would then continue to grow until the solidification was complete. A similar procedure would obtain for a mixture, say of composition w, the only difference being that there would be formed a larger proportion of crystals of composition q rela- 150 WARREN. tive to those of composition i. Further lowering of the temperature would result in an unmixing, either gradually along the lines iN and gM or with a break at temperature C, with the development of a perthitic intergrowth, just as has been explained previously. If we assume that the equilibrium is not continuously adjusted during the first crystallization interval, as perhaps might be expected in the case with magmas whose earlier crystallizations formed with relative rapidity, developing phenocrysts of some size, it seems proba- ble that, in the case of mixtures like w, for example, zonal growths might be developed — alternati\'e crystallizations of the two phases of variable composition — until the solidification had been completed. The results of some such crystallization as this are probably repre- sented by the zoned feldspar of many granites of coarsely porphyritic habit as, for example, certain phases of the RapakJAvi (Finland) granite and those from near Jonesport, Me., U. S. A. At lower temperatiires the unmixing would occur in amount corresponding to the particular compositions of the different zones. With relative rapid cooling, prob- ably accompanied by undercooling as in the case of the porphyries, metastable mixed-crystals might be formed, at least during the earlier part of the period, and these would later unmix more or less completely forming intergrowths. From a melt having the composition of the eutectic a similtaneous crystallization of the two mixed-crystals of compositions i and q would result, and while these might, under very favorable conditions, develop as separate crystallizations, it seems much more probable that intergrowths would result. Furthermore, undercooling would be much more probable to occur in the case of melts of eutectic composition. The amount of the undercooling and the resultant eft'ect on the rate of crystallization and the equilibrium adjustment would doubtless de- termine whether a homogeneous metastable mixed-crystal, or a cryp- toperthite, or perhaps a perthite would form. In any case further cooling of the solid crystals would tend to bring about an unmixing which would change the metastable mixed-crystals into an inter- growth, or coarsen any intergrowth already formed, the full or partial accomplishment of such a process depending on local conditions. Moreover mixtures of compositions near to the eutectic, perhaps even somewhat removed from it, would in all probability behave in much the same manner, particularly where rapid cooling obtained, and little or no distinction could be made between the final products so far as structure is concerned. The general structure of the crypto- perthites and their approximation to a constancy of chemical composi- PERTHITIC FELDSPARS. 151 tion lends support to the idea of a eutectie composition. It will be evident, however, from what has just been said that the occurrence in a rock of richly perthitic feldspars is by no means an indication that such intergrowths are necessarily of eutectie composition, or even very near it, as appears to have been held by some petrographers. Mixtures of a composition like that represented by the line ii may be supposed to result vmder certain conrHtions in the so-called anti- perthite intergrowths, while a mixture of composition v, well to the left of the diagram, would yield finally a mixed-crystal of the albite type of the same composition as that with which we started. From what has been said above it will be clear that rapid cooling, accompanied probably by undercooling, is believed to be primarily responsible for the development of anorthoclase, cryptoperthite and the perthitic intergrowths generally, where the latter are relatively rich — above something like 28 per cent Ab. Inasmuch as the rate of cooling is dependent on the geologic position in which a magma con- solidates, it follows that the presence of such feldspars are an indication that the rocks containing them have consolidated relatively near or at the surface. If this is correct it is a point of some importance. The anorthoclase phenocrysts of many of our porphyritic rocks bear out this contention, and geologic evidence is by no means lacking that many occurrences of rocks characterized by the presence of richly perthitic or cryptoperthitic feldspars have reached relatively high levels in the earth's crust, for example, the alkaline rocks of the Chris- tiania region and the microperthitic rocks of eastern Massachusetts, particularly of Quincy and the Blue Hills, not to mention many other examples that come to mind. We also believe that microcline is the stable phase of the potassic feldspar. The tendency to invert to microcline, like the intimately associated tendency of the two mixed-crystals to unmix, is probably not a strong one, and may often not find conditions favorable to its accomplishment. The change may be suspended indefinitely and even not occur at all, or not until the necessary energy is supplied by long continued heat and pressure such as might result from meta- morphic processes. That microcline is of much more common occur- rence than is generally supposed, the writer firmly believes. It is probably true that many students of rocks when they do not see the characteristic "grating" structure in a potassic feldspar call it orthoclase, overlooking the fact, if they were ever aware of it, that microcline does not always, in fact rather frequently, does not show this structure, that it may show only the albite twinning, or no twin- 152 WAEREN. ning at all. The ^^Titer has studied a number of rocks in which or- thoclase was described as being present only to find, on careful study, that the alleged orthoclase was really microcline. In comparing the feldspars of the pegmatites with those of the granites with which they are connected genetically, or of (Hfl'erent granites, it is of course impossible, with the rather meager data now at our command, to estimate definitely the effect of constituents other than the feldspars which are present in the respective magmas on the equilibrium relations existing between the feldspars. The effect of varying pressures is also largely of unknown magnitude. In the case of granite magmas, relatively poor in anorthite and mafic minerals, the principle differences in chemical composition as compared with the pegmatitic magmas which are derived from them, are generally con- sidered to be the greater amounts of water or water vapor and other of the more volatile constituents — mineralizers — of the pegmatitic magmas. The presence of greater amounts of water would probably have the effect of collecting different relative amounts of the two feldspars in the one as compared with the other. Probably with the greater excess of water in the pegmatitic magma would go a relatively greater amount of the more soluble (under those conditions) feldspar molecules. That the excess water would have any great effect in modifying the general relations of the two feldspars during crystalliza- tion in the pegmatites as compared with the granites, the writer very much doubts. One effect of larger amounts of water etc. would very probably be to facilitate any adjustment of equilibrium that might be imminent, particularly where the adjustment had to take place in the solid state. In the pegmatites the possible presence of more active residual liquors in contact with the crystals might conceivably have the effect of carrying the unmixing somewhat further in the direction of extracting more of the more soluble phase. As this is probably the albite, it may be, that in the pegmatitic feldspars, such as those here studied, the microcline phase contains a somewhat lower percentage of the albite than is the case for corresponding feldspar in the granites. The differences in the two cases are, in the wi-iter's opinion, likely to be rather slight at least so far as the main crystallizations of pegmatites and granites in general are concerned. The residual liquors in all cases would carry some feldspar in solution along with other materials, and in special cases, perhaps considerable amounts. These liquors might segregate centrally or otherwise, or migrate, and in any case would finally deposite their dissolved material as separate crystalliza- tions, distinctive in character, of albite and microcline or, under cer- PERTHITIC FELDSPARS. 153 tain conditions, of albite or microcline alone, perhaps of a composition very near to that of the pure end members of the system. Such, probably, are the crystallizations v.hich we find about the pockets or elsewhere in certain pegmatites. We feel therefore justified, at least for the purposes of discussion, in assuming that the same general relations obtain between the potassic and sodic feldspars during their crystallization in the pegmatites as in the granites, and in looking upon the figures here obtained for the chemical composition of the two mixed-crystal phases, fron, a study of the pegmatitic perthites, as holding true, approximately of course, for the granitic feldspars. Much further work must be done on the feldspars and their rocks before the problem can in any way be looked upon as settled, but it is believed that the theory elaborated above may at least serve as a reasonable working basis. Summary. — The results of a micrometric and chemical study of six specimens of perthitic feldspars from granite pegmatites are recorded. To these have been added the results of a similar study by Makinen of the perthites from the granite pegmatites of Tammela, Finland. Preliminary studies of certain other perthites are also referred to. Taken together the feldspars studied represent widely separated localities and quite a diversity of composition. It is believed that they may be considered representative of the perthitic feldspars of granite pegmatites in general. The methods of study employed and the characteristics of the inter- growths are briefly described. The quantitative results are collected in three tables which show the chemical compositions of the various feldspars studied, the relative amounts of the two mixed-crystal phases which constitute them and the chemical composition of each of these phases. These results and their probable precision are briefly discussed. The older theory that the perthitic structure is due to the introduc- tion of albite from without into a previously formed microcline or orthoclase crystal is discussed and held to be untenable. The more recently proposed theory of Vogt that the potassic and soda-lime feldspars form a broken series of mixed-crystals with a eutectic point between them is discussed with particular reference to the alkalic members and is held to offer a satisfactory l)asis on which to explain their relationships so far as these can be judged from such evidence, chemical and petrographical, as is at present available. The progress of crystallization in mixtures of various composition is considered with the aid of a diagram, a), for the case where a perfect and continu- 154 WARREN. ous adjustment of equilibrium with falling temperature is assumed to obtain, and, b), for the case where more or less hindered or imper- fect equilibrium is supposed to obtain, chiefly as the result of rela- tively rapid cooling such as might happen in the case of magmas consolidating at, or relatively near, the surface. It is concluded that, as held by Vogt, the perthitic structure found in primary potassic feldspar where the amount of albite (plus a little anorthite) does not exceed say about 28%, is due to the unmixing of a pre\'iously homogeneous mixed-crystal. Where more considerable amounts are present it is held that the particular texture which results, whether intergrowth or separate crystals, is very largely determined by the rate of cooling, particularly during the period of initial crystal- lization. Relatively rapid cooling is held to favor in general the development of perthitic structures, or in extreme cases, of homo- geneous mixtures — metastable forms. The presence of such crystalli- zations in a rock is therefore held to be an indication of the relative conditions of cooling, and therefore of the relative geologic position in which the rock magma consolidated. It is held, that while the total composition of the perthitic feldspai's of the pegmatites may differ somewhat from those of the granites, the general relations of the two feldspars during crystallization will be much the same in both cases, and that the final composition of the two mixed-crystal phases in both the pegmatites and the granites is approximately represented by the figures found by this investigation on the pegmatitic feldspars. It is held that microcline is the stable phase of potassic feldspar at relatively low temperatures, including ordinary temperatures, and that its presence in rocks is much more general than is commonly supposed: its failure to appear is due to a feeble tendency to invert, permitting the inversion interval to be readily passed by under certain conditions. Department of Geology, Massachusetts Institute of Technology, Boston, Mass. April, 1915. Proceedings of the American Academy of Arts and Sciences, Vol. 51. Xo. 4. — Xovember, 1915. ^HE GLACIAL-CONTROL THEORY OF CORAL REEFS. By Reginald A. Daly, THE GLACIAL-CONTROL THEORY OF CORAL REEFS. By Reginald A. Daly. CONTENTS Page. Outline of theory 158 Earlier statements of elements of the theory 162 Pleistocene temperatures of the tropical ocean 166 Lowering of sea-level by Pleistocene glaciation 171 Diminished volume of ocean water 171 Gravitative influence of ice-caps 173 Conclusion 174 Islands and continental shores during the Glacial period 174 Character of the shore rocks 174 Heights of the Pleistocene islands 177 Conclusions ; 177 Origin of the coral-reef platforms 178 Size of the actual platforms 178 Pre-Glacial history of volcanic islands 178 Duration of Pleistocene abrasion 179 Rate of Pleistocene wave-benching 180 Depth of the Pleistocene benches below present sea-level . . . . 182 Depths of lagoons and of coastal shelves in stable areas .... 183 Testimony of islands uplifted in post-Pliocene time 199 Origin of the existing reefs 209 Colonization of the platforms 209 Upward growth of the reefs 210 Special development of reefs at the edges of platforms 21 1 " Drowned " atolls and other banks 212 Volumes of the existing reefs 218 Objections to the Glacial-control theory 220 Glacial lowering of sea-level within the tropics 220 Restriction of reef corals by Pleistocene cold 221 General crustal stability in the coral-sea areas 221 Sea-cut platforms and drowned valleys outside the coral seas . . . 223 Drowned valleys of the coral islands 224 Pleistocene cliffing of oceanic islands 229 Biology of oceanic islands 231 Difficulties of the subsidence theory 231 Its alternative statements 232 Uniformity of the assumed subsidence 233 Alleged proofs of current subsidence 234 Permanence of the Pacific basin . 234 Small maximum depth of lagoons . 235 Flatness of lagoon floors; comparison of depths in lagoons and on banks 240 Psychological influence of classic diagrams 245 The test by boring through reefs 247 General conclusion 248 158 DALY. Outline of Theory. A FIELD study in the year 1909 impressed the writer with the nar- rowness of the coral reefs about the Hawaiian islands. In A'ievv of the proved rapidity of coral growth, this narrowness suggested that the reefs are geologically very young. ■"• The discovery that a considerable glacier had left its traces on Mauna Kea, Hawaii, about 3,600 meters above sea-level, directly indicated a possible connection between the youthfulness of the reefs and the former climate of the archipelago. During the northern winter the surface temperature of the Hawaiian shore waters is but little above the minimum at which reef corals can thrive, namely, 20° Centigrade or GS° Fahrenheit. The northern limit of possible reef growth in these longitudes is a line only about 800 kilometers north of Hawaii and the line is still nearer the islands to the northwest. The mean annual temperature of coastal waters about Vancouver island is 10° C. In the Glacial period that tem- perature was nearly 0° C. The erratic boulders and striations on the bedrock floor of the Mauna Kea glacier appear to have an antiquity of the same order as that shown in the traces of Pleistocene sea-level glaciers in Washington State and British Columbia. The conclusion seems inevitable that corals could not thrive during the Glacial period anywhere in the Hawaiian group. Hence, the existing reefs must have been planted in the course of late Glacial or post-Glacial time, thus explaining their youthfulness. The principle involved should ob\iously be tested by reference to the facts known concerning the rest of the world's coral reefs. The writer's efi'ort to do this led to a somewhat elaliorate hypothesis covering the reef problem in general. After all of its essential ele- ments had been recognized, the writer found that some of them had already been described in published form. Yet no one had assembled all necessary features of the explanation and a brief statement of the whole, as then worked out, was published in 1910.^ The object of 1 Throughout this paper the expression "coral reef" signifies the usual complex of skeletal and shell growths, of which the frame-work is true coral in situ, though a large part may be composed of nullipore or other algal mate- rial, moUuscan or other debris of littoral species, shells of the plankton, and chemically precipitated carbonates of calcium and magnesium. The corals themselves may make up less than one-half of a reef, yet its existence and increase depend on the successful growth of these animals in spite of a constant battle with the surf. 2 R. A. Daly, Amer. Jour, of Science, 30, 297-308 (1910); cf. Science Con- spectus, pub. by the Massachusetts Institute of Technology, 1, 120-123 (1911). GLACIAL-CONTROL THEORY OF CORAL REEFS. 159 publication was to offer the hypothesis for discussion, especially by those who have a closer personal acquaintance with coral reefs. The problem is importan,t, as it vitally affects the physiography, geological history, and geological dynamics of about one-eighth of the earth's surface, as Avell as the recent history of the ocean as a unit. The available data seem to show that the whole of the ocean was chilled during the Pleistocene Glacial period. Over wide stretches of the tropical seas the reef corals were exterminated or greatly weakened in their reef-building power. The land-masses or shoals, which had been defended by the pre-Glacial living reefs in those regions, were now successfully attacked by the waves of the open ocean, and benched. At the climax of glaciation, the waves of the tropical seas ran over a surface lower than now: first, because w^ater had been removed from the ocean to form the ice-caps (located chiefly on the continents); secondly, because each ice-cap attracted the remaining ocean water to itself and thus lowered the level of the seas within the tropics. The depth of the platforms, cut-and-built by the waves during maximum glaciation, was estimated to be from 30 to 50 fathoms, or 55 to 90 m., below present sea-level. With the late-Pleistocene warming of the air, the surface water of the tropical seas grew rapidly warmer, the ice-caps were slowly melted, and general sea-level was correspondingly raised again. The warming of the tropical seas allowed the coral larvae, emanating from the limited reefs not entirely killed in spite of the Pleistocene chilling, to colonize the new, wave-cut platforms. Since reef corals thrive best on the outer edges of such benches, the new colonies there specially formed reefs, which grew upward as sea-level rose. Of course many larvae would settle elsewhere on the platforms, as well as in the shore breakers. In general, however, the colonies on or near the outer rims of the wave- formed platforms would thrive better than the inside colonies and the dominant reef would be linear, following the edges of the platforms. The fringing, barrier, and atoll reefs are thus explained as shallow crowTis recently built up on wave-formed platforms. The hypothesis implies that barrier reefs and atolls have not necessarily characterized the warm seas of the pre-Pleistocene periods but represent physio- graphic forms due to the highly specialized effects of a Glacial period. The offered explanation does not involve any vertical movements of the earth's crust and thus contrasts with the famous Darwin-Dana theory. It does not imply that any large proportion of the total erosion suffered by oceanic islands was accomplished during the Glacial period, but merely that the reef platforms were then finally smoothed 160 DALY. by the removal of thin veneers of relatively weak materials formed on the oceanic plateaus in Tertiary and pre-Tertiary time. In this respect the new theory is closely allied to that of Tyerman and Bennet, who, nearly ninety years ago, suggested that the existing reefs have grown on platforms cut by ocean waves and currents. Neither they nor their successors holding the abrasion theory of reefs, like Wharton and Agassiz, had explained how the abrasion could take place, for it was apparently assumed by each of these authors that the defending reef corals were living in the tropical seas continuousl}' and for an indefinite period. The Glacial-control theory emphasizes the Pleisto- cene as one period of inhibited coral growth, but the bulk of the ero- sion which has affected the oceanic plateaus is clearly pre-Glacial in date. The fuller statement of the theory permits an outlining of the reasons for belief that marine abrasion has largely truncated the older oceanic islands, long before the Glacial period. Because of that preliminary truncation, very extensive smoothing by Pleistocene waves and currents was possible. (See Figs. 1-4.) Further, the Glacial-control theory fully recognizes that there has been Recent crustal warping in certain oceanic areas affected by coral reefs. '^ Such local subsidence or elevation has influenced the growth Sections illustrating the development of barrier reefs and atolls. Figure 1. A normal volcanic island. Figure 2. The same island largely peneplained, with the necessary forma- tion of an encircling embankment of detritus (stippled). It is here arbi- trarily assumed that there has been no marine abrasion. Figure 3. The same island, extensively benched by the waves, involving some increase of the embankment. Such benching is expected in very old islands which have been exposed to active abrasion, either because of the Pleistocene chilling of the ocean or because of temporary failure of reef pro- tection in pre-Glacial time. Figure 4. Complete truncation of the same island by continued marine abrasion, with a slight broadening of the embankment. This is a stage that, in many instances, was possibly attained in pre-Glacial periods, as well as during the Pleistocene. In Figures 2, 3, and 4 the size of the embankment, as drawn, corresponds merely to the bulk of purely inorganic detritus. If intermixed reef and other organic material were allowed for, the embankment must be represented as broader. After the abrasion, fringing, barrier, and atoll reefs would be favor- ably located at X, Y, Z, respectively. Shifts of sea-level are not shown. The sections are drawn to scale and are also intended to show the great areal extent of the weak embankment materials, laid down around old oceanic volcanoes in pre-Glacial time. About one-half of the platform represented in Figure 4 is underlain by these materials, which must have offered little re- sistance to the benching surf of the Pleistocene period. 3 In this paper, "Recent" means "post-Glacial" and "recent" means "in late geological time." /^ ^ 161 162 DALY. of some of the existing reefs. Nevertheless, the bathometrie relation of platform to reef is normally so constant in all three oceans that a general explanation of reefs in terms of crustal movements seems im- possible. In other words, the coral-reef problem is really the problem of the platform represented in each of the many submarine shelves and lagoon floors of the coral seas. Most of the reef platforms, like many banks situated outside the coral seas, have such forms, dimensions, and relations to the sea-level that they appear to have originated during a long period of nearly perfect stability for the general ocean floor. That is a conclusion forced on the writer by a close study of the marine charts. Its validity is a matter quite independent of the Glacial-control theory. Local uplifts and sinkings of the sea-bottom have certainly taken place, at intervals during past geological time, but submarine topography seems impossible of explanation without assuming crustal quiet beneath most of the deep sea during at least the later-Tertiary and Quaternary periods. The new theory, there- fore, is based on the necessity of assuming general crustal stability in the coral-sea areas during the formation of the existing reefs and platform surfaces. Crustal uplift or subsidence must also be assumed as affecting local areas, like the southwest Pacific, within the same time interval, but these phenomena should not be allowed to obscure the main truth which is legible in the bathometry of the tropical seas. Finally, one cannot doubt that general sea-level has been aftected by crustal movements during post-Pliocene time. Recent uplifts have been demonstrated along great stretches of the continental shores. So far as these have not been matched by crustal downwarps beneath the ocean, such uplifts have tended to raise the surface of the ocean everywhere. Hence, post-Glacial time may have witnessed a positive shift of sea-level through a cause that has nothing directly to do with the mere addition of water to the ocean by the melting of ice-caps. A Recent rise of sea-level to the extent of a few meters, owing to post- Glacial warping of the earth's crust, is quite credible. Some of the submergence so conspicuous in the coral archipelagoes may therefore be due to two distinct causes, both rendering unsafe the drowned- valley criterion used by Dana in his advocacy of the subsidence theory. Earlier Statements of Elements of the Theory. Many authors, including Adhemar, Croll, Sir William Thomson (Lord Kelvin), Pratt, Heath, Upham, Penck, Hergesell, and Wood- ward, have shown the considerable deformation which must be pro- GLACIAL-CONTROL THEORY OF CORAL REEFS. 163 duced in the sea surface by the gravitative attraction of an ice-cap like that covering the northern part of North America in Pleistocene time.* The lowering of general sea-level by the abstraction of water from the sea, to form one or more ice-caps, and the corresponding rise chie to melting of the ice have been discussed by some of the authors men- tioned. In 1882, Penck estimated that the Pleistocene glaciation in the northern hemisphere alone sunk the general sea surface 66.5 m. below its present level, assuming that the Antarctic ice-cap was then as large as it is now. If that cap were then non-existent, the Pleisto- cene sea-level would have been about 50 m. below its present position.^ Von Drygalski calculated that the rise of general sea-level chie to melting of the Pleistocene ice-caps has been 150 m., a value later adopted by Penck. ^ After the publication of the writer's first paper (1910), Professor D. W. Johnson kindly drew his attention to Belt's statement of the relation between such shifts of sea-level and the origin of coral reefs. As this appears to be the earliest published remark on the subject, it is worthy of "quotation: "Another class of phenomena, usually as- cribed to a gradual sinking of the earth's crust, but which might also be produced by the return of the sea to the level it stood at before the Glacial period, is that connected with the growth of coral islands. Darwin's celebrated essay on their formation first proved that they were due to the gradual deepening of the water. Dana, closely following Darwin in his theory, estimates that this deepening of the ocean bed from which the coral islands rise has been at least 3,000 feet, and that the subsidence to which he ascribes it extends round one fourth of the earth's circumference in the Pacific, being indicated by atolls in that ocean for 6,000 miles in length and 2,000 in \\idth." '^ Four years later Upham briefly referred to the same theme. After showing that the ocean was diminished as a whole by the growth of 4 J. Adhemar, Revolutions de la mer, Paris, p. 28 (1840) ; J. Croll, Climate and Time, London, p. 368 (1875); W. Thomson, Phil. Mag., 31, 305 (1866); J. H. Pratt, ibid., 31, 172 (1866); D. D. Heath, ibid., 32, 34 (1866); W. Upham, Geology of New Hampshire, Concord, 3, Part 3. pp. 18 and 329 (1878); A". Penck, Jahresbericht, Geog. Ges. Miinchen, 6, 76 (1881), and Jahrbuch, Geog. Ges. Miinchen, 7 (reprint), p. 31 (1882); H. Hergegell, Gerland's Beitraege zur Geophysik, 1, 59 (1887); E. von Drygalski, Zeit. Ges. Erdkunde, Berlin, 22, 274 (1887); R. S. Woodward, Bull. 48, U. S. Geol. Survey (1888). 5 A Penck, .Jahrbuch, Geog. Ges. Mimchen, 7 (reprint), p. 29 (1882). 6E. von Drygalski, Zeit. Ges. Erdkunde, Berlin. 22, 274 (1887); A. Penck, Morphologie der Erdoberfliiche, Stuttgart, 2, 660 (1894). A. R. Wallace, in Island Life, London, p. 157 (1880), briefly refers to the principle. 7 T. Belt, Quart. Jour. Science, 11, 450 (1874). 164 DALY. the Pleistocene ice-caps, and that the ice attracted the remaining sea water, he wrote: " Such a rise of the sea, increasing in amount at high latitudes, is attested b}' the modified drift of both America and Europe; and coral islands afford proof of the corresponding depression of the ocean, succeeded by a gradual elevation to its present height, over large areas within the tropics. The coral islands of the tropics are witnesses of a depression of the sea, amounting to 3,000 feet or perhaps much more at the equator, while different proof shows that at the mouths of the Mississippi, Ganges, and Po rivers it was at least 400 feet. If we reflect upon the widespread changes of sea-level that marked the glacial period, occurring only where they would be produced by taking water from the sea to form the ice-sheets, and by gravitation through their influence, and if we compare these recent simultaneous changes with the general stability of the continents, it seems reasonable to attribute them to movements of the sea rather than of the land." ^ Apparently without knowledge of the writings of Belt and Upham, Penck threw out his suggestion in the following form (translated) : " The causes of the general rise of sea-level in the latest geological time might perhaps be connected with these climatic changes which the earth underwent in the Glacial period. If, during that time, northern Europe, northern North America, and the Antarctic regions were simultaneously glaciated, a considerable mass of water must have been removed from the ocean, and, if the thickness of the ice be assumed as 1,000 meters, the sea-level must have been 150 meters below its present position. However, it is conceivable that, in conse- quence of the considerable cooling, the sea bottom sank during the Glacial period, and since rose again, so that the size of the ocean basins as a whole was not lessened. Whatever explanation is shown in the future to be correct, it cannot be doubted that with a lowering of the sea-level the zone of reef -building must also sink; hence that banks, on which the corals formerly could not live, then became accessible to those animals and could be built up into atolls. Further, with a general lowering of the sea, many banks must become subject to wave abrasion, which truncated them unless they were protected by growing reefs. Thus the lowering of sea-level in the coral-reef region led to the transformation of banks into islands and elsewhere to a further cutting- away of the banks. In this way the fact is explained that the great majority of the oceanic islands are found in the coral-reef region, while 8 W. Upham, Geology of New Hampshire, Concord, 3, Part 3, p. 18 and 329 (1878). GLACIAL-CONTROL THEORY OF CORAL REEFS. 165 outside that zone, submarine platforms (Pfeiler), on which atolls could originate, are rather rare." ^ It was impossible that the ideas of Belt and Upham could be liter- ally accepted, because each implied that the late-Glacial swelling of the tropical seas is to be measured in thousands of feet vertically, to match the Darwin-Dana estimates of shifts of level in the coral seas. Geologists might well be sceptical that the formation of the Pleisto- cene ice-caps could produce an equatorial drop in sea-level of 3,000 feet, or more. At the present time even the von Drygalski-Penck estimate of 150 meters seems excessive. It will be noted that Penck offered his suggestion with reserve and he apparently rejected it finally himself, as shown b}^ his later, complete acceptance of Darwin's theory, in a Vienna lecture which reviewed the various coral-reef theories but quite failed to mention Glacial controls. ^° So far, the Avi-iter has found no earlier statement of the second, fundamental control of the Glacial climate, namely, that on the distribution of the corals which throve during the Pleistocene. The killing or great impoverishment of the reef-coral fauna, except in small sea areas protected from the comparatively cold water of the open ocean, is believed to be as essential a featiu'e of the Glacial- control theory as the shift of sea-level. The writer's 1910 paper was a preliminary note; partly on account of its brevity, the first announcement of the theory has been mis- understood in some particulars. Additional, prolonged study of ocean charts has led to tlie appreciation of many facts, especially quantitative data, which were unknown to the writer when the Hawaiian reef problem was undertaken. These facts seem powerfully to support the new theory and, at the same time, to represent strong objections to the subsidence theory of Darwin and Dana. Both theories postulate a recent rise of sea-level within the tropics, but they are utterly contrasted in their meaning for dynamical geology in general. This paper therefore offers a needed fuller statement of the Glacial controls, as well as an analysis of quantitative elements im- plied in the older theory of submergence. 9 A. Penck, Morphologic der Erdoberfliiche, Stuttgart, 2, 660 (1894). 10 A. Penck, Vortrage d. Verein zur Verbreitung naturwiss. Kenntnisse in Wien, 36 Jahrgang, Heft 1.3, (1896). 166 DALY. Pleistocene Temperatures of the Tropical Ocean. In spite of conflict of views as to the cause of Pleistocene glaciation, it is clear that it was accompanied by some fall of average air tempera- ture and of average ocean temperature, in the northern hemisphere at least. In that hemisphere the great ice-caps, then larger than the present ice-caps by nearly 16,000,000 square kilometers in total area, were not merely the result of an atmospheric condition very difi"erent from that of the present time. The ice-caps in their turn must have seriously affected the wind system and therefore the system of surface currents in the sea. The annual shifts of the currents and changes in the paths of great storms, characterized by extremes of air temperature, must have often lowered the sea temperature below 20° C, even in parts of the ocean where the mean annual temperature may have been above 20° C. Though occurring but once a year, a few days' exposure to a temperature below that point would seriously endanger the life of the i-eef-building corals. The growing l)elief among glacialists, that the southern hemisphere was locally glaciated at the same time as the northern hemisphere, is a second principal reason for postulating a great restriction of coral reefs in the Pleistocene. The glaciers of the Andes, from the equator to Cape Horn, were much larger than now, at a time which is most probably placed in this geological period. Of similar date are the formerly expanded glaciers of Central Africa and New Zealand, and the Antarctic ice-cap seems to ha^'e been much thicker and more extensive during the Pleistocene. The notable glaciation of southern regions now bearing no perennial ice, as New South Wales (35° S. Lat.), Western Tasmania (42^ S. Lat.), Campbell Island (52.5° S. Lat.), the Auckland islands (51° S. Lat.), Macquarie island (55° S. Lat.), the Falkland islands (51° S. Lat.), and the eastern highlands of South Africa (28° S. Lat.), has been recently referred to the same period. ^^ For the northern hemisphere special importance must be attached 11 C. A. Siissmilch, An Introduction to the Geology of New South Wales, Sydney, p. 152 (1911); W. H. Twelvetrees, Proc. Rov. Soc. Tasinania, p. 72, (1900); J. W. Gregory, Quart. Jour. Geol. Soc, 60, 37 (1906); E. J. Dunn, Proc. Roy. Soc. Victoria, 6, 133 (1894); P. Marshall, The Subantarctic Islands of New Zealand, Wellington, p. 689 (1909) ; R. Speight, ibid., p. 705; D. Maw- son, The Home of the Blizzard, Philadelphia and London, 2, 292 (1914), and personal communication; A. Supan, Grundziige der phvsischen Erdkunde, 3rd ed., Leipzig, Plate XIII (1903). Compare T. W. E. David, Comptes rendus, Cong. geol. internat., Mexico (reprint, 1907), pp. 31-38 (1906). GLACIAL-CONTROL THEORY OF CORAL REEFS. 167 to the proofs that glaciers, doubtless Pleistocene in date, existed in, or close to, the coral-reef areas of the ocean. Among the more telling instances are those in Hawaii, Japan, Mexico, and East Africa. Locating all these lands on a world map, the reader will note how inevitable is the conclusion that the tropical seas were considerably cooler during the Pleistocene. Philippi finds further evidence of that fact in the character of the deep-sea ooze collected in the Indian ocean (between 0° and 55° S. Lat.) by the "Gauss" expedition. He shows that the content of calcium carbonate in this deposit decreases with increasing depth and he attributes this decrease to special chemical conditions due to Pleistocene chilling. -"^^ Much more difficult is the question as to the actual amount of Pleistocene chilling of the tropical seas. The lowering of snow-line at that time was at least 900-1,000 meters in Hawaii and Japan, ^^ about 1,500 m. in equatorial America, and at least 2,000 m. in equatorial East Africa. With those estimates the Pleistocene lowering of snow- line 1,000 to 1,500 m. for Central Europe, about 2,000 m. for southern British Columbia, and nearly 2,000 meters for the northeastern United States, may be compared. In the lowlands of Java the Selenka expedition found the remains of Pleistocene species of plants wliich now grow in that island at levels from GOO to 1200 m. higher. ^^ The average decrease of air temperature with increase of altitude in mountainous districts, for the first 5,000 m. above sea, is al)out 0.56° C. per 100 m. The a^'erage decrease determined only from summit stations is nearly 0.65° per 100 m,, which is close to the gradient for free air. These mean values for the decrease practically apply both to the temperate and tropical zones. ^^ Since the relative precipitation, the relative effect of insolation, and other factors of Pleistocene climates are not determined, the beha\ior of the Pleistocene snow-line cannot, in general, give a direct value for the average lowering of air temperature within the tropics at that time. The best estimates are doubtless those to be derivefl from lands exposed, then as now, to an oceanic climate; for example, Auckland island, New Zealand, Hawaii, Vancouver island, the Coast Ranse of 12 E. PhiUppi, Zeit. deut. geol. Ges., 60, 354 (1908). 13 H. Simotomai, Zeit. Ges. Erdkunde, BerUn, No. 1, p. 56 (1914). 14 A. Tornquist, Grundziige der Formations- und Gebirgskunde, Berlin, p. 279 (1913). 15 J. Hann, Handbook of Climatologv, trans, bv R. DeC. Ward, Xow York, Part I, p. 244 (1903). 168 DALY. British Columbia, and the Cascade Range of Washington and Oregon. For such regions the free air of Glacial times seems to have been from 6° to 10° C. colder than now. Hann computed the minimum lowering for Spain to be 4.5° C.-^^ David estimated a similar lowering for New South Wales at not less than 5° C.^'^ The mean annual temperature, and the mean monthly temperatures of the surface water of the open ocean are almost identical with those respectively belonging to the overlying air. Hence the tropical seas, at the time of maximum Pleistocene glaciation, probably had an aver- age temperature of at least 5° C, and possibly as much as 10° C, below their present mean annual temperature. This result may now be confronted with the data of the following table, compiled from the atlases of the Deutsche Seewarte (Atlantischer Ozean, 1902; Indischer Ozean, 1891; Stiller Ozean, 1893). It gives the present mean monthly temperatures of principal parts of the coral seas. Temperatures in the Coral Seas, in degrees Centigrade. Region. February. May. August. November. West Indies 23-26 + 27-28 + 28-29 25-28 Brazilian coast 27-28 27-27 + 25-26 26-27 Gulf of Guinea 27-28 26-28 + 23-25 26-28 Laccadive Islands 27 + 29 26-27 27 Maldive Islands 27-28 28-29 26-27 27-28 Chagos Islands 27-28 27-28 26-27 28 Western part of Indian Ocean 27-28 26-28 24-26 25-27 Andaman-Nicobar Islands 26-27 29 27-28 27 Gulf of Aden 24-26 28-29 26-28 26-27 + China Sea 24-27 28-30 28-29 26-28 Sunda Sea 27 29 27-28 28 Celebes Sea 27-28 28-29 28 + 27 + Bismarck Archipelago 28-29 28-29 28-29 28-29 + Caroline Islands 27-28 28 28-29 + 28 Marshall Islands 26 27 27-28 28-29 27-28 Hawaiian Islands 23-24 24 25 25 Marquesas Island 26 27 28 26 Paumotu (Tuamotu) Islands 25-26 + 24-27 22-26 24-27 Austral Islands 25 24 21 23 Tonga Islands 26 25 23-24 25 New Caledonia 26-27 24-26 22-23 24-25 Great Barrier, Australia 25-28 23-27 20-25 23-27 16 Op. cit., p. 376. 17 Comptes rendus, Cong. geol. internat., Mexico (reprint, : 1907), p. 34 (1996). GLACIAL-CONTROL THEORY OF CORAL REEFS. 169' The numbers in bold type show the minimum mean monthly temperatures; slightly lower are, of course, the respective absolute minima. Remembering the 20° lower limit for reef corals, the reader observes that a mere general fall of only 6° C. in the minima must cause a very extensive destruction of the living animals. ^^ Further the favorable temperature conditions of the western tropical Pacific and w^estern tropical Atlantic are partly due to the westward driving of abundant warm water by the trade winds. In the Glacial period the trade-wind belt must have been much narrower than now, and the effect of these winds correspondingly less. At the same time the general storminess, correlated with rapid shifts of cold currents, was greater than at present. Finally, the Pleistocene extension of the Antarctic ice-cap must have caused some narrowing of the sea south of Cape Horn, with the probable result of increasing the volume of the cold Humboldt current, which now distinctly lowers the temperature of the central Pacific. For various reasons, therefore, the temperature conditions for lusty coral growth during the Glacial period are not fully suggested even by the statement that the mean annual temperature was then lower than at present by a half-dozen or more degrees. That lowering was but one of several associated causes for the inhibition of coral-reef growth. The writer believes it is not an extreme view to hold that practically the entire area now occupied by the oceanic archipelagoes and by the great barrier reefs of Australia and New Caledonia was, during the maximum Pleistocene glaciation, bereft of reefs growing rapidly enough to resist destruction by the waves. Though meager, slow growth of corals may have been possible in the open ocean, they could only thrive in sheltered ba^^s or seas, especially those along the eastern continental borders within the tropics. In these localities the corals perpetuated their kind, rendering possible a future, more favorable existence in the open ocean. The resulting Pleistocene reefs are, of course, now completely submerged. A likely place for their development was in the southern part of the Red sea, which, in the Glacial period as now, was doubtless particularly warm. The vast, rough plateaus at 60 m. to 120 m. below the surface of that sea, may represent places where reef corals then flourished. The tropical Atlantic water is to-day cooler than that of the tropical 18 J. D. Dana (Corals and Coral Islands, New York, 3rd ed., p. 108 (1890) states "that the temperature of 68° F. (20°C.) is a temporary extreme — • not that under which the polyps will flourish." 170 DALY. western Pacific or that of the Indian ocean, and, from the proximity of the huge ice-caps of lun-ope and North America, must have been more chilled during the Glacial period than were the other oceans. This conclusion affords a possible explanation of the thorough con- trasts between the reef-coral fauna of the Atlantic and that of the Indo-Pacific province. Hartme^'er says that the species are " niemals identisch," the West Indian species showing a strong excess of Gor- gonidae or horn corals. ^^ The Atlantic and Pacific were connected, across Central America, in the late Tertiary, when, therefore, the reef faunas should have been similar. The closing of this direct passage, during the Miocene or Pliocene, tended to separate the Atlantic reef- coral fauna from that of the greater ocean ; but even then, because of the failure of a marine-temperature barrier south of Africa, the separa- tion may not have been complete. However, since the last Pleisto- cene chilling, the low sea temperature has tended to isolate the relict coral fauna of the Atlantic, whereby it remains quite different from the fauna of the Pacific-Indian basin. Reef-coral larvae cannot now round Cape Horn. Except for a few weeks in the southern summer, the surface temperatures at the Cape of Good Hope are well below 20°C. Though it is theoretically possible for coral larvae to drift from parent reefs within the Indian ocean into the Atlantic basin, the chance that they could there settle, mature, and propagate is extremely small; for the shortest path of the drifting larvae would be from Madagascar to the Gulf of Guinea, a distance of about 7,000 km. The actual path must necessarily be much longer. One may well doubt that the few larvae which can round the Cape during the southern mid-summer would sur\'ive so long a journey. In any case, the thermal barriers between the Atlantic and Indo-Pacific areas of coral reefs have been nearly perfect for much of post-Glacial time. Special destruction of corals in the Atlantic during the Glacial period, together with the influence of thermal barriers, may, then, explain this faunal contrast. On the other hand, the general absence of identical species in the two oceans might be explained by a rapid evolution of types in this admittedly protean group of animals, since the late-Tertiary closing of the Panaman passage. The problem of the relation between Pleistocene chilling of the 19 A. Heilprin, The Geographical and Geological Distribution of Animals. New York, p. 248 (1887);—. Hartmever, Mitt. Geog. Ges. Miinchen, 3, 129 (1908). GLACIAL-CONTROL THEORY OF CORAL REEFS. 171 Atlantic and the character of the West Indian coral fauna is not solved, but it should be attacked, not only for its zoological significance but also as a future, more complete test of the Glacial-control theory of atolls and barrier reefs in general. Lowering of Sea-level by Pleistocene Glaciation. Diminished Volume of Ocean Water. The second main postulate of the Glacial-control theory is likewise difficult to state quantitatively other than in terms of an order of magnitude. The important Pleisto- cene ice-caps were five in number. Each of them meant a diminu- tion of the volume of the ocean water and also a gravitative lowering of sea-level in the tropical seas. If all the ice-caps reached maximum size simultaneously, and if the area, location, and average thickness of each were known, it would be a rather simple matter to compute the position of Pleistocene sea-level in relation to present sea-level. A similar degree of certainty as to relation of the former level to Plio- cene sea-level would still not be attainable, since the qviestion as to the size, and even the existence, of an Antarctic ice-cap during Pliocene time has not yet been answered. The depth of water off the edge of the ^Antarctic ice-cap is so great that one cannot assume a Pleistocene area for that sheet much larger than its present area (about 13,000,000 square km. or 5,000,000 square statute miles). ^° The probable maximum is about 16,000,000 square km. The present average thickness of the ice is, of course, unknown; the minimum possible estimate is doubtless 300 m. In a personal letter. Professor T. W. E. David states that the Great Ice Barrier has recently decreased much more than 600 feet (183 m.), probably 1,000 feet (305 m.), while the Beardmore glacier, a distributary of the ice- cap, has clearly shrunk 2,000 feet (610 m.) vertically. The least assignable average thickness for the whole ice-cap at its greatest strength is probably 600 m., and it may have been three times as great. The present area of the Greenland ice-cap is about 1,900,000 square km. It may not have been much bigger in the Glacial period; the greatest area of its non-floating portion was probably less than 3,000,000 square km. The present average thickness and the average thickness at maximum strength are alike unknown, but respectively 20 Any submerged portion of the ice-cap is negligible in connection with the main problem. 172 DALY. appear to be of the order of magnitude noted for these averages in the case of the Antarctic ice-cap. The dimensions of the vanished Pleistocene ice-caps are discussed in the wi'iter's 1910 paper.^^ Therein an average thickness of 3,600 feet (1,100 m.) and a total area of 6,000,000 square statute miles (15,500,000 square km.) were assumed, and these values still seem to the writer to be of the right order. The uncertainties as to extension and thicknesses of the great ice- caps, both Pleistocene and modern, make it unnecessary to consider at all the many smaller areas of glaciation, which are thus of no sig- nificance in the present problem. Similarly, the displacement of sea-water by those parts of the ice-caps which grew into shallow, epicontinental seas is relatively small and may here be neglected. For convenience the various rough estimates are indicated in the following table. Estimated areas, in square kilometers. Estimated average thickness, in meters. Present. Pleistocene. Present. Pleistocene. Minimum. Maximum. Minimum. Maximum. Antarctic 13,000,000 16,000,000 300 1,000 (?) 600 1,800 (?) ice-cap Greenland 1,900,000 2,500,000 (?) 300 1.000 (?) 400 (?) 1,500 (?) ice-cap Vanished ■ 15,500,000 1,000 1,500 (?) Pleisto- cene ice- caps 14,900,000 34,000,000 According to the estimates, the existing ice-caps represent approxi- mately 4,500,000 to 15,000,000 cubic km. of ice, corresponding to about 4,000,000 to 13,500,000 cubic km. of water. The area of the whole ocean is nearly 361,000,000 square km. Hence, to form the existing ice-caps, its average level has been lowered below that of an ice-free earth by an amount lying between 11m. and 37 m. Assuming simultaneous maxima for the major European, Labrador- Keewatin, Cordilleran, Greenland, and Antarctic sheets during the Pleistocene, the total ice formed was 26,000,000 to 56,000,000 cubic km., corresponding to about 23,500,000 to 51,000,000 cubic km. of water. The general sea-level would thereby be sunk below that of an 21 Amer. Jonr. Science, 30, 300 (1910). GLACIAL-CONTROL THEORY OF CORAL REEFS. 173 ice-free earth by an amount lying between 60 m. and 140 m. The larger limit is close to the estimate of von Drygalski and Penck, above mentioned. The different estimates of the last two paragraphs imply that the general sea-level has been raised by an amount ranging between 23 m. and 129 m., since the assumed synchronous, maximum development of the Pleistocene ice-caps. The extreme nature of that assumption makes it improbable that the Recent rise of sea-level has been as much as 129 m. On the other hand, the minimum estimate of the net volume of ice melted since the Glacial climax is likely to be too small. A revision of the evidence has led the writer to favor a rise of sea-level of the order of 50 m. to 60 m. (27 to 33 fathoms). If the climax occurred during the Kansan stage of the Glacial period, secondary maxima in ice-formation and corresponding shifts of level occurred during Wisconsin and other stages. In the 1910 paper an average lowering of 25 fathoms (46 m.) was assumed for all the stages of heavy glaciation. Gravitativc Influence of Ice-caps. Woodward's well known memoir "On the Form and Position of the Sea Level" contains the formulas necessary to compute the deformation of the sea surface due to attrac- tion by an ice-sheet. Specially simple and convenient are the equa- tions (67) on page 41 of his paper. The table and figure on page 70 are further aids to a quick understanding of the problem and its solution. ^^ The figure clearly shows the strong uplift of the sea water in the ^•icinity of the ice-cap and a maximum depression of the water surface at the point antipodal to the center of the mass of ice. The change from positive to negative values, for the case considered, occurs in a zone situated more than 100 degrees of arc from the antipodal point; and, for about 45 degrees from that point, the values for the negative deformation are not far from the maximum. From Woodward's equation on page 41 it is easy to calculate ap- proximately the antipodal sinking of level produced by an ice-cap such as covered northern North America in the Glacial period, pro- vided its thickness is known. If the average thickness was 1,000 m., and if the sheet held that thickness nearly to its edge, the antipodal sinking of level caused merely by attraction of the ice would be about 5 m., a value which is directly proportional to the thickness of the ice. The antipodal point for this greatest of the vanished Pleistocene ice- caps is in the Indian ocean, southwest of west Australia. That part 22 R. S. Woodward, Bull. 48, U. S. Geol. Survey (1888). 174 DALY. of the ocean would be specially affected by the gravitative depression of water level. Elsewhere the tropical seas would in less degree feel the same effect, to which must be added the lowering effect of contem- poraneous glaciers in Europe and other regions of the globe. The total gravitative effect of all excess Pleistocene ice was probably nowhere within the tropics sufficient of itself to lower sea-level as much as 15 m.; an average for the coral-sea zone of 10 m., or about 5 fathoms, may be assumed without involving serious error in the following discussion. Conclusion. Combining results, it is seen that, at the time of maxi- mum glaciation, the tropical seas probably had an average level which was 60 m. to 70 m. (33 to 38 fathoms) lower than at the present time. Islands and Continental Shores During the Glacial Period. The full consequences of Pleistocene chilling and lowering of the tropical-sea level represent a problem having to do with the nature of the Pleistocene land masses and shoals, and with the amount of abrasion by Pleistocene waves. Character of the Shore Rocks. The extent of wave-benching at the low sea-levels of Pleistocene time must have been highly variable, owing to the enormous differences in the resistance of the materials composing the shore belts of islands and continents alike. At the first low-water stage those materials would include the following: 1. Massive, strong lavas, among which were the flows erupted during the shift of level; 2. Weak pyroclastic deposits associated with those lavas; 3. Strong volcanic or other formations exposed to wave action through faulting or other t;ypes of crustal movement; 4. Very weak mud and sand deposits which had been formed under the sea by destruction of pre-Glacial lands and coral reefs ; 5. Similar deposits locally somewhat strengthened by interbedded coral reefs or by calcareous cementation; 6. Mixtures of weak sediments or pyroclastic deposits with occa- sional stronger lava flows; 7. Coral reefs of pre-Glacial age; 8. Weak deposits of shells etc., formed in offshore banks before the shift of level ; 9. Comparatively weak material residual after the secular weather- ing of the pre-Glacial lands. GLACIAL-CONTROL THEORY OF CORAL REEFS. 175 Regarding some of these shore-line materials special consideration is necessary. Numbers 4, 5, and 6 refer to continental shelves and to submarine embankments encircling islands, and were of breadth depending on the antiquity of the land masses concerned. The breadth of the continental shelf, such as that off eastern Australia, measured scores of kilometers; the embankments surrounding large, very old islands were generally narrower but doubtless often measured many kilometers in width. Initially, the plains underlain by the marine sediments, laid bare during the first Glacial maximum, would vary in elevation, the highest parts being above the new sea-level by" just the amount of the negative shift from the pre-Glacial sea-level. The coral reefs of pre-Glacial age formed steep initial bluffs over- looking the new sea-level. The average massive coral is more re- sistant to wave abrasion than the loose deposits just mentioned but is much less resistant than an unweathered lava flow. Observation shows an ordinary reef to be greatly weakened by lenses of sand and poorly cemented coral breccia. If the sea-level fell as much as 60 m., the massive coral, only 35 to 50 m. thick and resting on talus sand and blocks, would be liable to undermining, and therefore quick destruc- tion, by the waves. The friable nature of the talus and other material underlying the existing reefs is suggested by the logs of borings in Florida, Funafuti, Bermuda, Sumatra, and the Hawaiian islands. ^^ How wide the pre-Glacial reefs were, it is impossible to say. Among other things, their width depended on the antiquity of the present coral-reef fauna as a whole. There is no evidence that it dates back of the Jurassic period, nor, indeed, is it proved that the present coopera- tive habit of these species was well established before the Miocene. Hence one cannot assume the pre-Glacial islands or continents to have been fringed with indefinitely wide reefs. Paleozoic and younger organic growths of all kinds were liable to wave-benching until protected, apparently in late geological time, by the "invention" of strong coral reefs. Even now, many reefs are just able to hold their own against the breakers; others have been completely truncated during recent years; and still others, the so- called "drowned atolls," have long failed to reach the surface at all. 23 E. O. Hovey, Bull. Mus. Comp. Zool., 28, No. 3 (1896); The Atoll of Funafuti, published by the Royal Society of London, 1904; L. V. Pirsson, Amer. Jour. Science, 38, 191 (1914); C. P. Sluiter, Petermann's Geog. Mitt., 1891, Lit. Ber., p. 46; A. Agassiz, Bull. Mus. Comp. Zool., 17, 121 (1889). The figures stating the range of thickness for the exposed reefs are deduced from the well-known depth limits of vigorous growth of reef corals. 176 DALY. Moreover, it is not certain that, after the cooperative reef-building habit was well established, the tropical-sea temperature was not for a time too high for vigorous growth of reef corals. Mayer's experiments have led him to conclude that these would be killed if the ocean tem- perature were as high as 98° F. (36.7° C). According to some of the experiments certain species ceased to take food at temperatures below 98° F.24 This temperature is only 12° F. (6.7° C.) higher than that of the warmest surface water of the present ocean. The c^uestion arises as to the tropical-sea temperatures during that part of Tertiary time when Grinnell Land, at 81° 44' N. Lat., had a mean July tempera- ture about 27° F. higher than it enjoys now, while the mean January temperature was then at least 50° F. higher than now.^^ Of course one cannot assume that the tropical sea was then correspondingly warmer than at present, but the possibility of a temporary lowering of vitality in the reef corals through excessive heat cannot be easily dismissed. More certain is the fact that, though Tertiary coralliferous lime- stones of great extent and thickness appear in Fijian and other up- lifted islands, no very wide, typical coral reefs appear yet to ha^■e been found among them.^^ Thus, the idea is to be entertained that massive coral structures of Tertiary or earlier age did not greatly retard wave abrasion during the Glacial period. The degree to which the rocks of the oceanic islands were weakened by deep weathering during pre-Glacial time varied with the antiquity of those islands. They certainly had very different dates of origin. Like all the continents, the area of the tropical seas was probal)ly affected b^,' strong ^'olcanic action in the early pre-Cambrian, as well as occasionally through all subsequent time. Presumably the major- ity of the Pacific-floor volcanoes are pre-Pliocene, if not pre-Tertiary in age. If this be true, the oldest volcanic islands, long before the 24 A. G. Mayer, Pop. Sci. Monthly, Sept., p. 219 (1914). 25 According to J. Haun, Handbook of Climatology, trans, by R. DeC. Ward, New York, 1,375 (1903). 26 li. B. Guppy (Observations of a Naturalist in the Pacific l)et\veen the years 1S96 and 1899, London, 1, 7 (1903)) concluded that coral reefs do not appear to have existed in \'anua Levu (Fiji) when that island began its 2,000- foot uplift. He also notes that "coral reefs never have been very extensive at the sea-border during the last stages of emergence" of that island. The dating of the Vanua Levu uplift is evidently a matter of great importance, on which, however, information is scant. It doubtless began long before the Glacial period. GLACIAL-CONTROL THEORY OF CORAL REEFS, 177 Glacial period, must have suffered peneplanation or great reduction and then decomposition by down-seeping soil waters. The great depth — 100 m. or more — for the shell of decayed rock is well illus- trated in the southern Atlantic states of the North American Union, in Brazil, and in many other tropical lands of the present day. Heights of flic Pleistocene Islands. The rate of wave-benching was necessarily controlled in part by the ^'olume of rock to be removed by the waves; hence the range of heights for the different kinds of islands and coastal plains is an important element in the problem. For convenience, the symbol d may be used to represent the vertical (downward) shift of sea-level at a time of maximum glaciation. Simi- larly, h may represent the height of the land l)efore glaciation set in; and //, the height of the land above sea after the shift of sea-level. In the case of the young volcanic islands, h had values up to a limit of about 4,000 m. At the other extreme were the very old, stable islands, for which h approached zero in value; for them H and d were nearly equal. Other ancient islands, which had been affected by crustal uplift, may have passed through more than one erosion cycle, with final peneplanation. Still others, once peneplained or greatly reduced in volume, may have sunk well below sea, and then, in pre- Glacial time, received a veneer of organic debris, whereby the surface of each of these "banks" was brought close to the Pleistocene ocean level. For such islands also, H and d were of the same order of magnitude. The parts of the benches cut by pre-Glacial waves and not veneered with coral or other growths, would furnish islands with H less than d. For the continental shelves and purely fragmental embankments about islands, H varied in value from zero to that of d. For coasts surrounderl by pre-Glacial fringing reefs, the average value of // was nearly equal to d. For offshore banks of mud, shells, or ooze within the tropics, H was generally less than d, by at least 35 to 45 m., since, by hypothesis, they were covered in pre-Glacial tin\e by water too deep for vigorous coral growth. Conclusions. Reviewing the facts and reasonable inferences, it therefore appears that most oceanic islands, at the time of maximum glaciation, were (a) low, with // ranging between zero and a value little greater than d; and (h) composed of generaUij weal: material — detrital embankments surrounding or covering an eroded central mass of decayed volcanic rock, which itself was likely to be weak also because of the presence of ash-be Is. The encircling talus embank- ment carried veneers of coral-reef material of varying strength. The new coastal plains along the edges of continents and greater 178 DALY. islands were all low and mostly composed of very weak materials. It is probable that these gigantic embankments were only parth' veneered with massive reef material; such comparatively strong rock was in danger of undermining and rapid destruction as soon as the sea-level fell 45 m. or more, because of glaciation in higher latitudes. On the other hand, the high volcanic islands were young. Their rocks were still not essentially weakened by decay, and, as to-day, largely consisted of strong, massive lava flows. Wave-benching in such material would be incomparably slower than in that of the older islands or of the continental shelves. Origin of the Coral-Recf Platforms. Size of the Actual Platforms. The lengths and breadths of the larger reef platforms are typified in the following list : Ontong Java atoll (Solomon islands) Macclesfield bank (China Sea) Australian Great Barrier Chagos Bank (Indian ocean) Suvadiva atoll (Maldives) Miladummadulu-Tiladummati atoll (Maldives) Length Extreme breadth Km. Km. 80 30 150 55 2,000-f- 50-185 150 110 80 65 145 30 The large majority of the platforms have lengths less than 30 km. and widths less than 20 km. In a consideration of their possible smoothing by surf action, width is nmch more important than length. (See also Table I at page 187.) Pre-Glacial History of Volcanic Islaiids. Both a priori reasoning and direct deduction from the knowai submarine contours about islands like Hawaii and Tahiti, warrant the conclusion that the largest and middle-sized platforms could not be caused by the abrasion of young volcanic masses during the Pleistocene. To review the situa- tion for the volcanic islands of great age, an ideal case may be considered. Assume that a conical island of normal volcanic com- position (mixed lava flows and ash beds) was formed so long ago as to have been thoroughly planed down by erosion. The encircling embankment of debris would contain the insoluble material washed out of the island, and in this would be incorporated shells and skele- GLACIAL-CONTROL THEORY OF CORAL REEFS. 179 tons of pelagic and shallow-water organisms. On account of the length of time involved in the erosion (peneplanation aided by wave scour), the organic increment to the terrace would be very great, perhaps rivalling the inorganic material in bulk. If the island were circular, with an original slope of 1 : 6 (nearly the average slope of young vol- canic islands) above and below sea-level, the embankment would probably have an average width at least one-third as great as the initial radius of the island. The total area within the 50-meter isobath, including the central outcrop of undisturbed rock, would be nearly double the original area of the island. The result of the calculation suggests that a large percentage of the area occupied by each of many Pleistocene islands would offer rela- tively small resistance to the waves of that period. For the area of the initial island more than one possibility must be weighed. If sea-level had remained constant throughout the preced- ing erosion cycle, and if the island had escaped marine planation, the volcanic rocks must have been more or less deeply weathered and weakened. After the Pleistocene shift of sea-level, their surface would be nearly at the height d above the new level. On the other hand, if the island had been truncated by waves, because of a temporary failure of reef protection during pre-Glacial times, the height of its central lavas above the new Pleistocene sea-level would be less than d by some tens of meters. Duration of Pleistocene Abrasion. As the facts concerning Pleisto- cene glaciation become better known, geologists are becoming steadily more convinced of the notable length of the Glacial period as a whole. Because of their wide experience and deep study, the estimate of Chamberlin and Salisbury is worthy of special emphasis. They regard the time elapsing between the climax of the Kansan stage and the climax of the Wisconsin stage as probably between 280,000 and 960,000 years. For the entire period other milleniums must be added, to represent the time during which the Kansan ice grew to its full thick- ness, and for the Sub-Aftonian stage, if the latter really was a time of widespread glaciation. ^^ When Bain showed how thoroughly the Kansan drift has been dissected by post-Kansan streams, he laid the foundation for general belief that intense Pleistocene glaciation began at least 300,000 years, if not at least 500,000 years, before the ice-caps of the Wisconsin stage began to wane. Reliable estimates of the total duration of a greatly lowered sea- 27 T. C. Chamberlin and R. D. Salisbury, Geology, New York, 3, 420 (1906). 180 DALY. . level within the tropics are harder to make. They must depend on an assumption as to the life of each full-bodied mass of ice after it has been once formed. That the later ice-sheets long persisted with considerable thickness is shown by the depths of fiord and other basins which have been glacially excavated. The imposing depth and breadth of the Grand Coulee in Washington State, cut by the Columbia river during merely a sub-stage in the last Glacial maximum of the Cordillera, is another ciualitative proof. If the rate of ice movement at the climax of the Wisconsin stage were known, it might be possible to give a minimum estimate for the din-ation of that climax; the data would be found in the distribution of special types of erratics, especially either those derived from high nunataks, or those carried over high masses like the Adirondacks or White Mountains. Though the whole subject is very obscure, the general probabilities, viewed in relation to the estimates by Chamberlin and Salisbury, suggest a period of from 50,000 to 200,000 years for the time during which the Pleistocene ice-caps were nearly or quite at their greatest volume. During that total period the tropical ocean had a level lower than now by an amount ranging from 30 m. to 75 m. Then occurred the deeper benching and smoothing of platforms by waves and cur- rents. Nevertheless, the sea was actively attacking the islands and conti- nental coasts throughout nearly the whole Glacial period. The reef- building corals were largely killed off long before the ice-caps of the first (xlacial stage reached their full size. The succeeding Inter- glacial stage may have witnessed a partial re-establishment of reefs in the open ocean, but, if so, such reefs must have been relatively feeble and short-lived defenders of the islands. Similar reasoning- applies to the other recognized stages of the Glacial period. Hence, though sea-level swung down and up several times, lively wave al)ra- sion must have been almost continuous. Rate of Pleistocene Wave-benching. Unfortunately little is accurately known concerning the speed with which ocean waves can drive in shore cliffs on the deep-sea islands. Most of the measurements so far made refer to coastal points affected only by the less powerful waves of the North Sea, the Mediterranean, or the inner edge of the conti- nental shelf of Europe. The clayey cliffs of Yorkshire, England, retreat at the rate of 2 in. to 4.5 m. per aunum. Matthews states that the shore-lines of Suffolk and Norfolk (also on the east coast) are being driven in at rates of from 3.5 m. to 14 m. per annum, while the rate for the Welsh coast, between Llanelly and Kidwelly (Bristol Channel) GLACIAL-CONTROL THEORY OF CORAL REEFS. 181 is nearly 2 m. per annum. ^^ The average rates of annual recession for the chalk cliffs of Normandy and for those of Dover are' said to be, respectively, 0.3 m. and 1 m. Fischer found the rate for the hard rocks of Algiers to be 10 m. in 1,200 years. ^^ The average rock-strength of most of the Pleistocene islands and coastal plains was doubtless no greater, and probably less, than that of the Dover cliffs, which also have height of the same order as the quantity d. On the other hand, the energy of the waves annually breaking on the English cliffs is less than that of the waves annually breaking on an equal length of coast-line in the oceanic islands. ^° If, as is probable, the tropical ocean was more stormy in the Glacial period than it is now, the rate of cliff recession was correspondingly higher. The peneplanation of an oceanic island, initially as large and lofty as Hawaii, would produce a composite of volcanic and shelf material about as extensive as the Macclesfield bank, one of the very larg- est coral platforms known (a "submerged atoll"). Such a mass would be attacked on all sides by the strong Pleistocene waves. The pre-Glacial embankment of sand, mud, and organic debris would yield at least as fast as the clayey cliffs of Yorkshire are receding before the relatively small waves of the North Sea. Probably 2 m. per annum would be the minimum rate of recession for cliffs developed in these shelf deposits. If the lava flows of the central mass were deeply weathered, the rate of cliff' recession there might a\erage 0.5 m. or more per annum. As already observed, the wave abrasion began before the climax of the Kansan stage and continued without serious interruption until the Wisconsin climax — a period estimated as 280,000 to 900,000 years. Is it too extreme to hold that, during such a long period, the surf of the open ocean, sweeping in on all sides, would abrade every part of an island even so extensive? Is it too extreme to believe that 28 E. R. Matthews, Coast Erosion and Protection, London, pp. 11, 21, 22 (1913). 29 See E. Briickner, in AUgemeine Erdkunde (Hann, Hochstetter. and Pok- orny), Prag and Wien, Abt. 2, p. 260 (1897). 30 This statemnit holds true in spite of the fact that wind waves are specially aided by the tides in their attack on the cliffs of England. C. Darwin (The Structure and Distribution of Coral Islands, London, 3rd ed., p. 86 (1889)) remarked that, if the corals of any one of the many low coral islands were killed, the whole island "would be washed away and destroyed in less than half a century.", On page 129 of his book he notes that a single storm entirely truncated two of the Caroline islands and partly destroyed two others. Manj'^ similar cases are on record. 182 DALY. a relatively smooth surface of al)rasion was completed just below the last low, Pleistocene sea-level? To achieve that end the waves of the whole Glacial period cooperated, but the final smoothing was accom- plished at the last Glacial climax. The abrasion was, of course, accompanied by a new slight increase in the width of the encircling embankment. Whatever doubt may exist as to the ability of the Pleistocene waves to develop so large a platform, there can be little as to their power to truncate completely the average volcanic island which had been peneplained and deeply decayed before the Glacial period. In this case the width was less than 20 km., and about one-half of the area abraded was composed of weak shelf deposits. Still more clearly would the truncation be accomplished if the central volcanic mass had been already once truncated, by pre-Glacial waves. (See Fig. 4.) In striking contrast were such islands as Hawaii, Tahiti, Murea, or Rotuma, the charts of which show submarine benches so narrow as to prove the extraordinary power of fresh lavas to resist the Pleisto- cene breakers. The benching of the pre-Glacial continental shelves was in general the work of waves running in only from two quadrants, instead of four, as in the case of the oceanic island. On the other hand, these shelves were not usually composed of any other material than weak sediments, irregularly veneered with tougher masses of reef coral. Cliff recession should therefore be rapid, and some of the wave-cut benches might be expected to have widths measuring in the tens of kilometers. Depth of the Pleistocene Benches Below Present Sea-level. In weak materials open-ocean waves can quickly form a bench about 10 m. below low-water level, but abrasion at greater depths is indefinitely slower. In the course of 50,000 years the depth of the bench surface would, probably as a rule, not be increased to more than 20 m., though its outer part might be at depths of 30 m. to 40 m. If the maximum lowering of level in the tropical ocean during the Pleistocene brought it 55 m. (30 fathoms) below present sea-level, the bench surfaces then cut would not be deeper than 75 m. to 95 m. (40 to 52 fathoms) below the same datum. If the Pleistocene sea-level within the tropics were 75 m. lower than the present one, the benches might locally have the depth of 115 m. (63 fathoms). The corresponding minimum depth is naturally zero. Between these limits are to be found the facets cut by the Pleistocene waves. The facets cannot be perfectly smooth and level, nor even in the case of those due to the complete truncation of islands, are they all at GLACIAL-CONTROL THEORY OF CORAL REEFS. 183 the same depth below present sea-level. That depth depends: (a) on the power of the waves, varying with storminess, length of fetch, etc. ; (b) on the highly variable strength of the island formations ; (c) on the original areas and heights of the islands ; and (d) on the differ- ent values for the lowering of tropical-sea level because of gravitative attraction by the ice. The control last mentioned is slight and for general purposes may be neglected, though it may partly explain the unusual depth of water now in the main Chagos lagoon of the Indian ocean. Each of the other controls was doubtless of nearly equal efficiency in the three oceans, so that in general the platforms should be at depths ranging between 60 m. and 100 m. The zero depth for the facets is nearly illustrated in such a case as Manga Reva (Gambier Islands), where small masses of hard volcanic rock rise as islands in the midst of a barrier-reef platform, on which are all depths down to 73 m. (40 fathoms). These islands are clearly residuals of one or more larger volcanic masses. According to the Glacial-control theory they have been separated by long subaerial and marine erosion and not by crustal subsidence, as stated by Darwin. Yet the wave-cut. Pleistocene facet does not appear at the present shores of these and similar islands, partly because the latter were subaerially eroded during the lower stand of sea-level in the Glacial period ; some of the existing bays are drowned stream valleys. Depths of Lagoons and of Coastal Shclres in Stable Areas. The attempt to state quantitatively the effect of Pleistocene wave abrasion is seen to be laden with difficulties. However, a fair judgment seems to indicate that composite benches, of area and depth corresponding to the platforms from which the present coral reefs rise, were then actually cut in islands and coastal plains. This conclusion is subject to two different tests. Since their completion, the Pleistocene benches have been veneered with shells and skeletons of pelagic organisms, with debris of coral reefs, and with knolls and linear reefs of growing corals. Besides sporadic, steep-sided knolls of coral, the lagoons, covering 95 to 99 per cent of the larger atoll platforms, have a nearly continuous bottom layer of calcareous mud, sand, and organic remains. The smaller the platform, the higher was the proportion of reef deljris in the \eneer, and the more rapidly has the lagoon area been shallowed. (Figs. 5-11.) On the other hand, each large lagoon should be of nearly uni- form depth over its central part. Wherever two or more large atolls were subject to similar conditions for reef growth and for pelagic life, the average depths of their lagoons should vary only within small limits. 184 DALY. The following table (I) shows the depths in representati\'e lagoons and on large banks which lack rimming reefs. The mean width of the broader part (not the maximum width) is given in Col. 1. Column 2 gives the maximum depth of lagoon or bank. For lagoons, Col. 3 gives the mean depth for a considerable area where the water is deepest; for banks, it gives the mean depth on the larger part of the flat top of each bank. In this column the values given are merely estimates, but they assist in giving a mental picture of the submarine relief. Between the ordinary atoll and the reefless bank is a type sometimes called a "drowned atoll," whose rimming reef is submerged because of the periodic killing of its corals. (See page 214.) The depth data for " drowned atolls" are likewise entered in the table. Sections of small and middle-sized atolls. Figure .5. Peros Banhos, Chagos group. Figure 6. Salomon Islands cluster, Chagos group. Figure 7. Six Islands cluster, Chagos group. Figure 8. North Minerva, 23° 37' S. Lat. and 178° 56' W. Long. Compare Figure 11. Figure 9. Wataru, Maldives. Figure 10. Section through one of the rare cays of Loai ta "drowned atoll," China Sea. Figures 5 to 9 illustrate the rule that the lagoons of small atolls are more filled with detritus than those of larger atolls. Comparison of them with Figures 10, 15, 19, 20, 33, and 34 shows lagoons of "drowned atolls" to be deeper than the lagoons of atolls with reefs awash or reaching the sea surface,, lagoons of nearly equal diameters being compared in each case. Uniform scales; vertical scale is 5 times the horizontal scale. Depths in. fathoms. Water is shown in black; rocks, including the reefs, are lined. z m ^ rii — J CO tn m , < lO 0): (0: in 2 1^ CM i£ eOOe "^ (T) IfeM E o-" 03 in 185 186 DALY, 20 7 131 220 ISO c.s 53 ■55 #,X.* 40 ^°Obs"Spot 6t> ,5,'3;i'^ J'f 10 w^ 260 Trarnfiiil /^ Water 15 16 /4 15 14 12 8 ,,y^ ■■■■■■y....,., -3" ] '^ 's i \^ 55 200 c.s Figure 11. Chart of North Minerva atoll, 23° 37' S. Lat. and 178° 56' W. Long. Compare Fig. 8. Scale, 1 : 74,000. Depths in fathoms. GLACIAL-CONTROL THEORY OF CORAL REEFS. 187 TABLE I. Depths in Lagoons and on Banks. Width of Maximum Mean depth in platform depth deeper part (Km.). (M.). (if.). Barrier Lagoons. Pacific Ocean. Great Barrier, Australia 10-12° S. Latitude 130 59 30 12-14 65 49 30 14-16 61 62 45 16-18 65 68 55 18-20 120 77 50 20-22 185 83 55 New Caledonia (north end) 46 40 33 Fiji Group Viti Levu (northwest side) 37 75 64 Viti Levu (northeast side) 17 64 37 Viti Levu (south side) 4 40 27 Vanua Levu (west side) 56 84 57 Vanua Levu (southeast side) 13 62 59 Wakaya (nearly an atoll) 7 64 49 Mbengha 13 59 29 Kandavu 4 38 22 North Astrolabe (nearly an atoll) 4 27 18 Great Astrolabe 6 41 30 Ngau 6 53 38 Nairai 4 38 26 Makongai 4 39 29 Moala 4 44 18 Budd-Iambu (nearly an atoll) 8 91 73 Oneata 6 37 31 Society Group Tahiti (northwest side) 4 49 20 Borabora 2 49 20 Raiatea 2 70 33 Caroline Group Truk (Ruk) 28 66 46 Manga Reva (Gambler Islands) 7 69 27 Indian Ocean. Mayotta, Comoro Group 14 71 37 188 DALY, Width of Maximum Mean depth in platform depth deeper part (Km.). (M.). (M.). Atolls. Pacific Ocean. Off Great Australian Barrier "Great Detached Reef" 7 60 36 Flinders 24 60 53 Saumarez 22 69 46 Bampton 24 65 55 Solomon Group Tasman or Niumano 14 37 31 North Minerva, 23° 37' S. Lat. , 179°W. 5 31 22 Long. South Minerva 3 29 18 Fiji Group Thakau Matathuthu 4 35 27 Thakau Vuthovutho 6 29 27 Ngele Levu 9 29 22 Ringgold 7 94 60 Wailangilala 4 42 33 Duff 4 42 18 Reid 9 37 33 North Argo 7 38 27 South Argo 15 66 53 Thakau Tambu 4 20 15 Aiwa 4 42 33 EUice Group Funafuti 13 55 46 Caroline Group Namonuito 30 48 40 Ulie 4 46 33 Lukunor 7 53 38 Marshall Group Arno 11 60 38 Wotje (Romanzoff) 17 51 45 Phoenix Group Canton 4 20 15 McKean (dry atoll) 1 0 0 Birnie (reef awash) 1 0 0 Paumotu (Tuamotu) Group Hao 15 59 45 Mururoa 9 28 22 China Sea. Thi tu (west atoll) 4 35 27' Thi tu (east atoll) 3 27 11 GLACIAL-CONTROL THEORY OF CORAL REEFS. 189 Width of •platform {Km.). Maximum depth (M.). Mean depth in deeper part {M.). Atolls. China Sea. Crescent (Paracel Islands) 13 75 46 Vuladdore (awash or dry reef patch) 4 0" 0 North Reef, near last, reef patch 4 0 0 Indian Ocean. Keeling (Cocos) 11 24 9 He Desroches, 5° 40' S. Lat., 53° 40' E. 16 31 26 Long. Chagos Group Diego Garcia 9 33 26 Peros Banhos 20 69 45 Salomon Islands 5 31 20 Egmont (Six Islands) 3 24 15 Blenheim 4 18 11 Maldive Group Fua Mulaku 1 0 0 Addu 9 71 55 Suvadiva 56 88 69 Hadummati 31 77 64 Kolumadulu 37 87 70 Mulaku 23 73 60 South Nilandu 22 71 59 North Nilandu 22 69. 57 Wataru 6 38 31 Felidu 18 75 55 Ari 28 79 55 South Male 15 68 49 Rasdu 7 42 27 North Male 28 71 55 Gaha 7 40 33 Horsburgh 7 42 37 Fadiffolu 18 59 46 South Malosmadulu 26 69 46 Middle Malosmadulu 9 49 37 North Malosmadulu 22 57 42 Miladummadulu 28 60 42 Makunudu (Malcolm) 4 31 18 Tiladummati 20 . 57 38 Ihavandiffulu 11 62 46 Minikoi 5 15 9 190 DALY. Width of Maximum Mean depth in platform depth deeper part (Km.). (M.). (M.). Atolls. Indian Ocean. Laccadive Group Suheli Par Agatti Peremul Par Bitra Byramgore Cherbaniani Pacific Ocean. North of Fiji Group Alexa Penguin Turpie Waterwitch China Sea. Macclesfield Tizard Rifleman Loai ta North Danger Seahorse (Routh) Indian Ocean. Chagos Group Great Chagos Pitt Speakers Victory Pacific Ocean. Thikombia (Fiji Group) Cortez (off California) Tanner " " China Sea. Prince Consort Indian Ocean. Saya de Malha, south bank (rimmed on north ; greatest depth on south) Saya de Malha, north bank 6 11 4 4 7 13 6 9 7 11 4 6 Drowned Atolls." 15 46 12 48 18 55 7 53 55 110 15 87 20 82 13 62 7 49 7 57 100 90 25 55 22 45 4 6 Rimless Banks. 7 91 15 95 7 97 13 110 46 81 128 73 42 44 46 48 77 73 60 55 38 40 75 35 38 2 60 75 80 68 73 55 GLACIAL-CONTROL THEORY OF CORAL REEFS. 191 Width of Maximum Mean depth in platform depth deeper part (Km.). (M.). iM.). Rimless Ba7iks. Indian Ocean. Nazareth 130 82 46 Seychelles 150 73 58 Amirante (partly rimmed) 35 64 55 Bassas de Pedro (Padua) 22 69 51 Cora Divh 11 70 53 Sesostris 13 77 45 Rodriguez island (shelf on west side) 20 75 60 Wadge, (off west coast of India) 7 63 55 Ceylon, shelf of east coast 22 91 55 Ceylon, shelf of west coast (rim at 46- 24 77 60 55 m.) Atlantic Ocean. Hotspur, 18° S. Lat., 36° W. Long. 22 82 64 Rodgers, 17° S. Lat., 37° W. Long. 13 91 62 Victoria, 20° 30' S. Lat., 38° W. Long. 22 73 60 Dacia, 31° N. Lat., 14° W. Long. 7 107 100 Challenger, southwest of Bermuda 7 73 60 Argus, near last 10 73 59 Though the conditions aflfecting these oceanic areas are, and long have been, highly variable, it is instructive to note the relation of the maximum depth to width of the platform. Average values for the two elements have been computed and the results given in Table II. Similar averages for general depths on banks and in the deeper parts of lagoons are also there entered. TABLE II. Depths of Lagoons and Banks in Relation to Widths of Platforms. Widtl 1 of platform. Number averaged. Average of maxi- mum depths (meters) . Average of mean depths in deeper parts (meters) . Atoll Tiflgoons 1-5 km. inclusive 23 21 16 6-10 u K 19 38 28 11-20 (I a 15 57 41 21-30 tl (1 12 66 51 31-60 u (( 3 84 68 Barrier Lagoons 2-10 a (( 15 50 27 (excluding Great 11-20 u it 4 64 41 Australian Bar- 21-60 li u 4 66 50 rier) 192 DALY. Number Width of platform. averaged Average of maxi- mum depths (meters). Average of mean depths in deeper parts (meters). nied" Atolls 1- 10 km. inclusive 4 41 32 11- 30 " " 8 60 49 31-100 " " 2 100 76 Rimless Banks All widths 22 82 62 Fiji atolls only 1-5 km. inclusive 5 36 25 6-20 " " 6 59 44 Maldive atolls only 1-10 " " 9 37 27 11-20 " " 5 64 48 21-60 " " 11 73 56 Tables I and II illustrate the following rules. (a) Both maximum and general lagoon depths increase with the width of the platform, until that w^idth reaches a value of about 20 km. These rules apply not only for world averages but also for averages calculated respectively for the Fiji and Maldive groups. It appears, therefore, that, where the conditions are not far from uniform, the filHng of the lagoon is in direct proportion to the width of the platform; yet more clearly than when many archipelagoes are considered. Where the platform is extremely narrow, the lagoon is almost, or quite, filled to sea-level. (Compare P'igs. 5-20, 38-43.) (b) Atoll and barrier lagoons of similar widths usually have maximum and general depths of the same order of magnitude. (c) The lagoons of "drowned atolls" are deeper than those of normal atolls having respectively the same widths of platform. This fact is expected, since the periodic destruction of reef corals must mean slower filling of a lagoon with coral detritus. (See page 214.) (d) The new theory demands that lagoon depths shall be generally less than those on reefless banks; Table II shows the corresponding fact. (e) The maximum depth of ordinary atoll lagoons is almost never more than 91 m. (50 fathoms). In only one case is the lagoon depth of a "drowned atoll" more than 100 m.; the Macclesfield bank, which is not completely rimmed with a reef, bears water 110 m. deep. Since probably not more than 5 m. to 25 m. can be allowed for the thickness of the Post-Glacial calcareous veneer in the wider lagoons, the accordance of platform depths for the wider lagoons and reefless banks seems clear. Their range of 60-90 m. represents magnitudes in ■ CJ , — ~ 1 f>d 1^; CJ 1^^ ^E ■ in\ o-i ■"" P, ■ b£' ^13 ^ go a: :s S ^ ^ S r- C " S o cS ~ rt=3 bC_g a.3 OO " o 0/ o 5 -^ ij p S c3 bC - c c ?•-■" oJtL, rt g bC O^^.--— -iz • - "3 ;;s bC t- 3 > -J3 «" pti Q ^ E-i oT rS 5 im' CO •*■ LO M a; j-< ""^ c cji a B a a F O ^ « « Bi BJ b •^ ^ D D & t3=£ 0_i^ O O O 0'h_; - M a 193 194 DALY. of the same order as the depths computed for the Pleistocene, wave- formed benches. ^^ The agreement is visible in spite of possible, though necessarily slight, uplift or subsidence in the areas listed. The tables indicate that the new theory withstands a statistical test, which is as plainly damaging to the subsidence theory. Neither maximum nor general depths in atoll and barrier-reef lagoons of larger size should so nearly agree if subsidence has been the essential control in forming coral reefs. (See page 235 ff .) Concerning this subject a few statements from those who have had special experience with reefs are of moment. Darwin himself wrote: "The greater part of the bottom in most lagoons, is formed of sediment; large spaces have exactly the same depth, or the depth varies so insensibly, that it is evident that no other means except aqueous deposition, could have levelled the sur- face so equally." ^^ Dana remarked: "The bottom of these large lagoons is very nearly uniform, varying but little except from the occasional abrupt shallowing produced by growing patches of reef." ^^ After making many soundings in five different atolls of the Maldive group, Gardiner reported that the bottom of each lagoon "was found to be of an almost uniform dead-level between the reefs and Sections showing general flatness of lagoon floors, and their topographic unconformity with the reefs; also proflles of a bank and a "drowned atoll." Figure 16. Suvadiva atoll, Maldive group. Figure 17. Kolumadulu atoll, Maldive group. Figure 18. Western end of Seychelles bank, Indian ocean. Figure 19. Macclesfield bank, China Sea, where rimless. Figure 20. Macclesfield bank, showing the main reef and coral knolls of this "drowned atoll." Uniform scales; vertical scale 7 times the horizontal. Depths in fathoms. Water is shown in black; rocks, including reefs and coral knolls, are lined. 31 A rough idea of the average thickness of the clastic-organic-chemical veneer may be obtained by first calculating the maximum amount of calcium carbonate which has been formed in the ocean during post-Glacial time, estimated as from 20,000 to 50,000 years. This maximum would include calcium oxide imported by the rivers, as well as any initial excess of that oxide above the total now in the ocean. That excess can only be found in the cal- cium sulphate content, since the ocean is now nearly saturated with the car- bonate. Considering the needs of organisms outside of the reef areas, and also the small proportion of reef material that can enter a lagoon, the calcu- lation shows that the calcareous veneer in each of the larger lagoons can average only a few meters in thickness. A similar result is obtained by com- puting the maximum annual increment of calcium carbonate on a platform, assuming an extreme speed of growth for the whole living part of its reef, e. g., 3 cm. per annum. 32 C. Darwin, Coral Reefs, London, 3rd ed., p. 36 (18S9). 33 J. D. Dana, Corals and Coral Islands, New York, 3rd. ed., p. 183 (1890). 195 196 DALY. shoals, which, arising precipitously, uniformly reach to within a few feet of the surface. It was to me most remarkable that we did not meet wath a single knoll of any sort jutting up to indeterminate depths."^* Gardiner also speaks of the Seychelles bank as "extraor- dinarily level," with depths of from 55 m. to 65 m.; and of "the wonderfully constant depth" of about 60 m. on the great Nazareth bank (Indian ocean). ^^ (See Fig. 18.) Among the conclusions reached by Bassett-Smith after his long study of the Macclesfield bank, one reads: "The evenness of depth is the most striking feature of the chart." ^^ (See Figs. 19-20.) Wharton, one of the leading experts on oceanic bathometry, writes concerning large banks southwest and south of the Ellice islands: "The remarkable thing about these banks is the absolute uniformity of the depth of water over their areas, inside the low rim of growing coral which encircles their edges in \arious degree. This depth is 24 to 26 fathoms." (See Figs. 34 A and 34B.) In the same article he adds: " I have no hesitation in saying that a flat floor is an invariable char- acteristic of a large atoll and I cannot find his [Darwin's] 'deeply concave surface' in any large atoll. On the contrary, a flat surface is found in all these, whether the rim be above, or below the surface." ^^ Wharton explained the flatness by wave-cutting, with the sea surface at its present level, though he did not show why such abrasion was once possible, while the defending reefs now make it largely impossible. Notwithstanding the incompleteness of his theory of coral reefs, Wharton's choice of the agency which produced the flatness of lagoon floors and of banks seems irresistible. He rightly regarded this flat- ness as no less than fatal to the Darwin-Dana theor3^ (See page 240.) Cross sections of the Australian shelf, illustrating the superimposition of the existing coral reefs on a broad platform, which was developed before, and independently of, the growth of those reefs. Figure 21. At 13° 10' S. Lat., through the Great Barrier Reef. Figure 22. At 16° ;35' S. Lat., through the Great Barrier Reef. Figure 23. At 24° 30' S. Lat., outside the coral sea. Figure 24. At 25° 45' S. Lat., outside the coral sea. The shallowness of the shelf at 24° 30' S. Lat. is explained by Recent, rapid aggradation due to the local configuration of the coast, and by a corresponding, special abundance of sand. Depths in fathoms. Uniform scales; vertical scale 12 times the horizontal. Water shown in black; rocks, including reefs, are lined. 34 J. S. Gardiner, The Fauna and Geography of the Maldive and Laccadive Archipelagoes, Cambridge, England, 1, 9 (1903). 35 ,1. S. Gardiner, Geog. Jour.. 28, 330 (1906) ; Nature, Nov. 9, p. 44 (1905). 36 P. W. Bassett-Smith in Report on Dredgings obtained on Macclesfield Bank, etc., Hydrographic Office, Adrairaltv, London, 1894. 37 W. J. L. Wharton. Nature, 55, 390 (4S97). 'Gl 1H9IN 'i- m O CO m 00 22 22 iCMi «*• ro t=H X 1= j7 a ■« ii a3 33 JO ^^J . .^-^ tf '.i- ■■ ISO Figure 33. Chart of North Danger "drowned atoll," China Sea. Scale, 1:144,000. Depths in fathoms. subsidence, even one nearly instantaneous, is not proved. If the sinking took place at a rate comparable to the average rate of obserA^ed epeirogenic movements elsewhere, the coral species could easily keep pace with the deepening and retain their favored depth below sea-level. Moreover, the subsidence would have to be incredibly uniform over a vast area, to explain the nearly uniform depth of the present rim on all sides of the bank. 214 DALY. An easier explanation is found in the special weather conditions of this oceanic area. The hurricanes of the South Indian ocean often arise far to the east of the Chagos bank and sweep westward, creating tremendous seas, which attack the bank and stir its muddy surface. The region is also visited annually by more ordinary storms, in number unusual for most tropical seas. Danckelman states that this part of the Indian ocean has an average of 119 days of heavy weather ("Ge- wittertage") every year, while the sea 35-40° farther south has a corresponding average of only 15 days.*^ If the occasional hurricane stirs the mud so thoroughly as to kill off most of the corals, the surf must succeed in cutting down the top of the atoll reef, no longer de- 2«J _ — ■ie' \,g '" IS to V -20 ' i " IB<.23Z zo li^ \io 20 20 /s/-5r^-' " " \ X'o •%3 „ «/" " .^ ^O '* 25--.,; i/«\-v ...'2' .« -"' 2< _ _ 25 ,, V 22. -.24 '92 \ 248 24 1'? 25 5c»"-l,>?o,o -^ 23 X \2S 796 .,io 25 2*25 >"> 25 " "^ ^N ^f; - —..o^ 23 X la ■ •M'\ ** " ',, 26 « ^* ^''"■'OX ,^ 24 24 2^\ , '' 26 26 24 24 2« 25 23 " 2^. 25 24 25 27 ^^ 25 60 -.4 25 2« /OO 40 2/ .■2r;, " ^^ '' >^?' " 25* ^^ " '* l-H JO 26 scrl =■' 25 , 2S "28 ,, 25 — J' 23 5 22 3422^^ „ ^« 26 27 22 26 27 ,„ 2' 27 26 .'"'2o;,, 25 ,„I ^* '° '"' ^3 24 " 28 " " 27 =« ^' " 23 25 27 2j ' V 274 ^23 2.\.'/< s IS /e , 2'-."''^\237 !7 2i 22 22 24 lOI ^:-^4 2' ^* '* " =?2 ,, lOe " -5s, J, 26 27 165 #54 25 24 23 22 22,,} f - I3t 2-* 1 /9 /8\ 320 1 -'9 /50 iia 23Vo ,2, ^ .f2 23/" 4o- IS ,„\ i4i- Figure 34A. Turpie bank, north of the Fiji group. This bank is only partly rimmed, yet shows nearly the same depth of water as that in the lagoons of Penguin or Alexa bank (Fig. 34B). Such close agreement, like the flatness of each platform, is difficult to explain on the subsidence theory. Scale, 1: 307,000. Depths in fathoms. 46 See O. Kriimmel, Handbuch der Ozeanographie, Stuttgart, 1, 321 (1907). According to C. Darwin (Coral Reefs. London, 3rd ed.. p. 40 (1889)), the sea is sometimes discolored with sediment washed out of the entrances to the Chagos atolls. On pages 87 and 113 of the same book will be found his classic statements as to the fatal effect of sediment on growing corals. GLACIAL-CONTROL THEORY OF CORAL REEFS. 215 fended by a living bulwark. The surviving corals would reoceupy the abraded surface as fast as possible, making new increments, which in their turn would be subject to later successful attacks by the waves. This hypothesis is supported by the character and relative uni- formity of the depths on the rim. Depths of from 7 m. to 20 m. are exactly of the order expected under the conditions. Much of the sand and mud formed in the partial destruction of the reef must be thrown into the lagoon, and the broad terrace, adjoining the inside of the reef-rim and covered with about 30 m. of water, appears to have been thus formed. The Pitt, Speakers, and Saya de Malha banks are under nearly the same climatic conditions as the Great Chagos. The China Sea and the open ocean washing the banks north of Fiji (named in the foregoing table), are famous for destructive typhoons. It seems reasonable to explain the submerged rims of all these crowned banks in the same way. (Figs. 34 A and 34B.) However, the correctness of the hypothesis now offered is not fully established until the normal, surface reefs of the Chagos archipelago (e. g., Diego Garcia) and of the F'iji group can be explained. The prob- lem does not now seem to be capable of full solution. Factors other than those just enumerated are important and cannot yet be valued properly. Among them are: the prevailing, as contrasted with the oc- casional, weather conditions ; the kinds and numbers of reef corals at work in the respective seas ; the protection exerted by central islands in the case of the Pacific barrier reefs ; the fetch of the hurricane waves ; and the size of the platforms involved. The last-mentioned feature is doubtless quite important, since the density of the muddy clouds stirred up by waves is directly proportioned to the width of the plat- form, as measured along the line in which the waves are running. The Maldi^•es are occasionally visited by hurricanes, coming from the Bay of Bengal. However, such storms are already much weakened by their passage across India or Ceylon, and the heavy swell generated in the l)ay itself can have little effect on the Maldive islands. Hence there is good reason for the abundance of flourishing surface reefs in this archipelago, though it lies so near the Chagos group. '*^ 47 C. Darwin (Coral Reefs, London, 3rd ed., p. 50 (1889)) quoted Captain Moresby's observation that the southern atolls of the Maldives are more constantly exposed to a heavy surf than are the northern atolls. The most southerly of all, Addu, though only 14 km. long, has a maximum lagoon depth of about 70 m., and is thus much deeper than the other small Maldive atolls. Darwin remarks: " I can assign no adequate cause for this difference of depth " (page 47 of his book). Is it in part due to a more frequent killing of the Addu €orals, in post-Glacial time, by the stirring of lag-oon sediments? >o . — " 5 :-' 'V! "■■ ■'S t~ * ~ ' 1 s> •"/.■•o> <6— . » S "to ^■ ■.-1 ■■ot: GLACIAL-CONTROL THEORY OF CORAL REEFS. 217 A number of large banks in the coral seas have either no projecting reef rims at all or else have merely local raised patches of coral on their edges. Examples are here listed (Table V). TABLE V. Rimless Banks in the Coral Seas. Length Km. Extreme Maximum breadth depth Km. M. A verage depth M. Indian Ocean Seychelles (Fig. 18) 300 175 120 55-65 Amirante 170 55 60 4.5-50 Basses de Pedro or 110 24 77 45-55 Padua (Fig. 37) China Sea Prince Consort 25 13 81 64-72 These cases are more difficult of explanation. The Seychelles bank, for example, does not now lie in the path of frequent hurricanes, and, so far as the -^Titer can find records, the Seychelles area has moderate weather the year around.*^ Similarly, part of the Laccadive plat- forms are provided with atoll reefs reaching the surface, while others are still flat banks lacking peripheral reef crowns. It is possible that the rimless banks not now lying in hurricane paths have been affected by hurricanes during some earlier fraction of post-Glacial time, as the climatic zones of the Pleistocene slowly shifted to their present posi- tion. The unrimmed banks above listed, as well as others within the tropics, have average depths of just the same order as those found on many banks outside of the coral seas, e. g., the Tanner, Cortes, and Osborn banks off California. As already noted, such accordance finds no systematic place among the consequences of the subsidence theory, but is expressly demanded by the Glacial-control theory, if these intertropical banks have been little modified by post-Glacial growth of coral, and if sea-level has undergone little change since the 48 A. Voeltzkow (Geog. Anzeiger, Jahrg., Heft 1, p. 5 (1907)) states that, so far as he saw during his extensive travels, the whole western part of the Indian ocean is remarkably devoid of strong reefs composed chiefly of living corals, though limestones enclosing isolated corals do form banks. He speaks of the local patches of growing corals ("Korallengarten") as secondary forma- tions, having no close relation (" ohne jede nahere Beziehung") to the platforms on which they rest. 218 DALY. late Pleistocene. That the depths are relatively great as compared with most lagoon depths, is at once largely explained by the failure of continuous reef growth, on which the post-Glacial aggradation of the platforms has so much depended. Among the Solomon islands, Guppy found that the reefs either reached the surface or had 7 m. to 18 m. of water upon them. Reefs of intermediate depth were not found. ^^ His generalization suggests the necessity of a rather definite critical strength which must be sur- passed if a reef is to keep its summit at sea-level. If its power of resistance to the waves falls below that critical point, the top of the reef is kept down to levels ( — 7 to — 25 m.), where the abrading surf can no longer conquer in the struggle with living coral and its allies. The exact causes for varying success in the ceaseless combat are here again not determined, and evidently mud-control is only one of many factors, which are quantitatively varying from place to place in the oceanic area. Nevertheless, the clear possibility that mud-control really explains the Great Chagos -and similar "drowned" atolls, seriously affects what Dana described as " one of the best demonstrations of the sub- sidence theory." Volumes of the Existing Reefs. According to the new theory, the living coral reefs rest on platforms prepared in the Glacial period, and thus, in general, rest on pre-Glacial sediments or volcanic rocks. The greatest thickness possible for these reefs is about 110 m., assuming an extreme amount, 75 m., for the rise of post-Glacial sea-level within the tropics. Usually the thickness would be less. Once more the theory can be tested quantitatively. The boring at Funafuti showed massive coral to persist to a depth of about 46 m. Below that depth the log of the boring suggests that it passed through talus material all the way to the bottom, at a depth of 340 m. This conclusion was reached by the writer after a careful study of the Funafuti report, issued by the Royal Society of London; a subsequent inspection of a duplicate set of the core material has tended to confirm the opinion. Unfortunately, the hole bored in the lagoon at Funafuti was not deep enough to decide the nature of the rock beneath the lagoon detritus. As shown in a later section (page 247), the main bore at F'unafuti, useful as it has been in clearing up many important points, was badly located for its primary purpose of testing the Darwin-Dana theory. A truly valuable test can be made by boring on a coral islet, 49 H. B. Guppy, Proe. Roy. Soc. Edinburgh, 13, 867 (ISSti). GLACIAL-CONTROL THEORY OF CORAL REEFS. 219 situated within the lagoon of a typical atoll, about midway between the main reef and the lagoon center. Among the more favorable locations is that at Breakfast (Friihstuck) Island in the Jaluit atoll. A second test is to be found in the thicknesses of open-ocean reefs which have been uplifted in the post-Glacial period. By the new theory they should never surpass 75 m. to 110 m. If the uplift has been great enough, the unconformity of reef and basement should be visible. After a fairly complete review of coral-reef literature, the writer is convinced that these theoretical consequences accord well with the facts so far published. That the existing reefs are mere veneers is the accordant testimony of Agassiz, Wharton, Semper, Gardiner, Guppy, and others of those who have studied recently elevated islands. Speaking of the Solomon group, Guppy wrote: " Amongst the numerous islands that I examined, I never found one that exhibited a greater thickness of coral limestone [i. e., true coral- reef material in situ] than 150 feet [46 m.], or at the very outside 200 feet [61 m.j." ^° His statement in principle serves to express the view held by each of these experienced observers. Again, the Glacial-control theory implies that the areal extent of the existing reefs must be small. Post-Glacial time is now proved to be quite moderate, of the order of 20,000 to 50,000 years. Yet, if a period of 100,000 years measures more closely the life of many existing reefs, these still cannot cover more than very small portions of their respective platforms. As a matter of fact the dry land of the average great atoll totals less than one per cent of the platform area. The whole reef crown of such a typical atoll as Suvadiva totals only about 10 per cent of the platform area. Such proportions illustrate the youth of the existing reefs. The lateral spreading of each reef is chiefly effected by growth of the corals at depths of 9 m. to 35 m., the increase being, however, much the more rapid on the open-sea side. If Gardiner's estimates of the rate of coral growth apply to the whole Recent period, the observed widths of the reefs would appear to demand from 20,000 to 50,000 years for their development. In other words, the widths as well as heights of the existing barrier and atoll reefs are of the proper size, if these cal- careous rims originated on the platforms in post-Glacial time. (See Figs. 5-22, 38-43.) Originating in very shallow water, reefs of the fringing class must have smaller average thicknesses than either barrier or atoll reefs. 50 H. B. Guppy, The Solomon Islands, London, p. 71 (1887). 220 DALY. A favorably placed fringing reef may have a width greater than that of a typical annular reef, but that cannot be true of shore reefs which are much troubled by invasions of mud. The observed widths of fringing reefs, including their maximum values, also agree with the implications of the Glacial-control theory. Therein the writer's initial problem — an explanation of the youth of the HaM^aiian reefs — finds a solution. It is worthy of note that Darwin's recognition of the youthful appearance of atolls and other reefs, in both the Pacific and Indian oceans, was a leading reason for his invention of the subsidence hypothesis. ^-^ Objections to the Glacial-control Theory. The preferred explanation of coral reefs contains many elements involving definite quantities. These include: the actual temperature of the Pleistocene ocean; the volumes of water abstracted from the sea during the various maxima of glaciation; the effect of glacial attraction on tropical sea-level ; the length of the whole Glacial period ; the total duration of the greater ice-sheets ; the power of Pleistocene waves; the rock-strengths and heights of the Pleistocene islands; the length of time since the late-Glacial warming of the sea; the loca- tion and amounts of post-Glacial (Recent) deformation of the earth's crust underlying the coral seas; the growth rates of corals and of their reef -building allies; and the maximum and mean depths of reef lagoons, island shelves, and continental shelves. In a recent paper Davis has expressed doubts as to the validity of the new theory, on various grounds. Some of his reasons have to do with the quantita- tive, elements just noted, and may be first noted, along with related objections implied in other writings. ^^ Some repetition of statement is advisable, in the interests of clearness. Glacial Lowering of Sea-level within the Tropics. "The maximum number of feet by which the sea-level was lowered may have been less than the amount above quoted (200 feet), because the depression of certain glaciated lands, like Labrador and Scandinavia, while ice sheets lay on them, was presumably compensated by an uplift of the neighboring sea floor, and that would have tended to raise the sea- level." (Davis, p. 729.) The depression of these lands, if no greater than their post-Glacial uplift, would not aflect the order of magnitude 51 C. Darwin, Coral Reefs, London, 3rd ed., pp. 43 and 69 (1889). 52 W. M. Davis, Bull. Amer. Geog. Soc, 46, 728-734 (1914). GLACIAL-CONTROL THEORY OF CORAL REEFS. 221 assumed by the present writer for the lowering of tropical sea-level. There is no evidence of a much greater depression of Labrador or Scandinavia during the Pleistocene. It is, further, improbable that the depression was compensated only by uplift of the sea bottom. If the repeated assertion that an ice-cap cannot be more than 1,600 feet (488 m.) thick were true, the foregoing statement of the Glacial- control theory would need essential changes. The reasoning on which the assertion is based, is theoretically faulty, and its conclusion impossi- ble for any one who has seriously reflected on the plain facts of glacia- tion either in the Canadian Cordillera or in the Laurentian-New England region of the east. Martin's recent proofs that in 1892 the Muir glacier had a local thickness of 750 m. and that in 1894 the Grand Pacific glacier was locally more than 900 m. thick, are good grounds for admitting the possibility of ice-caps which are 1,000 m. or more in thickness. Restriction of Reef Corals by Pleistocene Cold. It has been doubted that the reduction of ocean temperature was sufficient to kill or greatly weaken the corals on most of the reefs. Until the last year or two, practically all supporters of the Darwin-Dana theory neglected to discuss this point, and assumed a tropical-sea temperature con- tinuously favorable for coral growth. Dana held that the coral-reef period "probably covered the whole of the Quaternary." ^^ The assumption is clearly unwarranted by the facts now known concerning Pleistocene climate, north and south of the equator. The burden of proof is really on those who hold that the winter sea temperature of the coral-reef areas was not 5° to 10° C. lower during the Glacial period than it is now. General Crnstal Stability in the Coral-sea Areas. Upholders of the subsidence theory will naturally question that the ocean floor has been undisturbed for a time long enough for the preparation of the reef platforms by erosion and deposition. According to the new theory, most of this work was done in pre-Glacial time. The work demands much of the later Tertiary, as well as the Pleistocene period, and thus, during several million years, the relation of sea bottom and sea surface was not significantly changed. However, such crustal stability is necessarily postulated only for the parts of the coral-sea areas where broad platforms, about 75 m. below sea-level, are now found. For those areas the assumption of prolonged crustal stability, except for minute oscillations, seems absolutely unescapable. All theories of 53 J. D. Dana, Amer. Jour. Science, 30, 169 (1885). 222 DALY. coral reefs must recognize it. As already noted, the presence of a wide shelf or bench, a few tens of meters below sea-level, really represents a criterion for crustal stability during the later geological periods, generally including at least the time since the mid-Pliocene. The existence of the broad plateaus, their accordant relation at present sea-level, and the impossibility of explaining them by any cause other than prolonged marine action, are the supreme facts emphasized in this paper. The weakest element in the subsidence theory is its failure to take proper account of them. Perfect crustal stability in the intertropical zone during the Pleisto- cene and Recent periods is obviously not implied in the Glacial- control theory. Ample illustration of local uplift and subsidence in the coral seas has been given. Yet a comparison of Table III with the charts of reef areas in general shows the exceptions to prove the rule of an essential lack of important crustal deformation in those parts of the ocean, since the beginning of the Pleistocene. The fact that the Pliocene beds of California and other regions are strongly folded, or the fact that certain continental areas have undergone considerable warping since the mid-Pliocene, may incline some geologists to doubt the postulate of crustal stability for the sea floor during the same in- terval of time. However, a serious appeal to the diastrophic record of the continents can have no other effect than to show their general freedom from strong warping since the mid-Pliocene. So far as a parallel between continental and oceanic areas may be drawn, it merely corroborates the idea of widespread crustal quiet in the sub- marine crust. As a result of studying Californian or other mountains, one cannot forecast crustal behavior in the middle of the Pacific, nor, from Pliocene upwarps in the Alps, can one deduce recent subsidence in the Maldive-Chagos region of the Indian ocean. Dana's own theory of great antiquity for the ocean basins tends to forbid such a priori reasoning. While the Glacial-control theory fully recognized local diastrophism, as well as general crustal quiet, in the coral-sea areas during late geo- logical time, the Darwin-Dana theory seems to allow no place in them for a stable sea-floor during the same period. If that floor, anywhere in the reef regions, was not moved since the early Tertiary, the surface reefs should there be much wider than in the areas supposed to be sinking. No such unusual reefs are to be found in the charts. That there has been no stable place in the coral seas is surely harder to believe than the view that a general still-stand of the sea bottom in late geological time cannot be assumed, simply because there has been recent diastrophism in the continents. GLACIAL-CONTROL THEORY OF CORAL REEFS. 223 Sea-cut Platforms and Drowned Valleys Outside the Coral Seas. The Glacial-control theory implies a lowered wave-base and a lowered base-level for rivers during much of the Pleistocene period. The objection has been offered that the expected topographic effects are not visible in extra-tropical regions. A reply to the objection is to be found in the world's maps and charts. More than twenty years ago, Penck recognized the abundance of drowned stream valleys along coasts which have not been glaciated or uplifted in Recent time. He referred this widespread phenomenon to a general rise of sea-level and found the cause of the rise in the melting of Pleistocene ice, as several other authors had stated before him. To the same cause he attributed the development of the "Flachsee," which rims the conti- nents and larger islands.^* The general failure of geologists to follow Penck's lead seems to be due to over-emphasis on crustal subsidence as a cause of positive movements of sea-level along shore lines. In many modern works no other condition for positive movements is even mentioned. Examples are plentiful in the state survey and other papers dealing with the Pleistocene and Recent geology of the Atlantic States south of New York. The many buried rock-channels of the Thames, Cam, Tawe, Neathe, Wye, Severn, Avon, Dart, Towy, and other rivers in England and Wales, like those found beneath the estuary muds at Milford Haven, Plymouth, and Falmouth, have depths of the order required, if these channels were cut during the Pleistocene time of lowered sea-level. ^^ The Recent submergence of the Dogger bank in the North Sea has been correlated with the drowning of the British valleys. ^^ Other cases of Recent drowning to the same moderate degree are abun- dantly described in the literature concerning the Atlantic coast from Maryland southward. ^^ Erosion while the level of the Gulf of Mexico was Glacially low^ered, may well account even for the buried channel of the lower Mississippi.^^ To turn to the other side of the world, Andrews and others state that the New South Wales coast 54 A. Penck, Morphologie der Erdoberflache, Stuttgart, 2, 580, 658-660 (1894). 55 W. Whitaker, Quart. Jour. Geol. Soc, 46, 333 (1890); T. Codrington, ibid., 54, 251 (1898), and 58, 35 (1902). 56 A. S. Kennard, in discussion of J. W. Stather's paper, Quart, Jour. Geol. Soc. 68, 327 (1912). 57 See particularly G. B. Shattuck's volume on the Pleistocene and Pliocene deposits of Maryland, published by the Maryland Geological Survey (1906). 58 See J. Leconte, Elements of Geology, New York, p. 558 (1892). 224 DALY. shows a positive shift of level to the extent of about 60 m., since the Tertiary.^^ In fact, the more carefully existing coast lines are studied, the more apparent is the correctness of Penck's generalization and the more unavoidable is the hypothesis that most of the world's drowned valleys were submerged because of a Recent, general rising of the ocean's surface. Drowned Valleys of the Coral Islands. Many volcanic islands sur- rounded by barrier reefs have partially drowned erosion valleys. Dana regarded such valleys as proofs of crustal subsidence and Davis has adopted the same view. Some of the island embayments may be correctly explained in this way. For example, New Caledonia and the Fiji archipelago are generally regarded as located in a region of continental fragmentation. During the Tertiary period the eastern part of the Australasian continent was much faulted and otherwise deformed; the already dissected region sank below the sea and many valley bottoms became covered with water, scores or hundreds of meters in depth. Such submerged portions of the valleys were par- tially filled with detritus and shelly material. In some instances, abundance of mud doubtless prevented the sealing of the bays by coral growths. The outer stretches of these bays were thus subject to aggradation by waves coming in directly from deep water. The aggraded parts of the l)ays would have depths of from 20 m. to 50 m. below the sea-level of that time. Very slight additional erosion during the Glacial period, when sea-level was lowered, would be necessary to account for present depths below sea-level. Farther up the bays, the weak materials of the Tertiary deltas were attacked by the Pleis- tocene waves and the sediment thus stirred was dragged out into deep water by currents, both tidal and wind-driven. Many of the Fijian and other islands have been uplifted since the time of continental fragmentation; their Tertiary drowned valleys were similarly subject to cleaning-out by marine action during the Glacial period. Drowning of the resurrected valleys is a final, expected result of the late-Glacial rise of sea-level. (See also page 227.) However, a similar explanation cannot be admitted for most of the coral archipelagoes. These lie outside of the Fiji-New Caledonia 59 E. C. Andrews, Proc. Linn. Soc. New South Wales, Part 3, p. 786 (1903); C. A. Siissmilch, An Introduction to the Geology of New South Wales, Sydney, p. 153 (1911). GLACIAL-CONTROL THEORY OF CORAL REEFS. 225 area, where evidences of crustal uneasiness during the later stages of the Cenozoic are independent of the drowned-valley criterion. As shown by the charts, island embayments of the class here con- sidered have everywhere depths no greater than those anticipated if the bays are due to subaerial erosion during the Pleistocene time of lowered sea-level. In spite of aggradation by waves and currents, it seems inevitable that a few of the larger valleys should have greater depths, if their drowning were caused by crustal sinking in the Recent period. The failure of these greater depths tends from the first to weaken Dana's criterion. According to the subsidence theory, the still unsubmerged portions of the valleys should be of the one-cycle type, unless post-Glacial sinking were too slow for the complete submergence of the sides of the inner valleys which were excavated below the Tertiary floors because of the lowering of base-level in the Glacial period. According to the Glacial-control theory, the drowned valleys in general should be of the two-cycle type, or should have been of that type at some stage since the first climax of glaciation. In attempting to test the two theories by the use of this principle, it must be remembered that in volcanic islands a broad erosion valley is not necessarily as old as an equally broad valley cut in non-volcanic rocks. Observers in the Hawaiian islands, for example, cannot fail to be impressed with the breadth, as well as depth, of obviously youth- ful valleys cut in the volcanic formations. An abundant development of amphitheatres and broad troughs is an early result of subaerial attack on the typical volcanic island. The causes for such abnormal physiographic development are not wholly understood. One of them may well be a special tendency to caving or landsliding along the walls of young valleys cut in volcanic masses. The latter are characteristi- cally interrupted by zones of weak ash-beds or weak scoriaceous phases of the lava flows — layers liable to water-soaking, with consequent danger of landslides. A second cause for rapid slumping is found in the unusual prevalence of vertical joints in massive lavas. Widen- ing of the Pleistocene gorges cut in the central islands may therefore be more pronounced than in the case of gorges simultaneously cut in continental rocks. Whatever be the reasons, valley-making in vol- canic islands gives results somewhat different from those usual in typical areas of the continents. The physiographic processes operat- ing on the oceanic volcanoes certainly need thorough study. At least until that is accomplished, it is not advisable to apply the continental chronometer to the islands. There, many broad valleys seem to be 226 DALY. young valleys ; a relatively broad bay may not be a drowned " mature " valley, if by that term is meant a valley of great absolute age. The point specially worthy of note is that one can not, in these cases, safely locate the original bottom of each drowned valley at the inter- section of the visible valley slopes, simply prolonged without essential change of angle to the horizontal plane. When the sea-level of the coral seas fell, during the first Glacial climax, the floors of the broader stream valleys were trenched by the now revived streams. An "edge," "shoulder," or break of slope was then formed where the Tertiary floor met the top of the incised Pleistocene trench. Assuming no crustal movement, this "edge" was a few meters higher than the present sea-level, and at first sight it seems necessary to expect the break of slope to be visible to-day. The failure to find such valley-in-valley remnants in the present topog- raphy of certain Pacific islands has led Davis to doubt the Glacial- control theory. Yet, if the inner valley were essentially completed during the early, Kansan stage of glaciation (probably the time of maximum ice in North America at least), it is unlikely that the "edge" would still be generally, if at all, preserved. Post-Kansan time has been long enough for the mature dissection of the Kansan drift. The rock-material forming the "edges" of inner valleys must have been somewhat weak- ened by weathering; otherwise no "edge" would have been devel- oped, since the widening of such valleys depends on the preliminary weathering of the rocks in the valley sides. Post-Kansan time, fav- ored by the rapid rock-decay and heavy rains characteristic of the tropics, seems long enough to have largely or quite obliterated such minute features as these valley-in-valley " edges." Upstream, each revived river or creek must have speedily cut a distinct, narrow gorge in the floor of the Tertiary valley, just as the rivers of England and many other countries have cut young gorges during the Pleistocene. Those gorges should still exist, but, in general, they must be largely filled with post-Glacial alluvium and be thus invisible at the present surface. That the Glacial period was long enough for the excavation of inner valleys 50 or 60 m. deep, is not an extravagant assumption. If the Antarctic ice-cap, the last surviving one of great size, was also the first to form, the axial parts of the valley floors of the islands, except the lower stretches now submerged, have suft'ered subaerial erosion during a period longer than all Kansan and later time. Reviewing the criterion, it appears, first, that some bays of central GLACIAL-CONTROL THEORY OF CORAL REEFS. 227 islands in the western Pacific are explained by the sinking of those islands. However, the dating of that subsidence is not yet established, and the actual bays may be due to the Pleistocene cleaning-out of unconsolidated sediments which had been deposited in valleys, drowned during the Tertiary fragmentation of the Australasiatic continent. Secondly, the Glacial period was long enough for some further deepening of the Tertiary valleys by subaerial erosion. Well marked "edges" of the resulting valley-in-valley topography should not appear generally in the central islands of the present day, if the " edges " were formed early in the Glacial period. Thirdly, the narrow rock gorges cut at the heads of the bays should be more or less com- pletely covered by post-Glacial alluvium. The drowning of stream valleys is not the only cause for embay- ments. In each case it must be determined whether the bay is due to irregular accumulation of volcanic products, to faulting, to volcanic explosion, or to erosion. In many instances the bays are clearly en axe with valleys cut by streams and are so located because of pre- liminary subaerial erosion. However, such bays may not all represent river valleys submerged by change of sea-level. Ocean waves usually tend to smooth continental coast-lines, faced by broad submarine shelves. The shelves have a double office. They furnish shallow platforms on which coastal detritus may be quickly aggraded to sea- level; and they lower the erosive energy of the surf by partly wearing- out waves from the open ocean. Smoothing of a coast-line is a direct function of offshore aggradation. As the latter is delayed because of great depth of water, the waves have a longer time to search out the weak places in the land mass attacked. The more steep-to the coast, the more powerful are the attacking waves. In both respects undefended volcanic islands, with very deep water close to their shores, are subject to specially rapid searching by the waves. Now the very existence of a main valley in a Pleistocene island implies that its flooring rocks were already somewhat weakened by weather- ing. The volume of rock above sea-level, per unit length of shore line, was smallest at the intersection of the shore line with the valley floor. For two reasons, therefore, the surging breakers must have tended to cut bays in the Pleistocene islands, just as they are now cutting baj's in the Algerian coast, in some parts of the North Atlantic coast, in Christmas island (Indian ocean), and elsewhere. ^° 60 See T. Fischer, Petermann's Geog. Mitt., p. 1 (1887); C. W. Andrews, Geog. Jour., 13 (1899) (map of Christmas island). 228 DALY. Central-island bays of this origin are probably shorter and less con- spicuous than those due to the drowning of stream -erosion valleys, but are none the less worthy of attention by the student of shore topography. Again, as noted on page 162, a complete analysis of the bay problem must take account of the possibility of a Recent shift of sea-level, owing to causes other than the melting of glaciers. Post-Glacial uplift of the sea floor, over extensive areas, has been proved. If it has not been wholly compensated by sinking of the ocean bottom elsewhere, a general rise of sea-level occurred in post-Glacial time. Such a positive movement would tend to drown the valley-in-valley "edges," the Pleistocene shore cliffs, and allied topographic features. A post-Glacial rise of a few meters is quite possible as the result of diastrophic processes. On the other hand, the level of the ocean to-day cannot be many meters from its position in the Pliocene period. (See page 198.) That conclusion follows from the facts expressed in the charts of the continental shelves. Whatever may be the shapes of the rocky terranes beneath the shelves, the surface of each shelf has surely been smoothed and greatly widened by waves and currents. Each repre- sents an embankment growing, like a delta, into deep water. Normal storm waves and ocean currents effectively transport bottom mud if the depth of water is 75 m. to 40 m. or less. Depths of 75 m. to 40 m. prevail in the outer half of each of the wider shelves throughout the ocean. The building out of the great embankments to their actual widths demands all the time from at least the mid-Pliocene to the present day. The continental shelves seem, therefore, to indicate nearly the same position for sea-level during the later Tertiary as for post-Glacial time. The Pleistocene shifts of level represent a com- paratively brief interlude, and there is no evidence that the major shifts of that period were essentially caused by any other process than glaciation and deglaciation. Finally, the absolute proof of bay-making by Recent subsidence would not establish general subsidence for all areas characterized by barrier reefs or atolls, nor would it invalidate the Glacial-control theory. Recent warping or faulting of the earth's crust, in moderate amount, is an obvious fact in the Fijis, in the Tonga archipelago, in New Caledonia, in Oahu of the Hawaiian group, and in some other oceanic localities. The elevated strand-lines of the uplifted parts have correlatives in the drowned valleys of the sunken parts. As above noted, the Pleistocene platforms have been simultaneously GLACIAL-CONTROL THEORY OF CORAL REEFS. 229 warped up or down, and the subaerial and submarine topography of the islands in warped areas have been accordingly explained. More- over, it will be noted that a moderate sinking of the surface of a central island does not necessarily imply subsidence of the earth's crust beneath. (See page 233.) Hence, several lines of reasoning show the grave danger of error in regarding drowned valleys as proofs of the general crustal subsidence postulated by Darwin and Dana for all barrier reefs and atolls. Ex- cept in regions showing deformation in Tertiary time, the present depths of the corresponding bays are no greater than those expected as the result of erosion in the Pleistocene time of lowered sea-level. Therewith a measure of positive support is given to the Glacial-control theory. Vastly more compelling is the assemblage of facts regarding the bathometry of barrier lagoons and of atolls. The whole physiog- raphy of reef-covered areas, both below and above sea, must be con- sidered in attaining a just estimate of the problem. Ninety-nine per cent of that physiography is submarine and it does not accord with the Darwin-Dana theory. In the writer's opinion, a few scores of drowned valleys in an admittedly unstable part of the western Pacific are far less worthy of emphasis; they can hardly be held to prove general crustal subsidence in the coral seas. Pleistocene Cliffing of Oceanic Islands. Objection has been raised to the new theory, on the ground that it does not agree with the observed topography of the shores between the bays of central islands. If platforms as extensive as the Maldive or Macclesfield banks were planed off by Pleistocene waves and currents, it is held that central islands generally should show strong cliffs at the shore ends of the erosion spurs. As a matter of fact, many of these islands are more or less cliffed. Sometimes the cliffs reach scores of meters in height, though usually the cliffs at the ends of spurs are only a few meters high.®^ That they are not often much higher would be a strong argument against one phase of the Glacial-control theory, if the Pleistocene islands were all of the same height and rock-strength. Their great variability in each of these respects has been described. The Pleistocene waves must have cut wide benches in the pre-Glacial mud banks, shell banks, and banks of loose volcanic debris. Just as clearly they have made little impression on the lavas of Hawaii or Tahiti. 61 For many examples see the splendid series of photogravures illustrating .Agassiz's various expeditions to the Pacific archipelagoes. 230 DALY. Consider the case of a late-Tertiary volcanic island, of normal type, composed of massive flows, with some interbedded layers of ash. Before the shore became well protected by growing corals, it was subject to some cliffing, a narrow bench being cut in the lavas. Out- side the rock bench was a narrow shelf constituted of the detritus washed from the land by streams and waves. On this composite terrace the corals settled, and, with the development of a fringing reef, the waves were no longer able to attack the island successfully. The killing of the corals in the Glacial period caused a resumption of wave- benching. The sea-level being then 50 m. to 75 m. or more lower than before the ice-caps were formed, the waves quickly benched the outer part of the terrace but soon discovered the hard lava underlying the terrace detritus. Thenceforth cliff recession must have been not only slow but increasingly slower, since the cliff grew higher as the line of surf advanced into the gently sloping volcanic foundation. Assuming the sea-level to have been then constantly 55 m. below its present position, the height of the sea-cliff would have to be more than 55 m., if any of that cliff should be now visible. The recession of the cliff to the point where it was 55 m. high might well occupy most or all of the time during which the sea-level had nearly its maximum depression. If the volcanic spur was cliffed to a greater height, post- Glacial weathering and washing might have much softened its crest, a few meters above present sea-level. Of course, the greater part of each Pleistocene sea-cliff, now below sea-level, is buried by post-Glacial detritus and reef material. Such a buried cliff, more than 20 m. high, has been demonstrated by borings through the coral reef and underlying mud at Brandy Bay, Sumatra. The cliff rock there is andesitic.^^ The Tertiary sea-clift's of the imagined island were subject to wast- ing through all the time following the original development of coral reefs on the island; that is, from the late Tertiary, or earlier, to the present day. Those cliffs would be expected to have been much softened in contour, if not wholly extinguished as distinct facets, before the present epoch. In general, the volcanic islands exist because they are composed of rocks that are relatively resistant to the weather and to wave abrasion. Their existence, as well as the usual absence of very high spur-cliffs, merely shows that the Glacial period was of limited duration. On the other hand, atolls give no direct evidence as to its duration. If 62 C. p. Sluiter, Petermann's Geog. Mitt., 1891, Lit. Ber., p. 46. GLACIAL-CONTROL THEORY OF CORAL REEFS. 231 the new theory is correct, the atoll platforms were probably not the loci of pre-Glacial atolls. There is, in fact, no apparent reason for holding that great atolls or barrier reefs ever existed in pre-Pleistocene time. The extent of the massive coral reefs in the Pleistocene islands cannot be definitely stated, though in no case may it have been many times greater than, for example, that of the fringing reef at Rodriguez island in the Indian ocean. (See Fig. 28.) In any case it is unwise to assume that massive reef-rock or any other strong rock capped the whole of any Pleistocene island, which was truncated to form the wide fiats bearing the present atoll reefs. In conclusion, the quantity of wave-benching implied by the Glacial-control theory does not appear to represent a fatal objection to that theory, if the varying nature of the Pleistocene islands and shoals is well appreciated. Biology of Oceanic Islands. Another published objection is that the existing faunas and floras of the oceanic islands do not accord with a theory involving general crustal stability for the ocean floors during the later geological periods. However, the literature of biogeography shows such a diversity of opinions on this subject that the criterion cannot be fairly described as now having any decisive value. Darwin, Hooker, Salvin, Griesbach, Engler, M. Wagner, Wallace, Peschel, Wolf, A. Agassiz, Stearns, Heller, Dall, and F. X. Williams have con- cluded that the organisms of the Galapagos islands are not such as to necessitate belief in a former connection of this purely volcanic archi- pelago with any continent. Baur, H. Milne-Edwards, Von Ihering, and Van Denburgh, after their biological studies, thought best to assume a former connection with South America. Inasmuch as sharp divergence of views affects one of the best known island groups, any definite opinion regarding the origin of island species throughout the wide coral seas can still have but little - value. Consensus among the expert investigators is reserved for the distant future. The lit- tle that is known about the matter does not appear to weaken the Glacial-control theory, nor to support the subsidence theory, as ap- plied to coral reefs in general. Difficulties of the Subsidence Theory. A discussion of all published hypotheses concerning the reefs will not here be undertaken. Murray's solution theory, formerly re- garded with favor by leading investigators, is now seen to be weak on the quantitative side and is discounted by recent discoveries in typical 232 DALY. lagoons. However, certain features of the popular subsidence theory may be reviewed, in order to show more clearly the special advantages of the Glacial-control theory, the only other one involving Recent change of sea-level v/ithin the tropics. Here again discussion will be facilitated by a certain amount of repetition in presenting salient facts. Its Alternative Statements. It is important to observe that the Darwin-Dana theory is not the only explanation through subsidence. The vievv' of those famous authors is illustrated in the following passage, taken from the last (1895) edition of Dana's "Manual of Geology,"' page 350. Taking the Pacific area of reefs as a type, he wrote: "If, then, the atolls are registers of subsidence, a vast area has partaken in it — measuring 6,000 miles in length (a fourth of the earth's cir- cumference), and 1,000 to 2,000 in breadth. Just south of the line there are extensive coral reefs; north of it the atolls are large, but they diminish toward the equator, and mostly disappear north of it; and,, as the smaller atolls indicate the greater amount of subsidence, and the absence of islands still more, the line AA [of his map of the ocean] may be regarded as the axial line of this great Pacific subsidence. The amount of this subsidence may be inferred, from the soundings near some of the islands, to be at least 3,000 feet. But as 200 islands have disappeared, and it is probable that some among them were at least as high as the average of existing high islands, the subsidence in some parts cannot be less than 5,000 feet. This sinking probably began in the Tertiary era." In postulating a general, prolonged sinking of parts of the sea- bottom, each 10,000,000 to 25,000,000 square km. in area, Darwin and Dana agreed. As to one leading point the principles of their books do not agree. Darwin indicated the possibility of one or more con- siderable pauses in the subsidence; Dana seems not to have considered that suggestion as worthy of emphasis. The necessity of assuming at least one very long pause, if the Darwin-Dana theory is to with- stand even preliminary criticism, will be noted in succeeding pages. But Gerland offered a very different version of the subsidence theory. According to him, the coral reefs do not show sinking of wide continuous areas of the ocean floor, but do show the indepenflent sinking of each island mass ("Sockel"). In each case the subsidence is quite local, but has taken place at thousands of different points. ^^ This idea is worthy of attention. Nearly all of the oceanic islands and 63 G. Gerland, Beitraege zur Geophysik, 2, 56 (1895). GLACIAL-CONTROL THEORY OF CORAL REEFS. 233 shoals seem to be of volcanic origin. Rising from a sea bottom, 3,000 m. to 7,000 m. deep, each volcano is very high in absolute measure and is also of notable area. The local extravasation of so much lava may well entail local, moderate sinking of the earth's crust. It is, indeed, possible that such sinking is very often caused directly by volcanic action on a large scale. ^* More certain is the fact that, after a great volcanic cone is formed, its upper and central part tends to subside. Probably owing to the slow compacting of its deeper tuffs and vesicular flows, as these are gradually heated and so softened by the magma of the central vent, the mass slowly settles. The loss of connate water in the deeper ash-beds, both because of the heating and because of mere dead-weight, is another cause for settling in a local area, below the highest part of the volcano. Actual subsidence of this kind is exemplified in the "volcanic sinks" located at many central vents of the first rank.^^ Possibly, therefore, some of the drowned valleys and other physio- graphic features showing submergence of volcanic islands are to be explained by local sinking to the extent of a few meters or a few scores of meters. Clearly such subsidences would have very different geological dates, according to the respective times of preliminary volcanism. Erosion valleys so drowned in pre-Glacial periods would be filled, below sea-level, with sediment, which would be cleaned out again by the Pleistocene waves. Conceivably, some broad bays in the existing central islands may have thus originated. However, Gerland's version of the subsidence theory does not account for the essential contemporaneity of the present atoll, barrier, and fringing reefs. That they have been developed nearly or quite in the same interval of time is shown by their sectional areas and their volumes, as measured, in each case, above the break of slope at the platform on which the crowning reef stands. If the simultaneous submergence of coral islands in general were really due to crustal subsidence, the Darwin-Dana postulate seems to represent the only possibility. Their view is the one now to be briefly examined. Uniformity of the Assumed Subsidence. As just hinted, the surface outcrops and volumes of the greater barrier and atoll reefs, measurefl from the levels of the lagoon floors, are respectively nearly equivalent in the Pacific and Indian oceans and in the neighboring seas. If these reefs were formed by subsidence, the earth's crust must have sunk at 64 R. A. Daly, Igneous Rocks and Their Origin, New York, p. 185 (1914). 65 R. A. Daly, ibid., p. 150. 234 DALY. a nearly uniform rate, throughout the enormous area described. Since all large-scale crustal movement, which has become well under- stood, is differential — one part of a crust block moving faster than other parts, the Darwin-Dana theory faces another strong antecedent objection. Their postulate also fails to account for the approximate equality of volumes characterizing the typical atoll crown and the typical fringing reef, each volume being measured above the break of slope at the platform. One can hardly assume that all coasts fringed with reefs have sunk in recent geological time at the rate supposed to explain the atolls; nor that most fringing reefs are located in rising areas. Somewhere or other, an area of essential stability must exist in the coral-sea region; and there, according to Darwin himself, the breadth of the fringing reef should be much greater than that of a normal atoll or barrier reef.^^ This consequence is not matched with fact. Alleged Proofs of Current Subsidence. So far as the writer has been able to cover the literature, no case is recorded where a region bearing an atoll or barrier reef has been shown, beyond question, to be now visibly sinking. At least some of the instances cited by Darwin and others have not survived destructive criticism. This fact does not disprove his theory, but it annuls one of the positive arguments put forward in support of the theory. Even if local current subsidence were demonstrated, the Darwin-Dana theory would not be specially favored, for local sinking, like local uplift, is to be expected on any theory. Darwin's related argument, derived from the "drowned" condition of the Great Chagos and other atolls, has already been discussed (page 213). It falls to the ground if the phenomenon is due to hurri- cane action and mud-control. Permanence of the Pacific Basin. Dana was a leading advocate of the antiquity of most of the depressions occupied by the present ocean. Many specialists in geological dynamics favor that view, at least as far as the Pacific basin is concerned. Assuming great antiquity for the Pacific basin, the Darwin-Dana explanation of its reefs implies a unique, or almost unique, behaviour of the intertropical part of the basin in recent geological time. The narrowness of the atoll reefs is interpreted as meaning relatively rapid sinking. According to each of the two authors, the floor of the reef-covered Pacific has sunk thousands of feet within a period which is probably not equal to 5 66 C. Darwin, Coral Reefs, London, 3rd ed., p. 43 (1889). GLACIAL-CONTROL THEORY OF CORAL REEFS. 235 per cent, of recorded geological time. Clearly, similar sinking could not have occurred often in the same area, during the remaining 95 per cent of that record. No reason for specially great and rapid subsidence since the beginning of the "coral-reef period "' (early Miocene?) has yet been given. If this late sinking were actually preceded by many similar ones in the same area, during pre-Cambrian and later periods, one must assume intervening epochs when the sea- bottom was upheaved ; so that the final depth of the ocean should be no greater than it actually is. Rhythmic diastrophism of the kind and scale demanded is improbable. It should have left traces in the bottom topography of the Pacific, which, however, seems to be lacking in such evidences. The antiquity of a deep Pacific basin may be false doctrine, but, so long as manj^ facts continue to require its assumption, the subsi- dence theory has to bear the heavy burden of explaining the recent character of the postulated sinking of the Pacific bottom. In less measure the difficulty also applies to the Indian ocean area of reefs. High antiquity for the basin does not exclude its progressive deep- ening, but the relative stability of most of its floor in the later Tertiary seems proved by the size of the larger reef platforms and other banks. In whatever way these plateaus have been formed, the process must have taken very much time, even measured by the geological scale. The Glacial-control theory holds that only the final touches in fashioning the platforms were applied during the Pleistocene. Banks of sand and mud and low islands of similar material or of weak rocks were then truncated by the waves. During the relatively brief Glacial period, the sea bottom was so nearly stable as to permit of the wide benching of such banks and islands. The whole period was but a minute fraction of recorded geological time, and therefore is likely to have been one of general sea-floor stability, if the ocean basin dates from an early stage in the earth's history. The problem is baffling because of insufficient data, but the conclu- sion remains that the subsidence theor}^ is at a disadvantage because of the difficulty of reconciling it with facts, independent of coral reefs and suggesting an immense age for most of the ocean basin. Small Maximum Depth of Lagoons. Dana remarks that his theory " explains all the varying depths of lagoons, from the condition of near obliteration to that of a basin one to three hundred feet deep." ^^ A few paragraphs be^^ond, he adds: "The coral-growing areas over the 67 J. D. Dana, Corals and Coral Islands, New York, 3rd ed., p. 272 (1890). 236 DALY. great lagoons of atolls and the barrier-bounded channels of the Feejees and other archipelagoes and those of the outer waters about islands or their barriers, show no tendency to grow with large depressed centres, but rather with flat tops, as vegetation might grow, or else with elevated centres .... It is only through continued subsidence under such conditions that the margins can be made to grow so much faster than the interior as to produce thereby a basin-like interior 50 to 300 feet deep." It is strange that his recognition of only 300 feet (91 m.) as approxi- mately the maximum depth of lagoons inside both atoll and barrier reefs did not lead Dana to question the subsidence theory more seri- ously. Darwin did anticipate this objection and tried to meet it, as shown in the following passages : " Another and less obvious objection to the theory may perhaps be advanced, namely, that, although atolls and barrier-reefs are supposed to have gone on subsiding for a long period, yet that their lagoons and lagoon-channels have only rarely come to exceed 40 and never 60 fathoms in depth. But if our theory is worth consideration, we already admit that the rate of subsidence has not ordinarily exceeded that of the upward growth of the massive corals which live on the margins of the reefs, so that we have only further to suppose that the rate has never exceeded that at which lagoons and lagoon-channels are filled up by the growth of the delicate corals which live there, and by the accumulation of sediment. As the filling-up process, in the case of barrier-reefs lying far from the land, and of the larger atolls, must be an extremely slow one, we are led to conclude that the subsiding movement has always been equally slow\ And this conclusion accords well with what is known of the rate of recent movements of elevation " And with respect to the whole amount of subsidence necessary to have produced the many atolls widely scattered over immense spaces, the movement, as alread}^ shown, must either have been uniform and exceedingly slow, or effected by small steps separated from each other by long intervals of time, so as to have allowed the reef-constructing polypifers to bring up their solid framework to the surface; and this is one of the most interesting conclusions to which we are led by the study of coral-formations." ^^ Statements of similar import have not been discovered in Dana's book. The multitude of new charts published since Darwin and Dana wrote their books, have essentially confirmed their generalization as 68 C. Darwin, Coral Reefs, London, 3rd ed., pp. 153-4 and 192-3 (1889). GLACIAL-CONTROL THEORY OF CORAL REEFS. 237 to maximum depths; it is not likely to be shaken by additional sound- ing. Yet, if the great atolls are due to the sinking of correspondingly extensive islands, it is truly incredible that the lagoon depth should nowhere greatly exceed 300 feet (91 m.). The subsidence in such a case has been estimated in terms of thou- sands of feet. As the initial fringing reef was converted into a barrier and that finally into an atoll, the broadening lagoon waters must cover a kind of encircling moat between the central island or shoal and the upgrowing exterior reef. The reef speedily reaching the sea surface through most of its length, the waves and currents of the open ocean can do very little toward aggrading the lagoon floor, which is little or not at all disturbed by any waves or currents generated in the lagoon itself. Some of the fragmental material flooring each lagoon is derived from the rimming reef. From the outer edge of that reef the heavy surf is constantly forming mud, sand, and loose blocks. Some of the detritus is thrown forward by the breakers, to compose reef islets or reef awash. A much greater proportion is dragged out into deep water. ^^ One often meets the statement that much material is washed over the reef into the lagoon, but obviously the amount so transported can be only a small part of the whole detritus. That fraction, together with material locally brought in through reef channels, must be deposited near the reef, for waves and currents inside the lagoon are usually powerless to move sand in water 40 m. or more in depth. Abrasion by the lagoon waves is very low. The resulting debris, of grain coarser than that of fine mud, is likewise deposited close to the reef. As stated in the Royal Society report on Funafuti (page 375) Sollas observed that coral debris forms " but an insignificant part" of the "sand" (loose material) flooring that atoll lagoon, the chief constituents being foraminiferal shells and calcareous algae. In the extensive interior of the lagoon, any clastic material derived from the main reef is mud, with a little sand distributed by occasional great storms. The filling and smoothing-out of the hypo- thetical "moat" about a subsiding island is evidently little aided by this mud. The coarser detritus should form a well-defined terrace slowly growing inward from the reef. Such terraces would also be expected as a consequence of the Glacial-control theory; they are, in fact, conspicuous, though narrow, in Curtis's wonderful model of Funafuti at the Harvard University Museum. Their volumes are exactly of the order demanded if the reefs are of modern origin. 69 Cf. V. Cornish, Geog. Jour., 11, 530 (1898). 238 DALY. As the island, of normal profile, sinks, the land detritus rapidly diminishes in volume. The heads of the island deltas retreat farther from the "moat," so that its filling cannot be essentially attributed to outwashing from the central island. The slowness of the filling-in process is further shown by the very common steepness of the inner reef slope (Fig. 35). Darwin speaks of such reefs as "like a wall." ^° Describing the Funafuti atoll, Gardiner WTites: "The bottom of the lagoon, if the shoals were re- moved, and the whole elevated, would be a great plain surrounded by a ridge sloping steeply up to a line of perpendicular cliffs broken only at the few ship's channels; on this plain the greatest heights would be from 20-30 feet." ^^ In his great monograph on the INIaldives, Gardiner describes the lagoon slopes of the reefs as "practically per- pendicular" and elsewhere states that this fact "is not consistent with La g oon 0 50 100 200 400 Figure 35. Section illustrating the common steepness of reef edges on the two sides. the possibility of the lagoon's having been filled in by detritus washed over their encircling reefs." ^^ Like many coral knolls dotting the lagoon floors, the inner edge of the main reef often shows soundings of 20 to 40 m., at very short distances from the reef edge visible at low tide. In such cases the rate of upgrowth for the living reef must have been greater than the rate of upgrowth for the sand terrace alongside. If the aggrading process is so inadequate at the source of supply, how much more inadequate is it to smooth the vast interior of the lagoon in any reasonable time! (See Fig. 36.) As shown by the normal deepening toward the lagoon centers and by the sparseness of the coral knolls, pelagic shells, coral-knoll detritus, and bottom growths would be still less effective in filling the "moat" 70 C. Darwin, Coral Reefs, London, 3rd ed., p. 67 (1889) 71 J. S. Gardiner, Proc. Cambridge Fhil. Soc, 9, 434 (1898) 72 J. S. Gardiner, Amer. Jour. Science, 16, 211 (1903). GLACIAL-CONTROL THEORY OF CORAL REEFS. 239 and smoothing the indefinitely varied floor of a lagoon surrounding a sinking island. One must conclude that the filling of the "moats," so that nowhere they shall be covered with water more than about 90 m. deep, means an exceedingly slow rate of subsidence; one may doubt that the amount of aggradation, matching the known flatness of the lagoon floors, is a physical possibility even if that process occupied all Miocene and later time. But, on account of the narrowness of reefs, Darwin himself inferred a geologically rapid rate of sinking for them.^^ The subsi- dence theory therefore faces a serious dilemma, and none of its up- holders has yet offered a reasonable explanation of the "extremely slow" (Darwin) filling-up process accompanying a rapid subsidence. Apparently the only conceivable way out of the dilemma is to assume 0 50 100 _6p0 |v) Figure 36. Section through a typical coral knoll in the "drowned atoll" Tizard bank, China Sea (after W. U. Moore and P. W. Bassett-Smith), show- ing steepness of the knoll slopes and the distribution of corals. L, live coral; D, dead coral; S, sand. that the filling of the "moat" occurred during a long pause in the sinking, while the narrow reef rim is due to a recent renewal of sub- sidence. Since the atolls and barrier reefs of all the world have virtually the same features, it would follow that, throughout the greater part of the tropical Pacific, the Indian ocean, and the East Indian archipelago, there was a simultaneous, long-continued pause in subsidence; and that the pause was followed by a recent, rapid sinking to about the same amount, everywhere in the same world belt. The manifest improbability of the assumption shows the necessity of adding some other one, as yet unimagined, if the older theory is to account for the maximum depth of lagoons.^* 73 C. Darwin, Coral Reefs, London, 3rd ed., p. 43 (1889). 74 Darwin did not believe that crustal movement could be uniform even over the comparatively small area represented by the West Indian sea (C. Darwin, Coral Reefs, London, 3rd ed., p. 269 (1889)). How much more improbable is the view that the vastly larger Indian and Pacific areas occupied by atolls and barrier reefs, have been uniformly depressed in recent time, i. e., after the "long pause " ! 240 DALY. Flatness of Lagoon Floors; Comparison of Depths in Lagoons and on Banks. The comparative flatness of lagoon floors has been empha- sized as a general fact of the first importance. It is a feature expected on the Glacial-control theory and quite unexpected on the older theory, unless the auxiliary hypothesis of a very long pause in subsidence be accepted. With the last suggestion goes the correlative hypothesis that the reef rims now visible at the water surface, above the great platforms, are very modern affairs, relatively high and narrow be- cause formed by a recent renewal of sinking. Before its last assumed sinking, the plateau, now crowned with an atoll, was flatter than at present because then it lacked the existing main reef as well as the coral knolls, also recently upgrown, in the lagoon. Therefore, at the close of the long pause in subsidence, the plateau was nearly or quite as even-topped as if it had been truncated by wave erosion. In this connection a passage in Darwin's book is significant: "Di- rectly north of the Laccadives, and almost forming part of the same group, there is a long, narrow, slightly-curved bank, rising out of the depths of the ocean, composed of sand, shells and decayed coral, with from 23 to 30 fathoms (42 to 55 m.) on it. I have no doubt that it has had the same origin with the other atoll-like banks; but as it does not deepen towards the centre, I have not coloured it. " ^^ This is one of the reefless banks (Fig. 37) referred to in an earlier section (page 198). For some reason corals have been unable to raise an atoll crown on this bank and its form could not have been changed essentially since the end of the assumed long pause in subsidence. If all the banks of the coral seas had a similar history, their topography would not sug- gest recent submergence to an amount greater than 55 to 90 m., or possibly 100 m. The principal ground on which the theory of deep subsidence has been founded would be entirely cut away. Yet Dar- win held that this Indian ocean bank did have the same origin as the banks crowned with atoll reefs. The subsidence theory was invented chiefly to explain the ground- plans, maps, of the surface reefs; that is, one topographic element was emphasized, and the evidence of submergence is certainly good. But the same principle of questioning the existing topography — por- trayed in charts, full of soundings — suggests as clearly that submerg- ence has been strictly limited. A section across the southern part of Suvadiva atoll (Fig. 16) illustrates the common, pronounced break of slope between the reef 75 C. Darwin, Coral Reefs, London, 3rd ed., p. 247 (1889). GLACIAL-CONTROL THEORY OF CORAL REEFS. 241 185 :92 34 205 ,'28 25 \ ,'

/3pW) has the same minimum 5, for the two parentheses take on con- jugate imaginary values for conjugate imaginary values of p. Let a circle be described with centre at the origin and radius so large that, in the part of the sector (4) which lies outside it, ] e-""^ I < P; and suppose p restricted for the moment to this part of the sector. If p is exterior to all the little circles, (p (p) -f e-""^ is surely not zero, because of the inequalities (6) k(p)|>5, k-'"^|\n{x2— x\) ^ dx' However large k may be, and however small numerically the negative quantity X2 — xi, the series which has this last expression for its 7ith term is convergent. It follows that the series obtained by differ- entiating (8) term by term k times is uniformly convergent ^* for 0 ^ .T ^ xo; and as .^2 may be any number between 0 and Xo, the theorem is proved. -^^ The question is now inevitable: Is it possible to develop every analytic function in a uniformly convergent series of the form (8), with the restriction, perhaps, that the function shall vanish at the ends of the interval? As in other problems of this sort, it is easy to deter- mine what the values of the coefficients must be, if such an expansion exists. The formal series being given, an answer to the question may be expressed as follows : Theorem II. // N is any positive integer, however large, it is possible to define a function f{x), analytic throughout the interval 0 ^ a; ^ ir, and vajiishing with its first N derivatives both for x = 0 and for x = r, such that its formal expansion in a series of the form (8) diverges at every interior point of the interval. The system adjoint to (1), (2) consists of the differential equation dx"^ and the boundary conditions «(0) = 0, (j(7r) = 0, »'(7r) = 0. Its characteristic numbers are the numbers p„, and its characteristic solutions have the form d2) Cr,{x) = ^P"(^--'-f a;c"''»(^-'^) + w-f?"'""'^-'^'. 14 Of course the terms of the series are continuous and in fact analytic func- tions of X. 15 Professor Birkhoff points out that it may be shown by a slight modifica- tion of this argument that the series converges uniformly throughout a region of the complex x-plane including the segment (0, Xo) of the axis of reals in its interior, and so represents a function analytic for 0 ^ x < Xq. 390 JACKSON. If a function /(.c) is the sum of a uniformly convergent series (8), the coefficients in the series must be the numbers (13) On = —^ A Un{x)Vn{x)dx The number N being given, let q be the first number greater than or equal to N which is congruent to 2 (mod. 3). We shall examine the terms of the formal series for the special function .T5+l(7r-x)«+2 (14) fix) TT^ + ^X^+l)! This function, a polynomial in x, is of course analytic from 0 to tt, and vanishes with its first q derivatives at both ends of the interval, while the (q + l)th derivative takes on the value 0 at the point tt and the value 1 at the point 0. To determine the order of magnitude of the numerator in (13), when f{x) is the function just defined, let us break up the integral into three, corresponding to the three terms in the right-hand member of (12), and consider for the moment the second of these. The index n may be supposed so large that pn is real and positive. Integrating by parts q -\- 2 times, we find that J"' /—IN «+2 / (.t) e"""^''—^^ dx = g-^Pn-T 0 \ wpn J + f ^1 '^^ fV ^'+^^ (^) e """^^-'^ dx, \ copn J J and hence, by one more integration by parts and the application of obvious inequalities, that L f (.r) e'^''''(^-'^) dx — g- a,p„^ 0 ' \wpnj < Pn g+3 c := p ^prni Pn^+-' ' where c is independent of n A corresponding inequality holds for the third integral, with the same value of c. As for the first, it ob- viously remains finite ^^ as n becomes infinite, and so it is certainly true that Jo / (x) gP''^^-'^) dx I < — x^ e*"'"' 0 I Pn^'^'^ 16 As a matter of fact, of course, it approaches zero. EXPANSION PROBLEMS. 391 from a certain point on. When n is large, therefore, Jf (x) Vn (x) dx — \ col 0 L \ ^Pn — 1 \ 9+2 / _ X \ 3+2 g-a.p„,r_|_^2 OO^Pn < 3c Pn 9+3' iPnTi Asq ^ 2 (mod. 3), the expression in brackets reduces to 9+2 / \ /_ IN 9+2 ^2 N( f J f e-'-Pn'^+e-'^P"'^) = 2( J e^''"'^cos — pnTT. V3 (\ V3 \ (\ V3 cos — pnTT = C0s[-+ W+ — enJTT = ± COS f - + — - fn ] TT, lim 71= 00 V3 cos — PnTT V3 2 ■ It follows that as soon as n is sufficiently large, the absolute value of the bracket is greater than 3 i a nd hence that from a certain point on j^f(.x)Vn(,x) dx > 1 Pn 9 + 2' 2« ipnir With regard to the magnitude of the denominator in (13), it suffices to observe that, when pn is real and positive, \un{x)\^3 e^""'', I Vn (x) \ ^ 3 e^'""('^- ^\ and so I Jo Un (x) Vn (x) dx \ ^ 9 IT 6^""^. This inequality, with the one obtained for the numerator, shows that as n becomes infinite \an\ > Ci p.«+2' where Ci is positive and does not depend on n. From this relation it is apparent that the general coefficient does not approach zero rapidly enough to enable the terms of the series l^anUn{x) 392 JACKSON. to remain finite as n becomes infinite. Let xo be any particular value in the interval 0 < .tq < t. Modifying the form of (7), we may write Un{xo) = e- p""^ -{- 2 c''"""^ sin (-^ pnXo - '^j. The sine in the right-hand member may be zero or nearly zero for some values of n, but there are certainly infinitely many values of n for which this is not the case. For 1^^- P„ + 1 ^0 - gj - ^^— Pn .To - g J = ^0 + — (en + 1 - en) .To, which ultimately becomes practically equal to .tq, so that the two parentheses on the left-hand side can not both be nearly equal to integral multiples ^"^ of tt. It follows that for infinitely many values of n V3 'sm, . ^,...>, „ 0 Un{xo) I > €36^""''°, and I anUn{,Xo) I > Q+2 Pn where Co, c^, and C4 are all positive and independent of 71. As the right- hand member of the last inequality becomes infinite with n, the proof of Theorem II is complete. It may be remarked that no use has been made in this demonstra- tion of the fact that all the characteristic numbers are given by the asymptotic formula that we have employed. The proof would be unimpaired if there were infinitely many others. It is sufficient to know that there exist an infinite number of real characteristic values distributed in accordance with the formula. And this latter fact is 17 More precisely, if 5 is the smaller of the quantities .xo, ir — xo, the index n may be made so large that and then the difference between the two parentheses is between ^5 and tt — ^5, and one or the other of the parentheses must itself differ from the nearest integral multiple of tt by at least \h. r5 0:0 <-, EXPANSION PROBLEMS. 393 readily established by writing the characteristic equation in the real form e-p^+2e'^p^s\n(^pTv - ^ j = 0, and observing the changes of sign of the left-hand member. II. Differential Equation of Order v with v — 1 Specialized Boundary Conditions. The properties of the series studied in the preceding paragraphs are typical of a large class of expansion problems connected with differ- ential equations of order higher than the second. We shall consider now the differential equation ^^ (15) ^ + 2>2(.f)^^+... +P.(^)" + P''w = 0, .^3, where the coefficients p are real or complex functions of the real vari- able X, continuous with their derivatives of all orders, in an interval which we may still take as the interval 0 ^ .r ^ tt. The v linearly independent boundary conditions associated with this equation are to be subjected to the essential restriction that v — 1 of them involve the values of the solution and its derivatives only at the point 0, while the remaining condition may involve both end-points or the point tt alone. ^^ These conditions can be reduced by linear combination to the form 1 8 It is understood that any or all of the derivatives of orders from 0 to ^ — 2 may occur in the equation; only the {v — l)th is assumed to be absent. As is well known, a more general differential equation can be reduced to this form by a change of variables; cf. Birkhoff II, p. 373, footnote. The differ- ential equation which Liouville considers in the paper II is specialized in a different way. 19 Of course it amounts to the same thing, if a number of the conditions involve the derivatives at tt in such a way that these can be eliminated from all but one of the conditions by linear combination. A similar remark is applic- able to the cases treated later. The conditions must include at least one which involves the point tt, otherwise they are equivalent to the set it(0) = u'ifi) = ... = w(''-i) (0) = 0, and no characteristic solutions are possible. In the system treated by Liouville, the j/th condition involves only the point tt, and the coefficients are subject to further restrictions. 394 JACKSON. ks-1 Ws (u) = u^'^^ (0) + Z «y w*'^ (0) = 0, s=l,2,...,v-l, ('«) i:-r W, (u) = uik.) (x) + 2: /3,wW (x) + 2: aju^^^ (0) = 0, j=0 ,7=0 where I' - 1 ^ A;i > A;2 > • • • > ^\-i ^0, v - 1 ^ ^^ ^ 0; and we shall assume them in this form. It would be possible to simplify them still further, if there were any object in doing so. The coefficients a^-, aj, /?y, are any real or complex constants, and in particular some or all of them may vanish. The facts which we shall need concerning the form of solutions of the equation (15), apart from the question of any particular boundary conditions, may be taken from the papers of Birklioff already referred to. Following his notation, we shall use the symbol [a], where a is any constant, to represent a function of p = re^*, or of p and a second variable, which approaches the limit a uniformly with respect to 6, or with respect to d and the other variable, when r becomes infinite; the same symbol will be used to represent as many different functions of this sort as may occur. Furthermore, the roots of the equation w" -\- 1 = 0 are denoted by wi, W2, . . . , i^s^[l], and at the same time (18) £, ys{x,p) = {pwsYe'^-'s- [1], k = \,2,. . .,v - \. These solutions have continuous derivatives of all orders with regard to .T, for each value of p. If I is any one of the numbers 0, 1, . . . , 2^ — 1, and p is restricted to the sector Si of the complex plane in which lir^ ^(Z+l)7r — ^ arg p ^ ^^ —^ V V the choice of this fundamental system may be so made that the func- tions t/g(.T, p) and their derivatives with regard to x shall be analytic functions of p for each value of x, at least throughout the part of the sector exterior to a circle of sufficiently large radius about the origin. EXPANSION PKOBLEMS. 395 The last statement is still true if the sector Si is replaced by another one, bounded by lines parallel to the sides of Si, and including the latter sector wholly in its interior.^" Cutting off a part of the new sector, if necessary, by a large circle about the origin, we will let Ti stand for the remaining portion, throughout which each function Vs {^> p)> with its derivatives, is analytic. Of course the regions Ti will overlap each other to some extent, and where they do, two differ- ent fundamental systems of solutions will have been defined; but as we shall not have occasion to use the solutions in different regions Ti simultaneously, no confusion will result. As the problem involves p only in the I'th power, attention may be restricted to one fth of the p-plane, consisting, for example, of the sectors Si, and iS2;'-i; if there should be characteristic values on the boundary lines for which argp = ± -J V the values on one of these lines are to be disregarded. We have to deal, then, only with the regions To and Top-i, and a third region U which is finite in extent. It follows at once from the general existence theorem that in this finite region a fundamental system of solutions of the differential equation, and so the determinant whose vanishing gives the characteristic values, may be taken as analytic in p. The region can contain only a finite number of characteristic values, and these need not concern us further. Let us consider the region To. Let 101 = 6", W2 = e" , while W3, . . ., wp, denote the remaining I'th roots of —1, arranged, for example, in cyclic order, though the convention as to the last v — 2 subscripts is quite inessential.^^ Geometrically, the requirement is that when the roots are represented in the complex plane, as the vertices of a regular polygon of v sides inscribed in the unit circle, wath 20 Birkhoff III, p. 117. The assertion is readily justified by replacing the parameter p, in the argument of his paper I, by p + r, where r is a constant. 21 Our use of subscripts here is different from Birkhoff 's; and in its next stages our discussion, though following his in general method, is vitally affected by the special nature of the boundary conditions. 396 JACKSON. a vertical side at the extreme right, the lower and upper ends of this side are denoted by u'l and w^ respectively. Suppose that 0 is a quantity in the interval The representation of the numbers Wie^\ . . ., Wve'^, is obtained by rotating the original polygon through an angle 6. It is clear that as d varies over the interval, the argument of the complex quantity {wi—iVs)e^^, where s is any subscript greater than 2, has a positive mini- mum. The latter statement remains true if 6 is allowed to range over an interval V where e is a sufficiently small positive quantity. Now if p = re^^ is in Tq, 6 does satisfy the last inequalities, at least when /• is sufficiently large, and this has the consequence that as r becomes infinite, the real part of {loi — Ws)e^^ remains uniformly away from zero, the real part of pa{wi — Wg) becomes uniformly infinite, if a is any positive real constant, and e''«(«'s— ^0 uniformly approaches zero. Furthermore, the real part of Wie^'' itself remains uniformly away from zero, and g_p«jio approaches zero uniformly, even when multiplied by any posi- tive power of p. These facts will presently find application. The determinant A(p), by the vanishing of which the character- istic numbers are recognized, has for the element in its 5-th row and ^-th column the quantity WXyd- By substitution of (17) and (18) in (16), it is found that Ws(ijt) = {pwtYsil], s=l,2,---,u-l, W, {yt) = (pWiY" C^t- [I] + a, (piVtY [1], where a^ is the last coefficient aj that is different from zero in (16); if every aj is zero, a^ is understood to have the value zero likewise. The factor p^«, which is common to all the elements of the 5-th row, 5=1, 2, . . ., V — 1, does not affect the vanishing of A, and may be divided out; and we will divide out also a factor p'^v^pwi^ from the elements of the last row, though this factor is not explicitly apparent in all the terjns. The problem is thereby reduced to that of the van- ishing of the determinant Ai = EXPANSION PROBLEMS. 397 wi^"!!] 1 f W2*^''e'"^(''^"") [1] \ + a,p^-^''wi''c-''"'"^[l] J \ +a«p''-^>W2''e-''"'"^[l] J ■"■ / w/j'ePT(M'i'~«'i) [1] the exponent k — /.> in the last row of this determinant may be posi- tive, negative, or zero. It is at this point that the import of the remarks made at the close of the last paragraph but one becomes apparent. As p becomes infi- nite in the region Tq, the first element in the last row of Ai approaches Uiiiformly the limit Wi^^, and the elements from the third to the last uniformly approach zero. The second of the two terms which make up the second element of the row is also uniformly infinitesimal. Let us expand the determinant according to the elements of the last row, and examine the cofactor of the first element. If we suppose each bracket symbol replaced for the moment by unity, the cofactor is a determinant whose value is surely different from zero, not because its rows consist of like powers of a set of distinct quantities (it is not assumed that all the exponents kg are sxccessive integers), but because, after division of the s-th row by ^^2*^ s = 1, 2, . . ., i/ — 1, an operation which of course does not affect the vanishing of the determinant, each column may be regarded as consisting of like powers of the v — 1 distinct quantities iVi^'"'^, iV'f^'^, . . . , loi'^v-i^ the exponents of these powers in the several columns being the integers from 0 to v — 2 inclusive, in some order; the numbers Wz/ W2,. ■ .w^/ w^, are even powers of w^. If the brackets are restored, the cofactor is a function which uniformly approaches a limit different from zero when p becomes infinite. Similar reasoning, with a like conclusion, applies to the cofactor of the second element. Representing the limiting values of the cof actors of the first and second elements by 5i and ^2 respectively, we see that A] has the form (19) Ai = [5i wi^^] + [52 w-^^] e "'^^ ""'=-"'1) ; the terms in the expansion which involve the last v — 2 elements and the second part of the second element have been merged here in the first bracket symbol. If A2 is the expression which results when the bracket expressions in (19) are replaced by their limiting values, the equation A2 = 0 can 398 JACKSON. be solved explicitly, having infinitely many roots given by the relation PnV (W2 - wi) = log (^- ^^^j + 2 rnri, where the first term on the right is understood to refer to a particular determination of the logarithm, and n is any integer. It appears from the determinant expressions for 5i and 59 that each of these numbers is either the conjugate or the negative of the conjugate of the other; consequently they have the same absolute value, SiWi^" and 52^2*" likewise, and the above logarithm is a pure imaginary. Hence, noting that W2 — Wi = 2i sm -•, V we may write Pn sin (jr/vy where ^uq is a real constant. These roots are all of the first order, for the derivative of ZX2 never vanishes. The passage from the roots of Ai to those of Ai is effected by reasoning of the same sort as that used in connection with the simple case of the third order. The conclusion is as follows: The equation A = 0 has an infinite number of roots, which can be written in the form (^^^ "'= sin(x/.) ' n = g,g+l,g + 2,..., where g is some integer, positive, negative, or zero,^^ and ^'^ (21) ^'"^ en = 0. With the exception of a finite number at most, the roots are of the first order. Of course it still remains to consider the possibility of the existence of roots in the region Tov^i. But an argument analogous to that just completed shows that, at least from a certain point on, this region can 22 Of course it would be possible to adjust mo so as to make g equal to zero. 23 For the sake of convenience, a definition of «„ is given here which does not exactl}' correspond with that used in the previous special discussion. EXPANSION PROBLEMS. 399 contain roots only in the strip along the axis of reals which r^^z-i has in common with Ta, so that a finite number of roots at most are to be added to those already found. By a change in the value of g, if necessary, the expression given above can be made to account for all the characteristic values. It is evident a priori that if the coefficients in the diflFerential equa- tion and in the boundary conditions are real, the conjugate of any characteristic value will also be a characteristic value, and it follows that all the characteristic values with the exception of a finite number will be actually and not merely asymptotically real. We do not restrict ourselves to this case, however. To each characteristic value from a certain point on corresponds just one characteristic function. For, ^* as we have already had occasion to observe, the determinant Ai has first minors which are always different from zero when p is sufficiently large, and it follows that the same is true of A. An expression for the characteristic func- tion un{x) may be obtained by replacing the t-th. element in the last row of A(pn) by ?/((.r, pn), t = I, 2,. . ., v; it is seen that the result of subjecting the expression so obtained to any one of the operations Wg, s = 1, 2. . ., V — I, may be written as a determinant in which two rows are alike, while the result of the operation Wv is the vanishing determinant A(p7.) itself. Inasmuch as a factor independent of x is immaterial, a characteristic solution may be obtained equally well by inserting the functions i/f {.v, pn) in place of the elements of the last row of Ai. It is the latter more convenient form of solution that we 24 As a matter of fact, the mere observation that the characteristic value is a root of the first order for A suffices to show that there can not be more than one characteristic function; cf., e. g., Goursat, Annales de la Faculte des Sciences de I'Universite de Toulouse, deuxieme serie, 10, 5-9S (1908); pp. 37-38; Birkhoff III, p. 118. A proof is as follows: The derivative of A with regard to p may be written as a sum of v determinants, in each of which v — 1 rows are identical with the corresponding rows of A, while the elements of the remaining row are the derivatives of the corresponding elements of A. The cofactors of these derivatives in the several determinants are of course first minors of A, except as to algebraic sign. If p„ is a characteristic value to which two linearly independent characteristic functions correspond, the first minors of A are all zero for p = p„, and the derivative of A consequently vanishes. That is, there can not be two independent characteristic functions for Pn unless p„ is at least a double root of A. This general fact will be assumed in the later discussion of other sets of boundary conditions. It is clear that similar reasoning shows that for a root of anv order of multiplicity the number of linearly independent characteristic functions can never be greater than the order indicates; cf. Goursat, loc. cit., p. 44. I do not know whether the proof indicated here is to be found in the literature or not; it is the one which Professor Bocher gives in his lectures. 400 JACKSON. shall use. Or, rather, we shall use this form if the numbers 5i and 82, which were introduced earlier and figure now in the coefficients of yi and yo, are conjugate imaginaries, but shall multiply it by i if 5i is the negative of the conjugate of 62. We have then in either case V (22) un (.r) = Z [(Jt] e'"'"''^ where di and do, in particular, are conjugate imaginary quantities different from zero. It is clear that if x is positive and n is large, all the terms after the second will be insignificant in comparison with either of the first two. Let where Jh and I12 are real. Then (23) di e""""^ + di e""^^^ = 2e^' + p,.xcos{^/r) ^^g | f^^ _^ ^^^ gjj-^ (^/j,) | _ This identity is of course none the less valid for the circumstance that the numbers pn are not necessarily real. Suppose that a succession of constants ^^ an, 71 = g, g -]r 1, g -^ 2, . . . , is such that anUn(x) remains uniformly finite throughout an interval included in (0, tt), or, in symbols, (24) I anUn (x) I < G for Xi < X < xo, where 0 < .I'l < .I'o ^ tt. It is to be shown that throughout the interval 0 ^ .r < xo the series (25) Y. ttnUnix) n = g must converge and represent a function possessing derivatives of all orders. Because of the relations (20) and (21), the difference between the last factor in (23) and the expression (26) cos { h2 + (mo + n) x ] 25 This notation is adequate only if all the characteristic values from the very beginning are simple roots of the determinant equation. In order not to raise questions irrelevant to the main issue, it will be assumed, in the contrary case, that the series (25) begins with such a value of the index n that only simple roots are involved. In the extension of Theorem II, where the formal expansion of a definite function is under discussion, it is possible to assign values to the first terms of the expansion immediately; cf. Birkhoff II, p. 380; but we shall not be in any way directly concerned with these terms. EXPANSION PROBLEMS. 401 approaches zero, uniformly for 0 ^ .t ^ tt, when n becomes infinite. When 71 is sufficiently large, hi + (/xo + n)x will vary by more than 2x as X increases from .I'l to .To, and there will surely be an intermediate value X = X7i' for which the expression (26) is equal to 1. If this value is substituted for x in (23), for the successive values of n, the last factor remains from a certain point on numerically greater than a positive constant. Independently of this particular choice of x, the quantity gPnX COS (ir/f) I g(/io +n) X cot (tt/j') _- a^nX COt (jr/y) remains numerically greater than a positive constant beyond a certain point. Consequently the whole expression (23) remains for .r = a-/ nu- merically greater than a constant positive multiple of g(w+n)a;/cot(ir/:')^ and the same is true of Un{xn'). The inequality is strengthened if xn in the exponent is replaced by .I'l; it follows from (24) that (27) \an\< c'e -('«+'») ^^ cot (^/.)^ where c' is independent of 7i. We have now to consider the order of magnitude of the derivatives of un{x). Suppose X restricted to the interval 0 ^ .r ^ X2, where X2 < xi. If k is one of the numbers 0, 1,. . ., v — I, it is at once deduced from (17) and (18) that (28) ^-3 '> • • • > A-; ^ 0. For each value of p, a fundamental system of solutions of the equa- tion (29), 2;j(.T, p), s = 1, 2,. . ., V, can be assigned in such a way that (31) ^^zsix,p)^{-pws)''e-'>^sxiii k^O,h...,p-l. It is important for us now, however, to notice that the bracket symbol may be given a much more precise interpretation than was done before. Both the relations (31) and the relations (17), (18), remain true ^° if [a], where a is any constant, is understood to mean an expression of the form I 'AiC^O , hi'i-) , , 'Am-i(.r) E(x, p) p p^ p^-^ p^ where m is an integer that may be assigned arbitrarily in advance, the functions \pj{x),j — 1, 2, . . ., m — 1, are continuous with their de- rivatives of all orders for 0 ^ .r ^ tt, and E{.v, p) is a function of x 29 The statement may also be proved by observing that on the contrary assumption the adjoint conditions could be reduced to the form i'(7r) = v'iir) = . . = I' (""!) (tt) =0, and the adjoint system could have no characteristic values. 30 See Birkhoff, I, II. EXPANSION PROBLEMS. 405 and p = re** which remains finite uniformly in x and 6 when r becomes infinite. We shall use the symbol with this meaning from now on; the functions \{/ and E will be different in different cases, of course, and are to be replaced on occasion by quantities independent of the vari- able X. The solutions described in (31), regarded as functions of x, are continuous with their derivatives of all orders. Their analytic character as functions of p need not be further specified, for the deter- mination of the characteristic values does not have to be repeated for the s.ystem (29), (30). It follows from a general theorem ^^ that the characteristic values for this system are the same as for the system (15), (16), and that the number of linearly independent char- acteristic solutions corresponding to any one characteristic number pn is the same for both systems. From a certain point on, the system (29), (30), has just one characteristic solution vn(x) for each number pn, and the corresponding coefficient in the formal expansion of a function f{z) is /^ f{x)Vn{x)dx (32) -^ ^ / Un{x)Vnix)dx We shall show how to define analytic functions /(.r) for which the general term anun{x) does not remain finite. Except for an arbitrary factor independent of x, which will be chosen later, the characteristic solution V7i{x) may be represented by a de- terminant made up as follows: The t-th element in the first row is Zt (x, pn), which, by (31), has the form e-"""""" [1]. The ^th element in the 5-th row, s = 2, 3, . . . , v, is Vs{zt), or {-PnWtf''e-p"'^t''{l]. On division of the s-th. row by (—pn)^', s = 2, 3,. . ., v, and multi- plication of the t-th. column by e''"^t'^, t = 1, 2, . . ., v, this determinant is reduced to the form gPnWliw—x) Ml gp„W2 (v— X) Ml ... gPnWpi-K— X) Ml Wi^'-[1] wo^-'^l] ••• w/"-[l] 31 See Birkhoff II, pp. 375-376; Bocher, loc. cit., p. 407. 406 JACKSON. the first row be denoted by 5'i and 5'2- As in the corresponding case which was discussed earlier, it is seen that neither of these numbers is zero, and that each is the conjugate or the negative of the conjugate of the other. We shall denote by vn{x) in the one case the value of the determinant as it stands, in the other, the determinant multi- plied by i. Then we have (34) vn{x)= Z [d't]e'"'^ti--=^), 1=1 where d\ and fZ'2 are conjugate imaginary quantities different from zero. Let us examine the order of magnitude of the denominator of the fraction in (32), using the expressions (22) and (34) for the character- istic functions. In consequence of (20), (21), there is a real number 7, independent of n, such that ^^ (35) R (pnWt) < (Mo+n) cot (tt/p) +7 for t = 1 and t = 2, and a fortiori for ^ = 3,. . ., v; the constant fM)Cot{Tr/v) might be merged with 7. Hence there exist numbers D, D', and c, all independent of n, such that I Un{x) I < De'^ ^cot (x/.)^ I ^^(^^) [ < 2)'e" ('^- ^) ^^o* ('^Z"), and ^^^^ \ /q '^n{x)Vn{x)dx < cfi"'^ cot (,rA) _ Now let us define a function /(.r) by the formula (14), denoting by q for the present an arbitrary positive integer, and seek information about the corresponding value of the integral in the numerator of (32), with the aid of the expression (34) for vn{x). By integrating by parts, as at the corresponding point in the preceding section, with attention to the special values of /(.t) and its derivatives at 0 and tt, and supple- menting this process by the application of (20), (21), and (35), it is seen that / V(a;) e"""" ^"""^^ dx — [ -^— ]' e"" Jo •'^ ^ \pnwj _ f 1 Y+2 ^ J^ pTlv cot {it/ v) where Ci, like every quantity denoted by the letter c with a subscript throughout the remainder of this section, is independent of n. 32 By R{z) is meant the real part of z. EXPANSION PROBLEMS. 407 It is a diflference between the present problem and the special one treated in the preceding section that the coefficient of e''''"'^'^-^) in (34) is not a constant, in general, but an expression of the form Pn Pn ypm-\{x) E {x, p) Pn Pn The additional terms, however, are readil}' taken into account. The arbitrary number vi has not yet been assigned; let it be set equal to g + 3. Every one of the functions f{x) \pj{x), j = 1, 2, ...,?«, — 1, vanishes with its first q derivatives at the points 0 and tt. Hence it may be shown by integrating by parts that X fix) \pj{x) e"""" ('^-^) dx .)5rC0t {-k/v) —{iio-\-in)Tri . The factor e'"""*, of course, is ec{ual to ( — 1)". The third factor approaches a limit Hi—Hd as n becomes infinite. The second term in the braces, without the power of pn, may be WTitten similarly as ^rnt cot (ir/;-) gTiTTi niultiplied by a factor which approaches the limit Hi + Hii. Consequently the whole expression in braces is equal to ( — l)"pn~^~^ gWJT cot u/;-) multiplied by a quantity which approaches 2Hi. It follows that the absolute value of the expression is from some point on greater than ?_ pmrCOt{Tr/v) where Cg is a positive constant, provided that Hi 9^0. The condition that Hi be ec{ual to zero is that the imaginary part of the limiting value approached by the exponent of the third factor in (38), namely, the expression (q + 2)Ti V be an odd multiple of ^iri. There is no apparent reason why this condition should not be satisfied, in particular cases But it can surely not be satisfied for two successive values of q. We may accord- ingly adopt the following rule, with the assurance that the indicated choice of q is always possible: Let iV be any positive integer. Then q shall be the first integer greater than or equal to N, for which Hi ^ 0. Assuming that q is so chosen, we infer from (37) that /„ / (x) Vn (x) dx 0 ''^ ,9+2 from some point on, and from this inequality and (36), that Clo an n' g+2' where Cg and Cjo are positive. The remainder of the proof that the series diverges need not be elaborated. The details are suggested partly by the special discus- EXPANSION PROBLEMS. 409 sion in the first section, and partly by the argument just completed. It is found that if .Tq is any fixed value in the interval 0 < 0*0 < tt, the absolute value of un{xo) is equal to f"^cot(x/.) multiplied by a quantity which may be small for particular values of n, but certainly does not approach zero as a limit when n becomes infinite. It is possible to assign a positive constant Cu, such that for infinitely many values of 11. It follows that anMn(.ro) does not remain finite, and The statement of Theorem II is valid for the series discussed in the present section. III. Differential Equation of Order v with Fewer than v — I Specialized Boundary Conditions. If two of the boundary conditions, instead of a single one, involve the point x, the resulting series still show essentially the same char- acteristics, provided that the order of the differential equation is greater than 4, so that the specialized conditions remain in the major- ity. We retain the differential equation (15) of the preceding section, assuming now that j' ^ 5, and associate with it the following boundary conditions : kg—i Ws (w) = u^'^s) (0) + £ asj u^'^ (0) = 0, s=l,2,...,v-2, (39) ' ^ ° /Cj— 1 v—1 Ws {u) = vS'^s) (tt) + X ^,jU^^) (tt) + Z ^sjU^^^ (0) = 0, 7=0 7=0 S = V — 1,V, where J' - 1 ^ />:i > /.•2 > • • • > k.-2 ^0, V -1^ K_x > k, ^ 0. The differential equation being the same, the asymptotic expressions (17), (18), are still available,^^ and the regions Si and Ti may be de- fined as before. In determining the distribution of the characteristic values, it is sufficient to consider those in the sectors »So and S^v-i; all but a finite number of these will be in the regions To and T^v-i- If 33 For the present we may interpret the bracket symbol as it was interpreted at the beginning of the preceding section, without the refinement which was introduced later. 410 JACKSON. characteristic values occur on the ray arg p = t^Iv, which bounds So, the corresponding values on the ray arg p = — tt/v may be left out of account. It will be convenient now to have a definite understanding as to four of the subscripts of the roots uy, we shall write iri W 3)r» Sttx Wi = e " , 102 = 6", wz = e " , Wi= e " , so that the first four subscripts are assigned to the four right-hand vertices of the j'-sided regular polygon w^hich represents the v roots. If the polygon is rotated in the positive direction through an angle 6 between 0 and ir/v, the vertices representing the numbers wie*', .... ?f"4e^^ will be those furthest to the right, in the order of decreasing abscissas. The order of the remaining subscripts is immaterial. The quantities Wg (yi) have the asymptotic expressions: TVs (yt) = (pwO*« e ""^t^ [1] + as {pwiYil], s=v-1,p, where Ks is the order of the highest derivative whose value at the point 0 actually enters into the s-th. condition, for the last two values of s, and as is the corresponding coefficient, or is zero if the condition does not involve the point 0 at all. Let the factor p ^s be divided from the s-th row of the determinant of these quantities, s = 1, 2, . . ., v, and in the last two rows let js he \\Titten for Ks — kg- Then the condition for a characteristic value is expressed by the vanishing of a determi- nant Ai of the following form : ^* V'[l] w-z'^'ll] ••• t^>[l] + a.-i p'^f-i wi""-' [1] ) ( + a,_i p-'v-i w-i'v-' [1] f W/"-! e""'"'' [1] 1 \ + a,_i p^j.-i MJ/'"-! [1] j w/"" e""'"^ [1] ) j 10'^^ C'"-'^ [1] + a, pT" Wi"" [1] j 1 + a, py w-^" [1] wj"" e''"''''^[l] + a. pyw/il] 34 This does not correspond precisely to the determinant Ai of the preceding section, as no exponential factor has been divided from the last rows. EXPANSION PROBLEMS. 411 Let p be restricted to the region To. Suppose first, for simplicity, that the last two of the boundary conditions do not involve the point 0, so that a^-i and av are zero. Then in each of the minors of the last two rows of the determinant, an exponential factor can be taken from each column, and the expansion according to these minors has the form (40) Z5y/f,ye''^"'^+"^')Ml], where 5^;^/ and ^jjf are determinants of i' — 2 rows and 2 rows respec- tively, with powers of the quantities lOt as constituents. Of these deter- minants, 5i2, 5i3, ^n, and fi3 are surely different form zero, having essentially the same form as 5i and 82 in the preceding section ; it is im- portant to observe that when the roots Wj are arranged in cyclic order, Wi and W2 are adjacent, and likewise Wi and W3. It is immediately recog- nized that if p is in the remote part of To, all the terms of (40) excepting possibly that with the subscripts 1, 3, are insignificant in comparison with the term that has the subscripts 1, 2, and may be merged with the latter term, by virtue of the latitude that is allowable in the interpretation of the bracket symbols. We may write (41) Ai = [dioSioleP^"^' + W2)n^ [Sisrisle"^"" + '^^''. This is on the assumption that a^^i and a^ are zero. Among the terms which have to be added to those already taken into account, if this restriction is removed, the worst that can occur is e"^'^ multiplied by a power of p, with a further factor which remains finite. Such a term will be negligible in comparison with e''^^i + ^^'^, provided that a power of p is negligible in comparison with e''^'^, which will be the case if the vertex W2 of the polygon representing the I'oots remains at the right of the axis of imaginaries after the polygon has been rotated through an angle of t/v, that is, if 27r/V < ^tt, v > 4:. As the condi- tion last stated was imposed at the beginning of the section, the terms containing power's of p may be incorporated in the expression [5i2fi2]e''^'^'+ ""'^^ and the validity of (41) is general. Suppose the bracket symbols in (41) replaced by their limiting values, and the resulting expression set equal to zero. Since W2 — Ws = e" — e " = e " \e " — e " j , the equation so obtained, which obviously has only roots of the first order, is equivalent to the relation (42) 2pTi e "sin — = log ( — r^^ ) -f- 2mri, V \ 612^12/ 412 JACKSON. where the first term on the right is a particuhir value of the logarithm, and n is any integer. As in the case previously treated, the logarithm is a pure imaginary. Consider, for example, the relation between the determinants 5i2 and Si^. The latter may be regarded as obtained from the former by replacing w'3 in the first column by w^. But it may also be regarded as obtained by replacing each of the roots from W3 to %v by the root next to it cyclically in the negative direction (whereby W3 goes into one of the roots with a subscript greater than 4, and Wi goes into w-z) and then interchanging columns. This process amounts to a multiplication of each row of 812 by a root of unity, and so to a multiplication of the whole determinant by a quantity of the same sort. Similar reasoning applies to f 12 and f 13, and it follows that the ratio 613^13/512^12 has its absolute value equal to unity, and the real part of its logarithm equal to zero. The successive values of p in (42) are not real, because of the factor e'"^", but follow each other at equal intervals along a ray inclined to the axis of reals at an angle of tt/v. The correspondence between these values and the roots of the equation Ai = 0 is established as in the earlier cases. The roots of Ai = 0 in the region T2V-1 are similarly determined. There will be infinitely many of them, situated asymptotically along the ray arg p = — tt/v. But since the original problem involves p only in the j'th power, as has already been pointed out, the existence of roots p along this ray implies the existence of roots pe-'^*^" which will be in the region To, at least from a certain point on, and so will be among those already found. Thus the consideration of the region T^v-i adds only a finite number to the list of independent character- istic values. Hence: The characteristic values of the system (15), (39), are simple roots of the determinant equation from a certain point on, and are expres- sible in the form ^^ n-f- en — 1110 Pn = ■ /.^ , r ^ ' n = g,g+l,g-\- 2,---, sm {zir/v) where yuo is real and independent of n, g is a suitable integer, positive, negative, or zero, and lim ^ en = U. n= 00 35 It may be that there are infinitely many characteristic values just inside the lower border of jS-jv-i, and then the formula will give instead of these the equivalent values in Si] but the latter will serve as well and will be more con- EXPANSION PROBLEMS. 413 When n is large, a characteristic solution may be obtained by setting p = pn in the determinant Ai, and replacing the t-th element of the last row by ijt{x, pn) = e""^''' [1], t = 1,2,. . ., v; and this is except for a constant factor the only characteristic solution ^^ for p = pn. The numbers pnWi are situated along the positive axis of reals, the numbers pmvo and pniv^ along the rays making angles of 2ir/v with the positive axis of reals, and so on. Hence, when the determinant is expanded according to the minors of its last two rows, two terms are predominant over the others when n is large and x is not equal to 0 or tt, namely the terms involving" gP"(«'iT+t^:x) ^^^ ^pn{mir+w,x) ^ ^pj^g coefficients of these expressions are quantities which have already been seen not to vanish for large values of p. The limits approached by these coefficients are equal in absolute value, and if the whole determinant is divided by either square root of their product, they will be replaced by limits which are conjugate imaginaries. Dividing out also the factor e"""'"^, we may write (43) Un (x) = [di] e""^^^ + [do] e"""'^^, where f/i and do are conjugate imaginary quantities different from zero,^^ and the limits indicated by the bracket symbols are approached uniformly in x for .r' ^ x ^. x", if 0 < x' < x" < tt. This expression for un{x) is sufficient to suggest the proof of the first theorem, that a series of the form ^anun{x) which converges uni- formly throughout an interval 0 ^ .r < .tq must represent a function continuous there with its derivatives of all orders. In calculating the order of magnitude of the terms of the derived series it will be necessary to go behind the relation (43) to the complete determinant expression venient. If the problem is real at the start, all the characteristic numbers from a certain point on will be actually on the ray arg p = tt/v, and not merely approach it asymptotically. For if p is a characteristic number, the conjugate of p is a characteristic number, and e^^^ /" times the conjugate of p is a character- istic number, and unless arg p = tv/p the number last obtained is distinct from p, and there are two characteristic values where, if p is large, there ought to be only one. It may be remarked that in the plane of the variable \ = p", the characteristic values in this case are real from a certain point on, and become negatively infinite. 36 Cf . footnote in connection with the corresponding passage in the preceding section. It is readily shown that the solution indicated can not vanish identi- cally. 37 When X = 0 a number of terms are of the same order of magnitude, and when X = TT the terms involving gp-Cwx+w^^) ^^^ ^Pn(w,x+W3^) become com- parable with those indicated above. 38 They are not the same as the numbers di and di of the preceding section. 414 JACKSON. from which it was obtained; but the argument is so similar to the corresponding ones already given that its details may be omitted. For the proof that the development of a particular analytic function is divergent, the form of the adjoint system is to be taken into account. The adjoint differential equation is of course the same as Ijefore. The boundary conditions may be discussed by an appropriate modification of the method used in the case oi v — l specialized conditions. It is found that v—2 of the adjoint conditions contain the point tt only, while the remaining two not only may but surely do involve the point 0, in such a way that the sets of terms relating to this point in the two conditions are linearly independent.^^ The conditions may be simplified and arranged in a manner corresponding to our previous practice. The characteristic functions may be written in the form (44) Vn (.r) = [d'l] gP-^^C'r- a;) + [d'2] fP-J^jU- x)^ where the numbers pn are those already determined, d'l and d'o are conjugate imaginary quantities different from zero, and the approach of the bracket functions to their limits is uniform throughout any closed interval interior to (0, tt) . This expression, however, is not sufficient for the present purpose; for in determining the order of magnitude of the coefficients in the series it is precisely at the point a; = 0, where the formula breaks down, that the value of I'i(.t) is needed. At this point, terms involving eP"(^'^-"'>^) and eP"(^'^-"'>^) must be explicitly taken into account. The characteristic function (44) was obtained from the determinant corresponding to*° (33) by dividing out a factor e''"^'^^, independ- ent of X, and a constant factor depending on neither p nor x. By divid- ing out the same exponential factor, but a different constant factor, a characteristic function is obtained which may be represented ade- quately, in a sense presently to be explained, by the expression Vn {x) = 52' g — pnWlX gPnlVliir — X) -53' g — PnWlX ppnlOliir — X) 39 The latter part of this statement may be proved either by making use of the fact that the matrix corresponding to that which was called /3 must be non-singular, or by observing that if the statement is assumed to be untrue a direct determination of the characteristic values for the adjoint system leads to results inconsistent with those found for the given system. 40 This determinant in the present case has elements of the form ^kz^PnWiTT ^jj _|_ ^/^^ 7 2j^,^K 2 |^-|^j jj^ j|.g gecond row, where k'2, y'2, and K'2 are integers, of which 7'2 may be negative, and 0'^ is a constant; cf. the deter- minant from which (43) was obtained. EXPANSION PROBLEMS. 415 consisting of tlie principal terms in the expansion of the modified determinant according to the minors of its first two rows; here ^'2 and o's are quantities having the same absolute value, different from zero, and the constant factor just referred to is assumed to have been determined so that they are conjugate imaginaries, and /c'2 is the order of the highest derivative whose value at the point 0 enters into the second of the adjoint boundary conditions. Let us change our earlier notation to the extent of representing the new characteristic function, instead of (44), by rn(.r). The claim that this function is sufficiently well represented by the function vn{x) is to be understood as follows: Let/(.r) stand for the function (14), the integer q being left arbitrary for the present. If the bracket symbols are given the more specific interpretation explained in the latter part of the preceding section, and the various terms are integrated by parts, it is found that ^ prnr cot (2ir/v) (45) I rf(x)Vn(x)dx- rf(x)Vn{x)dx Jo Jo where c is independent of n, like the quantities to be denoted by the same letter with a subscript in the remainder of this section. The value of/f{x)r^(x) dx difl'ers by less than cie""^ ^ot {2./") / ^g + s from (46) -^, e""---^ 5'3 ^ gPnWSTT Pn 3+2 in comparison with which expression the difference just mentioned and the quantity (45) are insignificant, unless the four terms obtained by expanding the two determinants nearly or quite destroy each other. It is to be shown that this mutual destruction will not take place, at least if q is suitably chosen. If the first determinant is divided by ic^"'^'-'^^'-, and the second by ii'r^~-+*^'2, the remaining factors will be equal in the two cases, except as to sign, since u\/wo is the same as li's/wi, each being equal to g-^"/". The common factor e2(g + 2)«/. _ ^-2k',.i/. ^.j]! vanish only if —q-2 is congruent to A- '2 (mod. v), that is, for only one of any v consecutive values of q. On the other hand, the expression (47) -^;- U'^ W2-«-2+^'2 f"""^- + b'z Wi-^-2+^'== eP-'^A, which forms the remaining factor of (4G), can be treated as the expres- sion in braces in (37) was treated; in any v consecutive values of q there can not be more than one for which its absolute value fails to 416 JACKSON. exceed a quantity d e""" *^°* (2ir/;')^^5+2 ,^g j; ^ 5, by hypothesis, there will be at least three of any v successive values of q for which ' /(a:)?)„(a;)rfa- >-^e"'^cot(2,r/.) 0 I n It is readily seen that Un (x) Vn (x) dx \ < C4 e^'^cot (2V)^ X" and the rest of the divergence proof follows closely the lines of those already indicated. It suffices novy- to give a brief description of the principal features of the problem which is presented if n of the given boundary condi- tions, instead of two, involve the point tt, the number fx being any number less than half as large as v. The subscripts of the roots Wt are to be assigned in such a way that if 6 is between 0 and w/u the real parts of the numbers iVfe^'' succeed each other in decreasing order of algebraic magnitude. In the expansion of the determinant from which the characteristic values are found, the largest exponentials will be common to the principal terms, and so will not have an influence on the vanishing of the function; the occurrence of roots will be ren- dered possible by the balancing of e''"'^'^ against e''"'M+i'^, when ^u is odd, and by the balancing of 6""'^'^ against e''^M+i'^ on the one hand and of e"^!^-^^ against c''^^-^^'^ on the other, when fj. is even. The characteristic values of sufficientl}' large index will be roots of the first order, and will occur at approximately equal intervals; for odd values of jj. they will be distributed asymptotically along the axis of reals, and for even values of fi, along the ray arg p = ir/p. They will be of amplitude 0 or t/v exactly, from a certain point on, if the problem is real at the start. The characteristic functions may be represented by the expression (48) Un (x) = [di] e"""'/'^ -t- [d2] e''''"'*'+i% where di and do are conjugate imaginary ciuantities different from zero, and the asymptotic representation holds uniformly throughout any closed interval interior to (0, tt). It will be observed that the numbers pnWf^ and pnw^+i are situated asymptotically along rays making angles of p-tt/v with the positive axis of reals, whether p is even or odd. By using a somewhat more explicit form for the characteristic functions, it may be shown that if a series of constant multiples of them con- EXPANSION PKOBLEMS. 417 verges uniformly, its sum must necessarily have continuous deriva- tives of all orders. Of the adjoint boundary conditions, v — fx, depend on the point TT only, while the remaining fjt, contain linearly independent combina- tions of the values of v and its derivatives at the point 0. The char- acteristic functions of the adjoint system have an expression Vn (x) = [d'l] e'"'«'M('^^) + [d'o] eP''«'M+i('^^), the interpretation of which is similar to that of (48). Finally, the general coefficient in the formal development of the function (14) may be treated as in the previous cases, and it may be shown that for infinitely many positive integral values *^ of q the series diverges at every interior point of the interval (0, tt). It has been pointed out that the numbers pnWf^ and pmOf^+i, which figure in the formula (48), are situated along rays which become less and less inclined to the axis of pure imaginaries as [x increases. If an analogous formula were to hold in the case that v is even and IJ. = ^v, which has been excluded by our hypotheses, the corresponding numbers would be distributed along the axis of imaginaries itself, and the formula would resemble those which Birkhoff finds in the case of regular boundary conditions.*^ As a matter of fact, the boundary conditions do become regular when (x = ^v, ii the order of the deriva- tives at the point 0 involved in the last /x conditions is not too high. This is precisely what happens in the case of Liouville's conditions, with /i = 1, when v is equal to 2. 41 In the formula corresponding to (46), there will be a determinant contain- ing wi,..., Wfi-i, w^, and one containing Wi,. . ., w^-i, w^i. When /x is odd, a common factor is disclosed by dividing the first determinant by a power of wi, and the second by the like power of w^; when jj. is even, the second determinant is to be divided by a power of Wi, and the first by this power of W2. Among V successive values of q, there will be just m — 1 values for which the determi- nants vanish. Not more than one other value is excluded by the requirement as to the order of magnitude of the quantity corresponding to (47), and there remain at least p — /j. oi every v consecutive values of q, which are available for the purposes of the demonstration. 42 In the paper II, p. 389. The analogy is not perfect; the determination of the characteristic values becomes less simple in the regular case. Proceedings of the American Academy of Arts and Sciences. Vol. 51. No. 8. — November, 1915. THE MECHANICS OF TELtlPHONE-RECEIVER DIA~ PHRAGMS, AS DERIVED FROM THEIR MOTIONAL- IMPEDANCE CIRCLES. By a. E. Kennelly and H. A. Affel. THE MECHANICS OF TELEPHONE-RECEIVER DIA- PHRAGMS, AS DERIVED FROM THEIR MOTIONAL- IMPEDANCE CIRCLES. By a. E. KJENNELLY AND H. A. Affel. Received June 11, 1915. The following research was carried on, at the Massachusetts Insti- tute of Technology, under an appropriation from the American Tele- graph and Telephone Co. during the year 1914-1915. The experi- mental work was carried out at Pierce Hall, Harvard University. This research constitutes a continuation and extension of that reported to the Academy in September 1912, under the title of "The Impedance of Telephone Receivers, as affected by the Motion of their Diaphragms," ^ by A. E. Kennelly and G. W. Pierce. In that paper of 1912, it was shown that the impedance of a telephone receiver is different, when the diaphragm is free to vibrate, from that which it offers when the diaphragm's motion is damped or prevented. The difference between the "free" impedance as the frequency is varied, and the "damped" impedance, is called the "motional impedance," and measures the velocity of the diaphragm's vibration. When plotted vectorially, this "motional impedance" is found to be a circle passing through the origin of coordinates, and with its diameter depressed through a certain angle. Every telephone receiver and diaphragm possesses its own characteristic motional-impedance circle. The characteristics of this circle, in regard to diameter, depression angle, and distribution of frequency positions, determine certain electrical and mechanical properties of the instrument. Examples of such circle diagrams appear in this paper in Figures 4 and 14. It was shown in the 1912 paper above referred to, that there are four constants of an ordinary telephone receiver which determine the essen- tials of its behavior, both electrically and mechanically, throughout the range of ordinary telephonic frequencies (100 to 2500 i^ ). If we consider the impedance of a telephone receiver with the diaphragm prevented from vibrating, and thus incapable of reacting electromagnetically on the coils, when the latter are excited by alter- nating current, we find that, as might be expected, the impedance of 1 See Bibliography, No. 11. 422 KENNELLY AND AFFEL. / / 1 15C 7tC / / / 30b: / 7 1190/cj' 1 1 37/n 1 1,0 8 /rv ' /?" / OOlf / 897/~ / / \blSr / / 80 > 4/" / ^7 l/^ / j 1 / 'L 9~ 150 200 RESISTANCE- OHMS Fig. 1. — Locus of Vector Damped Impedance »OR A Particular Receiver. TELEPHONE DIAPHRAGMS. 423 these coils increases when the frequency is increased, although not in direct proportion. Thus, Figure 1 represents the locus of the imped- ance of a particular receiver when measured by Rayleigh bridge, with constant sinusoidal alternating-current strength, but with varying frequency. Apparent resistances are measured parallel to the axis QR, and apparent reactances parallel to the axis QX. As the fre- quency of the exciting alternating current is increased from 400 ^ to 2500 c^J , the locus of the impedance is the curve A B C D. Thus at 804 f^^ , the impedance is found to be OB, and at 1190 c^^ , it is in- creased to OC, where O is the origin (not shown). The increase in apparent resistance is due to increasing power loss at rising fre- quencies. The increase in reactance j X = ; Leo ohms, is to be attrib- uted to increase in the angular velocity co = 2 tt/ radians per second; where / is the frequency. The apparent inductance L henrys actually diminishes as the frequency rises; so that the increase in the re- actance OX is less rapid than the increase in frequency. The imped- ance measured in this manner is called the "damped impedance." If now the measurement of the impedance is repeated over the same range of frequency' but with the diaphragm released, so as to be able to react electromagnetically on the winding, the locus is found to be a looped curve, such as is represented in Fig. 2; so that one and the same vector impedance OP is found at two distinctly different fre- quencies. Thus at 702 ~, the "free impedance" has a magnitude of 153 +i 161.5 ohms, and at 1050 ~ , 148 + j 144 ohms. This peculiar dual valued behavior of the free impedance is due to the motion of the diaphragm. If at each of a series of successive rising frequencies, we subtract the vector damped impedance from the vector free impedance, as indicated in Figure 3, it is found that the successive vector differences, due to the motion of the diaphragm, and, therefore, called " motional impedances," lie approximately on a certain circular locus as shown in Figure 4. This circular locus is called the motional-impedance circle for the particular instrument. The same facts may also be presented in a scalar or non-vector diagram. Figure 5. Here the abscissas represent impressed frequency. The ordinates represent, to the left-hand scale, the magnitude or "vector modulus" of the telephone-receiver impedance. To the right-hand scale, they represent the phase angle or "argument." Heavy curves represent measurements of free impedance, and broken curves corresponding measurements of damped impedance. The measurements made on the particular receiver C, referred to in Figures, 1 to 5, are recorded in Table I. 424 KENNELLY AND AFFEL. / '6 1507 / / / / i3oe / / / 1 f ";^ I 923. 945 6>— • a^gsor ^" ^yi K/ 897 ^ >S^8 \ / W. 7yVo \ ^94 ~ / 1082/ ' ! / \ / V/' // 10 H '^ iL "1 / Y/ f ^=J^ / ^ '\ 0070, L 7 1043 yoi / '429^ \ s. / 1036 \ S^ _>> y ' 101; .6~ V, Q28~ .^ -^ 20^ f ^ 1 1 150 250 200 RESISTANCE-OHMS Fig. 2. — Locus of "Vector Free Impedance, for the Same Receiver. TELEPHONE DIAPHKAGMS. 425 1507^ / / / 1306 /; i / 7 ^ / /4/ f)^ cflf^ ^r^^ &s — ^/^g s^ 9 /^ ^^u^ 'X \, , ^ ^ A < h/ ¥t V \ \ ^ --^ ^Or^ \ ^ Tli \ \ \ \ > \ u M \ \^ \ \ N "^ ] Vr \ \ \^ ^"^ \ j / 1 or I \ ^ \ N ) / // '\ / o Vo \ s. / y G \ ~Tl \ / / \ V \ \ / \ N \ y -E V 5F L _^ 140 150 150 170 180 190 200 210 220 230 240 250 260 270 280 resistance, ohms Fig. 3. — Motional Impedance Vectoks at Successive Observed Frequencies. 426 KEXXELLY AND AFFEL. MOTIONAL RESISTANCE, OHMS -20-10 0 10 20 SO 40 50 60 70 80 90 100 110 + X -rr 2^ , a-i e e 3 -R 0, > 31 0) j> 3) S^ r n'^ +R \^\ \ / ^fe'"^ /K X S V / ^)>^ c^^" U K \ \, / \ 1h '82^ \ \ y' ^ Ip- -c -1-oe N . / / yoo^ / \ \ P^°-^ / \ \ / y \ '007 "A.- \ r^ / v^V / % \ / ^0^ v^ V I Hii ^E V \, ? to O > H ~- "^ - X Fig. 4. — Motional Impedance Circle of Same Receiver from Vectors in Fig. 3. TELEPHONE DIAPHRAGMS. 427 FREQUENCY, CYC. PER SEC. 500 600 700 800 900 10001100 1200 1300 1400 1500 o CO o // CO p // o 1 / ^R^UME^ T i m i -^"7 / y- " o ,^^^ M^ / / / CO A i m / / o % \j ^ / • J 00 OJ X / / \ /J r o / 1 / 1 < Q / - O / / 1 1 CJ / / ft i o // 3 1 C-J > O c o <5 r 5 z' / C-J z o 7/^ '1 s O / 1 < o / / r ^~ Fig. 5. — Scalar Diagram of Free AND Damped Impedance. 428 KENNELLY AND AFFEL. Max. Cyclic Vibration Velocity kines ux 05 O CO O 1^ ■* t^ Oi ^ '^f' lO CO CO CO c^ CO CJ t>- 00 i-l CO LO t- — 1 CO t^ CO -H 00 ^ 0 CO CO X -O: t- 0 CO OS 0 X oc ic X ^ LO lO CO OOO-H-H ^ C-\ C^l OJ CO CO "^ Lt LO CO CO CO lO 't' ^ CO CI ^ 0 X Max. Cyclic Displace- ments microns observed 00 O 02 O «2 lO CO CO t^ 00 00 lO CO o c-i t^ CO 00 O OS O) >-i rt LO X ^ C iO ^ CO CI CO .-1 »n 0 CM m t-- X CI OS ^ -^ ^ 'M (N (M CO ^ '^ lO CO CO X OS O 0 OS X t^ CO -f 00 CJ 0 XI Motional React, ohms ic o o CO lO r^ CO t^ lO ^ ic i-H CO 00 o X CO 'f rH CM CM t^ -^ C5> 10 rlOO r^ ^ t^ ^ CO lO CO lO CO ^ CM CO X Ol ^ ^ CO t^ o 1 1 1 Y OS X LO t^ ^ ■-■ -H 0 Xl^ 7771 1 0 -* •^ !>• 0 10 CO CM ^ ^ 1 1 1 1 1 1 1 Li-L Motional Inductance miiUhenrys 00 as lO rH rf< t^ 00 CO (>] 00 ■* 1-1 .-H LO 00 t^ CO CM CO 00 ic ^ 10 10 ^ lO 00 (M O ^ 4 CO --H LO 1 177 X X CO CO 0 TTTTT t^ IC 00 CM ^ 1 1 1 1 1 00 1 1 L Inductance Damped miiUhenrys t^ t^ t^ 'M 00 CO lO rr CO M O-J .-H ,-( 0 0 OS XX t^ CO lO ^ CM 0 I-~ 00 00 lO ^' ^ CO CO CO CO CO CO CO CO CO CO CO CO CO CO CO CO CO CO CO CO CO 00 CO CO CO CO CM CI CJ CI CM CO CO 00 CO CO CI CI CM CM ^ CO CO CO CO CO -H 0 CO CO Inductance Free miilihenrys lO CO M CO 0 lO CO CO CO CO CO 00 >-H t^ t-H t^ 00 00 C^l C-1 --1 ^ Tfi CO OS ^ rH r-H ,-1 1— 1 C'l 10 I^ X OS 0 C) CM CM CI CO 0 OS CO CM Ri-R • Motional Res. ohms CO 00 CO t^ ^ i-H ^ (M .-1 -H -Tf lO UO CO ^ uO CO t^ CO CO CO C CO X 0 ^ --< OS lO -H CO 0 CO lO CM 1 CM C) ' 1 1 X LO CI X ^ CI CI '.1 r^ 1-1 1 1 1 1 1 X CO 1 1 R Res. Damped ohms •* lO CO t^ O i !>i ■* 0 CO X r-H LO t^ CO r^ CO •» C'l CM O) 01 ^ ^ ^ ^ 00 CI OS CO LO rr CI -^ ■-( ,-H ,-H "* X CO OS t^ Tf< ^ lO LO -0 X 0 ^ CJ B 1 3 -<> "^i O 'M lO o o ^ o 00 ^ CO -* O CO CO C^I -* UO lO lO ^ ^ CO 00 CO O TtH -* ^ lO 00 CS C r^ ^ 10 i-O CO CO CO 0 in CO X 0 0 '^ X CI t^ CM 0-) CI CO CO CO CO CO CO CO CI -H X CM t^ ^ CO 0 10 OS f TfH 10 lO 10 CO CO CO CO CD X 00 CD LO X OS OS CO -^ r- CD t-- OS "-1 •* CO CO CO t^ 1-- CO C 0 t^ CI T X OS |d2 O) C-J Tti t^ t^ (M O O lO OS "* t^ 30 CC 00 r^ CO CO lO k Diaphragm first, the pressure is not uniform during the period of testing, and secondly; because any pressure on the diaphragm, by diminishing the air-gap, alters, to some extent, the magnetic characteristics of the in- strument. A convenient form of damping device is indicated in Fig- ure 13. It virtually applies a relatively large mass to the diaphragm, without imposing any appreciable mechanical pressure; so that in- ertia is depended upon for extinguishing the vibration. The receiver is securely supported in a horizontal position. A small threaded metallic disk K, is attached, by means of shellac, to the center of the diaphragm. Thick shellac varnish appears to be a convenient material for cementing a load to the diaphragm. The cylindrical metallic mass M (4 cm. long and 2 cms. in diameter) is then connected mechani- 448 KEXXELLY AND AFFEL. cally with the disk K by means of a screw rod. The weight of the cyhnder M is carried by the flexible metallic strip s, and counter- balanced on the beam bb'^ by the weight w. When carefully applied, the receiver diaphragm D at rest, remains damped without being stressed. The vector differences between the free and damped impedances, ^^ over the range of impressed frequency, will enable the motional- impedance circle to be drawn. From the data thus obtained, all four constants of the instrument may be computed as indicated in Appendix I. The circle diagram supplies ojo, A, and Z^. The amplitude measurer gives .t^. The con- ditions of the Rayleigh bridge, and its voltmeter, give i^. Data Obtaixed on Particular Receivers by IMeans of Ampli- tude Measurements. The following are the results of tests on four telephone receivers, marked A, B, C and D respectively, using the amplitude method, in conjunction with the motional-impedance circle diagram. Two of these, A and B, were tested in a vibration explorer ^^ previously referred to, and the other two — C and D, — ■ were tested with the amplitude measurer above described. Although the tests with the vibration explorer interfered with the normal structure of the instru- ment, by reason of the removal of the receiver cap, and are therefore not completely representative of the normal operating conditions; yet it was deemed important to secure, from the vibration explorer, a check on the amplitude method, through an independent determina- tion of the equivalent mass. The Taljle is divided into three parts. The first gives the Data obtained from the dimensions of the poles and of the particular dia- phragm used; also the d. c. resistance of the winding at 20°C, by Wheatstone Bridge. The telephone receiver D, was a "1000-ohm'' instrument. The others were of the usual "75-ohm" type. 15 It has been suggested by the Engineering Department of the Western Electric Co., that if two receivers are available, with sufficiently similar con- stants, one (free) may be inserted in the arm bd, and the other (damped) in the arm cd. A single series of inductance and resistance measurements, over the range of impressed frequency, will then be sufficient to determine the- motional impedance directly. 16 Bibliography, No. 17. TELEPHONE DIAPHRAGMS. 449 TABLE V. Data Secured ox Particular Receivers. Receiver Number A B C D Type Bell Bip. Bell Bip. Bell Bip. Watch Case Bip Data Area of each Pole cm. X cm. 1.14X 0.199 1.40 X 0.225 1.14 X 0.199 1.15 X 0.1^ Distance separating poles cm. 0.686 0.85 0.686 1.10 External diam. of diapli. cm. 5.52 5.40 5.52 5.57 Diameter of clamp- ing circle cm. 5.00 4.94 5.00 4.95 Thickness of dia- phragm over japan cm. 0.0399 0.031 0.031 0.0244 IVt. of diaphragm gm. 5.979 4.181 4.397 3.365 Direct current Resist- ance of coils, ohms at 20° C. 87.1 73.0 86.7 1079 Test Data Temperature of Test Deg. C. 26.7° 27.8° 20.° 20° Current through Rec. absamps. r.m.s. I 0.000 202 0.000 200 0.000 204 0.000 116 Resonant Frog, of Rec. eye. /.sec. /o 993 1020.4 1015 898.5 Resonant Ang. Vel. rad /sec. wq 6240 6412 6378 5646 Motional Impedance Circle diameter ab- sohms. Zm 80.2 X lO'-* 70 X 10» 140 X 10'' 367 X 10» Max. Amplitude of Resonancecm.X 10"* (Microns) . .Tm 7.53 7.19 10.35 6.64 Velocity of Diaph. at Resonance cm. /sec. (max. cyclic), a-^ 4.70 4.62 6.6 3.75 Decrement per sec. A (Mean) 61.2 236 149 356 450 KENNELLY AND AFFEL. TABLE V {Continued). Receiver Number A B C D Type BeU Bip. Bell Bip. Bell Bip. Watch Case Bip. Calculated Data Equivalent Mass of Diaph. m, (gms.) 2.^ 0.53? 0.902 0.986 Equivalent Mass Factor 0.49 0.17 0.245 0.387 Equivalent Mass Factor by Explora- tion Method 0.53 0.18 Equivalent Elasticity, s, dynes per cm. 9^.0 X 10^ 22.9 X 10^ 36.7 X 10^ 31.44X 10" Equivalent Diaph. Resistance r, djTies per kine 295 262.3 268 702 Force Factor. A. dynes per absam- pere 4.87 X 10 « 4.28 X 10« 6.12 X 10^ 16.05 X 10" Mean Angle of Lag (9 degrees •29 . 7 37.3 25 3 37.0 The second part of the Table gives the results obtained from the motional-impedance circle, and from the amplitude measurer. The circle diagram for the case of receiver b, is given in Figure 14. The outer circle is the motional-impedance circle, plotted to a scale of ohms, from observations with the Rayleigh bridge (Fig. 12). The inner circle, is a circle of maximum cyclic vibration velocities, at the center of the diaphragm, as deduced from vibration amplitudes ob- served with the vibration explorer, plotted vectorially from the origin O. It is shown in Appendix I (40) that the motional impedance, when the sinusoidal testing current is maintained constant, becomes the product of the vibrational velocity x and a constant. The inner circle, thus conforming satisfactorily to the impedance circle, supplies a check upon the theory of the case. Theoretically, the mean vibra- tion amplitude should be measured over the air-gap, instead of at the center of the diaphragm; but the difference is probably not material. It will be noticed that the resonant frequency fo, on the diameter of the impedance circle, is 1020.4 ■^^ , with a corresponding resonant angular velocity wo, of 6412 radians per second. The frequencies at the TELEPHONE DIAPHRAGMS. 451 MOTIONAL RESISTANCE, OHMS -20 -10 0 10 20 30 40 Fig. 14. — Diagram of Motional-Impedance and Maximum Cyclic Velocity for Receiver B 452 KENXELLY AND AFFEL. quadrantal points of the circle are 983 and 1065 ~ respectively, or in angular velocity 6176 and 6692 radians per second. Half the differ- ence between these is 258, which represents A, the damping constant. When A is required with greater precision, however, it is better to take a number of frequency points around the circle into account, by means of formulas (20) to (23). The resonance curve for this case, giving central maximum cyclic velocities of the diaphragm, against impressed frequency, appears in Figure 15. It is a fairly representative curve for telephonic receivers as a class. The curve of diaphragm velocity is drawn through ob- servation-points with the explorer, while the small circles represent the corresponding points as computed by formula. The motional impedance curve is also drawn through the observations. Theoreti- cally, the ordinates of the two cm"ves should retain a constant ratio. It should be remembered, however, that at frequencies remote from resonance, not only are the amplitudes hard to measure, but the force constant a needs correction. The third part of the Table gives the results of the calculated data, in terms of A, m, r and s. The numbers of turns N, in the windings not being known, the value of a/x, which is a characteristic quantity for a receiver, is omitted. The equivalent mass factor vi/m was obtained in tests A and B by two different and independent methods ; namely, by the computation of m as in Formula (49) AppendLx I, and through exploration, by Dr. H. O. Taylor, of the amplitudes ■'•'^ over the surface of the diaphragm, in the manner described in Appendix II. The fact that these two values of the mass-factor compare favorably, constitutes a check upon the validity of the amplitude method of determination. Influences which Affect the Instrument Constants. The telephone receiver is so sensitive to external influences, which effect the motional-impedance circle, that there are numerous ways in which such influences might be exerted and their effects thus revealed. It is proposed here to consider, however, only a few of these influences, and their general effects on certain receivers as deduced from the changes produced in their motional-impedance circles. The influences selected were (1), variations in the screw-clamping of the cap, (2), 17 Bibliography, No. 17. TELEPHONE DIAPHEAGMS. 453. SlNHO '3D-NVa3dlAinVN0l±0lAI 03 ot- oe oz 9 f 2 Z •D3S/'IAI0 A1ID0"13A IMOVdHdVia 454 KENNELLY AND AFFEL. variations in temperature, (3), variations in air chamber between cap and diaphragm, (4), variations in atmospheric pressure, (5), variations in added mechanical resistance. Variations in the Clajviping Adjust:mext of the Cap under Varying Torque. Tests of several telephone receivers, made before and after a removal and replacement of the screw cap, were found, at times, to differ con- siderably. This led to an investigation of the influence of screwing on the receiver cap with varying degrees of tightness. The caps were of the same molded composite material as the receiver cases. A lever clamping device was designed and constructed as shown in Figure 16. It consists of a brass rod AB, with a known sliding weight applied at a measured horizontal radius arm r. The rod terminates in a brass fork containing notches, which engage with pins p p screwed into the cover. The receiver is clamped by its shell in a horizontal posi- FiG. 16. — Method or Applying Cap Torque. tion. The cap is then screwed on slackly, and the final screwing is accomplished with the measured torque. The instrument is then tested for motional impedance under these conditions. The torque is expressed in gram-perpendicular-meters; i. e., in grams weight act- ing vertically at a horizontal radius arm of one meter. The effects of varying the screwing-on torque upon the motional- impedance circle of the particular receiver tested, (B with No. 36 diaphragm), are shown in Figure 17. It will be seen that with zero imposed torque; i. e., with the cap laid on the diaphragm, but not screwed, the motional-impedance circle has the smallest diameter, and is nested within the others. The resonant frequency of the diaphragm was 859 <^ . As the screwing-on torque is increased, the TELEPHONE DIAPHRAGMS. 455 -20 -1 0 MOTIONAL RESISTANCE, OHMS 0 10 20 30 40 50 60 70 80 90 100 110 ^^ K ^ ^ ^=^ \N 5\ \, N /^ / \ \ \ \ \\ \. 1 > I \ ^^ o \ \ ^ il^ i ^ 1/ \\ Iv h ■^ s^ 'HI "/ 1 ' \ 1 X ^7 Oj f^^ p/ w \ \ [^ b>>\ J>^^^ iFl \ ^ .f> ^^ \ / \ X / ^ \ ::^ ^0-* f T=CAP TORQUE GM.PERP.MET. ./o=RESOMANT "^ :^ -- ::^ CYC./ SEC. Fig. 17. — Motional-Impedance Circles with Different Cap Adjustments. o ^ CM X -- °o H LU o ^ ■ r/ S CO < y -/ LU O /^ CM 100 200 300 400 cap torque, gm.p. met. Fig. 18. — Curve Showing Relation between Motional- Impedance Circle Diameter and Cap Torque. 456 KENNELLY AND AFFEL. diameter of the motional-impedance circle increases, as indicated in Figure 18, in nearly simple proportion, until a torque of 250 gm-perp- meters is attained. Beyond this torque, there is very little effect on the circle diameter. After a torque of 20 gm-p-m had been ap- plied, no appreciable effect was discovered on the resonant frequency of this particular instrument. At the time that the above torque tests were taken, the method of amplitude measurement had not been^developed ; so that an exact o o ID \ ^ r^ ■ — =- ^ z CO o LU ^ z > \ ■^ ^A" \ V Q - o - •* o \ \ t z < to CO LU CC \ \ \ < o ^ X CM < O o 100 200 300 400 CAP TORQUE GM. P. METERS Fig. 19. — Curves showing Values of A and r FOR Different Cap Torques. determination of all four constants throughout the series is not possible. If, however, we assume that the equivalent mass m of the diaphragm remained constant throughout, since the resonant frequency remained practically unchanged, we are able to evaluate the three other con- stants A, r and s. Of these, the s constant must have remained un- changed, and the only variations would be those of A and r, which are plotted in Figure 19 as ordinates against torques as abscissas. It will TELEPHONE DIAPHRAGMS. 457 be seen that the vakie of A remained but Httle changed throughout the series, diminishing from 0.575 to 0.51 megadyne per absampere. The mechanical resistance r of the diaphragm changes, however, largely, diminishing from about 600 to 200 dynes per kine. Fig. 20. Mkthub of Am Gai- Measurement. At first sight, it seems difficult to explain why the change in imposed screwing-on torque should affect the mechanical resistance so mark- edly. It was found, however, that a noticeable effect of screwing on the cap tightly, was to increase the air-gap between diaphragm and poles. This is a reasonable effect, if it is remembered that the magnet poles bow the dia- phragm down towards them, when the diaphragm is laid on the clamping ring, and that the application of cap pressure, under the influence of screwing torque, to the upper clamping ring, tends to lessen this bowing. The increase in air-gap accompanying increase in torque was measured by the device shown in Figure 20, where the micrometer-head depth gauge is applied to the upper surface of the dia- phragm, from a temporary brass frame attached to the io 20 30 shell Fie-iire '>! shows the CAP DEFLECTION, DEGREES snen. j^igure -i snows me khj. 21. _ cukve showing Kelai ion magnitudes of the deduced between Air-Gap in Keceiver air-gaps, as ordinates, against B and the Cap Adjustment. CAP TORQUE GM.'PERP. 5 10 VIET. CM 1 1 1 1 ' \ 1 1 0 0 CO y ^ 0 CO y V coo LU 0 / r do ^C^, < 0 «o / < o' / Y 0 / / d v 458 KENNELLY AND AFFEL. degrees of twist of the cap as abscissas, an approximate scale of cor- responding imposed torque being added. It is known that a considerable proportion of the mechanical re- sistance r offered by a receiver diaphragm, is due to the braking effect of eddy currents set up in the diaphragm, during its vibration in a strong magnetic field. It is reasonable, therefore, to suppose that the diminution in r, which was found to accompany the increase in screwing-on torque, and in air-gap, may have been due to a reduction in eddy-current damping action, since the variations of magnetic field within the diaphragm would be diminished. It may be safely inferred that the adjustment of a telephone-receiver cap is able to affect the receiver constants considerably. In this particular receiver, the most sensitive and satisfactory setting was with the cap screwed on tightly. Effect of Variations in Temperature. In order to ascertain the influence of temperature upon the char- acteristics of a receiver, a large electric oven ^^ was used, in which the receiver was placed, at a conveniently controlled temperature. A number of motional-impedance circles were observed, at oven tem- peratures from 16° C. to 50° C, all other conditions being maintained constant. It was found that two effects were produced with rise of temperature; namely: (1) A reduction in resonant frequency, amounting to about 2.5 cycles per second, per degree C. rise. (2) A slight reduction in circle diameter, which, however, was not always noticed. The results obtained are given in the following Table. They are given in the sequence of observation. The reasons for the above indicated effects of temperature on the receiver characteristics have not been analysed. They might be attributed to temperature changes in the mechanical elasticity con- stants of the diaphragm; or to expansional effects in the structure; or to both causes. It is evident, therefore, that, judging from this particular instru- ment, the influence of temperature on a telephone receiver's character- istics are very appreciable. Care should be taken to maintain the temperature of the instrument constant during any set of observations. With this object in view, it was customary to conduct the tests with the instrument inside the closed oven, and with the heat shut off. 18 Bibliography, No. 18. TELEPHONE DIAPHRAGMS. 459 TABLE VI. Temperature Effect on Receiver Characteristics. Series A Series B Series C Condition of Cap Adjustment Tight Tight Loose Loose Loose Tight Tight Tight Temperature deg. C. 19 3° 47.0° 16.0° 46.0° 31.5° 51.0° 36.0° 22.0° Resonant Fre- quency cyc/sec. /o 886.7 834.9 827 742 788 817.3 852.5 867.5 Resonant Ang. Vel. rad/sec. coo 5575 5248 5200 4666 4956 5140 5360 5450 Motional Imped- ance circle diam. ohms Zm 88.2 72.0 52.0 48.0 45.5 70.5 70.0 71.0 Effect of Variations in Air-Chamber between Cap and Diaphragm. In order to ascertain the influence of the air-chamber over the diaphragm of the ordinary receiver, a special cap was used, in which this air-chamber could be varied, by altering the position of a friction- tight cylindrical plug, of the same diameter as the clamping circle. Motional-impedance circles were observed under these different condi- tions, with the results given in the following Table: — TABLE VII. Results of Motional-Impedance Circles with Changes in Cap Air- Chamber. Air-Chamber Circle-Diameter Thickness ohms Resonant Frequency cycles per sec. Normal, 0.5 mm. Large, 5 mm.. Infinite, Plug Removed 139 150 1.55 872 892 927 460 KENNELLY AND AFFEL. The effect of the air chamber, as compared with an open diaphragm, was to lower the resonant frequency sHghtly, and to reduce the circle diameter. In order to ascertain the influence of the ordinary atmospheric pressure upon the characteristics of the diaphragm, a bipolar receiver was suspended in the glass bell-jar of an air-pump, the two wires to the receiver being carried through a seal at the top of the jar. One mo- tional-impedance circle was obtained with full atmospheric pressure, and another after the pressure had been reduced to about 1 cm. of mercury. The two circles are shown in Figure 22. It will be seen that the removal of atmospheric pressure decreased the resonant fre- quency and enlarged the circle diameter. The results are given in the following Table: TABLE VIII. The Effect of the Atmosphere on Receiver Characteristics. Receiver C. Condition of Receiver In Vacuum In Air Temperature, deg. Cent. 18.5° 18.5° Current, absamperes. i.m, s. I 0.000 05 0.000 05 Resonant Frequency, cyc/sec. /o 956 1020 Resonant Ang. Velocity, rad/sec. coo 6007 6409 Impedance Circle Diam. ohms Zm 177 116 Decrement per sec. A 120 176 Equivalent mass gm. m 0.902 0.902 Equivalent Elasticity, dynes/cm. s 32.5 X 10« 37.0 X 106 Equivalent Resistance, dynes/kine. r 216.5 317.5 Resistance due to Air, dynes/kine 101. = 31.8% of total resistance Assuming that the equivalent mass m remained unchanged, the equivalent resistance r diminished in vacuo by about 30 per cent. The power of the receiver as a sound-producing device is theoretically limited to (x)V or 101 (.r)^ abwatts. Owing to the fact that the underside of the diaphragm is cut off from free access to the air, only part of this power can be actually utilised for sound production in air. At certain frequencies near resonance, the Rayleigh bridge balance becomes very sensitive to variations in atmospheric pressure. From an examination of the balance at these frequencies, it was easy to ascertain whether air was leaking into the bell jar. It seems likely, TELEPHONE DIAPHRAGMS. 461 therefore, that sudden variations of barometric pressure, within the normal range, might produce perceptible changes in the free imped- ance, near resonance. MOTIONAL RESISTANCE, OHIVLS -40 -20 0 20 40 bO 80 100 120 140 160 o % o i •^ ^ .y CVJ ^ .^ "- -^^ .H -<] r^ CO 5 o \, / \ \ C Y*- X (N / \ \ \, \ of z < o 1- ID O ' < lU o joJjl \ f^ ■j 1 uu ^ ~ \945. .f^ 1000 —994 :[k \^ \ /'" '3.5^ -J < o [\ .> > RE 4N__ MR^ \ / L. Iry. O T S ■^'^ 'A ^ / O q." H s^ # v.y.l'^S \j . X o o,^ V ?o jN ^ fei. J^^^ ■\ "V *■ lO ' Fig. 22. — Motional-Impedance Circles of Receiver C IN Air and in Vacuo. Effects of Variations in Mechanical Resistance. Since changes in atmospheric pressure had been shown to effect a marked influence upon the mechanical resistance r of the diaphragm, some tests were made to ascertain the mechanical resistance offered by small circular aluminum vanes of different diameters, fastened by a small metallic tie-rod to the center of the diaphragm. The small vane was in each case held at its center, coaxial with the diaphragm, and with its plane parallel to that of the diaphragm (see Fig. 23). In changing from one size of vane to another, the mass of the tie-rod was so altered as to maintain the resonant frequency of the loaded diaphragm substantially unchanged, and therefore, likewise, the total equivalent mass. The results are given in the following Table. 462 KENNELLY AND AFFEL. Fig. 23. — Air Resistance Measurements. TABLE IX. Table of Vane Mechanical Resistance to Vibration. Test No. Added Mass of vane & rod gms. Resonant Frequency cyc/sec. /o Vane Area sq. cm. on one surface Impedance Cu-cle Diam. ohms K Decrement per sec. A Equivalent Total Mechanical Res. r. dynes/kine Extra Res. 1 2 3 3.00 3.11 3.18 704 689 685 0 5 10 26.5 21.2 17.0 157 167 173 1370 1490 1570 120 200 TELEPHONE DIAPHRAGMS. 463 The equivalent mass of the diaphragm alone was measured inde- pendently by the loading method. It would appear, from the above tests, that the extra mechanical resistance of a circular disk was about 20 dynes/kine for each sq. cm. of disk area, — (ira^). These results, however, can only be regarded as preliminary. A few similar measurements, made with circular disk vanes im- mersed in water instead of in air, gave results of roughly 500 dynes/ kine per sq. cm. area of disk, or some 25 times the vibratory resistance of air. In this case it was found that the water not only added mechanical resistance; but also an appreciable extra mass (about 0.5 gm. per sq. cm.) to the vibrating system. These measurements of extra mechanical resistances are reported, not as definite results, but as indicating the directions in which the motional-impedance circle method may be applied to the analysis of vibrating systems. In conclusion, the writers desire to express their acknowledgment to the Research Department of the Western Electric Co. for valuable suggestions, help, and special instrument parts. Summary. (1) The principal characteristic constants, defining the mechanics of a telephone receiver, are the force-factor a, the equivalent mass m, the equivalent mechanical resistance r and the equivalent elasticity s. They may all be determined from the motional-impedance circle diagram, if one additional independent relation can be secured. (2) Out of a number of possible additional independent relations, three are considered in detail; namely, the vibrational exploration method for measuring m, the loading of the diaphragm for obtaining m, and the use of an amplitude measurer for determining the maximum cyclic displacement x^- (3) While all three of the above-mentioned methods are capable of giving results, the last named is the recommended method. It consists in applying a simple form of amplitude measurer, to the center of the diaphragm, during motional-impedance tests, and observing the amplitude at resonance. (4) With the amplitude measurer, the characteristic constants are derived for several particular types of telephone receiver tested. (5) The effects of various influences upon the behavior of a receiver, and on its characteristic constants, are discussed. 464 KENNELLY AND AFFEL. Appendix I. Elementary Theory of Simple Vibration. We may suppose a material particle of mass m grams at the point P,. Figure 24, to rotate in a circular orbit, and in the positive or counter- clockwise direction, about the center O, with uniform angular velocity CO radians per second, as indicated by the arrow. If the radius of the orbit is x cms., we may assume that there is a constant force sx dynes in the direction PO, along the radius vector, so that this centripetal force is proportional to the displacement x from the center O. The vector displacement, measured positively outwards from O at time t seconds, is X = .To e^"' cm Z (12) the epoch being selected such that .Tq = OP, when i = 0. The instantaneous velocity of the particle will be directed along the tangent PP^ at P, and its vector value will be : X = jco.ro e^"' = jco.r cm/sec Z (13) The instantaneous acceleration of the particle will also be directed along the tangent QQ^ at Q, a point 90° advanced in phase beyond P, and .V = {juf .To e^"' = - co^.r cm/sec2 Z (14) The forces acting on the particle at any instant, such as that indicated s in Figure 25, will then be (1) the elastically restoring force —sx=jx CO dynes in the direction PO, or opposite to the direction of x. (2) the force opposing the velocity, or —rx dynes acting in the direction PT or OC, assuming that this force acts in simple proportion to the instantaneous velocity, and (3) the force opposing acceleration, or inertia force, acting in the direction Q^Q, or OB = — mx = —jmcox dynes. The vector sum of these forces, in conjunction with a rotatory impressed force F dynes along OD, which sustains the motion, must be zero. The reactive forces (1) and (3) must be in mutual opposition at any instant, because (1) is j x, and the other is — jnmx. Their relative CO magnitudes, however, depend upon the value of the angular velocity TELEPHONE DIAPHRAGMS. 465 (0 of the particle in the orbit. It is evident that (1) diminishes with increase of co, while (3) augments. The vector diagram of Figure 25 represents the system of forces acting on the particle for the particular value coo, at which the two reactive forces (1) and (3), of resilience and inertia respectively, equate and cancel. The direction of reference OD, or standard phase, is the vector of instantaneous velocity x, at the moment selected. Then OA represents force (1) to magnitude and phase, or the vector reactive force of resilience tending to bring the particle to the center O, Figure 24. The equal and opposite reactive force OB, of inertia, tends to move it centrifugally away from O, FIG.Z4- p '«-j"x Fio.ZJ C-«- '^y-i-D Fig. 27 Figs. 24, 25, 26, 27. — Diagrams of Motional Equilibrium. in the direction of instantaneous displacement x, Figure 24. The force OC is directed opposite to the instantaneous velocity. The impressed vector rotating force F = OD, is directed in phase with this velocity. It balances the retarding force OC. The active forces OD and OC therefore cancel, while we have seen that the reactive forces OA and OB also cancel ; so that the system will retain a steady state of motion, with the angular velocity wq. The power put into the s^'stem by the impressed force is Fx ergs per second, and this is equal to the dissipa- tory output of the system rx"^, through the action of the frictionally retarding force. 466 KENNELLY AND AFFEL. It is well known that the forces in the simple rectilinear vibratory motion of a particle about a position of rest may be regarded as the projection, upon a line through this position, of the corresponding rotating system in a plane. The vibratory motions and forces of a particle subjected to (1) a simple vibratory resilient force —5.r dynes, (2) a motional retarding force —rx dynes, and (3) an inertia force — vix dynes, together with (4) a simple harmonic impressed force main- taining the motion, may therefore be considered as the projection of a vector system like that in Figure 25, on the reference line COD, when that system rotates counter-clockwise about the center O, with angu- lar velocity co radians per second. The instantaneous projection of OB will then be the inertia force opposing acceleration of the particle, for the instant considered, that of OA the instantaneous resilient force, that of OC the instantaneous frictional retarding force, and that of OD the impressed vibratory force, or vibromotive force (vmf). The instan- taneous velocity will be the projection of OD, when divided by r. The instantaneous displacement will be the projection of OA reversed, when divided by s. The instantaneous acceleration will be the pro- jection of OB reversed, divided by in. Consequently, for the case indicated in Figure 25, with reactive equi- librium, i. e., equality between the opposing reactive forces s.r/co and wz^i, the vibrational velocity .r, will be in phase with the im- pressed vmf. OD; while the vibrational displacement x will be 90° retarded behind the impressed vmf. At the impressed angular velocity less than coo, of reactive equili- brium, the reactive force (1) of resilience Jsx/cj: dynes, will be greater than the reactive force (3) of inertia —jmoix dynes. Such a case is indicated in Figure 26; where OAi, the resilient force, exceeds OBi the inertia force. Their difference is Ox\i\ the resultant reactive force. The impressed force F = ODi must now equilibrate the resultant of OAi^ and the motional retarding or frictional force OCi. It will then be seen that, in the steady state, the velocity .i: leads the impressed force by an angle au The displacement, in line with OBi, is now less than 90° behind the impressed force. At any impressed angular velocity greater than that of reactive equilibrium coo, the inertia force will overcome the resilient force. This is the case represented in Figure 27; where the vector inertia force OBo exceeds the vector resilient force OA2. Their difference 0B2\ is the resultant reactive force, which, combined with the frictional force OC2, gives the resultant force Oc/o to be overcome, or to be equilibrated by the impressed vector force F = OD2, which now leads TELEPHONE DIAPHRAGMS. 467 the velocity x, at standard phase, by the angle az. The displacement along OB2 is now more than 90° behind the impressed force. Consequently, as the angular velocity of constant impressed vmf, increases from zero to infinity, the angular velocity co commences at 90° lead with respect thereto, and indefinitely small magnitude, later comes in phase, with maximum value, at the angular velocity of resonance, or reactive equilibrium, and ends at 90° lag, again with indefinitely small magnitude. In Figure 28, the line OX represents the value of r, the resistance to motion, taken along the axis of reals. Xp is taken equal to the reactive diflFerence ; {vioo — -), drawn parallel to OY, the axis of CO imaginaries. The vector Oj), which may be called the mechanical impedance, will make an angle a with OX, equal to the phase dis- placement between the velocity and the impressed vmf. / / \ Fig. 28. Fig. 29. Locus OF Mechanical Impedance and Locus of Velocity under Varying Impressed Frequency. The equilibrium of vector forces represented in Figures 25, 26, and 27, for any steady state of angular velocity, may be expressed algebraic- ally by the formula or sx — Tx — mx + / = 0 sx -{- rx -{- vix = f — Fe-'"' dynes Z (15) dynes Z (16) 468 kennelly and affel. Solution in Terms of Velocity. The solution of the above differential equation, with reference to .r, is x= / j^^^^-M+jH-^-^^ ^ / (17) , ./ 5A sec where Xi is an initial vector velocity, and j\ = r/{2m) a numeric, the damping constant per second. The first term on the right hand of this equation indicates the condi- tions for the steady state of motion, while the second term is the transient term; i. e., comes into action only during changes of motion in the system, and when/ is either varied or withdrawn, the coefficients m, r, and s, remaining constant. We may, therefore, at present con- sider only constant impressed vmf. and ignore the second or transient term. The formula (17) then reduces to X = ^^- = ■'- — Z (18) 2 z sec where z is the mechanical impedance at the impressed and sustained angular velocity co. It is evident from Figure 28, that as w varies, the vector impedance OZ travels over the straight line y'X.y^. It is well known that the reciprocal of a variable having a straight-line locus is a variable with a circular locus, so that as co varies from zero to in- finity, the vector locus of x will be a circle OXP, Figure 29, with its diameter OX on the axis of reals, and equal in magnitude to F/r cms. per second. That is, the linear velocity of a simple vibratory sys- tem, having a retarding force directly proportional to x, follows a cir- cular locus in regard to magnitude and phase, coming into phase with the impressed force at the resonant angular velocity coo of reactive equilibrium. Certain relations between the fundamental constants m, r, and s of the vibrator may be determined from the observed distribution of angular velocities around the velocity circle. Thus, the angular velocity of resonance coo is found at the point where the diameter intersects the circle. Since, at resonance, the two reactive forces — = Tncoo, we have COo radians . <^o = Vs/m ~^^^ (19) TELEPHONE DIAPHRAGMS. 469 This gives one relation between s and m, in terms of the observed angular velocity of resonance. Moreover, it can easily be shown that the damping factor r (iP- — ui^ , . A = -— = per second (20) 2m 2cotana *^ ^ where co is an angular velocity on the velocity circle, at a point making an angle a with the diameter. If we select the angular velocity coi at the upper quadrantal point on the circle, for which the impedance angle a = —45° in Figure 28, the above formula becomes 2 2 A = — per second (21) 2coi Similarly, the angular velocity 002 at the lower quadrantal point for which a = + 45° gives A = — ;r per second (22) 2aj2 Summing the last two equations, we obtain A = — - — per second (23) so that the damping factor is half the difference between the angular velocities at the quadrantal points. Formulas (20) to (23), or any of them, furnish one relation between r and m, in terms of angular veloci- ties at observed points on the velocity circle. Some third relation is, however, required in order to evaluate m, r, and s. Transient Motion. We have seen that the second term of (17) becomes involved at any sudden change or discontinuity in the impressed force /, except in the case when the impressed force / happens to have the resonant angular velocity coq. In such a case there is no disturbance in phase relations; but there will be an instantaneous change in the lengths of the vectors OC and OD, Fig. 25, the vectors OA and OB remaining in equilibrium. It seems that in all other cases, a change in the im- pressed force / must be accompanied by a transient change in the motion of the system, due to the introduction of the second term in 470 KENNELLY AND AFFEL. (17). This term is a damped, or logarithmically decaying, periodic velocity, effected not at the angular velocity co of impressed force, but at the free angular velocity of the system co^ = Vcoo^ — A^ = ojo sin7 radians per second, unless A is larger than coo, in which case the second term of (17) is an ultraperiodic velocity, which may be repre- sented by the projection of a uniformly damped uniform angular velocity in a right hyperbola. -^^ For the particular case, cuo = A, or the aperiodic case, the motion may be represented by the projec- tion of uniformly damped uniform angular velocity in a parabola. In every case, therefore, the second term of (17) corresponds to the pro- jection of uniformly damped uniform angular velocity in a circle, hyperbola or parabola, i. e., a conic section. In the cases of the tele- phone-receiver diaphragms examined experimentally, A was much less than coo; so that only the periodic interpretation of damped uni- form angular velocity in a circle needs here to be considered. Damped Free-Vibration Vector Diagram. In Figure 30, let a particle at p, of mass m, be attracted cen- tripetally toward the center O, with a force varying directly as the displacement x, and be subjected to a frictional retarding force, oppo- Q Fig. 30. Fig. 31. Locus OF Damped Oscillatory Displacement, Velocity and Acceleration Force. ' site to the instantaneous velocity; as well as to an inertia force opposite to the acceleration. Then because there is no impressed force to give energy to the particle, energy will continually be absorbed from it, and 19 Bibliography No. 2 and No. 10. TELEPHONE DIAPHRAGMS. 471 the orbit of the particle will dwindle, until finally the particle will fall into the center O. This means that the orbit, instead of being a circle, will be an equiangular spiral, in which the tangent PP^ at any orbital position P, makes an angle of less than 90° with the reversed radius vector PO. The instantaneous acceleration will be directed along the tangent QQ^ of the spiral at the point Q, (180° — y°) in advance of P. The centripetal force will be directed along PO, the frictional force in the direction P^P, or parallel to QO, and the inertia force in the direc- tion Q^Q, or parallel to VO. These conditions are represented in the instantaneous force diagram, Figure 31, where Op is the centripetal force —sx, Oq is the frictional force —rx, and Ov is the inertia force —mx, X being the instantaneous displacement: x= Xoe^-^+^"">^ cms Z (24) so that r.i- = r (— A + joo')xo 6^-^+-''"')* = r (— A + jo:')x dynes Z (25) and mic = m (- A + jw')^ Xo e(-^+^'"')« = m (- A + jo^'Y x dynes Z (26) For equilibrium we require that ^° — sx — rx — mx = 0 dynes Z (27) or -s -r(- A + jo:') - m (- A + jcoj = 0 '-^^"^ Z (28) cm radians whence w' = Vajg^ — A- = coo sin 7 (29) where — = cos 7 numeric (30) coo Each of the quantities x, x, '.i, pursues an equiangular spiral around the center 0, or may be considered to pursue a circular path with uni- form angular velocity, subject to an independent damping factor e"^'. In the case of simple rectilinear vibrations, the projections of the spiral motion may be taken on a reference axis COD. The initial value of the velocity .i'l must be such as meets the physical conditions of the system at the moment of a sudden change in the impressed vibromotive force. After the change, the vibratory motion will be the sum of the two terms in (17) or, the sum of the projections of the 20 Bibliography, No. 8. 472 KENNELLY AND AFFEL. respective vectors in Figures 26 and 31, the former being rotated at the impressed angular velocity co, and the latter at the free angular velocity oo^ The latter motion, however, speedily expires by damp- ing, leaving the former in the steady state without further interference. Solution in Teems of Displacement. We have' hitherto considered only the solution of (15) in terms of vibratory velocity^ a*. We may, however, find the solution in terms of the displacement x, by integrating (17) as follows: 1 \ F e^'"' 1 icoj , Y 6-\ -A + jco' cm Z (31) r + J mo: ' CO \J03/ Z V^O / .i-A+j^')tyy cm Z (32) that is, the vector displacement of steady motion lags 90° in phase behind the vector velocity, and is equal in magnitude to that velocity divided by co. Also the vector displacement for transient motion lags 180°-7° behind the vector velocity, and its magnitude is that velocity divided by coq. All of the preceding theory is immediately applicable to the case of a simple alternating-current circuit, containing resistance, inductance, and capacitance in simple series, with an impressed sinusoidal emf. when current strength i is substituted for velocity x, quantity q for displacement x, inductance L for mass m, elastance S for elastic factor s, and electric resistance R for mechanical resistance r. Applications of Simple Vibrator Theory to Telephone- Receiver Characteristics. It can be shown ^^ that on the simple vibrator theory of telephone diaphragm vibration, the vibromotive force may be expressed as /j = At instantaneous dynes (33) where f, is the alternating electromagnetic pull exerted on the equiv- 21 Bibliography, No. 11. TELEPHONE DIAPHRAGMS. 473 alent mass of the diaphragm by the pole or poles of the receiver, when sinusoidal alternating current, i = I^e^"' absamperes, passes through the winding. A is a constant of the receiver, depending upon its structure. It represents the force exerted on the diaphragm per unit current. It may be defined by the expressions: — dynes/absampere (34) . ^ N§So_ in a monopolar receiver, and A = " = 2N3g.o5 dynes/absampere (35) in a bipolar receiver, where N is the number of turns in the winding, including all coils, 3Jo is the mean flux density in the air-gap at the normal position of rest, due to the permanent-magnet system, in the absence of electric current excitation, and 31 is the reluctance of the magnetic circuit to alternating mmfs. 2 = l/si, is the permeance of the circuit to such mmfs. Consequently, from a magnetic point of view, the strength of the receiver is A/N dynes per absampere per turn of exciting winding. This is equal to the product of ^o"!, the permanent normal flux-density and the permeance of the a.c. mmf. The maximum cyclic pull on the diaphragm will be A I^ dynes, and the rms. pull A I^/ V2 dynes. Since, however, owing to the effects of hysteresis and eddy-currents, the alternating flux in the air-gap or gaps of the receiver will lag behind the exciting alternating current by some angle ^]°, the instan- taneous pull will not be in phase with the current, taken at standard phase, but will be expressed by fi = A.i \J, dynes Z (36) Consequently, to current as standard phase, the velocity x in (6) will be X = — \^i -— Z (37) z sec. The alternating emf. induced in the winding by the rate of alteration of air-gap, and flux in the permanent magnetic circuit, will be e^ = A.x \^2 ab volts Z (38) 474 KENNELLY AND AFFEL. where A is the same force-constant as is defined in (34) and (35); or, substituting from (37) : AH ex = — 2; \|8i + i82 ab volts Z (39) From such experimental observations as have yet been made, it appears that jSi = /32 = |8, say; or that the lag of flux in the a.c. magnetic circuit, associated with the alternation of air-gap reluctance, is nearly the same as that associated in the same magnetic circuit with alternation of current. It is hoped to make further investigations on this point. In any case, however, if we consider that /3i + /So = 2jS, we have for the motional impedance Br A . , AH , A2 Z = ^ = V .r \02 = ^- \2i8 = — \2|8 absohms Z (40) This is the vector change in the impedance of the receiver winding, occurring at any assigned frequency, between the conditions of free vibration and suppressed vibration. Since in (40) the denominator z, the mechanical impedance, follows a straight-line locus (Fig. 5), as the frequency is varied, the system otherwise remaining constant, the motional impedance Z must follow a circular locus, similar to that of Figure 29, except that the diameter of the circle will be depressed be- low the axis of reals by the angle 2/3. From this circle, corrected when necessary for the effect of change in impressed frequency on A, the natural frequency coo and the damping constant A can be determined according to formulas (20) to (23). Solution for Diaphragm Constants .1, m, r and s, when x^ is GIVEN. From (36) and (38), taking current phase as standard, . ^ fi ^^^ ^frn^ ej^'i ^ e:cm dyues ,^j. i Im X Xm absampere and from (18), at resonance. fm = rxm dynes (42) where /m is the maximum cyclic value of the impressed force; and x^ the maximum cyclic velocity in phase therewith, also from (40) Cxm = Im T^m abvolts (43) TELEPHONE DIAPHRAGMS. 475 where em is the maximum cycHe induced voltage, I^ the maximum cycHc current at resonance and Z^ the maximum motional impedance, or the diameter, in ohms, of the motional impedance circle. Substi- tuting in (41) we have XmT ImZm dynes ,^^. or T 2 'Z J. (45) and Im^Zm = (i'm)"^ = Pm abwatts (46) where P„ is the maximum cyclic mechanical power at resonance. In these last equations the phase angles disappear and the magnitudes or vector moduli only are presented. The maximum cyclic velocity has a magnitude represented by Xm = oooXm kines (47) Xjn being the maximum cyclic displacement as measured at the resonant frequency coo over an air-gap i. e., near the center of the diaphragm. Consequently, _ Im^Zm dynes , , -m Xiti absampere T 27 dynes ' " {XmY kine = (--i'm)" r = P« abwatts Xm^ oi .= V^T. kine 1 cm. per sec. (unit velocity in C. G. S. system). L Inductance of receiver winding damped, (abhenrys). Li " " " " free, (abhenrys). M Total active mass of diaphragm within clamping circle, (grams). m Equivalent mass of a diaphragm, (grams) . mi Massattached to the center of a diaphragm as load, (grams). M Symbol for 10~^ mm., or 1 micron. N Number of turns in winding of a receiver, including both poles, (numeric) . n Number of annuli into which a vibrating diaphragm may be sup- posed to be divided . (numeric) . Pm Maximum cyclic mechanical power in diaphragm, (abwatts or ergs/sec) . 5! Permeance of magnetic circuit to alternating mmfs. (oersted)"'. TT = 3.14159 R Resistance of receiver winding, diaphragm damped, (absohms or ohms). R' Resistance of receiver winding, diaphragm free (absohms or ohms). 482 KENNELLY AND AFFEL. eH Reluctance of Magnetic Circuit to alternating mmfs. (oersteds). r Equivalent motional resistance of a diaphragm, (dynes/kine) . r In theory of equivalent mass, the radius of a point on diaphragm. (cm). pi Superficialdensity of diaphragm, (gm/sq. cm.) s Stiffness constant of a diaphragm, (dynes per cm.) t Time elapsed from a certain epoch . (seconds) . vmf. Contraction for vibromotive force, or alternating mechanical force tending to set up vibration, (dynes Z ). W Maximum cyclic kinetic energy of diaphragm, (ergs). Wr Maximum cyclic kinetic energy of an annulus of width dr, and radius r. (ergs). X Reactance of receiver winding, diaphragm damped, (absohms or ohms). X* Reactance of receiver winding, diaphragm free, (absohms or ohms) . XY Cartesian rectangular coordinates in a plane. X Vibratory displacement of a diaphragm, at any time, from its static position of rest. (cm). Xm Maximum cyclic vibratory displacement of diaphragm at any frequency, (cm). Xo Maximum cyclic vibration amplitude at center of diaphragm. X Vibratory velocity of a diaphragm at any time, (kines Z ). Xm Maximumcyclic vibratory velocity of diaphragm, (kines). X Vibratory acceleration of a diaphragm at any time, (kines/sec). Xi Initial velocity of a simple vibrator subject to damping, (kines) . xh xh x\ Maximum cyclic displacements at mid-radii of annular diaphragm segments, (cm). Xr Vibrational amplitude of diaphragm at radius r, and any instant (cm). Z Motional impedance of a receiver at any frequency, (absohms Z ) . Zm Zms Maximum or diametral amplitude of motional impedance at resonance, (absohms). . z = r + jz Mechanical impedance of diaphragm, (dynes/kine Z ) • w Angular velocity of vibratory simple-harmonic motion (radians/ sec). wo Angular velocity of vibratory simple-harmonic motion at reso- nance (radians/sec). co' Free angular velocity in presence of damping, (radians/sec.) woi Resonant angular velocity of a loaded diaphragm, (radians/sec.) Z Sign of a complex quantity. <^5 Symbol for "cycles-per-second." Proceedings of the American Academy of Arts and Sciences. Vol. 51. No. 9. — December, 1915. CONTRIBUTIONS FROM THE ZOOLOGICAL LABORATORY OF THE MUSEUM OF COMPARATIVE ZOOLOGY AT HARVARD COLLEGE.— No. 264. ON THE DEVELOPMENT OF THE CORAL AGARIC Li FRAGILIS DANA. By J. W. Mayor. With Six Plates, and Five Figures in Text. CONTRIBUTIONS FROM THE ZOOLOGICAL LABORATORY OF THE MUSEUM OF COMPARATIVE ZOOLOGY AT HARVARD COLLEGE.— No. 264. ON THE DEVELOPMENT OF THE CORAL AGARICIA FRAGILIS DANA.i By J. W. Mayor. Presented by E. L. Mark. Received June 15, 1915. CONTENTS Page. Introduction 486 Part I. On the larval development 487 1 . General considerations 487 a. Breeding season 487 b. Extrusion of the larvae 487 c. Form of the larva 487 d. Fixation 488 2. Anatomv 489 a. Material, methods and general features 489 b. Description of the mesenteries, mesenterial filaments and gas- trovascular cavities of the larvae studied . 492 Group I. Larva A 492 II. Larva B 493 " " Larva C 494 " Larva D 495 " Larva E 496 " III. Larva F 497 Larva G 499 Larva H 500 " IV. Larva I 501 c. Conclusions as to the course of development of mesenteries, etc. 504 Part II. On the postlarval development 506 1. Form of the young polyp 506 2. The early development of the skeleton 506 a. Description of skeleton A 506 b. Description of skeleton B 508 c. Conclusions on the early development of the skeleton .... 508 d. Stage with twelve primary septa 509 3. Comparison of the development of the skeleton in Agaricia fragilis with that in other Hexacorallidae 509 Bibliography 510 Description of plates 511 1 Contributions from the Bermuda Biological Station for Research. No. 38. 486 MAYOR. 1. Introduction. Interest has been re-awakened of recent years in the development of the Hexacorallidae by the papers of Duerden (:04), who has estab- lished the order of development of the mesenteries in the larva and of the septa in the attached polyp. As regards the larval develop- ment his work agrees in the main with that of Wilson ('88) on Manicina areolata. In the skeleton Duerden is the first to find bilat- eral symmetry in the order of development of the primary exosepta. Our knowledge of the development of coral larvae is in the main confined to four species; Astroides calycidaris (Lacaze-Duthiers '73); Manicina areolata (Wilson '88) ; Caryophyllia cyathns (Von Koch '97) and Siderastrea radians (Duerden :04). The early development of the skeleton has been studied in Astroides calycidaris by Lacaze- Duthiers ('73), in CaryojihyUia cyathus by Von Koch ('97), in Caryo- phyllia clavns, C. smithii and Balanophyllia regia by Lacaze-Duthiers ('97) and in Siderastrea radians by Duerden (:04). In 1907 the writer went to the Bermuda Biological Station for the purpose of studying the development of a recent coral. ^ He was fortunate enough to find the common "hat" or "shade" coral, Agari- cia fragilis, breeding and was able to rear the larvae. As the development of the primary mesenteries of the larvae of corals has been worked out in comparatively few cases and a study of the soft parts seemed a necessary prelude to a study of the skeleton, free swimming larvae have been studied both in the living state and in paraffin sections. These observations form the first part of the paper. The writer was not very successful in rearing the young polyps. In consequence only a few skeletons were obtained. The second part of the paper is devoted to a description and discussion of these young skeletons. 2 The writer wishes to express his indebtedness to Dr. E. L. Mark for kind assistance and criticism while at the Biological Station and later in the Zoo- logical Laboratory of Harvard University. DEVELOPMENT OF AGARICIA. 487 PART I. ON THE LARVAL DEVELOPMENT. 1. General Considerations. a. Breeding Season. The jBrst larvae of Agaricia fragilis were obtained from a colony of the coral collected in a cave on the shore of Agar's Island, Bermuda, on July 8th, 1907. The colony, which was about eight or nine centi- meters in diameter, was brought to the laboratory in the morning and placed in fresh sea water in a battery jar. In the afternoon numerous pear-shaped, light brown, larvae were seen swimming about in the water. Other colonies collected from a cave on Tucker's Island on July 15th when placed in fresh sea water in the laboratory gave off similar larvae. In 1908 larvae were obtained from seven out of eleven colonies over 5 cm. in diameter collected from Long Island, Bermuda, between the 22nd and 30th of June. No larvae were obtained from nine colonies under 5 cm. in diameter collected on June 21st and 22nd from the same place. These observations show that Agaricia fragilis may be found breed- ing at the Bermuda Islands during the latter part of June and the first part of July. b. Extrusion of the Larvae. For the purpose of obtaining the larvae, adult colonies of Agaricia fragilis were collected in caves and brought to the laboratory in battery jars. During the transference to the laboratory it is probable that in many cases the temperature of the water containing the coral was raised above the temperature of the water in the cave. Larvae were often extruded in large numbers while the corals were being trans- ported to the laboratory and during the few hours immediately suc- ceeding this. What the factors were which produced their extrusion was not determined. Usually, however, not all of the larvae were extruded at this time, a certain number being seen to remain within the parent colony. c. Form of the Larva. The planula, which is light brown in color, is capable of considerable change of form. It may, however, be described as piriform, the 488 MAYOR. broader, rounded end being aboral and in advance during swimming, while the oral opening is situated at the more pointed posterior end. The alterations in form may be of two kinds; first, elongation, with corresponding decrease in the radial axis, or shortening, with corre- sponding increase in the radial axis ; or, secondly, a contraction of one side of the planula so that the aboral end, which in such cases is usually somewhat flattened, becomes turned to one side. In Sideras- trea radians Duerden (:04) found the broad rounded end of the larva to be the oral end. The elongated form is the one usually assumed by the larva when swimming rapidly through the water. The larva takes a more defi- nitely piriform shape when it swims slowly over the substratum. In the latter case there may be a slight in-pushing of the aboral end form- ing a hollow in the center. This hollowing out of the center suggests a mechanism working by suction. When the larva comes to rest and applies itself to the substratum, it becomes hemispherical, the aboral end being the flat side of the hemisphere and applied to the substratum. Later these larvae may become almost disk shaped. In the preserved specimens the shape is usually either piriform or almost hemispherical. After having become hemispherical and applied to the substratum, the planula may detach itself, elongate and swim away. d. Fixation. The planula swims with the broader, rounded, aboral end foremost, as already stated, rotating on its longitudinal axis as it does so. In Siderastrea radians, which has the mouth at the broader end of its piriform larva, Duerden (:04) found that, as is the case in Agaricia fragilis, the aboral end is kept in front. From this it would seem that while the form of the larva of Siderastraea is favorable for locomotion that of Agaricia must satisfy other conditions than that of offering the least resistance to forward motion. Larvae after they are expelled from the parent colony are usually elongated in shape and swim through the water rapidly. Later they become shorter and broader and swim more slowly. The normal course seems to be for them to affix themselves to the substratum within a few hours after they have been extruded. In the laboratory larvae which did not become fixed in about twenty-four hours after extrusion did not do so when kept for seven days, although during this time they continued to swim through the water and also, as flat DEVELOPMENT OF AGARICIA. 489 disks, to move over the substratum with the aboral end downwards. The failure of such larvae to become attached may have been due to the surface of the glass vessel being unsuitable for fixation. As, however, some of the larvae became fixed on glass while others did not attach themselves even to rough surfaces, such as stones placed in the vessel, the nature of the substratum does not seem to have been the only factor involved. The experiment was tried of keeping larvae in the dark and also under additional pressure, — eighteen inches of water, — but in both cases they failed to attach themselves. When the larva is about to fix itself it becomes flattened and its aboral end is applied to the substratum. If a stream of water from a pipette be forced against such a larva, the animal may be made to elongate its body again, provided it has not begun the formation of a skeleton. Such larvae may remain in the current of the pipette, appearing as if attached to the surface of the glass by an elastic strand. This suggests that an adhesive mucous substance may be secreted at the aboral end when it becomes applied to the glass. In the free- swimming piriform larvae, especially when they are swimming close to the substratum and apparently in a condition to affix themselves, a concavity may be seen at the aboral end, giving that end the appear- ance of a suction disk. Sometimes larvae flattened themselves out into disks at the surface of the water. Such larvae never attached themselves to the bottom or sides of the glass jar and almost all the individuals of such lots became flattened out under the surface of the water. Such larvae tended to fuse into " aggregations " (Duerden : 04) and also to go to the sides of the vessel (surface tension). Many of these larvae lived to secrete a skeleton with six well developed primary septa while still floating at the surface of the water. 2. Anatomy. a. Material, Methods and General Features. A considerable number of larvae, fixed in various mixtures, were embedded in paraffin and sectioned either transversely or parallel to the oral-aboral axis. Transverse sections were found to be by far the most suitable for the study of the general structure. Nine of the larvae so sectioned have been selected for detailed description in order that the reader may have in as concrete a form as possible the data on 490 MAYOR. which the conclusions of the paper are based. Although only indi- vidual differences have been observed between the opposite sides of the bilaterally symmetrical planulae, it has been thought best for convenience in description to distinguish a right and a left side. In this connection the aboral end of the planula, which is foremost in locomotion, has been considered anterior while the oral end is con- sidered posterior. Each larva was cut into a series of transverse sections of equal thick- ness, which varied from 5 /x to 7 ^i for the different series. In record- ing the position of structures in the larvae the thickness of the sections has been used as a unit of measure. The nine larvae fall into four groups: — First, larva A, which has four pairs of mesenteries, only two of which are well developed ; Second, larvae B, C, D, and E, in which there are six pairs of mesenteries, the fifth and sixth however being only slightly developed and the mesen- terial filaments not extending the whole length of the larva; Third, larvae F, G and H, in which all six mesenteries are well developed and the mesenterial filaments of the first two pairs of mesenteries extend through the greater part of the larvae; Fourth, larva I, in which, besides the greater development of the fifth and sixth pairs of mesenteries, there are developed mesenterial filaments on the third pair of mesenteries. The musculature of the mesenteries consists of fibres developed in endoderm cells where these abut on the mesogloea. These fibres stain deeply with haematoxylin and so are easily distinguished in the preparations. In the mesenteries the majority of these fibres run longitudinally. Some, however, especially those near the junction of the mesogloea of the mesentery with that of the body wall, run obliquely or even transversely. It is not always easy to deter- mine which side of the mesentery shows the large number of fibres; first, because the mesogloea in the mesenteries is usually very thin and does not take stain well, and, secondly, because there are always some fibres on each side. This is still more difficult in the newly formed mesenteries, as in these there is almost always an approxi- mately equal number of fibres on each side of the mesentery. The distribution of the muscle fibres given in the diagram (text Fig. 1, C) is based on careful examination with a 2-mm. apochromatic objec- tive and compensating oculars x8 and xl2. In the first and second pairs of mesenteries the muscle fibres are confined almost entirely to the ventral sides. In the third and fourth pairs, the directives, the muscle fibres are more numerous on the lateral sides. In the fifth DEVELOPMENT OF AGARICIA. 491 and sixth pairs, when a difference can be observed between the two sides of the mesenteries, the fibres are most numerous on the dorsal side (Fig. 2, E). In the mesenteries in which a mesenterial filament is developed (first and second pair), the muscle fibres are most numerous in that region of the mesentery which is about twice as far from the body wall as from the internal margin of the mesentery, in those sections in which the filament appears. In all other cases the muscle fibres are most numerous along the internal edges of the Figure 1. Agariciafragilis. Larva A. A , a longitudinal section through the first pair of mesenteries and their filaments, reconstructed from transverse sections. Viewed from the dorsal side. The numbers, and the dotted lines corresponding to them, indicate the numbers and positions of the transverse sections of the series. Ectoderm and mesoderm black. B, C, D, E, and F represent sections number 14, 18, 23, 30 and 41, respectively. The reader views the aboral face of the section, and dorsal is up, so that the right of the larva is on the left of the figure. In C, J, II, III, IV indicate the first, second, third and fourth pairs of mesenteries, respectively. X 90. mesenteries and are usually numerous on both sides of the mesentery along this edge. The musculature of the body wall consists of a layer of longitudinal fibres developed in the inner ends of the ectoderm cells where they abut on the mesogloea and of a layer of circular muscles developed in the endoderm cells where they reach the mesogloea of the body wall. The mesoderm of the body wall, although thin, is everywhere clearly to be seen. 492 MAYOR. At the oral end the ectoderm of the body wall is folded in to form the oesophagus and the filaments of the first two pairs of mesenteries. At its oral end the oesophagus forms a complete tube. Aborally it becomes divided dorsally so that it extends for some distance as a scoop-shaped structure, U-shaped in cross section. In the following descriptions the point where the oesophagus ceases to be a complete tube and becomes scoop-shaped is expressed by saying that it is interrupted dorsally. b. Description of the Mesenteries, Mesenterial Filaments and Gastro- vascular Cavities of the Larvae Studied. Group I. Larva A. In this larva (text, Fig. 1) the first two pairs of primary mesen- teries are well developed and there are indications of the third and fourth pairs. The larva was fixed in Flemming's chromo-aceto- osmic mixture and the sections were stained with Heidenhain's iron haematoxylin. The aboral end of the larva was sectioned first and the sections are almost exactly transverse, the ventral side of the oral end, however, extends six sections beyond the dorsal side, a deviation from the transverse which cannot be detected at the aboral end. The sections are 7 micra thick and there are 56 of them in all. Text-figure 1 shows a.t A, a reconstructed longitudinal section passing through the mesenterial filaments of the first pair of mesenteries. The horizontal dotted lines are drawn in the planes of sections and each of the numbers at the side of the figure indicates the number of the section to which the line corresponds, beginning with the oral end. The figure shows that the aboral end of the larva is flattened and contains a slight concavity. This shape may be due to con- traction on the application of the fixing fluid, but larvae of this form were observed in the living state. At B, C, D, E, and F, in Figure 1, are shown the 14th, 18th, 23rd, 30th, and 41st sections respectively from the oral end. Photomicrographs of sections 15, 24, 30, and 46 are shown in Plates 1 and 2. The oesophagus is irregularly circular in cross section (Fig. I, B) and extends aborally in the axis of the larva. The fact that it appears nearer to the dorsal side in the figure Is due to the section having been cut somewhat obliquely. The larva has been cut so that the oesoph- agus appears as a complete tube first in section 6, the ventral side DEVELOPMENT OF AGARICIA. 493 only appearing in the earlier sections. In section 8 the endoderm appears on the ventral side and in section 14 on the dorsal side for the first time. The oesophagus is interrupted first ventrally in section 16 and then dorsally in section 19. In the sections the oesophagus is cut obliquely on the ventral side while on the dorsal side it is sec- tioned transversely. In spite of this the oesophagus extends on the ventral side through 12 sections, while on the dorsal side it extends through only 6 sections. Although interrupted on the ventral side, the ectoderm of the oesophagus is continuous with the ectoderm of the mesenterial fila- ments of the first two pairs of mesenteries (Fig. 1, C). The mesen- terial filaments of mesenteries / and // of the right side of the larva, (left side of the figure) are fused together as far as section 19, after which they become separated; those of the left side are fused as far as section 21. The mesenterial filaments of the first pair of mesen- teries extend on the right side of the larva as far as section 43; on the left side of the larva as far as section 41. They are larger than the filaments of the second pair of primary mesenteries and there is no evidence of an aboral enlargement as found in the larvae of the later stages. The mesenterial filaments of the second pair of mesen- teries extend on the left side into section 28 and on the right side into section 34. The mesenteries of the first pair extend on the right side of the larva from section 9 to 46, and on the left side from section 1 1 to 48. Those of the second pair extend on the right from section 13 to 46, on the left from section 14 to 46. The beginning of the development of the third pair of mesenteries appears on the right of the larva be- tween sections 8 and 31 and on the left between sections 10 and 27. At their oral ends the third pair of mesenteries extends to the ectoderm of the oesophagus. Slight indications of the fourth pair of mesen- teries are found on either side in sections 12 to 18 (Fig. 1, C). Here also the mesentery as defined by mesoglea and muscles extends to the ectoderm of the oesophagus. The distance to which the muscles and mesoglea of the fourth pair of primary mesenteries extend into the endoderm is less than in the third pair. Group II. Larva B. This larva has six pairs of mesenteries all clearly shown. It was fixed in a saturated solution of corrosive sublimate to which 1% acetic acid was added, and the sections were stained in Heidenhain's iron 494 MAYOR. haematoxylin and congo red. The sections deviate very slightly from the transverse, the right dorsal side of the oral end of the larva being cut first. As the oral end was cut first the right side of the larva is to the right in the preparation. The sections are 7 )U in thickness and there are 57 of them in all. As in the previous larva, the aboral end is somewhat flattened. The oesophagus is circular in cross section, and extends on the ventral side through one more section than on the dorsal side. On the ventral side it extends into section 22. After leaving their junction with the oesophagus in section 22 the mesenterial filaments of the first and second mesenteries of either side are not fused as in larva A. Those of the first pair of mesenteries extend on either side into section 42 and those of the second pair are turned from the aboral end of the oesophagus outwards toward the oral end of the larva extending on the right into section 12 and on the left into section 17. The filaments of the first pair of mesenteries are enlarged on the right side between sections 30 and 37, and on the left side between sections 37 and 39, those of the second pair of mesenteries decrease gradually in size until they end. Only mesenteries / and // are complete. The lengths and positions of the mesenteries are given in the follow- ing table, which shows the number of the section where each mesentery begins and ends: Number of mesentery 1 II III IV V VI Right side of larvae Oral end section No Aboral end section No. Left side of larvae Oral end section No. Aboral end section No. Although traces of the fifth pair of mesenteries extend into section 44 these mesenteries are only slightly developed in the aboral half of the larva. Group II. Larva C. This larva is in about the same stage of development as larva B, but the fifth and sixth pairs of mesenteries show more clearly. It was fixed in a saturated solution of corrosive sublimate plus 1% acetic acid. The 90 sections were stained with Heidenhain's iron haematoxylin and congo red, each being 5 /x thick. The aboral end 7 9 10 11 8 1 45 38 44 36 44 31 11 11 12 11 12 12 4S 40 44 36 44 36 DEVELOPMENT OF AGARICIA. 495 of the larva was cut first and on the dorsal side, so that the ventral side of the oral end occurs in nine sections after the dorsal side has ceased to appear. The oesophagus is almost round in cross section and is shorter on the dorsal side than on the ventral side, the length of the dorsal side being approximately 50 /x, while that of the ventral side is 170 /x. The mesenterial filaments of the first and second pairs of mesen- teries are continuous with the ectoderm of the oesophagus. On the right side they are fused in section 24 but below that separate, the filament of the first mesentery extending into section 64, that of the second into section 57. On the left side they are fused as far as section 45, below which the filament of the first mesentery extends as far as section 63 ; the second mesentery has no separate mesenterial filament. The filaments of the first pair of mesenteries are enlarged and crescent shaped in cross section on the right between sections 43 and 61 and on the left between sections 40 and 59. The filaments of the second pair of mesenteries are not thus enlarged and are approximately uniform in cross section throughout their length. The lengths and positions of the mesenteries are given in the fol- lowing table: Number of mesentery / II III IV V VI Right side of larva Oral end section No. Aboral end section No. Left side of larva Oral end section No. Aboral end section No. Only mesenteries / and // are complete. Group II. Larva D. Larva D, which was in nearly the same stage of development as larvae B and C, was fixed in a saturated solution of corrosive sublimate, and the sections were stained in Mallory's phosphotungistic acid haema- toxylin. The plane of sectioning deviated slightly from the trans- verse, the first part of the larva cut being slightly dorsal to the middle of the right side. The 47 sections are each 7 /x thick. The oesophagus is elliptical in cross section, the long axis being dorso-ventral. 15 16 7 20 12 20 69 74 55 82 47 66 11 16 7 20 8 21 68 74 55 82 47 66 496 MAYOR. The mesenterial filaments of the first and second pairs of mesen- teries of the right side are not fused together after they leave the oesophagus. The first of these extends from sections 13 to 26, the second from sections 13 to 25. On the left side the filaments are fused from their junction with the oesophagus at section 13 as far as section 17, below which the filament of the first mesentery extends to section 29, that of the second to section 25. The filaments of the mesenteries are enlarged as follows: those of the first pair on the right between sections 19 and 22, on the left between sections 18 and 23, those of the second pair on the right between sections 19 and 23, on the left between sections 18 and 21. The length and position of the mesenteries is shown in the following table : Number of mesentery / // III IV V VI Right side of larva Oral end section No. Aboral end section No. Left side of larva Oral end section No. Aboral end section No. Only mesenteries /, // and /// are complete. Group II. Larva E. This larva was fixed and stained like larva D. The aboral end was cut first and the sections are so nearly transverse that all sides of the circular oesophagus appear in the next to the last section of the series. There are 55 sections, each 7 ^t thick and they are numbered from the oral end. The oesophagus is elliptical in cross section. Its dorsal side ends at section 7 and its ventral side at section 11. The mesenterial filaments of the first and second pairs of mesen- teries are separate where they leave the oesophagus. The filaments of the first pair extend on either side from the oesophagus, at section 11, to section 47. Those of the second pair extend on the right to section 31 and on the left to section 35. The filaments of the first pair are enlarged in cross section on both sides between sections 23 and 44. Those of the second pair on the right between sections 16 and 29 and on the left between sections 16 and 34. The position and extent of the mesenteries is shown in the following table: 7 6 9 6 14 5 35 31 38 34 37 34 9 10 9 7 13 18 14 39 40 34 42 38 4 4 5 5 6 6 49 49 49 44 49 44 6 6 5 5 9 9 50 49 51 44 50 44 DEVELOPMENT OF AGAKICIA. 497 Number of mesentery / II III IV V VI Right side of larva Oral end section No. Aboral end section No. Left side of larva Oral end section No. Aboral end section No. In section 49 mesenteries / and II of the right side are united at their internal ends. Only mesenteries / to IV are complete. Group III. Larva F. This larva (text, Fig. 2) is not so much contracted along its oral- aboral axis as the one last described, and the oesophagus is only slightly invaginated. It was fixed in a saturated solution of corrosive subli- mate and the sections stained in Mallory's phosphotungstic acid haematoxylin. The oral end of the larva was cut first. The sections are almost exactly transverse. The oesophagus appears as a com- plete circle of ectoderm in the third section, the first two sections showing the dorsal side only. The sections are 6 /x thick and there are 86 of them. Of the sections, Figure 2, A-I, shows numbers 9, 13, 18, 21, 33, 45, 61, 71 and 74 and Plates 3 and 4 show photomicro- graphs of numbers 12, 17, 61 and 73. The oesophagus is elliptical in cross section, the long axis being dorso-ventral. It ceases on the dorsal side at section 9 (Fig. 2, A). Below this its cross section is U-shaped until section 13 (Fig. 2, B) is reached. Below section 13 it is interrupted on the ventral side also. The invaginated ectoderm of the right side of the larva is joined by the first, second and third mesenteries, that of the left side is joined by the first and second mesenteries (Fig. 2, C). Below section 20 the ectoderm is continued on both sides as the separate mesenterial filaments of the first and second pairs of mesen- teries. The filaments of the first pair extend on the right as far as section 70 and on the left as far as section 73, which is practically the whole extent of the gastrovascular cavity. The filaments of the second pair extend on both sides as far as section 56. The mesenterial fila- ments are enlarged as follows: those of the first pair of mesenteries on the right between sections 49 and 66, on the left between sections 52 and 70, those of the second pair of mesenteries on both sides between sections 32 and 54 (Fig. 2, E-H). The six pairs of primary mesenteries extend practically the whole 498 MAYOR. length of the endoderm of the larva. The following table shows their positions and extent: Number of mesentery Right side of larva II III IV Only mesenteries I to IV are complete. V VI Oral end section No. 8 8 9 8 9 13 Aboral end section No. 75 75 72 70 72 75 Left side of larva Oral end section No. 9 8 9 8 10 12 Aboral end section No. 77 76 72 70 72 77 Figure 2. Agaricia fragilis. Larva F. Transverse sections the oral end having been cut first. The right of the larva is to the right in the figures. Ectoderm and mesoglea colored black. A, B, C, D, E, F, G, H, I sections number, 9, 13, 18, 21, 33, 45,61, 71, and 74, respectively. In E, I, II, III, IV, V and VI mark the first, second, third, fourth, fifth, and sixth sections. X90. DEVELOPMENT OF AGAKICIA. 499 Group III. Larva G. The chief difJerences between this larva and larva F pertain to the form of the oesophagus and gastrovaseular cavity. It was fixed in Flemming's fluid and stained with Heidenhain's iron haematoxyhn. The series consists of 154 transverse sections each 7 /jl thick. The aboral end was cut first, so that the sections are viewed from this end, but the numbering of the sections is from the oral end. The oesophagus is well invaginated and is circular or slightly oval in cross section (Fig. 3, ^) as far as section 20. It is then interrupted on the dorsal side, as seen in sections 21-23 (Fig. 3, B), after which the ectoderm becomes divided also on the ventral side. The slight deviation of the sections from the true transverse plane is such that their ventral edges are more aboral than their dorsal edges, a fact which can be learned from the table showing the extent of the mesen- FiGURE 3. Agariciafragilis. Larva G. Transverse sections, aboral aspect. Ectoderm and mesoglea colored black. A, section 19; B, section 23. X 90. teries, and from the fact that the last few sections cut the ventral side only of the larva. There can be no doubt, therefore, that the oesophagus is interrupted on the dorsal before it is on the ventral side. There are only two pairs of mesenterial filaments, those of the first and second pairs of mesenteries. The filaments of the first and second mesenteries are fused on the left from sections 24 to 31 and on the right from sections 24 to 33. The sections of the filaments of the first pair of mesenteries decrease in size until section 57 is reached, after which they increase in size on both sides up to about section 124, where they are somewhat larger than those of the second pair of mesenteries. After this they decrease in size till they end, on the right at section 135 and on the left at section 132. The filaments of the second pair of mesenteries increase in size up to about section 57. In section 40 they have about the same size as those of the first pair of mesenteries 500 MAYOR. in that section. From sections 57 to 61 the size remains nearly constant, but below this it decreases until they end, on the right at section 106 and on the left at section 102. The extent and position of the mesenteries is shown in the accom- panying table. Number of mesentery I II III IV V VI ight side Oral end section No. 13 14 9 12 12 14 Aboral end section No. 144 147 139 135 142 142 eft side Oral end section No. 8 9 7 11 8 8 Aboral end section No. 138 135 138 134 137 133 Only mesenteries / to IV are complete. The gastrovascular cavity extends orally between the mesenteries as in larva B. At the oral end however it is found between the first four pairs of mesenteries only (Fig. ^, A). Where these diverticula of the gastrovascular cavity occur the fifth and sixth mesenteries do not reach to the oesophageal wall. Group III. Larva H. This larva, which is in about the same stage as larvae F and G, is very much contracted along the oral-aboral axis, the aboral surface being retracted to form a deep concavity. The individual was fixed in a saturated solution of corrosive sublimate plus 1% acetic acid. The sections were stained with Heidenhain's iron haematoxylin and Congo red, and the oral end of the larva was cut first, the right side being slightly in advance. The 71 sections are each 7 ju thick. The oesophagus is elliptical in cross section, the long axis of the ellipse being dorso-ventral. It is interrupted on the dorsal side in section 14 and on the ventral side in section 21, its ventral side extend- ing therefore 7 sections beyond its dorsal side. The mesenterial filaments of the first and second pairs of mesen- teries, the only ones present, extend nearly the whole length of the larva, almost but not quite touching the mesoglea of the aboral end; their position is as follows: those of the first pair of mesenteries on the right reach from section 22 to section 41, and on the left from section 23 to section 44; those of the second pair of mesenteries on the right reach from section 22 to 41 and on the left from section 23 DEVELOPMENT OF AGARICIA. 501 to 41. The filaments of the same side are not fused, as in some previ- ous larvae, except that on the left side they are united at their origin, (section 23). The enlargement of the filaments of the first pair of mesenteries reaches from section 29 to 38; the sections of the second pair are of almost uniform size throughout their length. The mesenteries extend practically the entire length of the larva; their positions are shown in the following table: Number of mesentery / II III IV V VI Right side of larva Oral end section No. 8 7 9 7 17 17 Aboral end section No. 52 49 54 50 53 51 Left side of larva Oral end section No. 10 9 10 7 17 17 Aboral end section No. .56 54 56 51 56 55 Mesenteries I to IV are the only complete ones. Group IV. Larva 1. This larva is slightly larger than any of the previous ones, and shows an advance in the development of the primary mesenteries and of the mesenterial filaments. It was fixed and stained like larva G, and then cut into 133 transverse sections, each 5 fj. thick, the aboral end being cut first. Figure 4 (A-F) shows the 15th, 19th, 23rd, 25th, 73rd and 93rd sections. Photomicrographs of sections 19 and 26, are shown on Plate 5 and of 73 and 93 on Plate 6. The oesophagus is slightly more invaginated than in larva B. It is oval in cross section (Fig. 4, A) from the oral end to section 17. In section 17 it is U-shaped, the ectoderm being interrupted on the dorsal side. This is also the case in section 19 (Fig. 4, B). Beyond this section the ectoderm is interrupted also on the ventral side (Fig. 4, C). On the ventral side of the oesophagus in sections 16 to 19 a diverti- culum of the gastrovascular cavity extends forward in the region of the third pair of mesenteries (Fig. 4, B). The mesenterial filaments of the first and second mesenteries (Fig. 4, D) are fused on the right (left as seen in the figures) as far as section 29, on the left as far as section 38. The mesenterial filaments of the first pair of mesenteries extend on both sides into section 112. The transverse sections of these fila- ments are smaller than those of the filaments of the second pair of 502 MAVOR. mesenteries as far as section 80 on the right and on the left up to sec- tion 83. They are largest (Fig. 4, F) on the right between sections 87 and 108 and on the left between sections 87 and 1 10, where they are larger than any of the transverse sections of the filaments of the second pair of mesenteries. The mesenterial filaments of the second pair of mesenteries extend on the right into section 89 and on the left (Fig. 4, F) into section 95. The transverse sections of these filaments (Fig. 4, E) are largest on the right between sections 54 and 87 — where © ® Figure 4. Agaricia fragilis. Larval. Transverse sections of larva, the oral end having been cut first. Dorsal being up, the right of the larva is to the right in the figures. Ectoderm and mesoglea colored black. A, B, C, D, E, F, sections number 15, 19, 23, 2.5, 73, 93 respectively. In C, I, II, III, IV, V, VI mark the first, second, third, fourth, fifth, and sixth pairs of mes- enteries. X 90. up to section 80 they are larger than the filaments of the first pair of mesenteries in the same sections, — and on the left between sections 58 and 90, where they are larger than the filaments of the first pair of mesenteries as far as section 83. The filament of mesentery /// on the right side of the larva (left in the preparation) is not continuous with the ectoderm of the oesophagus, being absent in sections 20 to 26. On the left side of the larva the filament of the corresponding mesen- tery if present at all is reduced to a few scattered cells in these sections. Both filaments are enlarged beyond these sections but are always much DEVELOPMENT OF AGARICIA. 503 smaller than the filaments of mesenteries I and II. The filament on the left of the larva (right in the sections) extends to about section 73 (Fig. 4, E), that on the right of the larva to about section 77, the limits not being sharply defined. The filaments of mesenteries /// differ histologically from those of I and //. The ciliated cells are shorter and thicker and gland cells seem to be absent, resembling in this the endoderm cells, from which the filament is not always sharply to be distinguished. A table is given showing the extent of the six primary mesenteries in this larva. Number of mesentery I 11 in IV V VI Right side of larva Oral end section No. 9 12 9 15 13 23 Aboral end section No. 11.5 112 102 104 112 102 Left side of larva Oral end section No. 12 1.5 11 15 21 26 Aboral end section No. 11.5 112 102 104 112 102 Only mesenteries I to IV are complete. This table shows that all of the mesenteries extend practically the whole length of the larva. The third and fourth mesenteries are, however, considering their early appearance in development, some- what shorter than the others. The mesoglea of the mesenteries is very thin and in many cases can be seen only where it is cut obliquely. That it is somewhat tough and rigid, at least in fixed and stained preparations, is shown by its being slightly displaced in sections, leaving the less resistant entoderm cells behind it. The gastrovascular cavity is clearly in the process of being divided into compartments corresponding to the mesenteries. It will be noticed in Figure 4, E and F, that a cleft in the entoderm occurs on either side of each mesentery and that entoderm bulges inward be- tween the mesenteries and between these clefts. This would seem to indicate that the formation of the compartments is brought about by the formation of the mesenteries. There is evidence in the histol- ogy of the entoderm that the clefts referred to are formed by the separation and breaking down of cells rather than by a process of infolding. This would tend to support the theory that the clefts are due to the closer adherence of the contiguous entoderm cells to the more rigid mesoglea of the mesenteries. 504 MAVOR. c. Conclusions as to the Course of Development of the Mesenteries, Mesenterial Filaments and the Gastrovascular Cavity. If the assumption be made that the mesenteries develop in the order of their size, the order of their development in the larva of Agaricia fragilis is indicated by the numbers given the mesenteries (Fig. 2, E), with the exception, that numbers V and VI develop simultaneously. So far as the first four pairs of mesenteries are con- cerned the order of development is that found by Faurot ('95) in Adamsia palliata and HalcamjM chrysanthellum; by Wilson ('88) in Manicina arcolata; by McMurrich ('91) in Rhodactis sancti-thomae and Aulactinia and by Duerden (:04) in Siderastrea radians. The writer believes that the larvae studied show that in the develop- ment of the six pairs of primary mesenteries there may be recognized three periods as follows: first, a period in which there are two pairs (/ and //), a condition not represented by any of the larvae, but shown probably to exist by the great development of pairs / and // and the only slight indication of pairs /// and IV in larva A; second, a period in which there are four pairs, represented in an early stage by larva A, and shown probably to exist by the large size of pairs III and IV and the small size of pairs V and VI in larvae B and C; third, a period in which there are six pairs of mesenteries represented by larvae B to I. By dividing the development of the mesenteries into these periods the writer does not wish to deny the appearance of the pairs of mesenteries in succession, but merely to show their association into three sets of two pairs each. Further, it is to be noticed that this association in sets of two pairs becomes more intimate as develop- ment proceeds. In larva A pair I shows a very considerable advance over pair // and this difference persists to some extent in the older larvae. In the same larva pair III shows only a slight advance over pair IV and in the older larvae such a difference between pairs /// and IV is hardly to be seen. In larvae B to I pairs T" and VI seem to have appeared and developed simultaneously. If this interpretation is correct, the ventro-dorsal order of develop- ment which is evident in the first two pairs, becomes less marked in the next two pairs (/// and IV) and has entirely disappeared in the last two pairs (F and VI). In other words the bilateral symmetry- of the larval mesenteries has begun to give way to a radial symmetry while it is still free swimming. Evidence of close association in development between the first two pairs of mesenteries is seen also in the development of their DEVELOPMENT OF AGARICIA. 505 mesenterial filaments, which are already well developed when the third and fourth pairs of mesenteries have only begun to show (larva A). The mesenterial filaments of the third pair of mesenteries show only after the six pairs of primary mesenteries have reached a compara- tively advanced stage of development (Larva I). The bilateral symmetry of the planula is shown in the position, as well as the order of development, of the six pairs of primary mesen- teries. This is clearly shown in the position of pairs / and //. The two mesenteries of pair / lie almost in the same straight line in cross sections, while the mesenteries of pair // if produced would meet at an angle of about 45 degrees (Fig. 1, D; Fig. 2, /; Fig. 4, F). Further, if the transverse sections be examined, it will be noted that the length of the circumference on the ventral side between the peripheral ends of the mesenteries of pair / is greater than the length of the cii'cumference between the peripheral ends of pair //. Both these facts may be expressed by saying that the angle between the mesenteries of pair / on the ventral side is greater than the angle between the mesenteries of pair II on the dorsal side. The transverse sections show that in the older larvae the oesophagus is elliptical in cross section and that the long axis of the ellipse is dorso- ventral, a condition found in Manicina areolata by Wilson ('88). The oesophagus is continued further aborally on the ventral than on the dorsal side. The mesenterial filaments of mesenteries / and // are continuous with the ectoderm of the oesophagus and are formed by an aboral growth of this ectoderm. These filaments show the same histological structure as the ectoderm of the outer surface, lacking however nettle cells. The condition found in larva I seems to show that the mesen- terial filaments of mesenteries /// are developed from the endoderm without any connection with the ectoderm. In Manicina areolata Wilson ('88) has shown that the filaments of the first three pairs of mesenteries are developed from the oeso- phageal ectoderm. In Aulactinia McMurrich ('91) found that there was no reflection of the ectoderm as described in Manicina by Wilson ('88), and that the median streak (which appears before the lateral streaks) of the filaments of the first three mesenteries was developed from the endoderm. My preparations seem to show that the gastro vascular cavity in Agaricia is developed by a breaking down and splitting of the endo- derm and that the mesenteries, muscle cells, and the cells which will form or have formed the mesenterial filaments are the agents which determine its form. 506 MAYOR. PART II. ON THE POSTLARVAL DEVELOPMENT. 1. Form of the Young Polyp. At the time of the flattening out of the young polyp soon after fixation there is no evidence of tentacles. The six pairs of primary mesenteries show clearly as grooves on the surface and when viewed by transmitted light as dark lines. The oral aperture, as in the planula, is oval. The young polyps were not studied in sections as it was desired to preserve the skeletons. 2. The Early Development of the Skeleton. The study of the skeleton has been confined to the skeletons of the larvae which fixed themselves to the glass vessels in which they were kept. In spite of numerous efforts to rear young polyps and to obtain early stages in the development of the skeleton, only a few skeletons suitable for studying the development of the primary ento-septa were obtained. Two of these have been chosen for description. a. Description of Skeleton A. This skeleton (Plate 5, top figure) shows a thin calcareous layer, the basal plate, covering the area of the glass to which the polyp has attached itself. This layer is slightly thickened in the centre, appar- ently an unusual condition; for a columella is not developed in the older skeletons obtained. The six primary entosepta are arranged radially, extending from near the periphery about half way toward the center. The two entosepta on the left and the upper one on the right show forking at their peripheral ends. The basal plate is thickened near the entosepta in areas corresponding to the primary entocoels. A thin outer ring which extended round the basal plate was largely destroyed in the maceration, a small part of it is shown in the lower part of the photograph. It surrounded the outer limit of the soft tissues of the polyp. The opacity of the young polyp made it im- possible to study the relations of the skeleton to the soft parts while the animal was still alive. So an outline drawing was made of the DEVELOPMENT OF AGAKICIA. 507 living polyp showing the positions of the mouth and mesenteries. When the polyp was killed and the soft parts macerated it was found that this drawing could be fitted with certainty over a drawing of the skeleton (Fig. 5), owing to the exact correspondence between the periphery of the basal plate and the outline of the soft parts of the polyp. The drawings show that the primary entosepta lie between the primary mesenteries and, when compared with the photograph (Plate 5, top figure), that the thickenings of the basal plate re- ferred to above coincide exactly with the areas enclosed in the primary entocoels. It will be noticed in both photograph and drawing that the primary Figure 5. Agaricia fragilis. Outline drawing of a young polyp, showing the six pairs of primary mesenteries and the opening of the mouth, superposed on a drawing of the skeleton of the same polyp (septa in solid black) made separately after the polyp had died and disintegrated, vi, m, median septa, I, l, lateral septa. septa show bilateral symmetry in their arrangement and form. Un- fortunately it was not possible in the living polyp to distinguish a dorsal from a ventral side. That the axis of this symmetry is, how- ever, dorso-ventral is shown by the oral aperture, which has its long axis in this direction. The axis of symmetry is occupied by two en- tosepta, which will be called median (Fig. 5, m) ; the upper (in the figure) lateral septa (/) make equal angles with the median plane. This is also true of the lower lateral septa, but the angle which the latter make is less than that made by the upper septa. All four of these septa are concave toward the upper side of the figure. 508 MAYOR. Both photograph and drawing show two pairs of very small exo- septa. Owing to ignorance as to which of the median septa is dorsal, it is not possible to say whether the upper pair is the dorso-lateral or ventro-lateral pair. The photographs and drawing have been ori- ented, however, to correspond with the arrangement found by Duer- den (:04) in Siderastrea radians, where the dorsal septa are the first to develop. b. Description of Skeleton B. In this skeleton a distinct epitheca surrounds the basal plate. The entosepta are relatively larger than those of skeleton A. There is the same bilateral symmetry with regard to the angles which the lateral septa make with the median plane. The lateral septa, however, in this case are slightly concave on the lower side. Exosepta and columella are absent. c. Conclusions on the Early Development of the Skeleton. The early skeletons obtained show great variation in size and in the development of the septa. In some cases only five septa are developed, in others one or more septa are defective. This variation may be due to the fixation of the larvae at an earlier period in their development than is usual; the skeletons showing most variation are the smaller ones. The following conclusions may however be drawn. The basal plate and the six primary entosepta are the first structures to be developed. The primary exosepta do not arise simultaneously. Bi- lateral symmetry is frequently shown in the arrangement of the primary entosepta. Four possible explanations of this bilateral symmetry occur to the writer. (1) The polyps which formed these skeletons may have fixed themselves with the dorsal or possibly ventral side bent over toward the substratum. (2) One or more of the median entocoels may have been enlarged by the growth of the wall of the polyp in that region. (3) It may represent a persistence of the bilateral symmetry seen in the development of the mesenteries. (4) It may represent the tendency of the coral to grow upward at one point to form the frond- like corallum of the older coral. DEVELOPMENT OF AGARICIA. 509 d. Stage with Twelve Primary Septa. A number of skeletons in this stage were obtained (Plate 5, bottom photograph). They show six primary entosepta and six exosepta. The epitheca covers the outer portion and is in the form of the base of a cone. It is decorated with ridges running toward the oral end of the polyp. The theca is seen joining the entosepta and exosepta. The primary entosepta extend beyond the theca but do not reach the epitheca. In a few of the skeletons a slight indication of the bilateral symmetry described for the earlier skeletons could still be recognized. 3. Comparison of the Development of the Skeleton in Agaricia fragilis with that in other Hexacorallidae. In Astroides calycularis Lacaze-Duthiers ('73) found that each of the six primary entosepta was formed from three centres of calcifica- tion, the septa being in consequence Y-shaped with the upper part of the Y toward the periphery. The two septa on the left in Plate 5 (top figure) are forked at their peripheral ends, although the prongs of the fork are not so large as in A. calycularis. Lacaze-Duthiers found that the other septa appear irregularly. The same species has been studied by Koch ('82), who also finds the primary septa forked. Both these authors find that the primary entosepta are developed before the theca or epitheca. In Caryophyllia cyatlnis and C. claws Lacaze-Duthiers ('97) found similar stages in the development. In the development of Sidcrastrea radians, studied by Duerden (:04), six primary entosepta are developed as single continuous rods without connection with the theca or epitheca. An outer ring, the beginning of the epitheca, similar to the ring described in skeleton A, surrounds the basal plate. The primary exosepta are developed in dorso-ventral succession and independently of the entosepta. University of Wisconsin, April, 1915. 510 MAYOR. Bibliography. Duerden, J. E. 1904. The Coral Siderastrea radians and its Postlarval Develop- ment. Carnegie Institution, Washington. Publn. No. 20. v+ 130 p., 11 pi. Paurot, L. 1895. Etudes sur I'Anatomie, I'Histologie et le Developpement des Actinies. Arch, de Zool. Exp. et Gen., Ser. 3, Tom. 3, p. 43-262, pi. 1-12. Koch, G. von. 1882. Ueber die Entwicklung des Kalkskeletes von Asteroides calycularis und dessen morphologischer Bedeutung. Mitt. Zool. Sta. Neapel, Bd. Ill, p. 284-292, Taf. 22-24. 1897. Entwicklung von Caryophyllia cyathus. Mitt. Zool. Sta. Neapel, Bd. XII, p. 75.5-772, Taf. 34. Lacaze-Duthiers, H. de. 1873. Developpement des Coralliaires. Deuxieme Memoire. Actiniaires a Polypiers. Arch, de Zool. Exp. et Gen., Tom. II, p. 269-348, pi. 12-14. 1897. Faune du Golfe du Lion. Coralliaires. Zoanthaires Sclerodermes. Arch, de Zool. Exp. et Gen., Ser. 3, Tom. 5, p. 1-249, pi. 1-12. McMurrich, J. 1891. Contributions on the Morphology of the Actinozoa. II. On the Development of the Hexactiniae. Journ. Morph., Vol. 4, p. 303-330, pi. 13. Wilson, H. V. 1888. On the Development of Manicina areolata. Journ. Morph., Vol. 2, p. 191-252, pi. 15-20. DEVELOPMENT OF AGARICIA. 511 Description of Plates. Plates 1-6. Development of Agaricia fragilis. All, except top and bot- tom figures of Plate 5, are photomicrographs of transverse sections of free swimming larvae. X 175. Plate 1. Larva A, upper photograph, section 15; lower, section 24. Plate 2. Larva A, upper photograph, section 30; lower, section 46. Plate 3. Larva F, upper photograph, section 12; lower, section 17. Plate 4. Larva F, upper photograph, section 61 ; lower, section 73. Plate 5. Two middle figures are of Larva I, the left (C) being section 19; the right (D), section 26. The top figure (A) is a photograph of skeleton A, the bottom one (B) of skeleton C. Plate 6. Larva I, upper photograph, section 73; lower, section 93. Mavor.— Development of Agaricia Plate 1 J. W. M. Photo. Prog. Amer. Acad. Arts and Sciences. Vol 51 HFliOTVPE CO., BnsTON. Mavor— Development of Agaricia J. W. M. Photo. (Proc. Amer. Acad. Arts and Sciences. Vol 51 HELIOTYPE CO.. BOSTON. Mavor.— Development of Agaricia, Plate 3 J. W. M. Photo. Proc. Amer. Acad. Arts and Sciences. Vol. 51. Mavor.— Development of Agaricia Plate 4 W. M.. Photo Proc. Amer. Acad. Arts and Sciences. Vol 51 HELIOTYPE CO.. I BOSTON Mavor.— Development of Agaricia. Plate 5 ' '•••W f ■**^c' ^ f \ .'.>-'• '*'- n J. W. M. Photo. Proc. Amer. Acad. Arts and Sciences. Vol. 51, HfLIOTYPE CO. BOSTON. Mavor. -Development of Agaricia Plate 6 J. W. M. Photo. Proc. Amer. Acad. Arts and Sciences. Vol 51 HtLIOT'.r'E CO., BOSH Proceedings of the American Academy of Arts and Sciences. Vol. 51. No. 10.— January, 1916. CONTRIBUTIONS FROM THE GRAY HERBARIUM OF HARVARD UNIVERSITY. New Series. — No. XLV. I. Compositae new and transferred, chiefly Mexican. By S. F. Blake. II. New, reclassified, or otherwise noteworthy Spermatophytes. By B L. Robinson. III. Certain Borraginaceae new or transferred. By J. Francis Macbride. CONTRIBUTIONS FROM THE GRAY HERBARIUM OF HARVARD UNIVERSITY.— NEW SERIES, NO. XLV. Presented by B. L. Robinson, June 17, 1915. Received June 17, 1915. I. COMPOSITAE NEW AND TRANSFERRED, CHIEFLY MEXICAN. By S. F. Blake. Sericocarpus bifoliatus (Walt.) Porter var. Collinsii (Nutt.) Blake, n. comb. — Aster Collinsii Nutt.! Journ. Acad. Nat. Sci. Philad. vii. 82 (1834). Sericocarpus Collinsii Nutt. Trans. Am. Philos. Soc. ser. 2. vii. 302 (1841). Sericocarpus tortifolius (Michx.) Nees var. Collinsii (Nutt.) Torr. & Gray, Fl. N. Am. ii. 103 (1841).— Nuttall's type in the British Museum, collected by Ware in eastern Florida, has obovate leaves, 3-5-toothed above the middle, but is indistinguish- able in any other character from the ordinary form of the species. So noteworthy a variation, although not recognized by Dr. Gray in the Synoptical Flora, seems to merit varietal rank. Sericocarpus rigidus Lindl. var. calif ornicus (E. Dur.) Blake, n. comb. — Sericocarpus calif ornicus E. Durandl Journ. Acad. Nat. Sci. Philad. (PI. Pratt.) ser. 2. iii. 90 (1855).— All the Californian specimens of Sericocarpus rigidus examined, including the type of S. californicus in the herbarium of the Museum d'Histoire Naturelle at Paris, collected near Nevada City by Rattan, differ from the nearly glabrous northern forms of the species in their more or less densely short-hispid-pilose stems, and appear to constitute a well defined geographical variety. Gymnolomia obscura Blake, n. sp. Annua tenuis erecta sursum parce ramosa 1.9-3.5 dm. alta. Caulis leviter striatus hispidulus et strigosus pallide brunneus. Folia inferiora (3-7-juga) opposita superiora alterna oblongo-ovata vel oblongo-lanceolata vel lanceolata acuta vel subacuminata basi cuneata crenato-dentata (dentibus ca. 7-9-jugis appressis rotimdatis) vel superiora integra supra viridia subdense tuberculato-strigosa subtus distincte pallidiora subsparse strigoso-hispida (pilis venas secundum longioribus) et sparse glandu- loso-adspersa triplinervia 2.7-4.7 cm. longa 9-13 mm. lata, petiolis vix marginatis paten ti-hispido-pilosis 3-5 mm. longis; ea inflore- 516 BLAKE. scentiae multo reducta bracteiformia. Capitula 6-9 parva (2 cm. lata) in pedunculis axillaribus et terminalibus 2.3-8.5 cm. longis monocephalis nudis vel bracteatis interdum etiam (vix normaliter) ex axillis inferioribus orientibus insidentia. Discus 6.5 mm. altus 7 mm. crassus maturitate valde convexus. Involucri 2-seriati vix gradati 3.5-5 mm. alti squamae herbaceae praecipue superne hispido- pilosae lanceolatae vel ovatae (turn supra mediam partem abrupte angustatae acuminatae) acutae mucronulatae rarius obtusae. Radii ca. 6 flavi neutrales oblongi emarginulati dorso minute strigillosi 7 mm. longi 3.5 mm. lati. CoroUae disci flavae 1.6-1.9 mm. longae (tubo glanduloso basi non ampliato 0.5 mm. longo). Paleae receptaculi valde convexi subhyalino-scariosae in carina sursum et apice virides- centes supra sparse breviter pilosae et ciliatae apice abrupte acutae mucronulatae 3.3 mm. longae. Achaenia oblongo-obovoidea parum incrassata glabra striata albido-nigro-maculosa 2 mm. longa 1 mm. lata. Pappus nullus. — Vera Cruz: Maltrata, Jan. 1883, Kerber 211 (type coll.: Brit. Mus., Kew). Vernacular name "Acaguale." — Distributed as a Scleropus (i. e. Sclerocarpus) under an unpub- lished name attributed to Schultz Bipontinus, which as originally used by Schultz can have had no application to the specimens. The species comes near G. longifolia and G. annua Rob. & Greenm., but is distinct in its short involucre. Gymnolomia hypochlora Blake, n. sp. Herbacea alta erecta supra ramosa. Caulis tenuis purpurascens striatulus hispidulo-strigillosus. Folia (saltern media et superiora) alterna lanceolata vel oblongo- lanceolata acuminata basi cuneata obscure serrulata (dentibus ca. 11-jugis appressis) ca. 1-1.5 cm. supra basin valde trinervia reticulata supra viridia tuberculato-strigosa aetate lepidota subtus non palli- diora venas et venulas secundum hispidula inter venas plus minusve glandulari-granulosa 8-13.5 cm. longa 1.7-2.7 cm. lata. Petioli hispido-piloso-ciliati 3 mm. longi. Capitula ca. 11 apice caulis laxe cymoso-paniculata 3.5 cm. lata, pedunculis striatis sparse tuberculato- strigillosis nudis vel 1-2-bracteolatis 5-13 cm. longis. Discus 1- (fructu) 1.4 cm. altus 1.4-(fructu) 1.9 cm. crassus. Involucri 5-seriati gradati discum superantis squamae serierum 3 exteriorum herbaceae lanceolatae vel oblongo-lanceolatae acuminatae plus minusve involutae nigrescenti-virides hispido-piloso-ciliatae et intus strigillosae basi induratae apice valde reflexae, eae serierum interiorum oblongae obtusae vel obtusiusculae apice submembranaceo maturitate elon- gate discum superantes ciliolatae infra paullum induratae plus minusve strigillosae. Radii 12 neutrales flavi oblongi 1.5 cm. longi COMPOSITAE NEW AND TRANSFERRED. 517 4 mm. lati. CoroUae disci flavae in basi tubi ampliata etiam basi faucium puberulae 5.5 mm. longae (tubo 1.2 mm.). Paleae supra purpurascentes et strigillosae apice abrupte acuminatae 6.8 mm. longae. Achaenia glabra substriata oblonga 3 mm. longa 1 mm. lata. Pappus nullus. — Jalisco: mountains above Etzatlan, 2 Oct. 1903, Pringle 11537 (type coll.: Kew). — Distributed as G. Ghiesbreghtii Hemsl., but that has mainly opposite leaves whitened beneath with a dense strigillosity. Haplocalymma Blake, n. genus Helianth.-Verhesin., Hymenostepldo Benth. proximum, a quo involucro uniseriato 5-phyllo et foliis alternis imprimis differt. — • Capitula heterogama radiata, floribus radii neutris disci hermaphroditis fertilibus. Involucri subcampanulati bracteae 5 uniseriatae oblongo-lanceolatae acuminatae adpressae apice paul- lum patentes subherbaceae intus subinduratae 3-5-nerviae dense strigosae flores radii subtendentes. Receptaculum parvulum con- vexiusculum, paleis complicatis flores disci amplectentibus onus- tum. Corollae radii 5 ligulatae flavae parvae o vales ca. 6-nerviae emarginatae; disci ca. 16 regulares flavae puberulae tubo brevissimo faucibus cylindricis apice 5-fidae. Antherae basi subauriculatae apice appendice deltoideo-ovata flava munitae. Styli rami obtusiusculi hirti non appendiculati. Achaenia radii abortiva; disci parva ob- longa a latere subcompressa subsparse villosa. Pappus e squamellis 6 hyalinis latis profunde laceratis duabus in angulis achaenii paullo longioribus compositus. — Herba ramosa tenuis strigillosa, capitula numerosa parvula in paniculis cymosis subdensis 3-5-cephalis ramos dichotome ramosos terminantibus gerens. Folia (saltem media et superiora) alterna ovata grosse sinuato-dentata dentibus acutissime mucronatis. — Species unica Viguiera microcephala Greenm. Haplocalymma microcephalum (Greenm.) Blake, n. comb. — Viguiera microcephala Greenm.! Proc. Am. Acad, xxxix. 105 (1903). — Morelos: limestone hills near Yautepec, 7 Nov. 1902, Pringle 8717 (type coll.: Brit. Mus., Kew); near Cuernavaca, 1905, Lemmon 85 (Kew). The genus Haplocalymma (aTrXovs, simple, and Kokvixixa, covering, from the uniseriate involucre) differs from its near ally Hymeno- stephium Benth. of southern Mexico, Central America and Colombia in its strictly uniseriate 5-leaved involucre (of 2-3 series in Hymeno- stephium) and its alternate leaves, but in pappus characters agrees perfectly with Hymenostephium angustifolium Benth. From Viguiera, to which its relationship is rather less close, it may be distinguished by the character of its pappus and involucre. The group of Viguiera 518 BLAKE. most nearly approaching it, including V. tenuis Gray and V. gracillima Brandegee, has the phyllaries 2-sub-3-senate and distinctly more numerous, while the pappus is typical of that of Viguiera, consisting of two slender awns sharply differentiated from the four to six shorter squamellae. Viguiera adenophylla Blake, n. sp. Herbacea supra raniosa ramis erectis. Caulis purpureo-brunneus plus minusve glanduloso-puberu- lus sursum appresse pilosulus subflexuosus. Folia alterna late ovata acuminata basi cuneata vel abrupte truncata in petiolum cuneate angustata crenato-dentata (dentibus 17-21-jugis depresso-deltoideis mucronatis) apice acuminate integro trinervia supra subappresse pilosula (pilis basi glandularibus vix tuberculatis) infra non pallidiora glandulosa pilosula 8.5-10.5 cm. longa 4.6-7 cm. lata. Petioli glandu- lari-pubescentes 1.8-2 cm. longi. Capitula ca. 26 cymoso-paniculata 2.8 cm. lata. Discus 8-13 mm. altus 7-10 mm. crassus. Involucri 6 mm. alti biseriati subgradati squamae lanceolatae vel ovato-lanceo- latae acutae subherbaceae basi plus minusve induratae costatae in margine et plus minusve dorsaliter sordide albido-pilosiusculae et glandulosae. Radii 8 flavi ovales emarginati in venis dorsi puberuli 10-12 mm. longi 5.5-6 mm. lati. Corollae disci flavae subappresse pilosulae 6 mm. longae (tubo 1.3 mm.) faucibus campanulato-infundi- buliformibus. Paleae oblongae mucronatae acutae margine flavidae apice eroso-denticulatae dorso sordide pilosiusculae parce glandulosae 8-9 mm. longae. Achaenia (immatura) sericea. Aristae 2 paleaceae oblongo-lanceolatae acutae vel obtusae fimbriatae inaequales 2.8- 3.5 mm. longae; squamellae 4 longiores liberae oblongae lacerato- fimbriatae ca. 1.2 mm. longae et ca. 4 parvae intermediae laceratae. — San Luis Potosi: alt. 1830-2440 m., 1878, Parry & Palmer 467 (type coll.: Brit. Mus., Gray Herb., Kew). — Not closely related to any other species. Viguiera angustifolia (H. & A.) Blake, n. comb. — Tithonia pachycephala H. & A. ! Bot. Beech. Voy. 299 (1840), not DC. (err. iden.). T. angustifolia H. & A. ! 1. c. 435 (1841). Viguiera blepharo- lepis Gray! Proc. Am. Acad. xix. 5 (1883). — The type of Tithonia angustifolia H. & A. (Tepic, Sinclair, hb. Kew.) is identical with Seemann 1481, from Cerro de Pinal, type collection of V. blepharo- Icpis, and not at all the same as V. bucldleiacformis (DC.) B. & H. f., although considered by Bentham (B. & H. f. Gen. ii. 375 (1873)) to be identical with the latter species. The only previous Viguiera angustifolia, that of Glaziou, Mem. Soc. Bot. France, iii. 412 (1910), is merely a nomen nudum. COMPOSITAE NEW AND TRANSFERRED. 519 ViGuiERA bicolor Blake, n. sp. Frutex ramosus verosim. ad 1 m. altus. Ramuli juniores canescenti-strigillosi seniores cano-brunnei subglabrati. Folia subopposita vel superiora alterna rotundata vel rotundato-ovata apice retusa vel rotundata vel obtusa basi truncato- rotundata trinervia supra (siccitate) nigro-viridia aspere strigillosa siibtus densissime griseo-strigillosa venosa 1.2-1.8 cm. longa 8-18 mm. lata margine plus minusve revoluta. Petioli canescenti-strigillosi 5-7 mm. longi. Capitula solitaria ramulos terminantia 1.8 cm. lata. Pedunculi canescenti-strigillosi 2.8-4.3 cm. longi. Discus 6-7 mm. altus 11-12.5 mm. crassus. Involucri 3-seriati 3-4 mm. alti gradati squamae acutae densissime strigillosae exteriores oblongae 0.8 mm. latae interiores oblongo-ovatae 2 mm. latae. Radii ca. 12 flavi oblongo-ovales in dorso et tubo dense puberuli 6.5 mm. longi 3 mm. lati. Corollae disci flavae puberulae 3 mm. longae (tubo 0.5 mm.). Paleae subobtusae dorso et apice puberulae 5 mm. longae. Achaenia (immaturissima) subsericea 2.5 mm. longa. Aristae 2 inaequales ad 1.2 mm. longae; squamellae ca. 6-8 acutae inaequales laciniatae multo breviores. — Hidalgo (?) : between Rio Grande and Jamaltepec (?), Dec. 1829, C. Ehrenberg 1227 (types in the Berlin Herb, and the Gray Herb.). — From its nearest ally, V. brevifolia Greenm., the present species differs chiefly in foliar characters. The leaves of V. brevifolia are ovate to triangular-ovate, usually acute but sometimes obtuse, and usually strongly canescent above, on petioles 2-3 mm. long; the awns of the pappus about 1 mm. long, equalling the squa- mellae. Ehrenberg's plant in the Berlin Herbarium was marked as a new species under an unpublished generic name equivalent to Gym- nolomia by Schultz Bipontinus, who could scarcely have given it more than a very casual inspection to refer it to this genus the salient character of which is the absence of pappus. Viguiera Brandegei Blake, n. nom. — Asyilia hispida Brandegee! Univ. Calif. Pub. Bot. iv. 94 (1910), not Viguiera hispida Baker in Mart. Fl. Bras. vi. pt. 3. 220 (1884).— The pappus of this species, of about 4 oblong irregularly lacerate squamellae with 1 or 2 longer similar aristae, is that of Viguiera rather than Aspilia. Among Mexican species V. Brandegei is most closely related to V. tenuis Gray and V. gracillima Brandegee. The species is renamed in honor of its original describer, whose publications have for many years been adding largely to our knowledge of the flora of Mexico and Lower California. Helianthus leptocaulis (Wats.) Blake, n. comb. — Viguiera leptocaulis Wats. ! Proc. Am. Acad. xxvi. 140 (1891). — Watson's 520 BLAKE. Viguiera leptocaulis, the type of which {Pringle 2247, from near Mon- terey) was examined at the Gray Herbarium by the writer some months ago, in its involucre, wholly deciduous pappus, and general habit agrees much better with the genus to which it is here referred than with that in which it was originally placed. It seems quite distinct from any described Mexican Helianthus. Phoebanthus Blake, n. genus Helianih.-Verhesin. — Capitula heterogama radiata, floribus radii 1-seriatis neutralibus disci herma- phroditis fertilibus. Involucrum hemisphaericum, bracteis 3-seriatis vix gradatis eis duarum serierum exteriorum lineari-lanceolatis acumin- atis infra valde induratis costatis supra appendice herbacea laxa multo longiore praeditis sparse tuberculati-hispidis eis seriei intimae latiori- bus brevioribus solummodo acutis apice obscurissime subherbaceis appressis flores radii subtendentibus. Receptaculum convexum, paleis complicatis flores disci amplectentibus. Corollae radii ligulatae patentes anguste oblongae apice 2-3-denticulatae; disci regulares tubulosae, tubo brevi faucibus elongatis cylindraceis apice breviter 5-fidae. Antherae basi valde sagittatae apice appendice ovali vel oblongo-ovata obtusa munitae, filamentis vel in tubum clausum connatis vel liberis. Styli florum disci rami lineares elongati hirti appendice lineari-oblonga hirta muniti. Achaenia radii trigona abortiva; disci subglabra valde incrassata subquadrangularia (attamen a latere plus minusve compressa) in angulis submarginata. Pappus ex aristis 2 interdum basi profunde laceratis tenuibus vel subpaleaceis persistentibus interdum ad dentes reductis et squamellis numerosis minutis intermediis compositus. — Herbae perennes e radice tuberi- formi horizontali submoniliformi. Caulis tenuis simplex 1-3-cephalus foliosus 3^-pedalis. Folia linearia alterna (paucis inferioribus oppositis exceptis) Integra 1-nervia paullum revoluta plus minusve tuberculato-strigosa. Capitula magna flava. Flores radii 10-20 poUicares et ultra; disci numerosi.^ — Species typica Hclianthdhi grandijiora Torr. & Gray. — Helianthus L. subgen. Pseudo-helianthus Gray, Syn. Fl. i. pt. 2. 285 (1884). 1. Phoebanthus grandiflorus (Torr. & Gray) Blake, n. comb. — Helianthella grandiflora Torr. & Gray, Fl. N. Am. ii. 333 (1842). 2. Phoebanthus tenuifolius (Torr. & Gray) Blake, n. comb. — Helianthella tenuifolia Torr. & Gray, Fl. N. Am. ii. 333 (1842). The two species of Phoebanthus ($oT/3os, Phoebus, the sun, and apOos, floiver) form a small group very similar to Helianthus in habital characters, but technically more closely allied to Helianthella. From the former Phoebanthus differs sufficiently in the numerous short COMPOSITAE NEW AND TRANSFERRED. 521 persistent squamellae united into a low denticulate crown, from the latter in habit and achenial characters, as also in range, the species of Helianthella being West American and Mexican, of Phochanthus Floridan. The species of Helianthella, which after the exclusion of Enceliopsis and Pseudo-hcliantlms of the Synoptical Flora are restricted to the subgenus " HcUantheUa proper" of that work, are perennial herbs with rather few lanceolate to oblong-lanceolate usualh^ ample leaves, alternate or opposite, of which the basal are generally much larger, phyllaries for the most part herbaceous or often enlarged and foliaceous, and achenes very strongly flattened (as in Encelia), on the margin villous or glabrate, with 2 longer or shorter awns and a corona of thin squamellae of varying length, united below or distinct to the base. . Although the genus Helianthella on its first publication (Torr. & Gray, 1. c.) was divided into two groups, of which the first (§1, without name) contained only the two species here referred to Phocbanthns, while the second had three species {H. Dotiglasii, H. lanceolata, and H. uniflora) all still retained in Helianthella, I have not thought it advisable to consider the H. grandi flora group as typical, on the ground of priority of position, and to create a new name for the second group, l)oth because a considerably greater number of name changes would be involved and because the H. Douglasii group was considered by Gray himself (see Proc. Am. Acad. xix. 9 (1883) ; Syn. Fl. i. pt. 2. 284 (1884)) as typical of the genus. Pionocarpus Blake, n. genus Helianth.-Verhesin. — Capitula hetero- gama radiata, floribus radii 1-seriatis neutralibus disci hermaphro- ditis fertilibus. Involucrum hemisphaericum, bracteis 2-seriatis (exterioribus paullulo brevioribus) anguste ovato-lanceolatis acumi- natis herbaceis subcostatis plus minusve piloso-hispido-ciliatis. Re- ceptaculum convexum, paleis complicatis acuminatis flores disci am- plectentibus onustum. Corollae radii ligulatae patentes oblongae; disci regulares, tubo brevissimo faucibus cylindraceis limbo 5-dentato. Antherae basi cordato-sagittatae apice appendice ovata obtusa prae- ditae. Styli rami lineares hirti appendice lineari-lanceolata longa hirtella praediti. Achaenia radii inania; disci oblonga pubescentia apice truncata valde incrassata paullum a latere compressa immargi- nata. Pappus persistens ex aristis 2 subaequalibus subtenuibus et squamellis ca. 10 multo minoribus laceratis basi ima unitis interdum una duabusve elongatis compositus. — Herba perennis subscaposa e radice verticali crassa tuberiformi. Caulis parcissime ramosus ca. 4-cephalus. Folia maxima ex parte radicalia lineari-lanceolata 522 BLAKE. utrinique acuminata longe petiolata subintegra cavilina 3-4 (infra mediam partem caulis). Capitula majuscula longe pedunculata, pedunculis basi unibracteatis. Flores radii ca. 10 flavi, disci numerosi flavi. — Species unica Hclianfhella madrensis Wats. PiONOCARPus madrensis (Wats.) Blake, n. comb. — Helianthella madrensis Wats.! Proc. Am. Acad, xxiii. 278 (1888). — Chihuahua: pine plains, base of the Sierra Madre, 19 Sept. 1887, Pringle 1302 (type coll.: Kew). The genus Pionocarpus (inchu, fat, and Kapirbs, fruit) is distinguished from Helianthella chiefly by characters of the achene and style. The achenes of Pionocarjnis are strongly thickened, unmargined, and broadly truncate at apex, and the style-branches bear a long hirtellous appendage. In Helianthella, on the contrary, the achenes are very flatly compressed, as in Encelia, are more or less distinctly white- margined, and at the true apex are much narrower than the body, the more or less united squamellae being borne in a notch at the sum- mit of the achene between the two awns. From Phoebanthus, to which its relationship is rather closer, Pionocarpus is separated by its habit and by the much more conspicuous squamellae. Perymenium blepharolepis Blake, n. sp. Frutex ramosus. Cau- lis densissime strigillosus cortice cano-brunneo tectus aetate glabratus; rami tenues subquadrangulares strigillosi fusco-brunnei. Folia oppo- sita ovata subacuminata basi cuneata vel cuneato-rotundata trinervia appresse serrata (dentibus 7-8-jugis) supra viridia strigillosa pilis basi tuberculatis subtus pallidiora non canescentia strigosa et strigil- losa plus minusve glanduloso-adspersa 2.5-3.8 cm. longa 10-17.5 mm. lata. Petioli tuberculato-strigosi 2 mm. longi. Capitula 1.5-2 cm. lata in apicibus ramorum 5-7 sublaxe cymoso-paniculata in pedun- culis axillaribus et terminalibus strigillosis 1.2-4.5 cm. longis. Discus 7-9.5 mm. altus 6-8 mm. crassus. Involucri 3-sub-4-seriati gra- dati 7 mm. alti squamae exteriores ovatae interiores ovato-lanceo- latae plus minusve strigillosae et ciliatae infra subinduratae pallidae ca. 5-nerviae apice laxo angustato subacuto herbaceo. Radii 8 fer- tiles aurei oblongo-ovales bidenticulati in venis dorso plus minusve puberuli 7-5 mm. longi 3 mm. lati. CoroUae disci aui'eae in dentibus hirtellae 4.8-5 mm. longae (tubo 1.4-1.5 mm.). Paleae obtusae su- pra strigillosae in margine et carina spinuloso-ciliolatae 5.8-7 mm. longae. Achaenia juventate plus minusve piloso-hispida maturita- tem vergentia sparse hispidula. Pappi aristae ca. 12 caducae ad 1.3 mm. longae subaequales (2 longioribus 3 mm. longis exceptis). — Puebla: Coxcatlan, alt. 2135-2440 m., Sept. 1909, Pnrpus 4143 COMPOSITAE NEW AND TRANSFEKRED. 523 (type coll.: British Museum). — Most nearly related among de- scribed species to P. verhesinoides DC, but distinct in its longer looser phyllaries, fewer longer-peduncled heads, and in its distinctly paler leaves, which are much less reticulate beneath. Perymenium hypoleucum Blake, n. sp. Frutex ramosus. Cau- lis strigillosus cortice griseo-brunneo tectus aetate subglabratus ; rami sub-6-angulares subdense strigillosi purpureo-brunnei. Folia oppo- sita (paucis superioribus interdum alternis exceptis) ovata vel elliptico- ovata obtusa vel acuta basi cuneato-rotundata serrata (dentibus 5-7-jugis late triangularibus subpatentibus vel appressis) pinnati- nervia (venis supra impressis subtus reticulatis majoribus 3-4-jugis imis validioribus) supra viridia dense strigosa pilis basi tuberculatis subtus cum pilis densissimis appressis brevibus canescentia et sub- aspera 2-3 cm. longa 1.1-1.9 cm. lata. Petioli strigosi immarginati 1.5-2 mm. longi. Capitula 1 cm. lata ad apices ramorum 4-10 dense cymoso-paniculata in pedunculis dense strigosis 3-8 mm. longis disposita. Discus 6-8 mm. altus 5-7 mm. crassus. Involucri 4- seriati gradati 4 mm. alti squamae solummodo apice herbaceae infra induratae pallidae dense strigillosae obtusae vel rotundatae extimae elliptico-ovatae interiores deltoideo-ovatae vel late ovales intimae 3-nerviae. Radii fertiles ad 7 ovales aurei 2-3-denticulati in venis dorso hirtelli 3.5 mm. longi 2 mm. lati. Corollae disci flavae in dentibus et interdum tubo hirtellae 3.5 mm. longae (tubo 1 mm.). Faleae in carina infra glandulosae supra subspinulosae in margine denticulato-spinulosae apice abrupte acutae 3.7 mm. longae. Achaenia (immaturissima) in angulis hispidula. Pappi aristae ca. 8 caducae spinulosae subaequales (una duabusve duplo longioribus exceptis). — Puebla: vicinity of San Luis Tultitlanapa, near Oaxaca, 9 July 1908, Pur pus 3087 (type coll.: British Museum). — Distributed as P. rude Rob. & Greenm., from which it is very distinct. More closely related to P. croceum Rob. & Greenm. and P. angustifolium Brandegee, but quite distinct from both. Perymenium leptopodum Blake, n. sp. Frutex ramosus. Caulis tenuis quadrangularis minute strigillosus cortice brunneo vel pur- pureo-brunneo tectus. Folia opposita ovato-lanceolata acuminata basi cuneata trinervia appresse serrata (dentibus 8-10-jugis) supra obscure viridia minute strigilloso-scabra subtus non pallidiora glandu- loso-adspersa venas et venulas secundum strigillosa 4.6-6.3 cm. longa 1.5-2.3 cm. lata. Petioli vix marginati tenues 9-16 mm. longi. Capitula 1.6 cm. lata in apicibus ramorum et ramulorum ternatim vel quinatim cymoso-paniculata. Pedicelli 5-19 mm. longi strigillosi. 524 BLAKE. Discus 5-7 mm. altus 4-6 mm. crassus. Involucri 3-seriati 4 mm. alti paullum gradati squamae extimae breviores late ovatae obtusae strigillosae vix ciliolatae basi coriaceo-induratae apice herbaceae mediae et intimae late oblongae subobtusae minus induratae ciliola- tae et strigillosae apice subherbaceae. Radii 8 fertiles flavi ovales bidenticulati dorso glanduloso-puberuli 4.5-5.8 mm. longi 2.8 mm. lati. CoroUae disci flavae in tubo puberulae 3-4.2 mm. longae (tubo 0.6-1.6 mm.), faucibus valde ampliatis. Paleae subacutae scariosae flavae in carina hispidulae margins superne lacerato-dentatae 4.2-4.6 mm. longae. Achaenia brunneo-nigrescentia transverse plus minus ve rugulosa incrassata submarginata apice rotundata sparse strigillosa 1.8 mm. longa 1 mm. lata. Pappus fragilissimus ex aristis 2 tenuibus spinulosis 2.5 mm. longis et aristulis ca. 12 inaequalibus (longiori- bus ad 1 mm. longis) compositus. — Guatemala: shrubby growth, alt. 1311 m., near Coban, Jan. 1879, von Ttierckheim. 339 (type coll.: British Museum, Kew). — This number is referred in Robinson & Greenman's revision (Proc. Am. Acad, xxxiv. 526 (1899)) to P. gymno- lomioides (Less.) DC, originally described from Vera Cruz, but differs in several important features from the description there given of that species, which is said to have entire leaves, 3-5-headed corymbs, spreading-pubescent pedicels, and much shorter petioles (3 mm. long). Chrysactinia acerosa Blake, n. sp. Frutex ramosus pedalis. Caulis tenuis brunneo-griseus minute hirtellus aetate glabratus oppo- site ramosus. Folia opposita vel superiora alterna filiformi-subulata 2-4-glanduloso-punctata semiteretia supra complanata apice subulato- mucronata glabra laete viridia 6-8 mm. longa 0.25 mm. lata saepius ramulos brevissimos foliosissimos subtendentia. Pedunculi ramos terminantes monocephali glandulari-hirtelli lineari-subulato-bracteo- lati striati ad 1 cm. longi. Capitula (immatura) hemisphaerica 5-6 mm. alta. Involucri uniseriati 3.5 mm. alti squamae aequales 8 anguste oblongae (0.8-1.3 mm. latae) membranaceo-chartaceae viri- descentes obtusae subcarinatae in margine scarioso sparse eroso- ciliatae infra apicem glandula anguste oblonga brunnea praeditae. Radii (immaturi) aurei oblongi fertiles. Corollae disci (immaturae) aureae in dentibus minute hirtellae 4 mm. longae. Achaenia (imma- turissima) pubescentia. Aristae pappi ca. 24 tenues spinulosae subaequales ad 3.5 mm. longae. — San Luis Potosi: Sierra de Guas- cama, Minas de San Rafael, June 1911, Purpus 5136 (type coll.: British Museum, Gray Herb., U. S. Nat. Herb.). — Distributed as a narrow-leaved form of C. mexicana Gray, but seemingly quite distinct from that species in foliar and involucral characters. COMPOSITAE NEW AND TRANSFERRED. 525 The four species of Chrysactinia now known may be conveniently divided into two sections, habitally well marked but without differ- ential characters of technical importance. I. Chrysactinia Gray (PI. Fendl. 93 (1849)) sect. Euchrysactinia Blake, n. sect. Folia subulata integerrima. Corollae radii discique aureae. — Type C. mezicana Gray. 1. C. MExiCANA Gray, 1. c. (1849). — Leaves subulate, flattened above, hirtellous on the margin, mostly alternate, 4.5-12 mm. long, 0.7-1.2 mm. wide, with rather numerous glands. Phyllaries appar- ently uniformly 12 in number, 4-5 mm. long. — • Pedis taxifolia Greene! Leafl. i. 148 (1905): see Greenm. Field Columb. Mus. Bot. ii. 274 (1907). — Western Texas and New Mexico to Puebla. 2. C. ACEROSA Blake. — Leaves filiform-subulate, glabrous, mostly opposite, bearing 2-4 glands, 6-8 mm. long, 0.25 mm. wide. Phyl- laries 8 in number, 3.5 mm. long. — San Luis Potosi. IL Chrysactinia sect. Phylloloba Blake, n. sect. Folia 3-17- pinnatilobata. Corollae radii albidae et aurantiaco-suffultae vel au- reae; corollae disci aurantiacae vel aureae. — Type C. pinnata Wats. 3. C. PINNATA Wats. ! Proc. Am. Acad. xxv. 154 (1890). — Leaves opposite, oblong, pinnatilobed almost to the midrib with 9-17 oppo- site oblique oblong to (uppermost) deltoid mucronate acute lobes, sparsely gland-dotted, 2.5^.9 cm. long, 1.2-1.9 cm. wide, the lowest pair of lobes reduced and stipule-like. Rays 8, whitish, orange-tinged outside; disk orange. — Nuevo Leon: Pringle 2524 (type). — Described by Watson as herbaceous, but apparently frutescent like the other species of the genus. 4. C. truncata Wats. ! I. c. (1890). — Leaves opposite or the upper alternate, ovate to ovate-oblong in outline, pinnatilobed nearly to the midrib with 3-7 mostly alternate lobes, the lobes entire or with 1-few spinulose teeth, truncate and glandular-mucronate at apex (or the terminal sometimes acute), 1.5-2.4 cm. long, 8-12 mm. wide. — Nuevo Leon: Pringle 2601 (type). — The rays are described by Watson as bright yellow. Coreopsis basalis (Dietr.) Blake, n. comb. — Calliopsis hasalis Dietr. in Otto & Dietr. Allgem. Gartenzeit. iii. 329 (17 Oct. 1835). Calliopsis Drummondii D. Don ! in Sweet, Brit. Fl. Gard. ser. 2. iv. t. 315 (1838). Coreopsis Drummondii (D. Don) Torr. & Gray, Fl. N. Am. ii. 345 (1842). — The long and detailed description by Otto & Dietrich of their Calliopsis basalis shows it to be identical with the 526 BLAKE. plant soon afterward described and figured by David Don as Calliopsis Drummondii, which has generally passed in American literature as Coreopsis Drummondii. The seed from which the plant was grown at the Berlin Garden was stated by Otto & Dietrich to have come from Missouri, but as it was included in a large lot of seeds received at second hand from an unknown collector there was obvious chance of error in regard to its origin, and there can be no doubt that it really came from Texas, to which the species is apparently confined. The probable identity of Calliopsis basalis and C. Drummondii was first brought to my attention through a manuscript note on a sheet in the British Museum by Mr. S. LeM. Moore, to whom I am indebted for permis- sion to publish this note. The form with linear or linear-oblong leaf-lobes described by Gray (Syn. Fl. i. pt. 2. 291 (1884)) as Coreopsis Drummondii var. Wrightii becomes C. basalis (Dietr.) Blake var. Wrightii (Gray) Blake, n. comb. 11. NEW, RECLASSIFIED, OR OTHERWISE NOTEWORTHY SPERMATOPHYTES. By B. L. Robinson. Cleome aculeata L. Syst, ed. 12, iii. 232 (1768). To the synonymy of this species should be added C. sinaloensis Brandegee, Zoe, v. 198 (1905). Authentic material of the latter, now in the Gray Herbarium, appears to agree in all features with South American specimens of C. aculeata L. The species has also been found elsewhere in Mexico, in Guatemala (Beam, no. 349), and in Honduras {Thieme, no. 5135 of J. D. Smith's distrib.). It thus seems to be an annual weed of rather wide distribution in tropical and subtropical America. The species was originally described by Linnaeus from American material and has been generally regarded as exclusively of this continent, but it is diffi- cult to find a single character on which to separate from it the West African C. ciliata Schum. & Thonn. Besk. Guin. PI. ii. 68 (1828) and C. guineensis Hook. f. in Hook. Nig. Fl. 218 (1849). C. Fischeri, nom. nov. C. serrulata Pax in Engl. Bot. Jahrb. xiv. 293 (1892), not Pursh, Fl. Bor. Am. ii. 441 (1814). Pursh's homonym, rejected by Dr. Gray and by many other botanists owing to its in- appropriate and undescriptive character, is now being re-established according to the International Rules of Botanical Nomenclature. Consequently the later homonym of Pax must be renamed. The next three new combinations, found necessary during recent work of Mr. George Safford Torrey, temporary assistant at the Gray Herbarium, are here published at Mr. Torrey's request. HosACKiA AMERICANA (Nutt.) Piper, var. glabra (Nutt.) G. S. Torrey, comb. nov. H. elata, j3. glabra Nutt. in Torr. & Gray, Fl. N. Am. i. 327 (1838). PiRiQUETA CAROLiNiANA (Walt.) Urb., var. viridis (Small) G. S. Torrey, comb. nov. P. viridis Small, Fl. S. E. U. S. 794, 1335 (1903). Distinguished from varieties of the variable P. caroliniana merely by the degree of pubescence, and possessing no distinctive habitat or geographical range, this form seems hardly worthy of specific rank. Lyonia fruticosa (MicLx.) G. S. Torrey, comb. nov. Andromeda ferruginea /3. fruticosa Michx. Fl. Am. Bor. i. 252 (1803). A. ferruginea Pursh, Fl. Am. Sept. i. 292 (1814), not A. ferruginea Walt. Fl. Car. 528 ROBINSON. 138 (1788). Lyonia ferruginea Nutt. Gen. i. 266 (1818), as to plant, not as to name-bringing synonym. Xolisma fruticosa Nash, Bull. Torr. Bot. Club, xxii. 153 (1895). The Status of Convolvulus africanus. The binomial Convol- vulus africanu^ seems first to have been used by Choisy in DeCandoUe's Prodromus, Lx. 342 and 418 (1845), where it is attributed to "Nick." or "Nich.," said to have been published in "h. St. Dom.", and is referred to the synonymy of Pharbitis cathartica (Poir.) Choisy. In the Index Kewensis, i. 600 (1895), Convolvulus africanus Nickols, Hort. St. Doming. 260, is referred to Ipomoea cathartica. This dis- position of the species is also made by House in his North American Species of the Genus Ipomoea, 205 (1908). Both on nomenclatorial and geographic grounds the name Convolvulus africanus (1776) seemed so strange a synonym for the much later and strictly x\merican Ipo- moea cathartica (1816), that the writer made some search for the rare work in which the name in question was originally published. This proves to be an "Essai sur I'histoire naturelle de I'isle de Saint- Domingue," published at Paris in 1776. The name of the author does not appear on the title page, but toward the end of the volume, on page 374, there is printed a note of "Approbation" by Adanson, regarding the work itself, and in this note the name of the author, P[ere] Nicolson, is mentioned. It may be noted that neither the title of the work nor the spelling of the author's name is as given in the Index Kewensis. On the page cited in the Index Kewensis (260) the only Convolvulus is not C. africanus but C. americanus, a species treated as follows : "LiANE purgative. — Syn. Liane a medicine. Liane k Bauduit, Arepeea, Car. Convolvulus Americanus. — 06s. Ses tiges sont grimpantes, cylindriques, sans vrilles; elles s'entrelacent dans les branches des arbres voisins, s'y accro- chent, & se replient ensuite vers la terre, y prennent racine, & forment de nouvelles plantes. On en tire un sue r^sineux qui se coagule, & dont on se sert pour purger. Un habitant du cul-de-sac nomm6 Bauduit, en fait un syrop purgatif qui porte son nom. Quoiqu'il soit fort en usage parmi les habi- [here begins page 261] tans du pays, il ne laisse pas d'etre dangereux, en ce qu'il occasionne quelquefois des superpurgations. Ses feuilles sont taill^es en coeur, un peu rudes, unies, sans dentelure. — Loc. EUe se trouve sur les mornes dans les lieux humides. — Virt. EUe purge violement. Liane purgative du bord de la mer. — Syn Convolvulus marinus, Catharti- cus, PI. Soldanella, Marcg. — 06s. Sa feuille est arrondie, bien nourrie. — Loc. On ne la trouve que sur les c6tes de la mer. — Virt. EUe est purgative." NOTEWORTHY SPERMATOPHYTES. 529 On page 213 of the same work occurs the following: "Convolvulus. V. Patate. Convolvulus Americanus. V. Jalap, Liane purgative. Convolvulus marinus catharticus. V. Liane purgative du bord de la mer. Convolvulus tinctorius. V. Liseron." While on page 251 is found the following: "Jalap. — Syn. Jalapa, Mirabilis, Convolvulus Americanus, Ray." From the above we see that the references of Choisy in DC. Prod, ix. 342 and 418, of Hook. f. & Jacks. Ind. Kew. i. 600, and of House, North American Species of Ipomoea, 205, are incorrect in several particulars. The binomial is Convolvulus americanus not C. africanus. The name of the author is Nicolson, not Nickols, as given by the Index Kewensis, nor Nich., as given by Choisy (page 418). The name of the work is Essai sur I'histoire naturelle de I'isle de Saint- Domingue, obscurely abbreviated by Choisy (p. 342) to h. St. Dom., and mistakenly rendered by the Index Kewensis as Hort. St. Doming. Finally it remains to ascertain whether this early name really belongs to the species Ipomoea cathartica Poir., to which it has been referred by these authors, and whether if so it should replace this later name. The cross-references in Nicolson's work show an association by that author of his Liane purgative with Convolvulus americanus of Ray and also with the vernacular name Jalap. On referring to Ray's well known Historia Plantarum, iii. 372 (1704), we find three uses of the name Convolvulus Americanus, namely: " 17. Convolvulus Americanus, vulgaris folio, capsulis triquetris numerosis, ex uno puncto longis petiolis propendentibus, semine lanugine ferruginea vil- loso Pluk. Almag. Bot. Hujus species major habetur. Mock-climber Barba- densibus dicta." "19. Convolvulus AmencaMJiS sub Jalapiae nomine receptus Pluk. Phyt. T. 25. F. 1. An Convolv. colubrinus Pisonis Caapeba?" "22. Convolvulus Americanus, subrotundis foliis viticulis spinosis Pluk. Almag. Bot. T. 276. F. 4. An Convolvulus Peruvianas perpetuus seu Holo- liuchi Hort. Fames. Fig. 4? Reliqua vide apud Autorem." Of these three uses of Convolvulus americanus by Ray, it is clear that the first (Ray's no. 17) is Ipomoea polyanthes G. F. W. Mey., while the third (Ray's no. 22), described as having suborbicular leaves and being spiny must have referred to Plukenet's Plate 276, fig. 3 rather than fig. 4 as stated, and was Calonyction aculeatum (L.) House. That it was to the second, namely Ray's no. 19, to which 530 EOBINSON. Nicolson referred is clearly shown by the expression "sub Jalapiae nomine." Of this plant Plukenet gives a figure cited by Ray and doubtless in the opinion of Ray representing his plant. This figure shows a highly conventionalized twiner with alternate deltoid-ovate entire leaves and solitary short-peduncled axillary flowers. The flow- ers show no corolla but merely ^a smallcalyx with relatively short lobes which do not equal the tube. It is impossible to regard this figure as representing /. cathartico, Poir. which is characterized by calyx- lobes of unusual length, greatly exceeding the short tube. The writer after some search has failed to find that Linnaeus expressed any opinion in regard to the identity of this particular figure of Plukenet, and from its general lack of detail it would probably be impossible to place it with any certainty. It may be seen from the data here assembled 1) that Nicolson made no Convolvulus africanus, a name which seems to have arisen from a clerical error of Choisy. 2) that Nicolson in employing the name Convolvulus americanus had no thought of coining a new designation or describing a new species, but was merely applying — in all proba- bility erroneously — the pre-Linnaean Convolvulus Americanus of Ray to a purgative twiner of Santo Domingo. He gives no botanical characterization sufficient to give validity to the name and merely discusses briefly the pharmaceutical properties. Nicolson's type is not known and even if the plant he was treating could be ascertained it is doubtful if the botanical type of Convolvulus americanus would not have to be sought in the plant of Ray, whose name Nicolson was intending to apply, rather than the plant to which he perhaps mis- takenly applied it. As we have seen, the plant of Ray may be traced back to a figure of Plukenet's which cannot be I. cathartica Poir., a well known name that thus relieved of an earlier synonym retains its validity. Ipomoea crassicaulis (Benth.), comb. nov. Batatas ? crassi- caulis Benth. Voy. Sulph. 134 (1844). Ipomoea fistulosa Mart, ex Choisy in DC. Prod. ix. 349 (1845). /. tcxana Coult. Contrib. U. S. Nat. Herb. i. 45 (1890). Operculina ornithopoda (Robinson) House, var. megacarpa (Brandegee), comb. nov. Ipomoea sp. Rose, Contrib. U. S. Nat. Herb. i. 344 (1895). /. megacarpa Brandegee, Zoe, v. 218 (1905). Operculina Roseana House, Bull. Torr. Bot. Club, xxxiii. 500 (1906) ; Bot. Gaz. xliii. 414 (1907). Formae typicae floribus fructu etc. simil- lima differt foliorum segmentis multo latioribus(0.9-3.5 cm. latitudine). The writer agrees with Mr. House that this plant, differing in range NOTEWORTHY SPERMATOPHYTES. 531 and rather strikingly in foliage, should not be merged without distinc- tion in 0. ornithopoda, yet the differences are of slight moment and the divergence between the extremes is very largely bridged by intermedi- ates. Thus the specimen collected at Agiabampo by Palmer (no. 781) has the leaf -segments ranging down to 6 mm. in breadth, being therefore considerably nearer to the form from San Luis Potosi than to Mr. Brandegee's plant from Culiacan, which has leaf-segments no less than 3.5 cm. wide. Accordingly, the varietal rank here seems preferable to the specific. Lantana achyranthifolia Desf. Cat. Pi. Hort. Reg. Par. ed. 3, 392 (1829). To the synonymy of this species may be added Lippia fimbriata Rusby, Mem. Torr. Bot. Club, iv. 244 (1895), and Lantana macropodioides Greenm. Field Mus. Nat. Hist. Bot. Ser. ii. 339 (1912). Stachytarpheta fruticosa (Millsp.), comb. nov. Valerianodes fruticosa Millsp. Field Columb. Mus. Bot. Ser. ii. 178 (1906). Priva aspera HBK. Nov. Gen. et Spec. ii. 278 (1817). While this species is known to me only by descriptions, I fail to find any distinction by which P. orizabae Wats. Proc. Am. Acad, xxiii. 282 (1888) can be separated from it. Rhaphithamnus venustus (Phil.), comb. nov. Citharexylum ve- nustum Phil. Bot. Zeit. xiv. 64G (1856). Rhaphithamnus longiflorus Miers, Trans. Linn. Soc. xxvii. 98 (1870); Reiche, Fl. Chil. v. 306 (1908). R. serratifolius Miers, 1. c. 99. Citharexylon elegans Phil, ex Miers, 1. c. 98 (1870). Vitex Bakeri, nom. nov. V. diversifolia Baker in Thiselton-Dyer, Fl. Trop. Afr. v. 323 (1900), not Kurz, "Andam. Rep. App. A 45; B 14" (1870?); C. B. Clarke in Hook. f. Fl. Brit. Ind. iv. 585 (1885). Vitex viticifolia (DC), comb. nov. V. montevidensis ? viulti- nervis Cham. Linnaea, vii. 374 (1832). Psilogyne viticifolia DC. Rev. Fam. Bign. 16 (1838). V. multinervis (Cham.) Schauer in DC. Prod, xi. 688 (1847), also in Mart. Fl. Bras. Lx. 297 (1851). The earlier specific name is here restored in accordance with Art. 49 of the Inter- national Rules of Botanical Nomenclature. Caryopteris odorata (Hamilton), comb. nov. Volkameria odorata Hamilton ex Roxb. Hort. Beng. 46 (1814). Clerodendron odoratum (Hamilton) D. Don, Prod. Fl. Nepal. 102 (1825). Caryopteris Walli- ckiana Schauer in DC. Prod. xi. 625 (1847); Clarke in Hook. f. Fl. Brit. Ind. iv. 597 (1885). Sphenodesme involucrata (Presl), comb. nov. Congea ungui- culata Wall. Cat. n. 1736 (nomen). C. ferruginea Wall. 1. c. n. 1737 (nomen). C. paniculata Wall. 1. c. n. 1739 (nomen). Symphorema 532 ROBINSON. paniculatum h. Heyne ex Schauer in DC. Prod. xi. 623 (1847), in synon. Vitex involucratus Presl, Bot. Bemerk. 148 (1844). Sphaeno- desma unguiculata (Wall.) Schauer in DC. Prod. xi. 623 (1847). In a recent attempt to verify the identification and labelling of the Verbenaceae in the Gray Herbarium, corrections were noticed which should be made in the current treatment of two species of the Gala- pagos Islands, namely: 1) The plants which have been treated as Avicennia tomentosa Hook. f . Trans. Linn. Soc. xx. 195 (1847) ; Anderss. Om Galap.-oarnes Veg. 201 (1853), also reprint, 82 (1857); Robinson & Greenman, Am. Jour. Sci. ser. 3, 1. 147 (1895); and A. officinalis Robinson, Proc. Am. Acad, xxxviii. 194 (1902), not L., are referable to A. nitida Jacq. 2) The plants of the Galapagos Islands treated as Lippia lanceolata Rose, Contrib. U. S. Nat. Herb. i. 137 (1892), not Michx.; L. nodiflora Robinson & Greenman, Am. Jour. Sci. ser. 3, 1. 147 (1895), not Michx.; and L. canescens Robinson, Proc. Am. Acad, xxxviii. 196 (1902), not HBK., are all to be referred to L. keptans HBK. Nov. Gen. et Spec, ii. 263 (1817), a species fairly well marked as to its firmer veins and more salient teeth of the leaves. Ageratum Houstonianum Mill., var. muticescens, var. nov., sta- tura foliis pubescentia floribus etc. formae typicae simillimura differt squamis pappi flosculorum vel omnium vel plurium valde reductis muticis ca. 0.1-0.2 mm. longis.— Mexico: Wartenberg, near Tanto- yuca, prov. Huasteca, collected in 1858, L. C. Ervendberg, no. 100 (type, in Gray Herb.); without locality, from the herbarium of the late Dr. F. W. Klatt (Gray Herb.) ; cultivated in the Missouri Botani- cal Garden, from 1886 (when collected by Pammel) to 1896 (when a second specimen was prepared by H. C. Irish). The specimens, now in the herbarium of the Missouri Botanical Garden, show by their labelling that the plant has passed under several horticultural names, "Stella Gurney," "Cope's Pet," etc. The seed is said to have come from Haage & Schmidt's establishment. In these specimens the pappus, though for the most part short and muticous, shows some variability on one and the same plant or even in the same head, certain florets, especially the central ones and those of the terminal heads, tending to have awned scales in the manner of the typical form. Eupatorium brachychaetum, spec, nov., herbaceum vel cum cau- dice ramoso 1 dm. longo pauUo lignescenti; caulibus subscaposis 2-2.5 dm. altis gracilibus puberulis purpurascentibus basin versus solum foliosis apice 2-4-capitulatis ; foliis oblanceolati-oblongis tenuibus op- positis 8-11 cm. longis 2-2.5 cm. latis duplice crenato-lobulatis vel NOTEWORTHY SPERMATOPHYTES. 533 -dentatis apice obtusis vel rotundatis basi gradatim cuneatis pinnati- veniis in venis sparsissime pilosiusculis concoloribus, petiolo ca. 1 cm. longo gracili; capitulis oblato-subsphaericis 1 cm. diametro in pedi- cellis nudis vel bracteolatis gracilibus pilosiusculis purpurascentibus adscendentibus valde inaequalibus (1-10 cm. longitudine) gestis; involucri squamis oblongis vel obovato-oblongis subbiseriatis subae- quilongis striatulis glabriuscuHs ciliolatis vix acutis 2.5 mm. longis; flosculis 40-50; coroUis glabris 2.2 mm. longis, tubo proprio cylindrico fauces campanulatas subaequanti, dentibus limbi anguste deltoideis recurvatis; antheris apice appendiculatis ; achaeniis fuscis glabris 1.5 mm. longis; pappi setis ca. 20 inaequalibus pro genere brevibus corolla dimidio brevioribus. — Cuba: Rocky stream bed, Arroyo Cimmaron, alt. 470 m., Trinidad Mountains, Santa Clara, 5 March, 1910, N. L. & E. G. Britton, no. 5085 (type, in herb. N. Y. Bot. Card.; and in Gray Herb.). A slender and attractive species, which should be readily recognized by its exceptionally short pappus. Var. extentum, var. nov., valde caulescens; caulibus 3 dm. vel ultra longitudine gracilibus foliosis; foliis oppositis; internodiis plerisque 1-5 cm. longis; pedunculis 6-8 cm. longis; aliter formae typicae simillimum. — Cuba: Rocky and shady banks of Iguanojo River, Santa Clara, 11 Aug. 1915, Brothers Leon & Clement, no. 5419 (type, in herb. N. Y. Bot. Gard.; phot, in Gray Herb.). Eupatorium bullescens, spec, nov., fruticosum 6-10 dm. altum; caule tereti gracili virgato usque ad apicem folioso striatulo puberulo; foliis sessilibus oppositis ovato-lanceolatis acuminatis serratis firmius- culis basi rotundatis vel saepius brevissime cordatis a basi trinerviis glabris supra bullatis lucidis minute reticulato-venulosis subtus paullulo pallidioribus opacis nervosis puncticulatis 4-6.8 cm. longis 1.3-2.4 cm. latis internodia valde superantibus ; inflorescentia laxe paniculata terminali folioso-bracteata, ramis gracilibus obscure puberulis; patentibus pedicellis 2-8 mm. longis filiformibus; capitulis parvis 6-9-floris; involucri squamis ca. 7 subaequalibus oblongis obtusis dorso obscure puberulis ca. 2.2 mm. longis post fructus delap- sum persistentibus et deflexis, disco parvo levissime convexo glabro; corollis glabris tubulatis sursum leviter ampliatis ca. 2 mm. longis, dentibus limbi 5 deltoideis suberectis; achaeniis gracilibus fuscis minute granulatis ca. 2 mm. longis; pappi setis ca. 20 sordide albidis corollam subaequantibus. — Cuba: on rocks, gorge of the Rio Yamui, Oriente, 7-9 December, 1910, J. A. Shafer, no. 7808 (type, in herb. N. Y. Bot. Gard., photograph and fragments in Gray Herb.); also in pine lands, alt. 420 m., near El Cuero, Oriente, 10-19 March, 1912, 534 ROBINSON. A^ L. Britton & J. F. Cowell, no. 12,763 (herb. N. Y. Bot. Gard.; photograph in Gray Herb.). Eupatorium epaleaceum (Gardn.), comb. nov. Chromolaena epaleacea Gardn. in Hook. Lond. Jour. Bot. vi. 436 (1847). Eupato- rium lupulinvm Bak. in Mart. Fl. Bras. vi. pt. 2, 301 (1876). Eupatorium havanense HBK., var. domingense (DC), comb, nov. E. ageratifolium y f doviingense DC. Prod. v. 173 (1836). Eupatorium kleinioides HBK., var. lasiolepis, var. nov., formae typicae simile differt praecipue involucri squamis dorso conspicue pubescentibus, pihs tenuibus attenuatis curvato-adscendentibus vel irregulariter patentibus. — Tropical Brazil, Burchell, no. 6885 (type, in Gray Herb.). Eupatorium leucocephalum Benth., var. anodontum, var. nov., formae typicae habitu inflorescentia floribus etc. simile differt foliis paullo angustioribus lanceolatis omnino integerrimis. — White-flow- ered shrub 2-3 m. high in clayey soil, alt. 1000 m., at La Victoria, near th^ boundary between Michoacan and Guerrero, Mexico, 23 March, 1899, E. Langlasse, n. 961 (type, in Gray Herb.). Eupatorium Mairetianum DC, var. adenopodum, var. nov., caule apicem versus et ramis et ramulis et pedicellis ubique dense glanduloso-puberulis fuscescentibus nee ut apud formam typicam canescenti-arachnoideis ; aliter formam typicam simillimum. — Guatemala: Cerro Quemado, Quezaltenango, 21 January, 1915, alt. 2440 m.. Prof. E. W. D. Hohvay, no. 98 (type, in Gray Herb.). Eupatorium pulchellutm HBK., var. angustifolium Wats, in herb, according to Pringle, Plantae Mexicanae, 1889, 2nd [unnumbered] page of the printed list, also on the labels. — This marked variety, recognized by Dr. Sereno Watson, seems never to have been charac- terized. It may be described as follows: statura habitu inflorescentia formae typicae simile differt foliis elongato-lanceolatis 8-10 cm. longis 1-1.8 cm. latis adscendentibus vel modice patentibus nee deflexis inconspicue serratis vel subintegris sparse in nervis venisque hispidu- lis utrinque viridibus, supremis angustissimis. — Mexico: slopes of canons near Guadalajara, Jalisco, 4 October, 1889, C. G. Pringle, no. 2315 (type, in Gray Herb.); barranca of Rio Blanco near Guadala- jara, alt. 1375 m., 30 September, 1903, Pringle, no. 11,524 (Gray Herb.). Easily recognized by its long narrow ascending leaves 5-8 times as long as wide, while in the typical form they are mostly deflexed and rarely 3 times as long as broad, being furthermore inclined to be more deeply serrate and tending to be distinctly paler beneath. Eupatorium pycnocephaloides, spec, nov., suffrutescens in ar- NOTEWORTHY SPERMATOPHYTES. 535 bustis alte scandens pubescens ; caulibus teretibus striatis flexuosis pur- purascentibus internodiis ad 1.8 dm. longis; foliis oppositis petiolatis ovato-deltoldeis acuminatis 5-6.5 cm. longis 4-4.8 cm. latis crenato- serratis basi subtruncatis merabranaceis a basi trinerviis utrinque viridibus supra appresse puberulis subtus praecipue in nervis venisque laxe pubescentibus, petiolo 2-3.5 cm. longo; paniculis terminalibus oppositirameis foliaceo-bracteatis, ramis late patentibus; capitulis 15-20-floris in capitibus terminalibus convexis vel subglobosis laxe glomeratis, pedicellis 2-4 mm. longis filiformibus pubescentibus; involucri subcylindrici basi turbinato-campanulati 6-7 mm. alti squamis ca. 3-seriatis valde inaequalibus viridibus plus minusve pur- purascentibus striatis dorso pubescentibus, extimis brevibus ovatis acuminatis, intermediis oblongis acutis, intimis lineari-oblongis apice obtusis; corollis albis vel roseis glabris 5-nerviis (nervis a basi ad sinum inter dentes percurrentibus) anguste tubulatis sine faucibus distinctis 4 mm. longis, limbi dentibus 5 deltoideis 0.3 mm. longis; achaeniis atrobrunneis 1.7 mm. longis in angulis et inter eos sursum hispidulis; pappi setis ca. 25 albis tenuibus quam corolla paullo brevioribus. — - Guatemala: Climbing over shrubs, Solola, alt. 2135 m., 28 January, 1915, Prof. E. W. D. Holway, no. 144 (type, in Gray Herb.); Volcan de Agua, Antigua, alt. 2135 m., 13 January, 1915 (in bud), Holway, no. 83 (Gray Herb.). A species somewhat resembling E. pycno- cephalum Less., but much larger, the involucral bracts more pubescent, the outer narrower and more acute. Var. glandulipes, var. no v., pedicellis et involucri squamis glandulo- so-puberulis; flosculis roseis; achaeniis subglabris faciebus lucidis costis solis paullo sursum hispidulis.^ Guatemala : Totonicapam, alt. 2440 m., 24 January, 1915, Prof. E. W. D. Holway, no. 106 (type, in Gray Herb.). Eupatorium rhexioides, spec, nov., laxe caespitosum glabriuscu- lum stragulum 1 m. diametro formans; caulibus elongatis apicem versus ramosissimis flexuosis angulato-striatis; internodiis inferioribus elongatis ad 1.5 dm. longitudine superioribus multo brevioribus vix 1 cm. longis; ramis valde flexuosis plerisque alternis in inflorescentias racemosas terminantibus ; foliis parvis deflexis ovato-lanceolatis vel lanceolato-ellipticis, inferioribus oppositis, ramealibus plerisque alternis ca. 2 cm. longis 6-9 mm. latis crassiusculis utrinque glabris subtus paullo pallidioribus 3-nerviis puncticulatis integris vel utroque latere obtuse 2-3-dentatis acutiusculis margine plus minusve revolutis basi cuneatis, petiolo 1-3 mm. longo; racemis plus minusve compositis bracteatis; bracteis inferioribus foliis similibus superioribus gradatim 536 ROBINSON. reductis linearibus pedicellos (2-7 mm. longos) obscure puberulos longitudine paullo superantibus; capitulis parvis ca. 12-flons; in- volucri vix imbricati squamis 7-10 linearibus acutis minute puberulis crassiusculis maturitate stellatim patentibus ad 5 mm. longis; flos- culis 6.8 mm. longis; corolla tubiformi granulata sursum paullo am- pliata 3.6 mm. longa, limbi dentibus deltoideis brevibus patentibus; antheris apice cum appendice ovata membranacea munitis; achaeniis gracilibus deorsum attenuatis ca. 3.2 mm. longis 5-angulatis costulis secundariis 1-2 inter costas primarias hinc inde interjectis; pappi setis sordidis ca. 30 vix barbellatis tenuibus corollam aequantibus. — Cuba: moist banks, on the trail from Camp La Barga (alt. 450 m.) to Camp San Benito (alt. 900 m.), Oriente, 22-26 February, 1910, J. A. Shafer, no. 4105 (type, in herb. N. Y. Bot. Gard.; photograph and fragment in Gray Herb.). EuPATORiUM ScHULTZii Schuittspahu, Zeitschr. d. Gartenbauver- eins z. Darmst. 1857, p. 6. This species appears to have four dis- tinguishable forms, which may be briefly characterized as follows: Forma typicum, pedicellis dense cum glandulis capitatis numero- sissimis puberulis, pilis paucis non capitatis hinc inde interspersis ; involucri squamis albidis, intermediis elliptico-oblongis apice ro- tundatis glabris vel subglabris non ciliatis. — Represented in the Gray Herbarium as follows : material cultivated in the Darmstadt Botanical Garden and authoritively labeled by Schultz himself; also Vera Cruz: Mirador, Liehmann, no. 41, Sartorius (without number); Zacuapan, Purpus, no. 2372; Chiapas: C. & E. Selcr, no. 2219; Costa Rica: Ojo de Agua, Hoffmann, no. 394, Rodeo de Pacaca, Pittier, no. 3285, slopes of Rodeo, at 1100 m. altitude, Pittier, no. 1600; Guatemala: Cubelquitz, Depart. Alta Verapaz, alt. 350 m., von Tuerckheim, no. 7888. Forma erythranthodium, forma nova, formae typicae simile sed differt conspicuiter involucri squamis pulchriter purpureis. — Guate- mala: Coban, altitude 1350 m., von Tuerckheim, no. II. 2090 (type, in Gray Herb.). Forma velutipes, forma nova, formae typicae simile differt pedi- cellis velutino-tomentellis vix vel solum obsolescente glandulosis. — Guatemala: San Lucas Tollman, Solola, altitude 1560 m., 2 February, 1915, Prof. E. W. D. Holway, no. 170 (type, in Gray Herb.); Depart. Guatemala, alt. 1525 m., J. D. Smith, no. 2373; on the volcano of Tecuamburro, Depart. Santa Rosa, alt. 2135 m., Ileyde & Lux, no. 4515 (distrib. of J. D. Smith). Var. ophryolepis, var. nov., pedicellis dense cum glandulis capitatis NOTEWORTHY SPERMATOPHYTES. 537 puberulis; involucre paullo quam eo formae typicae imbricatiori, squamis purpureo-viridibus striatis, exterioribus intermediisque plus minusve acutatis, intermediis ciliatis. — Guatemala: in woods, Volcan de Atitlan, Solola, alt. about 2135 m., Prof. E. W. D. Holway, no. 187 (type, in Gray Herb.). Eupatorium Shaferi, spec, no v., herbaceum vel vix basi paullo molliter lignescens glabrum 3-4 dm. altum; caule basi paullo decum- bens fistuloso purpureo-brunneo ad mediam partem folioso; foliis oppositis ovato- vel lanceolato-oblongis utroque acuminatis serratis utrinque glaberrimis subtus paullo pallidioribus 4-8 cm. longis 2-3.5 cm. latis pinnativeniis; petiolo 1-1.5 cm. longo supra canaliculato; panicula parva 4 cm. solum diametro sub-6-capitulata vel ampliori 3-chotoma 1 dm. diametro multicapitulata; bracteis oblanceolato- linearibus, bracteolis lineari-subulatis; pedicellis ca. 1 cm. longis adscendentibus obscurissime puberulis; capitulis 1 cm. diametro 15-20-floris; involucri 2-3-seriati squamis subaequalibus laxe imbri- catis oblongis obtusis vel apice rotundatis ciliolatis quam flosculis dimidio brevioribus post fructus delapsum persistentibus et reflexis, disco maturitate leviter convexo glabro; corollis glabris tubulatis sursum modice gradatimque ampliatis 3 mm. longis purpureis; achae- niis 5-angulatis deorsum decrescentibus basi substipitatis glabris aetate fuscis pallide costatis 2.8 mm. longis; pappi setis simplici- bus tenuibus sordide albis achaenium longitudine subaequantibus. — Cuba: on rocks by water near top of falls, Arro,yo del Medio, Oriente, alt. 450-550 m., 20 December, 1909, J. A. Shafer, no. 3227a (type, in herb. N. Y. Bot. Gard., photograph and fragments in Gray Herb.); also on rocks by water, same locality, date, and collector, no. 3227 (herb. N. Y. Bot. Gard.; photograph and fragments in Gray Herb.). Eupatorium urticaefolium Reichard, var. angustatum (Gray), comb. nov. E. ageratoides L.f ., var. angustatum Gray, Syn. Fl. i. pt. 2, 101 (1884). E. angmtatum (Gray) Greene, Pitt. iv. 277 (1901). Kyrstenia angustata (Gray) Greene, Leaf!, i. 8 (1903). Eupatorium vitifolium (Sch. Bip.) Klatt, Leopoldina, xx. 90 (1884). Hebecliniuvi vitifolium Sch. Bip. ex Klatt, 1. c. To the synonymy of this rare and little known Mexican species may be added Bulhostylis triangularis DC. Prod. vii. 268 (1838) and Carphephorus triangularis (DC.) Gray, PI. Wright i. 86 (1852) ex Hemsl. Biol. Cent.- Am. Bot. ii. 109 (1881). The type-specimens, both of DeCandoUe's and of Klatt's species, are exceedingly fragmentary, yet material is sufficient to show close correspondence in general contour, dentation, size, and venation of the leaf, open divaricately branched panicle. 538 ROBINSON, largish heads with narrow unequal acute involucral scales passing inward into narrow somewhat readily deciduous paleae. The achenes are 4-5-angled as described by DeCandoUe and the species is there- fore referable to Eupatorium rather than to Brickellia. With Car- phephorvs, notwithstanding the presence of a few paleae on the outer part of the disk, the species appears to have no close affinity. The older specific name {triangularis) having been already several times employed for other species of Eupatorium, cannot be here revived. The only difference noted between the type of E. vitifolium and that of Bulbostylis triangularis is that in the single leaf of the latter, still preserved in the DeCandoUean herbarium, the base is essentially truncate, while in the Liebmann plant, upon which Klatt's species was founded, the leaf-base, though in the main outline truncate, is slightly cordate at the insertion of the petiole. In consideration of the close correspondence in all other features observed, this one slight difference is believed to be merely individual. At all events, until the plants can be examined in more complete specimens even a varietal separation seems unwise. Brickellia cymulifera, spec, nov., subherbacea 3-4 dm. alta: caulibus gracilibus teretibus foliosis plus minusve decumbentibus crispe puberulis, internodiis 3-8 cm. longis; foliis oppositis graciliter petiolatis hastato-deltoideis caudato-acuminatis grosse dentatis basi cordatis utrinque viridibus sparse crispeque puberulis, laminis 3-5.4 cm. longis 2.5-3.8 cm. latis, petiolo 1-4 cm. longo; cymis oppositis laxe patentibus 3-5-capitulatis in axillis a parte media caulis (vel etiam a parte inferiori) usque ad apicem orientibus, pedunculis sub- filiformibus crispe puberulis 2-3.5 cm. longis in media parte bracteas binas foliaceas ovatas acuminatas integras gerentibus et ad apicem V)racteolis binis parvis linearibus munitis; pedicellis filiformibus crispe puberulis saepe bracteolatis 4-15 mm. longis; capitulis (valde imma- turis) ca. 12-floris 6 mm. altis et diametro; involucri turbinato- campanulati squamis ca. 17 convexis viridibus vel atro-purpureis tenuiter striatis ciliolatis apice rotundatis, extimis dorso puberulis orbiculatis, intermediis ovato vel obovato-oblongis dorso subglabris, intimis oblanceolato-oblongis; corollis (immaturis) cylindricis sparse granulatis 5-dentatis; achaeniis (immaturis) obconicis tomentellis; pappi setis 40-50 valde inaequalibus barbellatis albis. — Mexico: Minas de San Rafael, San Luis Potosi, November, 1910, Dr. C. A. Purpus, no. 4802 (type, in Gray Herb.,'' the same number also ex- amined in herb. Mo. Bot. Gard.). CoNYZA soPHiAEFOLiA HBK, Nov. Gen. et Spec. iv. 72, t. 326 NOTEWORTHY SPERMATOPHYTES. 539 (1820). To the synonymy of this species may be added Erigeron canadensis as cited in J. D. Sm. Enum PI. Guat. iv. 78 (1895), as to no. 6152. Verbesina Holwayi, spec, nov., ut videturherbaceaerectarobusta; caiile tereti medulloso glabro purpurascenti-brunneo exalato levissime striatulo usqvie ad inflorescentiam laxe folioso; foliis alternis ovato- deltoideis utroque latere 1-2-lobatis supra rugosis cum venis impressis reticulatis cum pilis brevissimis albidis scaberrimis subtus paullo pallidioribus viridibus molliter tomentellis ad 1.8 dm. longis 1.4 dm. latis, petiolo ad 9 cm. longo late alato (2 cm. latitudine) basi auricu- lato-amplexicauli, lobis folii inaequalibus ovato-oblongis plus minusve falcatis 3-5 cm. longis 3 cm. latis apice rotundatis margine irregulari- ter serrato-dentatis, lobo terminali multo majori subacuminato, costa media subtus prominenti, venis pinnatis in petiolo transversis in lamina prorsus curvatis; foliis superioribus gradatim reductis lanceo- latis, bracteis ultimis spatulatis vel linearibus minimis; panicula ampla 3 dm. diametro plana multicapitulata sordide puberula vel tomentella, ramis et pedicellis (8-14 mm. longis) adscendentibus; capitulis (fructigeris) contiguis ovoideis ca. 8 mm. diametro 1 cm. altis ca. 25-floris ut videtur discoideis; involucri squamis extimis linearibus inaequalibus dorso sordide pubescentibus acutiusculis, intimis spatu- lato-oblongis carinatis mucronulatis in media parte subherbaceis pubescentibus margine utroque tenuioribus subchartaceis ; paleis similibus; disco conico; achaeniis nigris 3.2 mm. longis bialatis in faciebus pallide tuberculatis, alis albidis ciliolatis; aristis 2 erectis 1.5 mm. longis tenuibus barbellatis. — Guatemala: Quezaltenango, alt. 2290 m., January, 1915, Prof. E. W. D. Holway, no. 96 (type, in Gray Herb.). This species is probably referable to § Lipactinia Robinson & Greenman, Proc. Am. Acad. xxiv. 563 (1899), though in the fruiting material at hand it is difficult to be sure of the absence of rays. It would seem to be most naturally placed near V. auriculata DC. Prod. V. 617 (1836) with which it agrees in many points, though it is readily distinguished by its different inflorescence and lobed leaves. Liabum sublobatum, spec, nov., caule tereti (exsiccatione) striato- ruguloso puberulo brunnescenti ad inflorescentiam laxe patenteque folioso; foliis oppositis late ovato-rhomboideis acuminatis hastato- subtrilobatis mucronulato-dentatis (dentibus inaequalibus) basi inte- gris cuneatis in petiolum acuminate decurrentibus supra laete viridi- bus glabris subtus arachnoideo-tomentosis 1-1.3 dm. longis 7-9 cm. latis paullo supra basin 3-nervatis; petiolo ca. 5 cm. longo supra canal- iculato puberulo subtus rotunduto aetate suberoso-ruguloso ; panicula 540 EOBINSON. laxe fastigiata 1.5 dm. diametro convexa vel subpyramidali multi- capitulata, ramis ramulisque adscendentibus fusco-puberulis, bracteis inferioribus foliaceis, bracteolis multo reductis lanceolato-subulatis plus minusve connatis sordide tomentosis ca. 3-4 mm. longis; pedi- cellis 4-5 mm. longis gracilibus rectis; capituHs discoideis 1.3 cm. altis ca. 9 mm. diametro ca. 8-floris; involucri turbinato-campanulati 4-seriati squamis oblongis apice rotundatis dorso striatulis purpureo- viridibus vix puberulis margine obscure ciliatis extimis brevibus sub- oibiculari-ovatis; corollis laete flavis 8 mm. longis glabris, tubo proprio gracili in fauces gradatim ampliato, dentibus limbi 5 lanceolatis recurvatis; achaeniis glabris crassiusculis; pappi setis albidis obscure barbellatis ca. 6 mm. longis (paucis extimis ter brevioribus).^ — Guate- mala: San Lucas Toliman, Solola, 2 February, 1915, alt. 1665 m., Prof. E. W. D. Holway, no. 179 (type, in Gray Herb.). A species near L. glabrum Hemsl., but readily distinguished by its somewhat lobed leaves and puberulent inflorescence. III. CERTAIN BORRAGINACEAE, NEW OR TRANSFERRED. By J. Francis Macbride. In the course of ordering up portions of the Borraginaceae at the Gray Herbarium it has become necessary from time to time to make new names and new combinations of names in order to have the work conform to the International Rules of Botanical Nomenclature. In addition, an attempt to classify the unnamed material in some of the groups has led to the discovery of a few species and varieties appar- ently undescribed. It seems advisable, therefore, to place these matters on record at this time. Tournefortia Miquelii, nom. nov. — T. syringaefolia Miquel, Stirp. Surin. 137 (1850), not T. syringaefolia Vahl, Symb. Bot. iii. 23 (1794), a name which must be revived to replace the more generally used but later synonym T. laurifolia Vent. Choix PI. 2 (1803). Tournefortia Aubletii, nom. nov. — T. glabra Aubl. PI. Guian. i. 118 (1775), not T. glabra L. Sp. PI. 141 (1753), which must replace T. cymosa L. Sp. PI. ed. 2, 202 (1762). Heliotropium fragrans, nom. nov. — H. odorum (Fres.) Giirke, Nat. Pflanzenf. iv. Ab. 3, 96 (1893). Heliophytum odorum Fres. in Mart. Fl. Bras. viii. pt. 1, 45 (1857), not Heliotrojpiwn odorum Balf. f. Proc. Roy. Soc. Edinb. xii. 81 (1884). Article 53 of the Inter- national Rules states that: "When a species is moved from one genus into another, its specific epithet must be changed if it is already borne by a valid species of that genus." Therefore H. odorum Fres. requires a new name on being transferred to Heliotropium because of the presence there of H. odorum Balf. f., a valid species which cannot, according to these rules, be renamed H. Balfouri as has been done by Giirke, 1. c. Heliotropium foliosissimum, spec, nov., multicaule decumbens subgriseo-pubescens; radice et caudice lignescentibus atrobrunneis; caulibus 5-14 flexuosis gracilibus basi ad apicem aequabiliter foliosis- simus 3-12 cm. longis; foliis elliptico-oblongis margine vix revolutis obtusis nunc alternis nunc suboppositis vel irregulariter dispositis 5-10 mm. longis 2-4 mm. latis; racemis bracteatis brevibus; calycis laciniis oblongo-ovatis; corollae tubo calycem non superante; nuci- bus strigosis. — Southern Mexico in the State of Oaxaca: Hacienda Blanca, July 25, 1895, L. C. Smith, no. 627 (type, in Gray Herb.) ; 542 MACBRIDE. sterile hills, Telixtlahuaca, July 27, 1895, L. C. Smith, no. 471; near Oaxaca, July 26, 1896, C. Conzatti, no. 157, in part; gravelly soil near Oaxaca, July 3, 1900, Charles C. Deam, no. 11; Cerro San Antonio, June 26, 1906, C. Conzatti, no. 1411. These specimens were labeled H. limhatum Benth., but that species is a more canescent plant of rigid erect habit, and with narrower longer leaves (10-15 mm. long, 1.5-2 mm. wide) and almost glabrous nutlets. The aspect, too, is very different both from the dissimilar manner of growth and because the stems of H. limhatum are leafiest at the base, where the leaves persist, while in H. foliosissimwn the stems are equably leafy and the lower leaves soon die. Heliotropium jaliscense, spec, nov., suffruticosum erectum, ramis hispidis et adpresse strigillosis; foliis petiolatis ovato-lanceolatis subacuminatis basi attenuatis integerrimis 5-10 cm. longis 2-3 cm. latis utrinque strigillosis et subtus in nervis hispidis; racemis flexuo- sis gracilibus ebracteolatis -pedunculatis ; pedunculis subterrainalibus; calycis lobis hispidis latitudine inaequalibus subacuminatis; corollae tubo calycem ca. 2 mm. superante; corolla 3.5-4 mm. longa; an- therae media in parte tubi insertae; stigmate late conico basi annu- lato stylum vix superante; nuculis 4 glabris forsan maturitate reticulatis. — Mexico: bushy slopes near San Sebastian, Jalisco, March 16-19, 1897, E. W. Nelson, no. 4083 (type, in Gray Herb.). A species bearing a superficial resemblance to H. parviflorum L. but by style and fruit characters a member of the section Euheliotropium. Heliotropium phyllostachyum Torr., var. erectum, var. nov., caulibus erectis 1-4 dm. altis; foliis oblongo-lanceolatis 1-3 cm. longis ca. 3 mm. latis; corollae tubo calycem superante, limbo 3-5 mm. lato. — Mexico: CuHacan, Sinaloa, Oct. 24, 1904, T. S. Brandegee (type, in Gray Herb.); between Guichocovi and Lagunas, Oaxaca, June 27, 1895, E. W. Nelson, no. 2743; Real de Guadelupe, Sept. 14, 1898, E. Langlasse, no. 351; near Cuernavaca, Morelos, July 25, 1896, C. G. Pringle, no. 7183; near Iguala, Guerrero, Sept. 22, 1905, C. G. Pringle, no. 13,681; Yucatan, 1895, G. F. Gaumer, no. 790. H. phyllostachyum Torr. in its typical form is a low (rarely 1 dm. high) diffusely spreading plant with short broad leaves and inconspicuous flowers, the corolla 1.5-2 mm. long, scarcely exceeding the calyx. It is mostly of more northern range than the variety, although it has been secured at Manzanillo, Colima, by Dr. Palmer (no. 891) and at Guaymas (no. 232). No. 891 is quite typical but no. 232 represents a transition to the variety in its erect habit. Because of these facts it seems best to give the southern plant varietal rather than specific rank. CERTAIN BORRAGINACEAE. 543 Omphalodes lateriflora (Aubrey), comb. nov. — Cynoglossum lateriflorum Aubrey, Prog. Morb. x. 25 (1801-1803). 0. littoralis Lehm. Neue Schrift. Nat. Fr. Berl. viii. 98 (1818). Solenanthus turkestanicus (Reg &Srairn.), comb. nov. — Kv^- chakewiczia turkcstanica Regel & Smirn. Act. Hort. Petrop. v. 626 (1877). Solenanthus Kuschakewiczi Lipsky, Act. Hort. Petrop. xxiii. 182 (1904). As Lipsky has well shown, this plant possesses no charac- ters which justify its being maintained as a genus distinct from Sole- nanthus Ledeb. He, however, as indicated above, failed to retain the original specific name. Solenanthus stamineus (Desf.), comb. nov. — Cynoglossum stami- neum Desf. Ann. Mus. Par. x. 431 (1807). Solenanthus Tourncfortii DC. Prod. X. 164 (1846). DeCandoUe rightly gives Cynoglossum stamineum Bieb, Fl. Taur.-Cauc. iii. 127 (1819) a new name under Solenanthus, namely S. Biebersteinii, but there seems to be no need for discarding the much earlier C. stamineum of Desfontaines. Lappula laxa (G. Don), comb. nov. — Cynoglossum laxum G. Don Gen. Syst. iv. 356 (1838). C. uncinatum Royle ex Benth. in Royle, 111. i. 305 (1839). Echinospermum glochidiatum A. DC. Prod. x. 136 (1846). Paracaryum glochidiatum Benth. ex Hook. f. Fl. Brit. Ind. iv. 161 (1885). Rindera glochidiata^NaW. Cat. no. 926, nomen nudum. DeCandolle (1. c.) was the first to assign this plant to its proper genus, but it had been previously published by George Don as indicated above. It is of interest that the specimens in the Gray Herbarium are marked "Good Echinospermum" in Dr. Gray's handwriting. Lappula Redowskii (Hornem.) Greene, var. Karelini (Fisch. & Mey.), comb. nov. — Echinos'permum Karelini Fisch. & Mey. Ind. Sem. Hort. Petrop. xi. 67 (1846). E. Redowskii (Hornem.) Lehm., var. Karelini (Fisch. & Mey.) Regel, Act. Hort. Petrop. vi. 341 (1880). As indicated by Regel (1. c), like the typical form of the species, but having the sides and faces of the nutlets nearly or quite smooth. The related American species, L. texana (Scheele) Britton, shows a varia- tion analogous to this. Lappula omphaloides (Schrenk), comb. nov. — Echinospermum omphaloides Schrenk, Bull, phys.-math. Acad. Sci. St.-Petersb. iii. 211 (1845). I must concur in the opinion expressed by Lipsky, Act. Hort. Petrop. xxvi. 567 (1910), that this is a good species of the genus Lappula {Echinospermum). The correct combination, how- ever, does not seem to have been made. ^ Allocarya glabra (Gray), comb. nov. — Lithospermum glabrum Gray, Proc. Am. Acad. xvii. 227 (1882). Allocarya salina Jepson, 544 MACBRIDE. Fl. West.-Middle Calif, ed. 1, 442 (1901). Mrs. Brandegee, Zoe, v. 94-95, called attention to the true relationship of this plant as long ago as 1901, suggesting that it might be an introduction. More recently Prof. Jepson (1. c.) redescribed it from the Alvarado salt marshes. Although the label on Lemmon's specimen (the original) bears the notation "Arizona," the specimen probably came, as Mrs. Brandegee remarks, from California. Dr. Gray compared his species to L. incrassatum Guss. which is a good Lithospermum and which consequently bears only a superficial resemblance to A. glabra. The Old World plant at maturity develops a similarly fistulous- enlarged rhachis and callous-thickened calyx, but it has the fruit, the flowers and the aspect of other members of the genus. The near- est relative of A. glabra is A. stipltata Greene. Mrs. Brandegee doubts if the former is anything more than "a swollen form" of the latter. The swollen character is a very noticeable, but not by any means, it would seem, the strongest difference. However this may be, glabra is the older name and must be used regardless of the dis- position one may make of A. stipltata. Allocarya tenuicaulis (Phil.), comb. nov. — Eritrichium tenuicaule Phil. Linnaea, xxix. 18 (1857). E. uliginosum Phil. Anal. Univ. San- tiago, xliii. 519 (1873). Krynitzkia trachycarpa Gray, Proc. Am. Acad. XX. 266 (1885). Allocarya diffusa Greene, Pitt. i. 14 (1887). When Dr. Gray described this plant (1. c.) he referred to it two Chilian speci- mens remarking that "it may be suspected to be the Lithospermum muricaium of Ruiz & Pavon, and probably it may have other specific names; none of them, however, can be safely adopted." Two years later Dr. Greene (1. c.) transferred the Krynitzkia species belonging to his new genus and maintained the name trachycarpa "as to the Cali- fornian plants only," at the same time making the new combination A. idiginosa (Phil.) Greene, with the notation "Krynitzkia trachycarpa Gray as to the Chilian specimens doubtless." Reiche in his Flora de Chile (1910) has defined the Allocaryas of that country, and has definitely shown that Ruiz & Pavon's plant is not ours (a conclusion reached by Dr. Greene, 1. c). He treats the North American plant, however, as a synonym of the earlier E. uligiyiosum, thus following the opinion of Dr. Gray, who evidently assigned the new name trachy- carpa because he had at that time no means of knowing what name should be rightly taken up. The Reed specimen, which he cited, is probably A. sessiliflora (Poepp.) Greene, but the Harvey one corroborates Reiche's treatment. Unfortunately this much named plant has never been properly christened even yet. We are given the CERTAIN BORRAGINACEAE. 545 synonym E. tenuicaule Phil, in the Flora de Chile, but for some reason the author of that work used the much later E. uliginosum Phil. It is true that the former name is not desirable but since it is perfectly tenable, it must be used. For the complicated synonymy see the Flora de Chile, where Reiche gives the citations of some of the named forms of this rather variable species. Allocarya linifolia (Lehm.), comb. nov. — Anchusa linifoliahehm. Asperif. 215, no. 158 (1818). A. oppositifolia & pygmaea HBK. Nov. Gen. et Spec. iii. 91-92 (1818). Krynitzkia linifolia (Lehm.) Gray, Proc. Am. Acad. xx. 266 (1885). From these names of the same date between which priority cannot be determined I have used the name selected by Dr. Gray (1. c.) and have followed his interpretation of the species. Our specimens are from Peru, Ecuador, and Bolivia. Allocarya linifolia (Lehm.) Macbr., var. Kunthii (Walp.), comb. nov. — Anchusa Kunthii Walp. Nov. Act. Nat. Cur. xix. 372 (1843). Antiphytum Walpersii A. DC. Prod. x. 122 (1846). Eritri- chium Walpersii (A. DC.) Wedd. Chlor. And. ii. 90 (1859). The foliar characters given by the authors cited — the much longer and more uniformly linear leaves — seem to be the only differences between this plant and A. linifolia; the nutlets are the same. Eremocarya micrantha (Torr.) Greene, var. lepida (Gray), comb. nov. — Eritrichium viicranihum Torr., var. lepidum Gray, Syn. Fl. ii. pt. 1, 193 (1878). E. lepida (Gray) Greene, Pitt. i. 59 (1887). The variety is confluent with the species, as pointed out by Dr. Gray, Proc. Am. Acad. xx. 275 (1885). The nutlet variation is nicely illustrated by Abrams's no. 2904, Aug. 5, 1902, which is typical of the variety as first described except that some of the plants have smooth and lustrous nutlets. The description of the species given in the Synoptical Flora calls for either "smooth and shining or dull and puncticulate-scabrous " fruits. In the type-specimens these are smooth and Dr. Rydberg has segregated those having rough nutlets as E. muricata Rydb. Bull. Torr. Bot. Club, xxxvi. 677 (1909). Un- fortunately a co-type specimen, viz. Parry, no. 164, collected in 1874, has perfectly smooth nutlets. Evidently the character has no spe- cific value in this genus, since the large-flowered plant (var. lepida) shows the same variation, and since herbarium material seems to indicate that the smooth- and rough-fruited forms grow intermingled. Furthermore, if one maintains the rough-fruited form of the small- flowered plant as a species (E. muricata) we need yet another species for the rough-fruited form of the large-flowered plant. 546 MACBRIDE. Greeneocharis dichotoma (Greene), comb. nov. — Krynitzkia dichotoma Greene, Bull. Calif. Acad. i. 206 (1885). The original collec- tion from western Nevada is the only representation of this species at the Gray Herbarium; other specimens so referred belong rather to the widely distributed and somewhat variable G. circumscissa (H. & A.) Rydb. The latter is canescent with a more or less appressed-strigose pubescence, especially on the stems and branches. A plant with fine widely spreading hairs and scarcely, if at all, strigose-canescent has been collected at an elevation of 3050 m., while the typical form seldom attains half this altitude. This high-mountain variation may be known as Greeneocharis circumcissa (H. & A.) Rydb., var. hispida, var. nov., hispida vix strigoso-canescens; pilis patentibus. — Specimen examined: California: trail to Mt. Whitney, August 13, 1904, Cvlhcrtson, no. 4243 (type, in Gra.y Herb.). Plagiobothrys catalinensis (Gray), comb. nov. — F. arizmiicus (Gray) Greene, var. catalinensis Gray, Syn. Fl. ii. pt. 1, 431 (1886). Be- sides differing from P. arizonicus in the open fruiting-calyx with ovate lobes and the duller rougher nutlets (as pointed out by Dr. Gray, 1. c), P. catalinensis has other distinguishing features. Mature nutlets are only 1.5 mm. long, dark in color, the rugae obscure and not at all acute, the ventral keel low and narrow, and the caruncle small. Mature nutlets of the former plant are nearly or quite 2.5 mm. long, light (almost white) in color, the rugae very distinct and acute, and the ventral keel and caruncle usually prominent. Moreover the spikes of the mainland plant are usually interruptedly bracteate or even naked above; the spikes of the insular species are uniformly bracteate throughout. Oreocarya virgata (Porter) Greene, forma spicata (Rydb.), comb, nov.— 0. spicata Rydb. Bull. Torr. Bot. Club, xxxvi. 678 (1909). Although the surface-character of the nutlets is generally diagnostic in this genus, the smooth-fruited plant represented by the above name is surely not worthy even varietal rank, let alone specific. The nutlets of 0. virgata vary greatly in the degree of roughness; and plants with more or less roughened fruits and those with per- fectly smooth fruits that grow together in the region of Pike's Peak are otherwise indistinguishable. Oreocarya multicaulis (Torr.) Greene, var. cinerea (Greene), comb. nov. — 0. cinerea Greene, Pitt. iii. 113 (1896). The only character that distinguishes this is the pubescence. As in the typical form the color of the nutlets and the height of the stems amount to CERTAIN BORRAGINACEAE. 547 nothing. It is very doubtful if the several segregate species proposed in this group can be maintained as they are founded on these or other characters equally trivial. However, the variation treated here is so striking in its extreme form that it is worthy varietal designation. Since Dr. Greene failed to indicate any definite specimen, the following representative collections are noted. Specimens examined: Colo- rado: plains, Pueblo, 1873, Edward L. Greene (type). New Mexico: Mogollon Mountains, on the middle fork of the Gila River, Socorro Co., August 9, 1903, 0. B. Metcalf, no. 431. Arizona: vicinity of Flagstaff, June 4, 1898, Dr. D. T. MacDougal, nos. 40, 204. Mexico: Casas Grandes, Chihuahua, May 13, 1899, E. A. Goldman, no. 407. Oreocarya suffruticosa (Torr.) Greene, var. abortiva (Greene), comb. nov. — 0. abortiva Greene, Pitt. iii. 114 (1896). Krynitzkia imdticaulis Torr., var. abortiva (Greene) Jones, Contrib. W. Bot. xiii. 5 (1910). Jones (1. c.) has pointed out that the incurving of the nutlets is a characteristic common to all members of the group. When only one nutlet forms (as is sometimes the case in this plant and also in others) the ventral keel is larger than when more mature. It then, of course, seems to end even more abruptly. The Californian plant simply represents an extreme in this matter. It is otherwise allied to 0. suffruticosa rather than to the other species of the group. See the remarks by Parish, Eiyth. vii. 95 (1899), which further prove the plant to be unworthy specific rank. Oreocarya virginensis (Jones), comb. nov. — Krynitzskia gloincrata (Pursh) Gray, var. virginensis Jones, Contrib. W. Bot. xiii. 5 (1910). Very distinct from 0. glomerata, which has narrowly ovate not at all winged nutlets. Besides the specimens from La Verkin and Diamond ^^alley, Utah, cited by Mr. Jones, another from the same region, viz.: no. 173 by Dr. C. C. Parry, 1874, is of this species. Oreocarya sericea (Gray) Greene, Pitt. i. 58 (1887). — 0. humilis (Gray) Greene, 1. c. iii. 112 (1896) ? Krynitzkia sericea Gray, var. fulvocanescens Jones, Proc. Calif. Acad. Sci. ser. 2, v. 710 (1895). Eritrichium. ghmeratum (Pursh) DC, var. "! fulwcanescens Wats. Bot. King Exped. 243 (1871) in part, not E. fulvocanescens Gray, Proc. Am. Acad. x. 61 (1875) i. e. Krynitzkia echinoides Jones, 1. c. 709. Mr. Jones (1. c.) assigned a new name to the plant collected by Fendler in New Mexico and labeled in herb, by Dr. Gray " E. fulvocanescens," on the ground that the name must be applied to a very different plant collected by Watson in Nevada (no. 853), because this was the plant for which the name was first published. It is true that Watson took his no. 853 to be Gray^s fulvocanescens in herb.; but the first specific 548 MACBRIDE. ^ use of the name was by Dr. Gray (1. c.) and although he cited Watson's variety as a synonym his description is entirely based on Fendler's plant. Furthermore, Article 47 of the International Rules states, " When a species .... is divided into two or more groups of the same nature, if one of the forms was distinguished or described earlier than the other, the name is retained for that form." The name fulvo- canescens must apply, then, to Fendler's plant, since it was first dis- tinguished and first described as a species. Accordingly it is rather the plant collected by Watson and wrongly included by him in his description of fulvocanescens as a variety of glomerata which needs the new name unless already described. The latter alternative seems to represent the truth. Jones (1. c.) and Greene (1. c. Ill) were evidently writing about the same plant; and when Dr. Gray proposed the name sericea he included under it his earlier Eritrichiwn glomera- tum, var. humile. The material in the Gray Herbarium would indi- cate that he was justified in this; but Dr. Greene in using the name specifically, wrote " E. glomeratum, var. humile Gray in part." There- fore, if 0. humilis Greene is distinct from 0. sericea, the Watson plant from Nevada discussed above must bear the former rather than the latter name. Oreocarya oblata (Jones), comb. nov. — Krynitzkia oblata Jones, Contrib. W. Bot. xiii. 4 (1910). Very distinct from all other species having long white corollas. 0. Shockleyi Eastw. and K. me^isana Jones are the only other members of its immediate group. The latter is probably a good species, nearer the former than is 0. oblata, but I have seen no specimen. 0. oblata probably is not uncommon in Arizona, New Mexico, and Texas. Specimens examined: Texas: among rocks (corolla white), El Paso, March, 1851, George Tlmrber, no. 147, Sept. 1884, Marcus E. Jon.es, 1881, G.R. Vasey, March, 1885, Asa Gray. New Mexico: 1851-52, C. Wright, no. 1566, in part. Cryptantha barbigera (Gray) Greene, var. inops (Brandegee), comb. nov. — Krynitzkia barbigera Gray, var. inops Brandegee, Zoe, V. 228 (Sept. 1906). Mrs. Brandegee on one of her labels has rightly cited us synonyms of the above variety, C. nevadensis Nels. & Kenn. and C. arenicola Heller, published two and three months later respec- tively. The very slender acuminate nutlet is the principal character of the variety. The muriculations, especially near the tip of the fruit, are often very sharp. A specimen collected by Dr. Gray in the Grand Canon in 1885 and included by him in the species must now be referred to the'variety. Proceedings of the American Academy of Arts and Sciences. Vol. 51. No. 11.— April, 1916. CONTRIBUTIONS FROM THE ZOOLOGICAL LABORATORY OF THE MUSEUM OF COMPARATIVE ZOOLOGY AT HARVARD COLLEGE.— No. 269. ON THE LIFE-HISTORY OF CERATOMYXA ACADIENSIS, A NEW SPECIES OF MYXOSPORWIA FROM THE EASTERN COAST OF CANADA. By James W. Mayor. With Three Plates, and Three Figures in Text. CONTRIBUTIONS FROM THE ZOOLOGICAL LABORATORY OF THE MUSEUM OF COMPARATIVE ZOOLOGY AT HARVARD COLLEGE— 269 ON THE LIFE-HISTORY OF CERATOMYXA ACADIENSIS, A NEW SPECIES OF MYXOSPORIDIA FROM THE EASTERN COAST OF CANADA. By James W. Mayor. Received, December 8, 1915. Presented by E. L Mark. CONTENTS. Page. I. Introduction 551 II. Diagnosis of Ceratomyxa acadiensis, n. sp 553 III. Material and Methods 554 IV. Stages in the life-history of C. acadiensis found in the gaU bladder . 556 1. Trophic stages 556 2. Sporogony ■ 559 A. Formation of the sporoblasts 559 B. Development of the sporoblast into the spore . . . .561 C. Structure of the fully formed spore 566 3. Abnormalities in the structure and development of spores . . . 568 V. Summary 569 VI. Bibliography 570 VII. Explanation of plates 573 I. Introduction. Uncertainty still exists concerning the main phases of the life- history of the Myxosporidia. Since the two papers by Gurley ('93, '94), the only paper published on the Myxosporidia of American fishes is a short one by Tyzzer (:00). The present paper is a result of the study of the Myxosporidia of the gall bladders of the fishes of the eastern coast of America. A general report of the work has already been published by the author (Mavor, : 15). Work was begun on a species of Ceratomyxa from the gall bladder of Zoarces angularis, which showed abundant spores and stages in spore formation. Later a species having spores of the same form and size was found in the gall bladder of Urophycis chuss. A study of the myxosporidian stage of the two parasites showed them to be similar. It was therefore decided to regard them as being the same. 552 MAYOR. species. As the species is a new one, it has been named Ceratomi/xa acadiensis. The only detailed observations on the life-histor>' of a disporous form, since the paper of Doflein ('98), are those of Awerinzew (:08, : 09) on Ceratomyxa drepanopsettae Awer. Awerinzew confines his observations in the main to the stages leading to the formation of the spores. He found that a binucleated stage of the myxosporidium, in which the two nuclei were exactly alike, was followed by a stage in which there were four nuclei, two of which were entirely trophic in function, while the other two formed, by division, microgametes and macrogametes. After a reduction in the chromatin of their nuclei, first the protoplasm and then the nuclei of a macrogamete and a microgamete fuse. This condition differs fundamentally from the conditions in the Polysporea found in Sphaeromyxa sabrezesi by Schroder (:07) and in M^xobolus pfeifFeri by Keysselitz (:08), where a fusion of the germ-nuclei occurs in the fully formed spores. The writer has made a careful stud\- of the very early stages of the myxosporidium and of the later stages in the spore formation of C. acadiensis in order to extend and to verify in another species of Cera- tomyxa Awerinzew's observations. He has found that in C. acadien- sis — as was conjectured by Awerinzew for C. drepanopsettae — a stage occurs with a single nucleus and that this nucleus divides into two nuclei, which, however, differ from each other in size, staining, reaction, and function. The writer, then, has found that there occurs in C. acadiensis, even in the binucleated stage, the differentiation of the nuclei which Awerinzew found first in the stage with four nuclei. Schroder (:07) has described the fusion of the two nuclei in the sporoplasm of the full\' formed spore, and finds there the caryogamy of the life-cycle. On the other hand, Awerinzew (:09), who has found a reduction of chromatin followed by a fusion of nuclei in the early stages of spore-formation in Ceratomyxa depanopsettae, believes that the two nuclei in the germ of the spore do not fuse. Experiments were made calculated to ascertain whether or not there occurred a fusion of the two germ nuclei in the fully formed spore when this had remained for a length of time in the intestine of a host fish. These experiments have, however, been unsuccessful, owing probably to the spores used not being ripe. for transference to the intestine or to the conditions of the experiment not being those required for the continued life of the parasite. It may, howe\'er, be said that the constant occurrence of the two germ nuclei side by side throughout the process of spore formation suggests a later fusion rather than a separation. CERATOMYXA ACADIENSIS. 553 II. Diagnosis of Ceratomyxa acadiensis, n. sp. Myxosporidium, typically club shaped, with a long tail (PI. Ill, Figs. 42, 45, 50), often many times the length of the thicker part of the body (PI. Ill, Fig. 50), but remarkably variable in shape (PI. Ill, Figs. 41, 43, 44, 47, 48, 49). Large individuals may be irregularly stellate (PI. Ill, 48). The pseudopodia often show a rigidity, as if possessing an endoplasmic axis. Certain of the pseudopodia may show clumps of protoplasm along their length, the clumps being connected by thin hyaline filaments of ectoplasm (PI. Ill, Fig. 49). Division into ectoplasm and endoplasm, though not always clear, is often to be made out in the anterior region. In the parasite of Urophycis chuss the myxosporidia were very often found attached in large numbers to the myxosporidium of an undetermined species, described by the author in a previous paper (Mavor, : 15). The examination of detached individuals showed the connection to be brought about by short pseu- dopodia at the anterior end. In the parasite of Zoarces angularis the attachment is probably to the epithelium of the gall bladder, as the fine pseudopodia are present, but the undetermined myxosporidian present in U. chuss seems to be absent. In the parasite of Pseudo- pleuronectes americanus no attachment has been observed. The dimensions of a typical myxosporidium of the species are : Length, excluding tail 12-25 /x Width . 10-20 M Tail up to 60 /x Sporc,^ having the form typical for the genus, but very wide, short, and slightly compressed dorso-ventrally, with very long fine lateral filaments (Fig. A). Polar capsules spherical. Polar filaments not visible in the fresh state, but extruded in either concentrated sulphuric acid or a solution of iodine in potassic iodide. The average measure- ments of spores are as follows : Length — sutural axis 7- 8 /u Width — bivalve axis 40- 50 ^l Diameter of polar capsule 3- 4 /i Length of lateral filaments 250-300 n Length of extruded polar filament 70 ^ Occurrence: (1) in the gall bladder of Urophycis chuss on the coast of New Brunswick, Canada. Here the myxosporidium is usually 2 A more detailed description of the spore will be found on page 566. 554 MAYOR. found attached by its anterior end to an undetermined parasite, prob- ably some species of Myxidium or Chloromyxum, which itself is attached to the epithelium of the gall bladder. Nine out of ten specimens of U. chuss examined were found to be infected; (2) in the gall bladder of the eel pout, Zoarces angularis, from the same waters. The attachment of the myxosporidium was not observed in this host, although the modification of the anterior end of the parasite for attachment, as in the parasite of U. chuss, was seen. Each of the eight specimens of Z. angularis examined was found to be infected; (3) in the gall bladder of the winter flounder, Pseudopleiu'onectes americanus. Here no attachment of the myxosporidium was seen. Vegetative forms were found in abundance in the fishes examined, but spores only rarely. Of twenty-five flounders examined, all contained the parasite. Comparison toith other sjiecies. In size the spores resemble most closely those of C. appendiculata Thel. (Thelohan, '95, p. 337). As Thelohan does not give a figure of the spore, and as the only measure- ments given are those of the length and sutiu-al diameter, it is impos- sible to carry, the comparison further. The myxosporidium differs from that of C. appendiculata as described by Thelohan in being often found attached. III. Material and Methods. The material for the present investigation was collected in Passa- maquoddy Bay at or near the mouth of the St. Croix River during August and September, 1913.^ The hosts of the parasites, in the case of the common Hake, — Urophycis chuss, — and the Eel pout, — Zoarces angularis, — were caught on the "trawl" * with clam, whelk or herring as bait, and in the case of the winter flounder, — Pseudo- pleuronectes americanus, — in the seine on low sand flats. The 3 It is a pleasure to express my obligation to the Board of Directors of the Marine Biological Stations of Canada for kind permission to work at the Biological Station of the Canadian Government at St. Andrews, New Brnns- wick, Canada, and my great indebtedness to Dr. Huntsman, the ciuvitor of the Station, for putting the resources of the Station at my disposal, and for many personal services in obtaining and arranging for the preservation of the living material. 4 The trawl used on the coast of New Brunswick is a set-line, a mile or so in length, with hooks, spaced about two fathoms apart, set on the .sea bottom between two anchors. CERATOMYXA ACADIENSIS. 555 living fish were brought in a " car" to the laboratory, where they were kept alive in tanks supplied with running water. The study of the living parasite was made during the months of August and September, and all the preserved material was collected during the same period. For smear preparations cover glasses were prepared as follows: After being cleaned in a mixture of 1 part bichromate of potash and 1 part concentrated sulphuric acid to 12 parts water, the cover glasses were washed, first, in tap water, and then in distilled water, and stored in 95% alcohol. AVhen required for use, the alcohol was burned from them by passing them through the flame of an alcohol lamp. The bile duct of the fish was ligatured and the gall bladder removed to a carefully cleaned watch glass, where it was cut open. Into a pipette, freshly made from new glass tubing, a small quantity of bile was drawn up and thence dropped on the cover glass. Most of the bile was then sucked back into the pipette so as to leave on the cover glass only a very thin film. The cover glass was then inverted and allowed to drop on the fixing fluid in such a way that it was supported by the surface tension of the liquid. In this manner the preparations were given no opportunity to dry. This is practically the method of Dofflein ('98), but in all cases the addition of blood to the gall was avoided. Although the gall is at times very poor in coagulative material, it is nearly always possible by using a perfectly clean cover glass to get a good smear preparation. The fixing fluids employed were (1) Schaudinn's fluid, consisting of two parts saturated aqueous solution of corrosive sublimate to one part absolute alcohol, used either hot or cold, and (2) Hermann's fluid, consisting of 75 cc. of 1% platinic chloride, 4 cc. of 2% osmic acid and 1 cc. of glacial acetic acid. These fluids were allowed to act for from five to ten minutes, and the cover glasses were then transferred (after Schaudinn's fluid) to 60% alcohol containing iodine, or . (after Hermann's fluid) to distilled water. The stains used w^ere Giemsa's azur-eosin or Delafield's haematoxy- lin. Both were diluted before use to one or two per cent, and allowed to act for from twenty-four to forty-eight hours. After staining in Giemsa's mixture, the smears were washed in tap water and destained in a mixture containing 95% acetone and 5 per cent, xylol. When sufficiently destained, they were passed in succession through the fol- lowing mixtures: (1) acetone 70 cc. and xylol 30 cc, (2) acetone 50 cc. and xylol 50 cc; (3) pure xylol, and were finally mounted in Canada balsam. For the details of this method of using Giemsa's stain, Kisskalt und Hartmann (:10, p. 14) may be consulted. After stain- 556 MAYOR. ing in Delafield's haematoxylin, smears were either first destained in acid alcohol or mounted directly in Canada balsam. For the study of attached stages, the wall of the gall bladder was sectioned. Pieces of the bladder, opened in a watch glass as described above, were fixed in Schaudinn's fluid, imbedded in paraffin, and cut into sections from four to seven ;u in thickness. The sections were stained in Giemsa's mixture or in Delafield's haematoxylin diluted as for the smear preparations, or in Heidenliain's iron haematoxylin. In the case of Giemsa's stain the best results were obtained by wash- ing in water rapidly, for twenty seconds or so, and then destaining in a mixture of acetone 95 cc. and xylol 5 cc. for eight to ten minutes. IV. Stages in the Life-History of C. acadiensis found in the Gall Bladder. 1. Trophic Stages. The earliest stage of C. acadiensis which could be recognized as such in the gall bladder contained a single nucleus (PI. 1, Fig. 1). It is possessed of a characteristically elongated body with a long tail. Such uninucleated myxosporidia were only rarely to be found. There can be no doubt, however, from their form and structure, that such organisms represent a stage in the life-history of C. acadiensis. There is no possibility of confusion with any of the tissue cells of the host, e. g., the blood cells. The size of the nucleus, relative to that of the succeeding stages, is rather large. It shows a reticular structure, in which chromatin granules are embedded. The single nucleus of the myxosporidium divides by unequal divi- sion giving rise to two nuclei differing in size and staining reaction (PI. I, Figs. 2, 7). The exact nature of this division, whether mitotic or amitotic has not been ascertained. It is possible that it resembles the "Heteropole Teilung" described for Trypanosoma noctuae by Schaudinn (: 04, p. 397, Fig. 2c). The small size of the object makes the determination of this question unusually difficult. One of these nuclei, usually the larger, stains faint red, or almost pink, with Giemsa's stain, while the other stains a deep crimson. Both nuclei show an alveolar structure of the achromatin, in the walls of which relatively large chromatin granules are found. When stained with Delafield's haematoxylin, these nuclei show at one end, under certain conditions, a " Binnenkorper," distinguishable as a granule slightly larger than the others, and surrounded by a clear area (PI. I, Fig. 4) ; CERATOMYXA ACADIENSIS. 557 but the intensity with which the two nuclei are stained is usually, as in the case of Giemsa's stain, markedly different. The stage with two nuclei is one of the most, if not the most, fre- quently found of all the stages in the gall bladder. Hence, either this stage must be of long duration or such myxosporidia must be formed in abundance. The next stage found is one with three nuclei (PI. I, Fig. 8). Usually two of these are similar in structure and staining reaction to the small dark nucleus of the binucleated stage. On account of the small size of the object and the close proximity of the nuclei to one another, it has not been possible to observe stages in the division of these nuclei. Certain appearances, however, where the two deeply staining nuclei lie close together, lead one to the belief that they have arisen from the division of a single darkly staining nucleus. This stage is followed rapidly by a stage with four nuclei (PI. I, Figs. 9, 10), in which two of the nuclei stain very faintly and are usually the larger, while the other two stain very deeply and are usually somewdiat smaller. When the stage with fom* nuclei is reached, the protoplasm of the myxosporidium around the two deeply staining nuclei becomes denser and alters its staining reaction (PI. I, Fig. 9), so that corresponding to each of these nuclei a spherical cell containing a single nucleus is formed. These cells may be called " sporoblast-mother-cells " since they later form the sporoblasts. The tW'O faintly staining nuclei remain in the endoplasm of the myxosporidium, and take no part in the formation of the sporoblasts. They may be called "tropho- nuclei." During the stage when the myxosporidium contains four nuclei, and possibly in later stages, small chromatin granules are given off from the nuclei of the sporoblast-mother-cells (PI. I, Fig. 10), first into the protoplasm of these cells and later into the surrounding endo- plasm of the myxosporidium. Here, apparently, they are absorbed, since they are not usually found during the later stages of spore formation. The sporoblast-mother-cells continue to divide until twelve are formed (PI. I, Figs. 13-17). These twelve cells then come together and form the sporoblasts. If this account be compared wdth that of Awerinzew (:09) for C. drepanopsettae, a number of differences in the two species will be seen. In the binucleated stage described, but not figured, by Awerin- zew the two nuclei are similar in all respects. He says (:09, p. 78): "Das Amoboid mit zwei Kernen bildet sich ohne alien Zweifel aus 558 MAVOR. einem einkernigen durcli karyokinetische Teilimg des urspriinglichen Kernes; die beiden sich auf diese Weise bildenden Kerne sind einander voUig gleichwertig vind in niorpliologischer Hinsicht in keiner Weise voneinander verschieden." The two nuclei in the l^inucleate stage of the myxosporidium of C. acadiensis are not only different in size but they also show a different staining reaction. As one of these nuclei, the more faintly staining, probably gives rise to the two trophonuclei and the other, the more tleeply staining, to the nuclei of the sporoblasts, there are already at the binucleate stage two kinds of nuclei. This condition was not found by Awerinzew until the stage with four nuclei was reached. The fact that the two nuclei of the myxosporidium ha\e, the one a trophic and the other a propagative function, makes it improbable that the binucleated myxosporidium is formed by the fusion of two uninucleated individuals as has been supposed by some authors. A fusion at this stage is rendered still more improbable by the occurrence of stages in the division of the single nucleus. If a fusion of two myxo- sporidia really takes place in the life-history, it would seem more probable that it occurs between two binucleate individuals, each pos- sessing a vegetative and a propagative nucleus. Such a fusion between binucleate individuals would be an interesting parallel to the condi- tions found by Keysselitz (:08) and Schroder (: 10), where two pairs of cells come together to form a pansporoblast. Of the four cells found by these authors, one from each pair forms the en\elope and takes no part in the formation of the sporoblasts, while the other two form the two sporoblasts. The nuclei of the two cells which form the envelope could then be compared with the two trophonuclei of C. acadiensis, while the two cells which form the sporoblasts could be compared with the two sporoblast-mother-cells of C. acadiensis. The writer has, however, no evidence that such a fusion of myxosporidia occurs, and indeed the conditions in the stage with three nuclei is strong evidence against this. The two accounts (Awerinzew's and my own) agree in the main for the stage with four nuclei, there being in the myxosporidium of both C. acadiensis and C. drepanopsettae two trophic nuclei and two nuclei of propagative function. The formation of microgametes and macrogametes and their subsequent fusion to form zygotes, I have not observed. It must be said, however, that the small size of the object and the close prox- imity of the nuclei renders the observation of any such phenomena very difficult. ceratomyxa acadiensis. 559 2. Sporogony. A. Formation of the Sporoblasts. The first indication of the sporoblasts occurs when twelve of the nuclei of the myxosporidium with the more deeply staining protoplasm around them arrange themsehes in groups of sLx (PI. I, Fig. 17). The formation of the sporoblasts by the coming together of cells originally separate agrees with the observations of Awerinzew ( : 09) on the formation of the spores in C'eratom^xa drepanopsettae Awer. The young sporoblasts (PI. I, Figs. 17, 20, 25) lie side by side in the myxosporidium. They are ellipsoidal and the sides in contact are usually somewhat flattened (PI. I, Figs. 18, 20). The protoplasm of the sporoblast is usually" clearly to be distinguished from that of the myxosporidium by reason of its different reaction to stains, as seen for example in Plate I, Figure 17, where the protoplasm of the sporo- blast is deep blue, while the surrounding protoplasm is faintly stained. A sharp differentiation of the protoplasmic masses surrounding the separate nuclei of the sporoblast is not to be seen in Figure 17. Such a differentiation exists, however, in later stages (PI. I, Figs. 18, 20, 23, 24), where capsulogenous cells ^ and valve-cells^ are clearly to be distinguished from each other. It has not been possible to distinguish any difference in the nuclei which are brought together in the formation of the sporoblasts. The}' are uniform in size and stainability and show from four to ten deeply staining granules, mostly at the periphery, connected by achromatic strands passing through the nuclei. With Giemsa's stain these gran- ules are deep purple or mauve and the achromatic strands are faint red (PL I, Fig. 17). In certain cases when these nuclei are stained with Delafield's haematoxylin a slightly larger granule is separated from the rest. This granule probably represents the Binnenkorper of Schroder (:07). The nuclei resemble those of the sporoblasts of Sphaeromyxa sabrazesi Laveran et Mesnil as described by Schroder (: 07), in their uniform size and chromatic structure, but differ in con- taining a relativelv smaller number of chromatin granules. 5 The term "capsulogenous cell" is used for the French "cellule capsulogene" (Thelohan, '95, p. 280) and for the German "Polkapselzelle." 6 The term "valve cell" is used for the French term " cellule d'enveloppe " (CauUerv et Mesnil, :05) and the German term "Schallenzelle" (Keysselitz, :08). 560 MAYOR. Two nuclei are found in the myxosporidium in excess of the twelve concerned in the formation of the two sporoblasts (PI. II, Figs. 28 and 32, where in each case only one of the two nuclei is shown. These are the trophonuclei of the earlier stages. They gradually degenerate as spore formation advances, but often are still to be seen when the myxosporidium contains two almost completely developed spores. Since these nuclei degenerate and no other nuclei are found in the myxosporidium outside the sporoblasts, the myxosporidium must cease to exist at the end of spore-formation. These nuclei are possibly to be homologized with the residual nuclei of the polysporea. The two sporoblasts are formed in juxtaposition, and remain to- gether throughout their development into spores. In the parasite of U. chuss the fully formed spores, even after there is little trace of the myxosporidium left, are nearly always found in pairs, the members of which are in a definite relative position with regard to each other (PI. II, F'ig. 39). This is not as evident in the case of the parasite of P. americanus. Their position would suggest a close relation between the sporoblasts and perhaps the presence of some structure enclosing them both. However, no membrane or other limiting structure could be seen surrounding the sporoblasts. The pair of sporoblasts may lie anywhere in the myxosporidium, with their long axes parallel to, or making any angle with the long axis of the myxosporidium. The absence here of any structure suggesting a pansporoblast — Pans- poroblast, Gurley ('94); Sporoblast erster Ordnung, Butschli ('81); Sphere primitive, Thelohan ('95) — agrees with the conditions de- scribed by Awerinzew ('09) for Ceratomyxa drepanopsettae Awer.; but in the case of his species the sporoblasts seem not to preserve any definite position in relation to each other (cf. Awerinzew : 09, Taf. 7, Fig. 26, 27, 30). Neither Dofiein ('98) nor Thelohan ('95), both of whom studied the formation of the sporoblasts in Ceratomyxa, men- tion the presence of a pansporoblast, and their figures show no such structure. The absence of a pansporoblast from Zschokkella hildae Auer. is also mentioned by Auerbach (:09), and from Chloromyxum cristatum Leger, by Leger ( : 06) ; but in these cases, the sporoblasts are formed singly. The pansporoblast of the Polysporea contains, in the case of forms with two polar capsules, fourteen nuclei. When these nuclei, with the differentiated protoplasm surrounding each of them, become asso- ciated in groups of six to form the two sporoblasts, two of the four- teen nuclei remain in the pansporoblast as "Restkerne," or residual nuclei (cf . Schroder, : 07, Taf. 15, Fig. 32, and Keysselitz, : 08, Taf. 13, CERATOMYXA ACADIENSIS. 561 Fig. 76, Taf. 14, Fig. 77). These two residual nuclei are obviously to be compared with the two trophonuclei of the Disporea. This close parallelism between the Disporea and the Polysporea in the formation of the sporoblasts has been remarked by Doflein ('98, p. 309) . He says in regard to these residual nuclei in the Disporea : " Ich glaube, dieses regelmassige Vorkommen in Verein mit dem unten geschielderten Verhalten verschiedener disporen Formen erlaubt uns in dieser Kern- ausstossung den Ausdruck einer Reduction zu erblicken, und ich werde die beiden Kerne in dem Nachfolgenden als 'Restkerne' bezeichen." Doflein (: 09, p. 762), however, compares the sporoblast of the Disporea with the entire pansporoblast of the Polysporea, a comparison which seems justified neither in the light of the very close parallelism in the number and arrangement of the nuclei in the m;y^osporidium of Cera- tomyxa and in that of the pansporoblast of polysporic forms, nor in the light of the constant relative position of the two sporoblasts in Ceratomyxa, which corresponds precisely to the arrangement in the pansporoblast. B. Development of the Sporoblast into the Spore. An early stage of the development of the sporoblasts into spores is shown in Plate I, Figures 18, 19, 20, 25. In each of the two sporo- blasts the two cells wliich will later form the valves of the spore-shell are seen enveloping the other cells. These cells will henceforth be called valve-cells. The nuclei of the valve-cells are situated at the opposite ends of the sporoblast (PI. I, Figs. 18 and 20), and are already somewhat flattened. Such a cellular origin of the valves of the spore-shell among the Cnidosporidia was first found by CauUery et Mesnil ( : 05) in one of the Actinomyxidae, Sphaeractinomyxon stolci CauUery et Mesnil. A little later this condition was also found in the Myxosporidae simul- taneously and independently by Leger ( : 06, in Chloromyxum truttae Leger) and by Mercier (:06, in Myxobolus pfeifferi). These obser- vations were confirmed by Leger et Hesse (:06) for the genera Myxi- dium, Henneguya, Myxobolus, and later by Auerbach (:07) for different species of Chloromyxum and Myxidium. Leger et Hesse ( : 07) have further found this to be the condition in the new genus Coccomyxa. Sclirdder (:07) has also found it in Sphaeromyxon sabrazesi Laveran et Mesnil, and Awerinzew (:09) in Ceratomyxa. In the preparation shown in Figures 18, 19, 20, 25 (PI. I), the last 562 MAYOR. two figures being difTerent views of the same sporoblast, no division into separate capsulogenous cells could be made out in the mass of protoplasm which contains the two nuclei; but in another sporoblast of about the same stage (PI. I, Figs. 23, 24) a differentiation into two capsulogenous cells could be seen. As compared with the conditions shown in Figui-e 17 (PI. I), the nuclei of the capsulogenous cells have become slightly larger, and the nuclear net does not show clearly, prob- ably on account of a movement of achromatin toward the periphery of the nucleus. The formation of the polar capsules in cells of their own was first shown by Biitschli ('81), who was also the first to realize their true nature. In some cases, however, the formation of the separate cells occurs later in the development when the sporoblasts are already formed, as in the polysporic genera Myxobolus (Keysselitz, : 08) and Sphaeromyxa (Schroder :07). The two nuclei of the sporoplasm are seen close together on the left in Figure 18 (PI. I), in the middle in Figure 23, and alone in Figure 25. Henceforth these two nuclei will be called the germ-nuclei,^ since they are the nuclei of the germ carried in the spore-shell. The germ-nuclei have retained the size and deeply staining character they had in the sporoblast represented in Fig. 17 (PI. I). A differentiation of the sporoplasm into two regions corresponding to the two nuclei was not seen; these nuclei were usually found close together and, indeed, almost touching each other. The examination of a large number of stages in the development of the sporoblast has revealed the constant presence of six nuclei in every sporoblast and consequently of two nuclei in the future sporoplasm. It seems to be a universal condition for the sporoplasm of the myxo- sporidian spore to contain two germ nuclei, at least during the later stages of the formation of the spore. Thelohan ('95, p. 270) writes "Le protoplasma renferme toujom's deux noyaux qui sont le plus souvent accoles I'un a I'autre, d'une fa^on plus ou moins etroite, de telle sorte que, dans certains cas, il parait plutot avoir un noyau unique un peu allonge et un peu etrangle vers sa partie moyenne." In more recent years two nuclei have been demonstrated in the spores of Myxol)olus (Auerbach :06), Myxidium (Auerl>ach :07), Sphaero- myxa (Schroder :07), Chloromyxum (Auerbach :07), Ceratomyxa (Awerinzew :09), and Zschokkella (Auerbach :09). 7 The term "germ-nuclei" is used for the French " noyaux du sporoplasma, noyaux du germe," and for the German "Kerne des Amoeboidkeims." CERATOMYXA ACADIENSIS. 563 Figures 21 and 22 (PI. I), different views of the same sporoblast, show a sHghtly more advanced stage than the previous one. The nuclei of the valve-cells and the nuclei of the capsulogenous cells have increased slightly in size and the chromatin is nearly all at the periphery of the nuclei. The polar capsules have begun to develop, as is to be seen by the presence of two clear areas, each containing a stained central mass. In preparations stained with Heidenhain's iron haematoxylin there is an evident connection of this mass with the protoplasm forming the periphery of the clear area. It would there- fore seem that the polar filament is developed from a club-shaped mass of protoplasm, which grows out into the vacuole which ulti- mately forms the cavity of the capsule. It may, therefore, be that the polar filaments, like the valves of the spore envelope, are formed from metamorphosed protoplasm. The comparison can be carried further, as it is found that the filaments, the walls of the capsules and the valves of the spore-shell form continuous structures, which ad- here when the two valves are separated and the other parts of the spore have disappeared. The stage in the development of the polar filament described above resembles the stage found in Sphaeromyxa sabrazesi Laveran et Mesnil by Schroder, that formed in Myxobolus pfeifferi Th. by Keys- selitz (:08) and that in Zschokkella hildae Auerb., by Auerbach (:09). Any connection of the nucleus with the formation of the vacuole, such as Awerinzew has described for Ceratomyxa drepanopsettae Awer., I have not observed. A myxosporidium containing two sporo blasts at a slightly more advanced stage than that represented in Figures 21 and 22 (PI. II) is shown in Figure 28. The separate sporoblasts are represented in Figures 26, 27, 29, 30. Here the sporoblasts have elongated without increasing greatly in size and their long axes are becoming curved. The nuclei of the cells which form the valves show the chromatin at the periphery and a clear interior. The nuclei of the capsulogenous cells are approaching this condition. The nuclei of the sporoplasm show the same condition as in the previous figures. A somewhat older stage is shown in Figures 31-34, which represent, in the same way as Figures 26-30, a myxosporidium containing two sporoblasts. Each sporoblast has increased considerably in size and shows a greater curvature of its long axis. The sporoblasts are so placed that their concave surfaces are in contact, and they lie usually with their bivalve axes ^ parallel, and therefore so that one sporoblast 8 8ee page 566. 564 MAYOR. has both its ends on the same side of the other sporoblast (PI. II, Figs. 38, 39); but their bivalve axes may make an angle with each other, so that the two ends of one sporoblast are on opposite sides of the other sporoblast. This close contact of the sporoblasts is con- tinued throughout their later development, and even the fully formed spores found floating in the bile are usually associated in pairs, and in the same relative position as that described for their sporoblasts. In each of the sporoblasts one of the polar capsules has reached a more advanced stage than the other (PI. II, Figs. 31, 34), for while one preserves the condition of the previous stage, the other appears as a deeply stained spherical mass, having retained the stain owing to its thicker membrane. This inequality in the rate of development of the two polar capsules of the same spore seems to be of frequent occur- rence in this species, a fact which would indicate an independent action on the part of the two capsulogenous cells. The nuclei of the sporo- plasm — the germ-nuclei — lie close together in the sporoplasm and are so arranged that the nuclei in one sporoblast lie in the end diagonally opposite to the end in which the germ-nuclei of the other sporoblast lie (PI. II, Fig. 38). This relation of the nuclei of the sporoplasm of a pair of sporoblasts is the one usually found, but other arrangements occur; e. g., the nuclei of the sporoblasts may lie in the corresponding ends of their sporoblasts or one pair may lie in the middle of its sporoblast while the other pair lies nearer one end of its sporoblast. The latter condition is realized in Figure 39 (PI. II), where the nuclei of the sporoplasm in the left sporoblast, not drawn in the Figure, lie in the middle of the sporoblast, and so that a line joining their centres would correspond very nearly to the antero-posterior axis of the sporo- blast, whereas in the right sporoblast the germ-nuclei occupy the half of the sporoblast which is below in the figure. Other cases were found where the nuclei of one sporoblast were far apart and in opposite halves of their sporoblast (PL II, Fig. 35), while those of the other pair were near together in the same half. A myxosporidium con- taining a pair of sporoblasts in a still later stage of their development into spores is represented in Figure 38 (PI. II). The nuclei of the sporoblasts are represented diagrammatically by circles, those at a deeper focus being indicated by a paler line. In the sporoblasts (PI. II, Figs. 37, 36) the nuclei of the valve-cells show the chromatin granules collected in clumps at the periphery of the nucleus, which has a clear interior. The nuclei of the capsulogenous cells show a similar condition in a slightly less advanced state. The structure of the germ-nuclei is as before. The pairs of these nuclei, however, CERATOMYXA ACADIENSIS. 565 have moved in opposite directions along the antero-posterior axes of their respective spores. A similar state of afYairs is seen in the slightly more advanced sporo- blasts shown in Figure 39 (PI. II). The nuclei of the valve-cells show the same condition as that in the stage last described. The regular arrangement of the granules at the periphery of the nucleus suggests the presence of a nuclear membrane, and in some cases a faint sugges- tion of such a structure is to be seen as a continuation of the contour of the nucleus between the principal granules (PI. II, Fig. 39, upper left hand nucleus). The changes in the nuclei of the valve-cells fol- low closely those described for the nuclei of the valve-cells of Sphaer- omyxa sabrazesi by Schroder (:07, p. 368), who writes: "Die Schalenkerne werden etwas grosser, langlich und flach; ihr Chromatin sammelt sich mehr und mehr unter der Kernmembran an." In the right hand sporoblast of Figure 39 (PI. II), at the lower end of the figure, a little above the nucleus of the valve-cell, may be seen a faint transverse line, which probably marks the internal boundary of the valve-cell; this line is also to be seen in later stages (cf. PI. II, Fig. 40, upper end of spore). cps. pol. fill kit. Figure A. Ceratomyxa acadiensis, n. sp., spore drawn from a fresh preparation to show method of orientation and the form; a., anterior margin; p., posterior margin; s., left valve; dx., right valve; cps. j)ol., polar capsule; fil. lat., lateral filament. X 1800 diameters. A considerably later stage in the development of the spore, still in the gall bladder, is shown in Figure 40 (PI. II). The sporoblast has now become a spore. The valve-cells extend out in a lateral direction on either side. They are shaped like a cone with a cavity which ex- tends about half way towards its apex. The nuclei of the valve-cells stain \ery faintly, only one or two faint granules showing at the periphery, the greater part of the chromatic substance having dis- appeared. The nuclei of the capsulogenous cells also have shrunk somewhat and one of them, the lower in the Figure, is divided into two parts. The chromatin substance is still peripheral in position, but 566 MAYOR. separate chromatin granules are not to be distinguished. The two germ-nuclei remain unaltered. In spores which have been for some time in sea water or in the stomach of another host, the nuclei of the valve-cells are no longer to be seen and the nuclei of the capsulogenous cells appear as small deeply staining masses near the polar capsules. The nuclei of the capsulogenous cells in the fully formed spore are described by The- lohan ('95), Doflein ('98), Plehn (:04), Schroder (:07), and others, as small deeply staining bodies adherent to the polar capsules. C. Structure of the Fully Formed Spore. In studying the structure of the spore it is convenient to use the method of orientation employed by Thelohan ('95, p. 250-251) and generally adopted by subsequent writers. Where there is a single polar capsule (cps. poL), or two (Fig. A), or more, close together, the part of the spore in which the capsules lie is called anterior (a in Fig. A). The plane, a., p. (Fig. A), passing through the suture separating the two valves, is called the sutural plane. The spore is oriented by placing it with the polar capsules in front, and the sutural plane vertically (Fig. A). Then the front is anterior (a. Fig. A), the upper surface dorsal, and the lower surface ventral, the right side the right, and the left side the left. The sutural diameter (Thelohan, '95, p. 251) is the greatest diameter in the sutural plane. The bivalve axis (dx., s., Fig. A) is the line w^iich measures the greatest distance be- tween the two valves, perpendicular to the sutural plane. The general shape of the spore of Ceratomyxa acadiensis, may be described as that of a spindle of which the longitudinal axis has been bent into the arc of a circle. The chord of this arc is the bivalve axis, and may be called the width of the spore. The convex side of the arc is anterior, the concave side, posterior. The sutural axis extends in the antero-posterior direction and is equivalent to the length of the spore. The two valves are cone-shaped, the pointed ends being directed one to the right and the other to the left, and the bases meet along the plane of suture. The spore is slightly compressed dorso- ventrally. A slight variation in the foim and dimensions of the opposite valves was often noticed. This was not, however, constant, nor was it sufficient to show the degree of asymmetry found by Doflein ('98, p. 284) in such spores as those of C. inaequalis Doflein. The lateral filaments extending outwards from the tips of the valves CERATOMYXA ACADIENSIS. 567 on either side are very long and thin. Their exact length in the spore of the parasite from Urophycis chuss was not measured. Their extreme fineness and great length make this very difficult, except in very favorable preparations. They were measured in the case of the parasite of Zoarces angularis (Fig. B), where they were found to be 300 ju in length, and in the case of the parasite of P. ameri- canus, where they measured about 250 /z in length, or from five to six times the width of the spore, exclusive of the filaments. In Plate III, Figure 51, is shown a myxosporidium in which two spores have developed side by side in the relative positions described in a previous section. The protoplasm of the myxosporidium extends out along the lateral filaments of the spores. The lateral filaments were not seen at any time wound around the developing spores as described by Doflein ('98) for the spores of Ceratomyxa linospora. The cavity of the valves does not appear to be continued into the filaments. The length of these filaments is greater, both rela- tively to the width of the spore exclusive of them, and absolutely, than the length recorded for the lat- eral filaments of any other species of Ceratomyxa. The longest lateral filaments hitherto described are those of C. linospora Doflein, which are 20 ^t in length, or twice the width of the spore excluding filaments. In their length these lateral filaments may be compared with the long posterior filaments found by Nemeczek (:11) attached to the spores of Henneguya gigantea Nemeczek, which attain a length of 77-100 /x, or about nine times the length of the spore. Long filaments are most common in the two genera Ceratomyxa and Henneguya. It is generally believed that the filamentous appendages of myxosporidian spores aid the Figure B. Ceratomyxa acadiensis, n. sp., spores drawn with the Abbe camera lucida from fresh prepara- tions; a., from the gall bladder of Zoarces angularis showing lateral filaments, lower filament full length, upper one cut off; h., from the gall bladder of TToph>cis chuss showing polar filaments extruded. X 270 diameters. 568 MAYOR. distribution of the spores, both by retarding the rate at which they sink and by rendering them more easily carried by currents. The polar capsules are almost spherical and lie close together at the anterior end of the spore. They are so oriented that the polar filaments when extruded cross each other (Fig. B, b). The extrusion of the polar filaments was effected by the action of concentrated sul- phuric acid, and by a solution of iodine in potassic iodide, but they were not extruded by ammonia-water. The failure of the latter reagent may have been due to the spores not having been ripe. When extruded the filaments appear as very fine threads of uniform thick- ness. The sporoplasm as seen in fixed and stained preparations is eccen- trically placed (being nearly all in one valve), and contains, in all the spores studied from the gall bladder, two compact darkly staining nuclei (PI. II, Fig. 40). The dimensions of the spores are given under the diagnosis of the species. 3. Abnormalities in the Structure and Development of Spores. Triradiate spores of C. acadiensis were of frequent occurrence in all tliree hosts of the parasite. These spores may show a fairly regular radial symmetry, both as regards the valves and the polar capsules (Fig. C, a), or one of the valves may be smaller than the other two, while the three polar capsules are of equal size and symmetrically arranged (Fig. C, b). Cases where a triradiate spore and a normal spore were developing in the same myxosporidium were found (Fig. C, c), also cases where two triradiate spores were developing together in one myxosporidium. The possession of a triradiate spore-shell formed from three cells is characteristic of the group Actinomyxidae. The frequent occur- rence of triradiate spores, whose shell is presumably formed by three cells (since in many cases three polar-capsule nuclei can be seen), is interesting as indicating a possible relationship between these two groups. CERATOMYXA ACADIENSIS. 569 Figure C. Ceratomyxa acadiensis, n. sp., abnormalities in spores and their development, a and h, triradiate spores from the gall bladder of U. chuss; c, myxosporidium containing two sporoblasts, one forming a normal spore, the other forming a triradiate spore with three polar capsules, from the gall bladder of P. americanus. DrawTi from fresh preparations. X 390 diameters. V. Summary. 1. A new species of Ceratomyxa from the gall bladder of Urophycis chuss, Zoarces angularis, and Pseudopleuronectes americanus is described. 2. The earliest stage of Ceratomyxa acadiensis n. sp., found in the gall bladder, contains a single nucleus. 3. By a heteropolar division of this single nucleus a trophic and a propagative nucleus arise. 4. The stage of the mj^xosporidium with four nuclei probably arises by the division of the trophic nucleus to form two tropho-nuclei and the division of the propagative nucleus to form two propagative nuclei. 5. The origin of the sporoblasts by the coming together of cells originally separate, as described by Awerinzew for Ceratomyxa dre- panopsettae, is confirmed for C. acadiensis. 6. The presence of valve-cells and capsulogenous cells is established for C. acadiensis. 7. The two germ-nuclei can be distinguished in the early stages of spore-formation and until the spore is completely formed. 570 MAVOR. Bibliography. Auerbach, M. :06. Ein Myxobolus im Kopfe von Gadus aeglefinus L. Zool. Anz., Bd. 30, p. 568-570. :07. Bemerkungen iiber Myxosporidien heiinischer Siisswasser- fische. Zool. Anz., Bd. 32, p. 456-465. :09. Die Sporenbildung von Zschokkella und das System der Myxosporidien. Zool. Anz., Bd. 35, p. 240-256. Awerinzew, S. :08. Studien iiber parasitische Protozoen I-VII. Trav. Soc. Nat., St. Petersburg, Vol. 38, fasc. 2, pp. V-XII + 1-139. 4 pi. (Russian; resume in German, pp. 135-139.) :09. Studien iiber parasitische Protozoen. 1. Die Sporen- bildung bei Ceratomyxa drepanopsettae mihi. Arch. f. Protist., Bd. 14, p. 74-112, Taf. 7 u. 8. Butschli, O. '81. Beitrage zur Kenntnis der Fischpsorospermien. Zeitschr. f. wiss. Zool., Bd. 35, pp. 629-651, Taf. 31. Caullery, M., et Mesnil, F. :05. Recherches sur les Actinomyxidies. 1. Sphaeraetinomyx- on stolei Caullery et Mesnil. Arch. f. Protist., Bd. 6, p. 272-308, Taf. 15. Doflein, F. '98. Studien zur Naturgeschichte der Protozoen. III. Ueber Myxosporidien. Zool. Jahrb., Abt. f. Anat., Bd. 11, pp. 281-350, Taf. 18-24, 20 Textfig. :09. Lehrbuch der Protozoenkunde. Aufl. 2, Jena, 914 pp., 825 Textfig. Gurley, R. R. '93. On the classification of the Myxosporidia, a group of pro- tozoan parasites infesting fishes. Bull. U. S. Fish Comm. for 1891, Vol. 11, p. 407-^20. '94. The Myxosporidia, or psorosperms of fishes, and the epidemics produced by them. U. S. Comm. of Fish and Fisheries. Report of Comm'r for year ending June 30, 1892, pp. 65-304 + i-v, pi. 1^7. CERATOMYXA ACADIENSIS. 571 Keysselitz, G. :08. Die Entwicklung von Myxobolus pfeifferi Th. Arcli. f. Protist., Bd. 11, p. 252-308, Taf. 13-16. Kisskalt, K., und Hartman, M. :07. Praktikum der Bakteriologie und Protozoologie. Zweiter Teil. Protozoologie. Jena. Aufl. 2, Teil 2, vi + 106 pp. Leger, L. :06. Myxosporidies nouvelles parasites des poissons. Ann. Uiiiv. de Grenoble, Tom. 18, p. 267-272. Leger, L., et Hesse, E. :06. Sur la structure de la parol sporale des Myxosporidies. C. R. Acad. Sci., Paris, Tom. 142, p. 720-722. :07. Sur une nouvelle Myxosporidie parasite de la sardine. Ann. Univ. Grenoble, Tom. 19, No. 3, p. 703-706. Mavor, J. W. :15. Studies on the Sporozoa of the Fishes of the St. Andrews Region. Contributions to Canadian Biology. Supple- ment to the 47th Annual Report of the Department of Marine and Fisheries, Fisheries Branch, Ottawa, Sessional Paper No. 39b, pp. 25-38, pi. 4. Mercier, L. :06. Contributions a I'etude du developpement des spores ch°z Myxobolus Pfeifferi. C. R. Soc. Biol., Paris. Annee 1906 (58) Tom. 1., p. 763-764. Nemeczek, A. :11. Beitrage zur Kenntnis der Myxo- und Microsporidien der Fische. Arch. f. Protist., Bd. 22, p. 143-169. Taf. 8, 9. Plehn, Marianne. :04. tJber die Drehkrankheit der Salmoniden. [Lentospora cere- bratis (Hofer), Plehn.] Arch. f. Protist., Bd. 5, p. 145- 166. Taf. 5. Schaudinn, F. :04. Generations- und Wirtswechsel bei Trypanosoma und Spiso- chaete. Arb. a. d. kais. Gesundkeitsamt, Bd. 20, p. 387- 439. Reprint in Fritz Schaudinns Arbeiten. Schroder, O. :07. Beitrage zur Entwicklungsgeschichte der Myxosporidien. Sphaeromyxa labrazesi [proprie sabrazesi] (Laveran et Mesnil). Arch. f. Protist., Bd. 9, p. 359-381, Taf. 14-15. :10. tJber die Anlage der Sporocyste (Pansporoblast) bei Sphaeromyxa sabrazesi Laveran et Mesnil. Arch. f. Px-otist., Bd. 19, p. 1-5. 572 MAVOR. Thelohan, P. '95. Recherches sur les Myxosporidies. Bull. Sci. France et Belgique, Tom. 26, pp. 100-394, pi. 7-9. Tyzzer, E. E. :00. Tumors and Sporozoa in Fishes. Jour. Boston Soc. Med. Sci., Vol. 5, p. 63-68, pi. 6. I CERATOMYXA ACADIENSIS. EXPLANATION OF PLATES. All the figures in Plates I to III were drawn from smear preparations with the aid of an Abbe camera lucida, a Zeiss apochromatic 2 mm. oil immersion objective and compensating ocular No. 18 at a magnification of 2950 diameters; and are reproduced without reduction. PLATE I. Myxosporidia of Ceratomyxa acadiensis; Figures 1-4, 8 from preparations of the bile of Pseudopleuronectes americanus; Figures 5-7, 9-12 from prepara- tions of the bile of Urophycis chuss. Figures 1, 2, 4, 8 are from preparations stained with Delafield's haematoxylin ; Figures 3, 5-7, 9-10, 17, 18 from preparations stained with Giemsa's azur-eosin. Figure 1. Individual with a single nucleus. Figure 2. Individual containing a large vegetative and a small propagative nucleus. Individual with two large vegetative and two smaller propaga- tive nuclei. A myxosporidium containing four nuclei,— Figure 4, showing one large vegetative nucleus and two small propagative nuclei; Figure 5, at a different level, showing one large vegetative nucleus. An individual containing a larger vegetative nucleus, stained red, and a smaller propagative nucleus, stained dark mauve. Myxosporidium as in Figure 6. An individual with one large vegetative nucleus and two small, more deeply stained, propagative nuclei. A myxosporidium containing two larger more faintly staining vegetative nuclei and two smaller more deeply staining propagative nuclei each surrounded by a protoplasmic area stained blue. Individual similar to that seen in Figure 9. Just outside one of the blue protoplasmic areas can be seen a chromatin mass prob- ably extruded from the propagative nucleus nearest to it. Figures 11-25. The development of the sporoblast into the spore in C. acadiensis from the gall bladder of Urophycis chuss. Figures 11, 12, 16. Three views of a myxosporidium containing eight nuclei, each surrounded by a differentiated area of protoplasm, blue in the preparation. Figure 16 is the appearance at a high focus showing one of the nuclei; Figure 11 at the middle focus showing five of the nuclei; and Figure 12 at a low- focus showing two of the nuclei. Figures 13-15. Views at three levels of a myxosporidium containing eleven nuclei, each with a differentiated area of protoplasm around it; Figure 13 at a high focus. Figure 15 through the middle and Figure 14 at a low focus. A myxosporidium containing twelve nuclei with their differ- entiated areas of protoplasm (blue in the preparation) arranged to form a pair of sporoblasts. Figure 3. Figures 4, 5. Figure 6. Figure 7. Figure 8. Figure 9. Figure 10. Figure 17. 574 MAYOR. Figures 18-20, 25. Individuals containing two sporoblasts. Figure 18. The left sporoblast. At opposite ends are seen the two valve-cells containing their flattened nuclei. The germ-nuclei are the two more deeply-staining ones nearest the left side of the sporoblast. The two nuclei nearest the right side are the nuclei of the capsulogenous cells. Figure 19. Diagram showing the position of the left and right sporo- blasts and their nuclei in the myxosporidium. Figures 20, 25. The right sporoblast; Figure 20 showing the appearance at a high focus, at which the valve-cells and the nuclei of the capsulogenous cells are seen; Figure 25 showing the appear- ance at a low focus, at which the germ-nuclei become visible. Figures 21, 22. Two views of an isolated sporoblast; Figure 21 at a high focus, showing near the ends the two nuclei of the capsulogen- ous cells and at the left the two more deeply-staining germ- nuclei; Figure 22, at a deeper focus, showing the two nuclei of the cells which form the valves of the shell. Figures 23, 24. Two views of an isolated sporoblast; Figure 23 at a high focus, showing at the ends the two cells which form the valves of the shell and near the center the two germ-nuclei; Figure 24 at a lower focus, showing the two capsulogenous cells. Mavor.— Ceratomyxa Acadiensis I r.^) Ci^ # Plate I i 1 i^. ^^ 10 ,PJ :^^^ v^> '. * 1 1 12 13 14 15 ^. 16 17 .E 18 20 # .■rf-> I't 21 J. V/. M. DfcL. ^ 22 ~-^-7> 23 ^ 24 Proc. Amer. Acad, Arts and Sciences. Vol. 51 25 I 576 MAYOR. PLATE II. Figures 26-30. A myxosporidium containing two sporoblasts. Figures 26, 27. Left sporoblast; Figure 26 at a low focus, showing the nucleus of a valve-forming cell (above) and the nucleus of a capsulogenous cell (below); Figure 27, at a higher focus, showing the nucleus of a capsulogenous cell (above), the nucleus of a valve-cell (below), and the two germ-nuclei. Figure 28. A diagram of the myxosporidium to show relations of sporo- blasts. On the upper left side is shown a nucleus lying outside the sporoblasts, one of the degenerating trophic nuclei. The other nuclei are shown diagrammaticaUy as irregular circles. Figures 29, .30. The right sporoblast; Figure 29, at a low focus, showing the two nuclei of the capsulogenous cells; Figure 30, at a higher focus, showing the two nuclei of the valve-cells at opposite ends and the two germ-nuclei near the centre. Figures 31-34. A myxosporidium containing two sporoblasts. Figure 31. Left sporoblast showing (near opposite ends) the two nuclei of the valve-cells, the tw^o nuclei of the capsulogenous cells and, below the middle, the two deeply-staining germ-nuclei. One polar capsule, drawn as a circle, showed in the preparation as a deeply-stained mauve sphere, the other, higher up in the figure, as a clear unstained area. Figure 32. The myxosporidium showing the positions of the sporoblasts and their nuclei. The small nucleus near the centre, drawn in detail, lies outside the sporoblasts; it is one of the degen- erating trophic nuclei. The other nuclei are drawn diagram- maticaUy as irregular circles. Figures 33, 34. Right sporoblast; Fig'iro 33, at a low focus, showing the two germ-nuclei; Figure 34, at a higher focus, showing, at the ends, the two nuclei of the valve-ceUs and, in the middle, the nuclei of the capsulogenous cells. The polar capsules show the same diiTerence in development as in the left sporo- blast. Figure 35. Advanced stage in the development of a spore, showing germ- nuclei far apart. Figure 36. Upper of the two spores shown in Figure 38, showing the nuclei of the two valve cells at the ends, the polar capsules and near to them the nuclei of the capsulogenous ceUs. The germ nuclei lie together on the left side and are smaller and more deeply stained than the others. Figure 37. The lower of the two spores in Figure 38. Figure 38. A myxosporidium showing the relative positions of the two young spores. Figure 39. Two almost fully developed spores. The germ-nuclei are seen in the right hand spore; but not in the left. Figure 40. A fully developed spore from the gall bladder. The capsulo- genous nuclei are degenerating as are also the valve nuclei. Mavor— Ceratomyxa Acadiensis Plate II C 26 mi "M:^ • 31 29 r 1 33 li 30 .'*A 34 J. W. M. Del. Proc. Amer. Acad. Arts and Sciences. Vol 51 578 MAYOR. PLATE III. Myxosporidia of Ceratomyxa acadiensis drawn from life. All the figures, with the exception of Figure 51, were drawn with a 4 mm. objective and ocular X 6 at a magnification of 660 diameters. Figure 41. Myxosporidium containing two sporoblasts, and showing in the anterior part a nucleus. From the gall bladder of Pseudo- pleuronectes americanus. Figure 42. Myxosporidium containing two sporoblasts. From gall bladder of P. americanus. Figure 43. Myxosporidium containing two .sporoblasts. From gall bladder of Zoarces angularis. Figure 44. Myxosporidium containing two sporoblasts. From gall bladder of P. americanus. Figures 45-47. Myxosporidia. From the gall bladder of Z. angularis. Figures 48, 49. Myxosporidia each contaming two sporoblasts. From the gall bladder of P. americanus. Figure 50. Myxosporidium. From the gall bladder of Z. angularis. Figure 51. Myxosporidium containing two spores and showing proto- plasmic prolongations in which are spore filaments. From the gall bladder of P. americanus. X 320. Mavor— Ceratomyxa Acadiensis Plate III 41 42 o'o 44 -**% /Ty^ 43 \ • •'• 45 51 . •• • 46 <• <•.' 47 »-^ o' o 48 \ 49 50 J. W. M. Del Proc. Amer. Acad. Arts and Sciences. Vol. 51 HELIOTYPE CO., BOSTON i Proceedings of the American Academy of Arts and Sciences. Vol. 51. No. 12. — April, 1916. POLYMORPHIC CHANGES UNDER PRESSURE OF THE UNIVALEN T NI TRA TES. By p. W. Bridgman. Investigations on Light and Heat made and published with aid from the RuMFORD Fund. POLYMORPHIC CHANGES UNDER PRESSURE OF THE UNIVALENT NITRATES. By p. W. Bridgman, Received, January 13, 1916. TABLE OF CONTENTS. Page. Introduction 581 Silver Nitrate 582 Caesium Nitrate . . . . • 587 Rubidium Nitrate 590 Thallium Nitrate 593 Potassium Nitrate 599 Ammonium Nitrate 605 Nitrates with No New Forms 618 Discussion 619 Summary 624 Introduction. In this paper the thermodynamic elements involved in the poly- morphic changes of a group of closely related univalent nitrates are given over a considerable pressure and temperature range. The apparatus, the experimental methods, and the methods of computa- tion are in every way the same as were described in a previous paper ^ on a similar subject. Particular interest attaches to the group of the univalent nitrates because of the wide occurrence of polymorphism in this group, the unusual number of polymorphic forms, and their complicated rela- tionships. Here, one would expect, is a particularly favorable field in which to look for new forms, stable only at high pressures, which will supply the gaps at atmospheric pressure. It turns out, however, that there are comparatively few new forms in the range reached here : one new form for ammonium nitrate and two for potassium nitrate; 1 P. W. Bridgman, Proc. Amer. Acad. 51, 53-124 (1915). 582 BRIDGMAN. The data are presented in the order of compHcation of the phase diagrams, and inehide the coordinates of the transition curves, the change of volume, latent heat and internal energy, and in those cases where the measurements were sufficiently accurate, an estimate of the difference of compressibility, thermal expansion, and specific heat of the different modifications. In the discussion of the relations of the various forms it will appear that the three simplest nitrates, those of Rubidium, Caesium, and Thallium, are much alike in their behavior under pressure; the nitrates of Ammonium and Potassium are much less simple in their relations, and Silver Nitrate is appar- ently not connected with these at all. Detailed Data. Silver Nitrate. — Two batches of this substance were used. The first was from Eimer and Amend, c. p., and was used both for the run at atmospheric pressure and for the runs over the entire high pressure range. For the determinations at high pressures it was hammered dry into an open steel shell. The high pressure points were found in the order of decreasing pressure and increasing tempera- ture. For the measurement at low pressure, the AgNOs was fused into a glass tube, in which it remained. The second batch was from the J. T. Baker Chemical Co., "analyzed chemicals," and showed by analysis less than 0.002% of impurity. This was fused into glass moulds, from which it was removed for the measurements. The moulds were of such a size that the sticks of AgNOs fitted loosely the inside of the pressure cylinder. In all cases, pressure was trans- mitted directly to the AgNOs by kerosene. With this second batch, five of the high pressure points were repeated, between 6000 and 12000 kgm., six months after the original determinations. Every part of the apparatus was different from the original, which had been de- stroyed piece by piece in the meantime by various explosions. The agreement of the p-t values of these two determinations is good. This repetition was made necessary because of an accidental distribu- tion of error in the two lowest Av points of the original measurements, such that there appeared to be a discontinuity in the A^ curve, so that there was consequently a possibility that another modification might exist. No perceptible decomposition of the AgNOs could be detected where it came in contact with the steel, although such might be EFFECTS OF PRESSURE ON NITRATES. 5S3 expected, because iron is known to deposit silver from solution. There was, however, some slight amount of decomposition somewhere, as shown by the escape of a few bubbles of gas when the apparatus was opened after the measurements at high temperatures. This decomposition might be either in the kerosene or the AgNOs. No such decomposition could be detected, however, after the measure- ments at the lower temperatures and higher pressures. Possibly this slight decomposition might explain the two discordant points which were discarded, although there are other possibilities, such as the fact that the change of volume is so small that a very slight amount of hysteresis would produce a comparatively large error. On repeating the measurements the possibility of hysteresis was avoided as far as possible by running the pressure back and forth several times over the transition before beginning the measurements. The quantity of AgNOs used in these runs varied from 61 to 103 gm. The direct experimental points are shown in Figure 1, the computed values of AH and AE in Figure 2, and the numerical results are col- lected in Table I. The only points of either determination which have been discarded are the two bad Av points already mentioned; one of these was about 30% and the other 12% too high. In view of TABLE I. Silver Nitrate. Pressure Temperature AV cm.3/gm. dr dp Latent Heat kgm.m./gm. Change of Energy kgm.m./gm. 1 159°. 4 .002.50 -.0075 1.442 1.442 1000 151 .8 254 77 1.402 1.427 2000 14.3 .9 259 80 1.3.50 1.402 3000 135 .8 265 83 1.305 1.385 4000 127 .4 272 86 1.267 1.376 5000 118 .7 279 89 1.228 1.368 6000 109 .7 287 96 1 . 145 1.317 7000 99 .0 296 125 .881 1.088 8000 83 .2 307 200 .548 .794 9000 .56 .8 320 415 .2,54 .542 9500 28 .3 326 990(?) (?) (?)^ 9770 0 .0 330 (V) (?)^ 584 BRIDGMAN. f^Momut 123456789 Pressure, kgm./cm.^ x lO"* Silver Nitrate Figure 1. Silver Nitrate. The observed equilibrium pressures and tem- peratures (circles) and the observed changes of volume (crosses) • 123456789 Pressure, kgm./cm.^ x 10* Silver Nitrate 10 Figure 2. Silver Nitrate. The computed heat of transition and the diflference of internal energy between the two phases. EFFECTS OF PRESSURE ON NITRATES. 585 the large probable error because of the great width of the region of dr indifference at the lower end of the curve, no values of AH or -t~ are dp tabulated for this part of the curve. The feature of particular interest for AgNOs is the remarkably sudden increase of curvature of the transition line beyond 6000 kgm., a feature unique to this substance. In fact, the increase of curvature is so great that when I first found the transition at 20°, it seemed as if there must be a third modification, and that the phase diagram would look like that of Agl. The whole of the field was very carefully explored for other modifications; the strongest motive for repeating the measurements was that the discordant At; points lent color to the suspicion that there might have been an undetected third modifica- tion near the bend with a small A?;. No other form could be found, however, to 12000 kgm. at room temperature, or at 200°, and on the repetition of the experiment none could be found in the neighborhood of the bend. It is possible to pass over the curve at any point without the reac- tion starting. The width of the region of indifference within which the reaction does not run after it has once started changes in a curious way with the pressure. The detailed data for this effect are to be given in the succeeding paper. There are several measurements of the transition data at atmos- pheric pressure. Schwarz ^ gives 159.2° to 159.7° by an optical method for the transition temperature, Hissink ^ finds 159.8° with a dilatometer, and Guinchant * gives 159°. The value found by extra- polation of the above data of my own was 159.4°. The change of linear dimensions during the transition has been measured by Guin- chant,* who found an increase of from 0.22 to 0.25% on cooling, and a decrease of 0.17% on heating. Assuming 4.3 for the density, this is equivalent to about 0.00047 cm^ per gm., against 0.00250 found above by direct methods. This shows that the change of volume during the transition does not take place equally in all directions, an effect shown by other substances also. Guinchant ^ has also found the latent heat of the transition to be 4.9 cal. against 3.4 found above. The agreement should be closer. The difference of compressibility of the two modifications was 2 W. Schwarz, Diss. Gott. 1892; Beibl. 17, 629 (189:3). 3 D. J. Hissink, ZS. phys. Chem. 32, 537-563 (1900). 4 J. Guinchant, C. R. 149, 569-571 (1909). 5 J. Guinchant, C. R. 145, 320-322 (1907). 586 BRIDGMAN. dii'ectly measured at several points. These values, which are some- what more self consistent than usual, are shown in Figure 3. At low pressures (high temperatures), the form with the larger volume is more compressible, but as we go along the transition curve to higher pressures and lower temperatures, the difference of compressibility become less, and finally reverses sign in the neighborhood of the region of rapid curvature. Below 70° the modification stable at the higher temperature is the more compressible. This is perhaps what one might be at first inclined to call the normal behavior, but it is to be noticed that there are two opposite factors here. It is natural to think that the phase of greater volume, as well as the high tempera- ture phase would be the more compressible. Evidently both of these things cannot happen at the same time on this curve. At low pres- <-.0„6 12345678 9 Pressure, kgm./cm.^ x 10^ Silver Nitrate 10 Figure 3. Silver Nitrate. The observed differences of compressibility between the two phases. sures it is the phase of greater volume which is the more compres- sible, but at high pressures it is the high temperature phase which is the more compressible. The relations in this regard at the high pressure end of the curve are similar to those on the melting curve of water and ice I, where water, although of smaller volume, is more compressible than ice. It should be remarked that this reversal in the sign of the compressibility is not to be regarded as remarkable in view of the sudden change of direction of the transition curve. The transition curve is unusual in this respect; one may expect unusual features in the behavior of the two phases to correspond. The differences of the expansion and of the specific heats may be calculated from the difference of compressibility, and are shown in Table II. The difference of expansion runs roughly parallel to the EFFECTS OF PRESSURE ON NITRATES. 587 TABLE II. Silver Nitrate. Pressure kgm./cm.s Aa A/3 ACp kgm.cm./gm. 1 -0.0^45 -O.O4.55 0..35 2000 - .06.35 - .O437 0.21 4000 - .0625 - .O422 0.11 GOOO - .0615 - .O56 0.13 8000 - .0,2 + .0,5 0.13 9000 + .07? + .0,5 0.08 difPerence of compressibility. On the low pressure end of the curve the high temperature phase is the less expansible, but at the other end of the cui"ve it becomes more expansible. The change in sign of Aj3 occurs at a lower pressure than that of Aa. Just what the struc- tural change in the crystal is which brings about these changes in sign of Aa and Aj8 is difficult to say; it may be a compacting together under pressure of the crystal framework of the low temperature form which makes it less responsive to changes of temperature and pressure. The compressibility of AgNOs (II) would be expected to show con- siderable variations with pressure just as ice I does. The difference of specific heat at atmospheric pressure, calculated with the above values of Aa and Aj8, is 0.0082 cal. per gm., which agrees almost exactly with the directly determined value, 0.008, of Guinchant.^ The difference of specific heat decreases at the higher pressures. There is no change of sign, however, but the high tempera- ture phase has throughout the greater specific heat. Caesium Nitrate. This substance was most kindly loaned by Professor Baxter, by whom it had been purified for his atomic weight work. Immediately before using it was dried in vacuum at 100° for several hours. This is a rare substance, and I was very fortunate to be able to get enough of it to work with; I take this opportunity of expressing my gratitude for Professor Baxter's kindness. 568 BRIDGMAN. Only one sample of the substance was used, 75 gm. in amount, and both the high and low pressure determinations were made with it. It was hammered dry into an open steel shell, and pressure was trans- mitted directly to it by kerosene. Two determinations were made at low pressures; one by ^•arying temperature at constant pressure to give Av, and one by varying pressure at constant temperature to give the transition point. Four points at higher pressures, all by the isothermal method, ere sufficient. The domain of indifference is nowhere wide, but it varies greatly at (liferent temperatures, a rather unusual effect. i\.t 100 kgm. it was possible to shut the equilibrium pressure within limits only 11 kgm. apart; at 177° the limits were 8 kgm., at 164° 17 kgm., and 189° 70 kgm., and at 202°, 125 kgm. It is unusual that the limits at 177° were narrower than at 164°, but it is to be noticed that the 177° point was determined fh-st; possibly the transition may have developed fissures in the solid which would account for the wider limits at 164°. It might be thought natural that along with the narrowness of the bund of indifference at the lower temperatures would go a high rate of reaction, but this was not so. The reaction velocity was distinctly slower than for most solids; at 100 kgm. about 90 minutes were TABLE III. Caesium Nitrate. Pressure Temperature AV cm.Vgni. dr dp Latent Heat kgm.m./gm. Change of Energy kgm.m./gm. 1 153°. 7 .00405 .0095 1.820 1.820 1000 163 .1 387 93 1.815 1.776 2000 172 .3 369 91 1.806 1.732 3000 181 .3 352 89 1.797 1.691 4000 190 .1 334 87 1.778 1.644 5000 198 .7 316 85 1.754 1.596 0000 207 .1 298 83 1.724 1.545 EFFECTS OF PRESSURE ON NITRATES. 589 required to reach approximately stationary conditions, and at each of the high pressure points about 50 minutes. As for most soHds, the velocity was distinctly greater with falling pressure than with rising. It is possible to pass across the equilibrium curve in either direction without the reaction starting, but only to a slight extent The experimental results are shown in Figure 4, the computed 0 12 3 4 5 6 Pressure, kgm./cm.^ x 10^ Caesium Nitrate Figure," 4. Caesium Nitrate. The observed equilibrium pressures and temperatures (circles) and the observed changes of volume (crosses). E 0 12 3 4 5 6 Pressure, kgm./cm.' x 10" Caesium Nitrate Figure 5. Caesium Nitrate. The computed heat of transition and the difference of internal energy. values of AH and AE in Figure 5, and the numerical values in Table III. The ciu-vatm-e of the y-t curve is normal ; the AV cu^^•e falls with rising pressure, as is normal, but the experimental accuracy was not great enough to surely determine the direction of cur\'atiu-e, so it is drawn as a straight line. There seems to be onlv one value for comparison at atmospheric 590 BRIDGMAN. pressure, 145° for the transition temperature by Wallerant.^ The value found above for the transition temperature was 153.7°. In view of the exceptional purity of the sample used above, it would seem that 153.7° should be given the preference. Direct measurements were made of the difference of thermal ex- pansion of the two phases at low pressures (77 kgm.), and of the difference of compressibility at higher pressures. The measurements are better than usual, and give more consistent results. Three measurements of the difference of compressibility at high pressures were made; one of these was somewhat too large, but the other two agreed within 5%. The values follow. At 77 kgm. Aa = 0.0649 A/3 - O.O433 At 5000 kgm. ACp= 1.0 Aa = O.OeSl A/S = O.O4I5 ACn= 0.28 At 77 kgm., A|3 was the observed value, and at 5000, Aa. We see that at both pressures the high temperature phase is the more com- pressible, has the greater expansion, and the greater specific heat. The differences between the two phases become less at higher pres- sures. The decrease of Aa with rising pressure was unmistakably shown by the direct measurements. This behavior of CsNOs is what one would perhaps be inclined to call entirely normal. The difference of compressibility between the two phases is of the order of the com- pressibility of platinum. Especially careful search was made for other modifications, since others were expected in analogy with the other nitrates. None were found to 12000 kgm. however, at 20°, 177°, or 200°. Rubidium Nitrate. I have to thank the kindness of Professor T. W. Richards for this rare substance. He personally prepared the samples by repeated crystallization. The purification was continued until the flame test, very sensitive in these cases, showed no trace of any of the other alkali metals. The salt as finally provided by Pro- fessor Richards was crystallized from acid solution, and gave an acid reaction. It was unfortunate that I did not succeed in removing all 6 F. Wallerant, Bull. soc. fr. min. (190,5) 311-374. EFFECTS OF PRESSURE ON NITRATES. 591 of the acid; even heating to over 200° in vacuum for several hours was not effective. The presence of free acid was shown by the corro- sion of the steel shell and cylinder after the runs were finished. Cer- tain very puzzling inconsistencies in the behavior under pressure are probably to be explained by the presence of acid. Professor Richards has therefore undertaken to prepare a second sample, this time crystallized from neutral solution. In the preparation of this sample several unexpected delays have arisen and the way does not yet seem clear to a successful outcome. I have therefore thought it best to publish tentatively the results already obtained, and to reserve for a future note the data on the perfectly pure sample. None of the con- clusions reached in this paper can be altered by a slight change in the numerical values. My unwillingness to delay longer is because the rest of the data of this paper have now been awaiting publication for somewhat over a year, and a succeeding longer paper is entirely finished, waiting only for the appearance of this paper. Runs were made at three different times. The first gave five points from 1000 to COOO kgm. These results were fairly regular; the Av point at the lowest pressure was lower than was to be expected, however. After two months the points at approximately atmospheric pressure were determined. The equilibrium temperature was nearly 1° low, judged by extrapolation from the previous high pressure points, and Av was 10% low. There was no possibility of accidental experimental error; two independent determinations at 78 kgm. gave very concordant results. Again after two months, four more points were redetermined at high pressures. The Av points of this run were very much lower than the previous points and more irregular; the greatest discrepancy being 30% at 4000 kgm. The equililjrium temperatures also did not agree, at 1000 kgm. being 0.4° low, and at 4000 kgm. 1.5° high. The transition line as redetermined w^ould, therefore, extrapolate to a lower transition temperature at atmospheric pressure. It seems probable that there was some slow permanent change taking place in the RbNOa due to the free acid. In the Figure and the Table, I have therefore given only the results of the first run, which seem more likely to be accurate. The equilibrium pressures and temperatures and the changes of volume are shown in Figure 6, and the numerical values are collected in Table IV. In view of the certainty of revision of these data, it did not seem worth while to give a diagram as usual for AH and AE. RbNOs is known to have another transition point at atmospheric pressure,^ near 219°, so that there is another modification not shown 592 BRIDGMAN. on the diagram. It is for this reason that the two phases shown above are numbered II and III. I made careful search for the transi- tion from I to II at 239° between 500 and 7000 kgm., but without success. It is very unHkely that the slope of this transition line is so oj-.ooe E .002^; u 1 2 3 4 i—^^O^;^ Pressure, kgm./cm.' x 10^ <] Rubidium Nitrate Figure 6. Rubidium Nitrate. The observed equilibrium pressures and temperatures (circles) and the observed changes of volume (crosses). It is to be remembered that these values are subject to revision. TABLE IV. Rubidium Nitrate.* Pressure Temperature Ar cm.3/gm. dp Latent Heat kgm.m./gm. Change of Energy kgm.m./gm. I I-III 1 164°. 4 .00688 .00990 3.04 3.04 1000 174 .1 646 955 3.03 2.96 2000 183 .5 603 924 2.98 2.86 3000 192 .6 561 898 2.91 2.74 4000 201 .4 518 875 2.80 2.60 5000 210 .1 476 857 2.68 2.45 GOOO 218 .6 434 843 2.53 2.27 * These values are subject to revision. EFFECTS OF PRESSURE ON NITRATES. 593 great that at 239° the transition pressure is below 500 kgm. Failure to find the transition means either that the transition point of the pure substance is appreciably higher than 219°, or else that II will support considerable superheating with respect to I. A special run ruled out the possibility that the transition from II to I is of the ice type and had been overlooked for that reason. I also made search for new modifications at room temperature to 12300 kgm. without success. During the first run I made measurements of the time rate of transi- tion and of the breadth of the band of indifference. The breadth of the band varied from 100 to 200 kgm. The transition velocity is unique in that the velocity with falling pressure is less than with rising pressure. This may be an effect of the free HNO3. The transition temperature II to III is given as 161° by Wallerant^ and 161°. 4 by Gossner.'^ Both of these values are lower than the value found above, 164.4°. It is significant that when I repeated the run, after the substance had become impure, I found a lower value than 164.4°, about 163°. It is therefore plausible to suppose that the lower values of Wallerant and Gossner may have been due to impurity. No previous measurements of the change of volume or of the heat of transition have been made. Thallium Nitrate. This salt was prepared by dissolving metallic Thallium in warmed dilute nitric acid, and by repeatedly crystallizing from aqueous solution. Preparation from the element was necessary because the nitrate is not kept in stock by any of the chemical supply houses. After each crystallization the crystals were washed; the end product was entirely free from acid reaction. The purification of this substance is particularly easy because of the great difference of solubility in hot and cold water; water at 100° dissolves about 100 times as much water at 0°. Only enough of the salt could be made for one filling of the apparatus. I used for this all the metallic Thal- lium available at the time in this country, and none was to be pro- cured from abroad. The salt was hammered cold into an open perforated steel shell, and pressure transmitted directly to it by kero- sene. With this single specimen a large number of readings were made. Measurements of this substance are of unusual difficulty because of the very small changes of volume and the large amount of super- pressure required to force the reaction to run. At least one unsuccess- 7 B. Gossner, ZS. Kryst. 38, 110-168 (1903). 594 BRIDGMAN. ful attempt was made before even the existence of the line II-III was estabUshed. Because of the unusual difficulties, a special procedure was necessary in getting the points of the II-III curve. After the reaction had started to run, and had been perhaps half completed, pressure was artificially changed in the direction of the suspected equilibrium, and after this change the direction of secondary pressure change noted. After each change of pressure it is necessary to wait only long enough to be sure of the direction of spontaneous pressure change before making another trial. It is thus possible to obtain results very much more quickly than by waiting for pressure to be brought back to the equilibrium value by the progress of the reaction, and in those cases in which Av is small, it is possible to shut the equili- brivmi pressure within much narrower limits. With the galvanom- eter and the bridge used the sensitiveness was great enough so that a change of j kgm. could be certainly detected. This made it possible to shut the equilibrium pressure within values differing by about 50 kgm. on the I-II curve, and about 100 kgm. on the II-III curve. One must be careful, when working with such small pressure changes, not to be deceived by the effects of heat of compression, which under proper conditions simulate the appearance of a reaction. The values of equilibrium pressure and temperature, shown in Figure 7 are fairly consistent. The values of AV, on the other hand, which are shown in Figure 8, were never satisfactory, although I used every wile at my command to get good values. The difficulty, which would be great enough in any event because of the small values of A^, is increased by the great sluggishness of the reaction between the two phases in each other's presence, as well as by the large superpressure required to start the reaction. If one tries to complete the reaction by running the pressure far beyond the transition point, error from hysteresis is introduced. Part of the difficulty is doubtless connected with the chemical instability of the salt itself. After the series of runs, some of the salt in the lower part of the cylinder was found changed to a brick red color. Thallium, of course, forms two series of salts, univalent and trivalent, and the trivalent form of some salts is the more stable. If there were a change from a thallous to a thallic salt, the reacting volume of the thallous salt would be less, and the change of volume less. If this change did take place, it must have been toward the end of the experiment, and would suggest a reason for the low points found for Av on the II-III curve at 11200 and 11700 kgm., and on the I-II curve at 7000 kgm. These points were measured with particular care, and there seems no doubt that they give accu- EFFECTS OF PRESSURE ON NITRATES. 595 rately the change of volume of whatever was in the apparatus at the time. The point at 11200 was measured in the regular way by de- creasing pressure at constant temperature; the point at 11700 was _ 3 4 5 6 7 8 9 10 11 12 Pressure, kgm./cm.^ x 10* Thaliium Nitrate Figure 7. Thallium Nitrate. The observed equilibrium pressures and temperatures. .0025 Id .0020 Q. .0015 E .0010 > .0005 U-l\ iliMi .000(^ 123456789 10 Pressure, kgm./cm.^ x 10* Thaliium Nitrate 'M 11 12 Figure 8. Thallium Nitrate. The observed changes of volume. obtained by changing temperature at constant volume, and is the mean of two points, practically the same, obtained with increasing and decreasing temperature. For this run, the temperature was lowered to 110°, corresponding to a superpressure of 6000 kgm., to be 596 BRIDGMAN. sure of completion of the reaction. The point at 7000 on the I-II curve is reallv three coincident points, two with falhng pressure, and one with rising pressure, at constant temperature. The pressure Hmits for this point were twice as wide as usual, to ensure completion of the reaction. , i • i There seems considerable reason to suspect that the high point on the upper end of the line for II-III belongs to another equilibrium line, and that there is a fourth modification. The consistently low points for the corresponding Av bear this out. These points, when compared with the low point on the upper end of the Av curve for I-ll, are too low in proportion to be entirely explained by a partial transi- tion to a thallic salt. This surmise was not verified; results would have been difficult of interpretation because of the known partial change of the thallous nitrate. It was also of no use to try for the usual data with the low pressure apparatus; the low pressure points are always made after the runs at high pressures, and m this- case it was known that the high pressure had at least initiated some sort of a permanent change. It may be mentioned that whatever the transi- tion product, it does not form mixed crystals with TINO3, for the p-t values were unaffected by the discordant Av values, and the tran- sition remained sharp throughout. Under the conditions, the best that I could do with the data was to draw a straight line through 23456 7 B9 Pressure, kgm./cm.' x 10^ Thallium Nitrate Figure 9. Thallium Nitrate. The computed heats of transition and the changes of internal energy. EFFECTS OF PRESSURE ON NITRATES. 597 the highest Av points. There are enough high points for each transi- tion to determine a pretty good Hne, and considering that practically all sources of error conspire to give a value of Av too low, there is considerable probability that the values adopted are near the truth. TABLE V. Thallium Nitrate. Pressure Temperature AV cms./gm. (It dp Latent Heat kgm.m./gm. Change of Energy kgm.m./gm. I-II 1 144°. 6 .00244 .00830 1.22 1.22 1000 152 .9 242 825 25 23 2000 161 .1 241 815 28 23 3000 169 .2 239 810 305 23 4000 177 .3 237 800 335 24 5000 185 .3 235 790 37 25 6000 193 .1 234 772 41 27 7000 200 .7 232 745 48 31 II-III 1 75°. 0 .00073 .00668 .38 .38 1000 81 .6 70 655 .38 .37 2000 88 .1 675 643 ..38 .365 3000 94 .5 65 632 .38 .357 4000 100 .8 625 624 .375 .350 5000 107 .0 60 616 .37 .34 6000 113 .1 57 611 .36 .325 7000 119 .2 55 606 .355 .315 8000 125 .3 52 603 .345 .30 9000 131 .3 495 600 .335 .29 10000 137 .3 47 596 .325 .275 11000 143 .2 445 .592 .313 .265 12000 149 .1 42 587 ..302 .25 The lat'mt heat and the change of energy, calculated with the values found above from the curves of t and Av against pressure are shown in Figure 9. The three modifications behave differently with 598 BRIDGMAN. respect to each other; AH and AE rising with increasing pressure on the I-II hne, and faUing on the II-III Une. The numerical results are collected in Table V. There are not many values for comparison. For the transition temperature I-II van Eyk ^ found 142.5°, Gossner '^ 151°, and Wal- lerant ^ 125°. For the transition temperature II-III Gossner gives 80° and Wallerant 80°. Van Eyk entirely overlooked this transition in his first paper; in a later paper ^ he corrected his earlier results and gives 72.8° by a dilatometric, and 79°-80° by an optical method. The values which I found above for the transition points were 144.6° and 75°. Although these were obtained by extrapolation, the range was small and the presumption is that they are at least as accurate as the other values. My values are means between limits reached by the reaction when running in opposite directions, both phases being present, whereas certainly Gossner and Wallerant did not shut their equilibrium point within values from the two sides, and sometimes give as the transition temperature merely the temperature of appearance of the new phase on heating. There seem to be no pre\ious measurements of the thermal effects of the transition, and the only measurement of change of volume is by van Eyk ^ who found 0.0004 ± cm.^/gm. for the transition II-III. The difference of compressibility of the phases may be found in the usual way from the difference of the slope of the isothermals above and below the transition. For this purpose only those runs were used which gave good values for Av. There are nine such good points for the change I-II. Of these, seven show no difference of compressi- bility and two show a slightly greater compressil^ility for II. The inference seems justified that within the limits of error the compressi- bility of I and II are the same. There are six good points for II-III, and all of these indicate that III is the more compressible. Further- more, these six points are all self consistent and show a decreasing difference of compressibility with rising pressure and temperature. At 75°, the compressibility of III, measured in cm.''' per gm. per kgm. per sq. cm., is 0.065 greater than that of II, at 105° it is O.OeS greater, and at 140°, 0.0615 greater. It is remarkable that the two phases with the smaller difference of volume should have the greater differ- ence of compressibility. Combining the difference of compressibility with the change of volume along the transition line, the difference 8 C. van Eyk, ZS. phvs. Cliem. 30, 4,30-4.59 (1899). 9 C. van Eyk, ZS. phys. Chem. 51, 721-731 (1905). EFFECTS OF PRESSURE ON NITRATES. 599 of expansion may be computed in the regular way. Ill turns out to be more expansible; the average difference over a range of 10000 kgm. is 0.00005 cm.^ per gm. per °C. Most careful search was made for other modifications to 12500 kgm. at 20°, 100°, 150°, and 200°, but no transitions except those found above were detected. Potassium Nitrate. — This substance was Kahlbaum's purest, used without further purification. It was fused into an open steel shell, thus effectually driving away the moisture, and then the sides of the shell were pierced with fine holes to ensure equality of pressure in all parts. Pressure was transmitted directly by kerosene. Two samples were used. With the first, 39 gm. in amount, all the high pressure points were determined, 21 in number. The second sample, weighing 32 gm., gave several points at low pressures. Considerable trouble was found in getting the Av points, and several repetitions were necessary. The discrepancies were always in the direction of too low values. The reason is probably to be found in a combination of the considerable lag which the transition shows, with the effect of fissures, which formed in great numbers on solidification from the melt. The effect of lag is that it is necessary to push the pressure considerably beyond the transition line to ensure the reaction starting, and the effect of fissures may be to prevent the reaction from running completely throughout the entire mass after it has once started in some parts. To be sure that the transition is complete, the pressure must be pushed far enough beyond the transition line to ensure kernels of the new modification appearing in all the parts. The point at 0°, for example, had to be repeated for this reason. To ensure completion of the reaction, when the point was repeated, pressure was maintained 1000 kgm. beyond the equilibrium line at room temperature for three hours, and only then was the temperature lowered to 0°. The first value of Av, before this procedure was adopted was 25% too low. The sluggishness of the reaction sometimes neces- sitated special procedure. Thus the lower point on the II-III curve was determined by maintaining temperature for some time at 90° to ensure completion of the reaction from II to III, and then lowering temperature at such a pressure as to land in the corner between the II-III and the III-IV lines. The lag varies considerably with the temperature on the several curves. Thus on the upper end of the I I-III curve, III could be considerably subcooled with respect to II, but II could be very little superheated with respect to III. The sluggishness of the reaction made it impossible to obtain good 600 BRIDGMAN. values for the change of volume between I and III or between II and III at low pressures. Three sets of measurements were made around the triple point I-II-III. The first gave the change of volume I-II by the method of changing temperature at constant pressure. The temperature could be raised so far beyond the transition line that the reaction from II to I was surely complete, and of course it was not ■difficult to be sure of starting with pure II at low temperature. But this method does not give the temperature of the transition point, and this transition temperature was especially needed in fixing the position of the triple point. To fix this point, pressure and tempera- ture values were obtained at low pressures on the I-III and the II-III lines, by the method of varying pressure at constant temperature. It was possible to shut the pressure within narrow limits, 20 to 30 kgm., but the domain of subcooling was so wide that the reaction was not completed even on reducing pressure to nearly atmospheric. The directly found values of AV at these points were 10 to 15% too low. These points were, therefore, discarded. The AV curves for I-III and II-III at low pressures were determined by extrapolation from the high pressure values together with the condition that the changes of volume should check with that for I-II at the triple point. In spite of the comparatively large amount of lag, the width of the band of indifference was not abnormal. No absolute value can be attached to these figures, because of the disturbing effect of fissures. The limits were notably wider on repeating a point, probal)ly because of the fissures formed by the intervening transitions, and were very much less in the low than in the high pressure apparatus. The figures pretend only to give roughly a comparative estimate. The band was widest for the II-IV curve, about 250 kgm.; on the II-III curve it varied from 200 kgm. at the lowest to 100 kgm. at the highest tempera- ture; on the I-III curve the variation with rising temperature was from 100 to 30 kgm., and on the III-IV curve from 140 to 0 kgm. with rising pressure. On all the curves, the band becomes narrower with rising temperature. The reaction, as usual, ran more rapidly at the high temperatures, and at least on the II-III and III-IV curves the transition ran more rapidly in the direction of falling pressure. The experimental values of pressure and temperature are shown in Figure 10, the experimental values of AV in Figure 11, the computed values of AH and AE in Figure 12, and the numerical results in Table VI. All of the pressure-temperature determinations are shown on the curves, but four of the AV points have been discarded for reasons already given. 123466789 Pressure, kgm./cm.^ x 10* Potassium Nitrate Figure 10. Potassium Nitrate. The observed equilibrium pressures and temperatures. 123456789 Pressure, kgm./cm.^ x 10^ X Potassium Nitrate Figure U. Potassium Nitrate. The observed changes of volume. 601 602 BRIDGMAN. The interesting features of the diagram, apart from the two new modifications, are the pronounced curvature of the II-III hne, and the great proximity of the I-II-III triple point to atmospheric pres- sure. It suggests itself that it might be possible under the proper conditions to realize III at atmospheric pressure as an unstable form, and as a matter of fact Wallerant has found an unstable form on 23456789 Pressure, kgm./cm.^ x 10^ Potassium Nitrate Figure 12. Potassium Nitrate. The computed heats of transition and the changes of internal energy. cooling I l)elow the triple point to al)out 114°. This unstable form is uniaxial, but Wallerant was not able to further characterize its crystal- lographical system. There are a large number of determinations at atmospheric pressure. For the transition temperature Bellati and Romanese ^° give the limits 128.7° and 121.7° on heating and cooling, Bellati and Lussana ^^ give 127.76°, Schwarz ^ gives the limits 121.5° to 129.5° by a thermal 10 M. Bellati and R. Romanese, R. 1st. Ven. 3, 653-669 (1884-85). 11 M. Bellati and S. Lussana, R. 1st. Ven. 2, 995-1023 (1890-91). EFFECTS OF PRESSURE ON NITRATES. 003 method and 129.5° by an optical method, van Eyk ^ gives the limits 125° to 127.8° on heating and cooling, Hissink ^ gives 127.9°, and Wallerant ^ gives 126°. The value found above by extrapolation was 125.8°. This value is probably as near the truth as any of the TABLE VI. Potassium Nitrate (KXO3). Pressure Temperature AY cm3./gin. Latent Heat kgm.m./gm. Change of Energy kgm.m./gm. I -III 1 125°. 8 0.01424 .02177 2.609 2.609 1000 147 .1 1.394 2091 2.800 2.661 2000 167 .7 1370 2038 2.964 2.690 .3000 187 .9 1360 2012 3.115 2.707 4000 207 .9 1354 1995 3.264 2.722 I-II 1 127 .7 .0060 .00486 4.95 4.95 II-III 510 120 .0 .01050 -.0221 1.867 1.921 1.300 100 .0 1286 290 1.654 1.821 1900 80 .0 1410 383 1.303 1..571 2350 60 .0 1480 509 .968 1.316 2690 40 .0 1.528 667 .717 1.128 2955 20 .0 1560 841 .544 1.005 II-IV 2665 o°.o .04474 .08032 1.522 ..330 2790 10 .0 4440 " 1.565 .326 2920 20 .0 4410 " 1.609 .322 604 BRIDGMAN. Table VI, Continued Pressure Temperature AV cm^./gm. dr dp Latent Heat kgm m./gm. Change of Energy kgm.m./gm. III-IV 3000 23°. 9 .02830 • .03844 2.187 1.338 4000 61 .4 .02740 .03655 2.508 1.412 5000 96 .8 2680 3438 2.883 1.543 6000 130 .0 2636 3210 3.311 1.729 7000 160 .9 2582 2964 3.780 1.956 8000 189 .2 2540 2680 4.381 2.349 9000 214 .5 2500 2350 5.186 2.936 Triple Point, II-III-IV II-III .01560 -.0830 .553 1.010 2930 21 .3 III-IV .02840 II-IV .03860 2 . 166 1.334 .04400 .08032 1.613 .324 Triple Point, I-II-III I-III .01420 .02166 2.631 2.615 115 128°. 3 II-III .00886 I-II .0200 1.778 1.788 .00534 .00486 4.409 4.403 others, because the extrapolation from the triple point is a very short one and does not admit of much question, and the transitions above the triple point, on which the extrapolation is based, do not show the lag effects so prominently as at atmospheric pressure. There seem to be no values of AV at atmospheric pressure for comparison. Bellati and Romanese ^° give for the latent heat of the transition 11.89 cal. against 11.6 cal. found above. The agreement is closer than usual. There seem to be no previous measurements of the effect of pressure on the transition. EFFECTS OF PRESSURE ON NITRATES. 605 Several direct determinations of the difTerence of compressibility were made, which were not, however, very accurate. At the triple point I-II-III, there are no data for the computation, and at the triple point II-III-IV the values are not accurate enough. This much may, however, be stated with some confidence. Ill is more compressible than I; the differe ice is of the order of O.O5I, and proba- bly the diflference decreases with increasing pressure: III is also more expansible than I, and the difference is of the order of O.O46. The specific heat of I is less than that of III, the difference being about 1 kgm. cm. per gm. II is probably less compressible than III, and it is probable that the difference of expansion between II and III changes sign along the transition curve; at the higher temperature II is less expansible, and at the lower temperatures more expansible. Ill is more compressible than IV and the difference is of the order of O.O5I. There is also a direct experimental determination of the difference of specific heats of I and II by Bellati and Romanese,^° at 0.014 cal., but the authors themselves say that the difference of specific heat cannot be expected to be accurate. Search was made for other modifications to 12000 kgm. at 20° and 200°, but without success. Ammonium Nitrate, The material used was Kahlbaum's purest, "zur Analyse," obtained from Eimer and Amend. Two series of runs were made, separated by an interval of a year. For each of these series of runs the material was hammered dry into a perforated steel shell, and the pressure transmitted directly to it by kerosene. For the second run, the salt was subjected to a preliminary drying in vacuum at 100°. No difference in behavior between the two runs to be attributed to this cause was to be detected. The second series, repeating the first, was made necessary by the fact that the measure- ments of Av given by the first series were irregular. The reason for this is that ammonium nitrate is a substance for which it is particu- larly difficult to force the reaction from one phase to another to run to completion. It is very curious that a reaction which will start to run with only a little subcooling or superheating will run to completion only with a much greater subcooling or superheating. I have found this same behavior several times with other substances. It has also been observed for ammonium nitrate by Behn; ^^ he found, for example, that for the transition IV-V it was necessary to lower the temperature fully 30° beyond the transition temperature before 12 U. Behn, Proc. Roy. Soc. 80, 444-457 (1907-08). 606 BRIDGMAN. he could be sure that the reaction was complete, and on raising the temperature he found the same lag also. The phenomenon is more unusual with rising temperature. Upon the repetition of the experi- ment, error from this effect was avoided only by very careful work, profiting by the experience of the first run. It is necessary to run the pressure several thousand kilograms beyond the equilibrium point in order to be sure of the completion of the reaction. The completion of the reaction is also greatly facilitated by running the pressure back and forth several times over the transition before beginning the measurements. In the neighborhood of the triple point II-III-IV, where it is not possible to secure completion of the reaction by running the pressure over a considerable range, because the pressure of equili- brium is too near atmospheric pressure, the same end was attained by raising or lowering the temperature, as occasion might require, and then bringing it back to the temperature of the measurement. I had 3456789 10 Pressure, kgfm./cm.^ x 10* Ammonium Nitrate U 12 Figure 13. Ammonium Nitrate. The observed equilibrium pressures and temperatures. EFFECTS OF PRESSURE ON NITRATES. 607 not thought of this trick of procedure at the time the measurements in the neighborhood at the triple points I-II-IV-VI were made, and some error is introduced into the actual data which it would have been possible to avoid if the measurements had been repeated. However, the values of the changes of volume at these points seem to be indi- cated with sufficient certainty. At low pressures, where the method is that of varying temperature at constant pressure, lag is not so troublesome, and repetition was not necessary. The low pressure point on the curve I-II was not repeated; the repeated points at low pressure for the transitions II-III and III-IV were in substantial agreement with the first determinations. The experimental \'alues for the eciuilibrium pressures and tempera- tures and the transition curves are shown in Figure 13, the experi- mental points and the curves for the change of volume in Figure 14, the computed values for the latent heats and the change of internal energy in Figure 15, and the numerical values are collected in Table VII. In the equilibrium diagram the points of both runs, in 1913 and 1914, have been included, but in the change of volume diagram, only the points of 1914 have been given. The irregularities of the early points for Av were so great that it was almost hopeless to attempt to obtain any information from them. Several features of these diagrams require special comment. The phenomenal steepness of the melting line is to be noticed. The point on the melting line at 200° was determined experimentally, but for the point at atmospheric pressure I have adopted the value of Lehmann.^^ An attempt to determine this point directly did not give satisfactory results, because of decomposition. This decomposition could be plainly detected at 200°, producing a rounding of the corners of the curves, so that the value for the change of volume cannot be counted on with much certainty. At the melting point at atmospheric pres- sure the decomposition was so much greater that it did not seem worth while to attempt to calculate a probable value for the change of volume at atmospheric pressure. The decomposition at atmospheric pressure has been found by other observers also; Behn ^^ states that it is noticeable at temperatures as low as 100°. It would seem, how- ever, that this decomposition is an effect chiefly found only when the liquid phase is present, or at least that it is effectively prevented by the higher pressures, because no trace of it was ever found on any of the transition curves between solids, even as high as 200°. 13 O. Lehman, ZS. Kryst. 1, 97-131, (1877). 4 5 6 7 8 9 10 Pressure, kgm./cm.' x 10^ Ammonium Nitrate Figure 14. Ammonium Nitrate. The observed changes of volume. % Hi tf^" '""'''"''''' 2 3 4 5 6 7 Pressure, kgm./cm.' x 10 Ammonium Nitrate 8 9 10 U 12 Figure 15. Ammonium Nitrate. The computed heats of transition and the changes of internal energy. EFFECTS OF PRESSURE ON NITRATES. 609 The curve drawn for the change of Nohrnie between II and VI perhaps requires special comment. This has been drawn as falHng with rising pressure, although the points actually obtained would indicate the reverse state of affairs. But the relations at the two triple points terminating this curve demand that the curve shall be as TABLE VII. Ammonium Nitrate. Pressure Temperature AV cm^./gm. dp Latent Heat kgm.m./gin. Change of Energy kgna.m./gm. Liquid - 1 1 168° .034 6.9 6.6 1000 202° .051± I-II 1 125°. 5 .01351 .00974 5.53 5.53 1000 134 .7 .01171 875 5.46 5.35 2000 143 .0 .01028 797 5.37 5.16 3000 150 .7 .00916 738 5.26 4.98 4000 157 .8 828 691 5.16 4.83 5000 164 .5 751 651 5.05 4.67 6000 170 .8 679 613 4.92 4.51 7000 176 .7 610 562 4.88 4.45 8000 182 .0 542 496 4.97 4.54 9000 186 .6 476 409 5.35 4.92 II-III 1 82°. 7 -.00758 -.0159 1.70 1.70 200 79 .2 797 190 1.48 1.50 400 75 .1 836 221 1.32 1.35 600 70 .4 874 253 1.19 1.24 soo 65 .0 913 284 1.09 1,10 ()l(l BRIDGMAN. Tabic VII, Cnntiinied. Pressure Temperature AV cm^./gm. dr Jp Latent Heat kgrn-m./gm. Change of Energy kgm.ixi. /gm. II-IV 1000 65°. 3 .01210 .0146 2.81 2.69 2000 79 .6 1211 141 3.04 2.80 3000 93 .4 1212 136 3.27 2.91 4000 106 .8 1215 131 3.52 3.03 5000 119 .7 1220 127 3.77 3.16 GOOO 132 .2 1228 123 4.05 3.31 • 7000 144 .3 1238 119 4.34 3.48 8000 156 .0 1250 115 4.66 3.66 9000 167 .4 1265 111 5.01 3.87 III-IV 1 32°. 0 . 02026 .0311 1.99 1.99 200 38 .5 2051 336 1.91 1.86 400 45 .4 2077 360 1.84 1.76 600 52 .9 2102 3S5 1.78 1.65 800 60 .8 2128 410 1.73 1.56 IV-V 1 -18°. 0 -.017± -.063 0.69 0.69 I-VI 9000 186°. 6 .00858 5.26 4.49 10000 194 .1 785 .00750 4.89 4.11 11000 201 .6 740 4.68 3.87 II-VI 9154 170°.0 .00312 -.125 .111 .296 9034 185 .0 373 .137 .474 EFFECTS OF PRESSURE ON NITRATES. Table VII, Continued. ()11 Pressure Temperature AV cm3./gm. dr dp Latent Heat kgm.m/.gm. Change of Energy kgm.m. gm. VI-IV 9000 167°. 9 .00959 5.16 4.29 10000 176 .1 955 5.23 4.28 11000 184 .3 952 .00820 5.31 4.26 12000 192 .5 950 5.40 4.25 Triple Point, II-III-IV III-IV .02135 .0417 1.72 1.54 860 63°. 3 II-III .00925 .0294 1.06 1.14 II-IV .01210 .0146 2.78 2.68 Triple Point, I-II-VI. I-II .00475 .00406 5.38 4.95 9020 186°. 7 I-VI .00855 II-VI .00750 5.24 4.47 .00380 -.125 .14 -.48 Triple Point, II-IV-VI II-IV .01267 .0111 5.06 3.91 9160 169°. 2 II-VI .00309 IV-VI -.125 .11 .38 . 00958 .00820 5.17 4.29 612 BRIDGMAN. drawn, and considerations of the effect of incompleteness of the reaction also indicate that the curve as drawn corresponds to the actual facts. Thus, the change of volume between VI and IV found at 173° is too small to lie on the curve through the points at higher pressures by precisely the same amount that the change of volume II-VI at the same temperature is too high to fall in with the values demanded by the necessary conditions at the triple points. The explanation is obviously that the reaction from IV to VI was not complete because it was not possible to lower the pressure far enough, and that the reaction was completed when the pressure was carried across the line VI-II, the rest of the change from IV to VI appearing as part of the change from VI to II. The data for the transition IV-V shown in the table and the diagram were taken directly from the paper of Behn.-^^ Because of the great stickiness of the reaction it would have been necessary to set up special apparatus for reaching temperatures at least as low as — 50° if accu- rate measurements of this transition were to be expected, and it did not seem worth while to do this. The slope of the transition line IV-V was not given by Behn, but has been computed by me from his data. Behn states that he was not able to get values for the change of volume which were self consistent to better than 20%, so the value given for the slope may be in error by as much as this. The results for ammonium nitrate are different in one or two inter- esting particulars from those found for most other solid transitions. The change of volume II-IV increases with rising pressure instead of decreasing, as it has for every other substance previously investigated. The change of volume between III and IV also shows the same effect. This is probably to be attributed to unusual properties of IV, as will be brought out in a later discussion of the relations of the compressibili- ties, thermal expansions, and specific heats. It is also very unusual, although this is not the only example, that the III-IV line is concave upwards. That this is true is indicated by several independent lines of evidence. A good deal of work has been done on this substance by other experimenters at atmospheric pressure, and there are at least two pieces of work on the effects of pressure. The existence of four modifications at atmospheric pressure has been known for some time, but it is not so commonly mentioned that there is a fifth modification. This seems to have been first discovered by Lehmann, was then studied by Wallerant,^ and more lately by Behn^^; its existence seems absolutely beyond question. The problems offered by this low EFFECTS OF PRESSURE ON NITRATES. 013 temperature form are of great interest. Wallerant was able to find that its crystalline system was the same as that of II, and he suggested that it was in fact the same modification. This appearance of the same phase in two isolated parts of the phase diagram is unique, if it should be proved, but seems to be not contrary to any of the thermo- dynamic necessities. Behn has subjected the question to a very careful experimental examination, measuring the various thermal and crystallographical properties of II and V to see whether it is possible to state an indentity. His data did not allow of a definite answer to the question, but seem to suggest that II and V are not the same phase. I did not make any measurements on this transition for the reasons already stated, but did investigate one questionable point. One might expect, since the transition IV-V is of the ice type, that at higher pressures another phase would exist, so that the falling transi- tion curve between IV and V would be replaced by a rising curve between IV and the new modification. I did not find any such new modification out to 12500 kgm. at room temperature. The following transition temperatures are found in the literature. For I-II Lelimann ^^ gives 127°, Bellati and Romanese ^* give the limits 124° to 125°, Lussana ^^ found the limits 124.9° and 125.6°, and Schwarz ^ has found the value 125.6° by one method, and the limits 123.5° to 125.5° by another. In the table I have given 125.5° as agreeing best with my own data and the best results of other observ- ers. For the transition II-III Lehmann ^^ gives 87°, Behn ^^ gives 83°, Bellati and Romanese ^* give the limits 82.5° to 86°, Schwarz ^ gives the limits 82.5° to 86.5° by one method, 82.8° to 82.7° by another, and 83° by a third, Wallerant ^ gives 82°, Tammann ^^ gives by extrapolation from his measurements at higher pressures the value 84.6°, and Lussana ^^ finds that the transition takes place at 85.85° on heating, but does not give the value on cooling. I have adopted the value 82.7° as agreeing best with my own data and those of others. For the transition III-IV Lehmann ^^ gives 36°, Bellati and Romanese^* give the limits 31° to 35°, Schwarz^ gives the limits 31° to 35° by one method, 32.4° (best) by another, and 35° by a third, Wallerant ^ gives 32°, Behn ^^ 32°, Tammann ^^ finds by extrapola- tion from his high pressure measurements 31.8°, and Lussana ^^ finds the limits 35.45° and 30.55° on heating and cooling. I have adopted U M. Bellati and R. Romanese, R. 1st. Ven. 4, 1395-1420 (1885-86). 15 S. Lusgana, Nuov. Cim. (4), 1, 97-108 (1895). 16 G. Tammann, Kristallisieren unci Schmelzen, Barth, Leipzig (1903), p. 299. 614 BRIDGMAN. the value 32.0°. There are only three measurements of the transi- tion IV-V. Lehmann (in a personal letter quoted by Behn) ^^ gives — 16°, Wallerant gives the limits — 14° to — 16°, and Behn finds — 18°. I have adopted this last value as more probably exact. For the melt- ing point I have adopted the value of Lehmann, 168°, which is some- what higher than the value given by Behn, 166°, as the average of several other determinations. As already stated, I made no attempt to measure this point myself. The change of volume at atmospheric pressure has been measured by Bellati and Romanese ^* for the transitions JV-III and II-III. For III-IV they give 0.01964 cm.^ per gm., but from a recomputation of their data I should prefer the value 0.01955, although their data show enough irregularity to admit a value as low as 0.01915. Behn ^^ gives for this transition 0.022 (he pretends to give only two significant figures), and I find by extrapolation from 77 kgm. (the extrapolation correction is less than one percent) the value 0.02026. For the transition II-III Bellati and Romanese give —0.00854 cm.^ per gm., Behn gives 0.0081, and I find by extrapolation from 77 kgm. a some- what lower value, 0.00758, the total extrapolation correction in this case being about 3%. The change of volume I-II was not measured by Bellati and Romanese. Behn gives 0.0132, and I find by an extra- polation from 77 kgm. 0.01351. Lussana ^^ did not measure the change of volume, but calculates from his other data for the pressure effect the value 0.01465, assuming Bellati and Romanese's value for the latent heat. The change of volume IV-V seems to have been measured only by Behn, who found that it lies between the values 0.0155 and 0.0185, but was not able to get it more accurately because of the lag already mentioned. The latent heats of transition have been measured only by Bellati and Romanese/* who give for III-IV 5.02 gm. cal. per gm., for III-II 5.33 cal., and for II-I 11.86 cal. I find by calculation from Clapeyron's equation the values 4.66, 4.00, and 13.0 cal. respectively. The only measurements of the eft'ect of pressure are by Lussana ^^ and Tammann.^® Tammann's measurements extended to 2800 kgm. and include points on the II-III, III-IV, and II-IV curves, but he made no measurements on the phase I. Lussana's pressure range was only 250 kgm.; he measured the three transitions III-IV, III-II, and II-I. Tammann states that there is no possibility of agreement between his values, those of Lussana, and those computed from the thermodynamic data of Bellati and Romanese. The discrepancy stated by Tammann as existing between Lussana's data and those to EFFECTS OF PRESSURE ON NITRATES. G15 be computed from the data of Bellati and Romanese appears to be due to an error of Tammann's in converting Lussana's data, which are given in atmospheres, to kilograms. Lussana himself states that the agreement of his experimental values with the computed values is almost perfect, and I have verified this, at least within one of two percent. The apparent discrepancy between Tammann's and Lussana's observations may furthermore be at least partly explained by the difference in the pressure range, because of the fact that the slope of the transition lines III-IV and III-II changes markedly with pressure. The slope of the line II-III decreases algebraically, the normal direction of variation, but the slope of the line III-IV increases with rising pressure, a somewhat unusual effect. These two equili- brium lines are so short that it would be a matter of considerable experimental difficulty to accurately measure the variation of slope with pressure. That there is a variation with pressure was first suggested to me by an examination of my own equilibrium points. Furthermore, on making the computations, it appeared that such a variation was imperatively demanded by the conditions at the triple point. With the help of the known relations at the triple point it was possible to get a fairly good idea of what the actual variation of slope must be. The values given in the Table were computed so as to satisfy the conditions at the triple point, and with no other condi- tions in mind. In fact, I did not know of the discrepancy between the values of Lussana and Tammann until after I had made the computations. All of the discrepancy between these observers cannot be explained in this way, however. Lussana gives for the slope of the transition III-IV over a range of 2.50 kgm. 0.0287, Tam- mann giAcs over the range up to the triple point, that is up to 860 kgm., the average slope 0.0346. I find for the average slope up to the triple point 0.0364, and for the average over the first 2.50 kgm. 0.0318. The variation of slope with pressure is fairly large and will explain a large part of the discrepancy between Tammann and Lussana, although it will evidently not account for all of it. The agreement of Tammann's and my results is better than that of Tammann and Lussana. It must be said that Tammann did detect the curvature of the transition lines, and states that the II-III curve is concave downwards and the III-IV concave upwards, but he makes no allowance for this fact in comparing his results with those of Lussana. In spite of the sur- prisingly good agreement found by Lussana between his results and those computed from the thermal data of Bellati and Romanese, this agreement must be recognized as accidental, and not necessarily 616 BRIDGMAN. indicating the correctness of the data of either BeUati and Ronianese or of Lussana. Lussana's method was not one adapted to give accurate results. He measured the time rate of heating or cooling and took the arrest points as the points of transition. These points differed by 5°, and the mean of the two points was taken as the true temperature of transition. The total effect of pressure over the range of 250 kgm. was little more than the width of the band of indifference of the individual measurements. The insensitiveness of the method is indicated by the fact that he found no variation of slope with pres- sure, whereas the variation ought to haAC been easily perceptible even over this low range. Lussana nevertheless, gives the results to three significant figures. On the III-IV curve the conditions are similar to those on the II-III curve. Lussana gives for the slope over the first 250 kgm. 0.0137 against 0.0135 computed, and Tammann gives for the range up to the triple point the average slope 0.0220. I find for the average slope over the first 250 kgm. 0.0178, and for the average slope up to the triple point 0.0226, in sul)stantial agreement with Tammann. A thoroughgoing comparison of my results with those of Tammann is not possible, because he has not taken any account of the known curvature of the transition line in computing the values at atmospheric pressiu'e. It is evident, howe^•er, that both his results and mine would difter in the same direction from the only other data at at- mospheric pressure, namely those of Bellati and Romanese, and also from those of Lussana. It seems probable that there may be some considerable experimental error in the data of Bellati and Romanese, and that the curvature of the III-IV cur\e, although large, can ac- count for only part of the discrepancies. The coordinates that Tammann finds for the triple point II-III-IV are considerably diflferent from mine. He finds 64.16° and 930 kgm. against my values 63.3° and 860 kgm. For the average slope of the II-IV line between the triple point and 90° Tammann finds 0.01404 and I find 0.01416. The agreement is better than usual. The pressure effect on the transition I-II has been previously measured only by Lussana. He finds 0.0116 for the average slope over the first 250 kgm. against my value 0.00962. The measurements of the difference of compressibility of the differ- ent phases at high pressures did not give any ver}- regular results. This much may be stated, however; I is more compressible than II, of the order O.OeS, and is more compressible than VI of the order O.OeS, and IV is more compressible than II of the order 0.068. This last is an EFFECTS OF PRESSURE ON NITRATES. 017 unusual feature, because IV is the phase stable at higher pressures and lower temperatures. That the relation of the compressibilities of II and IV is unusual might be also suggested by the fact, already men- tioned, that the change of volume becomes greater at the higher pressures. The other differences between I, II, IV, and VI may be obtained by a combination of the values given above. The experi- mental data were not accurate enough to allow a statement as to the variation of difference of compressibility with pressure on the transi- tion line; the above values are average values. No direct values were obtained on the short lines II-III and III-IV. I was also able to obtain fairly satisfactory measurements of the difference of thermal expansion of III, IV, and II at 77 kgm. II is more expansible than III, the difference being 0.000038 cm.^ per gni., and IV is more expansible than III by 0.000115. Here again, the behavior of IV is anomalous, being more expansible than a phase stable at a higher temperature. The thermal expansions at atmos- pheric pressure have been measured by Bellati and Romanese ^* and by Behn.^^ One may deduce from the data of Bellati and Romanese, although the authors themselves do not give the computa- tion, that the expansion of IV is 0.000222 cm.^ per gm., that of III 0.000134, and that of II 0.0001165. The differences between these vahies do not agree at all with the values which I have found, in fact they would make II less instead of more expansible than III. The values of Behn I have computed from a diagram which he gives; he himself does not give the computations. From his data I would make the expansion of II 0.00041 cm.^ per gm., that of III 0.00036, that of IV 0.00048, and that of V 0.00037. The differences between II and III and between III and IV will be seen to agree remarkably well with the values which I found above. There seems to be no doubt that IV is more expansible than III, and that II is more ex- pansible than III. The specific heats at atmospheric pressure has been measured by Bellati and Romanese.^* They find for IV 0.407 cal. per gm., for III 0.355 cal., and II 0.426. Here again IV is abnormal, its specific heat being greater instead of less than that of the high temperature phase. If now we compute in the regular way the differences of compressi- bility and specific heats at atmospheric pressure, using the values which may be deduced from Table VII for the variation of the change of volume and the latent heat with pressure along the transition curve, and assuming my direct experimental values for the differences of thermal expansion, we shall find that II is more compressible than 618 BRIDGMAN. Ill by O.O5I3, and III is less compressible than IV by O.OgSS. This verifies the abnormally high compressibility of IV again. Computa- tions made in the same way give for the difference of specific heats between II and III 0.16 cal. against Bellati and Romanese's ^^ direct value 0.071, and for the difference IlI-IV 0.048 cal. against their value 0.052. This verifies the abnormal sequence of the specific heats Il-llI-IV, and the numerical values agree as well as one could expect when one considers that the method of computation employed above throws by far the larger part of the error into the specific heats. This has already been fully explained in a preceding paper. The question of the time rate of the various transitions and the lag phenomena will be dealt with in greater detail in another paper. It will not be out of place to mention here, however, that many meas- urements were made of the time rate, and at all points it was found that the rate is greater with decreasing pressure, as has also been foiuid for most other substances. Nitrates with no New Forms. The following nitrates were examined without result for other forms between 1 and 12000 kg. at 20° and 200°; NaNOg, LiNOs+Aq, Hg2(N03)o, Hg(N03)2, Pb(N03)2, A1(N03)3. Sodium, although chemically similar in many respects to potassium and the other alkali metals whose nitrates are polymorphic, is in its crystalline properties known to be so much dissimilar that its salts do not form isomorphic mixtures. From this point of view, it is therefore not surprising that no new forms were fovmd. NaNOs (which is trigonal) does, however, crystallize isomorphously with the high temperature form (trigonal) of AgNOs, and will crystallize in small percentages isomorphously with the low temperature form. We suspect, therefore, that NaN03 has a second modification, stable at low temperatures at atmospheric pressm-e. It would be interesting to search for this. Hissink ^ has found no new form between 270° and — 50°. The transition, if it exists, is by analogy with AgNOs of the ice type, the transition curve falling with rising pressure. It is natural that no new form was found at high pressures at room temperature. LiN03 is probably not isomorphous with AgNOs or NaNOa, al- though Retgers says that it is with the latter. There is no necessity for expecting other forms. Arzruni, however, on authority which goes EFFECTS OF PRESSURE ON NITRATES. 619 back to 1857/^ says that LiNOs is trimorphic. Other authorities, Groth included, do not mention this, however, and the supposed polymorphism is doubtless to be explained as due to the transition points of the hydrates, because LiNOs crystallizes with H2O. I made especially careful search for other forms, isothermally at 20° and 200° to 12000 kg., and at constant volume between 100° and 160° at ap- proximately 3000 kg. and between 30° and 160° at 8000 kg. Hg2(N03)2+H20 does not crystallize isomorphously with the other univalent nitrates. The anliydrous salt showed no new form at 20° or 90°. At the higher temperature there was probably some decom- position; temperature was not raised higher for fear of further de- composition. The salt crystallizing with one molecule of water probably has a transition point at 20° at 8000 kg. ; no other tempera- ture was tried. The substance is a difficult one to investigate, because beside containing water of crystallization it is hygroscopic. It is difficult to remove the absorbed moisture without dehydration. The material deserves further investigation, however. Hg(N03)2 was the first of the nitrates tried which was not univa- lent. I know of no cases of polymorphism among this class; there is no particular reason to expect other forms at high pressures. Hg(N03)2 showed nothing new at 20° to 12000; on heating to 180° at 1000 kg. it decomposed with almost explosive violence, depositing metallic mercury throughout the apparatus. Pb(N03)2 is stated by xA-rzruni ^^ to exist as a mineral in two forms. I found no new form up to 12000, however, at 20° or 200°. A1(N03)3 was tried at 20°, 100°, and 200° with no effect. After the run it was found to have decomposed into a brownish mass, actively deliquescent. It was somewhat moist before the run. Discussion. IVIost of the facts described in this paper are collected and exliibited qualitatively in Figure 16, which shows the phase diagrams, pressure against temperature, and indicates other data also. The drawings are made with some exaggeration, so that the direction of curvature of each transition line may be readily' noticed. The arrows on the lines show the directions in which Av decreases numerically. An a, (3, 17 P. Kremers, Pogg. Ann. 92, 520 (1854) and 93, 23 (1854). 18 A. Arzruni, Phj'sikalische Chemie der Krystalle. Vieweg und Sohn, Braunschweig (189-3), p. 42. 620 BRIDGMAN. or Cp placed near a transition line, on one side, shows that the phase on that side of the line has the greater compressibility, thermal expan- sion, or specific heat. In addition, the Arabic numerals placed in the regions of the several phases show the crystalline systems. The Roman numerals give the designation of the phases used in the early part of this paper. The information about the crystalline 4 G 8 10 0 2 4 i; S 10 0 2 4 6 Pressure in Thousands of Kilograms per Sciuare Centimeter Figure 16. Collection in one diagram of the several phase diagrams. The Arabic numerals refer to the crystalline systems as foUows: 1, Quasi-trigonal. 2, Orthorhombic, quasi-tetragonal, optically negative. 3, Orthorhombic, quasi-tetragonal, optically negative. 4, Monoclinic, quasi-tetragonal, optically positive. 5, Tetragonal, optically positive. 6, Rhombohedric, quasi-cubic, optically positive. 7, Cubic. 8, Rhombohedric, type of calcite, optically negative. systems has been taken directly from Wallerant's Cristallographie, a work which deserves to be better known in this country. It is evident in the first place that AgNOs belongs in a class apart from the other five univalent nitrates. Its phase diagram is entirely different in character, and it is known not to crystallize isomorphously I EFFECTS OF PRESSURE ON NITRATES. ()21 with any of them. AgNOa may be disregarded in any comparative study, therefore, and in the following only the five alkaline nitrates are considered. A comparison of the phase diagrams is more significant when we combine with it a discussion of the mixed crystal relations. Wallerant has given in his book a most elaborate account of the mixed crystal situation, and I shall draw my data from him. One conclusion to be drawn from this discussion will be that simi- larity of phase diagrams is evidence of the most complete similarity of structure, more complete even than the ability to form continuous series of mixed crystals. Judged by this latter test, the five alkaline nitrates are very similar. Wallerant has proved that every one of the five can form mixed crystals in limited proportions with every one of the atmospheric polymorphic forms of all the other nitrates. The phase diagrams are not the same, however, for all five nitrates; those of Rubidium, Caesium, and Thallium are most similar. Each of these three diagrams contains a transition line between the cubic form (7) and the orthorhombic, pseudo-hexagonal, form (6). The use of the same numerals for these phases in the different diagrams is justified by the fact that within the proper temperature range the phase 7 or 6 of any of the three salts will form a continuous series of mixed crystals with the phase 7 or 6 of any of the others. We may call the line 7-6 the "same" line in the various diagrams. This line preserves various of its characteristics with very little change from substance to substance, as is shown by the summary in Table VIII. The lines 7-6 correspond so closely that we expect further similarity of the phase diagrams of these three salts. We should expect for TABLE VIII. Properties of Traxsitiox Line 7-6. Substance Average slope Average Av cm.3/gm. % Curvature Variation of Av RbNOs .009 .0052 1.0 Normal Normal CsXOs .009 .00.3.5 1.3 u « TIXO3 .008 .0024 1.3 " " 622 BRIDGMAN. CsNOg and TINO3 a line 8-7 like that of RbNOg. Such a line has not been found, however. Failure to find it for TINO3 is not strange ; the transition 8-7 may well have risen above the melting point, which is comparatively low, only 205°. But failure to find it for CsNOs is not so easy to explain. If the transition point has risen above the melting point, which for CsXOs is at 414°, the effect of increasing molecular weight on the transitions 8-7 and 7-6 is very different in going from RbNOa to CsNOa; the latter is depressed from 164° to 154°, while the former would be raised from 219° to over 414°. It seems quite possible that the transition for CsNOs may have been over- looked, and that a careful search up to 414° would be worth while. The mixed crystal diagrams of RbNOs with CsNOs and TINO3 suggest that both the latter salts would show the modification 8 if it were not for the interference of the liquidus line. With regard to the transition 6-2 of TINO3 there is certain indirect evidence that such transitions exist for both CsNOs and RbNOs, although their existence has not been directly proved. Wallerant states that CsNOs passes to another modification on cooling in liquid air. The crystalline form of this new modification he found to be rhombohedric, approximately cubic. This is not the same as the crystalline system of TINO3, 2, which is orthorhombic, but the diffi- culty of crystallographic observation at very low temperatures does not make it absurd to suppose that the system may have been in- correctly determined. The existence of another modification is of itself significant. With regard to RbNOs, Wallerant's observations on mixed crystals of the system NH4NO3 - RbNOs make it almost certain that at low temperatures RbNOs has another modification. In the mixed crystal diagram there is a range of concentration, varying with temperature, within which RbNOs and NH4NO3 crystallize together in a form with all the characteristics of TINO3, 2, a form which has not been found for either pure RbNOs or NH4NOS. An easy extrapola- tion of the region of the mixed crystals indicates at low temperatures the existence of the phase as pure RbNOs. I am not aware that Wallerant made the same examination of RbNOs down to liquid air that he did of CsNOs. If it should turn out that RbNOs and CsNOs have the phase 2, the transition point to 6 would be quite differently situated for these two salts than it is for TINO3. This need not be surprising; the work of Tutton on the sulfates, for example, has shown that although the Thallium salts are closely related to those of the alkaline metals proper, yet a detailed analysis shows abrupt dis- continuites in the numerical magnitudes. EFFECTS OF PRESSURE ON NITRATES. 023 The diagram of KNO3 next demands attention. The phase I or 8 is the same as the phase 8 of RbNOs. It is most tempting- to call the new high pressure phase III the same as 7 and IV the same as 6, for this would mean that at high pressures, by a simple displacement of the origin of pressure, the diagram of KNO3 takes its place with those of higher molecular weight, RbNOs, CsNOs, TINO3. Further- more, a shift of the pressure origin in this direction, corresponding to a decrease in molecular weight, seems a not unnatural result of the corresponding decrease of internal pressure. It is unfortunate that a comparison of numerical results is not favorable to this identification. The only data which we can use are for the line III-IV; we are to inquire whether this can correspond to the line 7-6 of RbNOs, CsNOs, and TINO3. The slope of the line III-IV is 0. 3, while that of 7-6 varies from 0.008 to 0.009 and the change of volume on the line III- IV is 5.5% against 1.6, 1.3, and 1.3% respectively for the other three salts. Another bit of presumptive evidence against the identification of III with 7 is that the unstable form of KNO3 found below the transi- tion I-II at atmospheric pressure is rhombohedric. It is plausible, but not necessary, to suppose that this unstable form is the same as III. But the form of 7 is cubic. The evidence is not conclusive either way, however, and it would be worth while to make special effort to find the system of KNO3 III; the triple point I-II-III is at such a low pressure that it would not be out of the question to try for a microscopic examination in a heavy glass capillary. In the case of NH4NO3 the probability is fairly good that the new phase VI stable only at high pressures is the same as 6 of Rb, C's, and Tl. This is indicated not only by its position with respect to 7 (NH4NO3, I) but also by the numerical values. The average slope of 7-VI is 0.0075 and the change of volume 1 .4% ; both of these are very close to the corresponding values for the Rb, Cs, and Tl salts. The probability is further much increased by the mixed crystal diagrams of NH4NO3 with RbNOs and CsNOs. Mixed crystals of RbNOg or CsNOs, 6, are stable over a very wide range of concentration, which comes closest to pure NH4NO3 at the same temperature as that indicated by an easy extrapolation of the line I-VI for NH4NOS to atmospheric pressure. The same argument cannot be applied to the mixed crystal diagram of NH4NOS with TINO3 because of the anomo- lous behavior of NH4NO3, II. To make the correspondence of NH4NO3 with the heavier nitrates complete, the phase IV of NH4NO3 should be identical with III of TINO3. But both Groth and Wallerant treat these two phases as 624 BRIDGMAN. not the same, calling the first 3 and the second 2. It is significant, however, that crystallographically, both these phases are remarkably similar; Wallerant describes both as ortho rhombic, quasi-tetragonal, optically negative. The chief argument against their identity seems to be that they do not form a continuous mixed crystal series, but such a continuous series would be impossible in any e\'ent because of the anomolous behavior of the tetragonal form II. In fact it is the peculiar behavior of the tetragonal form that supplies Wallerant with the argument that the forms II and V of NH4NO3 are identical. On the whole, the evidence does not seem to me conclusive that 3 and 2 are not identical; in view of the suggestiveness of the phase diagrams, further crystallographic investigation would be desirable. At any rate it is significant that the phases 3 and 2 are so much alike. Further identification of the phases of NH4NO3 is probably not possible; the phases 4 and 5 are not likely to exist in the other nitrates. These phases, if they are capable of existence at all elsewhere, are probably to be found at considerable negative pressures, which we cannot realize experimentally. On the other hand, it would not be strange if NH4NO3 had modifications peculiar to itself. The Radical NH4 cannot be surrounded by so simple a field of force as a single atom of Potassium, Rubidium, Caesium, or Thallium, and it is not surprising if in virtue of its greater complication larger numbers of stable forms are possible. Tutton has found the same behavior among the sulfates. Summary. The phase diagrams between 0° and 200° and fron^ 1 to 12000 kg. have been determined for NH4NO3, KNO3. RbX03, CsNOs, TINO3, and AgN03. One new phase has been found for NH4NO3 and two new ones for KNO3. The usual thermodynamic data, which include change of volume, latent heat of transition, difference of compressi- bility, thermal expansion, and specific heat, are given. AgNOs stands in a class by itself. The nitrates of Rb, Cs, and Tl have closely similar phase diagrams. The more complicated dia- grams of KXO3 and NH4NO3 may be brought into relation with these by a proper identification of the new high pressure forms with the atmospheric forms of RbNOs, CsNOs, TINO3. Additional crystallo- graphical work seems necessary, however, before this identification can be established. It is significant that at high pressures the dia- EFFECTS OF PRESSURE ON NITRATES. 625 grams of all five nitrates tend to the same simple type; the disturbing complications are found only at low pressures. Acknotoledgment. I take this occasion to make grateful acknowl- edgment of liberal grants from the Bache Fund of The National Academy of Sciences and from the Rumford Fund of The American Academy of Arts and Sciences, with which the expense of mechanical assistance and materials has been met. The Jefferson Physical Laboratory, Harvard University, Cambridge, Mass. Proceedings of the American Academy of Arts and Sciences. Vol. 51. No. 13. — July. 1916. THE PATHOLOGICAL EFFECTS OF RADL4NT ENERGY OiY THE EYE AN EXPERIMENTAL INVESTIGATION F. H. Verhoeff, A.m., M.D. Pathologist and Ophthalmic Surgeon, Massachusetts Charitable Eye and Ear Infirmary; Assistant Professor of Ophthalmic Research, Harvard University. Louis Bell, Ph.D. Consulting Engineer; Past President Illuminating Eiigiaeeriiig Society. WITH A SYSTEMATIC REVIEW OF THE LITERATURE BY C. B. Walker, A.M., M.D. Assistant in Ophthalmology, Harvar(j University. Associate in Surgery, Peter Bent Brigham Hospital. li'BOM THE Pathological Laboratort of the Massachusetts Charitable Eye and Ear Infirmary. THE PATHOLOGICAL EFFECTS OF RADIANT ENERGY UPON THE EYE. TABLE OF CONTENTS. Page Introduction 630 Photophthalmia 634 Determination of Liminal E>q)osure 638 Verification of Law of inverse squares 640 Effect of Repeated Exposures 641 Limit of Abiotic Action with respect to Wave Length 645 Record of Experiments (Nos. 37 to 93) 652 Histological Technique 660 Reactions of Ocular Tissues to Abiotic Radiation 662 Clinical; Conjvuictiva and Cornea 662 Histological Changes Found 665 Cornea 665 Conjunctiva 669 Iris 670 Lens 671 Possible Abiotic Effects on Retina 677 Experiments on Eyes of Rabbits 679 Experiments on Eyes of Monkeys 681 Experiment on Human Eye 684 Possible Abiotic Effects on retina of Aphakic Eye 686 Refutation of Birch-Hirschfeld's Findings 687 Thermic Effects of Radiant Energy 692 Cornea 692 Iris and Lens 696 Retina ' 697 Record of Experiments with concentrated sunlight (Nos. 95 to 101) . 699 Theory of Action of Radiant Energy on the Tissues 703 Abiotic Energy in the Solar Spectrum 705 Snow Blindness 706 Solar Erythema 708 Erythropsia 710 Vernal Catarrh 713 Senile cataract 715 Concentration of Energy in Images 716 General Nature of Absorption of Radiant Energy 717 Eclipse Blindness and Allied Phenomena 720 Possible Specific Action of Infra-Red 728 Experiment relating to accuracy of fixation 732 Glass Blowers' Cataract 734 Applications to Commercial Illuminants 737 Experiment with nitrogen lamp 738 Experiment with quartz mercury lamp enclosed in globe . . . 742 Protective Glasses 744 Ultra Violet Light as a Germicidal Agent. Experimental Investigation of its Possible Therapeutic Value 749 General Conclusions 756 Systematic Review of the Literature — C. B. Walker 760 Bibliography .... 794 Plates 811 '1630 VERHOEFF AND BELL. The fundamental purpose of this investigation has been to dis- cover what if any pathological effects can be produced upon the structure of the eye by exposure to artificial or natural sources of light. That such action may occur under sufficiently powerful exposure to radiant energy is certain, but the essential fact is the discovery of the quantitative relations between the amount of incident energy and the effects. These relations have generally been left quite out of the reckoning in discussing the subject, with the result of leading to vague and often quite unwarranted conclusions as irrelevant as if one should condemn steam heating as dangerous because one can burn his finger upon a radiator. The quantitative phase of the matter is, from a practical standpoint, all-important since on it depends the actual effects to be expected from the exposure of the eye to powerful natural or artificial sources of light. Although the literature of the suljject is very extensive, the appended bibliography covering over 450 titles, practically none of the work done has been quantitative in the sense of connecting the amount and kind of the energ\' received with the effects produced, and hence, despite the work of many careful investigators, the results have been singularly discordant and inconclusive, so that a coordi- nation of the facts from the standpoint of energy has seemed impera- tive. One of us ^^ has investigated recently the energy relations of the radiation from various sources of light both natural and arti- ficial, and the intent of the present investigation has been to determine by actual experiment on the eye the quantitative and qualitative effects of radiant energy on the conjunctiva, cornea, iris, lens, and retina. It has long been known that excessive radiation of one kind or another produces pathological changes in the eye, of many kinds and greatly var;)dng degrees of intensity. So far as natural light is con- cerned the well known effects of powerful solar radiation in producing snow blindness and allied troubles have been long familiar as also have been the severe scotomata due to direct observation of the sun, familiar in the literature under the general name of eclipse blindness. With the introduction of the electric arc, mild cases of ocular trouble due to over exposure to the arc began to attract attention, at first nearly a half a century ago, and the subject has occupied an increasing space in the literature ever since. More recently attention has been particularly drawn to the ultra violet radiation as productive of these pathological conditions, and most of the investigations bearing on the general subject have been directed toward the study of the specific EFFECTS OF RADIANT ENERGY ON THE EYE. G31 action of the ultra violet. It, therefore, becomes of fundamental importance to examine the effects of radiant energy with special refer- ence to their relation to the wave length of the radiation. Nature and Distribution of Radiant Energy. All radiant energy is at present believed to consist of transverse vibrations in the hypothetical ether, all propagated at the same rate and differing only in amplitude and wave length, hence in frequency, which is the reciprocal of wave length. The uniform propagation rate in vacuo is very nearly 300,000 km. per second and the wave lengths so far as ordinarily dealt with range from about .01 to about .0002 mm. For ordinary purposes no attention need be paid to the extremely long wave lengths ranging to .1 mm., to the extremely short ones between .0001 and .0002 mm., or to the enormously shorter one still of the order of magnitude of .0000001 mm. such as the X-rays are believed to be. For the very long waves are not present in material amount in the radiation from ordinary sources. The very short ones are absorbed by a few cm. or dm. of air, and the X-rays are practically only produced in apparatus intended for that purpose. The spectra given by all ordinary sources range between the more modest limits just given. In the earlier literature this spectral range used to be divided into heat rays, light rays, and actinic rays, a distinction wholly artificial since the three effects implied are far from being sharply defined. More generally the whole range is divided into the infra red portion, not ordinarily visible and extending from the longest waves to those of about 760 ixfj., the visible spectrum, extending from about 760 fj-fx to about 395 ix/jl, and the ultra violet portion reaching from 395 nij. to the neighborhood of 200 mm- This distinction is not rigorous or with sharp limits. While artificial distinctions have led to many misunderstandings, all radiation of whatever wave length is convertible into heat when absorbed by material bodies and may produce chemical changes as well. As a matter of fact these latter show a general tendency to increase with the frequency of the oscilla- tions, so that chemical changes are rare in the infra red and increasingly frequent as one approaches the extreme ultra violet. It is this tend- ency that is shown in the pathological changes which may be caused in living cells by the incidence of radiation. The rationale of the chemical effect of radiation seems to be that while all radiation transfers energy to the molecules which absorb it 632 VERHOEFF AND BELL. and produce heat, certain particular wave frequencies fall into step, as it were, with the oscillation periods which depend on the mo- lecular structure, and so break up the molecules when the energy absorbed is sufficient. The particular kind of radiation which pro- duces this direct action depends on the character of the molecules. Thus, for instance, the green modification of silver bromide is readily broken up by radiation of wave length as great as 1 /jl, while it requires radiation of double this frequency to affect ordinary silver bromide, and the molecules of living protoplasm begin to break up only when the wave length is down to about 300 yu/x as we shall show. But most chemical compounds are unaffected by any practical amount of radiation which may fall upon them except as they may be heated to the point of decomposition. Any effect which is due to radiation is in the last analysis dependent on the absorption of that radiation, in that there is involved a transfer of energy to the molecules or their parts in order that they may be heated or shaken apart. Every sub- stance absorbs radiant energy in greater or less degree, and the amount of absorption bears a definite relation to the thickness of the body as well as to the particular wave length of the incident energy. Certain substances, like fluorite and to a less degree quartz, let pass with very little obstruction radiation from far into the infra red to wave length 200 iJLiJL. Water absorbs the longer wave lengths of the infra red up to about 1.2 /i powerfully, and transmits nearly everything else up to the extreme ultra-violet, while pure air, generally speaking extremely transparent, produces some small but sharp absorption in the visible spectrum and completel}^ wipes out the extreme ultra violet. But whatever the wave length, the law connecting the absorption of energy with thickness of the medium is extremely definite. If a layer of unit thickness transmits a certain fraction T then any other thickness x will transmit a fraction T'^ of the incident energy. Thus, if a substance transmits badly, leading to a low value of T, very little energy gets through the outermost layers, while if it be fairly transparent a con- siderable amount of energy penetrates deeply. For example, a cer- tain Jena glass transmits violet light through 1 mm. of thickness with only a small fractional per cent of loss. It transmits the same light througli a cm. of thickness with the loss of only 2%, while near the extreme ultra violet of the solar spectrum it still transmits a little over 90% for a mm. of thickness, but barely 3S% through a cm. Where, therefore, radiant energy falls on a solid upon which through absorption it produces powerful chemical action, the immediate effect will be almost wholly superficial, and only by prolonged and intense EFFECTS OF RADIANT ENERGY ON THE EYE. G33 radiation can enough energy be communicated into interior layers to affect them in a similar manner. To put the thing in another way, it is only relatively inactive rays that through consequent lack of absorp- tion penetrate a medium easily. With sufficient incident intensity, however, enough energy may penetrate the outer layers to produce definite action within them. In ordinary transparent media the loss of energy by real absorption in the substance is less than the loss at the surfaces by reflection. This loss depends on the refractive index of the medium with respect to the entering or emerging ray, and for nearly normal incidence the coefficient of reflection, that is the proportion of the ray transmitted through a single reflecting surface is K = (n - 1) — . This coefficient is the same for each successive surface of transition, so that for rn surfaces the coefficient of transmission are Km = K™. Thus, if a glass plate has an index of refraction n the light transmitted is K^ ■^v-- s \ \ s \ \, N, \ s s \ X \ \ \, a' \, s \ \ s \ X b 1.0 I.l 1.2 1.3 1.4 1., n Figure 1. Transmission of glass surfaces. for the two surfaces. In case of an optical system having several lenses, the reflection losses may be severe, particularly if some of the glasses are of high index. Figure 1 shows in curve a the transmission of a single surface for various indices of refraction, and in curve b transmission of a double surface like that presented by a transparent 634 VERHOEFF AND BELL. plate. The reflection is a function as well of the angle of incidence, but for the ordinary angles up to 30 degrees or so the variation is negligible. At large angles of incidence such as would be presented, for instance, by the marginal rays of a beam incident upon the cornea the loss by reflection may be considerably more than doubled so as to materially reduce the amount of energy absorbed. In any case the surface density of the energy received by the cornea 80 70 bn ^ r.o c V > a: ■"« ■20 10 \ \ \ s \ \ \ \ \ \ \ Distance from axis Figure 2. Distribution of radiant energy on cornea. under such circumstances is diminished, following Lambert's law, in direct proportion to the cosine of the angle of incidence. The net result is that from parallel rays the cornea receives a much greater incidence of energy per unit area in the centre than toward the margin, which accounts for some of the results to be recorded later. Figure 2 shows for the average rabbit's cornea the approximate variation in the intensity of energy per unit area from centre to periphery. Photophthalmia. Inasmuch as most of the pathological changes in the eye observed, after exposure to light, either clinically or experimentally, have been ascribed to the action of the ultra violet part of the spectrum, it is with this that our work has chiefly been done, although we have also EFFECTS OF RADIANT ENERGY ON THE EYE. 635 examined the effect of the other radiations which are received from natural or artificial sources. Our first aim Avas to ascertain what quantitative relations existed between the incidence of energy on the eye and the pathological effects which might follow. Especially we desired to ascertain whether these effects were proportional to the incident energy and hence to such primary lesions as might be produced by it, or serve to set in train pathological changes of an extent not proportionate to the primary inducing cause. To this end we first turned our attention to the so-called ophthalmia electrica or photophthalmia (Parsons), at once the earliest known and commonest of the superficial pathologi- cal effects of radiation. Probably first observed by Foucault and Despretz about sixty years ago, it received its first notice from the medi- cal standpoint in a paper by Dr. Charcot. ^^ His brief clinical observa- tions are here reproduced in full as they are typical so far as the external effects go of a mild case of this particular affection. The luminous effects as described, are not characteristic and were no doubt purely psychical, and due perhaps to undue attention having been called to the sensations of light normally arising in the dark adapted eye. The fusion and vitrification of refractory substances produce far more intense eft'ects of this kind than would have been noted in the other experiments cited by Dr. Charcot. Erythema Produced by the Action of the Electric Light. By Dr. Charcot. — The fourteenth of February last two chemists were cooperating in making some experiments on the fusion and vitrification of certain substances by the action of the electric battery. They made use of a Bunsen battery of 120 elements. The experiments lasted about an hour and a half; but during this time the action of the battery was frequently interrupted and it was not working in all more than twenty minutes. At the distance of the experi- menters from the arc, about fifty cm., they were not sensible of a rise in tem- perature. Nevertheless, that evening and during the whole night which they passed without sleep they found in their eyes a feeling of severe irritation and saw almost continually flashes and colored spots. The next day both had upon their faces erythema of a purplish color with a feeling of pain and tension. In the case of M. W., the right side of whose face alone was exposed to the luminous arc, the reddening covered that whole side from the roots of the hair to the chin, and the sparks were only seen as if before his right eye. In the case of M. M. who had held his head lower and whose face had been protected against the arc by his brow, the brow only was affected with erythema. Upon both the experimenters the appearance of the skin in the parts affected was exactly that of sunburn; a slight desquamation was established at the end of four days and lasted in all five or six days. 636 VERHOEFF AND BELL. This effect of the electric light is very curious and in its pathology one may perhaps find the rationale of sunburn properly so-called. Everybody knows a high temperature is not a necessary condition for the production of this last affection, for there are some people who are attacked in the cool weather of the first days of spring, a fact analogous to those which we here report. Both concur in showing that in the radiation of the light it is not the calorific rays which attack the skin. Must one then invoke the action of the luminous rays themselves? No, or at least the intensity of the light seems to play only a secondary role. Indeed in the experiments made by M. Foucault in coupling several Ruhmkorff coils to produce sparks of which the length increased with the number of bobbins and where he had been able by the means of a double action interrupter to double the number of these sparks without diminishing their energy, this observer was attacked by headache, very marked and persistent troubles of vision and erythema, although the light was not more intense than that of a star which one looks at without fatigue. M. Despretz has noted that light obtained with 100 Bunsen elements produces eyeache and that from 600 ele- ments very rapidly produces erythema. There remain the so-called chemical rays and it is this sort of rays which seems to be the principle essential agent of the accidents. To protect the eyes, it suffices as M. Foucault has several times noted, to let the electric light pass through a uranium glass screen which absorbs a large proportion of the chemi- cal raj^s. Doubtless by protecting the face with this same uranium glass one would avoid also the production of erythema. The very rapid and energetic action of the electric light upon the skin and upon the retina one can under- stand the better since the chemical rays in it are as is well known relatively more abundant than in the solar light. An ordinary clinical case of photophthalmia as observed after exposure to arc lights, short circuits, and the like, commonly takes the following course. After a period of latency, varying somewhat inversely with the severity of the exposure, but usually several hours, conjunctivitis sets in accompanied by erythema of the surrounding skin of the face and eyelid. There is the sensation of foreign body irritation, more or less photophobia, lacrimation, and the other ordi- nary symptoms of slight conjunctivitis. Occasionally there is some chemosis. The symptoms usually pass off in two or three days, and in severe cases there may be desquamation of the affected epidermis around the eye. In a very few instances the cornea has been slightly and temporarily affected. There is almost always immediately fol- lowing the exposure, and quite unconnected with the photophthalmia proper, the ordinary results of a glare of light in the eyes, persistent after images, occasional scotomata, erythropsia and xanthopsia. EFFECTS OF RADIANT ENERGY ON THE EYE. G37 The diagnosis and prognosis we can hardly better state than in the words of Van Lint ^^^ in his report to the Belgian Ophthahnological Society. " Diagnosis: The symptoms and evolution of the malady character- ize very clearly accidents provoked by electric light. Nevertheless, the diagnosis is sometimes delicate. Certain people, especially employees w^orking habitually by electric light, complain of ocular troubles which they assign to the influence of electricity. These troubles have for the most part no relation to the cause invoked. In the patient affected by conjunctivitis one generally finds a sliglit infective conjunctivitis, if by troubles of vision one finds asthenopia due either to a local cause, hypermetropia or astigmatism, or to a general cause anemia, fatigue or the like. One must consequently eliminate all these outside causes before concluding that the troubles are chargeable to the electric light." "Prognosis: As one is able to see after a study of the symptoms the prognosis is always favorable. A duration of about five days seems to be necessary for the course of the malady. In case of nervous asthenopia the prognosis is equally favorable provided one protects the patient against the luminous sources. A case cited by Fere endured six weeks, but the patient was affected by nervous symptoms which had very remote relation with those provoked by electric light." Our first series of experiments was concerned with the relation between cause and effect in photophthalmia of rabbits following exposure to a powerful source of ultra violet radiation. As the source of energ}' we employed a quartz mercury lamp operating on 220 \olt circuit and normally taking 3.5 amperes with about 90 volts across the terminals of the tube. This was the same lamp of which the radia- tion has already been studied by one of us ^^ and which furnishes by far the best source of energy for such experiments, inasmuch as its ultra violet radiation is powei'ful and the light after running twenty to thirty minutes to heat up is extraordinarily steady. It is also remark- ably advantageous in the distribution of energy in its spectrum, since it gives off relatively little radiation of long wave length, the nearer infra red region being particularly weak, so that results obtained by it are not complicated, save in some experiments with bacteria, by any effects due purely to temperature. Although there is consider- able heat loss in the lamp it is nearly all in the form of heat waves of very long wave length which are wholly cut off by a cell containing pure water, the infra red lines of the spectrum being ^•ery few. As respects the radiation from this lamp, therefore, it is practically all 638 VERHOEFF AND BELL. in the visible and ultra violet portions of the spectrum, 35% being in the visible spectrum itself and 65% in the ultra violet between wave lengths 400 fxij, and 200 /x/j.. This 65%, is eciually divided between wave lengths 400 ju/i to 300 /jlh and 300 /xju to 200 fxfx, as one of us has already shown (loc. cit.). As the lamp was run the total radia- tion of energy having wave lengths less than 400 (jlijl at a distance of 50 cm. from the tube was to a very close approximation 11,000 ergs per second per square cm., of which 5,500 ergs per square cm. of energy were of wave length less than 300 /x/z. At distances other than the standard one here noted, the radiation follows the law of inverse squares with substantial precision. A small correction should theo- retically be applied because the radiating body is approximately a cylinder instead of a point. But for all practical purposes this correction may be neglected, since for a radiative body of the dimen- sions used it amounts to less than one quarter of 1% at all distances greater than 50 cm., and does not exceed 2% even when the distance is reduced to 20 cm. Plate 5, Fig. 1, shows the actual spectrum of the quartz lamp taken with a rather wide slit and prolonged exposure with wave length scale annexed, and Figure 2 shows the stronger lines of the visible spectrum alone. The lines toward the right of A, which do not appear in B, are merely the ultra violet lines of the over- lapping second order spectrum, the photograph having been taken with a concave grating of 1 meter focal length and 10,000 lines to the inch. Attention should be called to one interesting feature of this mercury arc spectrum. It will be seen that there is but a single ultra violet line between the strong double line at 313 /x/z and the strong group at 365 nix and this line is relatively weak. There is also a conspicuous gap between 313 and the next line at approximately 303 jjLiJi. These gaps in the spectrum are of some significance in inter- preting bactericidal experiments in this region of the spectrum. Determination of Liminal Exposure. As a starting point in our experiments it was necessary to determine,, using the standard source just described, how long exposure at some known distance was necessary in order to produce clearly marked symptoms of photophthalmia. Our experimentation throughout the work has been chiefly with rabbits, since these animals have been generally used by other experimenters and the characteristics of their eyes have therefore become fairly well known. EFFECTS OF RADIANT ENERGY ON THE EYE. 639 The method of experimenting was as follows: The animal was enclosed in a box without a cover through one end of which the head protruded, being held in place by a sliding end piece. The eye was held open b}' a Murdoch specuhmi made of proper size for the pur- pose, and was then exposed at a known distance from the quartz lamp working at standard intensity for a given period. A few hours later, usually the next day, after the course of the experiments was settled, careful examination was made of the external eye for any signs of effect from the radiation, the unexposed eye being used as a check. After a few preliminary trials we found that the occurrence of slight conjunctivitis was less readily determinable than the damage to the corneal epithelium showing in the reflection of light from the cornea by a slight irregular crackled appearance giving way after stronger exposures to faint stippling. A still greater severity of exposure produces faint haziness. We also tried staining with fluo- rescine as index of damage to the epithelium, but found in the first stages disturbance of the corneal light reflection a more reliable guide. This indicates a somewhat more severe exposure than produces the first trace of conjunctivitis, but its presence or absence is quite definite wdiereas the conjunctivitis may not be easy to determine if the rabbit's eyes are naturally somewhat reddened. The minimum exposure at .5 meter required to produce the first signs of pathological change on the surface of the cornea was deter- mined to be six minutes. The following experiments show the results leading to the establishment of this minimum, and it will be seen that the liminal period was sometimes about a minute shorter or longer, various animals differing somewhat in sensitiveness. It will be observed that to produce loss of corneal epithelium required an expo- sure about 2h times that necessary to produce slight photophthalmia. Experiments. Negative results. Exposures in minutes : 1; 2|; 3; 3; 3; 5; 5; 5; 7|. Slight but definite conjunctivitis with impairment of corneal light reflection. Exposures in minutes : 7|; 5; 7^; 7|; 6; 6; 6; 7|; 6; 4 (albino). Marked conjunctivitis with slight haze of cornea but without loss of corneal epithelium. Exposures in minutes: 10; 10. Marked conjunctivitis with haze of cornea and loss of corneal epithe- lium. Exposure in minutes: 15. 640 VERHOEFF AND BELL. Verification of Law of Inverse Squares. The outcome of this series of experiments was that radiation from the mercury vapor lamp to the amount of 4 X 10 ® erg-seconds per square cm. is required to set up the first definite symptoms of photoph- thalmia. This assumes that the effect is proportional to time, in other words, that the pathological results are determined by the total amount of energy, and the next series of experiments was directed to the establishment of the truth or falsity of this assumption. For this purpose, having ascertained the liminal exposure for a single distance, ..5 meter, exposures were made at various distances for times computed for equal total radiation, assuming the law of inverse squares to hold for the relative intensities. For example, at 1 meter the time required to produce the determining symptoms, assuming the law of inverse squares, sliould be four times that required for .5 meter, which was found to be closely the case. By repeated experi- ments at distances varying from about 20 cm. to 2.5 meters, the inverse square law was verified over a range of radiation intensities in the ultra violet varying from 72,000 ergs per square cm. per second down to 455 ergs per second per square cm., at a range in other words of 156 to 1. For any source yielding rays capable of producing patho- logical effects on the cornea therefore, the exposure time required to produce symptoms of photophthalmia is inversely proportional to the intensity of the radiation of such rays, and can be definitely deter- mined when the intensity of the damaging radiation is known, subject to the condition that if the computed time reaches many hours it may be even further lengthened by the intervention of physiological repair. This conclusion is of fundamental importance, since it shows that the symptoms are due to or are proportional to, the direct and primary effect of the energy. Since, as we shall show later, the rays which are able to injure cells by chemical action are only those of wave lengths below 305 ix/jl, these present experiments of ours show that the critical amount of such radiation required to set up well marked photophthalmia is approximately 2 X 10 ® erg-seconds per square cm. In other words only half of the total ultra violet already specified is effective in producing such symptoms. A close general relation l)etween the amount of incident energy and its effects on the cornea was beautifully shown by the results obtained after relatively severe exposures. In such cases after the symptoms had developed there was a distinct haziness confined chiefly to the EFFECTS OF RADIANT ENERGY ON THE EYE. 641 central portion of the cornea and rapidly shading off toward the periphery, where, as is shown in Figure 2, the energy received per unit area is greatly decreased. Thus the mere appearance of the affected area shows in a qualitative way proportionality between the exposure and the following lesion, a proportionality shown to be definite in the experiments we have described. The Effects of Repeated Exposures to Abiotic Radiations. A natural corollary of the proposition that the pathological effects of abiotic rays on the cornea are proportional to the energy is that at least for brief intermissions the effects of repeated short exposures are equivalent to their sum in a single long exposure. This is, of course, subject to the general qualification that reparative processes are steadily going on, tending rather gradually to the healing of injured tissue. An ordinary case of photophthalmia completely disappears in less than a week and repair is going on all through this period. It is obviously possible also that apparent complete recovery may still leave the tissues slightly hypersensitive to further injury. We therefore set about investigating the effects of repeated exposures, both liminal and subliminal, to ascertain the additive effect of short exposures, the rate at which the reparative processes proceeded, the completeness of their work, and the possible effects of secondary reactions incidental to the main pathological effects. There was a bare possibility that something akin to anaphylaxis might occur owing to the development of toxins, and this phase of the matter had also to be investigated. In the case of abiotic radiation affecting a large portion of the body it seems possible that a general constitutional effect might occur owing to the absorption of the toxic substances produced. Such an effect is known to occur after severe burns from heat. In the case of the eye, however, the amount of tissue affected is of course too slight for any such effect to be expected. The experiments on subliminal exposures repeated at intervals of a few minutes to an hour or more, here summarized, show clearly that within 24 hours the energy effects are simply additive, intermissions within this time evidently being too short for reparative action to take place. The discovery of this fact is important since it shows that with any source of abiotic rays it is the total exposure that counts, and that the effect of this total exposure, if within 24 hours, can be calculated from the data already given. 642 VERHOEFF AND BELL. Experiments. effect of repeated exposures. quartz mercury vapor lamp distance .5 meter. Siibliminal Exposures Repeated within 2.'+ Hours. Experiment 23. Right eye. Exposure 3f minutes. Interval 10 minutes. Exposure 3f minutes. Left eye. Exposed 7| min- utes continuously. Result: Reaction in both eyes, more marked in right. Experiment 24. Right eye. Exposed fi\'e minutes with four intervals of one minute each. Left eye. Exposed 5 minutes continu- ously. Result: Very slight reaction in each eye. Experiment 25. Right eye. Exposed 3 minutes. One hour inter- val. Exposed 3 minutes. Result: Marked reaction. Experiment 26. Right eye exposed 3 minutes. Four hours inter- val. Exposed 3 minutes. Result: Marked reaction. The next phase of the investigation dealt with subliminal exposures at intervals of one or more days, such as might occur in actual use of sources rich in abiotic radiation. The results show that an exposure of one-sixth the liminal repeated every 24 hours for 52 days has no visible effect on the cornea or conjuncti\'a. An exposure of one-third the liminal repeated every 48 hours has a slight eifect on the cornea after seven to nine exposures, which however gradually disappears in spite of the exposures being continued. A daily exposure of one- third the liminal begins to produce a reaction after six exposures. The conjunctivitis disappears, but the corneal eflFect gradually in- creases until after thirty-four exposures there is a marked central haze. This leaves a slight corneal scar which is barely visible forty days after the last exposure. An exposure one-half the liminal repeated at the end of 24 hours produces the effect of a single liminal exposure. A single exposui'e just subliminal, increases the sensitiveness of the eye to abiotic radiation for over two weeks. On the other hand an exposure one-sixth the liminal every 24 hours, or one-third the liminal every 48 hours repeated for a long period of time, has the effect of rendering the eye somewhat less sensitive to abiotic action. EFFECTS OF RADIANT ENERGY ON THE EYE. 643 Experiments. Subliminal Exposures Repeated after 24 Hours. Quartz Mercury Vapor Lamp. Distance .5 Meter. Experiment 27. Right eye. Exposed 1 minute every day except Sunday for 52 days. (44 exposures.) Result: No reaction through- out the experiment. Three days after last exposure; each eye exposed six minutes. Result: Moderate reaction in each e3'e, but greater in the left eye. Experiment 28. Left eye. Exposed 2 minutes every other day, except when Sunday intervened. Results: After 7 exposures, slight stippling of corneal surface. After 9 exposures, slight haze of cornea. After 1 1 exposures, haze of cornea gone. After 28 expos- ures, cornea clear, no stippling. After 33 exposures cornea clear, exposures discontinued. Nine days later, each eye exposed six min- utes. Results: Right eye, marked reaction. Left eye, much less reaction. Experiment 29. Left eye. Exposed 2 minutes every day except Sunday. A speculum was used at first but caused ectropion of the lid and was soon dispensed with. Results: No reaction until sixth exposure when there was slight conjunctivitis and stippling of the cor- nea. After 14 exposures the conjunctival reaction had disappeared but the cornea was distinctly hazy in the centre. After .34 exposures the central haze of the cornea was marked and the epithelial surface showed a number of fine irregular ridges. Exposures discontinued. Forty days later only a barely visible opacity of the cornea remained. Enucleation. Microscopic examination shows Bowman's membrane absent in places, and proliferation and irregular arrangement of the superficial corneal corpuscles. Experiment 30. Left eye exposed 3 minutes. After 24 hours, no reaction. Left eye exposed 3 minutes. Right eye exposed 6 minutes. Results: Reactions equal in two eyes. Experiment 3L Right eye exposed 5 minutes. Result: No re- action. Two weeks later. Right eye exposed 4 minutes. Result: Slight reaction. The next series of experiments had to do with the effect of previous reactions upon the sensitiveness of the eye to subsequent exposures. It was found that previous reactions rendered the eye more sensitive for at least one month, thus reducing the time of exposure necessary 644 VERHOEFF AND BELL. for a liminal reaction. It was found also that if an exposure sufficient to produce a slight reaction, was followed within 24 hours by a sub- liminal exposure, the total effect was considerably greater than that produced in the control eye by a continuous exposure of the same total length. Experiments. effect of previous reactions upon sensitiveness of eye to subsequent exposures. quartz mercury vapor lamp. Experiment 32. Distance .5 meter. Albino rabbit. Right eye exposed 4 minutes. After 24 hours, slight reaction (animal unusually sensitive). Right eye exposed 2 minutes. Left eye exposed 6 min- utes. Results: Right eye, increased reaction with loss of corneal epithelium. Left eye, moderate reaction without loss of corneal epithelium. Experiment 33. Distance .5 meter. Right eye exposed 5 minutes. Result: Very slight reaction. One month later. Right eye exposed 3| minutes. Result: Moderate conjunctivitis, marked stippling of cornea. Experiment 34. Right eye exposed 5 minutes at .5 meter. Left eye exposed 7| minutes at .5 meter. Results: No reaction in either eye. 14 days later. Right eye exposed forty minutes at 35 cm. through crown screen. Left eye exposed 4 minutes at 35 cm. with- out screen. Results: Slight reactions, more marked in left eye. 8 days later. Right eye exposed 3 minutes at .5 meter. Left eye exposed 4 minutes at .5 meter. Results: Slight reaction in each eye. More marked in right. Experiment 35. Distance .5 meter. Left eye exposed 7| minutes. Result: Reaction. 26 days later. Left eye exposed 4 minutes. Result: Slight reaction. Experiment 36. Right eye exposed 1| hours at 20 cm. through crown screen. Result: Marked reaction with keratitis, lasting over 2 weeks. 5 weeks later. Left eye exposed 2 minutes at .5 meter. Result: No reaction. EFFECTS OF RADIANT ENERGY ON THE EYE. 645 Determination of the Limit of Abiotic Action with Respect TO Wave Length. The critical wave length at which abiotic action on tissue cells ceases has not hitherto been accurately determined. For bacteria it has been found to be about wave length 295 mm- In this connection the observations of Henri ^^^ and his wife are important. These observers determined the coefficient of absorption of egg albumin for various wave lengths and found that the results corresponded closely q: 3UUU 2500 2000 IJOO \ !\ \ \ \ 1000 \ \ \ \ \ 500 \ \ \ \ ^^ '—■ , 2o0 Wave length SOOfJifl Figure 3. Variation of abiotic power with wave length. (Plotted from Henri's results). with the time value'of^bactericidal action for the wave lengths tested. The curve of absorption plotted from their results (Fig. 3) shows that the abiotic action of light with reference to wave length may be expected to diminish rapidly and terminate at about 310 mm- The experiments of Widmark, Hess, and Martin, in which the lens epi- 64G VERHOEFF AND BELL. thelium was injured by exposures through the cornea, prove con- clusively that 295 MM is not the limit for human cells, since the cornea obstructs all waves less than 295 ^i^t in length. It has frequently been assumed that there is no actual limit of abiotic action but that the latter exists in a diminishing degree through the entire syectrum. Theoretically this may be true, but practically it is not, as our experi- ments show, and for the following two reasons, namely, first, that in the case of the longer waves and moderate light intensities the abiotic action is so slight as to be readily overcome by the physiological activi- ties of the cells, and second, that in the case of the longer waves and intensities theoretically sufficient to produce abiotic effects the cells are destroyed by heat action, so that there is no opportunity for abiotic effects to become manifest. The real problem may therefore be stated to be the determination of the critical wave length for abiotic action with light intensities just below those sufficient to pro- duce injurious heat effects. For the experimental investigation of this problem the cornea and lens are of all the tissues of the body the most suitable. This is so because, owing to their great transparency, extreme light intensities are required to produce heat effects in them sufficient to mask abiotic effects. The conjunctiva and skin are far less suitable for this purpose # because when the limit of abiotic action with respect to wave length is approached the hyperemia due to heat action overshadows that due to abiotic action. In this investigation it was necessary to abandon the use of the quartz mercury lamp employed in the earlier experiments. As already noted the spectrum of this source has conspicuous gaps in the very region to be examined for the purpose in hand. It has no lines of perceptible intensity between 334 ^i/x and the strong double line at 313 jLt^t. Then there is a further gap extending down to the group having its center about 302.5 /x^ and another between this group and 297 iJ.fx. In fact there are only three rather widely separated lines in this entire debatable region within which the limit sought was known to lie. We therefore turned to the commercial magnetite arc as the most convenient available source since this had already been found by one of us to be particularly rich in the extreme ultra violet. This lamp uses as active electrode an iron tube carrying a compressed mixture of magnetite and of titanium oxide, in a pro- portion of about 3 of the former to 1 of the latter, opposetl to a copper positive electrode. The light is practically all derived from the arc stream produced by the magnetite electrode. The spectrum of this EFFECTS OF RADIANT ENERGY ON THE EYE. 647 source is enormously rich in lines due to a complex mixture of those belonging to iron and titanium, reaching down to v/aves below 230 mi. Moreover, the spectrum is particularly rich in lines between 330 n^i and 290 jiix. Beyond this the intensity falls oflF noticeably. The source is thus particularly well adapted for work in the region here investigated. The lamp used took approximately 9 amperes at the arc which consumed approximately 750 watts. The energy in the ultra violet from wave length 390 mm. was about 15,000 ergs per second per square cm. at .5 meter standard distance, of which approxi- mately 3500 ergs was below 300 ^l^t, as against very nearly 5700 ergs per second per square cm. for the quartz lamp in the same region. The latter source, however, as just pointed out has relatively more energy in the shorter wave lengths. Plate 5 shows side by side the spectra of the two sources in the ultra violet. For determining the wave length at which abiotic effects on the cornea and lens cease, the use of suitable screens is very much prefer- able to attempts at using the spectrum formed by a quartz prism as the source of energy. This is for the reason that with screens one can obtain an enormously greater amount of energy than it is practicable to get by passing the radiation through a slit, coUimating lens and prism, especially in cases where a considerable area like that of the cornea must be covered. In our experiments seven screens were employed of which the absorption was definitely ascertained by the spectrograph. These screens were of various sorts of optical glass and mostly in the form of discs 43 mm. in diameter and 2 mm. thick. They were as follows: — Limits of absorption 1. Extra dense flint Nd 1.69 335 ju/x 2. Medium flint Nd 1.62 315 mm 3. Medium flint Nd 1.616, 1 mm. thick. 310 mm 4. Light flint Nd 1.57 305 mm 5. Medium crown Nd approximately 1.52 300 mm 6. Extra light flint Nd 1.54 298 mm 7. Light crown Nd 1.51 295 mm The absorption of these seven glasses for the magnetite spectrum is shown in Plate 6 with the scale of wave lengths subjoined. These limits are taken at the point where the transmission somewhat ab- ruptly ceases. They do not run to the last lines of which traces are visible, since these are so immensely reduced by the absorption as to have little if any effect bearing on the experiments. In any case the 'B48 VERHOEPF AND BELL. error would be in the direction of safety, that is, it would tend to set the Umit of abiotic action at too long a wave length. It will be noted that in this series of screens the transmission grades off with considerable regularity. In addition to these screens the following absorbing media were used in some of the experiments. These did not prove of value in determining the limits of abiotic action, but are described here because the experiments in which they were used are important in reference to possible retinal effects, owing to the long exposures given. (Plate 5.) Material Limit of Absorption 8. CuClo 1.5% solution in 5 cm. quartz cell 320 ^iju, beyond 700 ^t/x 9. Blue Uviol glass 2 mm. 285 /x/x and beyond 470 ju^t 10. Auramine O .001% solution in 5 cm. quartz cell 250 nfj., 400 fx/j. to 450 nn 11. (9+10) Auramine O, which we tried in several concentrations, is remarkable for the freedom with which it transmits the extreme ultra violet while absorbing the violet end of the visible spectrum rather strongly. For producing intensive exposures, and particularly for work on the retina the magnetite arc here described was reenforced by the use of a quartz lens system. For one set of experiments we employed two piano convex quartz lenses each of 42 mm. diameter and 18 cm. focal length. These two were generally employed placed with the plane faces in contact either with each other or with one of our screens, making in fact a single lens of 9 cm. focal length for parallel rays. This lens was placed 20 cm. from the arc, an image of which was formed 14 cm. beyond it, at which point the eye was placed. In this set of experiments in addition to abiotic effects in the cornea and lens, small circumscribed heat effects were obtained in the retina analogous to those of eclipse blindness. These will be discussed later (page 697). In another set of experiments the apparatus was assembled as shown in Figure 4. The lenses referred to, B, were placed at 12 cm. from the arc flame A. In the converging cone of rays produced by B, was placed at a distance of 12 cm. therefrom a double convex lens C of quartz cut across the axis, 23 mm. in diameter and of 14 mm. focal length for parallel rays. In the combination used, C brought the rays to a focus at about 12 mm. from its outer apex at or near which point the cornea of the eye D was located in the experiments. The EFFECTS OF RADIANT ENERGY ON THE EYE. G49 path of the rays is shown by the dotted hnes in the figure. The effect of the arrangement was to pass through the cornea a strongly diverg- ing pencil producing a circle of intense light on the retina. The axial length of the ordinary rabbit's eye is about 16.5 mm. and the ordinary diameter of the area of intense illumination produced by our appara- tus was about 11 mm. at the retina as was determined by actual experiment upon a freshly removed eye. With this apparatus a large amount of energy could be concentrated on any required area at or within the surface of the cornea or lens, and by aid of these lens sys- tems we were able to obtain exposures enormously more severe than could possibly be obtained from artificial light sources in ordinary use or than have ever been obtained by previous experimenters in this field. The image was kept fixed during the exposure by slight shifting of the source or lens, since the arc itself tends to wander. By the use of photo paper the following data were obtained con- cerning the relative intensities of radiation at the focus and on the retina with the double lens system. The size of the area of most Figure 4. Quartz condensing system. Screens omitted for simplicity. intense illumination at the focus was 2 X 4j mm. The diameter of eft'ective illumination at a distance of 16 mm. from the focus, corresponding to the position of the retina, was 11 mm. Through a euphos glass screen an exposure of 5 seconds at the focus closely corresponded in intensity to an exposure of 75 seconds at the position of the retina, so that neglecting absorption by the media of the eye the intensity of the illumination of the retina with this lens system was about fifteen times less than that of the cornea. This ratio was con- firmed by photographs taken with a model schematic eye without screen and with a picric acid screen, so that it may be assumed to hold over a very wide range of wave lengths. ' The results of these experiments are given l)elow. For purposes of comparison other experiments with the magnetite are also given here, although they have no direct bearing on the determination of the critical wave length of abiotic action. For the same reason experi- ments are given showing the length of exposure to the quartz mercury 650 VERHOEFF AND BELL. vapor lamp through a crown screen (295 fxfx) necessary to produce photophthalmia. The thermic effects produced in these experiments as well as the character of the abiotic effects are discussed elsewhere (pages 662 and 692). For determining the critical wave length of abiotic action it will be seen that the crucial experiments were Experiments 81 to 85. These showed that for light of extreme intensity no effects either abiotic or thermic, were produced on the cornea or lens epithelium by an exposure of 1| hours to waves over 315 mx in length, and that no abiotic effects but slight thermic effects were produced by light con- taining wave lengths of 310 /x/x and longer. Light containing wave lengths of 305 fx^ and longer, produced marked thermic effects but no abiotic effects after one hour exposure, but after 1| hours exposure produce marked thermic eft'ects and the slightest possible trace of abiotic effects. That is to say, the limit of abiotic action with refer- ence to wave length for corneal cells is almost exactly 305 (jl/jl. In Experiment 82, in which the light was focussed on the cornea, the only evidence of abiotic action was the loss of corneal epithelium, while in the Experiment 83, in which the light was focussed on the anterior siu-face of the lens, the only evidences of such action were the slight but characteristic changes in the lens epithelium. In any other available tissues than the cornea and lens such slight abiotic effects would undoubtedly have been completely masked by extreme heat effects, but in none of these experiments did the lens show the slight- est thermic effects. The insignificance of the abiotic action at wave length 305 ^t^t becomes more apparent when the equivalent critical time is computed for exposure to the direct radiations from the magnetite arc. Our experiments show that the intensity of light at the focus of the double lens system is at least eighteen times the intensity of the bare arc at a distance of 20 cm. This means that to obtain slight loss of corneal epithelium with direct waves of 305 ^t/i in length from the magnetite arc, an exposure of 27 hours at a distance of 20 cm. or an exposure of 28 days at a distance of 1 meter would be required. To produce mild photophthalmia the time required would be about one-third these figures. As a matter of fact however, our experiments on frequently repeated subliminal exposures, already given, prove that at 1 meter no eft'ects whatever would be produced by such slight abiotic action per unit of time owing to the vital activities of the cells. Similar calculations for the portion of the spectrum including only waves longer than 295 fxfx, as well as direct experiments, show that EFFECTS OF RADIANT ENERGY ON THE EYE. 651 an exposure of 20 minutes at 20 cm., in case of the magnetite arc, is required to produce photophthalmia, or an exposure of 85 hours at 1 meter. The ratio of the abiotic activity of the whole spectrum to that of the portion exceeding 295 /x/x is y. This holds approxi- mately true also for the quartz mercury vapor lamp. In regard to the lens epithelium, it should be noted that only abiotic changes were obtained, and these with wave lengths as long as 305 iJLix as already stated. With light of this wave length, however, as shown by the above computation it is evident that they could not be obtained by means of the direct light from any artificial light sources at any distance at which the eye could bear the heat. With wave lengths of 295 fx/j. and over, the lens epithelium was affected by an exposure of 5 minutes to the double lens system, the liminal expo- sure probably being about 3 minutes. Since the cornea itself cuts off the waves at this point, it follows that an exposure to the bare mag- netite arc of 54 minutes at 20 cm. or of 22 hours at 1 meter would be required to affect the lens epithelium. Hess, strangely enough, was unable to obtain lens changes through a screen transparent to waves of 2S0 iJLn, and drew the inconsistent conclusion, in view of the fact that the cornea was known by him to obstruct still longer waves, that lens changes are produced only by the very short waves of the spectrum. The character of the lens changes produced in our experi- ments is described on page 671. Judging by the effects on the cornea, the abiotic intensity at the focus for the single lens system was about ^ that for the double lens system. Comparing the results obtained with the quartz mercury vapor lamp and those with the magnetite arc it is found that the abiotic activity of the entire spectrum of each is in about the ratio of 6 for the mercury lamp to 5 for the magnetite arc. This ratio holds with and without the crown screen (295 ^t^u) . These ratios obtained from the pathological effects are in fairly close accord with those derived from the experiments of one of us by purely radiometric methods. Taking average conditions of the two sources here referred to, the abiotic radiations from the quartz lamp should aggregate about 4200 ergs per second per square cm. at the standard distance of .5 meter. The magnetite arc as used gave abio- tic radiations aggregating about 3300 ergs per second per square cm. at the same distance. The ratio between these two quantities is 5 to 6.35 as compared with the 5 to 6 of the pathological results, an agreement quite as close as could reasonably be expected considering the nature of the case. 652 VERHOEFF AND BELL. Full grown rabbits were used in all the experiments and the light was focussed upon the cornea unless otherwise stated. In all experi- ments relating to the retina the pupil was previously dilated by a mydriatic. In the experiments with the doul)le lens system the con- junctiva was not directly exposed to the light so that the conjuncti- vitis noted was largely secondary. The iris however was sometimes more or less exposed. Experiments, quartz mercury vapor lamp. crown glass screen (295 /iju). Distance 35 cm. Experiment 37. Exposed 20 minutes. No reaction. Experiment 38. Exposed 40 minutes. Little if any reaction. Experiment 39. Exposed 40 minutes. Little if any reaction. Experiment 40. Exposed 60 minutes. Moderate conjunctivitis. Cornea clear. Iris congested. Distance 20 cm. Experiment 4L Exposed 1^ hours. Marked purulent conjuncti- vitis. Cornea very hazy in central area. Iris congested and hem- orrhagic. The conjunctivitis persisted about 10 days. The corneal haze and a few minute iris hemorrhages were \isible over 4 weeks. MAGNETITE ARC. NO LENSES OR SCREENS. Experiment 42. Distance 50 cm. Exposed 6 minutes. Slight conjunctivitis. Stippling of corneal surface without loss of epi- thelium. (Liminal reaction.) Experiment 43. Distance 20 cm. Exposed 1 minute. Slight conjunctivitis. Corneal epithelium intact. Experiment 44. Exposed 2| minutes. Moderate purulent con- junctivitis. Cornea stippled but epithelium intact. Experiment 45. Exposed 4 minutes. Marked conjunctivitis with edema and punctate hemorrhages. Cornea hazy and epithelium lost from f of its surface. Microscopic examination (48 hrs.) shows leucocytic infiltration of cornea, but corpuscles and endothelium normal. Lens epithelium normal. Experiment 46. Exposed L5 minutes. Marked conjunctivitis. Haze of cornea with loss of epithelium. Microscopic examination EFFECTS OF RADIANT ENERGY ON THE EYE. 053 (48 hrs.); shows leucocytic infiltration of cornea. Corpuscles slightly affected in outermost layers. Endothelium normal. Lens epithe- lium normal. Iris normal. Serum in anterior chamber. Experiment 47. Exposed 16 minutes. Result same as in previous experiment. MAGNETITE ARC. CROWN GLASS SCREEN (295 Mm)- Experiment 48. Distance 20 cm. Exposed 20 minutes. Slight conjunctivitis. Cornea clear, not stippled, epithelium intact. Experiment 49. Distance 20 cm. Exposed 40 minutes. Well marked conjunctivitis with edema and slight purulent discharge. Cornea clear. Reflex impaired, but epithelium intact — does not stain. Experiment .50. Water cell. Distance 14 cm. Exposed 22 min- utes. Moderate conjunctivitis. Corneal epithelium intact. Micro- scopic examination (48 hrs.). Cornea, iris, and lens epithelium normal. No serum in anterior chamber. MAGNETITE ARC. QUARTZ SINGLE LENS SYSTEM. WATER CELL. Experiment 51. Albino. Crown glass screen (295)Uju). Exposed 20 minutes. Immediate enucleation. Microscopic examination shows cornea, iris, and lens epithelium normal. Experiment 52. Pigmented eye. Crown glass screen (295 ixfx). Exposed 6 minutes. Moderate conjunctivitis. Slight haze of cornea without loss of epithelium. Microscopic examination (4 days) : Corneal stroma shows slight leucocytic infiltration. Corpuscles and endothelium normal. Lens epithelium normal. Experiment 53. (PI. 4, Fig. 12). Albino. Crown glass screen (295 iJLn). Exposed 12 minutes. Purulent conjunctivitis. Cornea hazy, shows loss of epithelium. Iris congested and hemorrhagic. Microscopic examination (48 hrs.) : Corneal corpuscles show marked abiotic changes. Endothelium absent. Iris shows interstitial hem- orrhages. Lens epithelium shows moderate abiotic changes. Retina shows burned spot. Experiment 54. (PI. 3, Fig. 9). Albino. Crown glass screen (295 M/x). Exposed 20 minutes. Marked reaction with loss of corneal epithelium. Microscopic examination (2 days): Most of the corneal corpuscles destroyed in central area, endothelium absent. ^Marked abiotic changes in lens epithelium. Iris shows interstitial hemorrhages. 654 VERHOEFF AND BELL. Experiment 55. (PI. 2, Fig. 5). Albino. Crown glass screen (295 iJLn). Exposed 1 hour. Marked reaction with loss of epithelium. Microscopic examination (24 hrs.): Corneal corpuscles destroyed in central area, endothelium absent, marked abiotic changes in lens epithelium. Iris shows interstitial hemorrhages and slight purulent exudation from vessels. Retina shows burned area. Experiment 56. Pigmented eye. No screen. Exposed 20 minutes. Marked reaction with loss (|) of corneal epithelium. Beginning vascularization of cornea on 6th day. Microscopic examination (6 days): Corneal epithelium reformed, corpuscles largely destroyed, endothelium absent. Lens epithelium shows marked changes. Iris: Most of the stroma cells destroyed in anterior half of iris for a distance of 1.5 mm. from the pupillary margin. Some of the stroma cells show characteristic granules. A few mitotic figures seen. Endo- thelium entirely absent from some of the blood vessels. Interstitial hemorrhages. Experiment 57. x\lbino. Crown glass screen (300 ^^u). Exposed 15 minutes. Little if any reaction. Microscopic examination (6 days): Cornea, iris, and lens epithelium unaffected. Retina shows heat effect involving pigment epithelium only. Experiment 58. Pigmented eye. Flint glass screen (315 /x/i). Ex- posed 1 hour. No reaction. Microscopic examination (6 days) : Cor- nea, iris, and lens epithelium unaffected. Retina shows burned spot. Experiment 59. Albino. Flint screen (335 ^/x). Exposed 1^ hours with 5 minutes intermission at end of 30 minutes. No reaction. Microscopic examination (2 days) : Cornea, iris, and lens epithelium unaffected. Retina and chorioid show burned area. MAGNETITE ARC. QUARTZ DOUBLE LENS SYSTEM. WATER CELL. NO SCREEN. Experiment 60. Exposed 5 seconds. No effect. Experiment 61. Exposed 10 seconds. Slight reaction. Corneal epithelium lost in exposed area. Experiment 62. Exposed 10 seconds. Result same as in Experi- ment 61. Experiment 63. Exposed 30 seconds. Marked reaction. Slight haze of cornea with loss of epithelium. Experiment 64. Exposed 5 minutes. Marked reaction. Soften- ing of corneal stroma. Translucent corneal scar at end of two months. EFFECTS OF RADIANT ENERGY ON THE EYE. G55 In the following experiments (see page 668) the exposure was sufficient to produce marked photophthalmia with loss of corneal epithelium and endothelium, destruction of corneal corpuscles, soften- ing and swelling of the stroma, and marked changes in the lens epi- thelium. The retina was normal in all. Experiment 65. Exposed 6 minutes. Enucleation at end of 4 days. Experiment 66. Exposed 20 minutes. Enucleation at end of 10 hours. Experiment 67. (PI. 2, Fig. 6). Exposed 20 minutes. Enucleation at end of 48 hours. Experiment 68. Exposed 20 minutes. Enucleation at end of 4 days. Experiment 69. (PI. 1, Fig. 1). Exposed 20 minutes. Enucleation at end of 12 days. Experiment 70. (PI. 3, Fig. 7). Exposed 20 minutes. Enucleation at end of 2 months. MAGNETITE ARC. QUARTZ DOUBLE LENS SYSTEM. WATER CELL. ]]lfh croum screen (295 ixji) : Experiment 71. Exposed 2 minutes. Slight conjunctival reaction. Corneal reflex impaired but epithelium intact. Experiment 72. Exposed 3 minutes. Loss of corneal epithelium in exposed area. In the following experiments there w^as marked keratitis and abiotic changes in the lens epithelium. The retina was normal. Experiment 73. Exposed 5 minutes. Enucleation at end of 48 hours. Experiment 74. Exposed 20 minutes. Enucleation at end of 5 days. Experiment 75. Exposed 20 minutes. Enucleation at end of 10 days. Experiment 76. Exposed 20 minutes. Enucleation at end of 34 days. With .001 cmramine 0 solution in quartz cell 5 cm. thick (substituted for w^ater cell), and blue uviol screen, the two together obstructing all waves less than 250 nfx and longer than 470 /jlij. in length (see PI. 5). Experiment 77. Exposed 45 minutes. Enucleation at end of 3 656 VERHOEFF AND BELL. days. Marked keratitis and abiotic changes in lens epithelium. Retina normal. With flint glass screen (298 hijl) : Experiment 78. (PI. 4, Fig. 11). Exposed 1 hour. After 20 minutes: Cornea shows well marked central haze (heat effect). Epithelium intact. After 24 hours: Slight conjunctival reaction. Haze of cornea greater. Central loss of epithelium (5 mm.). After 72 hours: Corneal epithelium reformed. Haze of stroma persists. Enucleation. Eye immediately opened. Fundus bisected vertically. One half fixed in saturated solution of mercuric chloride. The other stained by the vital methylene blue method. Lens fixed in Zenker's fluid. Microscopic examination : Corneal stroma swollen to 3 normal thickness. Corneal corpuscles completely destroyed in posterior two-thirds, present, but show characteristic abiotic changes in anterior third. Endotheliimi necrotic and absent in places. Lens epithelium shows marked abiotic changes. Retina normal. mth crown glass screen (300 /zyu) : Experiment 79. Exposed 1 hour. Slight conjunctival reaction. Marked central haze of cornea with loss of epithelium. ^Microscopic examination (48 hrs.): Corneal stroma swollen to twice normal thick- ness. Epithelium, corneal corpuscles, and endothelium completely destroyed in central area. Towards the periphery corpuscles first show abiotic changes then proliferative changes (heat effect). Lens epithelium shows slight but definite abiotic changes — swelling of cells and characteristic granules. Iris normal (unexposed). Retina normal. Experiment 80. Exposed 20 minutes. Slight central haze of cornea persisting over 9 days. Epithelium intact. With flint glass screen (305 nn): Experiment 81. (PI. 1, Fig. 2). Exposed 1 hour. Immedi- ately after exposure ; cornea perfectly clear. After one hour : Cornea shows distinct central haze. After 24 hours: Cornea more hazy, epithelium intact and normal. After 3 days: Haze of cornea persists. There has been no loss of epithelium on daily examination. Micro- scopic examination (3 days) : Corneal epithelium intact. Stroma swollen to § normal thickness. Endothelium absent. Corpuscles invisible in posterior \ of exposed area, present and actively proliferat- ing in anterior portion. No evidences of abiotic action. (Iris not exposed). Lens epithelium normal. Retina normal. Experiment 82. Exposed 1| hours. Immediately after exposure: cornea shows faint haze. After 45 minutes: Corneal haze more EFFECTS OF RADIANT ENERGY ON THE EYE. 657 distinct. After 1 hour: Haze more distinct. Epithelium intact. After 2 hours. Epithehum intact. Shght diffuse deep staining of stroma with fluorescine. Animal avoids obstacles by sight (other eye absent). After 24 hours: Cornea hazy, shows loss of epithelium (3 mm.). After 48 hours: Epithelium reformed. Microscopic ex- amination (48 hrs.); Corneal epithelium intact but thin. Stroma swollen to almost twice normal thickness. Corpuscles completely destroyed in posterior portion of exposed area, a few present in anterior portion show no abiotic changes. At periphery of exposed area the cells are actively proliferating. Endothelium absent. (Iris not ex- posed). Lens epithelium normal. Retina normal. Experiment 83. Pigmented eye. Light focussed on anterior pole of lens instead of on cornea as in previous experiment. Pupil not fully dilated. Total exposure 1| hours. Exposed 30 minutes. During the exposure, the pupil became contracted. Immediately after exposure: Cornea clear. After 50 minutes: Cornea shows dis- tinct haze. After 1 hour: Exposed 1 hour. Immediately after exposure: Iris congested, pupil contracted (about 1^ mm.) and irregu- lar, does not dilate in the dark. Cornea more hazy. After 2 hours. Corneal haze very marked. Epithelium intact, does not stain. Pupil slightly larger. After 24 hours: No conjunctival reaction. Pupil dilated, but not fully so. Cornea shows marked haze without loss of epithelium. After 48 hours : Conditions about the same. Epithe- lium intact, does not stain. Microscopic examination: (48 hrs.): Corneal epithelium intact. Stroma swollen to f normal thickness, and corpuscles aft'ected as in previous experiment. Endothelium absent. Iris normal. Retina normal. Lens epithelium shows slight but definite abiotic changes. Cells are slightly swollen in exposed area and a few of them contain characteristic granules. ]]'itli flint glass screen (310 fxfx) : Experiment 84. Exposed 1| hours. After 24 hours: Marked central haze of cornea. Epithelium intact, does not stain. After 2 days: Haze of cornea about the same. Epithelium intact, does not stain. x\fter 4 days: Cornea clearer. Epithelium intact. Micro- scopic examination (4 days): Epithelium intact. Corneal stroma swollen to about f normal thickness. Corneal corpuscles nowhere destroyed, show marked proliferative changes, especially in posterior portion of cornea. No abiotic changes seen. Endothelium absent in places. Lens epithelium normal. Retina normal. With flint glass screen (315 fxn) : Experiment 85. Exposed if hours. After 24-48 hours: No re- 658 VERHOEFF AND BELL. action. Cornea clear. Microscopic examination (48 hrs.): Cornea, lens epithelium, and retina normal. Jl'ith 1.5% copper chloride solution substituted for water in quartz cell, crown glass screen {295 fifx) and blue uviol screen, the whole obstructing all leaves less than 320 /xju and longer than 700 ^t/x in length. Experiment 86. Exposed 1 hour. No effect. Microscopic exam- ination (48 hrs.): Cornea, lens epithelium and retina normal. With 1.5% copper chloride solution and blue uviol screen, obstructing all waves less than 320 fifi and longer than 700 ju/z in length: Experiment 87. Exposed 1 hour. No effect. IVIicroscopic exami- nation (3 days): Cornea, lens epithelium, and retina, normal. With flint glass screen {315 iin). {The tvater cell leaked so that for an unknown length of time the eye received also infra red rays) : Experiment 88. (PI. 1, Fig. 3). Albino. Exposed 1 hour. After 24 hours: No conjunctivitis. IVIarked haze of cornea, epithe- lium intact. After 48 hours: Haze persists. Epithelium intact. Microscopic examination (48 hrs.): Corneal epithelium intact. Stroma swollen to double normal thickness, stains faintly in eosin. Corpuscles and endothelium completely destro\ed in exposed area. At periphery of affected area corpuscles show active proliferation with mitoses (page 693). Iris shows a few minute hemorrhages near pupil. Lens epithelium normal. Retina: Pigment epithelium shows distinct heat effect over an area 6 mm. in diameter. The cells are swollen or stain deeply in eosin, and their nuclei are often pycknotic. Otherwise the retina is normal. With crown glass screen {295 fxfx): Experiment 89. Aphakic eye. Pigmented. Exposed 35 minutes with intermissions of 1 minute every five minutes. Light focussed on opening in lens capsule. After 24-48 hours: Marked keratitis with loss of epithelium. Microscopic examination (48 hrs.): Cornea shows typical abiotic changes. Epithelium and endothelium absent. Iris (only slightly exposed) shows a few minute hemorrhages. Pig- ment epithelium of retina shows marked heat effect over an area 4 mm. in diameter. Ganglion cells and other retinal elements normal. JVith flint glass screen {315 fxii) but without tvater cell: Experiment 90. Pigmented eye. Atropine mydriasis. Exposed 30 minutes. Immediately after exposure: Cornea clear. Pupil con- tracted. After 20 minutes: Distinct haze of cornea. No conjuncti- vitis. After 4 hours: Haze of cornea marked. Epithelium intact, does not stain. No conjunctivitis. Pupil larger, vertically oval. Prompt lid reflex to light. After 24 hours: No conjunctivitis. EFFECTS OF RADIANT ENERGY OX THE EYE. ()59 Marked haze of cornea. Epithelium intact. Pupil widely dilated. After 48 hours : Condition about the same. Microscopic examination (48 hrs.) : Corneal stroma swollen to | normal thickness. Epithelium intact. Corneal corpuscles completely invisible in central portion of exposed area. At the periphery of the latter they show active pro- liferation many of them being in mitosis. Endothelium absent beneath exposed area. Iris and lens epithelium normal. Retina normal — pigment epithelium unaffected. With flint glass screen {335 h/jl), hut without ivater cell. Experiment 91. Pigmented eye. Exposed 30 minutes. After 24-48 hours: No reaction, cornea clear. [Microscopic examination (48 hrs.) : Cornea, lens epithelium, and retina normal. Pigment epi- thelium of retina normal. SUNLIGHT. BLUE L"\"IOL SCREEN AND .001% SOLUTION OF AURAMIN O IN QUARTZ CELL 5 CM. THICK. Light focussed on cornea by large quartz lens 12 cm. in diameter and 25 cm. in focal length. Atmosphere not perfectly clear (April 2, 1913): Experiment 92. Albino. Exposed 45 minutes. After 24 hours: ^'ery slight conjunctival reaction. Cornea hazy, epithelium intact. Iris congested. After 48 hours: Condition about the same. Corneal epithelium intact. Microscopic examination (48 hrs.): Corneal epi- thelium intact. Stroma swollen to about f normal thickness. Cor- puscles and endothelium completely destroyed in exposed area. At periphery corpuscles show proliferative changes. Affected area wider posteriorly than anteriorly. Xo characteristic abiotic changes made out. Iris normal. Retina congeni tally defective (coloboma of optic disc, ganglion cells few in number) shows no abiotic or heat effects. Pigment epithelium normal. MAGNETITE ARC. QUARTZ SINGLE LENS. WATER CELL. CROWN GLASS SCREEN (295 fXfJ.). Experiment 93. Right e^^e exposed 8 minutes. Left e^e exposed 4 minutes. One hour later; Left eye exposed 4 minutes. After twenty-four hours : Slight reaction in each eye without loss of epithe- lium. After 3 days: No reaction. Magnetite arc. Double lens system. Water cell. Crown screen (295 /x/x). Exposed left eye 060 VERHOEFF AND BELL. 3 minutes. After l^ hours: Exposed right eye 6 minutes; left eye, 3 minutes. Marked reaction with haze of cornea and loss of epithe- lium in each eye. Microscopic examination: (5 days after first expo- sures) : Cornea of each eye shows marked abiotic changes with loss of epithelium and endothelium. Lens epithelium shows abiotic changes in both eyes, but more marked in left eye. Histological Technique. Fixation: In a few of the earlier experiments the eyes were fixed in a warm saturated mercuric chloride solution as recommended by Birch-Hirschfeld. This was not found, however, superior to Zenker's fluid for demonstrating the structure of the ganglion cells, particularly the Nissl bodies, and Zenker's fluid at room temperature was there- fore used for fixation in all except one of the experiments relating to the retina. Before opening the eye it was usually placed in the fixing fluid for about ten minutes. This pre^'ented the cornea from losing its shape and the sclera and retina from becoming distorted as happened when the eye was immediately opened. The eye was then incised all around at the ora serrata, the vitreous body gently lifted out, the lens removed and the two portions replaced in the Zenker's fluid for four to six hours. Longer fixation gives less brilliant results. After fixation the tissues were washed in running water twenty-four hours. Embedding: Celloidin embedding was employed in all except one experiment in order to avoid the shrinkage that results from the paraffine process. Cornea and Iris: Meridional sections 8 to 10 /x in thickness were always made, passing through the middle of the most affected part of the cornea and the centre of the pupil. Tangential sections of the cornea 6 to 8 yu in thickness were also frequently made. The sections were stained in alum hematoxylin followed by .2% solution of water soluble eosin in 80% alcohol. Lens Capsule: The most satisfactory method of demonstrating changes in the capsular epithelium is by means of flat preparations. This method was used by Hess ^^^ and later by Martin,^^^ but Birch- Hirschfeld ^^ speaks of using flat sections. The method as we have carried it out is as follows : The eye is opened as already described, by an incision passing just behind the ciliary body all arountl. The zonule is then cut all around by means of scissors, care being taken EFFECTS OF RADIANT ENERGY ON THE EYE. 661 not to rupture the capsule, and the lens removed and placed in Zenker's fluid for two hours. The lens may be fixed in situ and removed after- wards, but this causes the iris epithelium to adhere to it. Birch- Hirschfeld mistakenly regarded such adhesions, which he found in the exposed eyes, as pathological. They may be removed by gently rubbing the capsule with wet filter paper. The lens is now rinsed in water and the capsule incised all around the equator with a sharp knife. The anterior capsule is now readily stripped off, floated in water and treated as follows: Lugol's solution (1%) a few seconds. Water. 95% Alcohol two minutes or longer. Water. 10% aque- ous solution sodium hyposulphite until color of iodine is removed. Water. The capsule will be found to curl toward the cell free side. It is now floated upon a piece of paper and by means of scissors five radial incisions are made through both paper and capsule reaching to within a short distance of the centre. It is then freed from the paper and floated upon a cover glass with the curled edges up, so that the epithelium is in contact with the glass and thus will be nearest the lens of the microscope. The curled edges are flattened out by stroking with bits of filter paper, which removes the excess of water and prevents the edges curling again. The preparation is now blotted firmly with filter paper. Alum hematoxylin until deeply stained. It is best to use a sharply acting hematoxylin solution and avoid differ- entiating in acid alcohol as the latter is apt to act unevenly. Water. 0.2% solution of water soluble eosin in 80% alcohol, 30 minutes. W^ater. The preparation is now thoroughly dehydrated in absolute alcohol, cleared in oil of origanum followed by xylol, blotted again if necessary, and mounted on a slide in xylol-balsam. Retina: Vertical sections of the retina 6 /x to 8 /u in thickness, were made in all cases. These always included the optic disc and the area below it that had been exposed to the light during the experiment. This area contains a much larger proportion of ganglion cells than any other part of the retina and may be regarded as analogous to the human macula, although it is much larger and less sharply defined. The ganglion cells are similar to those of the human macula, but never occur in more than a single row. Plane sections of the retina were also often made, and these were found to give the best demonstration of the ganglion cells. Sections were always stained in eosin and thionin, which is proba- bly the most satisfactory method for demonstrating Nissl bodies and at the same time gives a beautiful general stain of the retina. Dilute aqueous solutions of thionin rapidly lose in staining power, so that it is important that they be always freshly prepared. The fol- 662 VERHOEFF AND BELL. •lowing carbol-thionin solution devised by one of us retains its prop- erties indefinitely and from it a powerful staining solution may be made at once by the simple addition of water: Thionin to saturation, about .3 gm. Absolute alcohol, (>0 cc. Phenol crystals (melted) 30 cc. For use, add one full drop of this solution to 2 cc. of distilled water. Sections are stained as follows : (1) Lugol's solution 1:2:100, 1 minute, followed by water, 9»/o alcohol and sodium hx'posulphite solution, to remove mercurial precipitates. Water. (2) 0.2% solution of water soluble eosin in 80% alcohol, 5 mmutes. Water. (3) Carbol thionin diluted immediately before use as above, 5 min- utes. Water. (4) Differentiate and dehydrate in 95% alcohol, two changes, until excess of thionin is removed and sections show well marked eosin stain, about 30 seconds. (5) Oil of origanum. (6) Place on slide, blot, wash in xylol, blot, xylol-balsam. To obtain the most brilliant results it is important not to overstain the sections in thionin solution as it is then impossible to produce sharp differentiation of the Nissl bodies by treating with alcohol. The results also are more brilliant the shorter the time that has elapsed between the fixation of the tissues and the staining of the sections. The Character of the Reactions of the Ocular Tissues to Abiotic Radiations. conjunctiva and cornea. Clmical: Our experiments show that the effects on the conjunctiva and cornea of moderate exposures to waves less than 295 /xM m length do not differ qualitatively in their clinical aspects from those produced by longer exposures to waves from 295 MM to 305 mm m length. Severe exposures, however, produce markedly different effects on the cornea in the case of the short waves than in the case of the longer waves owing to the fact that the latter are not fully absorbed by the corneal stroma. The effects of severe exposures to very short waves is there- EFFECTS OF RADIANT ENERGY ON THE EYE. QQ3 fore not included in the following description, but will be given sepa- rate consideration. A description of combined thermic and abiotic- effects on the cornea in certain experiments, resulting from prolonged intense exposures is given on page 694. After exposure of a rabbit's eye to light containing abiotic rays,, no immediate changes take place, however great the intensity, pro- vided a heat effect is not produced, and symptoms of irritation do not usually appear for several hours. In other words, there is a latent period before any visible effects are produced. This exists not only as regards clinical symptoms but also as regards histological changes. In a general way it varies inversely as the severity of the exposure, but in no case is the first appearance of symptoms delayed longer than twenty-four hours. That is to say, a latency longer than this corresponds to an exposure too slight to produce any demonstrable effects. The shortest latent period observed by us was thirty min- utes. This occurred after intense exposure to the short waves of the magnetite arc, as described later. The least effect that occurs after exposure to abiotic radiation consists in slight hyperemia of the conjunctiva. After more intense exposures the congestion is corre- spondingly greater and is associated with edema and purulent exuda- tion. There also may be conjunctival ecchymoses. The cornea, after exposures sufficient to produce slight conjunctivitis, remains clear and shows only slight stippling of the. surf ace. After longer exposures the cornea becomes hazy in a rather sharply defined central area. This delimitation is no doubt due chiefly to the fact that the rays strike the periphery of the cornea obliquely so that there is less light here per imit area, and to a less extent to the greater loss by reflection at the periphery (see diagram, page 634). Over the central area the epithelium shows marked stippling and is then cast off, usually, however, not until about 24 hours. The loss of epithelium sometimes cannot be determined without the use of fluorescine stain- ing, owing to the margins of the defect not then being sharply defined. This is due to the fact as shown by microscopic examination, that the epithelium usually becomes thinned by desquamation before solution of continuity occurs.* The haziness of the cornea usually reaches its * The cornea of a rabbit's normal eye often shows punctate spots and irregu- lar lines after staining with fluorescine that closely resemble the lesions of dendritic keratitis. These are due to defects in the epithelium so small that they do not easily become visible until the stain has diffused through them into the corneal stroma, which requires one or two minutes. They are possibly due to the infrequent winking for which rabbits are noted. They cannot be mistaken by anyone familiar with their appearance for erosions due to ex- posure to abiotic radiations, because the latter stain almost instantaneously and are much larger and sharply defined. 664 VERHOEFF AND BELL. maximum in about 48 hours, when, as will he pointed out, there is some leueocytic infiltration. After 3 days the purulent conjunctival discharge becomes less, but it may not entirely subside for about 9 days. The corneal epi- thelium is usually reformed on about the 4th day. Haziness of the cornea noticeably begins to subside in 3 to 10 days. After five weeks only a slight central haze remains. Following sufficiently intense exposures, new vessels are seen extending into the cornea from the hmbus in about six days. The conjunctival reaction that occurs after moderate exposure to abiotic radiations, is only in very small part reflexly due to irrita- tion of the cornea. This is proved by several experiments in which the cornea was exposed through a diaphragm which protected the conjunctiva. Here, although the cornea was markedly affected, and the epithelium destroyed, the conjunctiva showed no reaction until after about 48 hours, and then only slight hyperemia. The foregoing description applies to the effect produced on the cornea by moderate exposures to the bare mercury vapor quartz lamp, or bare magnetite arc, and by relatively long exposures (5 to 20 minutes) to the magnetite arc through a water cell, quartz lens system, and crown screen. The latter absorbs all rays less than 295 ju/z in length and thus protects the corneal stroma from injury. With the magnetite arc, and quartz lens system, but without any screen, a very much greater as well as different effect may be produced. With this arrangement and an exposure of 20 minutes a dosage is obtained that is more than one hundred and fifty times as great as that of a liminal exposure necessary to produce slight keratitis. Following such an exposure the following changes occur. Immediately after the exposure the cornea is perfectly clear. At the end of thirty minutes there is slight hyperemia of the conjunctiva and central haziness of the cornea. At the end of four hours the conjunctivitis is marked and the corneal haze much greater. The exposed area is completely anaesthetic. The epithelium is intact, but stains slightly in flourescine. The iris is highly congested. At the end of twenty- four hours there is a marked general inflammatory reaction of the conjunctiva with oedema and purulent discharge. The epithelium is lost from the exposed area in twenty-four hours, and reformed about thirty-six hours later. On the fourth or fifth day the cornea, without becoming more hazy, begins to swell in the exposed region. This swelling increases and the aft'ected area becomes softened until an appearance is produced on about the eighth day of a large flaccid EFFECTS OF RADIANT ENERGY ON THE EYE. 665 vesicle involving two-thirds the area of the cornea. This condition remains almost unchanged until about the thirteenth day, except that on al)out the sixth day vascularization of the cornea is observed. On about the fourteenth day the inflammatory reaction, which has almost completely subsided, begins again. This is probably a reaction of repair. During this time the process of vascularization makes rapid progress and the new vessels invade the central area which is now somewhat firmer, but still pits when touched by a probe. On the twenty-fifth day the inflammatorv' reaction is again almost gone and the new ^•essels have begun to disappear. The exposed area is now only slightly swollen and no longer pits, but is very cloudy. On the thirty-third day the vessels have largely disappeared. The exposed area is no longer swollen and presents a translucent appearance. After two months the surface of the cornea is smooth and there is a translucent interstitial opacity. The repair of the injury is much more complete than could be expected in the case of a human cornea. After an exposure of five minutes i. e. one-fourth the former dosage, to the magnetite arc through the quartz lens system and water cell the cornea undergoes softening in the exposed area as in the case of the longer exposures. The injury, however, is repaired without vascularization of the cornea, leaving a central translucent scar. The Histological Changes Produced in the Cornea by Abiotic Radiations. The histological changes produced in the cornea by abiotic radia- tions were studied chiefly in eyes exposed to the magnetite arc with and without interposition of quartz lenses and various screens. Corre- sponding to the differences in the clinical effects, different histological effects were obtained when a crown glass screen was used than when it was omitted. The chief difference was that with the crown screen the corneal stroma escaped injury, due to the fact that it was then protected from all waves which it strongly absorbed, namely, waves less than 295 fjifx in length. With the crown screen, exposures sufficient to destroy the epithelium always severely injured the corneal cor- puscles. Without the crown screen, on the other hand, owing to the greater abiotic activity of the short rays stopped at the surface of the cornea, the epithelium was destroyed by exposures too short to have any visible effect on the corneal corpuscles. On account of these differences the histological effects produced by the short waves and relatively long waves will be described separately. 666 VERHOEFF AND BELL. The Histological Changes Produced in the Cornea by Abiotic Waves over 295 mm in Length. Since, for reasons already given, the central portion of the cornea under the usual conditions of the experiments is much more strongly affected than the periphery, the various degrees of injury produced are easily made out by examining the cornea from the periphery towards the centre. Examined in this way twenty-four to forty-eight hours after exposure, it is found that the epithelium first shows spacing out of its basal cells, and then in addition descjuamation of the super- ficial layers until finally the epithelium is more or less abruptly cast off. At the margins of this erosion the individual epithelium cells show changes similar to those met with in the case of the lens capsule, that is, formation within the cytoplasm of eosinophilic and basophilic granules. Swelling of the cells, however, is not noticeable, possibly because the cells are cast off when this occurs. The nuclei are rela- tively little affected, although some of them are pycknotic. Mitotic figures are observed only in the apparently normal epithelium at the periphery of the cornea. After exposures through a crown screen (295 fxix) sufficient to produce injury to the lens capsular epithelium, the corneal lamellae show slight if any changes; possibly they stain less deeply in eosin. The corneal corpuscles, however, show marked changes. Just as in case of the lens epithelium, all of the cells are not equally injured and certain cells here and there entirely escape, which are fewer in number the more severe the exposure. In the most exposed region, after twenty- four hours many of the nuclei are barely or not at all visible, while most of the others are in various stages of pycknosis and fragmen- tation. The cytoplasm often contains eosinphilic and basophilic granules similar to those seen in the lens epitheliimi. These are more abundant after twenty-four hours and are best seen in thin tangential sections. The eosinophilic granules are less readily seen in the cornea than in the lens epitheliimi, probably because they are to a greater or less degree masked by the eosin stained stroma. The effect on the corneal corpuscles is progressively less the deeper they lie, but an exposure of five minutes to the double lens system through the crown screen (295 /i/x) is sufficient completely to destroy all the corpuscles in the entire thickness of the cornea and also to destroy the endo- thelium. Polymorphonuclear leucocytes begin to invade the cornea in about twenty-four hours, reaching their maximum number in about EFFECTS OF RADIANT ENERGY ON THE EYE. 667 forty-eight hours. The purulent infiltration is greater the nearer the exposed area lies to the limbus, but is never sufficient to account for more than a small part of the haziness of the cornea. It is also greater the larger the area affected by the exposure. The corneal endothelium in the most exposed region is entirely cast off within twenty-four to forty-eight hours. At the margins of the defect the nuclei show pycknosis and the cytoplasm often con- tains the characteristic basophilic and eosinophilic granules. These changes are also found after somewhat less intense exposures, in cells that remain adherent in exposed regions. Repair of the Corneal Injury. Five days after exposure the epithe- lium is usually found reformed but thin. The visible corneal corpus- cles are still further reduced in number, and of those visible many still contain eosinophilic and basophilic granules. Towards the periphery the nuclei are abnormally rich in chromatin and many of them en- larged. Some of them show direct division and budding. Occa- sionally a mitotic figure is seen here. The endothelium has not reformed. In places on Descemet's membrane there are eosinophilic and basophilic granules evidently left by necrotic endothelial cells. After ten days the epithelium is still thin. The number of corneal corpuscles in the exposed area has slightly increased. The basophilic granules are apparently unchanged, but the eosinophilic granules in some cells stain less deeply and in others have apparently become confluent causing the whole cytoplasm to stain reddish. The nuclei are rich in chromatin, often polymorphous in shape, and sometimes show direct division and !)udding. Few if any mitotic figures are seen. The endothelium is completely reformed. After five weeks the cornea presents an almost normal appearance. The corneal corpuscles now slightly exceed the normal number. Many of the nuclei are abnormally large and a few cells contain double nuclei. The cause of the slight corneal opacity seen at this stage during life is not evident from the microscopic examination. Histological Changes produced in the Cornea by Light rich IN Abiotic Waves less than 295 n/j. in Length. With the bare magnetite arc, to destroy the epithelium of the cor- nea requires an exposure only one-eighteenth of that required when a crown screen (295 iifx) is used. In the former case it is evident 668 VERHOEFF AND BELL. therefore that the effect is due ahiiost entirely to waves shorter than 295 /i/i. After an exposure of four minutes to the bare magnetite arc, at a distance of 20 cm. the epithehum, at the end of forty-eight hours, is entirely lost from the central two-thirds of the cornea. At the periphery the epithelium shows gradually increasing desquama- tion of its cells until it is reduced to a single layer for a variable distance and then abruptly ends. The cells even in the single layer are appar- ently not severely injured, and occasionally one is found in mitosis. They do not contain basophilic and eosinophilic granules, probably due to the fact that the shortest waves were absorbed by the super- ficial cells, while the remaining waves were not sufficiently intense at the periphery of the cornea to injure the deeper cells. The corneal corpuscles, lamellae, and endothelium are normal. There is, however, considerable purulent infiltration of the cornea. This is fully as great as in the case of exposures through a crown screen sufficient to injure the corpuscles. Following an exposure of 20 minutes to the rays of the magnetite arc passing through a water cell and concentrated by the quartz double lens system, the following changes are seen: At the end of 10 hours the epithelium towards the periphery of the exposed area shows changes similar to those seen after exposure through a crown screen. As the central area is approached more marked changes occur; the nuclei are seen to become extremely pycknotic and the cytoplasm to stain intensely in eosin. Within the central area itself the super- ficial layers have become desquamated, leaving usually only the basal cells, which now consist of cylinders deeply stained in eosin from which the nuclei have entirely disappeared. Within the most exposed area at this stage the corneal corpuscles are still present in normal numbers. Their nuclei show marked pycknosis, but the cytoplasm contains no granules. Towards the periphery of the exposed area a few corpuscles containing granules are seen. The corneal stroma is swollen to a third more than its normal thickness, and stains less deeply in eosin. Unless the sections are very thick the stroma is apt to fall out of them. The individual lamellae are still recognizable but are greatly distorted, due no doubt to not holding their positions in the cutting of the sections. The endothelium is still adherent, but appears com- pletely necrotic in the exposed area, the neuclei being pycknotic and the cytoplasm staining deeply in eosin. After four days the epithelium is found to be reformed. Within the most exposed region the corneal corpuscles are completely invis- ible and the endothelium is absent. At the periphery of the exposed EFFECTS OF RADIANT ENERGY ON THE EYE. G69 area, many of the nuclei of the corpuscles are pycknotic or fragmented, and cells often contain eosinophilic and basophilic granules. Further away, the nuclei are enlarged and some of them show direct di\ision. A few mitotic figures are also seen. The stroma in the exposed area is still more swollen, stains still less in eosin, and shows evidences of injury down to Descemet's membrane. The individual lamellae are no longer recognizable and the stroma appears as an almost homo- geneous substance pervaded by indistinct wavy lines. There is a moderate leucocytic infiltration. After twelve days (PI. I, Fig. 1) the stroma is still more greatly al- tered. In the centre of the exposed area it has lost its normal structure and has undergone semi-liquefaction almost down to Descemet's mem- brane. This softened area contains a large amount of fibrin and a considerable number of pus cells and endothelial phagocytes. The leucocytes, however, are too few in number to cause an appearance in any way resembling an abscess. Around the area of softening groups of corneal corpuscles are actively proliferating, forming cells similar to fibroblasts. The epithelium is intact although altered in appearance. Numerous vessels are making their way into the cor- nea from the limbus. The endothelium has been almost completely reformed, but presents an abnormal appearance due chiefly to ine- qualities in the sizes and shapes of the cells. After two months the cornea has returned to its normal thickness. The epithelium and endothelium are normal. The stroma in the affected region presents an abnormal appearance, but less so than might be expected. The corneal corpuscles are greatly increased in number and their nuclei are abnormally rich in chromatin. The new formed corneal lamellae are less regularly arranged than in the normal cornea and here and there occur areas of hyaline tissue that has not yet become definitely laminated. Blood vessels are still present but are small and few in number. The Conjunctiva. The clinical effects of exposure of the conjuncti\'a to abiotic rays have already been described. Histologically the following changes were noted in the bulbar conjunctiva 24 to 48 hours after exposure to abiotic radiations: Necrosis and desquamation of the epithelium. Infiltration of the epithelium with pus cells. Congestion, edema. 670 VERHOEFF AND BELL. interstitial hemorrhages, and shght purulent infiltration of the sub- epithelial tissue. Basophilic and eosinophilic granules were not ob- served in the epithelium, possibly due to the fact that the cells were cast off when this degree of injury was reached. These changes were obtained after exposure to the magnetite arc with and without the single quartz lens. In the experiments with the double lens system the conjunctiva was not exposed. The Iris. Clinical. Twenty-four to forty-eight hours after an exposure sufficient to injure the lens epithelium, the pupil becomes contracted and the iris shows marked congestion and minute interstitial hem- orrhages in the exposed region. The congestion quickly subsides^ but the hemorrhages may remain visible for several weeks. Hi,stoJogicaI. The iris is directly affected only after exposures sufficient to injure the lens epithelium. After exposure to the bare magnetite arc sufficient to produce marked conjunctivitis and kera- titis, but insufficient to produce apparent injury- the lens epithelium or corneal corpuscles, the anterior chamber may contain serum and fibrin, evidently the result of an indirect effect on the iris vessels. After exposures sufficient to injure the lens epithelium, there is seen, in addition to congestion and interstitial hemorrhages, an insignificant exudation of pus cells from the iris vessels. With these changes, few if any individual cells of the iris may show signs of injury. After an exposure of 20 minutes to the magnetite arc and lens system, the albinotic iris in one experiment (Exp. 68) shows marked cell changes similar to those of the lens capsule. The cells affected are the stroma cells, the endothelial cells of the vessels, and the posterior epithelium. From some of the blood vessels the endothelium is com- pletely lost. Tlu'ombosis, however, is not observed. The character- istic basophilic and eosinophilic granules are most noticeable in the cells of the posterior epithelium, no doubt due to the fact that these cells are most abundant. Similar changes are found after 6 days in a lightly pigmented iris (Exp. 56) but here the pigment hides any possible change in the pigment epithelium. In most of the experi- ments with the double lens system the pupil was widely dilated so that the iris was only slightly exposed to the light. Posterior synechiae were not observed in any of our experiments. EFFECTS OF RADIANT ENERGY ON THE EYE, 671 Birch-Hirschfeld states that adhesion of the pigment epitheUum to the lens occurred after fixation in some of his experiments, although the light intensities used were far less than those used by us. As already pointed out, the adhesions noted by Birch-Hirschfeld were undoubtedly artefacts due to the action of the fixing fluid alone, since they occur in the case of normal eyes. In view of the numerous control eyes ex- amined by this observer, it is difficult to understand why he was not aware of this fact. The Character of the Changes Produced in the Lens by Abiotic Radiations. The light intensities and wa\'e lengths necessary for the production of abiotic effects in the lens epithelium have already been given (page 651). In none of our experiments was an opacity of the lens produced sufficient to be visible through the cornea. Even when the lens was examined in air after its removal from the eye it appeared perfectly clear. If, however, it was placed in normal salt solution it showed a delicate haziness in the pupillary area 48 hours after a severe exposure. Histological. In all except one experiment upon the lens, the cap- sule was removed and examined as a flat preparation, so that it was impossible to make a satisfactory examination of the lens substance. To determine the effect of the abiotic radiations upon the latter, the lens in one experiment (Exp. 67) was fixed in formalin and hori- zontal sections made of it. The magnetite arc, water cell and system of quartz lenses were used without a screen, and the exposure was 20 minutes. This was the exposure that had been found to produce extreme changes in the capsular epithelium. The eye was enucleated at the end of 48 hours. On microscope examination the lens capsule proper is found unaltered, while the epithelium shows the marked changes described below. The lens substance is definitely affected but only for a microscopic depth, the distance beneath the capsule by actual measurement nowhere exceeding 20 fi. In this narrow zone it stains much more intensely in eosin than the rest of the lens substance and is highly vacuolated. Occasionally it contains an epi- thelial cell which has evidently been forced into it. Lens Capsular Epithelium. This is the best possible tissue in which to study the cell changes produced by abiotic radiations because of 672 VERHOEFF AND BELL. the simplicity of its structure and abundance of its cells, and because, by means of the flat preparations described (page 660) the whole of the exposed area may be examined at once. Moreover, the effects produced in it are not complicated by the presence of leucocytes, since these cannot penetrate it. Following are the histological changes produced in the epithelium by abiotic radiations: If the capsule is fixed immediately after exposure, even if the latter has been prolonged, the cells appear absolutely normal. After 24 hours, changes are well marked, and reach their maximvmi in from 48 to 72 hours. After severe exposures, the cells may be so greatly affected that many of them no longer adhere to the capsule unless the latter is fixed within 24 hours. It is noteworthy that the cells in the exposed area are not all affected alike and one cell, or group of cells, may be markedly affected while the neighboring cells are only slightly affected. The chief changes noted consist in (a) swelling of the cells, (b) the appearance of granules in the cyto- plasm, and (c) the formation of a peripheral wall of cells. (a) After short exposures swelling of the cells may be almost the only change noted. It is plainly evident after an interval of 20 hours, but does not reach its maximum until after 48 hours. It is associated with increased transparency of the cytoplasm. All the cells do not swell to an equal extent, and as a result of the inequalities in compression the cells become misshapen in an irregular manner. (b) The granules (PI. 2, Figs. 5 and 6) in the cytoplasm first appear before the cells become much swollen. They are present within 10 hours but are more abundant after 48 hours. While evidently in the case of any individual cell a greater exposure is necessary to produce them than is required to produce swelling alone, a few cells contain- ing them may always be found if the epithelium is affected at all. The longer the exposure the greater the number of cells containing them, and also the greater the number of granules in each cell, so that after prolonged exposures almost every cell may contain them. The granules are of two kinds. The more abundant are more or less strongly eosinophilic, usually round in shape and varied in size, the largest exceeding half the size of the nucleus. One cell may contain from one to over twenty granules. Each usually appears to be situ- ated in a vacuole which it does not quite fill, but this may be due to shrinkage as a result of fixation. Close examination shows that they have a reticulated and subgranular structure. The other granules are intensely basophilic, and smaller than the eosinophilic granules. EFFECTS OF RADIANT ENERGY ON THE EYE. 673 the largest being about one-fifth the diameter of the nucleus and the smallest immeasurably fine. They also are usually round, but some- times irregular in shape. Often they are contained within the eosin- ophilic granules. Owing to their strong basophilic character, the natural assumption would be that the,y represent chromatin extruded from the nuclei. Such an origin however, cannot actually be traced. On the contrary, the impression is given that the cytoplasm first breaks up into, or is transformed into the eosinophilic granules, and that the basophilic granules are formed primarily within the latter. After intense exposures, as will be pointed out, the nucleus may undergo disintegration, in which case some of the granules in the cyto- plasm are undoubtedly nuclear fragments. (c) The wall, first so named by Hess,^''^ consists of a ring of deeply staining closely packed cells at the periphery of the exposed area, that is in the position corresponding to the pupillary margin at the time of the exposure (PI. 3, Figs. 8 and 9). The cells are evidently in a state of compression and in marked cases may be heaped upon each other. The wall is visible after 19 hours but later becomes more evident. The cells within it show only to a slight extent the changes seen in the central area. Martin ^^^ assumed that the wall was due to "submaximal damage at the pupillary margin." This, however, is certainly not the case since submaximal exposures or any other exposures do not give rise to a similar condition of the cells within the pupillary area itself.* Hess explained the wall as a result of the compression of the marginal cells by the sheet of swollen cells in the pupillary area. This explanation seems undoubtedly correct. We have found a similar if not identical wall four days after the injection of staphylococci into the anterior chamber. The cells in the pupillary area were swollen but did not contain basophilic and eosinophilic granules. We have found such a wall also 24 hours after the injec- tion of Lugol's solution into the anterior chamber, as described below (page 676). As will also be pointed out, a somewhat similar but yet different wall may be produced by the action of heat transmitted by the iris (page 696). In spite of the marked changes in the cytoplasm of the exposed cells, the nuclei remain comparatively normal in appearance except * Martin described in the capsule of one rabbit repeatedly exposed, a zone of proliferated cells which somewhat resembled the wall of Hess. The pupil- lary area, however, was otherwise free from abiotic changes. The condition was attributed to the effects of abiotic radiations, but in our opinion was almost certainly a congenital malformation such as we also have seen. 074 VERHOEFF AND BELL. after the most intense exposures. Some nuclei show distortion, due possibly to the uneven compression of the swollen cells, and some stain less deeply than is normal, but following exposures through a crown screen fragmentation is seldom observed. Marked nuclear changes are seen after long exposure to the magnetite arc through the double lens system without a screen, but even then only relatively few nuclei are affected. Ten hours after such an exposure nuclei here and there show the following changes : The nucleus becomes transparent and its chromatin converted into coarse deeply staining granules attached to the nuclear membrane. The transition of a normal nucleus into this state is evidently very alirupt. The nucleus then becomes polymorphous in shape and undergoes fragmentation. Usually the fragments are each bordered by nuclear membrane and contain one or more coarse chromatin granules. Mitotic figures are first seen after about 48 hours among the unexposed cells just outside the wall where they occur in large num- bers. After 5 days they are greatly diminished in number here. After 3 days a few may also be found in the wall itself. Within the exposed area mitotic figures are not seen until about the fifth day when they occur in considerable numbers. At this time the cells are still swollen. The basophilic granules are little if any changed except possibly they are more often irregular in shape, but the eosinophilic granules have largely become confluent and are apparently under- going solution. The mitotic figures are never seen in cells containing granules. Many of the nuclei are abnormally large and show early stages of direct division and budding. At the end of ten or twelve days the cells have almost entirely lost their swollen appearance and the basophilic and eosinophilic granules have almost entirely disappeared. The most striking feature now consists in the inequalities in sizes and shapes of the nuclei. Most of the nuclei are abnormally large; occasionally one has three times the diameter of a normal nucleus. Some are abnormally small. Many of the nuclei evidently are undergoing direct division, as all the stages in this process can be seen, from a slight constriction of the nucleus to two nuclei connected by a delicate strand. In addition to this, a process of budding can similarly be traced, the nuclei becoming poly- morphous in shape and constricting off buds varying in size from that of a normal nucleolus to half that of a normal nucleus. Some cells contain as many as twelve of these free buds. The buds have the reticulated structure, staining reaction and general appearance of the nucleus proper, and each most often contains a nucleolus. Cells EFFECTS OF RADIANT ENERGY ON THE EYE. 675 that contain two nuclei of nearly equal size always contain smaller buds in addition. At first glance the nuclear buds may be mistaken for persisting basophilic granules, but careful examination shows that they bear no relation to the latter either in appearance or origin. Few if any mitotic figures can now be seen in the exposed area or else- where. At the end of 5 weeks or 2 months the capsule shows about the same appearances as after 10 days (PI. 3, Fig. 7) . There is perhaps still greater variation in the sizes of the nuclei, and a greater number of the excessively large ones. The cells with double nuclei and nuclear buds are still present. In case of the extremely severe exposures, a few cells are found still containing basophilic granules after two months. In connection with the foregoing observations on the lens capsule several interesting questions arise. In the first place how is the abun- dant mitotic division of the unexposed cells in and around the wall to be explained? This proliferation is not due to minimal exposure to the rays for it does not occur in the pupillary area 48 hours after liminal or subliminal exposures. It is also not due to heat trans- mitted by the iris, because when a flint screen is substituted for a crown screen it does not occur after exposures more than four times as long. The only remaining possibility seems to be that it is due to toxic substances diffused from the injured cells of the exposed area. If this is the case why are not mitotic figures seen at the same time in the exposed area? The answer to this is probably that the cells are here so greatly injured that they cannot respond at once to the irritation of the toxic substances, which, moreover, may at first be so concentrated as to inhibit rather than stimulate the nuclei. This brings up the question whether abiotic radiation is a direct stimulant or depressant to mitosis. It certainly is not a direct stimu- lant because, as just stated, after liminal or subliminal exposures mitosis does not occur. On the other hand it probably is a depressant because following intense exposures mitosis occurs in the exposed area only after relatively long intervals (four to five days) and then only in cells that have escaped apparent injury. This is in marked contrast to the action of heat, which, as will be shown, produces abundant mitosis in 48 hours and is evidently an active stimulant to cell pro- liferation. Whether or not abiotic radiation is a direct depressant to mitosis 676 VERHOEFF AND BELL. it is certain that repair of the injury to the lens epithelium takes place largely without the aid of this process. This is proved by the fact that mitosis does not occur in tlie severely injured cells, that is in the cells containing granules. Each of these cells, therefore, if it imdergoes recovery as usually is the case, must do so without indirect division. It is evident that the eosinophilic and basophilic granules finally' become dissolved out. The enlargement, direct division, and budding of many of the nuclei probably represent the response of the latter in the process of cell repair. Similar nuclear changes are sometimes seen in malignant tumors. The nuclear buds are still present at the end of two months and their ultimate fate is prob- lematical. Finally the question arises whether or not the cell changes de- scribed are characteristic only of the action of abiotic radiation. As will be pointed out later, experiments on the cornea prove that the basophilic and eosinophilic granules are not produced by heat, and thus their occurrence in cells constitutes a distinct difference between heat and abiotic effects. On the other hand, the following experi- ment proves that the same cell picture may be produced by chemical agents. A few drops of Lugol's solution containing 25% iodine were injected into the anterior chamber of a rabbit's eye. The injected fluid became mostly precipitated so that its action on the lens surface was not uniform. On examining the lens capsule 24 hours later there were found, in addition to more extreme changes, areas in which the cells showed identically the same changes, including the basophilic and eosinophilic granules as are produced by the action of abiotic radiation. It is therefore obvious that these changes are not characteristic of abiotic action alone, but may be produced by other forms of chemical action as well. It is interesting that in this experiment, as previously mentioned, a wall was formed similar to that produced l)y abiotic rays, evidently due to the pressure of the injured cells within the pupillary area on the peripheral cells which were protected from injury by the contact of the iris with the lens (PI. 3, Fig. 10). The changes just described occurring in the lens capsule after expo- sure to abiotic rays, are essentially the same as those described by Hess ^^^ who used much longer exposures but a light source of much less intensity than employed by us. Hess does not describe the granules in the cytoplasm, although they are shown well in his excel- lent illustrations. He also does not describe direct division and budding of the nuclei, although the latter process likewise seems to be EFFECTS OF RADIANT ENERGY ON THE EYE. 677 shown in one of his ilhistrations. Apparently he attributed the repair of the injury chiefly to mitosis and not to recovery of the injured individual cells. He states, however, that he has no evidence that ultra violet light is a direct stimulant to mitosis. Widmark,^^* strangely enough, found mitotic figures only in the exposed area and regarded ultra violet light as a direct stimulant to cell proliferation. Birch-Hirschfeld ^^ states that by means of a 20 diopter glass lens he focussed the light of a 5 ampere arc light through a euphos glass screen upon the eye of a rabbit for five minutes for three successive days and on the day after the last exposure obtained the changes described by Hess. The euphos screen obstructed all rays less than 400 ij-ix in length. II is not stated that a water cell was used, and the diameter of the lens was not mentioned. In spite of such a remarkaljle result it is not stated that the experiment w^as repeated. We have been unable to obtain such a result through a light flint screen trans- parent for waves down to 315 ^i/x with the magnetite arc and still greater concentration of energy. ^Moreover in an experiment in which we focussed sunlight upon the lens by means of a large mirror no changes in the lens capsule resulted within the pupillary area, although there was complete necrosis of the iris due to heat. The lens capsule was affected only beneath the pupillary margin where it had been in contact with the heated iris and even here the changes were not such as are produced by abiotic action. We are therefore compelled to believe that Birch-Hirschfeld was in error. Possibly he mistook a heat effect similar to that just noted for the changes descrilied by Hess. He had never previously obtained the latter changes in any of his experiments and hence from personal observation was no doubt unfamiliar with their appearance. POSSIBLE ABIOTIC EFFECTS OF RADIANT ENERGY ON THE RETINA. It might be supposed that if a source of light is not sufficiently rich in abiotic rays to damage the cornea, the retina could not be injured by these rays. This, however, is not necessarily true because if the source of light is so small in size that the area of its retinal image is less than that of the pupil, the intensity per unit area as concerns transmissible rays will be greater on the retina than on the cornea. 678 VERHOEFF AND BELL. In fact under certain conditions, and witli a moderately dilated pupil the intensity of the light reaching the retina will he enormously greater than the same light as it passes through the cornea. For this reason it will be seen that if the transmissible rays were capable of injuring tissue cells, the macula of the eye might be seriously damaged in spite of the fact that the cornea and lens remained unaffected. This, of course, actually happens in eclipse blindness in which, how- ever, as will be pointed out, the effect is due entirely to heat generated in the pigment epithelium. There are two conceivable ways, exclusive of heat effects, in which the retina coidd be injured by light. If the light were sufficiently intense it might overstimulate the physiological mechanism upon which the perception of light is dependent and thus lead to more or less permanent impairment of this mechanism. It is obvious that such an effect could not readily be produced by light of wave lengths less than 400 /x/x since the latter has relatively little power to stimulate this mechanism even in aphakic eyes. The other possibility is that intense light might injure the cells of the retina by abiotic action in the same way that light rays of short wave length injure tissue cells in general. In connection with this possibility two facts previously established by us must be taken into consideration, namely that within wide limits discontinuous exposures to abiotic rays have the same total effect as a continuous exposure of the same total length, and that there is a limit below which such siunmation does not occur. Thus it would a priori seem possible that if an indi\idual fixed a bright source of light many times daily, serious damage to the macula might result. The problem in regard to the retina that chiefly concerns us in the present investigation may be briefly stated thus: exclusive of a heat effect, can the retina of the human eye be injured by light of an}' or all wave lengths that can possibly reach it through the cornea and lens? In attempting to answer this question it is important first to inquire whether or not the waves that are able to pass through the dioptric media are injurious to tissue cells in general. If they are so injurious the question is obviously to be answered in the affirmative. If they are not, the question is in all probability to be answered in the nega- tiv^e, but not perhaps with absolute certainty, since it is conceivable that the retinal cells are more susceptible to injury by light than are other tissue cells. It has been shown by Hallauer ^^^ and others that the adult human lens always absorbs all waves less than 376 mm in length, and usually EFFECTS OF RADIANT ENERGY ON THE EYE. 679 all those less than 400 njx in length. Now we have already shown tiiat the corneal epithelium and lens capsule are not affected in the slightest degree when exposed one and one-half hours to rays as short as 310 AtjU even when the intensity is considerably greater than that to which the retina is ever subjected in the case of any of the known artificial light sources. This exposure is at least forty-five times greater than that required to affect the corneal epithelium by waves of 295 /X)U and less. For the retina therefore to be affected by the abiotic action of light transmitted by the lens, it would have to be many times more sensitive to such action than the corneal epithelium. There is no reason to believe however, that this is the case, but on the contrary, since the abiotic effect depends upon the amount of absorp- tion of the waves, there is strong reason for believing that the corneal epithelium and retina are about equally sensitive to abiotic action. Assuming this to be so, these experiments show conclusively that the human eye could be fixed steadily and at close range upon the magne- tite arc certainly for over two hours and probably for many hours without suffering damage to the retina from abiotic action. Since as already pointed out the intensity of the image of a source of light of such small size as that in question decreases as the square of the dis- tance, the danger of injury to the retina at ordinary distances would be absolutely negligible. In the case of the lenses of some children, Hallauer found a very weak transmission band at 315 to 330 ju/x. This, however, does not invalidate the application of the above argument to the case of chil- dren, since we have shown that such waves are without abiotic eft'ect. While it seems to us that the foregoing facts prove conclusively enough that the lens affords complete protection to the retina from the abiotic action of light, in view of the fact that Birch-Hirschfeld claims to have produced pathological changes in the retinae of normal rab- bit's eyes by exposure to ultra violet light (cf . page 687), we have under- taken to investigate this question by direct experiment upon the retina itself. Such an investigation presents several difficulties. In the first place it is impossible to reproduce with animals exactly the conditions that obtain when the human eye is fixed upon a small intense source of light. This is so because it is impossible to insure in the case of an animal that the small image of the light source will always fall upon the same spot in the retina during the exposure. Moreover, even if this were possible it would be difficult if not impossible to find with certainty such a small area on microscopic examination unless fiSO VERHOEFF AND BELL. the lesion produced was well marked. It is therefore necessary to illuminate a large area of the retina. This we have done by means of a suitable system of quartz lenses used in connection with the magne- tite arc as described on page 648. Intense illumination of such a large area, however, for a long period of time entails a danger of over- heating the fundus of the eye. This we have successfully obviated by interposing a quartz, cell 5 cm. thick filled with distilled water to absorb most of the infra red rays. If this had not sufficed, heat effects could probably still have been prevented by interrupting the expo- sures at intervals to allow for cooling to take place, a procedure that no doubt would be necessary for light intensities only slightly greater than that used by us. In fact in one of these experiments in which the water cell leaked, a heat effect on the pigment epithelium was actually noted (Exp. 88). As will be seen the system of quartz lenses employed concentrated the light more intensely upon the cornea and lens than upon the retina. Advantage was therefore taken of this fact to determine at the same time the effect of exposures through various screens upon these struc- tures, the results of which have already been given. None of the screens obstructed any waves longer than 305 /x/x to 315 ^lyit, that is, any waves that otherwise could have reached the retina through the lens. The screens also pre^■ented excessive keratitis, which we de- sired to avoid since it would have prevented us from later making satisfactory tests of the lid reflex and pupillary reaction to light. If these reflexes had been abolished this fact alone would have fur- nished sufficient proof of the deleterious action of the radiations on the retina. As a matter of fact, except immediately after exposure a lid reflex was always obtainable. The details of these experiments are given on pages 655-658. (Experiments 65 to 90). It will be observed that the exposure was as long as one and one-half hoiu's in each of four experiments and one hour in each of six experiments. In all except one experiment the retinae were prepared for microscopic examination in the manner already described (page 661). In Experiment 78 the eye was immedi- ately opened and the retina bisected vertically through the optic disc, one half being fixed in a saturated solution of mercuric chloride and embedded in paraffin. The sections, 2 /x in thickness, were stained in thionin as in the case of the other experiments. The other half of the retina was used for vital methylene blue staining, the results of which will be mentioned later in commenting on Birch-Hirschfeld's observations (page 687). In none of the experiments could any EFFECTS OF RADIANT ENERGY ON THE EYE. 681 apparent changes be found in the exposed retinae that could not be found in unexposed retinae prepared by the same method. Certainly if there were any differences in regard to the Nissl bodies of the gang- lion cells they were too slight to be of any pathological significance. These experiments thus show that, so far as can be determined by histological examination, the retina of the normal eye, exclusive of heat effects, is fully protected from abiotic action by the lens. Since however, the objection may be brought forward that the retina may be injured so far as its function is concerned without showing any histo- logical evidence of the fact, we have endeavored to exclude this possi- bility also. For this purpose we employed the monkey instead of the rabbit because this animal possesses a macula similar to that of man. With the lens system described it is easily possible to illuminate intensely a sufficient area of the retina to insure that the macula is always included. If under these circumstances the light has an injurious action on the retina it will be rendered evident, since the macula is injured, by marked impairment in sight and particularly by a loss or impairment of the pupillary reaction. To avoid injury to the cornea by abiotic rays and injury to the retina by heat we made use of a H% solution of copper chloride in a quartz cell 5 cm. thick. The spectrum of this solution (PI. 5, Fig. 4) shows that it absorbs all waves shorter than 320 /x/x as well as all the so-called heat waves. It thus does not obstruct any short waves that could otherwise reach tlie retina through the lens. In these experiments two monkeys were employed in each of which the left eye was blind. One was an old female monkey, whose left eye had been rendered blind by an experimental Kronlein operation involving injury to the optic nerve, one year previous to the first of the present experiments. The otlier was a young full grown male monkey, whose left eye had been rendered blind by injection of alcohol into the orbit nine months previous to the first of the experi- ments. Ophthalmoscopic examination showed complete optic atrophy in the left eye of each. Neither monkey could find the way about when the right eye was excluded from vision. Direct pupillary re- action to light was absent, but the consensual reaction was well marked. This made it possible to determine the presence of a pupil- lary reaction while the right eye was under the influence of the mydri- atic, while the fact that the left eye was blind made it easy to detect any impairment of vision of the right eye. In each animal the visual acuity of the right eye was high, as shown by the ease with which it was able to catch flies and lice. The absence of binocular vision did 682 VERHOEFF AND BELL. not seem to hamper either animal in its judgment of distance except for a short time after the left eye had been made bhnd. In all four experiments the magnetite arc was used and the same arrangement of lenses employed as in the previous experiments with rabbits, the quartz cell, as stated, being filled with a 1^% solution of copper chloride. The light was focussed on the centre of the cornea. The animal was placed in a box which allowed only the head to protrude, and the eyelids kept open by means of a small speculum. The head was forcibly held in position by the hand of the observer. For the first five minutes the animal was difficult to control, but after this no great difficulty was experienced in keeping the eye in place. No local or general anaesthetic was employed. Normal salt solution was dropped on the cornea from time to time. The pupil of the right eye was previously dilated by homatropine except in the third experiment in which atropine was used. To give some idea of the light intensity and duration of the exposures in these experiments, it may be well to state that one of us exposed his eye with undilated pupil to these conditions for fifteen seconds, and obtained an absolute scotoma which gradually disappeared within five minutes. Erythropsia persisted about three minutes and was followed by xanthopsia which lasted the remainder of the five minutes. Experiments. Experiment 94. March 7, 1913. Young monkey, Macacus Rhe- sus. Right eye exposed 1^ hours. Immediately after exposure there is a lid reflex to a new 2| volt tungsten flash light, and within five minutes the consensual pupillary reaction is apparently normal. Within ten minutes the animal is able to see an apple five feet away, which he approaches and takes from the hand. March 8. Cornea clear. Consensual pupillary action normal. Slight direct pupillary reaction in right eye in spite of mydriasis. Owing evidently to the cycloplegia, the animal cannot catch flies readily. After several days, the mydriasis having disappeared, the animal is able to catch flies with his usual dexterity. Experiment 94a. March 28, 1913. Old female monkey. (Java.) Right eye exposed 1| hours. Immediately after exposure the lid re- flex to the flash light is absent, but is present in five minutes. Con- sensual pupillary reaction not determinable owing to some of the EFFECTS OF RADIANT ENERGY ON THE EYE. 683 mydriatic having accidently gotten into the left eye. Animal has great difficulty in getting around, vision evidently being much im- paired. After one hour vision is much improved; the animal follows the observer with the eye. March 29, 1913. Cornea clear. Well marked lid reflex to flash light. Consensual pupillary reaction also well marked. Visual acuity of animal apparently normal except that animal has difficulty in catching flies owing to eft'ect of cycloplegia. After several days, the mydriasis having disappeared, the animal is able to catch flies with her usual dexterity. Experiment 94b. December 4, 1913. Old female monkey. Right eye exposed 1^ hours. One minute after end of exposure there is a barely perceptible consensual pupillary reaction. After six minutes the reaction is well marked. Animal now released. Cannot see approach of observer's hand. Is compelled to feel her way to her perch in the cage. After one hour she is still apparently blind ; can- not see a carrot held near her, although she takes it when placed against her mouth. After seven hours, vision is still impaired. December 5, (18 hours). Cornea clear. Vision apparently normal — sees carrot, avoids hand movements etc., even in poorly illuminated cage. Consensual pupillary reaction normal. After the mydriasis has disappeared the animal catches flies as usual. Experiment 94c. February 5, 1914. Young monkey. Right eye exposed 1| hours. Three minutes after beginning of exposure the consensual pupillary reaction, tested with flash light, is absent. Immediately after end of exposure the lid reflex to flash light is present, but the consensual pupillary reaction is absent. At the end of three minutes the latter is distinctly visible, and in six minutes is well marked. Animal now released, finds his way at once to perch, avoids hand of observer — evidently sees well. One hour after exposure the eye lids of right eye are sewed together. When released the animal cannot find his way about and is easily caught, thus showing that if the sight of the right eye had been affected the fact would have been easily determined. February 6. There is a small abra- sion of the cornea probably due to the animal having frequently rubbed his eye as a result of a slight irritation produced by the sutures. Cornea clear. Consensual pupillary reaction intact. Animal sees well. February 7. The abrasion of the cornea is healed. Consensual pupillary reaction intact. Animal shows no evidences of poor vision. After the mydriasis has disappeared the direct pupillary reaction to light is normal and the animal seems to have normal vision. The results of these experiments show that even with exposures of 684 VERHOEFF AND BELL. extreme intensity and length, but insufficient to produce heat effects, it is impossible to injure the retina by light containing any or all rays capable of reaching it through the lens. They exclude both the pos- sibility of injuring the retina by over stimulating its perceptive mechanism, and also of injuring it l)y the abiotic action of light. Most surprising was the rapidity with which the retina regained its function. Thus in all four experiments within six minutes the con- sensual pupillary reaction was fully reestablished. There was in both sets of experiments, however, a marked difference between the young and the old monkey in regard to the time required for the restoration of usefid vision. In the case of the young monkey sufficient vision to enable him to see his way about, avoid hand movements, etc., was present in ten minutes after the exposure ended. The old monkey on the other hand was practically blind for an hour or more. In fact her visual acuity did not seem to l)e fully restored until the morning following the exposure. Both animals were able to catch fiies with their usual expertness after the mydriasis had disappeared.* The results obtained in these experiments would also seem to be of some significance in regard to the question of light adaptation. They suggest that after a certain state of retinal fatigue is reached no further effect is produced, however long the exposure. In fact it would seem, in young individuals at least, that after this stage is reached the recuperati^'e processes begin while the retina is still exposed. This aspect of the question, however, does not concern us here and further experiments would be necessary to elucidate it fully. In addition to the experiments on the eyes of monkeys we have availed ourselves of an exceptional opportunity to make a similar experiment upon a human eye. The subject was a female patient aged 50 years affected with carcinoma of the eyelid and orbit, the growth being so extensive as to necessitate removal of the eye. The left eye itself was apparently normal, the media being clear and the fundus normal. The visual acuity was reduced to fg— (unimproved by lenses) for some reason not definitely determined, but probably due to some irregularity in refraction resulting from the pressure of the upper lid. The lower lid was almost completely destroyed, while the upper lid was somewhat drawn down by cicatricial tissue at the outer canthus. It was therefore necessary for the observer to hold up the eyelid by finger pressure during the experiment. The right * Both of these monkeys were later killed, one after seven months, the other after fourteen months, and on microscopic examination the eyes that had been exposed wer3 found normal. EFFECTS OF RADIANT ENERGY ON THE EYE. 685 eye was normal and had normal visual aeuty. Before the experiment the pupil of the left eye was dilated with atropine, but the visual acuity remained the same. The total exposure was less than in the case of the monkeys, owing to the patient becoming somewhat fatigued, and for the same reason also the exposure was not continuous, but otherwise the conditions of the experiment were the same. The total exposure was 55 minutes, and the interval between the separate exposures was about 1| minutes. The first three exposures were 3, 9, 12 minutes respectively, the remainder were 5 minutes each. At the beginning of each exposure the patient stated that the "light was like the sun." At the end of the sixth exposure there was ery- thropsia and the visual acuity was reduced to counting of fingers at one foot. Within 2^ minutes after the last exposure the consensual pupillary reaction was well marked, and the patient could with diffi- culty count fingers at six feet. Three minutes after the last exposure there was only slight erythropsia. Xanthopsia was not noted at any time but may have been unrecognized by the patient. After 10 minutes the visual acuity was ^§q. There was an appearance of a mist before the eye, but no erythropsia. After 1^ hours the visual acuity was y§q, and a slight mist still persisted. After 3 hours the visual acuity was ^q-\-, and a white surface seemed almost but not quite as white as with the right eye. After 22 hours (in the morning), the visual acuity was fg— as before the experiment. There was no erythropsia, and central color vision was perfect for red, blue and green. 24 hours after the exposure the eye was enucleated. On microscopic examination the cornea, iris, lens epithelium (flat prepara- tion), and retina were found to be normal. The result of this experiment confirms those obtained with the monkeys. It is obvious that the retina could not have been injured by abiotic action of light, since the visual acuity was fully restored within 3 hours and remained so the following morning. The rapid- ity with which the erythropsia disappeared was unexpected, and indicates that duration of exposure is equally as important as its intensity in the production of persistent erythropsia. 68G VERHOEFF AND BELL. POSSIBLE EFFECTS OF ABIOTIC RADIATIONS ON THE RETINAE OF APHAKIC EYES. Since it has been shown experimentally that abiotic waves may pass through the cornea and injure the lens epithelium, it would seem that exposure of an aphakic eye to a light source rich in such waves might seriously damage the retina. Assuming as is probable, that the retina has the same susceptibility to abiotic action of light as the lens epi- thelium, the minimal exposure to the bare magnetite arc necessary to injure the retina of an aphakic human eye may be closely approxi- mated from the data of our experiments. The working diameter of the single quartz lens was 4.2 cm., and the w^orking focal distance 14 cm., making the working aperture 1/3.3 This corresponds to the aper- ture of a human aphakic eye with a pupil 4.5 mm. wide. Now we found that the lens epithelium of a normal rabbit's eye was unaffected by an exposiu-e of 6 minutes to the single lens system through a crown screen (29.5 ^i^), but moderately affected by an exposure of 12 minutes. The liminal exposure may therefore be taken as 8 minutes. The total loss by reflection etc. from the surfaces of the lenses, screen, and water cell, amounts to about 50%. Deducting this percentage, the mini- mum exposure to the magnetite arc necessary to affect the retina of a human aphakic eye would therefore be about 4 minutes, providing that the eye was close enough for the formation of a distinct image, and ignoring the blurring due to the lessened refraction of an aphakic eye. The absorption of the cornea is allowed for in this calculation, since in the experiments the light passed through the cornea, but the general absorption of the vitreous humor is not. Assuming this to be about the same as that of the cornea (although it probably ii'- greater) the calculated exposure would be increased to about 6 minutes. Since beyond 1^ meters, owing to the small size of the source, the intensity of the light on the retina would diminish as the square of the distance, it is safe to say that under the most favorable conditions, it would require fixation of the bare magnetite arc at a distance of 3 meters for almost | hour to injure the retina of an aphakic eye. According to our experiments on the effects of repeated exposures (page 641) a daily total exposure of 5 the liminal, which in the present case would be 8 minutes at a distance of 3 meters from the mag- netite arc, would produce pathological effects in the retina of the EFFECTS OF RADIANT ENERGY ON THE EYE. 687 aphakic eye in 6 days, while a total daily exposure of g the liminal, in this case 4 minutes, would produce no effects even if inflefinitely continued. These estimates do not allow for the pupillary contrac- tion, which would result from the fixation of such a bright source and which would in most cases increase the necessary exposures three or four times, or for imperfect fixation. They also do not allow for the thick cataract glasses which in most cases would be worn and which w^ould increase the necessary exposures many times, since a 10 dioptre lens would be almost impenetrable to abiotic radiations. For the quartz mercury vapor lamp still longer exposures would be required owing to the size and shape of the light source giving less concentration in the image. It is, therefore, now apparent why there is no known case of a human eye from which the lens has been removed in which the retina has been injured by exposure to artificial light, and why such injur}' is in the highest degree improbable. In endeavoring to demonstrate by direct experiment the possibility of injuring the retina of the aphakic eye by abiotic radiation we have found it difficult to obtain a satisfactory eye for the purpose. While it was easily possible to remove the lens from the rabbit's eye, the pupil became more or less completely obstructed in almost all cases. In one animal, however, we finally obtained by means of repeated discissions, a clear pupillary opening sufficient to admit the cone of light from the quartz double lens system. According to our calcu- lations an exposure of 35 minutes with the light focussed upon the pupillary area should have been sufficient to produce abiotic effects in the retina. No allowance, however, was made for absorption by the vitreous humor. As a matter of fact no abiotic effects could be demonstrated in the retina although marked heat effects were obtained in the pigment epithelium (Exp. 89). This experiment thus goes to show that the danger to the retina from exposing the aphakic eye to abiotic radiations is even less than is indicated by the above calculations. Birch-Hirschfeld's Observations. Since the results of our experiments especially in regard to the retina are so greatly at variance with those of Birch-Hirschfeld ^^ it may be well to review his experiments in some detail. This is all the more necessary because his results and conclusions have not hitherto either been confirmed or refuted. His experiments consist of two series. In the first series he separated out the ultra violet 688 VERHOEFF AND BELL. rays from a 15 ampere carbon arc lamp by means of a quartz lens and quartz prism, and concentrated them upon the anterior focal point of the rabbit's eye by means of a second quartz lens. The diam- eter and focal length of the latter he did not mention. He exposed both normal eyes and eyes from which he had extracted the lenses. The latter were seven in number. The length of the exposures were from one-fourth hour to 6 hours. Following the exposure there was only slight hyperemia of the conjunctiva which disappeared in 24 hours. The cornea and lens were unaffected even after the 6 hours'^ exposure. The retina on microscopic examination showed the follow- ing changes: chromatolysis and formation of vacuoles in the cyto- plasm of the ganglion cells. Loss of chromatin in both nuclear layers, the nuclei of the oviter layer becoming homogeneous and their cross striations almost completely obscured. These changes were found just the same immediately after exposure as in the course of the next 12 to 24 hours. After a few days they disappeared and the ganglion cells showed an increased amount of chromatin. In the animal wliich was exposed for 6 hours, however, vacuoles were found in the ganglion cells at the end of 6 days. In the case of normal rabbits' eyes exposed to the same conditions, retinal changes were found only when the eye was removed immediately after exposure and were said to be simply those of light adaptation. In the second series of experiments he exposed the rabbits' eyes to a 3 to 4.5 ampere Finsen light. No statement is made as to whether or not a quartz lens or water cell were used, so it is to be presumed they were not. Also no statement is made as to the distance between the eye and the light. Ten eyes altogether were exposed, two being aphakic. The time of exposure was from five to ten minutes. In all cases there was marked conjunctivitis, keratitis, and iritis (?), but no changes were ever found in the lens capsular epithelium. In the retina the following changes were foimd in both the normal eye and the aphakic eye, but were more pronounced in the latter: chi*o- matolysis and formation of vacuoles in the cytoplasm of the ganglion cells with changes in the nuclei of the latter. Swelling and begin- ning collapse of the nuclei of the inner nuclear layer. Loss of chro- matism in the outer layer. The vacuolization of the ganglion cells in some cases persisted several weeks.* When a thick glass plate * In a footnote Birch-Hirschfeld stated that in one aphakic eye after an especially severe exposure to the iron arc he obtained well marked myelin degeneration of the optic nerve. He also stated that he would later give the details of this experiment, but we are unable to find that he has done so up to the present time. EFFECTS OF RADIANT ENERGY ON THE EYE. 689 was placed before the eye these changes did not occur. The exact length of time after exposure when the eyes were examined is not stated. It also is not stated whether or not the changes could be found if the eyes were removed immediately after the exposures. It is impossible for us to accept the findings in these experiments for the following reasons. In the first place the retinal changes described were widespread. To obtain a widespread illumination of the retina, however, necessitates the use of a quartz lens of extreme aperture. Birch-Hirschfeld does not state that he used such a lens. The widespread illumination of the retina also necessitates a greater intensity of illumination of the cornea and lens than of the retina. This together with the fact that the lens capsule receives in addition rays of much shorter wave length than can reach the retina makes it inconceivable that the retina could be injured under these conditions without the lens capsule also being affected. Yet Birch-Hirschfeld states that in neither series of experiments was the intensity and dura- tion of exposure sufficient to injure the lens. In fact the abiotic intensity was so slight that the corneal epithelium was destroNcd only in one experiment in which the cornea apparently became infected. On the other hand if we assume that Birch-Hirschfeld used no lens or a lens of ordinary aperture, the retinal lesions, if any, would have been circumscribed and would not often have been found by his method of examining the eyes. As we shall show, with sufficient light intensity small retinal lesions can be produced under these condi- tions, but they are due to heat and are entirely different from those described by Birch-Hirschfeld. The radiant energy used b.y Birch- Hirschfeld, however, was undoubtedly insufficient to produce such an effect. In his first series of experiments it is stated that the changes occurred immediately after the exposures. This is inconsistent with an abi- otic action of light, since with this there is always a latent period. Thus we found that the epithelial cells of the lens capsule showed absolutely no change if examined immediately after severe exposure to abiotic rays. Birch-Hirschfeld holds that the ganglion cell changes he describes represent a further stage of light adaptation. Yet he maintains that they are due to the direct action of the light on the ganglion cells themselves. He states that there is no reason to believe that certain cells are more susceptible to ultra violet light than others, yet he found changes in the retinal ganglion cells and none in the capsular epithe- lium in spite of the fact, just pointed out, that the latter must have 690 VERHOEFF AND BELL. received light not only of greater intensity but also of shorter wave length than did the retina. We cannot then accept Birch-Hirschfeld's findings because we were unable to obtain retinal changes, although we used light intensities and exposures sufficient to injure the epithelium of the lens capsule, the stroma cells and endothelium of the cornea (which he did not). Judging by the relatively slight histological changes found in the cor- nea by Birch-Hirschfeld, the intensity on the cornea of the light used by us must have been over fifty times as great as that used by him, while some of our exposures were nine times as long as his maximum exposure (10 minutes) to the iron arc. In the case of the aphakic eye, we obtained no ganglion cell changes in 48 hours although the light in- tensity was so great that the pigment epithelium showed heat changes in spite of an interposed water cell. In Birch-Hirschfeld's experiments the pigment epithelium was uninjured although a water cell was not used. Finally, we cannot accept Birch-Hirschfeld's findings because our experiments on monkeys and on a human patient prove conclu- sively that the function of the ganglion cells is not injured by light of the same wave lengths and vastly greater intensity than that reach- ing the retina in Birch-Hirschfeld's experiments. In connection with Birch-Hirschfeld's findings the following observa- tions relating to the ganglion cells of normal rabbit's eyes may be of significance. In the first place, within the same retina there is great variation in the amount of chromatin substance in the individual ganglion cells; two cells side by side may show a great difference in this respect.* The sharpness with which the Nissl bodies stain in thionin varies considerably with slight variations in the staining pro- cedure, particularly as regards the length of time the sections have been immersed in the thionin solution and the degree of dift'erentia- tion in alcohol. The same statement applies also to the intensity with which the nuclear layers stain. While the ganglion cells of a normal retina probably never contain actual vacuoles, the arrange- ment of the chromatin particles is not infrequently such that they enclose spaces which bear considerable resemblance to vacuoles. Occasionally a ganglion cell may contain an apparently degenerated nucleus, and occasionally also a more or less disintegrated ganglion cell is seen. Possibly the injury to the latter is produced by the microtome knife. * Nissl bodies cannot be seen in fresh ganglion cells so that it is possible that they are formed after death. The term chromatolysis may therefore be misleading inasmuch as it means solution of substances during Ufe which may have never actually existed. EFFECTS OF RADIANT ENERGY ON THE EYE. G91 These observations are in accord with those of Bach made twenty years ago. Investigating the possible effect of fatigue upon the ganghon cells of the retina, Bach ^^ made a careful comparison of retinae of rabbits exposed and unexposed to light. In some cases eyes were exposed to a Welsbach light for 20 hours. Alcohol or sublimate fixation was used and the sections stained by the original Nissl method or in thionin. Contrary to the previous observation of Mann^^^ and the later observations of Birch-Hirschfeld,^^ he was unable to find that the ganglion cells of the exposed eye differed in any way from those of the unexposed. He says: "Ich geh zu, dass ich langere zeit im Zweifel war und bald diese bald jene Veranderimgen gefunden zu haben glaubte, jedoch alles anscheinend Gefundene liess mich die controle wieder als Irrthum erkennen. Es ist eben zu bedenken, dass trotz gleicher Schnittdicke, trotz des genau gleichen Verfahrens beim Farben etc. immerhim sich tinctorielle Unterschiede ergeben konnen Ich muss bemerken, dass auch in normalen Netzhauten an den Ganglienzellen sich Unterschiede besonders hin- sichtlich der Menge vmd Anordnung, der Form der farbaren Plasma- schollen ergeben, das auch normalen Weise Vacuolen in dem Zellleib gefunden werden, das die Kerne sich verschieden verhalten konnen — kurz ich konnte an den beleuchteten Netzhauten Nichts warhnehmen oder vermissen was ich nicht an normalen, an verdunkelten Netz- hauten auch wahrgenommen oder vermisst hatte." Birch-Hirschfeld ^^ states that also by means of the vital methylene blue staining method he found chromatolysis, vacuolization, and other changes in the retinal ganglion cells of eyes that had been exposed to ultra violet light, and that such changes were absent in normal eyes. I find, however, by this method in the normal rabbit's retina, appear- ances that correspond exactly to those described and depicted hy Birch-Hirschfeld, including particularly the "vacuoles" in the gan- glion cells which are abundantly present. The "vacuoles" at first glance appear to be really such, but a careful study of them strongly- suggests that they are here due to the cell reticulum staining more promptly and deeply than the cytoplasm proper, and thus producing an appearance of rounded spaces. I have also examined the retina by the vital methylene blue method 48 hours after the exposure of one hour to the magnetite arc and lens system (Exp. 78). The results obtained were identical with those obtained in the case of an un- exposed normal retina. In addition to the experimental investigation just discussed, Birch- Hirschfeld^^ has reported clinical observations in five cases of pho- 692 YERHOEFF AND BELL. tophthalmia following exposure to the mercury vapor lamp that he claims demonstrate the pathological action of ultra violet light upon the retina. In these cases he found, for colors only, para- or peri- central scotoma, central relative scotoma, and constriction of the peri- pheral field. Later ^* he reports that after an exposure of less than I hour to the Schott uviol lamp he himself was affected with mild photophthalmia followed by color field changes. He found a relative color scotoma in each eye beginning 15° from the fixation point that persisted 6 days. After this had completely disappeared he exposed his left eye to the same light through a colorless glass obstructing all waves less than 330 fxn in length and obtained no changes of any kind. He regarded this as proof of his contention that the field changes previously obtained were due chiefly to waves between 300 n/j, and 330 /XM- -"^s a matter of fact, however, Hallauer ^^^ has shown that the adult human lens absorbs all waves less than 376 mx and most of those less than 400 fxij. so that this experiment of Birch-Hirschfeld proves, if it proves anything, that the field changes obtained in his clinical cases and in his own case were chiefly subjective or at least did not represent pathological conditions. Moreover, as pointed out elsewhere (page 721) Birch-Hirschfeld*^ himself has recently taken exception to the similar field changes reported by Jess ^^^ in cases of eclipse blindness on the ground that they might well have been obtained in normal eyes. Thermic Effects of Radiant Energy on the Eye. The Cornea. In passing through the cornea, light of any wave length is absorbed to some extent. Waves less than 295 fxij. are completely absorbed while those over 315 m/x in length (judging by the results of our experiments) are very slightly absorbed. The absorption of the latter is no doubt due in part at least to the lamellae of the cornea and the corneal corpuscles, which cause internal reflections and refrac- tions, especially of the relatively short waves. With ordinary light intensities the amount of energy absorbed is so slight that no heat effects are produced, but with extreme intensities it is obvious that the latter could be produced even in the case of visible rays. In five of our experiments definite heat eftects were observed in the cornea. That the effects were due solely to accumulated heat and not in any degree to abiotic action, is proved by the character of the changes EFFECTS OF RADIANT ENERGY ON THE EYE. 693 produced, and by the fact that the epithehal cells of the cornea and lens were unaffected. The screens were such that the lens received waves of the same wave lengths as did the cornea. The corneal epithelium was unaffected probably owing to its being cooled by con- tact with the air. In no instance did the heat reach sufficient inten- sity as to cause pain. The most marked heat effect on the cornea was obtained in Exp. 88 in which the rays from the magnetite arc after passing through a flint screen and water cell were concentrated for one hour sharply upon the cornea by means of the quartz double lens system. Toward the end of the experiment it was discovered that the water cell had leaked, so that for an unknown length of time the eye had been exposed to infra red rays in addition to the shorter waves. This undoubtedly accounts for the fact that in no other experiment was such a marked heat effect produced, and that no effect was produced in Exp. 85 in which the same conditions obtained except that the water cell did not leak and the exposure was longer. 24 hours after the exposure the affected area was hazy and swollen but the eye was free from inflammatory reaction. On microscopic examination 48 hours after the exposure the epithelium was everywhere normal. The stroma was swollen to over twice its normal thickness and stained faintly in eosin. Within the central portion of the exposed area not a single corneal corpuscle could be seen. At the periphery the transi- tion into normal cornea was abrupt as regards the corpuscles but relatively gradual as regards the stroma. In the transition region the corpuscle towards the normal side were in active proliferation, many of them showing mitosis, while from here inward they suddenly became invisible. The endothelium in the exposed region was for the most part completely absent, but in some places a few faintly stained cells still adhered to Descemet's membrane. The cornea was ever^-where practically free from leucocytic infiltration. The iris showed a few minute hemorrhages around the pupil undoubtedly due to heat, since, as stated, the lens capsule was unaffected. (PI. 1, Fig. 3.) In the second experiment (Exp. 92) sunlight was focussed 45 minutes upon the cornea by means of a large quartz lens after passing through a blue uviol screen and a .001% aqueous solution of auramine O. Here the heat effect was similar but less marked than that just described. The effect on the corneal corpuscles was about as great, and the appearance of the stroma about the same except that it was much less swollen. The corneal epithelium, the iris, and the lens epithelium, were unaffected. ^^694 VERHOEFF AND BELL. The third experiment (Exp. 81) was similar to the first except that the flint screen allowed waves down to 305 ijlh to pass and that the water cell did not leak. The exposed corneal area was clear immedi- ately after the exposure, but 20 minutes later was found to be distinctly hazy. The epithelium at no time stained with fluorescine. On microscopic examination of the eye, enucleated tliree days after the exposure, the corneal stroma was found to be swollen in a rather sharply defined area. The epithelium was normal. The corneal corpuscles in the middle third of the cornea showed active prolifera- tion, but in the posterior third had for the most part disappeared. The endothelium was absent behind the exposed area. The iris and lens epithelium were normal. In the fourth experiment (Exp. 84) a flint screen transparent to waves down to 310 /x/x was used and the exposure was one and one half hours. The exposed corneal area was found to be hazy within one hour after the exposure. On microscopic examination of the eye, enucleated four days after the exposure, the corneal stroma was found very slightly swollen and to stain less strongly in eosin in its posterior layers. The corneal corpuscles showed marked proliferation in the posterior portion of the stroma and the endothelium was absent behind the exposed area. The iris and lens epithelium were normal. In the fifth experiment (Exp. 90) a flint screen (Slii fxfx) was used and the conditions were the same as in the first experiment with the important dift'erences that the water cell was omitted and the exposure was only 30 minutes. Distinct haziness of the cornea was observed within 20 minutes and within 24 hours became very marked. On microscopic examination (48 hours) the cornea showed changes similar to and almost as marked as those of Experiment 88. The corneal epithelium and lens epithelium were unaffected. Combined Thermic and x\biotic Effects of Radiant Energy ON THE Cornea. In four other experiments in which the exposures were prolonged, both abiotic and heat effects were obtained in the cornea. The screens used were transparent to waves less than 30.5 /jl/j. to 298 fx/j. in length and the exposures were from one to one and a half hours. In two of the experiments (Exps. 78 and 79) abiotic effects were indi- <;ated by loss of corneal epithelium and characteristic changes in the EFFECTS OF RADIANT ENERGY ON THE EYE. 695 lens epithelium, and heat effects by the haziness of the cornea occurring within 30 minutes after the exposure as well as by the slightness of the conjunctival reaction. Combined effects were also shown by the microscopic examinations, the corneal corpuscles being completely invisible in the posterior layers of the cornea, but present and showing characteristic abiotic effects in the anterior layers. The other two experiments were of especial interest because they showed the limit in respect to wave length beyond which we were unable to obtain abiotic effects. The same screen 305 /x^u was used as in Exp. 81 just referred to, in which only heat effects were obtained, but the exposures were longer. In Exp. 82 the exposure was 1| hours and haziness of the cornea was noted within 30 minutes afterwards. The lens epithelium was unaffected, and the only evidence of abiotic action was a slight loss of corneal epithelium occurring after 24 hours. The heat effects on the other hand, were very marked. The corneal endothelium was destroyed in the exposed region and only an occa- sional corpuscle could be seen even in the anterior layers of the stroma, so that any possible abiotic effects on the corneal corpuscles were masked by the heat effects. At the periphery of the exposed area the corpuscles were in active proliferation at the end of 48 hours. In Exp. 83 the conditions were the same except that the light was focussed upon the surface of the lens instead of upon the cornea, and that the exposure was interrupted for an hour at the end of the first 30 minutes. Haziness of the cornea was noted within 50 minutes after the first exposure. Following the second exposure there was no loss of corneal epithelium and the only evidence of abiotic action was a slight effect on the lens epithelium. The cells were slightly swollen in a small area and a few of them contained characteristic granules. The heat effect on the cornea was about the same as in the preceding experi- ment. According to the foregoing experiments with the double lens system, after an exposure of if hours through a water cell and flint screen (315 /xyu) no changes are produced; after an hour's exposure through a screen (310 /x^t), slight heat changes; through screen (305 mm) marked heat changes; and through still more transparent screens, marked heated changes combined with abiotic effects. With the flint screen (315 nfi) but without a water cell, marked heat changes are produced after 30 minutes exposure, the heat effect of the short waves here being reenforced by infra red waves. It is evident from these results that the specific absorption of the cornea with respect to wave length does not end abruptly, but gradually diminishes from 295 nfx to some- 696 VERHOEFF AND BELL. what beyond 315 /x/i. It is also evident that the energy absorbed from waves of 305 /x/x or over in length, is converted almost exclusively into heat, only the slightest traces of abiotic action being obtained with weaves of 305 jUjU in length after the most intense and prolonged exposures. It will be seen that the abiotic effects and heat effects of radiant energy upon the tissues are essentially diiferent. In the case of heat, a certain critical temperature is required before any effect is produced. This is shown by the sharp transition from normal into injured corneal corpuscles at the periphery of the exposed area, and also by the fact that the epithelium, being kept cool by contact with the air, remains unaffected. The heat effect therefore does not vary in direct ratio with the intensity of exposure, obviously due to the fact that dissipa- tion of heat enters into the equation. In the case of abiotic action on the other hand the effect varies directly with the intensity of the exposure. Heat of an intensity just below that sufficient to cause cell destruction, causes cell proliferation. Abiotic action does not directly cause cell proliferation no matter how intense or how slight the exposure. Lastly, heat does not produce the eosinophilic and basophilic granules in the cytoplasm that are produced by exposure to abiotic radiation. On the other hand, while it is evident that heat does not produce effects similar to those produced by moderate exposures to abiotic waves, extreme exposures to the latter may produce effects not unlike the severe effects of heat. Thus we have shown that se\'ere exposures to waves shorter than 295 /x^t in length may lead to complete disap- pearance of the corneal corpuscles and marked swelling of the corneal stroma. In the case of heat, however, the posterior layers of the cornea are more affected than the anterior la.^ers while in the case of abiotic action the reverse is true. Thermic Effects of Radiant Energy upon the Iris and Lens. In Experiment 97 in which the eye was exposed for one minute through a uviol screen to sunlight concentrated by the large mirror, the pigmented iris was severely burned in the exposed area, showing complete hyaline necrosis. The lens epithelium examined, after 48 hours, was unaffected in the pupillary area, but beneath the pupillary EFFECTS OF RADIANT ENERGY ON THE EYE. 697 margin it showed an incomplete ring which under the low power of the microscope resembled the wall produced in other experiments hy abiotic radiations. Examination under a higher power however, showed that the appearance was due chiefly to the fact that the cells were here in a state of active proliferation, almost every cell being in some stage of mitosis. It was evident that the heat from the pigment layer of the iris, where the latter was in contact with the lens capsule, had stimulated the cells of the latter to proliferation. It is note- worthy that in Experiment 99 in which the exposure was 1| minutes but in which the iris was unpigmented, neither the iris or lens capsule was affected. In none of our experiments was the lens injured by the heat gener- ated by the stoppage of rays within its own substance. That clouding of the lens can be so produced, however, even by visible rays alone, with sufficient intensity and prolonged exposure, has already been demonstrated by Czerny ^^ and Deutschman ^^ in the case of sunlight, and by Herzog ^^^ who used the carbon arc and suitable filters. The iris in no other experiment showed heat effects comparable to those just described. In most of the experiments with the magnetite arc and double lens system the iris was not greatly exposed to the light owing to the artificial mydriasis, but in Experiment 88 in which the most intense heat effect was obtained in the cornea, the iris showed hemorrhages near the pupil. In Experiments 83 and 90 the iris became greatly contracted towards the end of the exposures, and remained so for several hours, but again diluted within 24 hours. Thermic Effects of Radiant Energy upon the Retina. In a number of our experiments, some of which were made with other purposes in view, we obtained heat effects in the retina in spite of an interposed water cell 5 cm. thick. They were obtained mainly in two ways, one by the use of sunlight reflected from a silvered glass concave mirror 26 cm. in diameter and 1.5 meters in focal length, and the other by the use of the magnetite arc light concentrated l)y the single quartz lens system. A full description of the mirror and the calculated energy derived from it is given on page 721. The calcu- lated energy on the retina given by the c^uartz single lens system is given on page 724. The burns were obtained through screens that obstructed all waves less than 335 fxfx in length as well as througli 698 VERHOEFF AND BELL. more transparent screens. In the case of sunlight the exposures were from one-fourth second to one and one-half minutes, and the resulting burns were always severe, the retinal tissue being actually coagulated as will be described. In the case of the magnetite arc the exposures w^ere from ten minutes to one hour and the burns were much less severe. In addition to these, heat effects involving the pigment epithelium alone were obtained in two experiments with the quartz double lens system (Exps. 88 and 89) in each of which a large area of the fundus was illuminated. One of these was in the case of an aphakic eye, and the other in a case of exposure without a water filter. The significance of these experiments in connection with the questions of eclipse blindness and allied phenomena is discussed elsewhere (page 720). That the severe effects produced by concentrated sunlight were due to heat was obvious from their histological appearances and from the fact that the light intensity at the focus was found in all cases to be sufficient quickly to ignite a match or piece of paper. That the relatively slight effects produced by the magnetite arc were also due to heat, was obvious from the fact that only the pigment epithelium and outer retinal layers were affected and sometimes the pigment layer alone. If the effects had been due to the abiotic action of light the inner nuclear layer and ganglion cells would necessarily have been equally or even more greatly affected. Moreover, as we have already shown, when the corneal epithelium and lens epithelium were exposed to light of greater intensity and shorter wave lengths than was the retina in these experiments, and for a much longer time, no changes were produced in them. Thus in Experiment 53 a heat effect in the retina was obtained after 12 minutes exposure to light passing through the lens of the eye, that is, to waves longer than 330 /x/x, whereas in Experiment 85 no effect was produced on the cornea after an exposure of if hours to light of greater intensity containing wave lengths as short as 315 juju. This is easily explicable on the assumption that the retinal changes under consideration were due to heat, since the cornea and lens must each absorb a far less proportion of visible and infra red rays that reach them than does the pigment epithelium of the retina. On the other hand it is absolutely inexplicable on the assumption that the retinal changes were due to abiotic action, since it is inconceivable that the corneal and lens epithelium would be un- affected by abiotic action of light sufficient to produce nuclear frag- mentation in the outer muclear layer and pigment epithelium. The character of the histological changes clearly indicates that the heat EFFECTS OF RADIANT ENERGY ON THE EYE. 699 conversion took place chiefly in the pigment epithelium and inner layers of the chorioid, and that the outer layers of the retina proper were affected b\^ the heat conducted therefrom. An interesting problem is the exact determination, under various conditions, of the minimum intensity and duration of exposure to visible and infra red rays necessary to produce heat effects on the ret- ina. A discussion of this problem will be found on pages 721 and 732. The experiments in which heat effects on the retina were obtained by means of the magnetite arc were Experiments 53, 55, 57 58, 59, 88, 89. Following are the experiments with sunlight concentrated by the large mirror. Experiments. SunligJit Focusscd on Cornea by Large Mirror. Experiment 95. Without water cell or screen. Pigmented eye. Three exposures, j second, | second, and 10 seconds respectively. No inflammatory reaction. Enucleation at end of 33 days. Lens epithelium normal. Three burned areas in fundus of different grades of severity. Experiment 96. Water cell. Flint glass screen (335 ^t^t). Albino. Exposed 4 seconds. No inflammatory reaction. Lid reflex to light abolished. Enucleation at end of 48 hours. Two contiguous bin-ned areas in fundus, one exactly on disc. Marked hemorrhagic retinitis. Slight hemorrhage from retina into vitreous. Experiment 97. Blue uviol screen. No water cell. Pigmented eye. Exposed 1 minute. After 1 hour: Pupil contracted to | normal size, does not react to light. After 24 hours: Lid margins inflamed, lower one ulcerated, no lid reflex to light. Cornea clear. After 3 days : Cornea shows purulent infiltrate below (infected from ulcerated lid). Enucleation. Fundus shows two contiguous burned areas, one at margin of disc. Microscopic examination: (3 days): Slight purulent infiltration of cornea. Hyaline necrosis of iris. Lens epi- thelium normal in pupillary area, shows proliferative changes beneath pupillary margin (heat effect due to contact with heated iris. See page 696). Experiment 98 (PI. 4, Fig. 14). Water cell. No screen. Albino. Exposed 14 seconds (misty day). No inflammatory reac- 700 VERHOEFF AND BELL. tion. Enucleation at end of 6 days. Fundus shows burned area just beneath optic (Hsc. Lens epitheHum normal. Experiment 99. Blue uviol screen. Water cell. x\lbino. Ex- posed 1| minutes. No inflammatory reaction. Enucleation at end of 12 days. Fundus shows burned area undergoing repair. Cornea, lens epithelium, and iris normal. Experiment 100. Blue uviol screen. Water cell. Albino. Total exposure, 10 minutes, one second on, one second off. No inflamma- tory reaction. Enucleation at end of 7 days. Retina shows no burned areas. Cornea, lens epithelium, and iris, normal. Experiment 101. Blue uviol screen. No water cell. Pigmented eye. 220 exposures, I second each, with intervals of 1 to 3 seconds. No inflammatory reaction. Fundus normal. Character of the Thermic Effects Produced in the Retina. In the experiments in which the pigment epithelium alone was affected no changes were noted macroscopically. In most of the other experiments the lesions could be seen with the ophthalmoscope or better still on opening the eye after enucleation. They appeared as sharply defined reddened spots. Some of those obtained after ex- posure to sunlight showed blood extending from them into the vitreous humor. Some of the spots were observed only after the eye was placed in Zenker's fluid. In Experiment 96 in which a burned area involved the optic disc, there was intense hemorrhagic retinitis ap- parently due to thrombosis of the central vein. The spots produced by sunlight measured about 2.5 mm. in diameter. Those produced by the magnetite arc and single lens system were about 3 mm. in diameter as measured under the microscope with reference to the effects on the pigment epithelium, but only about 1 mm. in diameter as measured with reference to the effects on the retina proper when this was involved. This concentration of the effects in the center of the area was no doubt due to two facts, one being that the light was actually more intense here, and the other that towards the periphery of the area the heat generated in the pigment epithelium became rapidly dissipated. Microscopical: The most striking feature of all the burned areas whether due to long or short exposures was their sharp demarcation, illustrating again here as in ^e case of the cornea how sharply critical EFFECTS OF RADIANT ENERGY ON THE EYE. 701 is the temperature required to injure tissues. In all cases the pigment epithelium was the most severely affected of any portion of the retina, and in the slightest burns it alone was affected. This was true also in albinotic rabbits in which the epithelium was free from pigment. The other structures were affected in the following order according to the intensity of the action, the rods and cones, the chorio-capillaris, the outer nuclear layer. The inner nuclear layer, the ganglion cells, and nerve fibre layer, were affected only after extremely intense exposures and in our experiments were not affected after exposures to the magne- tite arc but only when concentrated sunlight was used. The slightest change that can be definitely made out in the pig- ment epithelium 48 hours after exposure consists in the cytoplasm of a greater or less number of cells staining intensely in eosin. When the effect is somewhat greater, vacuoles appear and may be so large and numerous that the cells appear almost completely transparent, the cytoplasm showing a delicate reticulum with the minute nodes at the junction points. In case the epithelium is pigmented the pigment appears to be separated from the membrane of Bruch by large vacuoles, and the nucleus may show marked pycknosis. When the effect is still greater the cell reticulum completely disappears leaving only the nucleus in the clear space thus formed. The nucleus may show fragmentation or simply chromatolysis. The basophilic and eosino- philic granules characteristic of abiotic action are not seen. In the somewhat more severe burns the pigment cells entirely disappear leaving only the pigment. In one eye examined six days after ex- posure the injured epithelium is found replaced by epithelium which has evidently grown in from the periphery, but between the new layer and the rods and cones numerous swollen vaculated and otherwise altered pigment cells remain. The changes in the pigment epithelium are best seen in plane section, and to determine the character of the slightest changes it is important to compare the appearances seen in the exposed eye with those of a normal eye. Forty-eight hours after exposures sufficient to affect the rods and cones, the outer limbs of the latter are found to be broken up into coarse granules, while the inner limbs are swollen to large bladder-like structures each containing a few fine granules. There may also be a greater or less number of red blood corpuscles among the rods and cones due to diapedesis from the chorio-capillaris, and also a certain amount of serum. When the outer nuclear layer is aft'ected, the nuclei lose their peculiar cross striations, becoming intensely pycknotic and some of them undergoing fragmentation. With this degree of 702 VERHOEFF AND BELL. injury the pigment cells have disappeared, and fragmented nuclei can sometimes be seen in the inner layers of the chorioid. The inner layers of the retina, including the ganglion cells, remain normal in appearance. (PI. 4, Fig. 12.) In the experiments in which sunlight was used, as already noted, much more intense heat effects are found in the retina and chorioid. Here the inner layers of the retina are disintegrated while the outer layers appear intact. This is undoubtedly due to the fact that the latter have been coagulated and thus fixed by the heat. At the periphery of the area the appearance is reversed, the inner layers being normal while the outer layers show the less marked changes already described. Two to six days after such an exposure the coagulated rods and cones maintain their normal appearance except that they stain abnormally deeply in eosin. The nuclei of the external nuclear layer are likewise coagulated, but have lost their cross striations and stain more deeply than normal. The internal nuclear layer shows different appearances in different places evidently according to the heat intensity. In some places the nuclei still take the basic stain and show fragmenta- tion. In others they stain in eosin and are not disintegrated, while in still others they have entirely disappeared. The nerve fibre layer is completely disintegrated and the ganglion cells have entirely disap- peared or stain only in eosin. At the periphery, the ganglion cells with their Xissl bodies stain less and less in thionin as the burned area is approached until they become entirely eosinophilic. The inner surface of the retina is in some cases coated with a thick layer of fibrin. The pigment epithelium is coagulated and retains its normal position in the exposed area. At the periphery it is disin- tegrated. The chorioid behind it shows large extra vastions of blood and marked nuclear fragmentation. There is no cellular infiltration, purulent or otherwise, of either the chorioid or retina. (PI. 4, Fig. 14.) After two months the retina is found replaced by neuroglia con- taining migrated pigment cells. In some cases the chorioid is appar- ently normal and the pigment epitheliiun reformed. In others the latter is absent and the chorioid replaced by two or three layers of a vascular fibrous tissue. EFFECTS OF RADIANT ENERGY ON THE EYE. 703 Theory of Action of Radiant Energy on the Tissues. A useful conception of the effects of radiant energy upon the tissues of the body is that the heat effect is due to increased molecular motion while the abiotic effect is due to direct atomic disintegration of the molecules with immediately resulting chemical changes. The first effect of the increased molecular motion is to produce a physical change analogous, for example, to the melting of ice. When the motion reaches a certain critical rate the molecules are broken up and various chemical changes result. Both heat effects and abiotic effects theoretically may be produced by rays. of any wave length, but practically in the case of short waves the heat effect is generall\' negligible, while in the case of long waves the abiotic effect is negli- gible. Our experiments show that for human cells the abiotic effect becomes negligible within a ^'ery short range of wave lengths, that is between 305 and 310 /i/x. For bacteria it becomes negligible still sooner, at less than 295 /x/i. Under ordinary conditions heat effects are also negligible here, and in fact all through the visible spectrum, although with extreme intensities such as afforded by concentrated sunlight they may be produced, as in eclipse Ijlindness, for instance. It is in reality due to the fact that abiotic effects and heat effects are negligible in the region of the spectrum indicated, that sunlight under usual conditions is not destructive to human life. This fact, considered from the standpoint of evolution, suggests a relation of light to the origin and structure of living matter, but a discussion of this aspect of the subject would lead too far. Since according to this conception the abiotic action of light is directly upon the structure of the molecules, slight chemical changes are produced after almost infinitesimal exposures. Theoretically, of course, there is a limit of exposure below which no disintegrating effect is produced upon the molecules, so that a series of such short expo- sures would produce no summative effect. Practically, however, this would be impossible to demonstrate in the case of li\ing cells. On the other hand, in the case of living cells summation of the effects of a series of exposures, if the intervals were too long, would not accurately occur, since the repair of the injury would take place to a greater or less extent. Thvis we have found in the case of the corneal cells that summation of effects becomes much less exact when the intervals of exposure are over twenty-four hours. 704 VERHOEFF AND BELL. In the case of heat, deleterious effects on the tissues must Hkewise be due to chemical changes. Since, however, these changes take place only when the molecules have reached a certain rate of motion, under ordinary conditions a measurable time interval must elapse before they begin. The length of the time interval depends upon the intensity of the light and upon the rapidity with which dissipation of heat occurs, and thus varies greatly under different conditions. Under ordinary conditions, however, the time interval is of consider- able length, so that a series of exposures does not produce a total effect equal to that of a continuous exposure of the same total length, and may not produce any effect at all. From a practical standpoint therefore this fact constitutes a fundamental difference between the abiotic effects and the heat effects of radiant energy. Unless light rays are absorbed by substances they can of course produce no effect upon them. Thus, as we have shown, waves over 295 iJLfjL in length unless extremely intense have no effect on the corneal stroma which is relatively transparent to them but have a markedly deleterious effect on the corneal corpuscles which absorb them. It does- not necessarily follow, however, that because light rays fail to pass through a given substance they must produce an effect upon it. For they may simply be changed into light waves of longer wave length (fluorescence) or their energy dissipated in the form of heat of an in- tensity too low to produce any changes. Both of these transforma- tions must take place in the case of the lens of the eye since light waves are constantly being stopped in it. That fluorescence actually occurs in the lens is, of course, well known and easily demonstrated. Assuming that the abiotic action of light of given wave lengths upon protoplasm is directly proportional to the coefficient of absorp- tion of the protoplasm for, that wave length, Henri ^^^ and his wife have determined this coefficient for egg albumin and a large number of waves. The curve plotted from their results given elsewhere (page 645) shows that the absorption becomes practically nil at and near 310 /x/x, so that the abiotic action must be very slight here. Moreover, since this method does not allow for the fact that the absorbed rays produce heat as well as abiotic effects, the abiotic action is undoubtedly less than is indicated by the curve of absorp- tion. These results, therefore, confirm in a striking manner those obtained by us by actual experiments on the cornea. The preceding discussion has concerned mainly the direct effects of light upon the molecules of the tissues without reference to histo- logical and clinical manifestations. The latter are of course, too EFFECTS OF RADIANT ENERGY ON THE EYE. 705 complicated for complete analysis since fully to understand them would require a knowledge not only of the chemical changes originally produced, but of the vital processes of the cells. An interesting question in regard to them is that concerning latency of their appear- ance. In the case of abiotic action, as has been pointed out, abso- lutely no visible change either histological or clinical takes place immediately after the exposure, and usually not for several hours. This is no doubt due to the fact that time is required for the chemical changes to produce physical alterations. In the case of heat effects, it is a matter of common experience that latency also occurs and that the time interval varies with the intensity of the exposure, but it is a far less striking phenomenon than in the case of abiotic effects. This may be due, among other causes, to the fact that the energy required to produce chemical changes by heat is so great owing to the rapid dissipation of the latter, that under ordinary conditions the critical point is quickly passed and an excessive effect produced. Abiotic Energy in the Solar Spectrum. As has already been noted the solar spectrum when filtered through a thick layer of atmosphere as at sea level when the sun is low fades out at about 305 njx. At high altitudes and with the sun running 0 -x-lO^^ a /^ "\ 4 3 / / /-^ ^^-^ <\ 1 / ^ k. / ' r^^^^=- 1 / ^^"^ ^^^ n y """^''''^ 700 800 900 Mt. Whitney z. d. 1300 yU/U 0° h. Mt. Whitney z. d. = 60° Figure 5. Distribution of energy in solar spectrum. high, it extends to about 295 /x^t- Under extremely favorable condi- tions some very faint traces of the spectrum were registered by Cornu ^^ down nearly to 292 ix\x. 700 VERHOEFF AND BELL. But substantially the whole of the solar spectrum which is capable of producing abiotic action lies between 295 ^t^t and 305 ixjx, is evanescent under most conditions, and only possesses pathological significance at high altitudes and especially in extreme cold. Thei'e is good reason to believe that the atmosphere is considerably more permeable to ultra violet radiations at low temperatures than under ordinary con- ditions, particularly as regards the extreme radiations. Figure 5 shows from the data of Abbot the distribution of energy in the so- lar spectrum in curve (a) at Mt. Whitney for a zenith distance of 0°, in curve (b) also at Mt. Whitney (14,000 ft.) but for zenith distance 60°. Near the latter limit lies the general range of solar radiation as ob- served at the surface. Two things in these curves are particularly noteworthy, first that in both and especially at the higher altitude the maximum radiation and indeed the bulk of the radiation in general lies within the visible spectrum. Second, the maximum energy lies not in the red, but in the case of the high altitude energy fairly in the blue at about wave length 470 ixix and at the lower altitude in the green at wave length about 500 /x/x. So far as the solar spec- trum is concerned, therefore, the heat energy is chiefly within the visible spectrum. No distinction therefore can be drawn between the visible and the infra red spectrum on the ground of heat radiation and all attempts to separate thermic effects by cutting out the visible spectrum are therefore futile. So long as this reaches the eye it carries with it the solar heat in its greatest intensity. From the area of the curves here shown it appears that of the energy at high altitudes only a very small proportion, of the order of magnitude of one quarter of 1% lies within the region 295 to 305 ^ipt. Even this small quantity is evanescent at the sea level and at ordinary temperatures. It is to the small remaining trace of abiotic rays here noted that the phenomena of snow blindness are due. From the clinical standpoint snow blind- ness is found to occur only as a photophthalmia of relatively very mild degree and under exposures usually for a long period and either at very high altitudes or very low temperatures or with both these condi- tions concurring. On snow fields the exposure of the eye to solar radiation, ordinarily greatly ameliorated by the obliquity of the inci- dence, is rendered much more severe by the reflection from the snow which is a good reflector down to the extreme ultra violet of the solar spectrum. One would not go far wrong in estimating that the radiation reaching the eye under such circumstances is of the order of magnitude of a million ergs per square cm. per second. A single square meter of snow at 2 meters distance would reflect to the eye EFFECTS OF RADIANT ENERGY ON THE EYE. 707 almost a tenth of this amount with the aUitude and sun favorable. Assuming now that one quarter of 1% of this quantity, that is 250 ergs per square em. per second is within the abiotic region 295 to 305 /x/x it is easy roughly to determine the exposure which is likely to produce snow blindness. We have already seen that a well marked pho- tophthalmia can be set up by a radiation in abiotic rays of about 2,000,000 erg seconds per square cm. Now assvnning that of the total radiation which would be received direct, half, through direct and reflected action, reaches the eye of one traveling among the high snow fields. The energy in total abiotic radiations would be about 1250 ergs per square cm. per second. If all of this cjuantity had the average abiotic effect on the conjunctiva and the cornea a little less than 27 minutes exposure would be required to make up the 2,000,000 ergs seconds just referred to and to produce symptoms of photophthalmia. As a matter of fact the region from 305 /x/x to 295 fxfj, has much less than the average abiotic effect. Our crown glass screen ^ 7 cuts ofP the ultraviolet at 295 /i/x substantially just at the end of the solar spectrmn. Experi- ments made with this screen on the magnetite arc which is fairly strong from 295 ^i/x to 305 ju^u showed that this screen increased the exposure necessary to produce photophthalmia eighteen times. It therefore appears that at a high altitude in the snow fields an exposure of 7 to 9 hours under extreme conditions would be required to produce photophthalmia as severe as that which we have here recorded as typical, i. e., involving stippling of the corneal epithelium, Clinically snow blindness very rarely reaches this phase, since, although the exposures may be long the intensity of abiotic solar radiations reaching the eye would be seldom as great as the maximum amount just mentioned. For instance Schiess-Gemusens ^^* reports two cases of ordinary snow blindness which fell under his own observation in which the ordinary symptoms occurred after practically all day exposures showing very marked conjuncti^'itis without any visible effect on the cornea. Inasmuch as the exposures in casual climbing on an all day trip are considerably less severe than with steady full exposure to the snow fields, it is fair to assume that this latter condi- tion might produce snow blindness in perhaps half the time previously mentioned. This checks well with our experiment on solar erythema where an exposure of 6 minutes at .5 meter from the magnetite arc unshielded gave a slight l)ut definite erythema of the skin. At sea level and under ordinary circumstances the critical exposure for snow blindness would undoubtedlv run to manv hours. It is well 708 VERHOEFF AND BELL. known clinically that snow lilindness has often been reported in polar exploration. In high latitudes at sea level the abiotic energy is greatly reduced but three circumstances enter the case to increase the danger of snow blindness. First the hours of sunlight are very long, second, intense cold is believed to decrease the atmospheric absorption for the extreme ultra violet, and third, the exposure of the eye to prolonged and intense cold, while it may not actually lower the vitality of the cells to render them more easily attacked by abiotic radiation, unquestionably would tend to lower their recuperative power and so effect the summation of exposures which ordinarily would be relieved by continuous repair. Solar Erythema. These data on solar energy at once call up the question of solar erythema generally attributed to the effect of ultra violet radiation.* Clinically this bears a suggestive resemblance to photophthalmia in that it has a period of latency and a similar period of duration. Further it is well known to occur easily at high altitudes with the sun running high, that is under circumstances which afford a fair amount of abiotic rays. The best recent investigation of this matter is that by Dr. deLaroquette,^^° Surgeon Major of the French Army in Algiers. His experiments under the intense tropical sun show the connection of solar erythema with the abiotic rays very clearly. In the first place in most cases he noted a primary erythema clearly due to temperature and perhaps associated with heated air as well as radiation, occurring only when the temperature is 30 degrees C. or more. This is followed after a period of latency of an hour or two by a secondary photochemical erythema going on under severe expo- sures to hemorrhagic pigmentation, local oedema and subsequent desquamation. Experiments in exposure of the skin under screens showed under layers of quartz and water, both of which are highly transparent to abiotic rays, the secondary erythema was well marked. * It may be mentioned here that there are certain rare chronic affections of the skin, notably xeroderma pigmentosa, that are believed to be due to exposure to day light, chiefly because they involve only the exposed surfaces of the body. It is supposed that for some unknown reason the skin is in such cases abnormally sensitive to, abiotic radiations. Possibly other slight irritants applied for the same length of time would produce similar effects. Two cases have been recorded in which the cornea was involved, and one of us has per- sonally examined such a case. EFFECTS OF RADIANT ENERGY ON THE EYE. 709 Under window glass and violet or blue glass it was slight, even after considerable exposures, while with yellow, red, green or black glass it was absent, although in each case the primary erythema was marked. No investigation was made of the absorption of the various glasses, but from our experiments the clear, the blue and the violet glasses are likely to let through the margin of the abiotic radiations in the thickness, 2 mm., here employed. Yellow, red, green and black glasses would certainly cut these off. The skin in open exposure to sunshine is very much more exposed to the full energy of the solar radiations than is the surface of the cornea and conjunctiva and for abiotic effects on the skin practically the full strength of solar radia- tion is available. One would therefore expect to get action from the abiotic rays in, at most, half the time noted with respect to the cornea and conjunctiva for a similar degree of effect. In other words, one should get in a couple of hours well marked effects and undoubtedly slight erythema in an hour or so, as experience well shows is the case, assuming somewhat similar degree of sensitiveness in the epithelial cells. In case of extreme exposure to heat radiation distinct heat effects may be found in either case. Dr. deLaroquette's observations on the human skin were fully checked by exposures on shaven areas on the skin of a guinea pig showing the same general phenomena. Dr. deLaroquette also suggests that low temperature and wind drying the epidermis and provoking intense superficial vaso-constriction tends to exaggerate solar erythema. In this way some rational account can be given of its occurrence under conditions of cold and severe wind alone w^hen the abiotic action of the solar radiation would be small or even wanting, in which case the effect would be a primary rather than a secondary one. Finally, in solar erythema, as in pho- tophthalmia, repeated exposures of somewhat subnormal intensity give acquired tolerance, while the skin is, as well known, somewhat hypersensitive to severe exposures following each other without time for the lesions to undergo repair. Our experiments with the bai'e magnetite arc as source indicate that the liminal exposure for perceptible abiotic effects is practically the same for the more sensitive parts of the skin as for the conjunctiva. The inner portion of the forearm was the portion of the body exposed in our w^ork on liminal exposures. Here with 6 minutes at .5 meter, which corresponds very well with the production of mild photoph- thalmia, a slight reddening of the skin appeared some few hours after exposure, rose to its maximum inside of the first 24 hours and vanished within a day or tw^o leaving no trace. Through the double 710 VERHOEFF AND BELL. lens system and crown glass screen (295 (x/jl) already described, the lim- inal exposure was between 15 and 30 seconds, the former figure giving no traces and the latter slightly more than a liminal exposure. In all exposures over half a minute there was immediate heat erythema and a subsequent development after a period of latency of a few hours. There was a distinct but slight feeling of heat during the exposure and a rather rapid extension of the erythema somewhat beyond the limits of the 5 mm. stop which limited the area exposed. In cases of severe exposure to the sun we are inclined to think that this primary erythema due purely to the effects of heat is of consider- able importance in the total results experienced. We found, as did Dr. deLaroquette, that vaseline acted as a fairly complete preventive as regards both primary and secondary erythema, particularly the latter, while glycerine gave a slight protective action in our results, more than would seem to be warranted in view merely of its trans- parency to abiotic rays. From these observations and from the clinical facts, often showing erythema greatly disproportionate to the intensity of abiotic radiation likely to be present, it seems prob- able that ordinary sunburn is due to a mixture of thermic and abiotic effects of which the former are often the more prominent, although they generally cannot readily be separated from the secondary abio- tic effects, the development of which they tend to mask. Erythropsia. So-called erythropsia is the name of a phenomenon rather than of a pathological condition. The clinical records are numerous but vague. They all indicate a condition, generally very temporary, in which the patient finds a more or less ruddy tinge in everything seen. There is nothing definite in the tint of the coloration or the period through which it is observable. It apparently runs from vari- ous shades of orange and rose to a fairly full red. The most definite description given of the apparent color, which evidently pertains to a rather extreme incidence, is given by Fuchs ^^®, who compares it to a strong fuchsin solution with a trace of eosin solution. A cursory ^-iew of the clinical records indicates that the cases cited fall into three general divisions. First, cases associated with neurosis such as those gi\'en by Charcot and others (cited by Wyeller^^^). These clearly cannot be associated with any pathological condition of the visual EFFECTS OF RADIANT ENERGY ON THE EYE. 711 apparatus. Second, there are many recorded instances of traumatic erythropsia some of which at least evidently are associated with the actual infiltration of the eye media with blood. Third, one finds a vast majority of instances which one may term photo-erythropsia in which the observed appearances, one can hardly dignify them by the name of symptoms, are associated with over exposure to light. These are so entirely without pathological significance that we should hardly consider them here save for the fact that the phenomena have been by some writers like Widmark *^^, Fuchs ^^^, and others, subse- quently attributed to the effect of ultra violet radiations. As this erroneous conception of the fact still persists in spite of the admirable work of Vogt,^°^ it is desirable here to note the relation of this so-called erythropsia to the general phenomenon of color vision. The whole subject was thoroughly investigated recently by Wydler ^^^ who very plainly showed that erythropsia is due to the red phase of the negative after image following over-exposure to light, ordinarily brilliant white light, although green and blue green illumination is even more effective. The association of the phenomenon with ultra violet radiation appears to be due to the fact that photo-erythropsia has been often observed after the intense glare which produces snow blindness and not infre- quently in the aphakic eye after an operation for cataract. That the ultra violet really has nothing to do with the matter is clearly shown by Vogt ^°^ who found that erythropsia could not be produced experi- mentally by the ultra violet rays alone, but very easily by light rays containing no ultra violet. We need only add here that it is possible to produce marked erythropsia through euphos glass which transmits no ultra violet, through B-naphthol-disulphonic acid which also cuts off the ultra violet, and through dense flint glasses which eliminate all of the ultra \aolet which could with any certainty get through the lens. Also it is produced with great facility by radiation through green and blue green media which intercept the ultra violet com- pletely, but flood the eye with light of a color certain to produce a strong red phase in the after image. The truth seems to be that the so-called photoerythropsia is merely the result of such unequal fatigue of the primary color sensations as leaves for a greater or less time there- after a color sensation predominantly red. This conception clears up at once the difficulty of accounting for the partial erythropsia which has been noted by Purtscher ^^^ and others, since in cases of unequal exposure of various parts of the retina to brilliant light the fatigue effects necessarily must vary over the field of vision. A glance at figure (6) will render the situation clear. The curves in this figure 712 VERHOEFF AND BELL. are those of the three normal color sensations reduced to equal areas as determined by Exner. If any of these primary sensations are fatigued that of the remaining color or colors becomes the predomi- nant tint seen. This has been beautifully shown by Burch who by suitable means fatigued to complete exhaustion each one of the sen- sations and various combinations of them. Burch ^^ found that of the three the red first regained its sensibility, after perhaps ten min- utes, followed by the green and last of all by the blue where marked fatigue might persist for several hours. Red vision is therefore normally to l)e expected after fatigue of all three sensations since the red recovers first and in case the green and blue are more fatigued than the red the latter will be more notably predominant. As the maxi- mum luminosity of the spectrum lies in the green and as at high alti- tudes under a clear sky the blue is relatively strong, exposure in the Blu / \ / \ Gl ee I f / \ 1 ^ \ ^1 ii-d l< / / \ \ / / s \ / V / ^ \ \ ^ / A \ \ U- r ^1 —- k- __ , ^■<-'^ Wave length Figure 6. Primarv color sensations, after Exner. high snow fields necessarily fatigues these two sensations predomi- nantly, and photoerythropsia in greater or less degree may reasonably be expected. As a corollary we may note the reputed activity of the quartz mer- cury arc in producing erythropsia. On figure 2, Plate 5, marked in their proper positions are the three chief lines of the mercury spectrum at waves lengths 454 /x/x, 546 /xjUj and the pair at 578 ix]x. It will be seen that they lie in positions which indicate stronger fatigue of the green than of the red and marked fatigue in the blue. The green line at 546 \x\x is by far the strongest of the three followed by the yellow pair at 578 /x^u and by the strong line of blue. The chief red line in the spectrum is relatively ver\' weak hence fatigue weakens the green most, red in the next degree and blue relatively little. After fatigue, EFFECTS OF RADIANT ENERGY ON THE EYE. 713 therefore, the red comes easily and quickly into activity and unites with the residual blue to produce very marked erythropsia of a dis- tinctly rosy tinge. This lasts for some minutes, while traces of the disturbance of the blue and green vision may still be found after pro- longed exposure for a period of several hours, and this like other forms of color fatigue takes place whenever the light which affects the three primary sensations has been active, quite irrespective of whether the ultra violet is present or absent. As to the aphakic eye Wydler*^^ has noted the probable effect of coloration of the lens in this connection. When an eye has been shielded, often for years, against any strong access of the blue and violet and against the extreme end of the green sensation as well and is then after recovery from cataract operation exposed to strong day- light, it is merely a phenomenon to be expected if there is predominant red vision after fatigue. It would not be surprising, for that matter, if after long disuse of the color sensations toward the blue end of the spectrum fatigue were easier and recovery less prompt than in the normal eye. Aphakic patients in whom the pupil is often greatly enlarged as a result of iridectomy are likely to receive extraordinarily intense illumination on the retina, and hence may show the phenome- non of color fatigue to an exceptionally great extent. Exposed in the open the color fatigue is very marked, and w^hen this wears off after the exposure ceases, the return of the red sensation may very naturally be accompanied by some degree of erythropsia for this reason alone. One of us has recently examined a case in which erythropsia so pro- duced was a characterisitc condition. Vernal Catarrh. Spring catarrh is an uncommon disease of the conjunctiva that most often affects the upper lid, much less often the bulbar conjunc- ti\'a around the corneal limbus, and almost never the two together. It is extremely chronic, lasting from 3 to 20 years, and is associated with the formation of peculiar granulation tissue, infiltrated with eosinophilic leucocytes to an unusual degree, containing downgrowths of epithelium from the surface. In the case of the conjunctiva of the lid, the new tissue forms within the papillae, thus giving rise to large flat papillary growths. The symptoms of irritation, photophobia, lacrimation, and itching, are most marked in the spring and warm 714 VERHOEFF AND BELL. seasons of the year, usually disappearing during the winter months. For this reason sunlight has been suggested as the etiological factor in the disease. This hypothesis was first advanced by Schiele ^^^ and advocated by Kreibich ^^^ who showed that an occlusive bandage had a favorable effect upon the symptoms. This effect, however, may be explained in other ways than by the shutting out of light. Birch-Hirschfeld ^^ repeatedly exposed the conjunctiva of a rabbit within a period of 18 months to the "Uviol lampe" of Schott and obtained changes stated to be not unlike those of vernal catarrh. He does not, however, accept the view that the latter is due to ultra violet light. No doubt similar changes could be produced by other irritants frequently applied. The evidence for the view that vernal catarrh is due to the action of sunlight, therefore amounts to little more than the fact that the symptoms are most pronounced in the spring and summer. This fact, however, is accounted for even better on the more recent theory that the disease is due to pollen. Moreover the following objections, that to us seem insurmountable, may be urged against sunlight as a cause. In the first place if vernal catarrh is due to sunlight the lower lid, which is not only more exposed, but thinner and more trans- parent, should be more affected than the upper lid, whereas, as a matter of fact, it entirely escapes involvement. In this connection it may be noted that in cases of trachoma, a somewhat similar disease, the lower lid also usually escapes and here the disease is undoubtedly due to some infectious agent. Similarly this theory is inconsistent with the fact that the bulbar conjunctiva, which is directly exposed to tiie light, is but seldom affected, and almost never affected in associ- ation with the palpebral form of the disease. Finally, the possibility of abiotic action is ruled out by the fact that it is impossible for abiotic waves to pass through the entire thickness of the lid, if only on account of its rich vascularization. This objection does not apply to possible heat effects produced by visible or infrared rays, but in this case it would be necessary to assume exposure of the eyelid to direct sunlight for considerable periods of time as well as special sensitiveness of the conjunctiva to heat, neither of which conditions seems possible. EFFECTS OF RADIANT ENERGY ON THE EYE. 715 Senile Cataract. The theory has been advanced (see page 780) that senile cataract is due to exposure of the lens to daylight, particularly that from the sky. This is based solely on the fact that the cataractous changes usually begin in the lower part of the lens. It is undoubtedly true that the changes do first appear below, but as a rule they are so far below that they are in a portion of the lens completely shaded by the iris. Thus it is most often necessary to produce artificial mydriasis before incipient lens changes can be seen with the ophthalmoscope. Moreover, if the cataract were due to exposure to light, the pupillary area should be the first affected, since from such an extended source as the sky it receives the greatest concentration of light, and since the chief absorption must occur here. We must conclude, therefore, that there is no sound evidence for this theory of cataract formation. A possible explanation of the fact that the lower part of the lens is usually first affected in senile cataract is that the structure of the lens ma.y normally be slightly diiferent here than elsewhere. From a developmental standpoint this is indicated by the fact that coloboma of the lens usually occurs below. Burge ^^'' has recently attempted to supply an experimental basis for the view that ultra violet light is responsible for cataract. He found that the rays from an unscreened quartz merciny vapor lamp had almost no coagulating efl^ect upon the lens protein even after an exposure of 72 hours at a distance of 5 cm. but that when acting in the presence of weak solutions of calcium chloride, sodium silicate, or dextrose, coagulation occurred. Since in senile cataracts calcium, magnesium, and sometimes silicates, are greatly increased, and in diabetic cataracts dextrose is presum- ably present, Burge assumes that these cataracts are due to the action of ultra violet light. That is, he assumes that these substance are present in undue quantities in the lenses of certain individuals and that this renders their lenses vulnerable to the short waves of daylight. This assumption is suflSciently controverted by the fact just men- tioned that senile cataract usually begins at the periphery below. But in addition, other serious objections to his argument may be pointed out. In the first place, in traumatic cataracts and cataracts due to inflammatory conditions, calcium salts, and no doubt mag- nesium and other salts, are deposited in great abundance, and the lens may even become completely calcified. In fact, the same thing occurs in dead tissues anywhere in the bodv, so that the reasonable 716 VERHOEFF AND BELL. assumption is that the presence of these salts in senile cataract is a result not a cause. Then, too, Burge made use of intensities of expo- sure and wave lengths to which the lens is never subjected during life. The cornea completely screens it from practically all the short waves found effective by him. The longest waves with which he could coagulate proteins were 302 [xfj. in length and the effect produced by these was insignificant. Burge suggested also that his results might apply to glassblowers cataract, overlooking the fact that the latter typically begins at the posterior pole of the lens, whereas in his experiments the part of the lens away from the light was little if at all affected. It is of course obvious that the slight loss of transparency he sometimes observed in this part of the lens could not have been due to the direct action of the light, since the effective rays could not have penetrated so far. The fatal objection to Burge's theory as applied to senile cataract is that the ultra violet solar rays cannot reach that portion of the lens where cataract generally begins, and that portion of the lens where ultraviolet light has the best chance of action is affected only at a late stage of development. Concentration of Energy in Images. We have already shown that superficial action of radiant energy on the eye depends on the actual energy in ergs per square cm., or other convenient measure, which falls upon the surface. Such value is directly as the energy of the source and inversely as the square of its distance. The density of incidence of energy at points within the eye is obviously dependent on the amount to which the superficial energy is concentrated by the refracting media, and at the retina the concentration of energy is determined by the size of the image and the aperture of the refracting system, which is determined by the area of the pupil. In dealing with an extended source the image is corre- spondingly extended and the surface density of energy in the image is correspondingly reduced. Hence it is that with sources like the tube of the quartz arc the image density is relatively small, while with point sources or those of very small area, like the electric arc, the retinal area is correspondingly small and of the total energy reach- ing the pupil there is a greater concentration in the image. Within limits the intensity of the effect on the retina is then directly pro- portional to the intrinsic brilliancy, or radiation per unit area of the EFFECTS OF RADIANT ENERGY OX THE EYE. 717 source. The mere lowering of the intrinsic radiation by spreading out the source therefore greatly lessens any possible effects of energy which may reach the retina, while it does not in any way affect the radiation which may reach the cornea and conjunctiva. There is, however, a very notable limitation to this principle which comes into play in considering small and intense radiants, as was long ago shown by Charpentier ^^. ^Yhen the image of a luminous object reaches the diameter of approximately 0.15 mm., variations of intrin- sic radiation at the source cease to be significant and the appar- ent intensity of the source varies simply with the inverse square of its distance. This corresponds to the visual angle of about 40 minutes of arc. For areas of greater dimensions one must reckon with the size of the image as determined by the ordinary laws of geometric optics, but for radiants of less than this dimension the image may be taken as of constant area corresponding to the circle of diffusion, and the energy concentrated in it varies as the inverse square of the distance of the radiating source. In any case the energy reaching the retina is diminished by the absorption in the media of the eye of which we will now take account. General Nature of Absorption. By absorption one means in general terms the stoppage of energy in any medium. This may be either specific, affecting only energy of certain ware length, or general, affecting more or less all energy what- ever. In the former case it is due to the molecular or atomic struc- ture of the material, in the latter to the fact that it is not physically homogeneous. In specific absorption such as takes place, for instance, in colored glasses, the molecular structure is such as to respond to and take up certain particular oscillation frequencies so that waves of these frequencies do not readily pass through the sub- stance. In general absorption the substance contains particles which reflect the energy from their surface or absorb it without definite regard to its wave length. Such absorption occurs, for instance, in some glass which is full of microscopic bubbles which reflect the energy at their surfaces, or in certain dark glasses which are filled with minute opaque particles. Both kinds of absorption may coexist in the same material, but. general absorption involves it only in a very indirect way due to the general properties of reflecting surfaces. The stoppage of radiant energ\- in the media of the eye is of two 718 VERHOEFF AND BELL. kinds. First, by absorption in the ordinary sense, and second, by reflection due to the structure of the eye. For instance, the cornea is somewhat lamellar in structure, built up of successive layers and cells and is therefore, owing to the differences in the index of refraction as the ray traverses the structure, somewhat less transparent than if it were homogeneous. There is also a slight loss of energy at the surface in passing into the aqueous. This is also practically homogeneous, but there is again a slight loss by reflection in passage from the aque- ous to the lens which has a materially higher index of refraction, and is of itself non homogeneous from the standpoint of refraction. Finally there is reflection passing from the lens into the vitreous and the vitreous itself is not without structure, so that in fact the path of rays through it can be traced by the faint diffused illumination due to its lack of homogeneity. It is quite impossible to determine accu- rately these losses, except for the initial loss which occurs by reflection at the surface of the cornea to which we have already drawn attention. These losses are greater for rays of short wave length than for those of long, and perhaps the most that can be said about them numerically is that a total loss is probably of the order of magnitude of 10% for rays of medium wave length. The general absorption of the media of the eye has been studied by Aschkinass ^ in connection with his determination of the absorp- tion spectrum of fluid water. He found that the transmission of the media of the eye for radiant energy in general was closely similar to that of water in a layer of equal thickness. The large proportion of water in these media would, of course, suggest a similarity and Aschkinass found the characteristic absorption bands of water in the experiments on the eyes of cattle and some control experiments on the human eye. The only notable discrepancy was in finding a con- siderably higher absorption in the cornea than would be warranted by its water equivalent which Aschkinass describes chiefly to a film forming very rapidly over the surface of the dead cornea. In examin- ing the bearing of these facts on the energy focussed upon the retina in any given case it should be noted that the absorption is chiefly in the infra red. Figure 7 shows the absorption curve of a 5'^'^ layer of water as found by Aschkinass ^. Hence in examining the absorption in any given source of energy it will be found relatively greatest for infra red radiation, except for the effect of the lens in cutting off a large part of the ultra violet. Luckiesh ^^'^ has made a study of the absorption of energy from various sources by the eye, based on Asch- kinass's results. From this it appears that from low temperature EFFECTS OF RADIANT ENERGY ON THE EYE. 719 sources like carbon incandescent lamps and ordinary flames the absorption of the total energy rises to nearly 90%. As we have ah'eady shown this must be increased by miscellaneous losses l)y reflection so that the amount of energy actually available in the image on the retina from such sources is very small. It is quite otherwise with radiants like the sun, which is roughly equivalent to a body of 5,500 to 6,500 degrees absolute as regards the char- acter of its radiation. Froni such a source the specific absorp- tion of water cuts off relatively little, and the total loss of ener'ijy ;^ 80 GO 40 ■JO k c 0 \ \ 0) (0 C \ \ / \ V / V .8 .9 1.0 1.1 1.2 1.3 1.1 1.5p, Wave length Figure 7. Approximate absorption of 5cm water, from data of Aschkinass. in the eye is of the order of magnitude of 25 to 30%. In phenomena like eclipse blindness therefore not only is the eye exposed to a very powerful radiating source, but the radiation is of such char- acter that it is not strongly absorbed and hence the energy in the image may rise to very great intensity. The solar radiation curves already shown make it plain that the proportion of energy cut off will be greater the greater the altitude of the sun and the less the general atmospheric absorption. Taking 30% as the total cut off in the eye one may obtain an approximate idea of the energy concen- trated on the retina in observing the sun unscreened, the total radia- tion being about 10^ ergs per square cm. per second; and assuming the pupillary diameter to be about 2 mm., approximately 3% of this energy will enter the eye, and subtracting 30% for absorption and reflection it results that the total energy concentrated in the image would be about 20,000 ergs per second. Taking the area aft'ected as approximately .15 mm. in diameter the concentration of energy in the image is on the basis of nearly 113 X 10^ ergs per square cm. per second. Even if only a quarter or a half of this amount is avail- able in the case of the partially eclipsed sun, it is evident that the immense concentration of energy in the image is sufficient to produce 720 VERHOEFF AND BELL. destructive effects such as have been often clinically noted and which we have observed in our experiments. From the relatively small absorption by the eye media in the case of solar radiation it is clear, however, that it is far more dangerous, in proportion to its intensity, than any artificial source of radiation. Passing from the general absorption of the eye for radiant energy here considered to the specific absorption of the several media, the facts have been pretty thoroughly established by the researches of Hallauer^^^, Schans and Stockhausen^^^ and Martin ^^^. As regards the general volume of radiant energy received by the eye there is no specific absorption except that already noted due to the aqueous content. Aside from this the numerical proportion of the energy from most soiu'ces specifically absorbed in the eye is very small and is con- fined to the ultra violet region. The human cornea cuts off practi- cally all the energy of w^ave length less than 295 mi. The lens wipes out the remaining ultra violet up to a point between 380 /x/x and 400 ^c/i. The vitreous absorbs strongly in the general region between 250 /jl/j and 300 h/jl in the thickness in which it exists in the human eye. Only a very minute proportion of energy within this range gets through so that the general effect of the absorption in the vitreous in the case of an aphakic eye is to re-enforce that of the cornea, as is well shown by the immunity from abiotic action of the retina in our experiment No. 89. In his experiment on the eye of a young rabbit, Martin found the limits of transmission to be about those here noted, except that the lens transmitted freely radiations longer than 350 (x/jl. As the human lens yellows with age its absorption reaches down into the violet, extending even to 420 fifx. Eclipse Blindness and Allied Phenomena. Every recent eclipse of the sun has given rise to numerous cases of so-called eclipse blindness, due to careless observation of the phe- nomenon in its partial stages, either with the naked eye or with altogether insufficient protection. We should not here consider the matter worthy of attention were it not for the fact that it has been loosely ascribed, like many other imperfectly investigated ocular injuries, to the malign effects of ultra violet light. Eclipse blindness appears in literature as far back as Plato's Phaedo, and is repeatedly mentioned through classical and post classical times as an appar- ently not vmexpected phenomenon. The eclipse of April 17, 1912, in EFFECTS OF RADIANT ENERGY ON THE EYE. 721 Germany produced a total of many hundred cases of more or less in- jury to the eyes, as noted by Wendenberg ^^°. Every eye clinic re- ceived its toll of more or less severe cases. Clinically the immediate eflFect is marked and immediate scotoma, which does not pass away promptly but leaves more or less seriovis cloudiness of vision and accompanying loss of acuity which may be temporary, lasting a few weeks, or in severe cases permanent. The scotoma is commonly central and generally of small extent, in a marked proportion of the cases corresponding fairly well with the dimensions of the sun's image, although wide variations from this may be due to repeated fixations overlapping or reenforcing each other. As it is generally impossible to tell just how long or how often the patient fixed the phenomenon nothing definite can be postulated concerning various varieties of scotoma which have been noted by various observers. The ophthal- moscopic observations usually show changes ranging from scarcely perceptible, to conspicuous and permanent pathological appearances involving lasting and destructive injury to the retina. Metamorphop- sia sometimes appears, the significance of which will be apparent in connection with some of our experiments, and diminution of visual acuity is fairly well marked, often falling below one third. With the progress of time the scotoma tends to contract and in mild cases normal vision is regained within some weeks, or in the most severe cases great reduction in acuity persists permanently.* Our experiments have been directed to the production of an artificial eclipse blindness in animals, and the examination of the lesions pro- duced, following up the work of Czerny ^^, Deutschman ^^, Herzog ^^^, and others with special reference to the intensity required to produce the lesions noted. The character of the lesions produced in these experiments is described elsewhere (page 697). The apparatus em- ployed was powerful enough to produce prompt and acute effects. For most of the experiment we employed the mirror apparatus shown in Plate 8 which consisted of a silvered glass mirror 26 cm. in diameter and 1.5 meters focal length, carried, as shown, in a fork mounting set up approximately in the meridian and fitted with slow motions in right ascension and declination so that the beam could be readily * Jess 200 describes relative ring scotoma for colors in a series of cases, but does not offer a convincing explanation for its occurrence. Boehm4:8 was un- able to demonstrate it in any of his cases, although he examined them with special reference to it. Birch-Hirschfeld^l suggests that a normal eye would show the same condition if examined in the same way. This criticism would seem to apply with equal force to the similar scotomata reported by Birch- Hirschfeld35 himself as occurring after exposure to ultra violet light. 99 VERHOEFF AND BELL. directed and kept in position. In use the mirror was slightly tilted so as to throw the focus just out of the path of the direct incident beam. The concentration of energy obtained by this instrument was enormously great, owing to the size of the mirror and its relatively short focal length. Its area was about 530 square cm. so that with an average reflective coefficient of 0.75, with good sunlight the re- flected energy would amount to some 4 X 10^ ergs per second. The image of the sun formed by this mirror is 13.2 mm. in diameter or about 1.36 square cm. in area. The energy at the focus then amounts to approximately 30 watts per square cm. A pupil expanded to say 10 mm. b\' the use of mydriatics would therefore take in a pencil equivalent to about 24 X 10^ ergs per second. Allowing as in other cases one-third for the energy absorbed by the media of the eye as a whole, the energy incident in the image would be approximately 16 X 10^ ergs per second. The diameter of the image in this case is just over 2.5 mm., corresponding quite exactly to an area of 5 square mm. The energy density in the retinal image therefore would be about 32 X 10^ ergs per second per sciuare cm. and it was found in our experiments that an exposure of \ second to this intensity was sufficient to produce a destructive thermic effect in the retina. This short period is very striking in comparison with the relatively long exposures necessary to produce typical eclipse blindness with the naked eye, although it agrees very well with the data which we later cite regarding energy burns from other sources. The secret of the relative resisting power of the naked eye is that usually in observations of the sun, the pupil is in extreme miosis, so that the amount of energy received is probably not more than 6% of that computed for the normal pupil, while the extremely small area of the solar image favors rapid dissipation of the energy not found when a considerable area is attacked, as in the case of the mirror experiments. The latter condition we have often noted in thermal experiments of other sorts with the big mirror and lenses of various kinds. A concentration of energy very much greater than that from the mirror, acts much more sluggishly on inflammable material when the focus is merely a minute point instead of an ap- preciable area. Another factor that tends to protect the human eye from the thermic action of light sources of small size, is the impos- sibility of perfect fixation for any length of time. This is well shown by some experiments on our own eyes (see page 732). The screens employed in the work are noted in connection with the various experiments. Inasmuch as silver reflects very badly in the region near the extreme ultra violet end of the solar spectrum and it EFFECTS OF RADIANT ENERGY ON THE EYE. 723 appeared not desirable to eliminate the possibility of specific injury due to such rays we employed in part of the experimental work as a substitute of the mirror in the concentration of the solar energy a quartz lens 12 cm. in diameter and 25 cm. focal length mounted on an adjustable stand so that we could work directly in its focus. A priori the climical evidence is strongly against any definite pathologi- cal effects due to the ultra violet radiation as such. Numerous cases are recorded in which typical eclipse blindness has been produced through ordinary spectacle lenses, through glass insufficiently dark, and through opera glasses and the like, in all of which cases the abiotic rays are, as we have already shown, cut off. Even very small thick- nesses of colored or even clear glasses are sufficient completely to ab- sorb these rays, which moreover are always cut off by the lens so that they cannot reach the retina where the lesions are found. This is in accordance with the conclusions reached by Parsons ^^^ in analyzing the evidence at hand. Attempts have been made by several investi- gators, notably Birch-Hirschfeld, to eliminate the infra red rays also by the use of thick water cells and other absorbing media, but these attempts so far as experiments with solar light are concerned are fu- tile, because, as a glance at Figure 5 (page 705) will show, the greater portion of the solar energy lies entirely within the visible spectrum with its intense maximum in the blue or green according to the effect of the atmospheric absorption, so that for solar radiation it is the light rays which are thermally effective, the energy radiation in the ultra violet and infra red being relatively insignificant. Our experiments show with the utmost distinctness that the effects known as eclipse blind- ness are wholly thermic, due to the intense concentration of the solar energy upon the retina by the refracting system of the eye itself forming an image of destructive energy intensity, the amount of which we have already computed in considering the energy concentrated in images. It is only the briefness of the casual fixations of the sun, and the great reduction of the size of the pupil in response to the intense illumination, that prevents the very common occurrence of such injuries. In the observation of an eclipse the patient is tempted to dangerously long fixation and the necessary results follow. With long fixation the typical retinal lesions of eclipse blindness may be produced by sources of moderate intensity. For instance in our experiment No. 53 an exposure of twelve minutes was made on the eye of an albino rabbit with the single quartz lens system and the magnetite arc as source, through the 5 cm. quartz water cell. Now from experiments previously made on the radiation from the magnetite 724 VERHOEFF AND BELL. arc by one of us the energy entering the pupil was of the order of magnitude of 444,000 ergs per second. Examination of the retina showed that the lesion produced was practically 3 mm. in diameter. Hence the concentration of the energy in the image allowing the same absorption as in the previous computation amounted to nearly 42 X 10^ ergs per second per square cm. This is roughly ^ of the concen- tration in a direct solar image and correspondingly the lesion produced was comparable with that of a typical case of eclipse blindness. It is quite impossible to get an accurate idea of the critical length of fixa- tion which appears in cases of eclipse blindness, since the observations producing it are generally discontinuous and not noted. This ex- periment, however, indicates, making due allowance for the extent of the experimental image and for the extremely small size of the pupil in looking at the sun with the naked eye, that the critical period for the development of eclipse blindness is, with close fixation, of the order of magnitude of a minute or less. An exposure of even a few seconds would be highly dangerous were it not for the extreme miosis set up and the usual wandering of the image upon the retina. Rapid shifting of the focal image on the retina gives the tissue an opportunity for cooling, so that if the fixation at a single point is not long enough to produce destructive effects little permanent damage can be done, although the scotomata may be severe. Our experi- ments Nos. 100 and 101, in which the exposure to the solar heat through a blue uviol screen was intermittent, show this excellently. In the first the exposure was for alternate seconds over a period aggregating ten minutes, or more than six times as long as necessary to produce burning of the retina in a continuous exposure. No damage was done. The second experiment, in which no water cell was used, consisted of 220 exposures of ^ second with one to three seconds interval between. In this case there was again no damage done although the exposure was three and two-thirds times as long as was required to produce destructive lesions of the retina in two differ- ent cases with a continuous exposure. In this connection one may note that the experiment of Best^^ in fixing the sun for ten seconds through a screen of blue uviol glass was a somewhat hazardous one since this glass lets through a very mate- rial proportion of the energy from a high temperature source like the sun. Best's purpose in making this experiment was to show that ultra violet light is not injurious to the retina of a normal eye. The exposure however, was too brief for the result to be of importance in this regard. It is not ultra violet energy which is to be feared in a EFFECTS OF RADIANT ENERGY ON THE EYE. 725 case like this as we have already shown, but the danger was from pure heat radiation of which 10% to 15% was due to pass through the uviol glass. We produced in experiment No. 97 most destructive effects from solar heat passing through this medium, and while Best's experiment of ten seconds was below the danger limit it should never be forgotten that the solar energy lies well toward the blue end of the spectrum, and media which successfully cut out the red and infra red are of very little service in protection against solar radiation. Thermic Effects on the Retina from short Circuits. It is worth noting in connection with eclipse blindness that sources of intense energy other than the sun may produce similar results. For example, Uhthoff ^^^ reports the case of a patient exposed to a violent short circuit in which a fortnight after the accident grayish spots due to alterations in the pigment epithelium were observable in the macula of the left eye and were still observable six months later. This ophthalmoscopic appearance is closely similar to that many times recorded in eclipse blindness. Still later Knapp ^°^ records a case of bilateral injury produced by a tremendously powerful short circuit occurring a scant half meter from the patient's face. There was complete temporary scotoma followed on the next day by some superficial symptoms indicating photophthalmia, and a week later by metamorphopsia, while the vision had been steadily sub-normal. In each fundus was a patch corresponding to the image of the short circuit flare, in which serious damage had been done. These injured areas were still obvious a year later. The retinal lesions described, and especially the metamorphopsia, are such as are typical in the case of eclipse blindness. Here the energy radiation of the short circuit was concentrated in the image to an extent sufficient to produce a typical thermic lesion. The slightness of the abiotic radiation re- ceived is evidenced by the very brief superficial symptoms, and the retinal injury, owing to the absorption of practically the whole ultra violet by the media of the eye, must have been due essentially to the pure energy radiation of which the amount, judging by the de- scription, was probably not less than 100 to 200 kw. A short circuit involving 100 kw. would give a superficial intensity at a half meter of over 30,000,000 ergs per square cm., that is, more than thirty times the intensity of solar radiation. The area of the scotoma produced 726 VERHOEFF AND BELL. was, from the description, in the neighborhood of 1 sq. mm. Assum- ing a pupillary diameter of 5 mm. likely to be found in working in a moderate degree of light when surprised by the short circuit, the energy entering the pupil would be at least 6 X 10® ergs per second concentrated in the image, that is an energy density amounting to in the neighborhood of 6 X 10^ ergs per second per square centimeter reckoned without regard for absorption. Allowing one third of the energy absorbed in the eye the energy density in the image should be 4 X 10^ ergs per second per square cm. two or three times, at least, greater than the corresponding energy density for a direct ol)servation of the sun, very possibly, owing to the intensity of the short circuit, even several times greater than this. It is little wonder then that although the exposure time is stated to be less than 1 second the results were serious. In true eclipse blindness the length of fixation is the chief factor in the damage. Thermic Effects on the Retina from Lightning Flashes. A consideration of these miscellaneous energy effects on the eye would be incomplete without referring to the injuries to the eye received from lightning. In such cases a sharp distinction must be drawn between cases in which the patient is actually struck by light- ning, with more or less serious effects, and those in which the patient is clearly not struck, but subject to direct radiation from a nearby flash of lightning. In the former class of injuries electrolytic action and exceedingly severe nervous shock generally occur and the final results may include various grave ocular symptoms sometimes ending in complete blindness due to cataract or atrophy of the optic nerve. In the second r class of cases the effects are usually limited to severe scotomata which may impair vision for some hours or days but as a rule there are no lesions visible either superficially or with the ophthal- moscope, and no permanent damage is done. This immunity is chiefly due to the usually considerable distance between the actual lightning bolt and the observer, since the amount of energy actually involved in a lightning discharge of the first order of magnitude may be enormously great. Sir Oliver Lodge estimates it as high as 10'-° ergs. There are a few instances, however, in which the energy re- ceived at the eye has been great enough to produce typical lesions both from abiotic action and probably also from purely thermal EFFECTS OF RADIANT ENERGY ON THE EYE. 727 effects. Silex337, Vossius^°^ Rivers ^76, Dunbar Roy302 and Le Roux et Renaud ^^^ have all noted superficial injuries of the cornea and the last named, as also Oliver 2®° have noted symptoms of chorio retinitis, which seem to indicate lesions due to thermic effects, in the latter case invoking the metamorphopsia frequently associated with eclipse blindness. The case of Le Roux et Renaud was a specially notable one in which the patient, on guard duty at night, was exposed to a very powerful flash. It was immediately followed by violent erythropsia which lasted for some two hours. The Gendarme re- mained at his post and went home and to bed about three hours later. The next morning he woke with acute headache, with substantial blindness in the left eye followed a few hours later by loss of sight in the right eye. There was double acute conjunctivitis with swelling and reddening of the lids and conjunctiva and marked chemosis. A little later there was diffuse interstitial keratitis, a change in the color of the iris from blue to greenish and a grayish haze on the lens front visible in a bright light. These affections of the anterior eye cleared later and when the ophthalmoscope could be used there was marked haze in the vitreous, which cleared very slowly and not completely even after three years. This was believed by Le Roux et Renaud to be associated with chorio retinitis and was certainly secondary to the original lesions. It is very difficult to make anything like an exact computation of the energy which produced these results, since not only is the total amount of energy in a lightning flash extremely variable and known only to a rough approximation, but the duration of the flash is also variable and uncertain. Thus much is clear, however, that the very heavy discharges, in which the length of flash is some hundreds or thousands of meters and the quantity of electricity discharged very large, are also the relatively slow flashes, since the equivalent con- denser capacity is very large. The estimates of frequency rising to millions per second can have no place here, since obviously the velo- city of free waves being only 300,000 km. per second, a flash of one or several km. in length cannot have a very high oscillation frequency even supposing it permits oscillations at all. Attempts to measure the frequency of the discharge have often led to results of less than .001 of a second and it is altogether probable that in these long flashes there is no oscillation at all on account of the resistance effects. Start- ing from the estimate of Sir Oliver Lodge of 10^" ergs per second and assuming an effective time of discharge .0001 of a second, and that of the total flash not over .1 is within the effective range of reaching the 728 VERHOEFF AND BELL. eye, the energy in the discharge may be reckoned as one or perhaps several thousand times that of the short circuit discussed in connec- tion with Knapp's case. With a nearby discharge occurring say within 10 or 20 meters, the quantity of energy received by the eye would be amply great to account for even severer results than those noted by Knapp (loc. cit.). Working back from our experiments with the bare magnetite lamp carrying about 500 watts in the arc, it again appears that the quantity of energy assumed by Sir Oliver Lodge is more than sufficient to ac- count for the results of a nearby discharge. It therefore must be admitted that the direct action of lightning in producing both abiotic and thermic lesions in the eye as in the cases of Le Roux et Renaud and of Oliver is well within the bounds of possibility although requiring very unusual proximity to a powerful discharge which may well have been the case in these instances. The extreme rarity of such clinical conditions is perhaps ascribable to the few instances in which there is close proximity to the stroke combined with free exposure of the eyes without the patient being actually involved in the shock. One must consider therefore that lesions produced directly by the radiating energy of a lightning discharge are extremely unusual and unlikely to occur, although well within the range of possibility, as the cases here referred to show. Possible Specific Action of Infra Red Radiation. As we have already stated, there was no segregation of various radiations in our experiments on thermic effects except in so far as abiotic radiations were cut off by certain screens. So far as all indi- cations go the effect of all other radiation than the abiotic is chiefly chargeable to thermic energy without respect to wave length. As already explained different sources present totally different distribu- tions of energy with respect to wave length, the lower the temperature the greater being the proportion of the so-called infra-red rays. In the case of the quartz mercury arc and the magnetite arc with which we chiefly worked, the spectra are essentially discontinuous and hence do not obey Planck's law, so that there is no definite relation between the temperature and the wave length of maximum radiation. The total energy spectrum of each of these sources, however, is exceedingly complex. Of the total energy spent in the arc a certain proportion goes to maintain the characteristic linear spectra, a cer- EFFECTS OF RADIANT ENERGY ON THE EYE. 729 tain other portion goes to heating the electrode or containing tube and the surrounding mechanism. While therefore the spectrum which can be seen or photographed is linear, there is superimposed upon it the continuous spectrum of the radiating solid at rather low tempera- ture and of relatively large extent, since the radiation is not only from the arc or its containing tube but from the immediately heated sur- roundings. Much of the loss of efficiency in both sources referred to comes from this secondary heat radiation. This is particularly the case in the quartz mercury arc of which the actual light-giving effi- ciency is very high, much higher than is indicated even by its really small specific consumption per c. p. The existence of this secondary radiation which is mainly of very long wave length, make comparisons between such sources and the ordinary radiating solids very difficult. The following table gives for the magnetite arc the transmission with respect to the total energy of the most important of the various media which we employed, as determined by a Rubens thermopile. Absorption of Certain Screens. Source Magnetite Arc. Filter 2 quartz plates each 3 mm. thick Same + 5 cm. distilled water Water cell and Dense Flint Nd " " " Medium Flint No " " Light Flint N^ " " Crown Nb Dense Flint N^ 1.69 alone Medium Flint N^ 1.63 " Light Flint No 1.57 " Crown Nb 1.51 " It will be observed that the actual transmission of the empty quartz cell consisting of two 3 mm. polished plates was only 53% of the total energy. Of the 47% lost, roughly 15% should be in the reflections from the four surfaces of the two 3 mm. polished plates. The re- mainder, that is more than a third of the total energy, is mainly Percentage of Transmission 53 33 1.69 (335 mm) 26 1.62 (315 mm) 28 1.57 (305 mm) 27 1.51 (295 u^fji) 28 40 45 40 43 730 VERHOEFF AND BELL. from the cut-off of secondary radiation of relatively very long wave length, 4 to 5 n and more, received from radiating surfaces at and below red heat. The addition of 5 cm. of distilled water in filling the cell reduced the transmitted radiation to 33%, which represents the radia- tion between the former limit and approximately 1 ^u. There is every reason to believe that most of this energy is from the hot body radia- tion rather than from the line spectrum of the arc itself, for in so far as known metallic spectra are not rich in intense infra-red lines. Screen No. 1 then, cuts off 22% of the energy between 1 /x and the extreme ultra violet. Screen No. 2, medium optical flint, is relatively transparent, almost as transparent as the light crown screen No. 7, while the light flint screen No. 4 cuts oft' more energy than either of these. Used without the water cell all the four screens mentioned cut off between 50 and 60 % of the total energy due to the large absorption of glass in the extreme infra red corresponding to the secondary hot-body radiation of the source. This secondary radiation being from a very diffused source cannot have a conspicuous effect in those experiments in which the light was concentrated through lenses although it comes into play in the free radiation from the arc. It is quite certain, for instance, that in the bactericidal experiments with the mercury arc which one of us has recorded, in which trouble was experienced from heating of the water in which the bacteria were suspended, this heating was mainly from the large secondary radiation which is readily stopped by water rather than from the characteristic radiation of the mercury vapor. The data heretofore given by one of us on the proportion of ultra violet energy in the quartz mercury lamp and the magnetite arc may be regarded as substantially correct for the metallic spectra as such, the quartz water cell of 1 cm. thickness employed in these experiments cutting off the secondary radiations rather completely without inter- fering materially with the energy of the line spectrum. In our experiments involving total thermic effects necessary corrections for the conditions of the experiment can be made by reference to the foregoing table. As regards the thermic action of radiation on the eye, there is no reason to suspect any specific effects with respect to wave length. So far as the action is not definitely abiotic or concerned with the stimulation of the light perceiving functions of the retina it seems to be purely a question of energy as in any other case of heating. The more violent phenomena produced by heating are considered in our discussion of eclipse blindness and allied phenomena (page 720). EFFECTS OF RADIANT ENERGY ON THE EYE. 731 The possible effects on the lens of heat radiation persisting over a very long period we have considered in discussing glass blowers cataract (page 734). Thermic effects on the cornea we have shown to be obtainable only under extreme experimental conditions (page 692). The media of the eyes generally, as we have already shown, are of substantially the same absorbing character as water so far as concerns visible and infra red rays and therefore take up chiefly radiant energy of long wave length such as is prominent in radiation received from low temperature sources like hot metals or molten glass. The pig- mented iris and the pigment epithelium of the fundus are exceptions in that they absorb quite completely radiant energy irrespective of wave length. There is, therefore, a tendency to produce localized thermic effects in both these structures, as we have shown in various of our experiments. The lesions occurring under these circumstances have already been described. In case of the iris, the heat effects may proceed all the way from moderate irritation to serious permanent injury (see page 696). This phase of the matter was recently inves- tigated by Reichen^^^. This observer studied the effects of con- centration of infra red radiation on the eye by cutting off the visible and ultra violet spectrum by a filter of iodine in bisulphide of carbon, and absorbing the extreme infra red by a water filter. In this way the rays with which he was concerned were substantially those be- tween 800 and 1200 /jl/jl. Using as source the electric arc between carbon electrodes, by means of a rock salt lens he concentrated the filtered energy upon the eyes of rabbits for periods from 1 to 33 min- utes. The only effect noted was contraction of the pupil lasting some hours and generally slight, evidently depending on the direct heat effect tl>rough the filters used, since the visible rays were practically excluded from the retina. Such pupillary contraction is reasonably to be expected after moderate irritation of the iris, such as would be furnished by this heat stimulus and may occur as some of our experi- ments show, in the absence of any recognizable histological changes. Violent and persistent effects were hardly to be expected inasmuch as the source used is not rich in rays transmissible through Reichen's filters. Most of the energy from a carbon arc is intercepted by a water filter and a good deal of the remaining energy by the iodine in carbon bisulphides. This selective action is well shown in one of Reichen's experiments in which the water screen, which was quite close to the arc, was boiling violently after 7^ minutes exposure. Reichen's experi- ments, therefore, merely show a mild irritation not in the least peculiar to the region of the spectrum employed. It would seem quite im- 732 VERHOEFF AND BELL. possible to obtain even this eflfect except through strong focussing of the hght upon the eye, a condition which is not found in the use of ordinary illuminants for any purpose. So far then as infra red radiation is concerned the eye is not subject to any special dangers, and the concentration of heat upon it in any way would inevitably set up a danger signal of painful sensation long before any definite heat effects could be obtained. In fact it is easily demonstrable that the full radiation that can under practical conditions be received from the most powerful illumi- nants is incapable of producing on the retina any lesions due to thermic action such as may be found in eclipse blindness. In our Experiment 53, already referred to, changes in the pigment epithelium and a definite burnt area were produced. The diameter of the area in which histological changes were clear was 3 mm. shading off toward the edges and having a central area about 1 mm. in diameter, in which the damage was serious and comparable with that in eclipse blindness. We have found that in this case the concentration of the energy in the image amounted to very nearly 4.2 X 10® ergs per second per square cm., roughly 2g of that found in the solar image formed directly through a 3 mm. pupil. The exposure to this intensity lasted 12 minutes. Here then is a definite case in which the result was positive. Negative results, however, of which a few were obtained in our experi- ments, are unsatisfactory since they do not take account of the wan- dering of the image, or of imperfect fixation which as we have shown would be likely to avert injury as in the typical case of intermitted exposure to heat. Our large magnetite arc gives at a distance of 2 meters a total radiation of about 2500 ergs per second per square cm. At this dis- tance from the arc we find the pupil of the human eye is narrowed to a scant 2 mm. and this area would intercept about 75 ergs per second of the total amount stated, allowing 5 absorption as heretofore. This means that the energy concentrated on the retina would be about 50 ergs per second. To compute the energy density requires a knowl- edge of the size of the image including the element of imperfect fixation. We have found from personal experiments that on fixing at 2 meters the magnetite arc for a few seconds and measuring the size of the scotoma produced by viewing at the same distance a card ruled to centimeters, that even for this short fixation period the area of the scotoma is nearly four times that of the geometrical image of the source. On fixation of 6 minutes the scotoma rises to twenty-five times the size of the geometrical image. This latter has an area of EFFECTS OF RADIANT ENERGY ON THE EYE. 733 about 0.01 square mm. as determined from photographic data, so that for a long fixation the energy received would be distributed over practically 5 square mm. of area. The density in the image as actu- ally obtained in a long fixation would amount to about 20,000 ergs per second per square cm., only about 200 P^^* of the energy density which produced the positive result in experiment No. 53. This difference is still further enhanced practically by considerable differ- ence in the size of the image areas, the smaller one being relatively less effective than the large on account of the more rapid dissipation of heat. It is therefore evident that an exposure of even 12 minutes with close fixation to a source as powerful as this arc, a thing which no rational human being would be likely to undertake, still involves only a small fraction of the minimum energy known to produce definite lesions. Further evidence of the harmlessness of such exposures even to very pow^erful artificial illuminants may be derived from noting that the heat concentration in this case, that is 20,000 ergs per second per square cm. is less than 3^, of that received by direct fixation of the sun with the same pupillary aperture (see page 719). Now it is well known that one can fix the sun for a very few seconds without danger of anything more than a temporary scotoma. From this we may conclude that the arc here considered may be fixed, as well as it is possible to secure fixation, for a considerable period without running danger of a permanent scotoma. In order to demonstrate beyond any possibility of doubt that the retina cannot be injured by exposure of the eye to a source such as the magnetite arc, one of us has actually fixed his eye upon this arc at a distance of two meters for 6 minutes on two occasions. The macula itself was not fixed upon the arc but upon an object of the same size placed 2.6°, to one side of it. In this way more perfect fixation of the image of the arc on the retina was ensured than if the arc had been looked at directly. The results were only of a tempo- rary character. There was considerable xanthopsia lasting about 3 minutes. The scotoma had disappeared after each observation in less than ten minutes and could only be detected for a few hours by careful dark adaptation of the eye. The following day no traces of it were determined. It is therefore clear from actual experiment as well as from general principles that even the extremely powerful arc which we have here used is from the standpoint of injury to the retina entirely harmless under any circumstances conceivable in practical use. Some large carbon arcs may yield even two or six times the total heat energy 734 VERHOEFF AND BELL. of the magnetite arc here in question while yet remaining within perfectly safe limits from any practical standpoint. No actual arti- ficial illuminant can fairly be considered dangerous from the standpoint of thermic action on the retina. It will be observed that since no screens were used, these experiments also afford evidence of the harm- lessness of abiotic radiations to the retina. Glassblowers' Cataract. The fact that glassblowers are subject to a special form of cataract has naturally raised the question whether or not the latter is due to radiant energy and if so, whether to abiotic or thermic action. Meyhofer ^*^ examined 506 glassmakers and found 11.6% affected with cataract. The cataract almost always appears first in the left eye which is the more exposed to the light. When it appears first in the right eye it has been found that the glassblower has been in the habit of turning this eye towards the oven (Stein ^*^) . The glassblow- ers show also a peculiar rusty brown spot on each cheek, more marked on the left. The length of time necessary for the development of the cataract has not been exactly ascertained, but it is evidently several years. The cataract usually begins before the age of 40. In Mey- hofer's series the youngest blower was aged 15 years and the youngest affected with cataract 17 years. Of 59 cases of cataract, 42 were under the age of 40. In the latter cases, both eyes were affected in 16, the left eye alone in 19, and the right eye alone in 7. The glassblowers are thin and delicate, and are subject to asthma and pulmonary tuber- culosis. Almost all of them have emphysema of the parotid gland. During their working hours they perspire excessively and in conse- quence drink enormous quantities of fluids, including beer, coffee, wine, and lemonade. The cataract begins as a rosette like or diffuse opacity in the cortex at the posterior pole of the lens, the remainder of the lens for a long time remaining clear. Later, striae similar to those of senile cataract may appear. At operation the nucleus is found larger than that of other individuals of the same age (Stein ^*^), and the capsule more fragile (Cramer ^^). Parsons ^^^ states that in only one-fifth of the cases does the cataract differ in appearance from a senile cataract, but this probably applies to the late stages. Hirschberg ^^^ states that for over 100 years it has been recognized by various observers that individuals exposed to intense heat and light EFFECTS OF RADIANT ENERGY ON THE EYE. 735 are especially liable to cataract. Peters ^^^ suggested that the cata- ract was clue to the venous stasis in the vortex veins associated with the forced expiration, analogous to the cataract produced experi- mentally by tying off the vortex veins. Leber ^^^ advanced the view that it was due to concentration of the aqueous resulting from evapo- ration from the cornea and the loss of water from excessive perspira- tion. Parsons ^^^ suggested that the cataract results from altered nutrition due to overheating of the ciliary body. Cramer ^^, Stein ^*^, and others believe that it is due to ultra violet light (chemical action), while Vogt ^^^ regards the infra red rays as chiefly responsible. The great frequency with which glassblowers' cataract occurs, its relatively uniform type, and the fact above all, that it occurs first in the more exposed eye, show clearly enough that it is chiefly due to the action of radiant energy on the eye itself. This is supported also by the fact that the cheek shows a more marked area of discoloration on the side of the first affected eye. The further questions whether the cataract is due to the direct action of the light upon the lens, or upon the eye as a whole, and whether it is due to abiotic or thermic action, are not quite so easily answered. The character of the radiation from molten glass is well known. It is that from a homogeneous body of relatively low temperature, 1200° to 1400° C. It is certain that the spectrum of a non gaseous body at this temperature does not include any of the so-called abiotic radia- tion since the extreme limit of the spectrum of molten glass found by any investigator is 320 nfx and estimates range from that to 334 /xju. We have already shown that the abiotic action cannot be traced beyond 305 nfj, and there is not the slightest indication from our researches or any predecessors that there is reason to suspect an extension of such activity to waves longer than 320 MM- Even if there were, such rays would be stopped at the front of the lens by its absorption and hence would be unable to affect the posterior cortex. Moreover the radiation of a body at such temperature is relatively very weak all through the ultra violet, the maximum, according to Planck's law, for a body at 1300° C. lying far in the infra red while the energy in the whole visible and ultra violet part of the spectrum is less than 1% of the total. Hence to ascribe injurious effects to the visible or ultra violet radiation without eliminating once and for all the 99% of infra red radiation is to lose all sense of proportion between cause and effect. To follow up the theory of the matter a little further we have shown that the specific .abiotic action is clearlv to be eliminated. Of the ravs whicii are 736 VERHOEFF AND BELL. absorbed by the lens, reaching from 300 ijlix to about 400 nfx, the chief absorption is, following the general theory which we have already explained, at the front surface, hence if by any stretch of the imagina- tion glassblowers' cataract could be assumed to be due to an indefi- nitely long application of such rays it should occur at the anterior and not at the posterior cortex. To rays in the ordinary visible spectrum the lens is notoriously transparent and in default of absorption of energy there is no reason to expect any specific effects from it. We have been able by the use of sources of extreme power greatly concen- trated, as our experiments show, to obtain specific action of the ultra violet rays only to a microscopic depth, 20 ju, so that the experimental evidence lies squarely against any lesions directly producible by such sources in the posterior cortex, particularly in the absence of any effects in the anterior cortex. This reasoning also holds for the time integrals of any effect of such raj's over periods however long, for whatever the aggregate effect might be it would always remain much greater at the anterior than at the posterior part of the lens substance. The absorption of the media of the eye for various wave lengths and especially those predominant in sources of the temperature considered has been investigated by Aschkinass ^ who has shown that the characteristic absorption of the long waves is essentially that of water. An analysis of Aschkinass' results by Luckiesh ^^° shows that for sources of moderate temperature the resulting absorption is chiefly in the cornea and aqueous, that is in the outer layers of the absorbing media. For a source of the tempera- ture of a glass furnace, between 80 and 90% of the energy will be absorbed by the cornea alone, and the amount stopped in the lens from all causes would not be over 3 or 4%, but this small fraction of the total is still absolutely a considerable quantity and is somewhat concentrated in the lens since not less than three-fourths of the total refraction of the eye is due to the cornea. There is therefore a slight tendency to concentrate energy toward the rear of the lens. Such concentration must be small, however, owing to the contracted pupil and the resulting narrow angle of the pencil of rays entering the lens. It is doubtful whether the actual concentration would reach more than a few per cent and the effect of this would be more than offset by the greater al)sorption in the anterior cortex. As regards the distribution of temperature in the eye resulting from the intense radiation most of the absorption takes place as stated, in the cornea, and a greater proportion in the aqueous than in the lens. The front of the cornea is of course rapidly cooled by convection and EFFECTS OF RADIANT ENERGY ON THE EYE. 737 the fluid of the anterior chamber also gets some facility at cooling by convection currents. The iris which strongly absorbs most of the energy which falls upon it especially if strongly colored, to a certain extent screens the front surface of the lens behind it especially since the circulatory system in the iris tends to prevent its temperature rising materially unless the access of energy is above the rate at which circulation can take care of it. At the rear of the lens the vitreous with its fibrillated structure effectively prevents, like all such sub- stances, the existence of convection currents. Just what the net effect of the structure is upon the steady distribution of temperature when the eye is exposed to radiation cannot be quantitatively de- termined, and while undoubtedly the heat reaching the rear of the lens from the energy transmitted to that point, or received from that taken up by absorption in the anterior part of the eye, cannot readily escape and hence tends toward concentration it seems somewhat doubtful whether this cause alone could determine the starting of cataract at the posterior cortex. We are inclined to attach more importance to the suggestions of Leber ^^^ that the effect is a secondary one due to the loss of water in the drain produced by the heat on the front of the eye and elsewhere, and especially to that of Parsons ^^^ that malnutrition due to interference with the functions of the ciliary body by the heat may be chargeably w^ith the malady. We are the more inclined to this opinion since the intense radiation acts through the sclera as well as through the cornea, thus aft'ecting the whole structure. It is also a well known clinical fact that diseases of the fundus produce at times cataractous changes in the posterior cortex that are held to be due to impaired nutrition of the lens. The develop- ment of glass blowers' cataract is so slow that it is quite hopeless to reach its cause experimentally, but from the facts here stated we incline to the opinion that these secondary effects of radiation are more important in producing it than the specific action of radiation in producing localized effect at the posterior cortex. In any case it is perfectly clear that abiotic radiations are not concerned, and have nothing to do with the matter. Applications to Commercial Illuminants. In considering any possible deleterious effects of radiation upon the eye there are certain pathological effects which can be at once eliminated, at least from any consideration of commercial illuminants. 738 VERHOEFF AND BELL. First, by the experiments which have been heretofore described we have made it clear that there can be no injury done to the retina by ultra violet light as such, even the most severe exposures failing completely to produce any effect whatever. Second, all thermic effects of energy from any source on the external eye are at once ruled out of consideration by the immediate discomfort produced by excessive heat. No person would tolerate extreme heat radiation on the external eye for a period long enough to produce the slightest damage. There remain therefore to be considered thermic effects within the eye, and specially those due to the focussing of intense radiation upon the retina as in eclipse blindness; over stimulation of the physiological processes in the retina, that is pathological effects due to light as such in its action on the retinal structure, and finally those abiotic effects of the extreme ultra violet rays on the external eye properly known as photophthalmia. It has been our purpose to ascertain the practical risks incurred in the use of artificial illuminants and the precautions required to avoid danger. To this end have we sought especially the quantitative relations in the action of radiant energy upon the eye. We have, therefore, experimented with the most powerful sources used for practical lighting under conditions of intensity immensely greater than occur in their every day use. We have shown (page 728) that as regards the general thermic effects of energy upon the eye there is no chance of damage to the retina or to the media of the eye under any practical conditions of use. With respect to damage to the retina in particular we have been unable to produce it except by exposures and intensities enormously greater than could possibly be reached in the use of artificial sources of light. We have used beside the quartz mercury arc which is not particularly strong in general radiation, a 750 watt magnetite arc which takes the greatest amount of energy of any arc light ordinarily used for illumi- nating purposes and the 750 watt nitrogen lamp which gives more ultra violet than any other incandescent source, and which failed completely to produce any specific damage to the eye, although in one experiment the animal was overcome by the general heat effect, as in sunstroke. This occurred after an exposure of 1| hours at 20 cm. from the filament. A second experiment with an exposure of 2 hours at the same distance, in which the animal was well protected from the heat and the head kept cooled with water, showed no damage to the eye of any kind. This source as a whole focusses less sharply than does the arc lamp taking an equal amount of energy, and in this case we have already shown that the retina would not be subject to damage EFFECTS OF RADIANT ENERGY ON THE EYE. 739 save by an impossibly long exposure with accurate fixation. These experiments were at far shorter distances and consequently of enor- mously greater intensities than any which could be found with such sources as illuminants, and we may hence conclude that so far as general effects of thermic energy are concerned no source used for illuminating purposes is capable under working conditions of produc- ing any observable deleterious results. Experiments with the arc lamp are crucial because this source is nearer to being a point source than any illuminant of similar power and hence gives much sharper concentration of energy in the image. It is with this concentration in the image that the possible damage to the retina is concerned. Diffuse sources which do not come to a definite focus may be entirely neglected for this particular purpose. Low temperature sources of which most of the energy lies in the ex- treme infra red are even less effective in concentrating energy upon the retina, because the eye never focusses such wave lengths on the retina except as a diffuse spot. The only sources from which there seems to be any material danger from the standpoint of thermic effects are certain very powerful high temperature sources used in the arts, such as heavy arcs used for welding purposes, furnaces, electric or other, where there is customarily great concentration of energy, and such purely accidental phenomena as short circuits. Perhaps some of the very powerful arcs used in searchlights might be included in this class of the possibly dangerous. Ordinary discretion in avoiding disagreeably powerful lights or suita- bly shading the eyes from them should avert easily all real danger from any of these sources of radiation, save the short circuits which are accidental rather than ordinary risks. Certainly no sources used for lighting purposes can be classified as dangerous from the standpoint of thermic effects. We wish particularly to emphasize the fact that so far as any possible temporary or permanent injury to the retina is concerned such action must depend on the concentration of energy in the image. Conse- quently, extended sources of moderate intrinsic brilliancy are to be preferred to intense sources. Hence it is desirable to protect all sources naturally of high intrinsic brilliancy by diffusing globes. As regards dangers of injury to the eye from light radiation as such, our experiments indicate that it has been very greatly exaggerated as regards its pathological possibilities. It is undoubtedly true that brilliant sources of light are disagreeable and that they produce unpleasant effects in temporary scotomata, disturbance of color 740 VERHOEFF AND BELL. vision, persistent and annoying after images and fatigue due to efforts to overcome the difficulties of vision under these disadvantages. As regards definite pathological effects or permanent impairment of vision from exposure to the luminous rays alone we have been unable to find either clinically or experimentally anything of a positive nature. The experiments on monkeys, which we have recorded, show very clearly that exposure to light of intensity many times greater than anything to be found in the use of commercial illuminants produced only temporary scotomata. The lid reflexes appeared within a very few minutes and the scotomata seemed to have worn away within at most a few hours. There was not the slightest sign of permanent impairment of vision. As noted on page 684 there were indications that the process of light adaptation may go on to a considerable degree even during very severe exposure to light. In the experiments on the human eye results were found closely comparable with those obtained in the earlier experiments on monkeys. The erythropsia passed away in a few minutes and the scotoma wore away rapidly so that after three hours the visual acuity was normal, although there still remained traces of color scotoma. After 22 hours visual acuity remained normal and the central color vision for red, blue and green was also normal. This intensity of the light in this case was far in excess of anything which could be reached in the use of commercial illuminants and there was a length of fixation many times greater than could ever be found in practical use of lights. These experiments seemed con- clusive in showing that the effect of even extraordinarily severe exposure to luminous rays produces only such temporary effects as might reasonably be expected and is followed by no lasting injuries of any kind. Whether frequent and long exposures of a similar kind might exhaust the extraordinary recuperative powers of the eye is a matter on which in the nature of things there can be no direct ex- perimental evidence and which is not of practical importance since under no working conditions could even a single exposure compar- able in severity with those obtained in our experiments be produced. There is, however, very strong clinical evidence that even severe daily exposures to intense light lasting over many years fails to produce material injury to the eye. For in the case of glass blowers there is extreme exposure both to the heat and light of the furnace occurring daily for many years. While glass blowers' cataract, probably arising as we have shown from secondary causes quite aside from the direct effects of radiation, may be produced under these circumstances, there has never been noted any injury to the retina. The fact that EFFECTS OF RADIANT ENERGY ON THE EYE. 741 the retina is uninjured under these extreme conditions seems to indi- cate as do our experiments that the eye is remarkably tolerant of intense light even under the circumstances of exposures of great severity lasting over very long periods of time. These results do not justify the use of powerful unscreened sources near the eye, since in this condition vision becomes difficult and the effects of eye-strain due to other causes than mere illumination of the retina become unpleasantly in evidence. They do show, however, that the eye in the process of its evolution has acquired the ability to take care of itself under extreme conditions of illumination to a degree hitherto deemed highly improbable, and that the effects on the retina due to any exposure to intense light in the least degree likely to be found in the use of practical illuminants are temporary and of no pathological significance. A fortiori, there is not even a remote chance of pathological effects on the structure of the eye due to light received from extended surfaces of low intensity like translucent globes and diffuse reflections as from paper. As for the ultra violet part of the spectrum to which exaggerated importance has been attached by many recent writers, the situation is much the same as with respect to the rest of the spectrum, that is, while under conceivable or realizable conditions of over exposure injury may be done to the external eye yet under all practical condi- tions found in actual use of artificial sources of light for illumination the ultra violet part of the spectrum may be left out as a possible source of injury. All illuminants possess an easily measurable amount of ultra violet radiation ranging, as one of us has already shown ^^ from about 4 ergs per second per square cm. per foot candle of illumi- nation in the quartz arc with the usual globe, to more than tw^enty times this amount in the enclosed carbon arc shining through a quartz window. Between these two lie the whole range of incandescent lamps both gas and electric, the ordinary mercury arcs and the ordinary Cooper Hewitt tube, flames, arc lamps of various sorts, and sunlight. The last mentioned occupies an intermediate position between the high efficiency electric incandescent lamps and the older incandescent lamps or ordinary flames. The ultra violet in these various sources is distributed in different ways. All the flames and incandescents give continuous spectra which die out for even the highest tempera- ture -of these sources at about wave length 300 ju^- Sources giving discontinuous spectra generally extend below this limit of wave length, but often with very feeble radiation in this region. Such, for instance, is the case with the carbon arcs, which show chiefly metallic 742 VERHOEFF AND BELL. impurity lines within the very short wave lengths and owe their considerable proportion of ultra violet to radiation just outside the visible spectrum. From the standpoint of effects upon the eye the ultra violet region may be divided into two sharply separated portions, one of which produces abiotic effects while the other does not. We have for the first time definitely established the line of partition between these two portions at 305 /x^t. Some of the earlier experi- menters in this field imagined that they had detected abiotic effects with slightly longer wave lengths, an error apparently due to insuffi- cient knowledge of the absorbing screens which they employed, which with rare exceptions are described, if at all, in very loose terms. No injurious effects have been attached with any reasonable degree of certainty to the ultra violet radiation which lies between the end of the visible spectrum and the beginning of the abiotic rays. Since this range of radiation is present in considerable amount in ordinary sunlight, it is sufficiently obvious that any definitely harmful results producible under ordinary conditions would have been eliminated by the ordinary progress of evolution. Artificial illuminants under any practical conditions of use expose the eye to much less severe radia- tion in this part of the spectrum than does ordinary daylight and a fortiori can be excluded as possible sources of harm. With respect to abiotic radiations we have every reason to acquit on sound experimental basis every known artificial illuminant when working under the ordinary conditions of commercial use. Even the c^uartz mercury lamp, which is per se richer in abiotic radiations than any other commercial source of illumination, when equipped with its ordinary globe is not only less rich in ultra violet per candle power given than any other source, but is as we have found by experiment incapable of producing any abiotic effects on the eye even after six hours exposure at 30 cm. from the tube. We have further shown that even where the lamp is used without its globe, a condition which is avoided in consideration of efficiency, long exposures would still be necessary to produce any injurious effects at any distance reasonably to be expected. The same immunity from danger attaches to all the sources at present in commercial use. It is well within the bounds to say that there is no commercial illuminant from which the least risk of abiotic radiation is incurred under the circumstances of practical use. The only sources used in the arts which have abiotic power enough to require special caution in their use are those not employed for the purpose of illumination. Such, for example, are the powerful arcs used in some electric welding processes, those employed in the EFFECTS OF RADIANT ENERGY ON THE EYE. 743 fixation of nitrogen from the air and lamps specifically designed for abiotic purposes, like the ultra violet lamp of Henri, Helbronner and de Recklinghausen.^^* The last mentioned is a quartz arc of peculiar form taking 1150 watts and giving approximately six times as much abiotic radiation as the lamp used in our experiments. Such a source would therefore give typical photophthalmia in about one minute at 50 cm. and milder symptoms of conjunctivitis and erythema in somewhat less than this time. A comparable degree of activity is indicated for the other sources here referred to. In these cases con- siderable caution must be exercised to avoid even short exposures to the unscreened source, but arcs of this character are highly special in their functions and have no connection with matters of illumination. As regards the general effects of the ultra violet portion of the spec- trum, it must be remembered that the abiotic action is chiefly super- ficial and we have shown that even under exposures of great severity there are no indications of any injury to the retina from ultra violet rays even in the aphakic eye. In order to make it clear that the results as to abiotic action, which we have obtained from animal experimentation, are substantially applicable to the human eye as well, we may from the standpoint of general theory point out that the effect of the abiotic rays we have shown to be definitely dependent on the quantity of the radiation, the action of which can be reckoned much as if it were a mere mechani- cal force. There is no reason from our experiments or those of others to suppose that such radiations act with much greater intensity on one kind of living cell than on another. We have in addition ample direct evidence that the effect on the human eye and on the rabbit's eye are entirely comparable, for we have shown experimentally that the critical amount of abiotic radiation for photophthalmia on the rabbit's cornea and for erythema on the human skin is the same. As a clinical fact, of which the early observations of Charcot ^^ are typical, in every case of photophthalmia the erythema of the skin surrounding the eye is quite as conspicuous as the conjunctivitis. In fact, wdiile in practically every case of photophthalmia erythema appears, it is very rare in clinical cases to find the stippling of the cornea taken as one of our characteristic symptoms, hence the amount of abiotic energy required to produce photophthalmia, since it will also produce erythema, must be substantially as great for the human eye as for the rabbit's eyes on which most of ovir experiments were performed. Our general conclusion, therefore, regarding the effect of radiation 744 VERHOEFF AND BELL. from practical illuminants on the human eye is that no sources com- mercially employed for such a purpose are to be regarded as dangerous and that the most ordinary care in providing illumination with which comfortable vision can be obtained is sufficient for complete security against all possibility of injury from radiation. Protective Glasses. As we have shown, the lens completely screens the retina from abiotic radiations so that attempts further to protect it from such radiations by means of glasses of any kind are superfluous. We have also shown that the retina even of the aphakic eye, imder ordinary conditions is in no danger of injury by any source of light in common use, and is no doubt completely protected by the thick cataract glasses usually worn. In addition we have shown that the retina under ordinary conditions is in no danger of injury from the heat generated within it by the light from such sources. Heat effects are to be feared only in the case of extreme light intensities such as direct sunlight, and, exceptionally, short circuit arcs and lightning flashes. Against these any of the extremely dark glasses are effective. As regards the external eye, as just pointed out, it also is in no danger from abiotic radiations from any of the usual light sources. Photoph- thalmia may of course readily be produced by sufficiently long ex- posures at close range to high power arc lights of any kind, or to the quartz or uviol mercury vapor lamps, but it is only under special conditions that such exposures would occur. Here ordinary spectacles of crown glass usually afford sufficient protection, and adequate pro- tection would certainly be afforded by any of the ordinary yellowish, greenish or grayish protective glasses in common use, preferably in the form of coquilles so as to exclude lateral light. These would also afford ample protection against snow blindness or the photophthalmia produced by short circuits. Birch-Hirschfeld ^^ states that from his own personal experience in protracted and regular working with the uviol lamp he found complete protection from photophthalmia with the smoke gray spectacles, and intimates that Stockhausen's claim of having photophthalmia after working with the electric arc lamp, in spite of the fact that he wore ordinary spectacles, is readily explained by the circumstances that common spectacle lenses do not adequately protect the eye from radiation entering laterally. For protection of the external eye against extreme heat such as EFFECTS OF RADIANT ENERGY ON THE EYE. 745 that to which glass blowers are exposed, few glasses except such special ones as have recently been devised by Crookes, are very effective, the deep green copper oxide glasses being perhaps as good as any. Such glasses combined with amber or yellow green tints so as to reduce the transmitted light to a moderate amount in the middle of the spectrum are probably the most efficient protection against extreme radiation of all kinds, except that of direct sunlight. In the case of glass blowers, it is difficult and perhaps futile for them to make use of any sort of glasses, owing to excessive perspiration. So far then as concerns actual injury by light, the eye under ordi- nary conditions of modern life is in no danger. The question of wear- ing protective glasses so far as concerns the ordinary individual there- fore narrows itself down to the determination of those best adapted to obviate the sensations resulting from too intense illumination. These are unpleasant in the same way as are extremely loud noises, and for protection against them one wears glasses that reduce the light as one plugs the ears with cotton in a boiler factory. Any glass that reduces the light is effective for this purpose, but preferably, perhaps, a glass that transmits light chiefly in the middle of the spectrum, for which the eye is customarily focussed. The question of the color of the glass is, however, of little importance and the personal idiosyncrasies of the individual may be safely allowed free play here. It is probable that if such glasses are too long worn they will increase the sensitiveness of the individual to light. The question of the use of protective glasses in pathological condi- tions of the eye does not specially concern us here, but it may be stated that it is one simply of reducing the intensity of light reaching the retina. In cases of iritis this is possibly of some importance since it favors dilatation of pupil. In cases of glaucoma, on the other hand, excess of light is desirable, since it contracts the pupil. As Hess ^^* has pointed out, the sensations incident to the so called photophobia, associated with keratitis and other irritable conditions of the eye, are present in the dark as well as in the light, so that it is evident that undue importance has been attached to the exclusion of light in these conditions. As regards fundus conditions, the use of protective glasses has no rational basis, except possibly in the case of retinitis pigmentosa and allied conditions, as suggested by Axenfeld^. The exaggerated attention that has been paid in recent years to the harmful effects of ultra violet radiations has had one good effect in modifying the character of the protective glasses prescribed by ophthalmologists. Earlier practice was based on a general desire to 746 VERHOEFF AND BELL. cut down the light received by patients whose eyes were for one reason or another sensitive. This requirement resulted in the production of glasses of more or less dark neutral tints and sometimes dark shades of colored glasses commercially obtainable like green and blue. A later phase of practice reflected the view common a quarter century ago that red light was a thing producing in some unknown way spe- cially bad effects and consequently was to V)e shunned. Hence it was not uncommon to prescribe glasses of cobalt blue, green, and various amethyst tints. It is interesting to note that such glasses, while they reduce greatly the red of the visible spectrum still transmit quite freely the nearer part of the infra red which carries a large amount of energy (Coblentz ^^). In fact, of this nearer infra red such glasses transmit almost as much as does ruby glass. Therefore, the earlier protective glasses were not effective in cutting off heat radiation and tended to transmit mainly light toward the blue end of the spectrum. Perhaps the chief benefit of the agitation that has taken place within the last decade on the possible, though as we have shown highly improbable dangers of the ultra violet, has been the bringing into prominence the new types of protective glasses. These, intended primarily for eliminating the ultra violet rays, have tended to types of selective absorption which give advantageous results in modifying the visible light, which is really the chief object of concern to the ophthalmologist. There is a close kinship in the absorption of most of this recent crop of glasses. The prototype runs back nearly tw^enty years to the work of Fieuzal ^^^ who produced, specially with a view to protecting the eye against glare in the high mountains, a grayish green glass which cuts off the ultra violet very completely and shades down the blue so as considerably to shorten the spectral range of the rays transmitted. As one of us has already shown ^^ the last line transmitted by this glass from the spectrum of the quartz arc is 404.6 jjlh very faintly. Its absorption much resembles (loc. cit.) that of ordinary amber glass, except that the latter carries somewhat heavier absorption into the blue green accounting for its yellowish rather than greenish tinge. Either of these glasses is substantially as efficient as any of more recent origin in cutting out the ultra violet. In more recent times Hallauer^^^ and Schanz and Stockhausen ^°^, have discussed at length glasses protective against the ultra violet and have brought out special protective glasses with this point in view% the former known by the name of the inventor, the latter under the trade designation " Euphos." At about the same period the firm of Rodenstock produced a glass of similar type under EFFECTS OF RADIANT ENERGY ON THE EYE. 747 the trade name of "Enixanthos" and various approximations to and imitations of these have from time to time appeared. A paper by Hallauer ^^^ gives the result of a rather thorough spectrographic study of all the protective glasses then in common use to which those inter- ested in the subject may be referred. The composition of most of these glasses is held as an unnecessarily solemn secret, but it is gener- ally understood that they are essentially iron-lead glasses. They run in color from a distinctly yellowish to a somewhat bluish green, varying in tone and transmissibility according to the composition of the various shades put on the market. An ordinary green glass bottle gives in the spectrograph about the same absorption as the medium densities of any of the glasses referred to. The deeper shades of any of them cut off the spectrum completely just about at the beginning of the ultra violet and weaken it well into the violet. The lighter shades transmit to wave lengths 360 ^^u to 320 ix^x with cor- respondingly less reduction in the general intensity. An^^ and all of them are completely effective against the abiotic radiations, even although the medium shades sometimes transmit very weakly in the region 3200 jjlix to 3400 mm as well as in the visible spectrum. To render the record of these recent protective glasses fairly complete there is shown in Plate 7 the iron arc spectrum and its transmission through the three ordinary grades of the Hygat glass of Rodenstock, an excellent type of this general group. Figure 6 is here the iron arc spectrum, 5, 4, and 3, respectively the light, medium and dark Hygat glasses, and Figure 1 a special protective glass of American manufacture designed specifically to reduce the red end of the spectrum beside cutting out the ultra violet and much of the violet and blue. In point of effectiveness in cutting out the ultra violet the differences between these various glasses of approximately the same shade are inconse- quential and the choice between them lies mainly in the matter of taste as regards their particular color and absorption in the visible part of the spectrum. Of the recent glasses exploited in America the so called Noviol glass is remarkable for its extraordinarily sharp cut off of the spectrum in the blue. Voege^^^ in answering the question as to what spectral range of radiation gave the most satisfactory results held that the light from the clouds of a clear sky, being that light to which the eye through evolution had become adapted, was on the whole to be preferred. This would indicate the use of neutral glasses without selective absorp- tion. Hertel and Henker ^^^ found that clouds and clear sky con- tain very little energy below 310 mm and considered that the onlv 748 VERHOEFF AND BELL. protection needed for artificial light sources was to reduce the light from them at the working distance to substantially the range of the sky spectrum. From these and other experiments they concluded that the best protective glass should preferably reduce the spectrum to approximately that of cloud or sky light. This again indicates the use of neutral non selective absorbing glasses. Against this view it may be properly objected that the eye in its evolution has rejected the whole infra red and ultra violet as ineffective and in fact derives very little useful illumination from the red at the one end and the violet and blue at the other. The luminosity values of these portions of the spectrum are very small and it may be added, fortunately very small, else the chromatic aberration of the eye would make distinct vision quite impossible. We are inclined, therefore, rather to the view that such radiation as produces the maximum reqviired lumi- nosity with the minimum energy access to the eye is best adapted to protect the eye from any and all injuries which may be due to exces- sive radiation. This indicates the use of glasses absorbing at both ends of the spectrum so as to bring the strongest light in the region of great- est luminosity, that is in the yellow green. As one of us has already shown ^^ composite spectacles reducing the spectrum to a nearly monochromatic stripe in this region actually enable one to view the most powerful sources without discomfort while yet transmitting enough light to permit writing or reading one's notes. The glasses of Plate 7 in the deeper shades all show something of this character- istic absorbing both ends of the spectrum and in so far represent a slightly different type from the glasses which have preceded them. Crookes found that suitable absorption at both ends of the spectrum could not be obtained without encroaching somewhat on the visible portion but rendered this encroachment rather inconspicuous by using a heavy didymium glass which cuts out the yellow and leaves a pinkish tinge. This, however, is not an objection in cases requiring thorough protection unless the encroachment is so great as actually to be inconvenient in seeing. Where, therefore, practical protection against powerful sources of radiation is necessary, glasses meeting the requirement of maximum luminosity with minimum energy present material advantages. These advantages become practically inconse- quential where the question is one of merely moderately reducing too bright general light, and the choice between such special protective media and ordinary neutral tint glass reverts again to a matter of taste. ULTRAVIOLET LIGHT AS A GERMICIDAL AGENT. Experimental Investigation of its Possible Therapeutic Value.* f. h. verhoeff, a.m., m.d. As is well known, light-waves of sufficiently short wave-lengths are highly germicidal to bacteria suspended in mediums which are trans- parent to these waves. The question has arisen, therefore, whether or not it may be possible to make use of ultraviolet light in the treat- ment of local infections. Ultraviolet light has long been successfully used by Finsen in the treatment of certain skin diseases, notably lupus vulgaris, and re- cently has been employed by ophthalmologists in the treatment of vernal catarrh and trachoma, also, it is asserted, with successful results. Its beneficial effect in these conditions, however, obviously is not necessarily due to a direct germicidal action, but possibly only to an irritant action on the tissues. Since the cornea compared to other tissues of the body is relatively transparent to ultraviolet light, it follows that if it should prove im- possible by this means to destroy bacteria within corneal tissue without at the same time producing undue injury to the tissue itself, the same negative results would be obtained in the case of all other tissues. For this reason the present investigation was confined to experiments on the cornea. These experiments were made in connection with an investigation by Louis Bell and myself on the effect of ultraviolet light on the normal eye, advantage being taken of the powerful light sources and apparatus therein employed. Hertel ^, in 1903, reported the successful use of ultraviolet light in the treatment of corneal ulcers, asserting that here it had a direct germicidal action on the infecting bacteria. He also made some interesting experimental observations in this connection, the most important one from a therapeutic point of view being that he was able to abolish the motility of cholera bacilli enclosed in a quartz cell and * Reprinted by permission from the Journal of the American Medical Asso- ciation, March 7, 1914, Vol. LXII, pp. 762-764. 1 Hertel, E.: Experimentelles liber ultraviolettes Licht, Ber. li. d. 31 Vers, d. ophth. Ges., Heidelberg, 1903, p. 144. 750 VERHOEFF AND BELL. placed within the anterior chamber of a rabbit's eye, by exposing them to the action of ultraviolet light passing through the cornea. The source of the light was a magnesium electrode giving off rays with wave-lengths from 0.28 to 0.309 microns, and the exposures were from twenty-five to thirty minutes. The current was from 3 to 4 amperes. Later he obtained the same result with his cadmium-zinc electrode. Hertel assumed that the bacteria were actually killed, but he did not state that this was demonstrated by means of cultures. He also did not exclude the possibility that the effect on the bacilli was due to heat. Hertel, in addition, tested the therapeutic action of ultraviolet light on a series of rabbits' eyes in which he had produced staphy- lococcic corneal ulcers and obtained " pleasing results." The resulting scars were slight and no changes could be found in the depths of the eyes. These results, however, it seems to me, lose any possible significance when it is considered that staphylococcic corneal ulcers artificially produced in rabbits as a rule promptly heal without any treatment, as I have frequently observed. Hertel maintains that light of short wave-lengths has a greater deleterious effect on bacteria than on tissue-cells. This may be true for very short waves, but it is certainly not true for waves which are able to pass through the cornea. Thus, I found that severe keratitis could be produced by exposing the cornea through a crown screen to a quartz mercury -vapor lamp at a distance of 20 cm. for one and one- half hours, whereas staphylococci suspended in distilled water and exposed under the same conditions were not killed in six hours ^. This experiment also would seem almost alone sufficient to prove the impossibility of destroying bacteria within the clear cornea without producing too much injury to the corneal cells. This being the case, it is almost inconceivable that bacteria could be destroyed in a cornea infiltrated with pus-cells and so made practically impassable to germi- cidal waves. Hertel also attached importance from a therapeutic point of view to the conjunctival hyperemia and cell irritation produced by ultra- violet light. The practical value of these factors is questionable, and the latter factor would seem more likely to do harm than good in the case of corneal ulcers in which the cells already have sufficient un- 2 It is important to note that for these long exposures it is necessary to keep the bacterial container surrounded by cool water, as otherwise the bacteria may be killed in an hour or so by the accumulated heat. The lamp and screen used in this experiment are described later. EFFECTS OF RADIANT ENERGY ON THE EYE. 751 favorable influences to contend with. In any case they would not justify the use of ultraviolet light in treatment of such conditions in the absence of any germicidal effect of the light. In a later communication, HerteP reported in detail a series of clinical cases of corneal ulcer treated by means of ultraviolet light from a cadmium-zinc electrode. The latter he recommended as being equal in efficiency to the magnesium electrode and at the same time more practical to use. Twenty-six cases were treated with light therapy alone, 8 cases with light therapy followed by Saemisch section, and 13 cases with light therapy followed by cauterization or the latter and Saemisch section. Thus in 21 out of his 47 cases of corneal ulcer, the result of the light therapy was so unsuccessful that cauterization or Saemisch section had to be undertaken. These results do not seem impressive for an agent that is supposed to kill the bacteria within the ulcers. Hertel exposed his patients from three to five minutes two or three times daily. At the most this was equivalent to a daily total exposure of only fifteen minutes. Now he had found that it required from twenty-five to thirty minutes to kill (or inhibit)? bacteria exposed through a perfectly clear cornea. How then could it be expected that an exposure of fifteen minutes would suffice to kill them in a purulent infiltrate which acts as a far more effective barrier to ultraviolet light? In a communication to appear later, Louis Bell and I show that interrupted exposures to ultraviolet light wdth intervals of less than twenty-four hours have practically the same effect on the cornea as a continuous exposure of the same total length. For this reason by frequently repeating his exposvu'es, Hertel undoubtedly increased the injury to the corneal tissue without at the same time, in all proba- bility, obtaining a corresponding increase in germicidal action. In his experiments and in the treatment of his cases Hertel employed no screens. Thus the cornea had not only to contend with the rays that could penetrate it, but also with those stopped within the stroma and at the surface. As the rays stopped near the surface are evi-, dently useless so far as killing bacteria within the stroma is concerned it occurred to me that by screening them out and so decreasing the damage to the cornea, longer exposures might safely be used, thereby increasing the possibility of a germicidal effect within the cornea. The screen selected for this purpose was a crown glass, which permitted 3 Hertel, E.: Experimentelles und klinisches iiber die Anwendung lokaler Lichttherapie bei Erkrankungen des Bulbus, Arch. f. Ophth., 1907, Ixvi, No. 2, p. 275. 752 VERHOEFF AND BELL. only waves greater than 0.295 microns in length to pass *. As will be seen, however, this procedure was unsuccessful. No germicidal effect on bacteria within the cornea could be noted even when exposures through this screen were used which were sufficient to produce severe keratitis and even injiu-e the epithelium of the lens capsule. As light sources in the following experiments the magnetite arc and the quartz mercury-vapor lamp were chiefly used. To obviate the remote possibility that the cadmium-zinc arc employed by Hertel might possess some special advantage, this arc was also used. That greater intensity was obtained with our cadmium-zinc arc and quartz lens than was obtained by Hertel is proved by the fact that not only severe keratitis but also marked changes in the epithelium of the lens capsule were produced. The mercury-vapor lamp used was the Cooper-Hewitt model with- out the globe (220 volts, 3.5 amperes). The magnetite arc was of the ordinary self-regulating type as known to trade, without the globe. The voltage was about 80, the amperage from 9.8 to 10. The light was passed through a quartz water-cell 5 cm. in thickness, and concentrated on the cornea by means of a quartz lens 4 cm. in diameter and 9 cm. in focal length, placed 20 cm. from the light source. In Experiments 6 and 7 still greater intensity was obtained by means of a second quartz lens 23 mm. in diameter and 15 mm. in focal length. In the case of the cadmium-zinc arc, the same apparatus was used except that the electrode consisted of an alloy of equal parts of cad- mium and zinc in a thin-walled copper cylinder, and was water-cooled. The water-cell was omitted. The voltage was about 80, the amperage about 6.8. A number of experiments were first made by injecting staphylo- cocci or pneumococci into the corneas of rabbits and after twenty-four hours exposing the resulting abscesses to the ultraviolet light. Healing did not seem to be hastened, but since recovery ultimately occurred, as it did also in the control eyes, these experiments are not regarded as sufficiently conclusive and are not given in detail. Experiment 1, however, in which tubercle bacilli were injected into the cornea, was 4 Since this wave-length has been found to be the hmit of transparency for the cornea, it would be expected that such a screen would protect the cornea from injury, the longer waves not being absorbed by the latter. As a matter of fact I have found that it does almost completely protect the corneal stroma, but permits severe injury or destruction of the epithelium, corneal corpuscles and endothelium. EFFECTS OF RADIANT ENERGY ON THE EYE. 753 perfectly conclusive since the resulting lesions continued to progress in both eyes alike. In the other experiments the exposures were made immediately after the injections, that is, with the corneas clear, so that the conditions were the most favorable possible for germicidal action of the light. The results in these experiments, moreover, are clear-cut, because if the light had killed the bacteria, abscesses would not have formed. This is proved by Experiment 2, in which in the control eye the bacteria were first killed by exposure to ultra- violet light before they were injected into the cornea. Experiments. Experiment 1. April 10, 1912, a suspension of virulent tubercle bacilli is injected into each cornea of a rabbit. May 1, each cornea shows a small tubercle. The right eye is ex- posed to the quartz mercury -vapor lamp through crown screen 1^ hours at 20 cm. May 21, both tubercles have developed as usual. The animal is killed. Experiment 2. June 22, 1912, a suspension of Staphylococcus^ aureus in distilled water is injected superficially into the left cornea of a rabbit. The remaining bacterial suspension is then exposed at 20 cm. for three minutes to the quartz mercury-vapor lamp. (Cul- ture taken proves that all organisms have been killed.) This suspen- sion of killed staphylococci is then injected into the right cornea of the same rabbit. Each cornea is exposed to the quartz mercury-vapor lamp at a dis- tance of 0.5 meter for fifteen minutes. June 23, both eyes show marked photophthalmia. The left cornea shows well-marked abscess. The right cornea shows only a faint haze along the tract of the needle. June 24, both eyes show increase in photophthalmia, with haze of corneal stroma and central loss of corneal epithelium. The abscess of the left cornea has increased in size and there is now hypopion. The right cornea shows no abscess. Enucleation is performed. Experiment 3. Oct. 21, 1913, a suspension of Staphylococcus aureus in distilled water is injected into each cornea of a rabbit, the amount injected into the right cornea being three times that injected into the left. The left eye is then exposed for thirty minutes to the 754 VERHOEFF AND BELL. cadmium-zinc arc through a quartz lens (no water-cell or screen of any kind being used). October 22, the right eye shows intense inflammatory reaction, with a large abscess of the cornea and pus in the anterior chamber. The left eye shows equally intense inflammatory reaction and a corneal abscess about half the size of that in the right cornea. The abscess shows discrete points evidently corresponding to colonies of bacteria. October 23, abscesses of the two corneas are now about equal in size, (3 mm. in diameter). The anterior chamber of each eye contains pus. Epithelium is entirely absent from the left cornea (this being confirmed by microscopic examination). The right cornea shows loss of epithelium onlj- in the vicinity of the abscess. Enucleation is performed. The lens capsule of the right eye after fixation in Zenker's fluid is examined in flat preparation and shows slight changes in the nuclei of the epithelium evidently due to the action of staphylococcus toxins, but no changes similar to those seen after exposure to ultraviolet light. The lens capsule of the left eye shows, in addition to these nuclear changes, well-marked changes characteristic of exposure to ultraviolet light — swelling and granular degeneration of the cytoplasm of the epithelial cells. Histologically both corneas present the same picture, and each contains numerous large masses of staphylococci. Experiment 4. Feb. 8, 1913, a suspension of Staphylococcus aureus in distilled water is injected into each cornea of a rabbit. The left cornea is then exposed for thirty minutes to the cadmium-zinc arc through a crown screen and quartz lens, the image being kept on the injected area. February 10, there are abscesses of equal size in the two corneas. The left eye in addition shows severe photophthalmia with marked general haze of cornea and large central loss of epithelium which in- cludes area of abscess. This eye also shows exudate adherent to the posterior surface of the cornea behind the site of the abscess, due e\idently to injury of the endothelium by the ultraviolet light. Enucleation is performed. The lens capsule of the left eye, after fixation in Zenker's fluid, is examined microscopically in flat preparation, and shows marked changes characteristic of exposure to ultraviolet light. Experiment 5. March 19, 1913, a suspension of Staphylococcus aureus in distilled water is injected into the cornea of a rabbit. The cornea is then exposed to the magnetite arc for forty-five minutes through the quartz lens, quartz water-cell, and crown screen. This exposure in the case of normal eyes had been found sufficient to cause EFFECTS OF RADIANT ENERGY ON THE EYE. 755 necrosis of the stroma cells and endothelium of the cornea, to cause hemorrhages in the iris, and to produce marked changes in the lens capsular epithelium. March 20, there is abscess of the cornea. Marked photophthalmia is noted with loss of epithelium. March 24, the abscess is larger. Enucleation is performed. Lens capsular epithelium (flat preparation) on microscopic exami- nation shows marked changes, and the iris shows numerous hemor- rhages characteristic of exposure to ultraviolet light. Experiment 6. Dec. 16, 1913, a suspension of Staphylococcus aureus in distilled water is injected into the cornea of a rabbit. The injected area is then exposed twenty minutes to the magnetite arc through a water-cell and a system of two quartz lenses. No screen is used. As previously determined, with this arrangement an exposure of thirty seconds is sufficient to cause marked keratitis and destruction of the epithelium, while an exposure of twenty minutes causes com- plete destruction of the corneal corpuscles and softening and swelling of the stroma down to Descemet's membrane, and ultimately leads to vascularization and cicatrization of the cornea. December 18, there is no abscess. Marked photophthalmia is noted. The cornea is hazy and epithelium is absent from three quarters of the surface of the cornea. December 26, there is no abscess. The inflammatory reaction is almost gone. The cornea is softened and swollen. December 29, the inflammatory reaction is increasing again (reaction of repair); the corneal tissue is very soft. Vascularization is well advanced. Jan. 5, 1914, vascularization of the cornea is complete. The in- flammatory reaction is subsiding. January 9, the vessels are beginning to disappear. The cornea is leukomatous. Experiment 7. Dec. 19, 1913, the suspension is injected and the exposure made as in Experiment 6, except that the time of exposure is six minutes. This exposure is sufficient to cause softening of the corneal stroma. December 21, marked photophthalmia is noted. There is an abscess at the site of the injection. December 23, the abscess is smaller. A culture is taken. Enu- cleation is performed. Culture shows abundant growth of staphylococci. Lens capsular epithelium shows marked changes. 756 VERHOEFF AND BELL. CONCLUSIONS. The results of these experiments prove conclusively that ultraviolet light cannot under any conditions destroy bacteria within the cornea, even when the latter is perfectly transparent, without at the same time severely injuring the corneal tissue. Destruction of bacteria within the transparent cornea was obtained only when a light intensity and exposure were employed sufficient to cause complete destruction of the corneal corpuscles and intense injury to the corneal lamellae (Experiment 6). Moreover, it does not seem possible that ultraviolet light could in practice be successfully used to destroy bacteria within a corneal abscess or ulcer, that is, when the cornea was no longer clear, even with the sacrifice of corneal tissue, as in the case of the actual cautery. For either the exposures would have to be impracticably prolonged, or such extreme intensity of light would be required that the heating effect would exceed that of the abiotic action. It is doubtful also if ultraviolet light of such intensity could be made available for thera- peutic purposes. It must be concluded, therefore, that so far as direct destruction of bacteria within any of the tissues of the body is concerned, ultra- violet light possesses no therapeutic value. GENERAL CONCLUSIONS. 1. The liminal exposure capable of producing photophthalmia to the extent of conjunctivitis accompanied by stippling of the cornea, is in terms of energy about 2 X 10^ erg seconds per square cm. of abiotic radiation of the character derived, for example, from the quartz lamp or the magnetite arc. About two and a half times this exposure, i. e., 5 X 10^ erg seconds per square cm. is required to produce loss of corneal epithelium. 3. The abiotic action of the cornea and conjunctiva produced by any radiating sources follows the law of inverse squares and is directly proportional to the total abiotic energy received. It can therefore be definitely predicted from the physical properties of the source. 3. After exposure of the eye to abiotic radiations there is a latent EFFECTS OF RADIANT ENERGY ON THE EYE. 757 period before any effects clinical or histological become perceptible. This period of latency in a general way varies inversely with the severity of the exposure, but a theoretical latency of 24 hours or more corresponds to an exposure entirely subliminal. 4. The combined effect of repeated exposures to abiotic radiations is equivalent to that of a continuous exposure of the same total length, provided the intermissions are not long enough to establish reparative effects. Approximately the exposures are additive for intermissions of somewhat less than 24 hours. Exposures of ^ the liminal given daily begin to show perceptible effect after about 6 exposures. Daily exposures of g the liminal repeated over long periods produce no effect whatever, except to give the external eye a degree of immunity against severer exposures. Actual abiotic damage to the external eye renders it temporarily more sensitive to abiotic action. 5. Abiotic action for living tissues is confined to wave lengths shorter than 305 n^i, at which length abiotic effects are evanescent, while for shorter wave lengths they increase with considerable rapidity. 6. For the quartz arc and the magnetite arc the abiotic activity of the rays absorbed by the cornea is eighteen times greater than those which are transmitted by it. To affect any media back of the cornea requires therefore at least eighteen times the liminal exposure hereto- fore mentioned. 7. Even with exposures as great as one hundred and fifty times the liminal for photophthalmia the lens substance is affected to a depth of less than 20 ^t, and this superficial effect undergoes in the rabbit complete repair. Such enormously intensive exposures, which we obtain with the magnetite arc and double quartz lens system may completely destroy the corneal epithelium, corpuscles, and endothe- lium. The corneal stroma may be strongly affected by waves shorter than 295 h^jl, which it completely absorbs, but is very slightly affected by the remaining abiotic radiation. 8. The histological changes produced by abiotic radiation are radically different from those produced by heat, and the cell changes are best seen in flat preparations of the lens capsule. The most char- acteristic change is the breaking up of the cytoplasm into eosinophilic and basophilic granules. 9. Changes in the lens epithelium like those following abiotic action, including the formation of a "wall" beneath the pupillary margin, are not exclusively characteristic of abiotic action, but may be produced by ordinary chemical reagents. They are, therefore, characteristic not of abiotic action alone, but of chemical action in general. 758 VERHOEFF AND BELL. 10. Abiotic radiations certainly do not directly stimulate, but on the contrary apparently depress mitosis. Their action in this respect also is materially different from that of heat. 11. The lens protects completely the retina of the normal eye even from the small proportion of feebly abiotic raj's which can penetrate the cornea and vitreous humor. 12. Experiments on rabbits, monkeys and the human subject prove that the retina may be flooded for an hour or more with light of extreme intensity (not less than 50,000 lux), without any sign of permanent injury. The resulting scotoma disappears within a few hours. Only when the concentration of light involves enough heat energy to produce definite thermic lesions is the retina likely to be injured. 13. The retina of the aphakic eye, owing to the specific and general absorption of abiotic radiations by the cornea and the vitreous body, ij adequately protected from injury from any exposures possible under the ordinary conditions of life, even without the added protection of the glasses necessary for aphakic patients. 14. To injure the cornea, iris, or lens, by the thermic effects of radiation, requires a concentration of energy obtainable only under extreme experimental conditions. 15. Infra-red rays have no specific action on the tissues analogous to that of abiotic rays. Any effect due to them is simply a matter of thermic action, and such rays are in the main absorbed by the media of the eye before reaching the retina. 16. Actual experiments made on the human eye show conclu- sively that no concentration of radiation on the retina from any artificial illuminant is sufficient to produce injury thereto under any practical conditions. 17. Eclipse blindness, the only thermic effect on the retina of common occurrence clinically, is due to the action of the concentrated heat on the pigment epithelium and choroid, this heat being almost wholly due to radiations of the visible spectrum within which the maximum solar energy lies. 18. The abiotic energy in the solar spectrum is a meagre remnant between wave lengths 295 (jlijl and 305 /xju, aggregating hardly a quarter of 1% of the total. At high altitudes and in clear air it is sufficient to produce slight abiotic effects such as are noted in snow blindness and solar erythema, the former only occurring with long exposures under very favorable circumstances and the latter being in ordinary cases complicated by an erythema due to heat alone. The amount EFFECTS OF RADIANT ENERGY ON THE EYE. 759 of abiotic energy required to produce a specific effect in solar erythema is substantially the same as that required for mild photophthalmia. 19. Erythropsia is not in any way connected with the exposure of the eye to ultra violet radiations, but is merely a special case of color fatigue, temporary and without pathological significance. 20. Vernal catarrh and senile cataract we can find no evidence for considering as due to radiations of any kind. 21. Glass blower's cataract often charged to specific radiation, ultra violet or other, we regard as certainly not due to ultra violet light but probably due to the overheating of the eye as a whole with consequent disturbed nutrition of the lens. 22. Commerical illuminants we find to be entirely free of danger under the ordinary conditions of their use. The abiotic radiations, furnished by even the most powerful of them, are too small in amount to produce danger of photophthalmia under ordinary working condi- tions even when accidentally used without their globes. The glass enclosing globes used with all practical commercial illuminants are amply sufficient to reduce any abiotic radiations very far below the danger point. 23. Under ordinary conditions no glasses of any kind are required as protection against abiotic radiations. The chief usefulness of protective glasses lies not so much in their absorption of any specific radiations, as in their reducing the total amount of light to a point where it ceases to be psychologically disagreeable or to be incon- veniently dazzling. Glasses which cut off both ends of the spectrum and transmit chiefly only rays of relatively high luminosity, give the maximum visibility w^ith the minimum reception of energy. For protection against abiotic action in experimentation, or in the snow fields, ordinary colored glasses are quite sufficient. 24. So far as direct destruction of bacteria within the cornea or any other tissues of the body is concerned, abiotic radiations possess no therapeutic value. This is due to the fact that abiotic radiations that are able to penetrate the tissues are more destructive to the latter than to bacteria. SYSTEMATIC REVIEW OF THE LITERATURE RELATING TO THE EFFECTS OF RADIANT ENERGY UPON THE EYE. By C. B. Walker, A.M., M.D. Chronological Account. Historically we find the first study of the properties of the ultra violet light in connection with the eye was commenced long before high power lamps were invented or the role of ultra violet rays in producing eye-injuries was established. The first work, was stimu- lated by purely scientific interest with no prophylaxis or therapy in view; in fact these latter factors did not enter for several decades. As early as 1845, Brucke ^^ laid the foundations for subsequent work, in his investigation of the reason for the invisibility of the ultra violet rays. In order to determine whether these rays failed to traverse the eye media or failed to stimulate the retina, he first studied the absorptive power of the eye media. He found that Gum Guaiacum had a characteristic bluish appearance in ultra violet light. By means of this substance he was able to say that the lens absorbed ultra violet rays strongly and the cornea and the vitreous humor to a less extent. Later with the assistance of Karstein he found with sensitive paper that a combination of lens, vitreous humor and cornea, diminished somewhat the intensity of the violet, began to absorb more just outside the visible spectrum, was especially strong on the "M" {372 n/j) line of the Draper spectrum, and practically total beyond, that is for rays less than 370 ^t/x. In 1852 Stokes ^^^ discovered fluorescence and thus afforded another means of studying the absorption of the eye media. Donders^^ with Rees in 1853 threw a solar spectrum upon a screen covered with qui- nine sulphate which by its fluorescence rendered the ultra violet rays visible. Various eye media, unfortunately enclosed in glass con- tainers, were then interposed in the path of the ultra violet rays. The absorption power of glass itself rendered these results of little value. In 1855 Helmholtz ^^^ studied the lower limit of the visible spectrum using a quartz prism, but his high-grade myopia interfered. He also used the fluorescing screen of quinine sulphate and studied the fluo- 760 EFFECTS OF RADIANT ENERGY ON THE EYE. 761 rescence of the crystalline lens, of various solutions and of the retina. The fluorescence of the latter he discovered in the morbid state with Setschenow. A number of observers, Eisenlohr ^°^ in 1856, Janssen ^^^ in 1860, Franz ^^i in 1862, Listing 228 in 1865, Mascart 2*° in 1869, Sekulic ^^* in 1872 and Sauer ^°^ in 1875 continued the study of the same question of the length of the visible spectrum in much the same manner. Their results, with the exception of those of Eisenlohr and Sauer, added little however, since, as the latter pointed out they were not free from certain objections which will be taken up more in detail under the discussion of the properties of the lens. As early as 1858, Charcot ^^ gave the first description of photoph- thalmia and erythema produced by a small electric laboratory furnace (cf. page 635). In 1867 Czerny ^^ made the first experimental observations on the effect of direct sunlight on the retina. Even through heat filters he found he could produce in the rabbit's eye marked destruction of retinal elements in 10 to 15 seconds exposure with concentrated sun rays. Deutschman ^^ later (1882) showed that these changes could be noted in 1 second in the same manner. Herzog ^"^^ confirmed these results in 1898 and reduced the exposure time to § sees. In some of these cases there wei'e cataractous lens changes but no outer eye disturbance. TyndaH ^^° in 1876, made an important contribution to the subject in establishing the fact that ultra\iolet rays are absorbed by the atmosphere, since the ultra violet content of the solar spectrum was found to be much greater on high mountains than on low plains. In 1806 *^^ Wenzel and later von Beer in 1817 ^^^ pointed out the disposition to cataract among glass blowers. The introduction of the arc light in 1879 and 1880 mark an epoch in the history of this subject, for at least two reasons; first, because a means was afforded for far more accurate and extensive study of ultra violet light which the electric arc so abundantly supplies; secondly because immediately after the use of the electric arc for lighting and high temperature furnaces became general, cases of what was later designated as ophthalmia electrica, began to make their appear- ance. Martin 2^7 and Nodier 2^8 in 1881 reported on these cases, and although they correctly described the symptom complex, their expla- nation of the phenomenon as a sympathetic reflex from the injured retina was soon proved to be incorrect. In 1882 Leber 221 after studying the cases of cataract formation after exposure to lightning came to the conclusion that such cataracts were produced electro-chemically. 762 WALKER. In 1883 de Chardonnet ^^ used the ultra violet rays from the arc light to study the absorption power of various parts of the eye, and emphasized for the first time the very important role of the lens as a determinant for the limit of visibility of the short waves of the spectrum, by virtue of having the highest absorption power of all the eye parts. He argued therefore that aphakic eyes should have a greater range of spectral vision than normal eyes. Accordingly he examined two patients who had clear eye media after the extraction of cataractous lenses. He asked these patients to observe an arc light through a quartz glass plate, thinly silvered so as to cut out all rays except those between 343 /x/u and 301 mm hues. Normal eyes could not make out on looking through this glass whether the arc light was burning or not, but the aphakic patients could even detect motion of the light. In 1886 Meyhofer ^*^ found 11.6% of glass blowers under 40 years of age had cataracts, and these were mostly left sided where heat exposure was greatest. In 1888 Hess ^^^ allowed an electric spark to impinge on the supra- orbital region of a rabbit and produced equatorial cataracts. There was more or less central distruction of lens capsule and vacuolization of anterior lens fibres with peripheral increase of mitotic figures in the capsule. This result was later confirmed by Chiribuchi, but was shown to be an electrochemical rather than abiotic effect. It remained for Widmark *^^ in 1889 to experiment with the effect of ultra violet light on the eyes of the laboratory animals. He repro- duced the stages of electric ophthalmia in the rabbits' eyes and con- sidered the reaction to be of the nature of an inflammatory erythema. He first demonstrated the protective power of the lens by interposing a fresh rabbit's lens in the path of the ultra violet rays to which the rabbit's eye was exposed. The rabbit's eye in this case failed to give the characteristic reaction. Hirschberg ^''^ in 1898 first suggested the possible influence of intense sunlight in producing early senile cataracts in India and in the country, though Schulek^^^ had in 1895 from the statistics of Grosz, already noted that the senile cataract was more common in people working on the hot plains than in city dwellers. Schwitzer^^^ was the first to incriminate the ultra violet portion of the sunlight as an etiological factor in these cases. Hirschberg ^^* in 1901 first noted that the senile cataract almost always began in the lower c^uadrant of the lens. Perhaps stimulated by the possibilities of protection to the eyes sug- gested by Widmark's experiments, Schuleks^*^ in 1900, examined a EFFECTS OF RADIANT ENERGY ON THE EYE. 763 great number of substances for protective properties. Unfortunately the substances he found to have the necessary transparency and ab- sorptive power were certain liquid solutions. His results for the absorptive power of the lens and the other parts of the eye were the same as those obtained by de Chardonnet. The study of protective glasses was later taken up by Staerkle^*^. and Vogt^^^, and quite recently with more success by Hallauer ^^**'^^^, Schanz and Stock- hausen^^^, and Birch-Hirschfeld^^ Widmark ^^^ in 1901 and 1902 continued to develop the experimental method of studying the problem on the eyes of laboratory animals. He introduced some very ingenious experimental arrangements and was the first to show with the aid of the microscope that ultra violet rays can produce definite pathological lesions of the corneal and lens epithelium as well as of the conjunctiva and the skin of the lids and face. Further he believed he had ascertained that the injuries to the lens can be readily aggravated until cataract formation is the result. He found that heavy glass (18 mm. thick) when interposed prevented these changes. Solutions of quinine sulphate were equally protective. He was the first to note the similarity of ophthalmia electrica and the outer eye trouble in snow-blinding. It was not till 1907 that these results received some confirmation by Hess ^^^. A number of observers had looked for lens changes both before and afterwards without success, or with variable results. Thus Ogneff 2^^ in 1896 using an arc light of .5000 to 8000 c. p. noticed no lens trouble but much outer eye trouble as Widmark *^^ had shown in 1889. Herzog ^^^ in 1898 repeated this work with a heat filter and a common glass optical system on young rabbits and considered that any small effect such as he found was due to heat transformation. Birch-Hirschfeld ^° in 1904-5 found no lens changes with 4-10 min. exposure to a 4 amp. Finsen light. HerteP^° in 1903 using the magnesium spark likewise noted no lens change; nor did Strebel using 5 min. exposures to a 6 amp. iron arc light. Hertel's -^^^ work in 1903 was based on his idea that the pathogenic range of ultraviolet rays should be determined on the living cell rather than on the photographic plate since there was a difference in the action of these rays on chemical and living substance. He therefore enclosed certain bacilli, in tiny quartz glass boxes which could be inserted into the aqueous or vitreous chambers of the eye. Exposing the eye, then, to the ultra violet rays from a magnesium electrode spark, he found that waves of 280 mm would not pass through the lens and kill the organisms (B. Coli) behind it even after 60 min. 764 WALKER. exposure, while bacteria in the anterior chamber were killed in 25 to 30 min. but not in the control with common glass interposed (cf. page 749) . In none of the eyes, 26 in all, was lens trouble noted. In 1904 definite pathological changes believed to be due to ultra violet light were noted by Birch-Hirschfeld ^^ in the finer structures of the retina. In 1906 and 1907 Vogt ^^^ and Hallauer ^^^ began the careful study of transparent and colorless protection-glasses, and the latter produced by a secret process the so called " Hallauer glass." In 1907 Schanz and Stockhausen ^°^ also invented and patented a new glass, which they called " Euphos glas." In 1908 Birch-Hirschfeld ^^ studied in five cases visual field changes produced by uviol lamps, showing sector and ring formed scotomata for red and green to be the predominant varieties, but later (1912) he objected ^^ to the ringscotoma found by Jess '^^^ in the same year. Voege^^^ in 1908 asserted that daylight might be taken as the ideal light, especially "cloud light." He compared the spectra of various high power lights protected with milk and opal glass coverings, with the spectra of cloud light and found them to compare favorably, and therefore concluded that these lights so protected are not to be con- sidered dangerous when properly used. In 1909 Schanz and Stockhausen^^® vigorously opposed this attitude and their view was supported in the same year, by the appearance of the statistical study of Handmann ^^^ showing that the senile cataract begins in the region of the lens most exposed to the short wave length light of the sky, that is in the lower half. In this year Birch-Hirschfeld^^, Schanz and Stockhausen^^®, and Hallauer ^^^ (on human lens only), by spectrophotographic method, measured with the greatest care, the absorptive power of various kinds of glass, the cornea, vitreous humor and lens of various animals and of the human eye. They also made careful measurements of the spectral range of a great variety of light sources with and without co^'ering of common glass, milk glass and opal glass. In 1910 Schanz and Stockhausen ^^^ made a very careful study of the fluorescence of the human lens by a hitherto unused method, and also examined more carefully than before the spectrum of the glass blowers' furnace and the conditions under which the glass blowers were forced to work. They contended that the glass makers' cataract is due the longer of the ultra violet rays with perhaps the assistance of the short- est visible rays. Also in 1910 Hertel and Henker ^'^^ accepting Voege's idea that the EFFECTS OF RADIANT ENERGY ON THE EYE. 765 ideal light is cloud light or skylight, used the most accurate instruments available in the laboratory of C. Zeiss, in Jena, to measure the percent- age absorptive power at different points in the spectrum, of various glasses. These glasses were all found to be inferior to opal and milk glass for the purpose of enclosing strong arc lights to produce a spec- trum most nearly approaching in quality and quantity, the spectrum obtained from sky light. Recently (1912) Martin ^^^ has verified some of the results of Widmark, Hess, and Romer, while Carl Behr ^"^ has reported some very interesting functional disturbances of light adaption power of the eyes in patients working by artificial light. These results will later be taken up more in detail. Having thus rapidly traced the important steps in the progressive development of the knowledge of ultra violet light in relation to the eye, the mass of findings may doubtless be rendered much more available by considering them separately and in more detail with reference to the A'arious parts of the eye. The Outer Eye. — Photophthalmia, Vernal Catarrh. Probably ophthalmologists have experienced less difficulty in reach- ing definite conclusions, concerning the condition called ophthalmia electrica or photophthalmia (Parsons ^^^), than with any of the other effects of ultra violet light. As to the symptom complex little has been added since the first report in 1858 (cf . page 635) of Charcot ®^ and the later observations of Martin ^^^ and Nodier ^^^ in 1881, shortly after the general introduction of the electric arc for lighting and furnaces, (in 1879 and 1880). The workmen most exposed to these arcs, particularly the furnace arc, began to complain of symptoms that we now know to be due to photophthalmia (see page 634). In a week the eyes were practically normal. The affection of the sur- rounding skin known as dermatitis electrica was not unlike that of sunburning of severe grade except in its origin. As in sunburn, a tanning was notable after the inflammation had subsided, for several weeks. Although retinal changes were seldom noted with the oph- thalmoscope, functional disturbances were observed such as temporary blindness or scotomata, floating spots of red, yellow or blue or occasion- ally erythropsia, or red vision. Therefore the very earl\- writers were inclined to believe the outer eye trouble followed sympathetically from the retinal injury. These retinal disturbances also led Terrier ^^^ 766 WALKER. in 1888 to divide the large number of reported cases into two classes; a mild group without retinal disturbances and of good prognosis, and a severe group with retinal disturbances and of bad prognosis. However, after the classical experiments of ^Yidmark these theories and classifications were no longer found to be useful. Widmark*^^ in 1889 exposed the rabbit's eye to various parts of the arc light spec- trum. He found that when a 1200 c. p. arc light was used for 10 min. on the rabbit's eye without screening out any ultra violet rays all the typical symptoms of electric ophthalmia appeared after a latent period of 6 hours. By varying the time of exposure any degree of injury could be produced from a mild erythema to ulceration of the conjunctiva and cornea. But if a common glass plate 0.5 to 1.0 cm. thick was interposed the rabbit was entirely protected. Thus he established for the first time that rays below 300 /XM in length were chiefly responsible for the outer eye trouble. This particular point was confirmed by Ogneff ^^^, Hess ^'^'^ , Kiribuchi ^°^ and subsequently by practically all observers. Further ^Yidma^k concluded that ultra red and the visible rays are entirely without effect, outside of common heating effects. Widmark made another important contribution to the knowledge of this subject when he drew attention to the striking resemblance of ophthalmia electrica and the disturbance found on the outer eye in cases of snow-blinding. He showed that they both had the same latent period and were ushered in with the same syndrome of symp- toms. Further that erythropsia, temporary blindness, or blind spots occurred in both. The fact that ultra violet light is stronger on high mountains and snow covered surfaces as had been shown by Tyndall and subsequently verified by Hehnholtz, Langley, Cornu and Mascart, was a further argument emphasized by Widmark in support of his contention that ultra violet rays are responsible for both ophthalmia electrica and snow blinding. Birch-Hirschfeld, Hertel, Best and others had up to 1907 found a thick plate of common glass to be sufficient protection from electric ophthalmia as Widmark had pointed out. But in 1907 Stockhausen after one half hour working with arc lights received a severe ophthal- mia electrica through glass protection. Schanz & Stockhausen ^^° therefore repeated Widmark's experiment and found common glass to be inefficient protection for long intense exposure. They were able to produce the characteristic symptom in a rabbit's eye through 18 mm. of common glass after 4 hours' exposure to a 15 amp. arc light. Thus stimulated they studied the manufacture of glass carefully, EFFECTS OF RADIANT ENERGY ON THE EYE. 767 and finally produced the yellowish colored "Euphos glas" which they recommend as very satisfactory, not only for protection glasses, but also for use in making arc light coverings or mantles. However Birch-Hirschfeld ^* in 1907 considered the above exposure so intense as to afford no criterion for cases as they usually occur. From his results he asserted that one need not be afraid to use the ordinary smoked, uviol, flint, or even common glass, in the great majority of cases. He exposed a rabbit's eye for 1 hour as close as 10 cm. to a uviol lamp (mercury vapor tube) protected only by a 2 mm. thickness of common glass. After the 6 hour latent period no symp- toms whatever developed although the control rabbit's eyes were badly damaged. Further he exposed his own eyes to a 3000 c. p. quartz mercury arc lamp at 1 meter distance using smoked glass goggles as a protection. Although the surrounding skin of his face was burned, no symptoms of ophthalmia electrica developed. Birch- Hirschfeld ^* also proved that by daily exposure of the rabbit's eye to ultra violet rays a chronic inflammation of the outer eye could be produced which was very similar in appearance, both grossly and his- tologically, with vernal catarrh, originally considered by Schiele^^^ in 1899 to be a result of exposure to the light rays of the sun. In this investigation Birch-Hirschfeld exposed the rabbit's eyes for 10 min. every day for 180 days at a distance of 10 cm. from the " Uviol lampe" of Schott. The eye lids of the rabbits were everted during the expo- sure. After passing through the usual acute ophthalmia electrica a chronic inflammation was established similar to vernal catarrh, but no trouble in the lens or retina was noted. Birch-Hirschfeld considered rays shorter than 330 /x^t to be prima- rily responsible for the outer eye disturbance in this case, still rays from 330 ju/x to 400 /x^ or more could not be excluded as etiological factors. However he agreed with Axenfeld and Ruprecht^, 1907, that these factors could not entirely explain vernal catarrh. Vogt*°^ in 1912 thought that exacerbations at least in the disease depended on thermic influences. The Cornea: — Absorption, Injuries. That the cornea might suffer severe injury in bad cases of oph- thalmia electjica was early noted by Terrier ^^^ in his report of 1888. In these cases a dull haziness of the cornea with perhaps a phylectenu- lar condition or bleb formation was first noted. This condition could either go on to ulcer formation l)y infection or to panuus formation 768 WALKER. by vascularization. The ulcer formation as is usually the case, often lead to, or was accompanied by, iritis. Corneal disturbance in snow blinding has been occasionally reported. Hildige^^^ in 1861 and Reich ^^* in 1880 saw small ulcers. Widmark *^^ in 1889 studied the progress of the earliest changes on the cornea due to ultra violet rays. With the aid of the microscope he found first in the corneal epithelium a swelling and necrosis of the nuclei leading to necrosis of epithelial cells, and small areas of desqua- mation followed sometimes by ulcerative conditions and usually by opacities. These findings were at once verified by Ogneff^^^ and Bresse ^^ and later by many others. Hertel ^P in 1903, repeating this experiment and with rays of 309 m/x to 280 ^/x from the magnesium spark, was able to produce the same corneal injuries, as well as to kill, or at least demoralize bacilli enclosed in quartz containers and placed in the anterior chamber. This could not be done when common glass was interposed in the control experiment. That this fact may be taken as evidence that rays of 280 mju were able to penetrate the cornea does not follow, was pointed out two or three years later by Birch- Hirschfeld, Schanz and Stockhausen, who considered that rays of greater length than 280 mm in sufficient amount to kill organisms could not be excluded (cf. page). Widmark made no attempt to determine the absorptive power of the cornea by spectrophotographic methods. Schanz and Stock- hausen ^^^ were among the first to attempt accurate measurements in this way on the human as well as on the animal cornea. They found that all rays below 300 mm are absorbed by the cornea. Hess, Birch-Hirschfeld and Herzog verified this measurement and again later Birch-Hirschfeld^^ attempting still greater accuracy, with the same method, placed the absorptive limit at 306 mm- Parsons ^^® in England in the same way found rays above 295 mm able to penetrate the cornea. Still later, in 1909 Schanz and Stockhausen reconsidered the limit of 300 ij.li for the absorptive power of the cornea placing it at 320 ijl/m for all practical purposes, since the spectrum was so weakened be- tween 320 fjLfjL and 300 hijl as to be without action on the lens. 300 nn was however still considered the point of complete absorption. Martin ^^^ in 1912 agreed with Parsons that the cornea offered no resistance to waves above 295 nix length but all beyond t^is limit were completely cut off. EFFECTS OF RADIANT ENERGY ON THE EYE. 7Q9 The Aqueous and Vitreous Humors. The humors of the eye seem to be the most silent regions as far as response to insult from ultra violet light sources is concerned. Since they have an absorptive power, never greater, and often less than that of the cornea, the latter apparently protects them from the action of the injurious rays. Bonders ^^ who made the first attempt to measure the absorptive power of vitreous humor alone was not aware of the fact that the containing vessel must be made of thin quartz glass, so that his re- sults were of no value. After him Soret ^*^ in 1879 reported the first reliable results. He found the vitreous humor able to absorb rays of lengths less than 294.8 fxfj, in thicknesses of 1 cm. and still smaller values for thinner layers. The values of de Chardonet^^ in 1883 were still lower. He found the absorptive value to lie between the 310 MM and 304 mm lines. Birch-Hirschfeld ^* in 1909 found that the vitreous humor in 1 cm. layers has an absorptive power, practically constant for all animals, of ravs less than 300 fi/j. thus being the same as common glass. Schanz and Stockhausen^oa, Vogt^^^, Hess^^s, Ogneff ^59^ Birch-Hirschfeld 3^- and all recent observers have also confirmed this value for 1 cm. layers of vitreous humor. Parsons ^^^ for thinner layers, — j^ of an inch, — found absorption to begin at 280 fxfjL and become complete at 270 mm- Martin ^^^ in 1912 confirmed the later results and found no change in the absorptive power of eye media as long as 8 hours after death. The Iris. The iris and uveal tr.act have long been noted to suffer in severe exposures to short wave lengths. Martin ^^^ and Nodier^^^ in 1881 noted inflammation of the iris in severe cases, confirmed by Terrier ^^* in 1888. The very short exposure with light rays by Czerny ^^ in 1867, Deutschman ^^ in 1898 and Herzog ^^^ in 1898 gave no notable iris changes beyond slight hyperaemia. Hess ^^^ in 1888 by use of the electric spark impinging in the supraorbital region, and Kiribuchi ^°^ in 1900 with the Ley den jar spark were both able to produce marked uveitis. Gardner ^^'^ in 1871, Berlin ^2 1888, and Ewald ^^^ 1891, have reported hyperemic and swollen iris in snow blinding. Widmark*^^ in 1889 noted microscopically in cases of 2-4 hours- 770 WALKER. exposure a marked swelling and hyperemia of the ciliary body, and in later experiments in the same way noted small hemori'hages in the iris. Gross examination showed myosis and discoloration of the iris. Ogneff^^^, Terrier ^^^ and Weiss ^^ confirmed these findings. Birch-Hirschfeld ^^ in 1904 with the Finsen 3.5 to 4.5 amp. arc light for 5-10 minutes noted iritis and cyclitis in 6-12 hours with fibrinous exudate into the anterior and posterior chambers. Further experi- ments in 1908 with the Schott lamp 10 min. exposures at 10 cm. daily for 180 days showed practically no effect on the iris though a chronic eonjunctival inflammation was produced. Martin ^^^ in 1912 noted, in rabbits exposed 1| to 2 hours at a dis- tance of 1 in. from a Kromayer mercury vapor lamp hyperemia and myosis of the iris but no exudates. Further by the hemolytic method of Romer he found that the iris showed evidence of damage with inten- sities above 1 hour exposure at 4 in. distance. Whether the injury to the iris produced when the light is of sufficient strength is a direct result of light rays or a secondary effect of the corneal and outer eye injury was not made clear. The Lens: — Absorption; Fluorescence, as a Determinant OF Visibility of ultra violet Rays; Injuries; Cataracts. The reports of different observers upon the absorption power of the crystalline lens have varied considerably. Birch-Hirschfeld ^"^ in 1909 accounted for the long list of previous variations in the following manner. Aside from the personal equation or individual variation in observation, there is a considerable variation in the absorptive power of lenses of different animals of the same species as well as of different species. His results may be tabulated thus: Range of Average Variation in Animal absorption Different Animals Swine 330 juju 15 /z/x Calf 328 MM 12 mm Ox 385 fj.fj. 30 MM increasing with age. To a less extent thickness plays a part, though not so very great, since 5 mm. of rabbit lens has about the same absorptive power as 10 mm. of ox lens, for waves less than 390 nn. The formula for the EFFECTS OF RADIANT ENERGY ON THE EYE. 771 effect of thickness or intensity before and after transit shows a varia- tion possibility of small degree thus, — Ji = Jo-e'^'^ where Ji = intensity after transit Jo = intensity before transit d = thickness and k = coefficient constant. The human lens he found to vary considerably, as will be shown later, with such factors as age, consistency and color. As has been stated the absorptive power of the lens and other eye media for ultra violet rays, and the limit of visibility of the spectrum in the ultra violet region are two problems whose investigation has been carried forward in the same stages since it was obvious from the start that the determination of one would throw much light on the other. Brucke^^ really opened the subject in 1845 when he speculated as to the range of the visible spectrum and the reason for the invisibility of the ultra violet rays. In the manner described he found the ox lens to absorb rays below 370 mm- Bonders ^^, using the method of fluorescing screens of quinine sulphate, discovered by Stokes, at- tempted to measure the absorptive power of the lens but the glass containers vitiated his results. Stokes ^^'^ by direct observation of the solar spectrum through a quartz prism, thought he could see as low as the 372 nfx, 358 ij-h, and 335 mm lines and perhaps further. In the same way Helmholtz ^^^ with a quartz optical system could see a few lines in the 372 mm and 318 mm region, although his eyes were very myopic. He also observed the ultra violet rays directly through holes in the fluorescing region of a screen of quinine sulphate upon which the spectrum was thrown. By similar methods Listing ^^^ placed the limit of visibility at the 372 iJLiJL Hne and Sekulic ^^* at the 358 hijl line. Mascart 2*° however, using high intensity of ultra violet illumination, considered lines as low as 313 niJt, to be visible. Soret ^*^ by photographic methods found the vitreous humor of the ox in 1 cm. thicknesses to have the same absorptive power as the cornea of 294.8 fxix. He found that the lens of the ox absorbs rays shorter than 383 iJ.fx, and the entire eye has the same limit. Nevertheless he maintained that the human eye could see rays as short as 294.8 iJ.fj.. Eisenlohr ^°^ threw doubt on these results when he pointed out that fluorescence alters the ultra violet rays so that the observation by means of or in the presence of fluorescent light, is not accurate. He found even on white paper screens that fluorescence rendered rays i t Z WALKER. visible as low as 354 \xix in length while observation of the same light through the spectroscope showed no rays visible less than 395.6 \x\i in length. Later Sauer ^°^ using metal electrodes came to the same con- clusion. With the use of the arc light de Chardonnet ^^ photographed the rays able to pass through the human lens and found absorption began at the "H" line (397 /x/x) increased to the "L" line (381 mm) ^.nd became total at the "M" line (372 \x\i). The absorptive power of the cornea lay between the 304 /x^i and 299 mx lines and for the vitreous humor between 304 ix\x and 310 \x\x. To de Chardonnet belongs the credit of properly emphasizing the significance of the higher absorption of the lens in determining the lower limit of visibility of the spectrum. He concluded that patients having clear media after cataract operations could see more of the spectrum than normal eyes. By thinly silvering a quartz glass plate he was able to prevent all except rays below the 343 ix^x line from passing through so that when normal eyes attempted to observe an arc light through the plate it was entirely invisible. He found two aphakic patients however who could tell when the arc light was turned on and off or when it was moved while lighted. Widmark*^^ next took up this question in a very thorough manner. He used Hasselberg's modification of Rowland's spectroscope using a grating with a radius of 1.6 meters. The light source was an arc light with iron poles. The discontinuous spectrum obtained in this way gives sharp well known lines to examine and is free from aberrant light. Widmark examined eight aphakic patients ranging from 59 to 68 years of age. Seven could see lower in the scale than he, himself, could. Four of these were examined roughly by observing the spectrum thrown on a screen. One of these could see no better than himself so that the above more accurate method was used on the second four giving the following results for the ultra violet limit, — 313 /x/x, 313 /xju, 342 MM and 344.5 mm- In order to test the normal range at various ages for comparison he examined 59 individuals ranging from 11 yrs. to 74 yrs. of age. The results are here tabulated. No. of Individuals Age in years Lower limit of vision in ,«/' 10 11-20 378-395 Average = 386 14 20-30 371-395 = 382.5 6 30-40 372-393 = 388.9 13 40-50 380-394.5 = 388.7 3 50-60 378-402 = 391.7 10 62-74 379-410.8 = 401.8 EFFECTS OF RADIANT ENERGY ON THE EYE. 773 After 55 years of age, only one out of 12 individuals could see rays below 395 fxjj., that is, in the ultra violet region. The only medium which changes at that age is the lens, so that it is still more definitely proven that the lens establishes the lower limit of spectral vision in the human eye. Birch-Hirschfeld ^^ compared the visual threshold or distinguishing power, for various intensities of the same wave length, in the ultra violet region, of the aphakic eye and the eyes of individuals ranging from 14 to 70 years of age. He was thus determining not the ultra violet limit of vision, but the intensity at which a definite wave length which both groups of eyes could see, would become visible. He found that the threshold of the lensless eye exceeded that of the normal eye not inconsiderably, except in the case of a red blind physician who had developed a power of distinguishing small intensity changes that almost equalled that for the lensless eye. As the wave lengths of the light used were diminished, he found that the lensless eye gradually gained more advantage until near 381 /x/x and below it showed far greater sensitiveness than the normal eye. His results thus agreed with those of Widmark, though the same accuracy was not attempted, — (the screen method was used) since his point was only to show the relatively greater sensitiveness of the lensless eye, in the short wave than in the long wave length ultra violet regions of the spectrum. The appearance of the ultra violet spectrum has been variously described. Helmholtz characterized it as deep indigo blue under weak illumination to silvery blue under stronger illumination. Seku- lic ^^* and Sauer ^°^ called it silver gray in color. "Widmark's aphakic patients described the first part, 340 nn to 370 fifx, as blue or violet and below that all described it as a weak light gray. More recently Schanz and Stockhausen ^^^, Birch-Hirschfeld^^, and others agree on lavender-gray as the best descriptive term. The question as to how this sensation is produced, whether by direct stimulation of the retina or by the intermediation of the phenomenon of fluorescence of the lens or retina, has been variously answered. Soret^*^ favored the latter view, but the work of Widmark*^^ and others shows that when the lens, which fluoresces more than any other part of the eye, is removed, still greater range of vision in the ultra violet region is obtained. Further as pointed out by Widmark*^® and Mascart^*° the ultra violet region appears in sharp lines and bundles, not as a blur of light impossible to focus such as would come from the fluorescing lens. Nor could the fluorescence of the retina make these rays visible and still give a sharp image. Thus according 774 WALKER. to Tigerstedt^^^ it is generally accepted that the retina is sensitive to such ultra violet rays, as are able to penetrate the eye media and be focussed upon it. The same does not hold for ultra red rays although Helmholtz thought it did from the work of Briiche and Knoblauch ^^ in 1846 who found that heat from the Argand burner did not penetrate the eye appreciably. Cima '^^, 1852, also Janssen ^^^, and Franz ^^^ in 1862, found only about 9% of the heat from a Locatelli lamp was trans- mitted. This was confirmed by Klug ^°* in 1878 with gas and sunlight. Tyndall ^''^, 1865, with a 650 c. p. arc lamp found that about \ of the dark heat rays were transmitted through the vitreous of the ox. Engelman ^^° in 1882 using the bacterium photometricum, — which always migrates to the infra red region when exposed to the spectrum, as an indicator, found the same phenomena when water glass vitreous lens or cornea was interposed. Hertel ^^^ in 1911 showed the lower limit of subjective and of objective stimulation of the retina were about the same, lying between 820 nn and 840 mm- Vogt *°^ however, in 1912 showed conclusively that a great amount of the ultra red light reaching the retina is not visible, in fact as much as 80% or more. Further on normal human eyes he found that 3% of the heat reached the retina and less than 1% passed on into the orbit. 20% to 25% passed through cornea or sclerotic. The aqueous absorbed 20-30 % of the heat transmitted by the cornea. The cornea iris and lens together transmit 6% of the heat falling on the cornea. The lens absorbs 30% of the heat transmitted by the cornea and iris. Vitreous absorbs nearly 60% of the heat falling on its anterior surface. The upper lid transmits 6%. Fluorescence has long been a subject of much interest and study. A. von Graefe knew that fluorescence of the lens was due to ultra violet light and Helmholtz ^^° after an extended study of the fluores- cence of the lens, quinine sulphate solutions and other fluorescing bodies, concluded that fluorescence in general is due to the appearance of rays of various length, and is therefore really mixed or white light. Fluorescence was then the result of a transformation of ultra violet rays to rays of greater wave length. He considered the rays between 400 ixjx and 300 ixfx, to be chiefly the ones transformed. Widmark noted an apparent decrease of fluorescing power of the lens as the age of the individual increased and the absorptive power increased. Schanz and Stockhausen ^^^ in 1909 took up the question of fluo- rescence in connection with their study of the properties of " Euphos- glas" which they found in certain grades to absorb rays below 400 /x/x EFFECTS OF RADIANT ENERGY ON THE EYE. 775 and still cut down the visible spectrum but little — 5% for a thickness of 10 mm. They found that fluorescence of the rabbit's lens was not diminished by 18 mm. of plate glass, therefore rays less than 300 nfx, could not be held responsible for the fluorescence. Nor did flint glass absorbing to 350 /xix decrease the phenomenon but when Euphos glas was interposed the fluorescence was stopped. Therefore they limited the range of fluorescing rays to 350 mm to 400 nix. If they allowed the light to traverse both a blue uviol glass and a Euphos glas before striking the rabbit's eye there was no fluorescence of the lens. Now when the Euphos glas only was removed after adaptation had taken place, much lid-spasm and blinking of the rabbit's eyes took place as the fluorescence began. They laid great stress on this occurrence as an indication of the painful and injurious effect of this group of rays, 350 MM to 400 niJL, on the retina. From a study of the spectrum in this region photographed through the whole eye media and from the appearance of the fluorescence itself, they came at this time to the conclusion that fluorescence was due not as Helmholtz said to trans- formation of short wave ultra violet light to longer waves of different length in the visible field, but due to the appearance of a new spectral color lavender-gray of definite wave length. Not only did they con- sider fiuorescence, as did Widmark, to decrease with age but also with length of time of exposure, since the fresh lens from a gliomic eye of a child diminished notably in fluorescing power after exposure of a few hours. This decrease of fluorescing power they attributed to some breaking down or change in chemical composition which they supposed, without analytical proof, to be involved in the production of fluorescence. Birch-Hirschfeld ^^ at once objected to these conclusions, consider- ing that nothing had been offered as real proof against the Helmholtz theory of fluorescence. He maintained that nothing had been shown as to the nature of fluorescence and even questioned whether the range of rays responsible could be established in the manner described. The fact that lid-spasm was elicited as described he could not consider as evidence of a retinal injury, since it is well known that such reflexes are easily produced by harmless light on the dark adapted eye. Aside, then, from injury to the retina by the range of rays mentioned, a mere change of intensity would account for the lid-spasm since Helmholtz has shown that fluorescent light is many times as intense, physiologi- cally, as the light producing it. As to the diminution of fluorescence with age Birch-Hirschfeld considered the question still open, since he found the fluorescing power of the lens of an individual 70 years old 776 WALKER. undiminished. Further he regarded fluorescence of the nature of a catalytic chemical reaction if it was to be considered chemical at all, emphasizing the fact that no chemical basis for this phenomenon had ever been established. Against the proposition that fluorescence decreased with the length of time of exposure, he cited an experiment in which he exposed and fluoresced the rabbit's lens continuously for five or six hours with no noticeable change in intensity of fluo- rescence whatever. He objected to drawing any conclusions from the lens of a gliomic eye since the presence of a pathological condition might readily alter the properties of the lens although it was appar- ently normal. Later in 1909 Schanz and Stockhausen^-^^ Ijy means of Wood's light filter which absorbs the rays below 37.5 fxfj. were able to make further investigations on the range of fluorescing rays in the same manner previously employed. This filter inhibited the fluorescence very little so that the range most effective in producing fluorescence was placed at 375 fxn to 400 /x/i. The fresh clear lenses from eyes removed by tumor or absolute glaucoma were vised. The absorption power of the cornea and lens in these cases were studied by photo- spectrographic methods with a quartz glass optical system in the usual manner. From a study of these photographs they agreed with Birch-Hirschfeld that the cornea had a greater absorption power than glass and considered that the effectual absorption amounted to 320 ytiju as mentioned under cornea. That absorption in the lens was increased by age was further confirmed. Schanz and Stockliausen^-^^ further investigated the phenomena of fluorescence with the result that they retracted their previous idea that fluorescence represents a separate spectral color lavender-gray, and returned to the theory of Helmholtz that it is made up of rays of various length as is ordinary mixed or white light. They used the crossed prism or crossed spectrum method of Newton to analyze the fluorescent light in the following manner. With a quartz glass optical system, the light from an arc light was concentrated on a prism, the resulting spectrum was rendered linear by focussing it on a screen with a cylindrical lens, axis parallel to the length of the spectrum. Now when thin layers of fresh human lens were laid on the screen in various regions of the spectrum fluorescence was noted to begin in the blue region, become more intense in the violet region, and was strong- est of all in the ultra violet region between 370 nfx and 400 juju. Below 370 nij. fluorescence diminished slowly. The maximum point of fluo- rescence was at 385 nfx. This fluorescent light was further analyzed EFFECTS OF RADIANT ENERGY ON THE EYE. 777 by a second prism placed at right angles to the first prism and parallel with the first spectrum. The second spectrum seen through the second prism showed at once that the fluorescent light was made up of rays of different length. Of these the greater portion were green waves, a less portion of blue waves, while a considerable amount of red was also present. In the light of these findings arid after a further study a few months later the following tabulation of the eflfect of light waA^es on the eye was prepared : Visible light red to green blue to violet I II 7G0mm-490mm 490mm-400mm Invisible ultra violet light III 400/x/i-375MAt IV 375mm-320mm V These rays proceed un- changed to the retina and are visible. A small part in- creasing with age is by the lens absorbed, and is concerned in its fluores- cence. Another part fluoresces the retina and the rest is seen by the retina as blue and violet. A part fluoresces the lens. A part fluoresces the re- tina. A part pro- ceeds unchanged to the light sensitive retina. Whether the appearance of the lavender-gray is due to a direct stimulation or by intermediation of fluorescence is un- kno\\Ti. Take little part in fluorescing the lens. Are intensely ab- sorbed by the lens, reaching the retina only in young eyes much weakened. Do not penetrate through the cornea but pro- duce outer eye trouble. Hallauer ^^^ in 1909 spectro photographically measured the absorp- ti\e power of over 100 fresh human lenses and found it to depend mostly on individual differences of thickness, color and consistency. For young lenses, while most of the rays were absorbed at about 400 nfx, a certain number of more or less weakened rays between 330 nfx and 315 ^i/x were able to pass throvigh. The effect of severe or chronic illness in these cases was to increase the amount of all to pass through in the latter region. Also in advanced age, where the absorption lay usually between 400 /x/z and 420 fxii, the effect of severe reducing diseases was to reduce the absorption power to about 375 ju^. Martin ^^^ in 1912 found the absorption power of lens suspended in normal salt solution to begin at 400 /x^t and become complete at 350 fxfx. 778 WALKER. Injuries to the Lens. Czerny ^^ in 1867 in his blinding experiments with sun's rays noted turbidity in the lens cortex but no change in the lens capsule. Deutsch- man ^^ repeating these experiments in 1882 got the same results. Herzog^^^ in 1903 obtained similar results with the carbon arc, glass lenses and heat filters. These lens changes were undoubtedly due to the thermic action of visible rays. Widmark*^'' in his experiments on the outer eye, 1889-1892, with the 1200 c. p. arc light noted lens changes microscopically. These he did not find when ultra violet rays were screened out by a quinine sulphate solution therefore he concluded they were the etiological factor. Ogneff ^^^ in 1896 repeated this experiment with a 6000 to 8000 c. p. arc lamp at a distance of 50 cm. to 1 meter for 15 to 20 min. but found no lens trouble though all the conjunctival corneal and iris troubles were present. Widmark^^^ repeated his work again in 1901 using a 4000 c. p. zinc arc in much the same way with the same results. Two rabbits A & B were exposed to the same arc light at the same time. In the case A the light traversed two glass lenses separated by 5 to 6 cm. of a 10% quinine sulphate solution. The distance from arc to lenses was 13.6 cm. and light was concentrated on the dilated rabbit's eye 6 cm. beyond the lenses. Ultra violet rays and heat rays were cut out by this method and no lens changes were found. In case B the conditions were the same except that the lenses were of quartz, separated by water. In this case in addition to the usual disturbance in the outer eye and iris, the lens capsule in the pupillary area showed at first intense staining of the nuclei, mitosis, cell proliferation and destruc- tion. There was swelling of lens fibre bundles with partial destruction. Also transudate between the cortex and capsule. In 1904 with a 3| to 4| amp. Finsen light, Birch-Hirschfeld ^^ could not get lens changes after 5 to 10 min. exposures. Hertel ^^^ in 26 rabbits used in the previously mentioned experiment, and also Strebel ^^^ failed to get lens changes, though the usual outer eye and iris changes were produced as previously observed. Hess ^^* in 1907 used a 3| amp. uviol mercury vapor lamp with a 65 cm. tube. He exposed 1 to 16 hours at a distance of 10 to 30 cm. The animals used were rabbits, guinea-pigs and frogs. He found outspoken lens changes as described by ^Yidmark. These lens changes appeared about 48 hours after a 6 to 12 hour exposure. Surrounding EFFECTS OF RADIANT ENEEGY ON THE EYE. 779 the central damaged area but under cover of the iris was a ring or "wall" of deeply stained cells crowded together perhaps by the swollen central cells, first damaged. These changes appeared in the pupillary region and would show regeneration as indicated by numerous mitotic figures in the course of 2 to 4 days, if no further or stronger exposure was made. He found that interposition of glass plates cutting out rays below 313 fxiJ, or even 280 ^t/x prevented lens trouble. He agreed with Widmark in thinking the glass blower's cataract due to ultra violet light. Birch-Hirschfeld '^^, however, took up this point in 1909. He ex- posed a rabbit's eye for 5 minutes at a time on three successive days to the light from a 5 amp. arc light which traversed first a " Euphos- glas" and was then concentrated with a 20 diopter common glass lens. No heat filter was used. On the 4th day he found on microscopical examination the same lens changes recorded by Widmark and Hess. Therefore he concluded that rays in the neighborhood of 400 ^t/x must be responsible, and that probably some of the shorter blue and violet rays were effective as well as the longer ultra violet rays. Further he argued that this same group of rays was in all probability responsible for the production of the glass blowers cataract and possibly also for the production of the senile cataract, though this latter he regarded as far from proven (cf. page 677). Without experimental evidence Wenzel *^^ in 1806, von Beer in 1817 and Plenk in 1877 ^* pointed out the disposition to cataract among glass blowers, and Meyhofer ^^^ in 1886, found the percentage to be 11.6 in glass blowers under 40 years old. These cataracts commonly began on the left side which was most severely exposed to the heat. Robinson ^^^ and Stein ^*^ later confirmed these findings. Schanz and Stockhausen ^^° in 1910, measured the quantity and quality of the radiations from the glass blowers furnace, and the temperatures to which his head was subjected at different stages of his work. By accurate spectro-photographic measurements of the light at the distance at which the glass blower worked, they found the spectrum to be especially strong in the region from 400 MM to 350 /x/i, shading down to 320 h/jl below which there were no rays. At once they considered they had the explanation why these people had lens trouble without anterior eye trouble. The worker's head was exposed to a temperature of 110 degrees C. in taking the glass from the oven and to 45 degrees C. during the process of blowing. This temperature, while it might be a factor, is not so great as that to which many iron and blast furnace workers are exposed without receiving any eye 780 WALKER. injuries. Temperature, therefore, they thought not nearly so much to blame as the ultra violet rays from 400 nn to 350 /x/^ which fluoresce the lens most strongly. The absence of rays below 320 nn accounted for the absence of outer eye trouble, in answer to the question raised by Birch-Hirschfeld during the previous year. In comparing the cataracts of glass blowers and those produced artificially by the arc light, they noted as has Widmark*^^, Hess ^^^, Cramer ^^ and others, as well as Stein ^*^ later, that they both begin in the pupillary region, but that while the artificially produced cataracts begin usually on the anterior pole, the glass maker's cataract starts usually on the poste- rior pole. For the posterior polar variety no better explanation could be offered at that time than that of Cramer ^^, who believed them to be the result of the greater concentration of chemical rays at that point due to the refractive power of the eye media anterior to that point. Other theories were concentration of the chamber fluids (Leber ^^^) and increased venous stasis (Peters ^^^). However Snell^*^ in Eng- land fovmd cataract no more common among glass blowers than among other laliorers, and Robinson ^^^ foimd the percentage increasing aljove normal only among the finishers working with very heavy metal glasses, after long service. In 1909, Handmann^'''^ submitted an extensive statistical study of senile cataracts. Hirschberg confirmed by Schulek ^^° had first sug- gested the intense simlight as a factor, in the country and India, and had noted the early appearance in the lower quadrant. Hand man was able to prove that the senile cataract particularly of India for the most part 81% (previously given by Greene ^^^ as 95%), begins in the lower quadrant of the lens. This region of the lens was found to be more deeply yellow colored, and therefore had a higher absorp- tive power for ultra violet. The ultra violet content of light coming from the sky at such an angle as to strike this quadrant is far greater than that reflected into the eye from the broken surfaces below, thus supporting the original idea of Schwitzer ^^^ that abiotic rays were an etiological factor. Presumptive evidence at least was furnished to explain the senile cataract. Schanz and Stockhausen ^^^ at once ac- cepted these findings as giving the key to the etiology of a large group of senile cataracts. But Birch-Hirschfeld^^ pointed out that dwellers in mountainous and snowy regions of high ultra violet content were not prone to cataract and considered that the shorter visible rays could not be excluded here as in other cases, and Hess^^* from a mathematical standpoint considered that rays not obstructed by the lids and eyelashes could not reach the lower quadrant of the lens in sufficient quantity to account for the condition. EFFECTS OF RADIANT ENERGY ON THE EYE. 781 Martin ^^^ in 1912, with a single high intensity exposure, — one half to two hours at a distance of one inch from a Kroraayer water-cooled mercury vapor lamp, found the lens capsule changes as described by Hess. Because the interposition of a benzol cell prevented these changes they were ascribed to ultra violet light. "With repeated exposures of moderate intensity without lid retractors — at a distance of 4 inches to 3 ft. at intervals of 1 to 2 weeks over a period of 2| to 12 months exposure times varying from 1 to 3 hours, the following changes were noted. In one rabbit of this latter series exposed every 10 days at 4 in. distance for 1 hour over a period of 3 to 12 months lens changes somewhat as described by Hess were found, but differed in that the "wall" was wider, the central cells were uninjured, and proliferation was 2 or 3 cells thick. There was present in all of this series slight corneal opacities. Others of the series, less severely exposed, had no corneal opacities or lens trouble, while those more severely exposed, — 3 hours every 2 weeks at 4 inches for 3 to 11 months, — showed dense corneal opacities which had undergone vas- cularization but the lenses were clear and capsule normal, supposedly protected by the dense corneal opacities. Sharply to be distinguished from these are cataracts experimentally produced by exposure to light, are those resulting from the actual transit of the electric current through or near the eye. Leber ^^^ in 1882 first explained, the long noted tendency to cataract in people struck by lightning, as due to the electro-chemical reaction. This was experimentally demonstrated by Hess ^^^ in 1888 who made an electric spark to impinge in the supraorbital region of a rabbit. He noted a central destruction of lens epithelium and vacuolization of lens fibres w^ith a marked secondary peripheral mitosis and prolifera- tion of the lens epithelium resulting in the formation of equatorial cataracts. Likewise Kiribuchi ^°^ in 1900 using the Ley den condenser spark produced the same results. Clinically many cases of cataract after lightning stroke or short circuit have been reported as due to ultra violet rays but really must be included in this last group. Birch- Hirschfeld^^ in 1909 found, in a study of all such cases reported up to that time, injuries such as burns and cicatricial formations or impair- ment of nutrition or nerve suppl}', which would readily account for cataract formation. These cataracts are further distinguished by the fact that clinically and experimentally they do not appear centrally as do those in purely blinding experiments but peripherally. Again on an experimental basis the length of exposure time is seldom if ever long enough. No cases of lens trouble from lightning without bodily 782 WALKER. injury have been reported. Essentially the same is true in blinding due to short-circuit arcs, though sometimes it is difficult to distinguish mechanical from light effects, as in cases reported by Grimsdale and James ^^^ and by Posey ^^^ in 1911. The Retina: Injuries, Scotomata, Erythropsia. That blindness may result from direct observation of the sun of eclipses, has been known without doubt for ages. Indeed Galileo ^^ is known to have injured his eyes by observation of the sun with his telescope. Galen ^^^ cites cases of blinding, with more or less subse- quent return of vision, in observers of eclipses of the sun. He also noted that central scotomata or blind spots often resulted in the same way. Reid^^* in 1761 and Soemmering'*^ in 1791 according to Hess ^^* probably gave the first accurate description of the subjec- tive phenomena of sun blinding. Less frequently the same ocular disturbances have also long been noted in seamen long exposed to strong reflection of the sun's rays from water surfaces. Likewise travelers over desert or glary plains are not uncommonly afflicted with these visual disturbances. Czerny ^^ as early as 1867 showed that a lesion of the retina of the rabbit visible with the ophthalmoscope could be produced by the sun's rays. Coccius, Ruete, in 1853, and Jaeger in 1854, had already described the ophthalmoscopic changes in the human eye. Czerny threw, by use of a concave mirror, and glass lens system concentrated sun's rays which had traveled a 20 cm. water heat filtering tube, into the eye of the rabbit for 10 to 15 sec. The region of the retinal image on exposure was found to be whitened and seared. A section under the microscope showed what he described as a coagulation of the albuminous substances of the retinal elements. On the 17th of May, 1882, there was an eclipse of the sun which in a few days brought four cases of sun blinding into Leber's clinic in Gottingen. Deutschmann ^^ at once reported them with a repetition of Czerny's experiment in which he fully confirmed Czerny's findings. One of the four cases used a smoke glass to observe through and another a blue glass, but each received severe injury nevertheless. On ophthalmoscopic examination all four cases showed a character- istic appearance of the macular region varying somewhat in degree corresponding with the degree of injury. In his experiments Deutsch- mann arranged a convex lens to transmit the sun's rays reflected from. EFFECTS OF KADIANT ENERGY ON THE EYE. 783 a concave mirror. The distance separating the two was equal to the sum of their focal lengths so that parallel light was thrown into the atropinized eye of the rabbit. Even after a second's exposure a silvery white spot round to oval in shape, covered the retinal image region. It was surrounded by a brownish ring. Longer exposure enlarged the silvery central spot and the surrovmding rings became paler and took on a silvery sheen. Microscopic section showed droplets and clumping of the coagulated albumins of the retinal cells. Surrounding and below these areas were exudative and then hypere- mic areas. The choroidal pigment was increased and showed a tendency to wander. Indeed there was so much similarity to early stages of choroiditis disseminata that he was the first to consider the possibility that heat and light rays may be an etiological factor in the latter disease. However Aubaret ^^°°, Hess ^^°^, and Garten ^^°^ have since attempted, without success, to prove this proposition. To determine the influence of heat Deutschmann passed the rays through a tube of clear running water 20 cm. long. The changes could be produced but it always took a few minutes longer. Therefore he concluded that both heat and light are active as etiological factors. In none of these cases with short exposure times were outer eye troubles noted. In 1892 Widmark*^^ repeated Czerny's and Deutschmann's work with a 1200 c. p. arc light. The exposure time was 2-12 hours, usually 4 hours. With the 10% quinine sulphate filter to remove ultra violet light he found much less retinal trouble than with the quartz glass system. Also less effect through yellow bichromate filters than through blue, so that he concluded that blue violet and ultra violet rays were most effective. Herzog ^'^^ in 1903 reported that he had repeated the work of Czerny and Deutschmann in 1898 and found that the circumscribed retinal lesions could be effected in | second. A similar but more diffuse retinal change could be produced in | to 2 hours with a 15 amp. arc light whose rays were concentrated on the rabbit's eye after traveling a 28 cm. tube of alum water. Further similarity in action was noted in the production in two or three old rabbits of an opaque cataractous condition of the lens visible to the eye as Czerny and Deutschmann had noted. When the cornea was continuously irrigated with normal salt solution to avoid overheating, the result was a cloudy swelling of the epithelium which was less transparent than the simple des- quamation that resulted without the excess of moisture. All these changes, including the cataract formation already described he ascribed to light transformation to heat, and not to ultra violet light. 784 WALKER. However, Aubaret ^ in 1900 was inclined to disregard heating effects in sunblinding since he found a thermometer held in sunlight concen- trated by 40 D diaphragmed lens, only registered 1° to 2° increase in temperature. But Birch-Hirschfeld ^^ showed that 50° paraffin in thin layers on black paper was melted in a few seconds when exposed to sunlight as the retina would be in sunblinding but wdien white paper was used under the same conditions several minutes were re- quired to melt the paraffin, thus demonstrating the effect of black retinal pigment in absorbing heat. The phenomenon of light adaptation and the effect of various wave lengths of light on the retina has been studied by a number of observers, Mann 2^^, Kuhne^^^, Pergens^^^, Nagel^^^, van Genderen-Stort ^^'', and more recently by Hess^^^, Birch-Hirschfeld^^ and others. Van Genderen-Stort ^^'^ in 1887 showed that pigment wandering in the choroid and retina was least active in yellow light and this finding called attention to the value of yellow glasses as protection for the eyes. He further amplified the knowledge of the effect of light on the finer structures of the retina of the dark adapted eye. When such an eye is exposed to light he showed that the pigment cells just out- side the rods and cones send down processes, apparently between the rods, which carry with them much pigment and thus tend to isolate optically the rods from each other. At the same time the cones tend to escape this isolation somewhat by moving in the same direction tow^ards the light. Another change found to take place was the bleaching of the visual purple found in the outer part of the rods in the dark adapted eye. The effect of ultra violet ray and light fatigue on the finer struc- tures of the retina has been studied by Widmark*^^, Czerny^^, DeutschmannS^ Ogneff^sa, Bach i*, Stebepsa, Kiribuchi 203, and Terrien ^^*. Except for the last three, who found a very slight gang- lion cell chromatolysis, their results were negative, which Birch- Hirschfeld ^'^ later thought was due to insufficient light intensities having been used. In snow blinding nothing more than hyperemia of the retina and optic disc have been reported (Reich ^^, 1880). Birch-Hirschfeld^'^ in 1904 made some experiments on rabbits in which he claimed to have produced definite pathological changes by exposing the eyes to ultra violet light. In his first series of experiments he separated out the ultra violet light from a 15 ampere carbon arc light by means of a quartz prism, and obtained retinal changes only in eyes from which he had removed the lenses. In his second series he used the direct light from a 3 to 4.5 ampere water cooled Finsen EFFECTS OF RADIANT ENERGY ON THE EYE. 785 iron arc and obtained well marked changes, notably chroraatolysis and vacuolization of the ganglion cells, in lens containing eyes as well as in two aphakic eyes. These experiments are described and dis- cussed in detail on page 687. From the foregoing experiments Birch-Hirschfeld concluded that the lens did not afford complete protection to the retina from the specific action of ultra violet light. Best ^^, however, took the view that no danger to the retina was to be feared from ultra violet light since he found he was able to look at the sun directly without deleteri- ous effects for ten seconds through a blue uviol glass wdiich transmits freely the ultra violet waves from the sun, but absorbs all visible waves longer than about 470 nn in length. Birch-Hirschfeld^^ in 1908 reported five cases of visual disturbance among workers in mercury \'apor illimiination. He concluded that after long occupation with unprotected eyes in mercury vapor arc- light a disturbance of retinal function may result with or without an electric ophthalmia or a similar conjunctival reaction. These in- juries took the form of pericentral scotomata for red and green. By the aid of a Priestly Smith scotometer these scotomata were mapped out. The affected region had a sector or ring-formed shape at a distance of 15 degrees to 20 degrees from the fixation point. Red usually appeared yellowish and green as a gray or even white. The central color vision was only in two cases injured in the sense of red- green blindness, though floating spots often temporarily obscured the fixation point. The color scotomata disappeared in the course of a few weeks if protecting glasses were worn or if work in ultra violet light was abandoned. The solar eclipse seen in Europe on the 17th of April, 1912, afforded excellent opportunities for observation of ophthalmoscopic and func- tional retinal changes due to sun blinding. In the series of 50 reported by Birch-Hirschfeld*^, four cases showed normal eye grounds, 19 others had increased f oveal reflex, with frequently a concentric irregu- lar reddish-brown area, which in eleven cases cleared up at the end of a week. In 16 other cases there was noted irregular pigmentation of the macula and small gray puntiform spots or globules which remained unaltered for months. In 31 cases an absolute central and in 19 cases an absolute paracentral scotoma was found which afterwards became relative scotomas. The rest had relative central scotomas to begin with. These were mostly eccentric downward extending 1° to lO'* The majority of cases with the milder injuries regained almost normal visual acuity in the course of a few weeks. In small numbers of single 786 WALKEK. cases numerous previous observers ^^*, Vinsonneau, Stocke, Pergens V. Pflugk, Arlt, Lesearret, Menacho, Villard, found similar changes. Lamhofer in 1912 found a chorioretinal exudate with pigmentation and at the same time Best reported a case with decreased peripheral field and adaptation. Erythropsia has also been noted by Birch- Hirschfeld and by Braunschweig. In 26 out of 36 cases of sun blinding in the eclipse of 1912, Jess ^°° working in Hess's clinic found a ring-scotoma 20° to 40° from the fixation point. In this ring formed area, white appeared gray, and colors were not seen in a few cases, in others red was called yellowish, green was called gray, and blue was called yellow. In the course of a week the damage gradually decreased until only a small semicircular area of scotoma was left below. Hess ^^* states that this finding has been verified by Peppmueller, Pergens, and Hoppe. Speleer found enlargement of the blind-spot ring-scotoma and concentric contraction of the field after sun blinding. Birsch-Hirschfeld*^, however, has taken exception to the ring-scotoma of Jess, finding that it was a normal phenomenon and not specially related to eclipse-blindness. The large number of functional impairments due to lightning and short-circuit flashes cannot be referred definitely to any given range of wave lengths. These derangements have been reported in great numbers and variations. There may be permanent blindness or temporary blindness in one or both eyes. Central and peripheral scotomata are more common and while usually temporary they may be permanent. Hancock ^^^ in 1907 reported a case of ring-scotoma. Disturbances of color vision are very common but temporary. These include erythropsia, red blindness, red green and blue green blindness and scotomata. Ophthalmoscopic examination may show nothing even in severe cases or there mav be punctate spots in the macula as in sun blinding, Uhthoff "5, Haab ", Terrien ^^s. Birch-Hirschfeld ^i considers rays from 350 fx^ to 400 nfj. most active in producing these disturbances with probably even more assistance from blue and violet rays because of the greater intensity. While usually there is outer eye trouble in these cases a few have reported functional disturbances without photophthalmia. Nelson Dering^^* Le Roux and Renaud ^^5 Maclean ^^^ and Purtscher^^^ In snow blinding functional disturbances have been noted to take the form of temporary amblyopia (Widmark *^^ citing Enald) night- blindness (Widmark*^^) and day lilindness. In 1907 Best & Haenel ^^ found after snowblinding a central scotoma for red and green extend- ing about 10° from the fixation point. This disturbance disappeared in the course of six weeks. EFFECTS OF RADIANT ENERGY ON THE EYE. 787 Very recently Behr ^^ has found marked reduction in the Hght adaptation power in four patients complaining of visual disturbances after continued work by arc light and strong incandescent lights. They noted that after working by these lights and then moving to a darker part of the room that they could hardly see to work at all, or on coming out into sunlight everything appeared gray or dark. In twilight or after a rest in darkness the vision improved slowly only to receive another setback on further exposure to the artificial light. These patients were tested on the Piper instrument for measuring the dark adaptation power. Normally as Piper has shown there is a slow increase of sensitiveness during the first 61 to 10 min. in the dark and then a sudden marked increase in 30 to 35 min. followed by a further slow increase until after 45 min. the maximum light per- ception sensitiveness is reached. Behr found in none of these cases was there much increase of sensitiveness after 10 min. and at the end of 45 min. was still | to | of normal. These cases were advised to work less by strong artificial lights and Euphos glas was prescribed. Mild light such as oil lamplight was advised where possible. Three of the patients were presumably cured because they made no further complaint, while the fourth whose case was followed, regained in a short time a light adaptation that was better than normal. Erythropsia or Red Vision. Hildige ^^^ in 1861 noted erythropsia in snow blinding. Mayer- hausen'^*^ and Steiner^^° in 1882 reported cases after blinding by lightning and short circuiting and by the sun's rays both directly and indirectly as from water surfaces or from snow. Cases which occurred after cataract operations were further reported by Dimmer ^^ and Putscher ^^^ in 1883. Widmark in his correlation of electric ophthal- mia and snow blinding noted that erythropsia occurred in both but did not investigate the subject further. Fuchs ^^^ in his monograph of 1896 emphasized the role of ultra violet light in all these cases espe- cially in cases of snow blinding, aphakia, and electric ophthalmia. Using the fact that established by Kuhne ^^^ and Konig two years before, that the visual purple is most rapidly bleached by rays below 500 jjLfx. Fuchs advanced the theory that the regenerating visual purple accounted for the red vision as an entopic phenomenon. Schulek^^^ produced Erythropsia by observing spectral ultra violet 788 WALKER. light, also Birch-Hirschfeld did the same while working with the Schott-light. However, Vogt ^^^ in 1908 dilated his own pupil and after producing erythropsia by observing a sunlit snow field found no decrease in the red vision when he went into a room illuminated only by a light whose red rays had all been screened out by an Erioviridin filter. Vogt 3" in 1908 and also Wydler ^^2 -^^ 1912 considered ery- thropsia as the red phase of the after picture of the intense white surface. Best^^ in 1909 agreed with Vogt's view and considered erythropsia due to visible rays since he could produce it by looking at a snow surface through a yellow glass cutting out rays below 400 nij. but not with a blue uviol glass. However Birch-Hirschfeld^^ could not consider that ultra violet rays were relieved of responsibility entirely since the wide pupil alone of the aphakic eye could hardly account for erythropsia, as Vogt held, by admitting a greater quantity of light. He therefore concluded that invisible as well as visible rays were active in the etiology of erythropsia. Rivers ^^^ in 1901, advanced a theory as to the red color based on an observation made by Briiche in 1851 that the eye under normal conditions is more or less completely adapted for red. Rivers attri- butes the color in erythropsia to the blood in the anterior retinal layers. Schoute ^^* objected to this theory on the basis that Pur- kinje's experiment of the eutopic vision of retinal vessels shows that light is absorbed by the blood and they give dark shadows. Protective Glasses. In 1900 Schulek ^^° first studied the means of protecting the eyes against ultra violet rays and found that solutions of Triphenylme- thane in xylol and Nitrobenzol in Alcohol had the highest absorptive power of the transparent media examined. These liquids absorbed practically all rays below 396 ju/z. He suggested that these solutions be enclosed in flat oval shaped glass chambers made to fit the eyes and to protect them from injuries due to ultra violet radiations. Stearkle^*^, Vogt^^^ and Hallauer ^^^ studied the absorptive prop- erties of blue uviol, yellowish and smoky gray glasses. The last named worker produced a glass mixture by a secret process the so called "Hallauerglas." After a similar study, Fieuzal ^^^ produced in like manner Fieuzelglass, which however, was not greatly used on account of its color. Also a yellow-green glass patented under the EFFECTS OF RADIANT ENERGY ON THE EYE. 789 name of Enixanthosglas was offered as well as a variety of the modi- fications by glass manufacturers. Here again the color was not satis- factory. Schanz and Stockhausen^°^, after finding that electric ophthalmia could be produced through 18 mm. of common glass as has been men- tioned, began to study glass manufacture in the hope of producing a colorless glass of high ultra violet absorption power for general use. In 1909 they produced and patented a glass of higher absorptive power than hard flint and called it "Euphosglas." At first it was made in grades 1, 2, 3, and 4, but recently other grades have been added. It has a light yellowish green tinge and fluoresces in ultra violet light. Birch-Hirschfeld^^ in 1909 studied photometrically the absorption power of these and other glasses with considerable accuracy. In the same year Vogt ^°* compared a new and very hard flint glass produced by Schott with his absorptive solutions and was surprised to find that it had about the same efficiency, beginning to absorb at 405 jUju and giving practically complete absorption below 396 mm- In 1909 Hallauer^^^ measured photometrically the absorptive power of the various protective glasses available at that time. The thick- ness varied from 1 to 3 mm. and exposure time of one minute. The average results follow. Common glass absorbs to 295 Blue glass " 300 Lead glass " 305 Smokey gray " 325 "Gonin" glass " 330 Schott's heavy flint " 340 Fieuzal yellow glass " 375 Enixanthos glass " 380 Euphos gray glass " " 390 Euphos green glass " " 390 Hallauer glass ^64 " " 420 As glasses became available to cut out various wave lengths the question arose as to what spectral range constituted the best illumina- tion. Voege's^^^ answer to this question in 1908 raised a considerable controversy. He maintained that the light from the clouds or clear sky had been for ages the normal illumination for the eyes, but never- theless contained a considerable amount of ultra violet light as low as 790 WALKER. 300 /JL^JL in wave length. He examined the spectra of various high power are lights with opal and milk glass shades or coverings. These spectra compared very closely with that of cloud light. Hertel and Henker ^^^ in support of this view carried out a very accurate set of measurements, in the laboratory of C. Zeiss, Jena, of the cloud and skylight spectrum, of variously covered or shaded high power arc lights and of the percentage of penetrability and absorp- tion power of various glasses for wave lengths in different parts of the spectrum. The spectrum of the clouds or clear sky was found to contain no rays below 300 nfx and very few below 310 ju/x. Having what they considered an ideal light the next question was in what manner should the artificial lights be compared with it. In producing injurious effects on animals they noted that unprotected lights of very great intensity had been used under conditions never found in modern lighting systems therefore previous observers had studied entirely atypical conditions and not the conditions to which human eyes are really exposed. Widmark had used a 1200-4000 c. p. arc light without covering at a distance of 25 cm. from the animal's eye for length of time ranging from 2-4 hours and longer. Likewise Birch-Hirschfeld had used similar arc lights with and without disper- sion through a prism usually at a distance of 10 cm. to 20 cm. Hess had used a Schott uviol lamp 65 cm. long of 3-3| amp. The animals were at a distance of 10 cm. to 20 cm. and the time of exposure 1-16 hours. These data represented the average experimental conditions necessary to produce injuries but not at all the conditions to which mankind is at present exposed. Therefore they considered a far better criterion for practical pur- poses would be accurately to photograph the spectrum of arc lights at the minimum distance of actual service in lighting, and to find whether or not such globes or mantles can be placed around the light source as to render the spectrum, in cjuantity and quality, within the range of the cloud or skylight spectrum. Accordingly the lights were placed 50 cm. to 100 cm. distant from the spectroscope opening. The optical system of the spectroscope was made of quartz glass. The effect of indirect illumination and of half spherical or inverted bowl shaped shades for increasing horizontal illumination, was also measured by raising the light and its shade about 40 cm. above the axis of the spectroscope condensor. In all cases under these conditions by use of the usual milk glass and opal glass shades they had no difficulty in getting a spectrum comparing in quantity and quality very closely with the spectrum of EFFECTS OF RADIANT ENERGY ON THE EYE. 791 cloud light and entirely within the spectrum obtained from ordinary Welsbach gas mantle lights with ordinary clear glass shades. They used a variety of lights including flaming arc and mercury vapor arc lights. The absorptive power of various other glasses proposed for this purpose, by their inventors was determined by use of the most ac- curate available method. In the visible part of the spectrum the absorpti^•e power was determined by use of a polarizing spectrophoto- meter with crossed Nicol's prisms, while in the ultra violet region the same instrument was used with optical parts of quartz glass. The results are tabulated below in terms of percentage of penetrability of various wave lengths. The absorption power is obtained by simply subtracting these results from 100. In all these cases a thickness of glass of 1 mm. was used indicated by A', except in the case of the very dense Neutralglas which was measured in thicknesses of 0.1 mm. indicated by A*''^ Concerning protective glasses for the eyes, Hei'tel and Henker believe the same criterion should be followed, namely, that the best glass is the one which will reduce the spectrum of the particular light to which the eyes are exposed to the closest possible approximation to the spectrum of cloud and sky light. For observation of the strongest arc lights at close range, the condition under which certain workmen are placed, they consider the Neutralglas F 3815, of Schott to be the best. With this glass, in layers thinner than any other glass, one may observe directly the bare 20 amp. arc light at 50 cm. without injury since the spectrum is about the same as that of cloud light minus the ultra violet portion. The thickness of Hallauerglas No. 64 necessary to give the same results as Neutralglas of 0.8 mm., was 9.0 mm. ; and of Euphosglas No. 4 was 38 mm. Next to Neutral glas, for this purpose stood the smoky or Rauch- glas No. 276 and Sonnenglas No. 66 of the Fredener glass works. After these came Hallauerglas No. 66, while Hallauerglas No. 62 and No. 64 and Euphosglas Nos. 1, 2, 3 and 4, were not strong enough in absorptive powei*. Schanz and Stockhausen^^^ at once criticised the above work, objecting particularly to the premises on which the decisions were based, namely that skylight or cloud light can be taken as the ideal light. Against this view was cited particularly the work of Hand- mann showing that a very large group of cataracts begin in the quad- rant of the lens most exposed to the light of the sky during life. Since it was definitely known and admitted by all that certain injuries to 792 WAJiKER. O < ex. CD lO CO '^ o CO T-H GO c-i Ol (M 00 Ci 05 00 c£> oi 1— 1 t--. CO o Ol 1— ( CO fO -* Ol 00 OI CD -t< 00 — ( CO o t^ ^ CD d ^ CO CO 00 CO CO o C3 t^ CO ^ 03 ^ Ol a 00 ■* ^ ^ d lO CO lO 00 CO Ol ^^ Ol LO 00 t^ to Ol CO T(^ ^ LO t^ Ol CO C5 CO ,—1 ^ •* iO rH ■:c CI ^ Ol iC CO CD' Ol ■ d -^^ l^ LO ^ Ol 00 CO ^ Tt< Q 00 01 ■^ _l CD CO CO CJ -*1 ^ 00 00 !>. t>- 00 CO 1—1 CO s o Ol o -* OS M t^ CO >o Oi Ol 00 00 00 lO o^ GO CO lO «p CO CO CO lO o Oi m CO t^ • -^ 05 05 00 00 o lO CO C5 CO" lO '-' 00 CO "+I Oi o o (M r>- r- >o I^ C5 Oi 00 Ci o CO CO C5 CO 1.0 .— I ■* 00 to CO Ol 05 ' lO CO CO" "lO -* 00 00 00 00 00 00 CO C5 CO : =9 ^ -H o <1 ^ ^ ^ ^ ^ ^ ^ <11 01 CO -t Ol rfi CD CO lO CO CD CD ^H 6 2; CO ■X a. rj< CO 5, _C ■XI c8 j: _bjD J§ p£4 -^ M iJ o fil _2 C a ;-< 75 tc ■jn a> eS '^ 0) 1 0 3 _ce 15 - ^ 3 1) ^ » ffi o z EFFECTS OF RADIANT ENERGY ON THE EYE. 7^3 the eye could be produced by ultra violet light, any light containing these rays could not be considered ideal light. The positive findings, previously mentioned of pathological changes due to ultra violet light, violet and perhaps blue rays, such as the glass blower's cataract, erythropsia, blindness permanent and temporary, and scotomata as well as other functional retinal disturbances, they held to be against the assumption that light containing considerable amounts of ultra violet, violet and blue rays, can be considered ideal and harmless. Further minor objections were made to the exposure time used in photographing the different light sources and it was pointed out that their newer glasses "Euphosglas A" and "Euphosglas B," more suitable for technical purposes, had not been examined. 794 VERHOEFF, BELL, AND WALKER. BIBLIOGRAPHY. 1 Abbot, The Sun's Energy Spectrum and Temperature, Astro. Jour., 24, 3 and 197 (1906). 2 Abelsdorf, Zur Anatomie der Ganglionzellen der Retina,Arch. f. Augenh., 42, 188 (1901). 3 Alexander, Ein Beitrag zur Ophth. Elect., Deutsche med. Wochenschr., No. 47, p. 779 (1899). S^i Allen, F., Injuries to the eyes caused by intense Hght, Science, 15, 109 (1902). 3b Andrews, J. A., The electric light as an illuminator; the effect of strong light upon the eye, Tr. Am. Ophth. Soc, 4, 228 (1885-7). 4 AscHKiNASS, Spectrobolometrische Untersuchungen liber die Durch- lassigkeit der Augenmedien fiir rote und ultrarote Strahlen, Zeitschr. f. Psychol, u. Physiol, der Sinnesorg., 11, 44 (1896). 5 AscHiNASS, tjber das Absorptionspectrum des flixssigen Wassers und fiber die Durchlassigkeit der Augenmedia fur rothe und ultrarothe Strahlen, Annalen der Physik und Chemie, (5) 55 (1895). 6 AuBARET, Sur les scotomes par eclipse solaire, Arch. d'Ophth., 1907, p. 96. 7 AxMANN, Schutzgliiser aus optischen Glas, Deutsche med. Wochenschr., 1909, p. 152. 8 Axenfeld, Bemerkungen fiber Retinitis pigmentosa, besonders solche ohne Hemeralopie, Klin. Monatsbl. f. Augen., Beilageheft, 42, 54 (1909). 9 Axenfeld und Rupprecht, Die Pathologie des Fruhjahrskatarrhs, Klin. Monatsbl. f. Augen., 45, 144 (1907). 10 AxMANN, Schutzglaser gegen ultraviolette Strahlen, Zeitschr. f. physik. u. diatet. Therapie, 12 (1908-09). 11 AxMANN, Die Uviol-Quecksilberlampe, Med. KHnik., 1906, No. 4. 12 AxMANN, Uviolbehandlung und Augenkrankheiten, Deutsch. med. Wochen- schr., 1907, No. 5. 13 Baas, Das Gesichtfeld, Stuttgart, Ferd. Enke, 1896. 14 Bach, Die Nervenzellenstructur der Netzhaut in normalen und patho- logischen Zustande, Arch. f. Ophth., 41, 3, 62 (1895). 14a Batten, D. R., Injury to an eye by direct sunlight, Tr. Ophth. Soc. Unit. Kingd., 21, 70 (1900). 15 Beauvais, Accidents oculaires consecutifs a I'observation des echpses de soleil, Rec. d Opht., 1906, p. 85. 16 Beauvais, Simulation d'amblyopie double attribuee a la lumiere des eclairs, Ann. d hygiene ed. de med. leg., 1896, p. 434. 17 Behr, Beitrag zu der Frage nach den Veriinderungen und Schiidigungen des Auges durch Licht, Arch. f. Ophth., 82, 509 (1912). 18 Bell, On the opacity of certain glasses for the Ultraviolet, Proc. Am. Acad. Arts and Sci., 46, 671 (1911). 19 Bell, On the Ultraviolet Component in Artificial Light, Proc. Am. Acad. Arts and Sci., 48, 1 (1912). 20 Bell, Star Colors: A Studv in Physiological optics. The Astrophysical Journal, 31, 234 (1910). 21 Bellile, Note sur des lesions determinees par I'emploi de la telegraphic sans fil a bord des batiments, 1909, Archive de med. navale, 81, No. 3 (1909). 22 Berlin, Om Snoblindhet, Nord. med. Archiv., 20, 3 (1888). 23 Berthier, Les nouveaus modes d'eclairage electrique, Paris, 1908. EFFECTS OF RADIANT ENERGY ON THE EYE. 795 23a Bertholet, Evil effects of ultra-violet light, Gas World, 1913, p. 2. 24 Best, tjber die Schadigung des Auges durch ultraviolette Lichstrahlen, Klin. Monatsbl. f. Augen., 48, 342 (1910). 25 Best, Bedarf das Auge des Schiitzes gegen leuchtende oder gegen ultra- violette Strahlen?, Munch, med. Wochenschr., 1909, p. 1710. 26 Best, 1st Schutz der Auges vor uv. Licht notwendig?, Med. Klinik, 1910, No. 7 and No. 297 27 Best, tJber die praktische Tragweite der Schadigungen des Auges durch leuchtende und ultraviolette Strahlen, Klin. Monatsbl. f. Augen., 47, 520 (1909). 28 Best, tJber die Dunkeladaptation der Netzhaut, Arch. f. Ophth., 76, 146 (1910). 29 Best und Haenel, Rot-Griin-Blindheit nach Schneeblendung, Klin. Monatsbl. f. Augen., Beilageheft, 45 (1907). 30 Von Bezold, tJber die Fluoreszenz der Retina, Sitzungsbericht der math. phys. Klasse der K. Bayer. Akad. des Wissens. zu Miinch., 7 (1877). 31 Birch-Hirschfeld, Die Wirkung der strahlenden Energie auf das Auge, Allgem. Path, des Auges, Lubarsch-Ostertag, 1910. 32 Birch-Hirschfeld, Bedarf das gesunde Auge des Schutzes gegen Licht, besonders gesen solches von kurzer Wellenlange?, Med. Klinik., 30-33 (1910). 33 Birch-Hirschfeld, Der Einfluss der Helladaptation auf die Struktur der Nervenzellen der Netzhaut nach Untersuchung an der Taube, Arch. f. Ophth., 63, So (1906). 34 Birch-Hirschfeld, Die Veranderung im vorderen Abschnitte des Auges haufiger Bestrahlung mit kurzwelligen Lichtes, Arch. f. Ophth., 71, 573 (1909). 35 Birch-Hirschfeld, Weiteres Beitrag zur Kenntnis der Schadigung des Auges durch ultraviolet tes Licht, Zeit. f. Augenheilk., 20, 1 (1908). 36 Birch-Hirschfeld, Entgegnung auf der Aufsatz von Schanz und Stock- hausen, Klin. Monatsbl. f. Augen., 1909, p. 608. 37 Birch-Hirschfeld, Die Wirkung der uv. Strahlen auf des Auges, Arch. f. Oph., 58, 469 (1904). 38 Birch-Hirschfeld, Zur Beurteilung der Schadigung des Auges durch Kurzwelliges Licht, Zeitschr. f. Augen., 21, 385 (1909). 39 Birch-Hirschfeld, Beitrag zur Kentnis der Netzhautganglienzellen unter phj'siologischen und pathologischen Verhiiltnissen, Arch. f. Ophth., 50, 166 (1900). 40 Birch-Hirschfeld, Die Nervenzellen der Netzhaut unter physiologischen und pathologischen Verhiiltnissen, mit besonderer Beriicksichtigung der Blendung (Finsen, Rontgen, Radium), Miinch. med. Wochen- schr., 1904, p. 1192. 41 Birch-Hirschfeld, Zum Kapitel der Sonnenblendung des Auges, Zeitschr. f. Augen., 28, p. 324 (1912). 42 Birch-Hirschfeld und Inouye, Weitere Versuche tiber die Wirkung des ultravioletten Lichtes auf die Netzhaut, Arch. f. d. ges. Phvs., 136, 595 (1910). 43 Birch-Hirschfeld, tJber Sonnenblendung des Auges, Ber. ub. d. 38. Versamml. d. ophthalm. Gesellsch., Heidelberg, 1912, p. 241. 44 BiRTis, Beitrag zur Kataraktbildung nach elektrischem Schlag, Zeitschr. f. Augen., 16, 525 (1907). 45 Bjerrum, Wie entstent der Schmerz bei Lichtscheu?, Centralbl. f. Augenh., April, 1903 46 Black, Artificial illumination as a factor in the production of ocular dis- comfort, Trans, of the Ilium. Engin. Soc, 6, 166 (1911). 47 Blessig, Uber Wirkung farbigen Lichtes auf das Auge und ihre hygienische und therapeutische Verwertung: Prasidialvortrag im Verein St. 796 VERHOEFF, BELL, AND AVALKER. Petersburger Aerzte, Jan. 10, 1906, St. Petersburger Med. Wochen- schr., 1906, No. 36. 47a Blondel, a.. Daylight and artificial light, effects on the eves, Elek. Zeit., Sept. 24, 1908. 48 BoHM, Blendungsretinitis, etc., Klin. Monat. f. Augenh., 51, 471 (1913). 49 Boll, Zur Physiologie des Sehens und der Farbenempfindung, Ber. d. Akad. d. Wissensch., Berlin, 11 (I), 1877. 50 BoRDiER, Absorption des rayons ultra-violets par le verre; transparence actinique des verres servant de " Conserves ' ' ; consequences pra- tiques de cette etude pour la meilleure protection des yeux contre la reverberation solaire (Societe de med. de Lyon), Annal. d. Oculist., 144, 351 (1910). 51 BoRSCHKE, Untersuchen iiber die Herabsetzungen der Lescharde durch Blendung., Zeitschr. f. Physiol, d. Linnesorg, 34 and 35 (1904). 52 BRANDE.S, Cataracterelectrique aux deux yeux. Bull. Soc. beige d'opht., Nr. 32 (1911). 53 Braunschweig, "Ulaer Schadigungen des Auges infolge der Sonnenfinsternis vom 17, 4, 1912, Klin. Monatsbl. f. Augenh., 54, 758 (1912). 54 Bresse, Del'ophthalmie electrique et du coup du soleil electrique, These du Nancy, 1891. 55 Briere, Neuro retinitis causees par la reverberation des eclaire, Gaz. des. Hop. 1876, No. 41. 56 Brixa, Eine Verletzung des Auges durch Blitzeschlag, Klin. Monatsbl. f. Augenh., 1900, p. 759. 56* B^ocA, Action of extreme red and ultra-violet light on the eye, Science Abstracts, 1913, p. 266. 57 Brose, Two cases of electrical flashing followed by severe retinal irritation and intense eye pain, Ophth. Record, 1909, p. 84. 58 Brucke, tjber das Verhalten des optischen Medien des Auges gegen Licht und Warmestrahlen, Arch. f. Anat. u. Physiol. 1845, p. 252. 59 Bruche and Knoblauch, Miillers Arch., 1846, 379, Poggendorff' s Ann., 65, 598. 60 Buller, Ein Fall von Verletzung des Auges durch Blitzschlag., Arch. f. Augenh., 21, p. 390, 1890. 61 BuRCH, On Temporary Artificial Colour BUndness, Phil. Trans., 191, 1 (1899). 6l» BuRGE, The effect of radiant energy on the lens and the humors of the eye. Am. Jour. Physiol., 36, 21 (1914). eib BuRGE, Ultra-violet radiation and the eye. Trans. 111. Engr. Soc, 1915.^ 62 Cassien, Accidents produits sur I'appareil de la vision par I'electricite a bord des navires de guerre, These Bordeaux, 1895. 63 Candon, Le coup de soleil electrique, Rev. gen. d'ophtalm., 7, p. 63, 1888. 64 Cernovodeanu and Henri, Effect of ultra-violet light on micro-organisms, Compt. Rend. Acad. Sci., 150, 52 (1910). 65 Chance, A Case of Electric Light Burn of the Eye, with Transient Blind- ness, Ophth. Record, 1907, p. 203. 66 Charcot, Erytheme de la face et ophtalmie produits par Faction, lumiere electrique, Comptes rendus des seances et memoires de la Soc. de Biologic, 1858, p. 63. 67 deChardonnet, Vision des radiations ultra violettes, Comptes rend, de I'Acad. des Sc, 1883, p. 509. 68 deChardoxnet, Penetration des radiations actiniques dans I'oeil de I'homme, Seances de I'Acad. des Sc, 1896, p. 441. 69 Charpentier, La Lumiere et les Couleurs, p. 138. 70 Charez, Cataracta producida par el rays. Arch, de Ophth. hispan.-Amer., 1904, p. 549. EFFECTS OF RADIANT ENERGY ON THE EYE. 797 71 Chevalier, Sur les modifications de la lumiere chromatique a travers les verres colores emploj^es en oculistique, C. R. Acad, science, 76, 177, 1873. 72 Chiarini, Cambriamenti morfologica che si verificano nella retina dei vertebrati per azione della luce e oscurita, Bollet. della R. Acad, med. di Roma, 1904. 73 CiMA, Sur Potere degli umori, etc., Torino, 1852. 74 Cobb, Physiological Aspects of Illuminating Engineering. 1910. Lectures on Illuminating Engineering delivered at the Johns Hopkins Uni- versity, 2, p. 525-575. 75 Cobb, Physiological points bearing on glare, Trans. Illuminating Engineer- ing^ See, 6, 153 (1911). 76 CoBLENTZ, Infra-red transmission Spectra, Carnegie Inst. Pub., 97, 54 (1908). 77 CoHN, Lehrbuch der Hygiene des Auges, 1892. 78 Cords, tjber Sonnenblendung, Zeitschr. f. Augenh., 27 (1912). 79 CoRNU, Etude du spectre solaire, Comp. rend., 1878. 80 CouLLAND, Huit cas d'ophtalmie electrique. Arch. d'Ophtalm., 29, 26 (1909). 81 Cramer, Entstehung und klinische Besonderheiten des Glasblaserstars, Klin. Monatsbl. f. Augenh., 45, 47 (1907). 81a Crookes, The preparation of eye preserving glass for spectacles. Read before Royal Society, Nov. 13, 1913. Reprinted in Chemical News, Nov. 28, 1913. 82 Crzellitzer, tJber eine Massenverletzung durch elektrische Strahlen, Miinch. med. Wochenschr., 1906, p. 2321. 83 Cutler, Cas remarquable de maladie oculaire produite par une decharge electrique, Ann. d'Ocul., 1849, p. 34. 85 CzERNY, iJber Blendung der Netzhaut durch direkter Sonnenhcht, K. Akad. d. Wissensch., 6, Oct., 1867. 86 Defontaike, Note sur le coup de soleil electrique. Bull, et mem. de la Soc. de Chirug. de Paris, p. 799, 1887. 87 Denig, Uber parenchymatose Triibung der Hornhaut infolge von Blitz- schlag, Munch, med. Wochenschr., No. 34, 1894. 88 Desbrieres and Bargy, Un cas de cataracte due a un decharge electrique industrielle, Ann. de Ocul., 1905, p. 118. 89 Deutschmann, liber die Blendung der Netzhaut durch direktes Sonnen- hcht, Arch. f. Ophth., 28, 241 (1882). 90 DiTTLER, tJber die chemische Reaktion der isolierten Froschmetzhaut, Pflii- gers Arch., 1907, p. 44. 91 DiTTLER, tiber Zapfenkontraktion an der isolierten Netzhaut, Pfliigers Arch., 1907, p. 1. 92 Dimmer, Zur Erythropsie Aphakisches, Wien. Med. Wochenschr., 1883, No. 15. 93 Dimmer, Zur Aetiologie des Friihjahrskatarrhs der Konjunktiva, Wien. Klin. Wochenschr., No. 2, 1905. 94 Dobrowolsky, Die Ursachen der Erythropsie, Arch. f. Ophthalm., 33, 2, p. 213. 95 Dobrowolsky, Les verres dites fumes, doivent etre preferes aux verres bleus pour preserver la vue de I'eclat de la lumiere, Ann. d'Ocul., 1873, p. 156. 96 Dolganow-Klinowitsch, tJber gelbe und gelbgriine Glaser, Russk. Wratsch, 20, 1900. 97 DoNDERS, tJber das Verhalten unsichtbarkeit in den Medien des Auges. Arch. f. Anat. u. Phys., 1853, p. 459. 798 VERHOEFF, BELL, AND WALKER. «8 Dor, Electrocution; atrophic partielle du Nerf optique droit; cataracte de I'oeil gauche; incapacity de travail, La Clinique ophtalm., 1909. 99 Dor, Comment garnautir I'oeil des rayons ultra violets?. Rev. gen. d'oph- thalm., 1905. •99a Dow, Some physiological effects of light, Ilium. Engr., April, 1908, p. 86. 100 DowNAR, Alteration de I'oeil apres un coup de foudre, Gazeta Lekarska, No. 9, 1877. 101 Drake-Brockman, Is the use of electric light injurious to the human eyes, Ophthalmoscope, 1906, p. 488. 102 DuNBAR-RoY, The effect of intense flashes of electricitj^ upon the eve. The Amer. Jour, of Ophth., 1897, p. 353. 103 DuFOUR, Affection retinienne produite par une eclipse de soleil, Bull, de la Soc. de la Suisse Rom., 1879, p. 321. 104Eder, Vergleichung der chemischen Leuchtkraft und photographischen Wirksamkeit verschiedener Lichtquellen, Handb. d. Photogr. 1, 1894. 105 Eder u. Valenta, Absorptionspektren von farblosen und gefarbten Glasern mit Berlicksichtigung des Ultraviolets, Denkschr. d. Akad. d. Wissensch, 1894. 106 Eisenlohr, Die brechbarsten Lichtstrahlen im Beugungsspektrum und ihre Wellenlange, Ann. d. Physik u. Chemie, 98, 353 (1856). 107 Ellet, Cataract caused bv a discharge of industrial electricity, Ophthalm. Rec, 1906, p. 4. 108 Engelhardt, Ber. iiber die 10 Versaml. d. Oph. Gesell. zu Heidelberg., p. 134. 109 Enge-lmann, Uber Bewegungen der Zapfen und Pigmentzellen der Netz- haut unter dem Einfluss des Lichtes und des Xervensystems, Med. Kongr. Kopenhagen, 1884. 110 Engelmann, Onderzoekingen gedaan in het physiol. Lab. der Utrechtsche Hoogeschool, 7, 2, 291 (1882). 111 ExNER, tjber den Sehpurpur. Wien. klin. Wochenschr., Xo. 8, 1877. 112 Feilchenfeld, Zur Frage des phvsiologischen Blendungschmerzes, Cen- tralbl. f. prakt. Augenh., 1908, p. 97. 113 Feldmann, tJber den Friihjahrskatarrh, Zwangl. Abhandlg. aus dem Gebiete d. Augenh., 7, 6 (1908). 114 Fere, Note sur les accidents produits par la lumiere electrique, Compt. rend, hebdom. des seanc. de la Soc. de biologie, No. 21, p. 365, 1889. 115 Ferentinos, tJber Sehstorungen infolge des Beobachtung einer Sonnen- fuisternis, Ophth. Klin., 1906, p. 1. 116 Fick, tJber die Ursachen der Pigmentwanderung in der Netzhaut, Viertel- jahrsschr. d. Naturf. Ges. Ziirich, 35, 83, 1890. 117 Fick, tJber die Lichtwerkungen auf die Netzhaut des Frosches, Ophthal. Ges. Heidelberg, 1889. 118 Fick, Gesundheitsplege des Auges. In Graefe-Saemisch Handbuch der gesammten Augenheilk, 2 Auflage, VX Ivap., 19, p. 16. 119 FiEuzAL, Des verres colores en hygiene oculaire. Bull, de la clin. nat. oph. de I'hospice des Quinze-vingts, 1885, No. 3. 120 Frank, Klinische Beobachtungen iiber die Wirkung des blauvioletten Endes des Spektrums auf Erkrankungen der hornhaut, Westnik Ophthalm, 1906, No. 1. 121 Franz, Uber die Diathermasie des Medien des Auges, Ann. d. Physik u. Chemis, 115, 266 (1862). 122 Freeland, a group of symptoms caused by the electric light, Brit. med. Journ., 1893, p. 234. 123 Freysz, Katarakt durch elektrische Starkstrome, Deutschmanns Beitragc z. Augenheilk., 1909, p. 1. EFFECTS OF RADIANT ENERGY ON THE EYE. 799 124 Fromaget, L'Ophtalmie electrique, Journ. de med. de Bordeaux, April, 1908. 125 Frost, Two cases of night blindness from exposure to a bright light, Trans. Ophth. Soc. Unit. Kingd., 188.5, p. 12-3. 126 FuCHS, tjber Erythropsie, Arch. f. Ophth., 42, (4), 207 (1896). 127 FucHS, Untersuchungen fiber die Gefolge der Blichtung auftretenden galvanischen Vorgange in der Netzhaut und ihren zeitlichen Verlauf , Pfliigers Arch. f. d. ges. Phys. Vol. 84, p. 425. 128 FucHSund Welponer, Zur Farbie der Netzhaut, Wiener Med. Wochenschr. No. 10, p. 221, 1877. 129 Galen, De usu partium, lib. 10, cap. 3. 130 Galezowsky, Des Cataractes traumatique, Recueil d'opht., 1881, p. 705. 131 Galezowsky, Des Ophtalmes electriques, Rec d'Opht., 1902, p. 521. 131a Galli, Snow blindness, Ann. D'Oculistque, 145, 197 (1911). 132 Gariel, Valeur comparative des divers modes d'eclairage, Soc. franc. d'Ophtalm., 1910. 133 Garten, Die Veranderungen des Sehpurpurs durch Licht, Arch. f. Ophth., 1906, p. 112. 134 Garten, Die Veranderungen der Netzhaut durch Licht, Handb. d. ges. Augenheilk., 1907, 119-121 Lief erg., p. 19. 135 Garten, Uber die Veranderungen des Sehpurpurs durch Licht, Arch. f. Ophth., 1906, p. 1. 136 Gayet, Sur le pouvoir absorbant du Cristallin pour les rayons ultraviolets, Societe francaise d'Ophthalmologie, 2, 188 (1884). 137 Genderen-Stort, Uber Form Ortsanderung der Netzhautelemente unter Einfluss von Licht und Dunkel, Arch. f. Ophth., 1887, Bd. 33. 138 GiNESTONS, Ophtalmie electrique, Gaz. hebdom. des sciences med. de Bordeaux, No. 48, 1906. 139 GiNSBURG, Bin Fall von Blitzstar, Westnik ophth., 1906, p. 14. 140 GoTSCH, The time relations of the photoelectric changes produced in the eveball of the frog by means of colored lights, Jour, of Phys., 31, No. 1, 1904. 141 GoNiN, Lesions oculaires causees par la foudre, Ann. d'Ocul., 1904, p. 81. 142 GoNiN, tJber Sehneeblendheit, Ann. d'Ocul., Sept., 1908. 143 Gould, Is the electric light injurious to the eyes?, Med. News Philad., 1888, p.. 634. 144 Gradenigs, Tiber den Einfluss des Lichtes und der Warme auf die Retina des Frosches, Wien. Med. Ztg., No. 29, 30, 1855. 145 V. GR.4.EFE, Anaesthesia retinae mit Konzentrischer Verengerung des Gesichtfeldes, Klin. Monatsbl. f. Augenh., 1865, p. 261. 146 Greene, (Incipient cataract etc.). Jour. A. M. A. 51, 400 (1908). 147 Grimsdale, Electric light conjunctivitis, Med. Press and Circular, 1902. 148 Grimsdale and James, A Case of Cataract from an Electric Shock, Ophthalmoscope, 1911, p. 315. 149 GuTZMANN, Zwei Falle von Blitzkatarakt, Wiener Klin. Wochenschr., 1906, No. 16. 150 Haab, Traumatische Makula-Erkrankung bewirkt durch den elektrischen Strom, Klin. Monatsbl. f. Augenheilk., 1897, p. 213. 151 Haab, Die Blendung des Auges durch Sonnenlicht, Korrespondenzbl. f. Schweiz. Arzte, 1882, p. 383. 152 Hallauer, tiber die Absorption von Kurzwelligen Licht durch die mensch- liche Linse, Khn. Monatsbl. f. Augenh., 1909, p. 721. 153 Hallauer, Einige Gesichtspunkte fiir die Wahl des Brillen materials, Ber. d. 34, Vers. d. Ophth. Ges. Heidelberg, 1907, p. 334. 154 Hallauer, Spectrographic Studies concerning the Limits of absorption of our protective glasses. Arch, of Ophth. N. Y., 39, 34 (1909). 800 VERHOEFF, BELL, AND WALKER. 155 Hallauer, Spectrographische Untersuchungen liber die absorption- grenzen unserer Shutzglaser, Arch. f. Augenh., 64, 257 (1909). 156 Hancock, Ring Scotoma, Royal London Ophthal. Hospital Reports, Vol. 16, part 4 (1907). 157 Handmann, Uber den Beginn des Alterstars in den unteren Linsenhalfte, Klin. Monatsbl. f. Augenh., 47, 692 (1909). 157a Hansell, The ocular sequelae of exposure to the sun. A report of two cases, Med. News. Phila., 62, 486 (1893). 15 7b Hartridge, G., The electric light and its effect on the eyes, Brit. Med. Jour., 1, 382 (1892). 158 Heckel, Electric ophthalmia, Opht. Rec, 1905, p. 500. 159 Heckel, Report of a Case of electric ophthalmia, Amer. Jour, of Oph., 1906, p. 29-180. 160 Heiberg und Gronholm, Histologische Untersuchungen iiber die Ein- wirkung des Finsenlichtes etc., Arch. f. Ophth., 80, 47 (1911). 161 Helmholtz, Uber die Empfindlichkeit der menschlichen Netzhaut fiir die brechbarsten Strahlen des Sonnenlichtes, Pogg. Ann. 94, 205 (1855). 162 Helmholtz, Handb. der physiologischen Optik, 1896. 163 Mme. et M. Victor Henri, Variation du pouvoir abiotique des rayons ultraviolets avec leur longueur d'onde Compt. rend, de I'Acad. des Sc, 135, 315 (1902). 164 Henri, Helbrouner, et de Recklinghausen, Nouvelle Lampe a rayonement ultraviolet tres puissant a la sterilization de grande quantites d'eau, Compt. rend, de I'Acad. des Sc, 155, 852 (1912). 165 Hertel, Einiges liber die Bedeutung des Pigments fiir die physiologische Wirkung der Lichtstrahlen, Zeitschr. f. AUg. Physiol., 6, 1 (1906). 166 Hertel, tjber die physiologische Wirkung von Strahlen verschiedener Wellenlange, Zeitschr. f. allg. Physiol., 6, 95 (1908). 167 Hertel, Experimentelles iiber ultraviolettes Licht, Ber. d. 31, Vers. d. Ophth. Ges. Heidelberg., 1903, s. 144. 168 Hertel, Uber Beeinflussung des Organismen durch Licht, speziell durch die chemisch wirksamen Strahlen Zeit. f. Allgem. Physiol., 1904. 169 Hertel, Uber den Gehalt verschiedener Spektralbezirke an physiologisch wirksamer Energie, Zeitschr. f. physik un diatet Therapie, 1905, p. 1. 170 Hertel, Einiges iiber die Empfindlichkeit des Auges gegen Lichtstrahlen, Ber. d. 34. Ves. d. ophth. Ges. Heidelberg, 1907. 171 Hertel, Experimentelles und Klinisches iiber die anwendung lokaler Lichttherapie bei Erkrankungen des Bulbus insbesondere beim Ulcus serpens, Arch. f. Ophth., 66, 275 (1907). 172 Hertel und Henker, tJber die Schadlickheit und Brauchbarkeit unserer modern Lichtquellen, Arch. f. Ophth., 1910, 73, 590 (1910). 173 Hertel, Uber lichtbiologische Fragen. Zeitschr. f. Augenh., 1911. 174 Hertel, Mitteilungen iiber die Wirkung von Lichtstrahlen auf lebende Zellen, Nachr. d. K. Gesellach. d. Wissensch., Gottingen, 1906. 175 Hertel, Uber die Einwirkung von Lichtstrahlen auf den Zellteilungs- prozess, Zeitschr. f. Allgem. Physiol., Vol. 4, 1904. 176 Herzog, Diskussion zum Vortrag Birch-Hirschfeld, Bericht d. ophth. Ges. Heidelberg, 1903, p. 164. 177 Hess, Experimentelles iiber Blitzkatarakt, Internat. ophthalm Kongs, 1888, p. 308. 178 He.ss, Uber die Wirkung ultraviolettes Strahlen auf die Linse, Miinch. Med. Wochenschrift, 1906, No. 36, 1788. 179 Hess, Versuche iiber die Einwirkung ultravioletten Lichtes auf die Linse, Arch. f. Augenh., 1907, p. 185. 180 Hess, tJber " Blaublendheit " durch Gelbfarbung der Linse, Arch f. Augenh., 1908, p. 29. EFFECTS OF RADIANT ENERGY ON THE EYE. 801 181 Hess, Weitere Mitteilungen liber die Gelbfarbung der menschlichen Linse und ihr Einfluss auf das Sehen, Arch. f. Augenh., 64, 293 (1909). 182 Hess, tJber den Farbensinn im indirekten Sehen, Arch. f. Ophth., 35, No. 4, p. 1 (1890). 183 Hess, Pathologie der Linse, Graefe-Samisch. Handbuch, p. 179. 184 Hess, tJber Schadigungen des Auges durch Licht, Arch. f. Augenh., 1913, p. 127. 185 Hessberg, Ein weiterer Beitrag zu den Augenverletzungen durch Blitz- schlag, Miinch. med. Wochenschr., 1909, p. 511. 186 Hewetson, Remarks on acute inflammation .... after witnessing electric welding, Amer. Jour, of Ophth., 1893, p. 297. 187 Hewetson, Danger to the eyes during electric welding etc., Lancet, 2, 946 (1897). 188 Heydt, a Case of so-called Bottle-makers' Cataract, Ophth. Record, 1911, p. 94. 189 Hilbert, Zur Kenntnis der Augenverletzungen durch Blitz, Wochenschr. f. Ther. u. Hyg. d. A., 1909, No. 12. 190 Hill, Two cases of electric blindness. Lancet, 2, 194 (1897). 191 HiLDiGE, Fall von Schneeblindheit, Med. Times and Gaz., 1861. 192 HiRSCHBERG, tJber den Star der Glasbliiser, Zentralbl. f. Augenheilk, 1898, p. 113. 193 HiRSCHLER, Zur Rotsehen der aphakischen, Wien. med. Wochenschr., 4, 1893. .. 194 Hoffman, tJber die Schneeblindheit, und einige verwandte Blendungs- erscheinungen, Mitteil d. deutschen u. osterr. Alpenvereins, 1886, No. 6. 195 HopPE, Augenschadigung durch die Sonnenfinsternis vom 17 April, Miinch. "med. Wochenschr., 1912, Nr. 45. 196 IsAKOWiTZ, Augenerkrankung durch Sonnenblendung, Deutsche med. Wochenschr., 1912, p. 1143. 197 IwANOFF, Les consequences de la foudre sur la vision, Soc. franc d'opht., 11, 463 (1893). 198 Janssen, Sur I'absorption de la chaleur rayonnante obscure dans les milieux des I'oeil, Comptes rend, de I'Acad. des Sciences (Paris), 51, 128-131 (1860). 199 Jess, Verh. d. physikal.-med. Gesellschaft Wiirzburg, 1912. Also Munch. med. Wochenschr., 1912, No. 20. 200 Jess, Die Ringskotome nach Sonnenblendung, Arch. f. Augenheilk., 1913, p. 78. 201 Jones, The effects of electric light on the eye, Ophthalm. Rev., 1883, p. 106. 202 JuNntJS, tJber Unfallverletzungen insondere Augenkrankungen durch electrische Starkstrome, Oph. Klinik, 1906, No. 11. 203 KiRiBucHi, Experimentelle L'ntersuchungen liber Katarakt und sonstige Augenaffektionen durch Blitzschlag, Arch. f. Ophth., 1900. 204 Klug, Du Bois,.Reymond's Arch. f. Physiol., 1878, p. 246. 205 Knapp, Bilateral Macular disease after Short Circuit, Zeitschr. f . Augenh., 29, 440 (1913). 206 Knies, Ein Fall von Augenverletzung durch Blitzschlag, Arc. f. Ophth. 33, 3, 236. 207 KoELSCH, Der Augenschutz in Glashlitten, Miinch. med. Wochenschr., 1911, p. 462. 208 KoLLNER, Die Klinische Diagnose der erwerbenen Violettblindheit, Berl. Ophth., Ges., 1907. 209 KosTER, Verslag over emige experim. betr. de erythropsie, Nederl. Tijdechr. V. Geneeek., 1, 1899. 210 Kramer, Ursachlicher Zusammenhang zwischen einer Augenkrankung (Ret. chorioditis) und Wirkung eines elektrischen Schlages durch den 802 VERHOEFF, BELL, AND WALKER. Korper bezw. Blendung der Augen bei Kurzschluss, Med. Klinik, 1908, p. 803. 210a Krall, Eyesight and artificial illumination, 111. Engr. Soc. 1908, p. 212. 211 Kreibich, Die Wirkung des Sonnenlichtes auf Haut und Conjunctiva, Wien. Klin. Wochenschr., 1904, p. 673. 212 Kretschmer, Verletzung durch elektrischen Strom, Zentralbl. f. Augenh.,. 1899, p. 286. 212a Krienes, Einfiuss des Lichtes auf das Auge, in physiologischer und patho- logischer Beziehung, 1897, 8°. 213 Krotow, Zur Wirkung der Hallauerschen Schutzglaser, Westnick Ophth.,. 26, 1006 (1909). 213a Kruckmann and Telemann, Investigation of temperature of various parts of the eye under "thermo-penetration," Arch. f. Ophth., 86, 395 (1913). 214 Kruss, Die Durchlassigheit einer Anzahl Jenaer optischer Glaser fiir ultraviolette Strahlen, Inaug. Dissert, Jena, 1903. 215 KuBLi, Vier Falle von Erythropsie, Arch. f. Augenh., 18, p. 258. 216 KuHNE, tjber das Vorkommen des Sehpurpurs, Zentralbl. f. med. Wissen- sch., 1877, p. 257. 217 KuwABAR.A., Zur Pathogenese des Blitzstares, Arch. f. Augenh., 62, 1 (1909). 218 Laker, Ein neuer Fall von Augenaffektionen durch Blitzschlag, Arch. f. Augenhoilk, 14, 161 (1884). 218a Laret, Sur la transparence des milieu del'oeil par les rayons u. v., Compt. Rend., 88, 1013 (1879). 219 Larsen, tJber die Intensitat der Sonnenstrahlen, Finsens Mitteil., 1, 1900, p. 99. 220 DE Laroquette, Solar Erythema and Pigmentation, Le Monde Medicale^ Feb., 1913, p. 43. 221 Leber, Uber Katarakt und sonstige Aifektionen durch Blitzschlag, Arch. f. Ophth., 28, 255 (1882). 222 Leber, Die Ernahrungs und Cirkulations Verhaltnisse der Auges, Graefe- Saemisch, 1903, p. 454. 222a Lefranc, Contribution a I'etude de lalumiere et de la Chaleur, consideres comme cause de maladies des yeux chez les verriers principalement. (1883) 4°. 223 Le Roux, Cataracte par decharge electrique. Arch, d'opht., 1902, p. 626. 224 Le Roux, Troubles oculaires d'origine electrique. Arch, d'opht., 1904, p. 727. 225 Le Roux et Renaud, Sur un cas de photo-traumatisme oculaire par la lumiere electrique. Arch. d'Opht., 1908, p. 377. 226 Lescarret, Des scotomes pas eclipse solaire. These Bordeaux, 1900. 227 LiNGSCH, Cataracta traumatica nach Blitzschlag, Wien. med. Wochenschr., 1904, No. 23. 228 Listing, Anitl. Bericht der 40 Versammlung, Deutscher Naturforscher und Arzte zu Hannover, 1865. 229 Little, The effect of electric light on the eye, Ophthalmic Review, 1883, p. 196. 230 LucKiESH, Radiant Energy and the Eye, Electrical World, Oct. 25, 1913. 230a LucKiESH, Glasscs for Protecting the Eyes in Industrial Processes, Trans. 111. Engr. Soc, 9, 472 (1914). 231 LuNDSGAARD, Zwci Falle von Verletzungen des Auges durch electrischen Kurzschluss, Klin. Monatsbl. f. Augenh., 1, 501 (1906). 232 Mackay, On blinding of the retina by direct sunlight, Ophthalmic Review, 13, 147 (1894). 233 Maklakoff, Influence de la lumiere voltaique sur les teguments du corps humain. Arch, d'opht., 1889, p. 97. EFFECTS OF RADIANT ENERGY ON THE EYE. 803 234 Maclean, Cas de foudre, cited in Canstatts Jahresber., 3, 1849. 235 Mann, Histological changes induced in sympathetic, motor, and sensory- nerve cells by functional activity. Journal of Anatomy and Physiol., 29, 1895. 236 Marenholtz V, Ein Beitrag zu den Blitzschiidigungen des Auges, Zeitschr. f. Augenh., 28, 1912. 237 Martin, Pathogenese de I'ophthalmie electrique, Annales d'oculistique, 100, 25 (1888). 238 Martin, E. K., The effect of ultraviolet rays upon the eye. 1912. Pro- ceedings of the Royal Society, (B) 85, No. B 579, p. 319. 239 Mascart, Remarques sur une communication de M. de Chardonnet relative a la vision des radiations ultraviolettes, Comptes rend, de I'Acad. des Sc, 1883, 96, p. 571. 240 Mascart, Sur la visibilite des rayons ultraviolets, Comptes rend, de I'Acad. des Sciences (Paris), 68, 402 (1869). 241 Mayerhausen, Zur Kenntnis der Erythropsie, Wien. med. Presse, 1882, No. 42. 242 McMuLLEN, The Injurious Effect of Light on the Eye, Ophth. Review, 1910, p. 97. 243 Mettey, Recherches experimentales sur le phototraumatisme oculaire par la lumiere electrique. Arch. d. opht., 24, 244 (1904). 244 Meyer, Amblyopie et amauro.se par discharge electrique. Arch. d. Ophth., 22, 1902'. 245 Meyhofer, Zur Aetiologie des grauen Stares Jugenliche Karakten bei Glasmachern, Monatsbl. f. Augenh., 24, 49 (1886). 246 Meyhofer, Ein weiterer Fall von Katarakt nach Blitzschlag, Klin. Monatsbl. f. Augenheilk, 1886, p. 375. 247 Michel, v., tjber objektiv Wirkungen des Lichtes und bestimmter Licht- quellen auf die Netzhaut des Auges, Deutsche Revue, Oktober, 1908. 248 MoNPiLLARD, Etude sur les proprietes des sources de lumiere artificielle et des verres de lunetterie, vis-a-vis des radiations ultra violettes, Recueil d'Opht., 1910, p. 43. 249 MoNTHUS, Brillenglaser zur Absorption ultravioletter Strahlen, Soc. d'Ophthalm., Paris, 4, 6, 1907. 250 Morgan, A case of traumatic cataract due to lightning stroke. The Ophthalmoscope, 1906, p. 166. 251 Nagel, Die Wirkungen des Lichtes auf die Netzhaut, Nagels Handb. 3, 1, (1904). 252 Nagel, tJber Blendungsschmerz, Klin. Monatsbl. f. Augenh., 61, 1 (1903). 253 Nagel, Uber den Ort der Auglosung des Blendungsschmerzes, Klin. Monatsbl. f. Augenh., November, 1901. 254 Nagel and Schaefer, tJber das Verhalten der Netzhautzapfen bei Dunkeladaptation des Auges, Zeitschr. f. Psych, u. Physiol, d. Sinnesorgane, 34, 271 (1904). 255 Nagel und Piper, tJber die Bleichung des Sehpurpurs durch Lichter von verschiedener Wellenlange, Zeitschr. f. Psych, u. Physiol, d. Sinnesorgane, 39, 1905. 256 Nesnamoff, tJber die Einwirkung der chemischen Sonnenstrahlen auf den verlauf der eitrigen Augenerkrankung, Westnik. Opht., 18, 1901. 257 NicoLAi, Ophthalmia photoelectrica, Weekbl. van het Nederl. Tydschr. voor gneesk., 2, 434 (1890). 258 NoDiER, Sur une ophtalmie causee par la lumiere Electrique, These de Paris, 1881. 259 Ogneff, Einige Bemerkungen viber die Wirkung des elektrischen Bogen- hchtes auf die Gewebe des Auges, Arch. f. Phys., 1896, p. 209. 260 Oliver, A clinical study of a case of double chorio-retinitis in the macular regions following a flash of lightning and a flash of burning Lyco- podium. Trans, of Am. Ophthalm. Soc, 1896. -804 VERHOEFF, BELL, AND WALKER. 261 OsswALD, tjber Lichtsinnstorung und Erythropsie bei operierten Myopen, Deutsch. med. Beitr. z. Augenheilk, lOOO, p. 464. 262 Pagenstecher, AugenafTectioii nach Blitzschlag, Arch. f. Augenh., 13, 146 (1884). 263 Paxus, Amblyopie et amaurose par decharge electrique, Arch. d'Ophth., 22, 1902. 264 Parker, Case of Electric Burns of the Eye, Ophth. Record, 1909, p. 85. 265 Parsons, Affections of the Eye produced by undue Exposure to Hght, 17th Internat. Congress of Med., Sect. 9, 199, 1913. 266 Parsons, Text book of Ophthalmological Pathology, 1909. 266a Parsons, J. H., The deleterious effects of bright light on the eyes, Science Progress, July, 1909. 266b Parsons, Glare, its causes and effects, Lond. 111. Engr., Feb., 1910. 266c Parsons, Some effects of bright light on the eyes. Jour. Amer. Med. Assn., Dec. 10, 1909. 267 Pasetti, Ottalmia foto-elettrica. Rivista storica e casi clinici. Annali di Ottalm., 38, 399 (1909). 268 Pastega, Accidenti da elettricita industriale e loro effetti suU' organismo e specialmente suU' organo visivo, Annali med. navale, e coloniale, XVI, 2 (1911). 269 Pergens, Action de la lumiere sur la retine, Ann. de la Soc. roy. des sc. med. Bruxelles, 1896-1897. 270 Pergens, tJber farbige und farblose Augenglaser, Klin. Monatsbl. f. Augeiih., 1897, p. 33. 271 Perlet, tJber den Einfluss des Lichtes auf die Netzhautelemente der Taube, Z, f. Biol., Bd. 52, 1908. 272 Peters, Weitere Beitrage zur Pathologic der Linse, Monatsbl. f. Augen- heilk, 42, 1904. 273 Pfahl, Erfahrungen liber Verletzungen durch Blitz und Elektrizitat, Deutsch. med. Wochenschr., 1908, p. 1267. 274 Philips, Multiple minute ulcers of the cornea follovvmg exposure to electric light. Lancet, 1887. 275 PiNCUs, Die wissenschaftlichen Grundlagen der Zeozontherapie, Arch. f. Augenh., 72, 1913. 276 Pino, Zur Kenntnis und Erkliirung der Erythropsis, Neederl. T. f. Geneesk., 1902, p. 161. 277 Piper, tJber die Funktionen der Stabchen und Zapfen und iiber die Physi- ologische Bedeutung des Sehpurpurs, Med. Klinik. No. 25, u. 26, 1905. 278 Piper, tJber die Lichtwirkung am normalen Auge, Med. Klinik., p. 42, 1907. 279 Plaut, liber die Ursache des Blitz-Keratoconus, Klin. Monatsbl. f. Augenheilk., 1900, p. 334. 280 Plitt, tJber den Friihjahrskatarrh der Augen und den iitiologischen Ein- fluss des Sonnenlichtes, Miinch. med. Wochenschr., 1909, j). 1923. 281 PopEN UND Klimowitsch, Die Durchlassigkeit der Augenmedien fiir ultraviolette Strahlen, Russk. Wratsch., 1910, p. 552. 282 Posey, Lenticular Opacities by Exposure to an Electric Spark, Ophth. Record, 1911, p. 207. 283 Power, Temporary complete loss of vision from exposure of the eyes to a flash of lightning, St. Georges Hosp. Rep., 5, 1870. 284 Prat, Observation d'un coup de soleil electrique. Arch, de med. naval, t., 50, 1888. 285 Praun, Die Verletzungen des Auges Wiesbaden. 1899. 286 Preindlesberger, Drei Falle von Katarakt nach Blitzschlag, Wiener Klin. Wochen., 1901, No. 13. 287 Purtscher, Zur Frage der Erythropsie Aphakischer, ZentralbL f. Augenh., 1883, p. 161. EFFECTS OF RADIANT ENERGY ON THE EYE. 805 288 PuRTSCHER, Neue Beitrage zur Frage der Erythropsie, Arch. f. Augenh., 3, 1887. 289 PuRTSCHER, Ein Fall von Augenaffektion durch Blitzschlag, Arch. f. Ophth.,29, 4, 195 (1883). 290 Rayner-Batten, Eclipse blinding with obstruction of a retinal artery and hemorrhage into the vitreous, Trans. Ophth. Soc. Unit. Kingd., 21, 70 (1900). 290a Recklinghausex, Actiou upon the eye of different sources of artificial light, Le Genie Civil, 1913, p. 434. 290b Recklinghausen, M. de, Hvgienic properties of lamps, Elec. World, 1913, p. 842. 291 Reich, Ein Bhtzschlag, KHn. Monatsbl. f. Augenh., 16, 361 (1878). 292 Reich, Die Neurose des nervosen Sehapparates, hervorgerufen durch anhaltende Wirkung grellen Lichtes, Arch. f. Ophth., 26, (3), 135 (1880). 293 Reichen, Experimentelle Untersuchungen iiber Wirkingen ultraroten Strahlen auf das Auge, Zeit. f. Augenh., 31, 20 (1914). 294 Reid, An inquiry into the human mind on the principles of common sense, 1801, p. 275. 295 V. Reusse, Beitrag zur Kenntnis der Erythropsie, Arch. f. Augenh., 62, 113 (1909). 296 Rivers, Injury to the Eye from a large charge of electricity. Arch, of Ophth. N. Y., 23, 34 (1894). 297 Roemer, Zur Frage des Blendungsschmerzes, Zeitschr. f. Augenh., 8, (1903). 298 Robinson, On bottle-makers cataract, Brit. med. Jour., 1907, p. 381. 299 Roches, Note sur deux cataractes electriques, Annal. d'Oculist, 141, 347 (1909). 300 RocKLiFFE, The effect of electric light on the eye, Ophthalm. Rev., 1883. 301 RoMER, Arbeiten an dem Gebiet der sympathischen Ophthalmologie, Arch. f. Ophth., 55 & 56, (1903). 302 Roy, Effect of intense flashes of light upon the eye, Alabama Med. Journ., 12, 532 (1899). See Dunbar-Roy 102. 303 Ryerson, Lightning stroke causing eye diseases, Med. Rec. April, 1899. 304 Saemisch, Lehstorung infolge von Blitzschlag, Klin. Monatsbl. f. Augenh., 2, 22 (1864). 305 Sauer, Experimente iiber die Lichtbarkeit ultravioletter Strahlen, Ann. d. Physik. u. Chem., 155, 602 (1875). 306 Sautter, Electric Injuries of the Eye, Ophth. Rec, 1911, p. 238. 307 Schanz, Veranderungen u. Schiidigungen d. Auges durch Licht, Deutsch. med. Wochenschr., 1913, No. 8, 308 Schanz, Demonstration des durch ultraviolette Strahlen zu erzeugenden Lidschlussreflexes und der durch diese Strahlen veranlassten Fluo- rescenz der Linse, Ber. u. d. 35, Vers. d. Ophthalm. Gesellsch., Heidel- berg, 1908, p. 353. 309 ScHANZ u. Stockhausen, Wie schutzen wir unserer Augen vor der Ein- wirkung der ultravioletten Strahlen, KHn. Monatsbl. f. Augenh., 1907, p. 454. 310 Schanz u. Stockhausen, Durch des uv. Licht. der modern ktinstlichen Lichtquellen eine Schildigung des Auges zu befiirchten?, Elektro- technische Zeits., 1908. 311 Schanz u. Stockhausen, Die Schadigung des Auges durch Einwirkung des uv. Licht. Elektrotechnische Zeitschr., 33, 777 (1908). 312 Schanz u. Stockhausen, Zur Beurteilung der Schadigung des Auges durch leuchtende und ultraviolette Strahlen, Khn. Monatsbl. f. Augenh., 1909, p. 452. 806 VERHOEFF, BELL, AND WALKER. 313 ScHANZ u. Stockhausen, Die Wirkung der ultravioletten Lichtstrahlen auf das Auge, Berlin Klin. Wochenschrift, 1909, No. 21. 314 ScHANZ u. Stockhausen, Schutz gewerblicher Arbeiten gegen Blendung Sociale Medizin un Hygiene, 4, 403 (1909). 315 ScHANZ u. Stockhausen, tJber die Wirkung der Ultravioletten Strahlen auf des Auges, Arch. f. Ophth., 69, 452 (1909). 316 ScHANZ u. Stockhausen, Wie schutzen wir unsere augen von die Ein- wirkung der uv. Strahlen unser kunstlichen Lichtquellen, Arch. f. Ophth., 69, 49 (1909). 317 ScHANZ u. Stockhausen. tJber Blendung, Arch. f. Ophth., 71, 175 (1909). 318 ScHANZ u. Stockhausen, tJber die Fluorescenz der Linse, Arch, f . Ophthal. 73, 184 (1909). 319 Schanz u. Stockhausen, Weiterer uber Blendung, Arch. f. Ophth., 73, p. 561. 320 Schanz u. Stockhausen, Zur Enstehung des Glasraacherstars, Arch. f. Ophth., 73, 553 (1910). 321 Schanz u. Stockhausen, tJber die Schadlichkeiten und Brauchbarkeit unserer modern Lichtquellen, Arch. f. Ophth., 75, 403 (1910). 322 Schanz u. Stockhausen, Schutz der Augen gegen die schadigenden Wirkungen der kurzwelligen Lichtstrahlen, Zeitschr. f. Augenh., 23, 399 (1910). 323 Schiele, Friihjahrekatarrh der Conjunctiva, Arch. f. Augenh., 19, 281 (1888). 324 Schiess-Gemusens, Uber Schneeblindheit, Arch f. Ophth., 25, p. 3. 325 ScHjERNiNG, tJber die Absorption der ultravioletten Strahlen durch verschiedene optische Gliiser, Diss. Berlin, 1885. 326 Schleicher, Ein Fall von Katarakt nach Blitzschlag, Mitteilung a. d. Ophth. Klin. Tubingen, 2, 3, 295 (1890). 327 Schneller, Zur Kasuistik der Chorioretinitis nach tJberblendung, Arch, f . Ophth... 30, p. 1. 328 Schueler, Uber Blendung nach Beobachtung einer Sonnenfinsternis, In.-Diss., Heidelberg, 1912. 329 ScHULEK, Die Erythropsie, Ungar. Beitr. z. Augenheilk., 1, 101 (1895). 330 ScHULEK, Schutzbrillen gegen Ultravioletten auf Grund photologischer Studien, Ungar. Beitr. z. Augenheilk., 2, 467 (1899). 331 ScHWiTZER, Beitrage zur Entstehung des grauen Alterstares, Ungar. Beitr. z. Augenheilk., 2, p. 291. 332 ScHWiTZER, Uber die Aetiologie des grauen Stares, Orvosi Hetilap. Szemes- zet, 1899, No. 3. 333 Seabrook, Amber yellow glass in the examination and treatment of eyes, Med. News. Aug., 1903. 334 Sekulic, Ultraviolette Strahlen sind unmittelbar sichtbar, Ann. d. Fhysik u. Chemie, 146, 157 (1872). 335 Servais, Observation de cataracte produite par la foudre, Ann. d'Ocul., 52, 185 (1864). 336 Setschenow, tiller die Fluoreszenz der durchsichtigen Augenmedien beim Menschen, Arch. f. Ophth., 1859, p. 205. 337 SiLEX, Beitrag zur Casuistik der Augenaffectionen in Folge von Blitzschlag, Arch. f. Augenh., 17, p. 65. 338 SiLFVAST, ELn Fall durch Blitzschlag hervorgerufener Lasionen der Augen, Finska. lakares, Zeitschr. f. Augenh., 9, 320 (1902). 339 Smith, A Scotometer, Ophth. Rev., 1906, p. 155. 340 Snell, Sun blindness of the retina, Brit. med. Jour., Jan. 18, 1902. 341 Snellen, Erythropsie, Arch. f. Ophth., 44, p. 1. 342 Soemmering, Pflichten gegen die Augen., 1791. EFFECTS OF RADIANT ENERGY ON THE EYE. 807 343 SoRET, Sur la visibilite des rayons ultraviolettes, Compt. rend, de I'Acad. des Sc, 97, 314 (1893). 344 SoRET, Sur I'absorption des rayons ultraviolets par les milieux de I'oeil et par quelques autres substances, Compt. rend, de I'Acad. des Sc, 97, 572 (1893). 345 SoRET, Sur la transparence des milieux de I'oeil pour les rayons ultra- violets, Compt. rend, de I'Acad. des Sc, 88, 1012 (1879). 346 SoRET, Archives des Sciences Phys. et. Naturelles, Recherches sur I'ab- sorption des rayons ultra-violets, 10, 429 (1883). 346a SoRSHKE, Untcrsuchungen fiber die Herabsetzungen der Sehscharfe durch Blendung, Zeitschr. f. Phys. d. Sinnensorg, 34 and 35, 1904. 347 Speleers, Gesichtsfeldaufnahmen bei Ringskotom durch Sonnenfinsternis, Klin. Monatsbl. f. Augenh., 1912. 348 Staerkle, tJber die SchadUchkeit moderner Lichtquellen auf das Auge und deren Verhiitung, Arch. f. Augenh., 50, 12 (1904). 349 Stein, Untersuchungen liber Glasbliiserstar, Arch. f. Augenh., 74, 53 (1913). 350 Steiner, Zur Kenntnis der Erythropsie, Wien. med. Presse, 1882, No. 44. 351 Steinheim, Zur Kasuistikder Erythropsie, Zentralbl. f. Augenh., 1884, p. 44. 352 Stern and Hesse, Uber die Wirkung des Violettlichtes auf die Haut, Mlinch. med. Wochens., 1907, p. 318. 353 Stigell, tjber Blendung der Netzhaut, Diss. Strassburg, 1883. 354 Stocke, Sonneneklipse Skotome. Viaamsch Natur en Geneeskundig Congres, Klin. Monatsbl. f. Augenh., November, 1912. 355 Stockhausen, Blendung, ihre Ursache und Wirkung, Zeitschr. f. Beleuch- tungsw., 1910. 356 Stockhausen, Die Beleuchtung von Arbeitsplatzen und Arbeitsraumen, Naturforscher-Vers. Dresden, 1907, Ophth. Sektion 3 Sitzung. 357 Stokes, Philosophical transactions, 1852, p. 558. Note 73. 358 Strauss, Resultate der Uviollichtbehandlung bei Hautkrankheiten, Dermat. Zeitschr., 13, 1907. 359 Strobel, Lichttherapie und Augenheilkunde, 75 Vers, deutschen Natur- forscher in Kassel, 1903. 359a Stubel, Fluorescence of animal tissues in ultraviolet light, Pfiuges. Arch, f. d. Ges. Physiol, 142, 1 (1911). 360 SuLZER, Vier Falle von Retinaaffektion durch direkte Beobachungen der Eklipse, Klin. Monatsbl. f. Augenh., 1883, p. 142. 361 Sweet, Electric Burn of Eyes. (Section on Ophth. College of Physic. of Philadelphia), Deutschr. med. Wockenschr., 1908, p. 212. 362 Talko, Contusion des Auges durch Blitz, Zeitschr. f. Augenh., 5, p. 484. 363 Terrier, Ophthahnie electrique. Arch. d'Ophth., 8, 1 (1888). 364 Terrien, Cataracte par decharge electrique, Arch. d'Ophth., 28, 679 (1902). 365 Terrien, Du prognostic des troubles visuels d'organism electrique. Arch. d'Ophthalm., 22, 1902. 366 TiGERSTEDT, Die Grenzen des sichtbaren Spektrums, Biophysikalisches Centralblatt, 1905, 1, 1 to 4 and 33-38. 367 ToczYSKi, Ein ungewohnlicher Fall von Augenverletzung durch Blitz- schlag, Arch. f. Augenh., 70, 1911. 368 Trendelenburg, Uber die Bleichung des Sehpurpurs mit spektrahlen. Licht in ihrer Abhangigkeit von der Wellenlange, Zentralbl. f . Physiol. 1904, No. 24. 369 TscHERMAK, Die Hell-Dunkel-Adaptation des Auges, Ergebn. d. Physiol. Asher- Spiro, 1902. 370 Tyndall, Das Licht, Deutsche Ausgabe, Braunschweig, 1876, p. 177-192. 371 Tyndall, Ueber leuchteude und dunkele Strahlung, Pogg. Ann. 124, 36 (1865). 808 VERHOEFF, BELL, AND WALKER. 372 Uhle, Anamie des Nervus opticus iind Retina durch Blitzschlag, Klin, Monatsbl. f. Augenheilk., 24, 379 (1886). 373 Uhthoff, Ein Fall von einseitiger zentraler Blendung-Retinitis durch elektrisches Bogenlicht mit nachfolgender traumatischer Neurose, Zeitschr. f. Augenh., 2, 341 (1899). 374 Uhthoff, Blitzschlagwirkung auf das Auge, Deutsch. med. Woehenschr., 1907, p. 1841. 375 Uhthoff, Zur zentralen Blendungsretinitis bei Beobachtung der Sonnen- finsternis 1912, Sitzungsber, der wissensch. Akad. d. Augenarzte Schlesiens, 1912. 376 Uhthoff, Frlihjahrskatarrh, Arch. f. Ophth., 29, 177. 377 Ulbrich, Optikusatrophe nach Einwirkung eines elektrischen Stromes, Zentralbl. f. Augenh., 1900, p. 264. 378 Valois, Ophtalmie electrique, Clin, ophth., 1904, p. 92. 379 Valois et Lemoine, Troubles visuels consecutifs a I'observation directe de la derniere eclipse de soleil. Rev. gen. d'ophth., Sept. 30, 1912. 380 Valtjde, L'erythropsie, Paris, Steinheil, 1888. 381 Van Lint, Accidents oculaires provoques par I'electricite, Rapport presente a la Soc. beige d'ophthalm. Seance, Bruxelles, Nov. 1909. 382 Verhoeff, Ultra violet light as a germicidal agent. Experimental in- vestigation of its possible therapeutic value. Jour. Amer. Med. Ass., 62, 762 (1914). 383 Verhaeghe, Cataracte par coup de foudre, Gaz. des Hopitaux, 1905, No. 89. 384 Vetsch, Frlihjahrskatarrh der Conjunctiva, Inaug. Disert., Zurich, 1879, p. 35. 385 ViLLARD, Troubles oculaires consecutifs a Tobservation directe des eclipses de soleil, Ann. d'oc, 1906. 386 ViLLARD, Augenstorungen nach direkter Beobachtung von Sonnenfin- sternis, Soc. d'Opht., 1906. 387 ViNSONNEAu, Scotome par eclipse solaire et lesion maculaire. Arch, d'opht., Sept., 1912. 388 VoEGE, 1st durch das uv. Licht der modernen kiinstlichen Lichtquellen eine Schadigung des Auges zu befiirchten, Elektrotech. Zeits., 1908, No. 33. 389 VoEGE, Bemerkungen zu den Aufsatz Schanz und Stockhausen: Uber die Wirkung der ultravioletten Strahlen auf das Auge, Arch. f. Ophth. 70, 403 (1908). 390 VoEGE, Die Ultravioletten Strahlen der kiinstlichen Lichtquellen und ihre augenblickliche Gefahr fiir das Auge, Verlag von J. Springer, Berlin, 1910, p. 14. 391 VoEGE, tjber den Schutz des Auges gegen die Einwirkung der ultravioletten Strahlen unserer kiinstlichen Lichtquellen, Elektrotechn. Zeitschr., 1909. 392 VoEGE, tlber Licht und Wiirmestrahlung der kiinstlichen Lichtquellen, Journ. f. Gasbeleuchtung und Wasserversorgung, 1911, No. 13. 394 VoEGE, A Simple Apparatus for the Examination of the Colour and Quality of Radiation furnished by Artificial Sources of Light, The Illuminating Engineer, January, 1909. 395 VoEGE, Need we fear the effects of the ultra violet constituent in the light from modern illuminants on the eye. The Illuminating Engineer. London, 2, 205 (1909). 395a VoEGEL, Lichttherapie in prakt. elektrotechnick, Electrotech. Zeitsch., 30, 32, Jan. 14, (1909). 396 VoGT, Schutz des Auges gegen die Einwirkung ultravioletten Strahlen durch eine neue nahezu farblose Glasart, Arch. f. Augenh., 59, 48 (1907). EFFECTS OF RADIANT ENERGY ON THE EYE. 809 397 VoGT, Ursache unci Wesen der Ervthropsie, Bericht. d. ophth. Ges Heidelberg, 190S. 398 VoGT, Bemerkungen zu der Replik von Schanz und Stockhausen, etc., Arch. f. Augenh., 74, 4, 408 (1910). 399 VoGT, Erwiderung auf die Arbeit von Best, 24 Klin. Monatsbl. f. Augenh., 68, 238 (1910). 400 VoGT, Experiment. Untersuchungen liber die Durchlassigkeit der durch- sichtigen Medien des Auges fur. das Ultrarot kiinstlicher Licht- quellen, Arch. f. Ophth., 81, 1 (1912). 401 VoGT, Einige Messungen der Diathermasie das menschlichen Augapfels etc., Arch. f. Ophth. 83, 1912. 402 VoGT, Kritik der Abhandlung von Schanz und Stockhausen und Best, Arch. f. Augenh., 64, 344 (1909). 403 VoGT, Beitrag zu der Frage der Entstehung der Blendungserythropsie, Arch. f. Augenh., 60, 91 (1908). 404 VoGT, Schadliche Lichtquellen und Schutzglaser gegen dieselben, Med. Klinik., 1910, No. 9. 405 Vossius, Ein FaU von Blitzaffektion des Auges, Beitr. z. Augenh., 1892, P- 42. 406 Vossius, tlber die durch Blitzschlag bedingten Augenaffektionen, Berl. klin Wochenschr., 1886, No. 19, p. 304. 407 Vossius, Fall von Ophthalmia electrica, Berhner klin. Wochenschr., 1886, p. 304. 408 DE Waele, Asthenopsie nerveuse par lumiere electrique. Emploi de verres jaunes. Bull, de la Soc. beige d'opht., 1909. 409 Webster, Fox and Gould, Retinal insensibility to ultraviolet and infra red rays, Amer. Jour, of Ophth., 1886, p. 345. 410 Wendenberg, Schadigungen des Sehorgans durch Blendung bei Sonnen- finsternis Beobachtungen, S. Ivarger, Berlin, 1914. 411 WiDMARK, ilber den Einfluss des Lichtcs auf die vordern Medien des Auges un Haut- Uber die Durchlassigkeit der Augen fiir ultravioletten Strahlen. Beitrage zur Ophthalmologic, Leipzig, 1891, p. 355-502. 412 WiDMARK, tjber Blendung der Netzhaut, Skand. Arch., 4, 281 (1893). 413 WiDMARK, tJber die Durchdringlichkeit der Augen Medien fiir ultraviolette Strahlen, Skand. Arch. f. Physiol., 3, 14016 (1892). 414 WiDMARK, tJber den Einfluss des Lichtes auf die vorderen Telle des Auges, Nord. med. Ak., 21, 1889. 415 WiDMARK, tJber den Einfluss des Lichtes auf die vordernen Medien des Auges, Skand. Arch., 1, 264 (1889). 416 WiDMARK, Uber die Grenze des sichtbaren Spektrums nach der violetten seite, Mitteil. aus der Augenklinik zu Stockholm, Jena., 1898, p. 31. 417 WiDMARK, Von den pathologischen Wirkungen starker Lichtquellen auf das Auge, II Nord. ophth. vers. Kopenhagen, July, 1903 418 WiDMARK, tJber den Einfluss des Lichtes auf die Linse, Mitteil. aus d. Augenklin. d. Carol, med.-chir., Inst, zu Stockholm, 3, 135 (1901). 419 Williams, Observations on the Effect of the Light of the Mercury-vapour lamp on the Eye, Electrical World, Sept., 1911. 420 W^olffberg, Electroopthalmie und Hysteric, Wochenschr. f. Ther. und Hyg. d. Auges, 1903. 421 WtJRDEMAN AND MuRRAY, Case of macular retinitis due to flash of electric light, Ophth. Rec. 8, 220 (1899). 422 Wydler, Experim. Unters. tiber Blendungsnachbilder u. deren Verhaltnis zur Blendungserythropsie, Zeitschr. f. Augenh., 27, 299, 428, 524 (1912). 423 ZicKENDRAHT, Notiz fur die Absorptiongrenzen emiger Gliiser im Ultra- violett, Zeitschr. f. wissenschr. Photogr., 7, 8 (1909). 810 VERHOEFF, BELL, AND WALKER. 424 ZiMMERMANN, Beitrag zur Kenntnis der durch intensives Licht hervor- gerufenen Verjinderung des Sehorganes, Festschr. d. Stuttgarter arztl., Vereins., 1897. 425 ZiRM, Ein Fall von bleibenden, ausgedehnten Veranderungen der beiden Maculae durch direktes Sonnenlicht., Arch. f. Ophth., 60, 401 (1905) 426 ZscHiMMER, tjber neue Glasarten von gesteigerter ultraviolett durch- lassigkeit, Zeitschr. f. Instrumenkunde, 1903, No. 12. 427 ZscHiMMER, Die Physikalischen Eigenschaften des Glases als Funktioneii chemischen Zusammenretzung, Zeitschr. f. Elektrochemie, 1905, No. 38. 428 ZscHiMMER, Diskussion zum Vortrag Schanz, Elektrotechn. Zeitschr., 1908, p. 848. PLATE 1. Figure 1. Exp. 69. Mag- netite arc, double lens system, no screen, exposure 20 min- utes. This exposure was about 200 times the liminal exposure for photophthalmia. Cornea 12 days after ex- posure. The epithelium ha^ reformed. The stroma is softened down to Descemets membrane. The unevenness of the corneal surface is due to the irregular shrinkage of the semiliquefied tissue in the process of fixation. Note that the effect on the stroma is progi'essively greater and more widespread towards the external surface. Photo. X 12. Fig. 1. Figure 2. Exp. 81. Mag- netite arc. Double lens sys- tem, water cell, flint glass screen (305 mm), exposure 1 hour. Relatively slight heat effect on the cornea after three days. The epithelium is intact but the endothelium is destroyed. The corneal corpuscles are present and actively prolif- erating in the anterior layers, while many of them are de- stroyed in the posterior layers. Photo. X 32. Fig. 2 Figure 3. Exp. 88. Mag- netite arc. Double lens sys- tem, no water cell. Flint glass screen (315 mm), exposure 1 hour. Marked heat effect on cor- nea after 48 hours. The epithelium is intact. I'he stroma is swollen to twice its normal thickness and the cor- puscles are completely de- stroyed in the most exposed area. At the periphery the reappearance of the corpuscles is abrupt and they are in active proliferation. The en- dothelium is destroyed over a large area. Photo. X 39. Fie. 3. PLATE -2. f<^ vr -. ^ ^ ^ % K^' ii --,v :/ai Fia. 4. Fii,'. Figure 4. Normal lens capsular t'pitlK'lium of ral:)bit. Fiat preparation. Photo. X 264. Figure 5. Exp. 5.5. Magnetite arc. Single lens system, crown glass screen (295 nix), exposure 1 hour. Lens capsular eiiithehum 24 hours after exposure, showing marked abiotic effects. The cells ai'e swollen and most of them contain granules. In the photograph the basojihilic and eosinophilic granules cannot readily be dis- tinguished from each other. Photo. X 264. "W /J5P- 4 Figure 6. Exp. 67. ALagnetite arc. Double lens system, no screen, exposure 20 minutes, about 200 times the liminal exposure for photophthalmia. Lens capsular epithelium 48 hours after exposure, showing marked abiotic effects. The darker granules in the cells are basophilic, the lighter, eosino- philic. The clear spaces in this and the ]irevious figure are due to some of the cells having failed to adhere to the capsule. Photo. X 264. 812 PLATE 3. » V mi ♦«.c t ^ ^ e t * # ijj a Fig. 7. Exp. 70. Fis. S. Magnetite arc. DouVjle lens system, no screen, Figure 7. exposure 20 minutes. Lens capsular epithelium 2 months after exi:)osure, showing rejiarative changes. The nuclei vary greatly in size and many of them are extremely large. iSome of the cells contain double nuclei. The small dots in the cells are nuclear buds constricted off from the main nuclei. They are not related in any way to the granules in Figs. 3 and 4. Photo. X 264. Figure 8. Magnetite arc. Double lens system. Crown glass screen (295 nn), exposure 20 minutes. Lens capsular epithelium 19 hours after exposure, showing wall of deeply staining cells, corresponding in position to the pupillary margin. Photo. X 42. Fig. 9. Figure 9. Exp. 54. Fig. 10. Single lens system, crown glass Magnetite arc. screen (295 mm), exposure 20 minutes. Lens capsular epithelium 48 hours after exposure, showing vail. Many of the cells in the unexposed zone outside the wall are in mitosis. Photo. X 42. Figure 10. Lens capsular epithelium 48 hours after injection of Lugol's solution in the anterior chamber, showing wall similar to that produced by abiotic radiations. (See page 676), Photo. X 106. 813 PLATE 4. f^90m^.^ iri '^ \U:\ §.3^-m* 9f^ Fig. 13. Fig. 12. Figure 11. Exp. 78. Magnetite arc. Double lens system, water cell, flint glass screen (298 /u/x), exposure 1 liour. Showing retinal ganglion cells of rabbit unaffected 48 hours after exposure. Thionin stain. Photo. X 706. Figure 12. Exp. .53. Magnetite arc. Single lens system, water cell, crown glass screen (295 n^x), exposure 12 minutes. Heat effect on retina after 48 hours. In the affected area the rods and cones are disintegrated and the nuclei of the nuclear layer are fragmented. The inner layers of the retina are normal. Photo. X 42. Figure 13. Same experiment. Affected portion of retina under higher magnification. Photo. X 190. Fig. 14. Figure 14. Exp. 98. Sunlight concentrated by large mirror. Water cell, no screen. Exposure 14 seconds. Intense heat effect on retina after six days. On the left, the retina is normal. On the right, the outer layers including the rods and cones are coagulated and retain their forms, while the inner layers are disintegrated, the heat here having been insufhcient to coagulate them. For the same reason the entire thickness of the retina is disintegrated at the margin of the affected area. The choroid shows a large hemorrhagic extravasation. Photo. X 28. 814 ■> a .o CO .S •I I tit oT j-j 815 ^•^ o PIH S CO ^ o a. ■-: a a fe c (M lO re s -< (^ r^ o Ph :r c^ ra ■^ lO :ti^ II '^ 816 K r^ r^ •^ r^ 0^ c bU Oj o S B K. He was a Corresponding Member of the New York, New Hampshire, Maine and Rhode Island Historical Societies. During a period of about thirty years Mr. Goodell was connected with the publication of the Province Laws and for seventeen years he filled the office and actually performed the functions of Editor. It was in deference to his ideas that the system of annotation was adopted which prevails throughout the five volumes of the public Laws. It was his purpose to furnish material illustrative of the conditions in the community which caused the legislation to which the notes were appended. In the search for material of this char- acter, he ransacked the Archives, searched the records of the Courts, examined contemporary publications and caused similar searches and examinations to be instituted at the Records Office in England. His interest in the work, great at all times, increased with its progress. His care and attention were directed not only towards the faithful reproduction of all this contributory material, but he also developed a corps of assistants whose work was characterized by phenomenal accuracy. He may be said to have become thoroughly identified with the work and upon its reputation among future students will depend his fame. It is quite certain that the careful study by way of preparation of these notes made him unquestionably the best in- formed man in Massachusetts on the customs and modes of life during the days of the province. Mr. Goodell's connection with the publication of the Province Laws was abruptly severed in 1896. This was brought about by impatience at delays in bringing forth the volumes, delays which were caused by his desire for completeness and accuracy. Even before the work was finally stopped, he was compelled to abandon his system of annotation and to revert to that set forth in the original resolution, marginal references. The methods by which the publication was closed were of a nature to leave ill-will between the parties concerned. Never- theless, three years thereafter, when the renewal of the publication was under consideration, the Governor and Council consented to the tender of the Editorship to Mr. Goodell. He declined however to accept the position. He was married in 1866, and when he died he left surviving him, a widow and two sons. His house was always open to his friends and here surrounded by the attentions of a devoted family, he passed JOHN CHIPMAN GRA.Y. the later days of his life. Those who were able to profit l)y his hospitality dwell upon his genial manners and the great power and interest of his conversational faculties. The attention of the iVcademy is of necessity concentrated upon his liistorical research in connection with the Province Laws. It is upon the character and value of that work that his fame must rest. While we can not at present predict with certainty what the verdict of posterity will be, we have at any rate at our command the pro- found respect with which students today treat these annotations, upon which we may predicate an opinion as to what is impending in the future. For the details given of Mr. Goodell's Salem life, this sketch is indebted to a Memoir published in the Bulletin of the Essex Institute, to which one mav turn for fuller details concerning this portion of his life. Andrew McFarl.'Vnd Davis. JOHN CHIPMAN GRAY (1839.-1915) Fellow in Class III, Section 1, 1878, Vice-President, 1902-1912. John Chipman Gray, long Fellow and for se^'eral years Vice- President of the Academy, was born in Brighton, July 14, 1839; received from Harvard University the degrees of A.B. in 1859, LL.B. in 1861, and A.M. in 1862; served in the army of the United States, 1862-65; practised law in Boston 1865-1915; and was lecturer and Professor in the Harvard Law School continuously, excepting two years before 1875, from 1869 until his resignation in 1913; died February 25, 1915. He received the honorary degree of LL.D. from Yale and Harvard. He had been President of the Harvard Alumni Association, of the Harvard chapter of Phi Beta Kappa, and of the Boston Bar Association, and a member of the Massachusetts Histori- cal Society. He was a legal author of eminence, his principal works being Restraints on Alienation (1883, 1895), The Rule against Per- petuities (1886, 1906, 1915), The Nature and Sources of the Law (1909). Professor Gray was a virile man, mentally and physically, one in whom wisdom, judgment, and probity were joined with illuminated HENRY WILLIAMSON HAYNES. 889 common sense. A distinguished and productive student and teacher, whose influence was profound upon forty classes of students, he was also a lawyer whose counsel was sought by clients of all sorts, and was desired and found acceptable by the courts. He was not merely a learned lawyer; he was besides a graceful as well as forceful writer, a classical scholar of no mean attainments, a man of ready wit, and an administrator of large affairs. To his friends and associates he was, in Dean Thayer's noble phrase, "a rock of trust." In him the Academy has lost a useful and faithful Fellow and a wise counsellor. J. H. Beale. HENRY WILLIAMSON HAYNES (1831-1912) Fellow in Class III, Section 2, 1880, Librarian, 1886-1899. Henry Williamson Haynes was born in Bangor, Maine, September 20, 1831, the only son of Nathaniel and Caroline Jemima (William- son) Haynes. He attended the Boston Latin School and was gradu- ated from Harvard College with the class of 1851. He taught for one or two years, then studied law, and was admitted to the bar in Boston on Sep):ember 28, 185G. In 1867 he was made Professor of Greek and Latin in the University of ^^ermont, and in 1869 Librarian of the same institution, positions which he held until 1873. From that year until his death on February 18, 1912, he made his home in Boston, though he several times spent many months in travel abroad. He was elected to membership in the Academy October 13, 1880. On May 25, 1886, he was chosen Librarian, an office which he held until 1899. The most striking characteristic of Professor Haynes was the l)readth of his intellectual interests. Always a lover of literature, especially of the literatures of Greece and Rome, he early interested himself in natural history, in archaeology, and in anthropology, and was one of the first Americans to devote himself to the study of pre- historic monuments. The papers and notes which lie contributed to many difterent journals cover an unusually wide field, ranging all the way from comments on Latin mottoes and literary and historical notes of many kinds to discussions of palaeolithic implements and 890 GEORGE WILLIAM HILL. aboriginal fire-making. He was an indefatigable collector and brought together a valuable library and an important collection of archaeologi- cal material, which he bequeathed to Harvard University, the Mu- seum of Fine Arts, and the Boston Society of Natural History. A comparati\'ely full account of his life, with a list of his published writings, was printed in the American Arithropologist, Vol. XV, 1913, pp. 336-346. G. H. Chase. GEORGE WILLIAM HILL (1838-1914) Fellow in Class I, Section 1, 18G5. Hill's work on the lunar theory justly entitles him to a permanent and important place in the history of celestial mechanics. In his two papers, "Researches in the Lunar Theory," and "The Motion of the Perigee of the Moon," he made contributions which have already created a fundamental change in the methods of viewing the problem of three bodies. In the first of these papers he outlined a plan for dealing with the lunar theory as a practical problem and worked out the first approximation. In the second paper the chief difficulty of the second approximation is solved in a manner which shows Hill's high capacity for algebraic analysis. This method, though based on one by Euler, was given a form which was capable of development into a complete theory. But to most students of the subject of celestial mechanics his initiation of the idea of the periodic orbit in the former of these papers is his greatest achievement. The great work of Poincare which has pointed out new regions of research is based mainly on this idea, and many others have followed in his footsteps. Another feature of the memoir is the stress laid on the zero velocity curve from which we may hope in the future to have some more exact ideas as to the stability of planetary and satellite systems. Ten years of his life were spent in constructing theories of the motions of Jupiter and Saturn, and in forming tables for the prediction of the positions of those bodies. This work, although laborious and carried out with great care and accuracy, contains little that is new. Hansen's method is used with almost no change. In spite of the success which attended his labors one cannot help a feeling of regret EDWARD SINGLETON HOLDEN. 891 that Hill, with his genius and originaHty, should have spent the best years of his life on computations which could have been carried through by others if the necessary delay in training them for the work had been granted. Hill published a number of other papers on a variety of astronomical subjects, many of them worthy of careful study for Hill always put the impress of his own mind into his memoirs, even when developing the ideas of his predecessors. The Carnegie Institution of Washington has published Hill's memoirs in four large quarto volumes. For practically the whole of his life Hill was a member of the staff of the Nautical Almanac office, doing his work as far as possible at his home in West Nyack, N. Y. He cared little for money or luxuries, and resigned his post as soon as he had completed the tables of Jupiter and Saturn. He was essentially a man who lived for the development of his own ideas and who cared little for contact with his fellow men. Nevertheless those who knew him well, and were in the habit of accompanying him on the walks or trips which he frec^uently took in the country, always speak of the pleasure of his companionship. But few people knew him personally, and his life was so uneventful that the biographer finds little to record, yet his name will live after those of most of his generation are forgotten. He was born in New York City on March 3, 1838, and died at his home in West Nyack, N. Y., on April 16, 1914. E. W. Brown. EDWARD SINGLETON HOLDEN (1846-1914) Fellow in Class I, Section 1, 1885. Edward Singleton Holden was born in St. Louis, Missouri, on November 5, 1846. He was a student in Washington University, St. Louis, in the years 1862-66, and was graduated in the latter year with the B. S. degree, William Chauvenet, author of the well known " Manual of Spherical and Practical Astronom}^," was Chancellor of, and Professor of Mathematics and Astronomy in, Washington Uni- versity during that period. It is probable that Mr. Holden pursued astronomical studies with Professor Chauvenet, but his interest in astronomical subjects had been aroused on the occasions of visits to 892 EDWARD SINGLETON HOLDEN. Harvard College Observatory while his cousin, Professor George P. Bond, was director of that institution (1859-65). Mr. Holden mar- ried Professor Chauvenet's daughter Mary in the year 1871. During the years 1866-70 Mr. Holden was a cadet in the U. S. Military Academy at West Point, graduating in the latter year. He ranked third in his class. In 1870-71 he was Second Lieutenant in the Fourth U. S. Artillary, in the following year Instructor in Natural Philosophy in the Military Academy, and in 1872-73 In- structor in Practical Military Engineering. In 1872 he published an octavo treatise on " The Bastion System of Fortifications, Its Defects and Their Remedies." In March, 1873, Lieutenant Holden resigned his commission in the army to accept appointment as Professor of Mathematics in the U. S. Navy, for service as astronomer in the U. S. Naval Observatory at Washington. There he came at once into close personal relations with Simon Newcomb, serving as assistant to Newcomb with the just completed 26-inch Clark refracting telescope. It is clear from historical developments, as well as from passages in Newcomb's "The Reminiscences of an Astronomer," that Newcomb was tre- mendously impressed with Holden's energy and ability. When Mr. D. O. Mills, President of James Lick's first Board of Trustees, went tc Washington in 1874 to consult with Newcomb concerning plans for the projected Lick Observatory, Newcomb "suggested that a director of the new establishment should be chosen in advance of beginning active work, o that everything should be done under his supervision. As such director I suggested that very likely Professor Holden, then my assistant on the great equatorial, might be well qualified." It is an illuminating comment upon Professor Holden's promise as an astronomer of the future that he should be recommended, and I think tentatively selected, as the director of the proposed Lick Ob- servatory, to contain the largest and most powerful telescope in existence, at a time when his astronomical experience had covered little more than one year. Professor Holden was then less than twenty-eight years of age. Professor Holden went to London in 1876 as a delegate from the U. S. Government to examine and report upon the South Kensington Loan Collection of Scientific Instruments, especially as to improve- ments in the astronomjcal and geodetical instruments. While on the staff of the Na^•al Observatory, Professor Holden mafle himself A'erv familiar with the literature of astronomy; he EDWARD SINGLETON HOLDEN. 893 observed comets, nebulae, satellites and double stars; he prepared and published bibliographies relating to nebulae and star clusters, to the transits of Mercury, and to other subjects; he prepared annual reports on the progress of astronomy; he published a critical descrip- tion of the 26-inch refracting telescope, a monograph on the Nebula of Orion, a volume on The Life and Works of Sir William Herschel, and, in collaboration with Professor Newcomb, a textbook on astronomy. Professor Holden was in charge of an expedition dispatched by the Naval Observatory to observe the total solar eclipse of July, 1878, at Central City, Colorado. In 1881 Professor Holden resigned his commission in the navy and accepted the directorship of the Washburn Observatory, University of Wisconsin. This position he held during the years 1881-85. At Madison he organized and pushed the observational work of the Washburn Observatory with vigor. Professor Holden was in charge of the expedition sent by the National Academy of Sciences to observe the total solar eclipse of May, 1883, in the Caroline Islands, where he was assisted by four young astronomers whose names later became well known: C. S. Hastings, E. D. Preston, S. J. Brown and Winslow Upton. An extensive program was carried through successfully. The volume containing the observational results is a model in form, and is fre- quently referred to by eclipse observers. In 1885 Professor Holden was appointed President of the Univer- sity of California and Director of the Lick Observatory, to serve in the former capacity until the Observatory should be completed, and thereafter in the latter capacity. He assumed the active directorship of the Lick Observatory on June 1, 1888. Professor Holden was consulted extensively by the Lick Trustees, beginning with 1874, and during the last three years of the construction period he was intimately associated with the Trustees as adviser. Professor Holden's term of office as the first Director of the Lick Observatory terminated b^y resignation on December 31, 1897. An astronomer has said that " the first requisite for the director of a great observatory is to have a very clear notion of just what kind of work ought to be done, how it should be done, and then to give all the aid in his power to the investigator." Director Holden selected the most promising men he could find in the United States to comprise the Observatory staff. He assigned them definitely to lines of research which the succeeding years have shown to be of the highest importance. He gave them such opportunities to succeed as no other astronomers 894 EDWARD SINGLETON HOLDEN. had ever enjoyed. In particular, he gave them great Hberty of action. To quote from Newcomb's "Reminiscences," page 190: "The institu- tion made its mark almost from the beginning. I know of no example in the world in which young men, most of whom were beginners, attained such success as did those whom Holden collected around him." The evidences of Professor Holden's organizing ability and energy are written all over the Lick Observatory. His own scientific work in the Lick Observatory related principally to the photography of the Moon, but the administrative duties did not leave him much time for personal research. The last years of his administration were marred by the existence of animosities in the observatory community, and by much ill-advised criticism in the newspapers. The time has not come for any member of the staff in Professor Holden's administration to attempt a published discussion of the subject. From November 1901 until the time of his death. Professor Holden was Librarian of the U. S. Military Academy. In this position he was extremely successful. Many distinguished honors were conferred upon Professor Holden. He was elected Foreign Associate of the Royal Astronomical Society in 1884; a member of the National Academy of Sciences in 1885; and later to membership in the Astronomical Society of France, in the Italian Spectroscopic Society, in the American Academy of Arts and Sciences, etc. He received the degree of LL.D. from the Uni- versity of Wisconsin in 1886, and from Columbia University in 1887; the degree of Sc.D. from the University of the Pacific in 1896; and the degree of Litt.D. from Fordham College in 1910. Professor Holden's interests took a wide range. He has published: on the bastion system of fortifications; on studies in Central American picture writing; "The Mogul Emperors of Hindustan," a delightful volume, dated 1895; "Mountain Observatories" (1896); "Pacific Coast Earthquakes" (1898) the "Centennial History of the U. S. Military Academy, 1802-1902"; and several other volumes, as well as many popular and semi-popular magazine articles. Professor Holden died at West Point on March 16, 1914, where he was buried with military honors. The event marked the passing of a remarkably able and interesting man. • W. W. Campbell. SAMUEL WILLIAM JOHNSON. 895 SAMUEL WILLIAM JOHNSON (1830-1909). Fellow in Class I, Section 3, 1871. Samuel William Johnson was born in Kingsboro, New York, July 3, 1830, and died at his home in New Haven on July 21, 1909. He entered the Sheffield Scientific School of Yale University as an ad- vanced student of chemistry in 1850; later studying abroad under such masters as Erdmann, Liebig, Pettenkofer, and Von Kobel. In 1856 he began his career as Professor of Chemistry in the Sheffield Scientific School, teaching Analytical, Agricultural and Theoretical Chemistry, in all of which he displayed that clear and concise knowl- edge which makes the great teacher. It was, however, in the field of agricultural science that his greatest interest lay, and it was mainly thi'ough his efforts that the State of Connecticut gained the honor of inaugurating in this country the work of Agricultural Experiment Stations, which are now in successful operation in every State of the Union. The Connecticut Station, founded through his efforts, served for many years under his management as a striking illustration of the many ways in which chemical science can give aid to practical agriculture. In 1868 he published a book, "How Crops Grow," which attracted wide attention, and in 1870 this was followed by another volume, "How Crops Feed." They were the first books printed in this country bringing together in related form such knowl- edge as then existed regarding the composition and physiology of plants In reality they furnished a new basis for instruction in agriculture, and that their worth was clearly recognized is shown by the fact that they were translated into German, Russian, Swedish, and Japanese, and the former book into French and Italian also. In analytical chemistry, between the years 1864 and 1883, he edited three editions of Fresenius' Manual of Qualitative Chemical Analysis. Also, in 1870, he edited Fresenius' System of Instruction in Quantita- tive Analysis. His writings on chemical subjects were numerous, he being for more than fifty years a constant contributor to agricultural and scientific journals. As the writer has stated in another connection. Professor Johnson's broad and keen grasp of chemical problems, added to his farsighted appreciation of the many advantages to be gained by a judicious application of the science of chemistry to agriculture, made him a power in his generation, and his services counted for much in the 896 WILLIAM THOMSON, LORD KELVIN. development of agTicultural chemistry. Further, in the early years of the Sheffield Scientific School he was a pillar of strength, an example of the highest type of productive scholar, and a forceful illustration of the power which a scientific man can wield for the good of the community. The life of Samuel William Johnson and the work he accomplished constitute a suggestive example of a form of high public service which the man of scientific training can render his country and humanity. Russell H. Chittenden. WILLIAM THOMSON, LORD KELVIN (1824-1907) Foreign Honorary Member in Class I, Seclion 4, 1872. William Thomson was born in Belfast, Ireland, on June 26th, 1824. He was knighted in 1866 and was raised to the peerage in 1892 with the title of Baron Kelvin of Largs. His ancestors had migrated from Scotland in the seventeenth century. His father, James Thomson, removed from Belfast with his family to Glasgow, Scotland in 1832 to occupy the chair of Mathematics at Glasgow University. William matriculated at that institution in 1834, being then a little over ten years old. He was a precocious boy, rapidly advancing in his studies and early giving promise of that great mathematical genius and grasp of scientific problems, which especially characterized his later life. At sixteen he had, as a voluntary task, mastered Fourier's methods and theorems in the original French, this study occupying a fortnight only. He graduated from Cambridge University in 1845 taking rank as second wrangler, and shortly afterwards became Smith's Prizeman. Many evidences of his special talents were shown during his student years at Cambridge, in papers and discussions. Leaving Cambridge he studied in Paris and there came into contact with such leaders as Pouillet, Regnault, Liouville, Biot, Cauchy, Foucalt, Dumas, Pelouze and others. In 1845 he was granted a Fellowship with rooms at Cambridge. He had early shown a deep interest in electrical problems and theories, in which field he afterward found opportunity to do his most important work. This interest brought him into contact with Faraday. The WILLIAM THOMSON, LORD KELVIN. 897 SO called Kerr effect, which on Thomson's suggestion was sought for unsuccessfully by Faraday was not actually discovered till 1876. In 1846 Thomson, then only 22, was appointed to the Chair of Natural Philosophy in the University of Glasgow. This position he held for fifty years. There now began a most active period of his life and many of his students afterward became noted in physical science. Thomson's studies covered the widest range including Kinematics and Thermodynamics. Important discoveries, such as the Thomson thermoelectric effect were made, and inventions of the greatest scien- tific and technical value worked out. Among these latter may be mentioned the famous mirror galvanometer, of exceeding sensibility and almost essential to ocean telegraphy. There was also the well known water dropper leading to the so called Mouse Mill Replenisher. His great invention of the siphon recorder for cable telegraphy was made in 1869, and later his improved deep sea sounding apparatus and his mariner's compass improvements, aided greatly the art of navigation, in which he had always possessed a keen interest. The briefest record of his great activity in various fields is permissible here. In 1851, elected to the Royal Society, his papers on Thermo- dynamics brought him into contact with Helmholtz, Joule and later Clerk Maxwell. It happened that a few years prior to 1858, Thomson had made a study of the conditions of transmission of signals in long submarine cables, using the Fourier mathematics to determine the factors of retardation, etc. In 1858 the first Atlantic Cable was laid, and for a time it was successful ; failing after several hundred messages had been transmitted. Such success as there was must be attributed to Thomson, and the failure was probably caused by faults of insula- tion made worse by allowing the cable to be manipulated at high voltages and under conditions directly at variance with Thomson's ideas and disapproved by him. The great treatise by Thomson and Tait on Natural Philosophy was projected about this time. Part I however was not published until 1879 and Part II not until 1883. The authors thereafter abandoned the publication of the originally contemplated two addi- tional volumes. Following the London International Exhibition, Thomson took an active part in the committee work in 1862 involving the establishment of values and names for the international electrical units, such as the ohm, the volt and the farad. This pioneer work has been continued at the later international congresses and conferences, several of which he attended. 898 WILLIAM THOMSON, LORD KELVIN. The task of again laying an Atlantic Cable was taken up in 1865 with Thomson as the responsible electrical head. As is well known the 1865 cable broke in mid-ocean necessitating a new cable. This was laid in 1866 under Thomson's careful supervision aboard ship, and with constant tests as to the integrity of the cable being paid out. The result was completely successful and was at once followed by the picking up at sea, the splicing and completion of the broken 1865 cable; thus rendering available two cables instead of one. It was characteristic that Thomson had previously renounced any claim for time or services or even expenses, unless success was obtained. It was this success which earned him his Knighthood. His attention had not, however, been turned away from his purely scientific work. His papers on the age of the sun's heat and on the secular cooling of the Earth, brought on the famous controversy with the naturalists, principally the biologists and geologists; a controversy which was continued until at last the discovery of radio-activity had introduced new, and before unknown, factors, which had not entered into his calculations, and which virtually destroyed their applicability to the case. His studies on the rigidity of the Earth did much to negative the old assumption of a fluid interior supporting a solid crust. His papers dealing with his vortex theory of atoms, and papers and addresses on the ether of space were not the less noteworthy, though the views advanced he later renounced. In spite of his firm belief in the doctrines of creation and design in nature, he was not prevented from suggesting that the first germs of life on the earth might have been brought by a meteor from the outside. His practical judgment caused him to realize at once the great value of Bell's telephone in 1876, he being a first witness of its performance at the Centennial Exhibition. His appreciation of the future great service to be expected from the larger electrical applications was another evidence of his exceptional practical insight. This is exemplified in a strikingly prophetic statement made by him on Jan. 22, 1878, before the Institu- tion of Civil Engineers of Great Britain. Later Sir William Thomson contributed markedly to the art of exact electrical measurements in large work by his inventions of electrostatic voltmeters and particularly by his electrodynamic bal- ances, now known as Kelvin balances. He was author of the article on Electricity and also the four principal sections on Heat in the ninth edition of the Encyclopoedia Brittanica. Among many subjects investigated mathematically and experi- mentally by him were gyrostatics and wave motions, on which he became an authority. HENRY CHARLES LEA. 899 Sir William was raised to the peerage in 1892 and was afterward known as Lord Kelvin. His retirement from his professorship in the University took place in 1896 and was the occasion of a jubilee in his honor after fifty years of service. In this celebration many of the foremost scientific men of the time took part. He died Dec. 17, 1907, and his remains now rest in Westminster Abbey, next to the grave of Isaac Newton. A memorial window was later placed in the Abbey in commemoration of his life and work. He was twice married, but there were no children. In him were united the greatest gifts, a mathematical ability of the highest order, a deep scientific interest, an originality and industry most exceptional, and a personality most modest and attractive. His great talents were exercised not only in pure science, but were brought to bear with equal fruitfulness on difficult practical problems, and led to the con- ception and development by liim of many valuable inventions, to some of which allusion has been made in the foregoing brief account of his life. Elihu Thomson. HENRY CHARLES LEA (1825-1909) Fellow in Class III, Section 3, 1870. Henry Charles Lea, son of Isaac Lea and grandson of Mathew Carey, was born in Philadelphia in 1825 and died there 24 October 1909. Of precocious ability, he was educated at home and b}' private tutors, and at the age of eighteen entered his father's publishing house, with which he remained actively connected until 1880. He took an active part in the business and political life of his city and in public matters generally, but his dominant interests were those of a scholar and man of letters, manifested first in literary and scientific studies printed in his early youth, and later in the monumental series of books on mediaeval law and ecclesiastical history and institutions which began to flow from his pen in 1866. The eighteen solid volumes of his published works of history comprise: " Superstition and Force" (1866, fourth edition, 1892); "History of Sacerdotal Celibacy" (1867, enlarged edition, 1907); "Studies in Church History" (1869); "History of the Inquisition of the Middle Ages" (1888; also in a 900 FRANCIS CAUOT LOWELL. German and a French edition) ; " ( 'hapters from the Religious History of Spain" (1890); "A Formulary of the Papal Penitentiary" (1892); "A History of Auricular Confession and Indulgences in the Latin Church" (1896); "The Moriscoes of Spain" (1901); "A History of the Inquisition of Spain" (1906-1907, also translated into German); "The Inquisition in the Spanish Dependencies" (1908). Taken as a whole, Lea's works represent the most considerable contribution made to European history by an x\merican scholar. His interest in the past was institutional and scientific, rather than biographical or dramatic, so that his writings lacked the element of popular appeal and were more widely read and more highly valued in Europe than in the Ignited States. He dealt with highly contro- versial subjects, and not all his conclusions have won universal assent, but he commanded general respect for his candor and judgment, his untiring industry, and the extraordinary range, depth, and solidity of his learning. Biographical and critical accounts lui\e been printed in the Pro- ceedings of the American Philosophical Society, L (1911), and the Proceedings of the Massachusetts Historical Society, XLIII, 183-188. Charles H. Haskins. FRANCIS CABOT LOWELL (1855-1911) Fellow in Class III, Sccticn 1, 1898. Judge Francis Cabot Lowell was born in Boston on January 7, 1855, son of George Gardner Lowell, and grandson of the member of this Academy for whom he was named. In boyhood he showed some of the qualities which distinguished him in later life, for in physical size and intellectual grasp he developed early. To friends of his youth he seemed colossal in both, and the extent of his memory appeared well-nigh prodigious. He was easily at the head of his class, and Mr. Noble, the headmaster of the school, said many years later that he thought him on the whole the best scholar he had ever had. His interest in politics and his knowledge of public events, began at a very tender age, and he discussed these things while he was scarcely out of the nursery. When hardly more than twelve years old the FRANCIS CABOT LOWELL. 901 Struggle between President Johnson and Charles Sumner so fixed his attention that he formed an ambition to be some day a Senator of the United States — a youthful dream that might well have been gratified had he not accepted a seat on the bench. Most normal boys have visions of holding great places in the world, and this one would not be worth recalling were it not that it showed already a sense of reality, a capacity for observing political facts, rare at that age. In later years he maintained that, at the time of Johnson's administration, he was quite justified in thinking an influential Senator more important than the President. In boyhood he was so large as to outgrow to some extent his strength, and although he rode a horse, fished, and sailed a boat well, — and indeed was an excellent boatman throughout his life, — he took little part in the rougher competitive sports, such as baseball and football, which throw boys together. In 1872 he passed the examinations for admission to Harvard College with honors in every subject in which honors could be obtained; but on account of his health, which was believed to be delicate, he did not go to Cambridge until the next year when he joined his class as a sophomore. While he never complained of his fortune at any period of his life, he always had a feeling that this delay at the outset did not give him a quite fair start in his college career. In fact, it was a couple of years before his classmates appreciated his merit and his force; but in time these were recognized and in his senior year he was chosen to preside at the meeting called for the discordant busi- ness of choosing class officers. In scholarship he naturally ranked high, graduating with honors in History, and the confidence of the authorities in his ability was shown by his being selected to fill a sudden vacancy in writing a Commencement Part with only a few days' notice. After graduating he spent a 3'ear in travel over Europe, returning to enter the Law School, where he was from the outset one of the leading men. The beginning of active life in the world was not wholly promising, for he had not the push or the good luck to attract business. He was for eighteen 3'ears in practice at the bar, first with the writer of this memoir, and later with Frederic J. Stimson also, who joined the firm in 1891. In all the intercourse of the office he was the most considerate, generous and wise of partners; but the clients were not numerous, and in fact the amount of work was neither exacting nor highly remunerative. He was not, however, discouraged, for he had within himself other resources. In the early years of practice the 902 FRANCIS CABOT LOWELL. two partners wrote a book on " The Transfer of Stock in Corporations," but thereafter his spare time was turned into other channels, three leading interests, outside of office work, commanding his attention, History, Harvard University, and Politics. History he had always cared for, and the quality of his mind, his powerful memory, and his judicial temper fitted him peculiarly to pursue it. He had studied it much in College, and became fascinated by the life of the Maid of Orleans. To gathering everything published about her life and times he devoted much labor for several years, and then wrote his " Joan of Arc." Towards the end of his life he was again at work on history, but did not live to write what he had planned. Harvard had always been very dear to him. He felt for the Uni- versity the affection that comes from an inheritance of generations, and from associations beginning in childhood, as well as from spending five years within its walls. In 1886 he was elected an Overseer and, except for one year, he sat on the Board until he was chosen in 1895 a member of the Corporation, an office he retained until his death in 1911. The University has had few members of its Governing Boards who worked harder, none more constant in interest than he. The third avocation, if one may call it so, was politics — a revival, or since it was never lost, a fruition of the interest of boyhood. His attitude toward public life was exactly what one would like to see. His leading motive was neither personal ambition, nor a stern desire to fulfill a disagreeable duty. He enjoyed each position he filled, and looked forward to a more important one without grave disap- pointment when he failed to get it. He did his duty diligently, faithfully, and with pleasure, a stranger alike to the passionate eager- ness of the reformer and the self-seeking of a more common type. He began modestly and dutifully in the Common Council of the City of Boston, did what he could for honest administration there from 1889 to 1891, and later went into the legislature. Here was a field better suited to his talents. His character commanded respect, his ability won confidence, and he became Chairman of the Committee on the Judiciary and the leading figure in the House. Some results of his observation of men and methods at the State House he embodied in an article on "Legislative Shortcomings" ^ full of a clear perception of actual facts unfortunately too rare in our political writings; a quality which appears also in another study on " The American Boss." * Published as these were only in the ephemeral pages of a periodical 3 The Atlantic Monthly, March, 1897. 4 Ibid. Sept. 1900. FRANCIS CABOT LOWELL. 903 they have been lost from sight, and he never pursued the subject in a more permanent form. His service in the House covered the three years from 1895 to 1897. How much farther he might have gone one cannot say, for in 1898 he was offered and accepted the position of Judge in the United States District Court for Massachusetts. Some of his friends saw that his pohtical prospects were bright, and that by accepting an appointment to the bench he would renounce them; but he knew how uncertain they must be and felt that, with his practice at the bar no better than it stood, to refuse would mean to leave the law and to give up his profession altogether. For judicial work he was by intellect and temperament well adapted, and whatever he may have abandoned, the suitors and counsel in his court gained by his acceptance of the place. Promoted in 1905 to the Circuit Court, he spent on the federal bench the last thirteen years of his life. To the questions that came before him he brought the ability, the careful thoroughness, the large knowledge that he did to everything else he undertook, and it is notable that of his many decisions an unusually small proportion were over-ruled on appeal. From childhood he had qualities that made his life a singularly useful and happy one. What he once said of the grandfather for whom he was named might have been said of him, — that he had weak appetites with a strong will — a rare and fortunate combination. He seemed to embody the Greek idea of temperance, the possession of all good tendencies in moderation and none of them in excess. His political attitude was typical. Early in life, before he became active in politics, he was, at the election of 1884, a Mugwump; but ever after he was a consistent Republican. He saw, and made no attempt to conceal, and scarcely to excuse, the faults of his own party, yet he adhered steadily to it in office and out. This was charac- teristic. Clear in perception, just in opinion, he was a partisan with- out blindness and almost without prejudice, and that from a naturally contented disposition. He liked the men and the things with which he was associated — felt kindly toward them and was happy with them. The serenity of his disposition never failed in public or in private life. Even during the last two years of growing weakness he worked to the utmost of his strength, enjoyed cruising in his boat in the waters of Vineyard Sound and Narragansett Bay that he loved so well and never in sickness or in health, in success or in disappoint- ment, did he show the fretfulness that is the blight of modern life. A. Lawrence Lowell. 904 FREDERIC WILLIAM MAITLAND. FREDERIC WILLIAM MAITLAND (1850-1906). Foreign Honorary Member in Class III, Section 1, 1897. Frederic William Maitland was born in London 28 May 1850 and died at Las Palmas, Canaries, 19 December 1906. The grandson of Samuel R. Maitland, the historian of the "Dark Ages," he was educated at Eton and Trinity College, Cambridge, where he came under the influence of Henry Sidgwick and won high distinction in philosophy. He entered Lincoln's Inn in 1872 and was called to the bar in 1876. His interests, however, soon began to turn from the practice of law to its history, and in 1884 he was appointed Reader of English Law in the University of Cambridge, and in 1888 Downing Professor of the Laws of England, a chair which he held until his death. It is, however, characteristic of the English university system that the duties of his professorship consisted of general lectures to undergraduates on the elements of law rather than of the training of scholars in his special field, so that he formed no school of disciples who could develop or continue his work. His professorship, however, gave him considerable leisure for writing, and in spite of the ill health which soon drove him southward in the winter and finally cut him off in the fulness of his activity, he accomplished an astonishing amount of productive labor. It is a curious fact that Maitland owed to a Russian historian, Paul Vinogradoff, his introduction to the original records of English legal history. The acquaintance ripened speedily into his first important publication, a roll of " Pleas of the Crown for the County of Gloucester" in 1884, followed in 1887 by " Bracton's Note-book." Then came "Select Pleas of the Crown" (1888); "Select Pleas in Manorial Courts" (1889); "Three Rolls of the King's Court, 1194-5" (1891); 'Records of the Parliament of 1305" (1893); "The Mirror of Jus- tices" (1895); " Select Passages from Bracton and Azo" (1895); and the "Year Books of Edward II," as far as 1310 (1903-06). Merely as an editor of records and as the prime mover in inaugurating the publications of the Selden Society, he would hold a high place among those who have advanced the cause of English history. He shrank from no editorial labor, such as the difficult problems of the Law- French of the Year Books, but his introductions also show the wide learning, the luminous view, and the brilliant style which characterize all his writings. Besides these editions and a number of scattered FREDERIC WILLIAM MAITLAND. 905 essays, most of which have been brought together into the three volumes of his "Collected Papers," his most important works are "Domesday Book and Beyond" (1897); "Township and Borough" (1898); "Roman Canon Law in the Church of England" (1898); a translation of Gierke's " Political Theories of the Middle x\ges" (1900) ; a brilliant lecture on "English Law and the Renaissance" (1901); a posthimious set of lectures on "The Constitutional History of P]ngland" (1908); and the classic "History of English Law before the Time of Edward I" (1895), published conjointly with Sir Fred- erick Pollock but chiefly the work of Maitland. His last weeks in Cambridge were given to the "Life and Letters of Leslie Stephen." A full bibliography of his writings is appended to A. L. Smith's "Frederic William Maitland" (Oxford, 1908), where many charac- teristic passages are quoted. A biography, with a number of letters illustrating his style and the charm of his personality, has been pub- lished by Herbert Fisher (Cambridge, 1910). As an historian of English law Maitland has never been equalled. He was a finished jurist without the lawyer's reverence for form and authority; he combined the philosopher's power of analysis with the faculty of seeing everything in the concrete; and he had the delicate sense of evidence, the flashing insight, the vivid imagination, and the human sympathy of the great historian. To him the history of law was the history, not of forms, but of ideas; through it "the thoughts of men in the past must once more become thinkable to us." Yet law is not something abstract: its records "come from life," as he said of the Year Books, and must return to life. "English law is English history," he wrote; yet, first of English scholars, he saw it clearly against its Continental background. Unlike many jurists, howev^er, he did not seek to reduce the manifold complexities of life to a few general principles and to clarify what had never been clear; he avoided too definite conclusions and rather let his mind play about a subject in all its variety and illuminate it from different angles. To a masterly gift of exposition and a talent for apt illustration he joined a marvellous style, pointed, witty, epigrammatic, lighting up the dullest and most technical subject, and adorning everything it touched. Confining himself to the history of institutions and ideas, he did not enter the field of the narrative historian, so that the absence of a common standard renders comparisons difficult; but the quality of his mind justifies Lord Acton's judgment that he was " the ablest historian in England." Charles H. Haskins. 906 BENNETT HUBBARD NASH. BENNETT HUBBARD NASH (1834-1906) Fellow in Class III, Section 2, 1876. Bennett Hubbard Nash was born at Bloomingdale, now a part of New York City, July 6, 1834, the son of Joshua and Pauline (Tucker) Nash, and died at Little Boar's Head, New Hampshire, July 20, 1906. His early education was received in Europe. It is perhaps an indica- tion of the seriousness of his character even in boyhood that he joined the Waldensian Church at the age of fourteen in Turin. A younger brother was born at Florence, Italy, in 1836, and both brothers were members of the class of 1856 in Harvard College, and both attained high rank, being among the first ten scholars of that class at graduation, and, while neither seems at that time to have contemplated adopting the profession of teaching, both later became college professors. Bennett entered Andover Theological Seminary in 1856 and was graduated in 1860. In the spring of that year he had been licensed to preach, and in the next few years he did occasionally preach in Boston and elsewhere. In 1866 he was appointed instructor in Italian and Spanish in Harvard College, and he became assistant professor in 1871 and professor in 1881, a position which he held till the summer of 1894, when his resignation, presented in December, 1893, took effect. During the long period, broken by severe illness in the academic year 1872-73, which was covered by his college teaching his work was untiringly faithful and conscientious, and it is easy to understand that he carried the same faithfulness to duty in all details into every- thing that he undertook, particularly in his later years his manage- ment of the financial affairs of family connections. He was interested not only in languages and in literature, but also in music and other things, as is indicated by noticing the various organizations of which he was a member. Besides being a Fellow of this Academy he was connected with associations, clubs, and benevolent societies, including the American Philological Association, the Modern Language Asso- ciation of America, the Dante Society, the American Dialect Society, the Bostonian Society, the Harvard Musical Association, the Apollo Club of Boston, the St. Botolph Club, the Colonial Club of Cambridge, the University Club of Boston. He was married Feb. 19, 1861, in Boston to Mary Pratt Cooke (daughter of Josiah Parsons Cooke). E. S. Sheldon. JOHN ULRIC NEF. 907 JOHN ULRIC NEF (1862-1915) Fellow in Class I, Section 3, 1891. John Ulric Nef was born in Switzerland (Herisau) June 14, 1862, but came with his parents to the United States as a very young boy. He graduated from Harvard University in 1884, receiving the award of the Kirkland Traveling Fellowship, which gave him the opportunity to study chemistry at the University of Munich. In 1886 he received the degree of doctor of philosophy at Munich. His chief teacher was v. Baeyer and undoubtedly Baeyer's great work on unsaturated valences in carbon derivatives was decisive in leading Nef to make the study of the nature and activities of carbon valences his own life work. But he had the fine courage, the critical judgment and the tremendous capacity for ardent work — earmarks of his genius — to strike out on independent paths and develop his own lines of thought. This was shown already in his very first investigations (on the structure of benzoquinone), undertaken at Purdue University where he held his first chair in chemistry (1887-'89), and continued at Clark University ('89-'92). His fundamental researches on bivalent carbon (as proved particularly for the isocyanides and the fulminates) followed imme- diately after his work on the quinones; it was started during liis stay at Clark University and completed at the University of Chicago, to which he was called in 1892 and where he remained, as head of the department of chemistry, until his death on August 13, 1915. This brilliant series of investigations brought to him immediate and widest recognition as a bold and successful thinker : he became a fellow of the American Academy of Arts and Sciences in 1891, a member of the National Academy of Sciences in 1904 and a member of the Royal Society of Sciences of Upsala, Sweden, in 1903. He received the honorary LL.D. degree from the University of Pittsburgh in 1915. Dr. Nef applied the new line of thought, thus developed, to the study of the mechanism of many classes of reactions of organic com- pounds and he was engaged at the time of his death in an exhaustive investigation of the sugars from the point of view of bivalent carbon reactions. His ideas have been used and found stimulating and fruitful in the work of a great many other investigators and there is little doubt that his view of the existence of bivalent carbon in iso- cyanides, fulminates and similar derivatives is destined to form a permanent part of the theory of chemistry. Julius Stieglitz. 908 SIMON NEWCOMB. SIMON NEWCOMB (1835-1909) Fellow in Class I, Section 1, 1860. Simon Newcomb may justly be compared with the great leaders of commerce and finance who have risen up during the past half-century, and who have had so large an influence on the development of Ameri- can industry. That he was early drawn to astronomical science may be regarded as an accident of his career. As soon as he found himself in a position where his ideas could have free scope he set to work to condense and organize the cloud of facts and observations which was rapidly producing chaos in the astronomy of position. Different investigators were using different values for the fundamental constants of astronomy and the catalogues were affected by personal equations and instrumental errors; Newcomb sought to eliminate these and to combine observations from every source to produce values which would be generally accepted, or which could be made the basis for all future investigations. In this aim he was successful. Practically all investigations of the present day which lead to the determination of the constants of the star system, and to those of precession, nuta- tion, aberration, etc., are compared with Newcomb's results and hence are comparable with one another. This achieved, he set to work on a much longer task, that of pro- ducing new values and tables of the motions of the eight major planets of the solar system. He soon recognized that this was too large an undertaking for one man, even with a considerable force of computers at his disposal. For the most difficult portion, namely the theories of Jupiter and Saturn, he was fortunate in securing the services of the one man who had the requisite knowledge and ability to perform the task, G. W. Hill. All the rest of the work was carried on under Newcomb's supervision. When the construction of the theories from gravitation had been completed, he carried through a comparison of them with over sixty thousand meridian observations in order to determine with the highest possible accuracy the constants of their orbits. The tables embodying these results were then formed and are used in most of the national ephemerides. The chief attraction to him, however, was the motion of the Moon. By comparisons with early occultations and eclipses for the records of which he ransacked many European libraries, he discovered the great fluctuation from its theoretical orbit which the Moon possesses and which is yet awaiting an explanation. From the modern occulta- ALFRED NOBLE. 909 tions he showed that there were also minor differences concerning the origin of which we are equally in doubt. Although errors crept into both his theoretical and numerical work on this body it will always have great value for all students of the subject, if only on account of the great wealth of astronomical knowledge that Newcomb brought to bear on his discussions of its motion. Newcomb was in the service of the Na^^al Observatory nearly all his life. From 1877 to 1897 he was director of the Nautical Almanac, and for eleven years of this period he was also professor of mathematics and astronomy at Johns Hopkins University, and editing the American Journal of Mathematics. He wrote many papers on economics of considerable value, and was interested in a variety of kindred subjects. His astronomical textbooks have had a wide circulation. E. W. Brown. ALFRED NOBLE (1844-1914) Fellow in Class I, Section 4, 1913. Alfred Noble, commonly referred to as the "Dean of American Engineers," was born on a Michigan farm on Aug. 7, 1844, and he died April 19, 1914. His fame as an engineer was not confined to the United States, for the Institution of Civil Engineers of Great 13ritain in 1911, elected him an Honorary Member of that body. During his long, honorable and active career in his own country, he had already received the highest honors from American engineering societies and in 1910 he was awarded the John Fritz medal for "Notable achieve- ments as a Civil Engineer." In his early life, he was known as a diligent student, modest, faithful and industrious. The work on his father's farm undoubtedly de- veloped a sound physique which enabled him later in life to endure long and severe demands on his strength. On Aug. 9, 1862, Noble enlisted in the Twenty-fourth Michigan Volunteers, later a part of Hooker's " Iron Brigade," and all through the war he was called upon to take an active part in the movements of the Army of the Potomac; he was never wounded but, like many of his companions, came very near falling a victim to disease. In June, 1865, he was mustered out of the service, at which time he had risen to the rank of Sergeant; then followed two years' service in the 910 ALFRED NOBLE. Adjutant General's office in Washington and studies in preparation for a college course at Ann Arbor, which resulted in the degree of C. E. in 1870, a degree which brought honor alike to the recipient and to the College which bestowed it. Alfred Noble is always looked upon by the University of Michigan as one of its most distinguished sons. On ]\Iay 31, 1871, Mr. Noble married Miss Georgia Speechly, of Ann Arbor; then came long years of hard work in the profession of his adoption and the steady climb to positions of greater and greater responsibility. We may only glance at his successes in connection with the canal and lock at Sault St. Marie, the building of important bridges both in the East and in the West, and his opening of a private office in Chicago for carrying on a practice as Consulting Engineer. It was at this time, 1894, that he became more widely known as an Engineer of remarkable ability and in many different lines of work. At the same time Noble was called upon to design hydraulic works of original nature in connection with the regulating works of the Chicago Drainage Canal, and to give advice in structural steel work of the most difficult character. He served on the Nicaragua Canal Board, the U. S. Deep Water Commission and in 1899 he became a member of the Isthmian Canal Commission. He visited Europe, he studied all the problems which were submitted to him with the greatest care and fidelity and his reputation widened so that he was called in every direction to give advice on the construction of work of the greatest magnitude, such as the Barge Canal of New York, the locks and dams at Panama, the Thebes bridge over the Mississippi, the extensions of the Pennsylvania R. R., with the difficult problems relating to the tunnels under the Hudson, the N. Y. Water Supply from the Catskills and man}' other important works in the United States and Canada. His resources never seemed to fail ; they were the results of long years of close study and experience, united with remarkable natural gifts and a calm and placid disposition. Alfred Noble's career was a fine example of the great engineer, with many specialities united in one practitioner. It is given to few in any profession to excel in many directions. In the early history of the country this was more often possible than at the present time, but a study of Alfred Noble's life and his wonderful achievements must convince a careful inquirer that, with a firm grasp of the funda- mentals, and a wide practice in many branches, a splendid type of engineer is developed. Mr. Noble was elected a member of the American Academy on January 8, 1913. Desmond FitzGerald. JOHN MORSE ORDWAY. 911 JOHN MORSE ORDWAY (1823-1909) Fellow in Class I, Section 3, 1861. John Morse Ordway was born at Amesbury, Mass., April 23, 1823. He was graduated from Dartmouth College in 1844, and immediately took up manufacturing chemistry, including a wide variety of materials. In 1869 he was appointed Professor of Metallurgy and Industrial Chemistry at the Massachusetts Institute of Technology, where he remained until 1884. During this period he not only showed great zeal and devotion in his work as a teacher, but continued his close association with the technical field. About 1871 he designed and constructed furnaces for assaying, smelting and refining metals, in which these operations could be carried on in the laboratories of the Institute upon such a scale as to be of practical value, this being the first time that such work had been introduced into any educational institution. He was much interested in the promotion of education of women in science, and in 1876 a special chemical laboratory for women was established at the Institute through his endeavors. He directed the work of this laboratory in addition to other duties. He was elected an Associate Fellow of the iVcademy in 1861, was a member of the Council from 1878 to 1881, and a member of the Rum- ford Committee from 1871 to 1881. In 1884 he accepted a professorship in Tulane University, and was also Professor of Biology in H. Sophie Newcomb College, which is a part of Tulane University, which position he held until a short time before his death in 1909. He was an enthusiastic teacher, a stimulating lecturer, and a versatile worker, who both through his students and his varied re- searches contributed much to chemical science. H. P. Talbot. 912 CYRUS GUERNSEY PRINGLE. CYRUS GUERNSEY PRINGLE (1838-1911) Fellow in Class 11, Section 2, 1901. Cyrus Guernsey Pringle, who won distinction as a botanical explorer, was born in East Charlotte, Vermont, May 6th, 1838, and died at Burlington, Vermont, May 25th, 1911. By descent he was of both Scotch and Puritan stock. His early years were spent in the simple and rugged conditions prevailing in small agricultural communities of northern Vermont in the middle of last century. Beyond a partial secondary schooling his education was self-acquired. He never had the advantage of collegiate training, yet in maturity he wrote well, in a finished, if somewhat formal, English, was familiar with the ele- ments of Latin and Greek, and had a good command of Spanish, not to mention his technical knowledge of botany, plant-breeding, and horticulture. He was early associated with the Friends and, sym- pathizing deeply with their principles and faith, became a member of their Society. In July, 1863, he and two other F'riends from his community were drafted for service in the Federal Array. An uncle offered to pay the sum needful to secure a substitute, but this he refused, feeling that it would have been a concession to an abhorrent principle. Notwith- standing their firm remonstrance against taking an}- part in warfare, Pringle and his associates were taken to conscript camps first on Long Island in Boston Harbor and then to Culpepper, Virginia. They suffered many indignities even to bodily torture, were mild and kindly toward their tormentors, and yet remained absolutely fixed in their determination not to sacrifice their principles. Their case came to the attention of Stanton, the Secretary of War, and they were sum- moned to Washington, where they were treated with more considera- tion. At length, through the intervention of Isaac Newton, the Commissioner of Agriculture, their case was brought before Lincoln, who promptly ordered their release. A diary of more than 300 pages carefully kept by Pringle during these trying experiences not only possesses considerable historic significance regarding conscription of the period but is in itself a human document of high interest.^ During the next sixteen years Pringle's attention was largely devoted to the improvement of agricultural methods and especially 5 See Atlantic Monthly, cxi. 145-162 (1913). CYRUS GUERNSEY PRINGLE. 913 to experiments in plant-breeding. With natural aptitude, unlimited patience, and increasing skill, his achievements were notable, re- sulting in improved strains of wheat, oats, grapes, and potatoes. It has been estimated that these improvements effected by Pringle in crops of fundamental importance have together brought to the Ameri- can farmers an increase of profit amounting to several millions of dollars. He was also greatly interested in horticulture and at one time cultivated a very large number of different species of Lilium and Iris, and experimented extensively in hybridization. He also turned his attention to certain practical aspects of plant-surgery and plant- pathology. During all this time he acquired an increasing interest in the native flora of his region and, after making some admirably selected and carefull}^ prepared collections in Northern Vermont, extended his explorations to the White Mountains and to the region of the lower St. Lawrence. About 1880 an attack of inflammatory rheumatism warned him to seek occupation in a milder and drier climate. After consultation with Dr. Asa Gray, Prof. C. S. Sargent, and others, who recognizing his ability gave him aid, counsel, and several botanical commissions, he started on what became his special career and chief life-work, namely botanical exploration. His first journeys to California and Arizona were undertaken in a variety of interests — in connection with the forestry survc}^ embodied in the United States Tenth Census, to secure wood samples for the American Museum of Natural History, and to obtain specimens and data helpful to Dr. Gray, then writing his Synoptical Flora of North America. Having accomplished notable results in all these directions, Pringle was encouraged by Dr. Gray to enter the more difficidt field of Mexico. This he did in 1885 and from that time until the end of his life in 1911 he made more than twenty journeys to that country. He repeatedly penetrated its wildest regions^ climbed lofty mountains, explored numberless canons, traversed deserts, and cut his way through dense tropical jungles. Each year he brought back collections of astonishing extent and excellence. The difficulties and perils of his travels were great. Without means to equip expeditions of size, in which safety may to some extent be effected by numbers and cooperation, he travelled with methods of Spartan simplicity, often alone, rarely with more than one assistant. The nature of his work took him into the most primitive regions, where neither the native food nor lodging was endurable, and in 914 CYRUS GUERNSEY PRINGLE. consequence he frequently slept in the open and subsisted on the simplest kinds of food, which he could bring with him, as for instance crackers and Vermont cheese. He early suffered from tropical fevers, treated his disorders himself, sometimes with heroic doses of powerful medicines, and after a few years seems to have acquired a certain immunity from malarial diseases. On these trips he took with him, one or two at a time, no less than thirty-one assistants. Nearly all of them suffered from fevers, and he nursed them tenderly and brought them all safely back to their homes. In selecting these assistants, he tlid not seek men of scientific attainments but chiefly husky young men from farms and lumber camps, who could help him in the manual labor of collecting and drying plants under his direction and especially in transporting heavy presses and other equipment often for many miles. Incidentally, it was in certain regions the duty of the assistant to carry somewhat obviously in his belt or hip-pocket a large cavalry revolver. This seems to have been strictly for moral effect, and so far as has been learned it was never employed otherwise. Indeed, there is a suspicion that it was not always loaded. Happily, Pringle was able to establish exceedingly helpful relations of enduring friendship with several of the high officials in the govern- ment of Diaz. These gentlemen, some of them connected with the department of public health, were of great assistance in many ways, particularly in sending him from time to time telegraphic reports of the distribution of epidemic diseases in the republic, thus enabling him to avoid infected districts, a precaution of undoubted importance to which may have been due his success in escaping during a long series of years the more dangerous of the tropical diseases. Pringle's first collecting in Mexico was in the northwestern states of Sonora and Chihuahua, a natural extension of his work previously done in Arizona. P'or some years thereafter he devoted his chief attention to the states of Nuevo Leon, Coahuila, and Jalisco, then to San Luis Potosi, particularly to the diversified and botanically rich southern extremity of that state, a region not penetrated by Schaffner, Parry & Palmer, or others who had previously done notable explora- tion farther north in the same state. From about 1890 Pringle devoted himself largely to southern Mexico, working with considerable thoroughness over Michoacan, the State of Mexico, the Federal District, Morelos and Puebla. Then with great enthusiasm he continued his labors in Oaxaca, "the garden state," most wonderful of all. CYRUS GUERNSEY PRINGLE. 915 It is far easier to indicate the regions in Mexico where Pringle did not collect than where he did. In the following states he collected not at all or only to slight extent: Taraaulipas, the coastal portions of Vera Cruz, Tobasco, Campeachy, Yucatan, and Chiapas. His work extended to practically all the other states and most of them he explored repeatedly and with all feasible thoroughness. He was very methodical in his work. At the outset he collected about forty specimens of each kind selected. Later, as his reputation increased and he was able to secure wider sales for his duplicates, he augmented this number to fifty and at length to sixty specimens of a kind. Returning to his home in Vermont after a season in the field, he would make up his sets, submit one complete set to specialists for expert determination, care for the printing of labels, and ship his duplicate sets to the numerous subscribers both in this country and to nearly all the important herbaria in other parts of the world. Although earnestly devoted to his work, he was very humanly subject to a sort of rhythmic ebb and flow of enthusiasm. Each year in returning from his long and arduous journeyings, he would assure his friends that he could never return to Mexico, that he had ex- hausted the botany of the country so far as it was accessible to a man of his age and strength, that he believed he would devote himself to indoor work in his herbarium. After several winter months, devoted most happily to this indoor work, he would again become restless, would earnestly study the map of Mexico, and as soon as sufficient money returns accrued from the sales of his plants of the previous season, he would be off on his way back to Mexico with even more ambitious plans than any he had previously ventured upon. It has been estimated by a close friend and excellent botanist that Pringle collected in all more than 500,000 specimens of plants, repre- senting about 20,000 different species and varieties, of which about 12 per cent were new to science. From 1898 until his death Pringle was the officially appointed Collector for the Gray Herbarium of Harvard University, and it has been at that establishment that the greater part of his remarkably interesting collections of plants have been identified and the numerous novelties have been given scientific study and publication, though in this work several specialists elsewhere, notably at the Department of Agriculture and United States National Museum, have given much expert aid. From the beginning of his botanical work to the end of his fife Pringle was greatly interested in developing his own herbarium and, 916 CHARLES PICKERING PUTNAM. as he included in it an essentially complete series of his own extensive collections and added much material received by exchanges effected both in America and other parts of the world, both with professional botanists and amateurs, with museums, large herbaria, and with dealers, he gradually brought together a really notable herbarium, one of the best private botanical collections in any part of the world. In 1902 the University of Vermont acquired this herbarium under conditions most happily arranged for the comfort of Pringle himself. Not only was the collection given the safe housing he had long desired for it in the substantial biological building of the institution, but he was appointed permanently its curator, with an appropriation for its care. Rarely does a man whose life has been spent in exploration attain a position so congenial and so considerately calculated to permit him in advancing years to organize and correlate the results of his life work. From the University of Vermont Pringle received the honorary degrees, first of Master of Arts, and later of Doctor of Science.^ Benjamin Lincoln Robinson. CHARLES PICKERING PUTNAM (1844-1914) Fellow in Class II, Section 4, 1912. Charles Pickering Putnam, M.D., — well known for many years as a practitioner of medicine, but perhaps more widely known, yet not more warmly remembered, as a devoted worker on the broadest possible lines of social service,- — was born in Boston, September 15, 1844, and died, April 23, 1914, in his seventieth year. His parents were Charles Gideon Putnam and Elizabeth Cabot (Jackson) Putnam. His paternal grandfather was Samuel Putnam of Salem, a well-known and honored member of the Massachusetts Bar and for a long time a Justice of the Supreme Court of Massachu- setts. His maternal grandfather was Dr. James Jackson, of Boston, 6 For further details regarding Mr. Pringle's life, see the biographical sketches by Dr. Ezra Brainerd, Rhodora, xiii. 225-232 (1911); by C. R. Orcutt, Science, new ser. xxxiv. 176 (1911); and by Prof. George P. Burns, ibid. 750-751 (1911). CHARLES PICKERING PUTNAM. 917 one of the first and best among the founders of modern medicine in this commonwealth, who Hved long enough to see his grandchildren attain to a ripe age and to make them familiar with qualities of mind and heart which impressed them deeply. Judge Putnam's wife, Sarah Gooll, belonged to the distinguished Pickering family of Salem, while Dr. Jackson's wife was a member of the Cabot family, which was then eminent, as it has been since, for the public and private virtues of its representatives. Dr. Putnam graduated from Harvard College in 1865, and from the Harvard Medical School in 1869. After this he studied abroad, giving special attention to the diseases of children, and in the latter part of 1871 began to devote himself to his profession in Boston. Although he always carried on a general practice, he paid especial attention to pediatrics, and did some excellent pioneer work in ortho- pedics, then a branch of medicine that was but little known. In 1898 he was President of the American Pediatric Society. He lectured at the Harvard Medical School on the diseases of children from 1873 to 1875, and was clinical instructor in the same branch from 1875 to 1879. So kindly was his disposition, so full was he of sympathy with others whose lot had been harder than his own, so ready to be a worker and, where need was, a fighter for the embodiment of good principles in good institutions, that he found himself, — almost of necessity, and from the very outset, — plunging more and more deeply into social work. The history of his private life and medical labors need not be re- corded here. It is enough to say that it was made up of a never- ending series of acts of untiring devotion, prompted by the warmest of feelings and the highest sense of duty. As for his social service work, this was described so well by his relative Mr. Joseph Lee, in a paper first published in the Boston Med. and Surg. Journal for May 7, 1914, that I will complete this brief record by the following quotations from that source: "Dr. Putnam had been since the beginning of his practice of medi- cine a leader in charitable and social work,— almost from the begin- ning the most important leader of such work in Boston, the first to take hold and the last to let go of each new and important enterprise. Dr. Putnam was one of the founders, in 1873, of the little-known but extremely important Boston Society for the Relief of Destitute Mothers and Infants, which was a pioneer in establishing the policy of keeping mother and child together, and was president of the society 918 CHARLES PICKERING PUTNAM. from 1904 until his death. In 1875 he became physician to the Massachusetts Infant Asylum, and from 1898 to 1910 he was also president of the board of trustees. The ordinary death-rate in such institutions was at that time something over ninety per cent a year. The Massachusetts Infant Asylum had already brought the rate down to less than a quarter of that figure when Dr. Putnam became connected with it, and he by his skill and devotion again reduced it by two-thirds or more. He was one of those who in 1879 took part in the movement for establishing the Associated Charities, the second charity organization society in this country; and he was always one of the sustaining members of that society in the real, not the conven- tional, sense, working in many capacities, as president of a conference, as director, as chairman of many committees, including the present important one on inebriety, and, since 1907, as president. From 1892 to 1897 Dr. Putnam took a leading part in the very important movement for the reorganization of the Boston Institu- tions for the care of prisoners, of the poor, and of poor, neglected, and delinquent children, being on the special committee appointed by Mayor Matthews in 1892, chairman of the board of visitors of 1893-94, chairman of the standing committee on pauper institutions of the advisory board appointed by Mayor Quincy in 1896, a steady fighter for the reorganization bill of 1897. When the new system of separate unpaid boards of trustees was established he was appointed a member of the Board of Children's Institutions, and was its chairman from 1902 to 1911, performing in that capacity a great and harassing, though invisible and unappreciated, service to his fellow-citizens. ********* He was active in the campaign against tuberculosis and a director of the Mental Hygiene Association. He was one of the first to take up broad social questions from the legislative end, was the first experienced charity worker to enlist in the Massachusetts Civic League, and helped secure the establishment of the State Board of Insanity. ********** Dr. Putnam was for a generation the backbone of social work in Boston. We have all looked to him to do the hard things, — to take ujj the new line at which the timid balked and which the unimaginative could not see, sustaining the old from which the glamor had worn off, stiffening up the weak places, making the hard decisions. He was here, as in all things, a man to accept responsibility, take the burden CHARLES PICKERING PUTNAM. 919 'cm "himself, and carry it, — a patient and successful physician to the community a,s well as to the child. Dr. Putnam's most distinctive characteristic was the power of enlistment. In each of the many services he undertook it seemed to ttose he served and to his fellow workers as if that must be the only thing he had to do. There are in every enterprise the helpful men, the wise, the brilliant men, the steady workers. And then there are the essential men, those without whom the thing will not be done. In an extraordinary number of instances Dr. Putnam was among these last. Whatever happened, however badly things might go, whoever else became lukewarm or discouraged, his associates knew that he, at least, would see the thing through, that he had enlisted for the war, intended doing as much, be it more or less, as might be necessary. Dr. Putnam was a remarkably resourceful man and would recon- struct his patient's world, physically as well as morally, by his calm assumption that anything needed could be done, and in hundreds of cases by doing the most impossible parts himself. Slower minds thought him slow at laying the first brick, whereas he had completed the whole structure in imagination, and was hesitating what kind of chimney-pot to use. And the best was the power behind it all in the great kind heart, that would see and know only the best, and, with a quality like the sun, could see only light wherever it was turned." Dr. Putnam's wife, Lucy Washburn, and three children, Charles Washburn, Tracy Jackson, and Martha, survive him. J. J. Putnam. ■920 FREDERIC WARD PUTNAM. FREDERIC WARD PUTNAM (1839-1915) Fellow in Class III, Section 2, 18G5. Frederic Ward Putnam, the son of Ebenezer and Elizabeth (Apple- ton) Putnam was born in Salem, Massachusetts, April 16, 1839, and died in Cambridge, August 14, 1915. Although his father, grand- father and great-grandfather were all graduates of Harvard College, he planned to take up a military career, having been promised an appointment at West Point. Through the influence of Louis Agassiz, however, his attention was turned toward the study of Natural History, and he accordingly entered the Lawrence Scientific School in 1856, graduating in the class of 1862. For some years he devoted himself to work in Natural History. In 1875, however, he was made Curator of the Peabody Museum of American Archaeology and Ethnology at Harvard University, and from that time on, his main energies and interests were centered on archaeology and anthropology. He applied himself to the building up of the museum collections by explorations in the field; to the training and teaching of investigators and students; and to the establishment and development of anthropological institutions throughout the country. His great services in all of these lines have led to his being regarded as one of the founders of anthropology in America. The professional and honorary positions held by Professor Putnam have been as follows; — Curator of Ornithology, Essex Institute, Salem, 1856-64; Assistant to Professor Louis Agassiz, Harvard LTniversity, 1857-64; Curator of Vertebrates, Essex Institute, Salem, 1864-66; Superintendent, Museum of the Essex Institute, 1866-71; Superintendent, Museum of the East Indian Marine Society, Salem, 1867-69; Director, Museum of the Peabody Academy of Science, Salem, 1869-73; Curator of Icthyology, Boston Society of Natural History, 1859-68; Permanent Secretary, American Association for the Advancement of Science, 1873-98; Assistant, Kentucky Geologi- cal Survey, 1874; Instructor, Penikese School of Natural History, 1874; Assistant, United States Engineers in Surveys West of the 100° Meridian, 1876-79; Assistant in Icthyology, Museum of Com- parative Zoology, 1876-78; Curator of the Peabody Museum of American Archaeology and Ethnology, 1875-1909; Honorary Curator, 1909; Honorary Director, 1909-1915; Peabody Professor of American FERDINAND, FREIHERR VON RICHTHOFEN. 921 Archaeology and Ethnology, Harvard University, 1886-1909; Emeri- tus; 1909-1915; State Commissioner of Fish and Game, Massa- chusetts, 1882-1889; Chief, Department of Ethnology, World's Columbian Exposition, 1891-1894; Curator of Anthropology, Ameri- can Museum of Natural History, New York, 1894-1903; Professor of Anthropology and Director of the Anthropological Museum of the University of California, 1903-1909; Emeritus, 1909. Vice-President, Essex Institute, 1871-94; Boston Society of Natural History, 1880-87, and President, 1887-89; President, American Folklore Society, 1891 ; of the Boston Branch of the Society, 1890-1915; President, American Association for the x\dvancement of Science, 1898; Vice-President, Numismatic and Antiquarian Society of Philadelphia, 1896-1915; Vice-President for the United States, International Congress of Americanists, New York 1902; Chairman, Division of Anthropology of the International Congress of Arts and Sciences, St. Louis, 1904; President, American Anthropological Association, 1905-6. Professor Putnam received the following honorary degrees ; — from Williams College, 1868, A.M.; University of Pennsylvania, 1894, S.D. From the French Government he received the Cross of the Legion of Honor in 1896. R. B. Dixon. FERDINAND, FREIHERR VON RICHTHOFEN (1833-1905) Foreign Honorary Member in Class II, Section 1, 1901. Ferdinand, Freiherr von Richthofen, Professor of Physical Geog- raphy in the University of Berlin, and eminent both as a geol- ogist and geographer, died on October 6, 1905, in his 73d year. He was best known for his researches in China, and especially for his solution of the problem of the origin of loess. Baron von Richthofen was born of a distinguished Silesian family at Karlsruhe on May 5, 1833. His early education was received in his native town and later in Breslau where he first studied geology. Two years of university studies in Breslau did not greatly interest him, so in 1852 he went to Berlin where he came in contact with Beyiich and with Karl Ritter. While at the LTniversity, working 922 FERDINAND, FREIHERR VON RICHTHOFEN. under the inspiration of Ritter, he became acquainted with a man who had travelled in the Orient and thus attained his first interest in that portion of the world in which he travelled later. Von Richt- hofen's thesis for the doctorate was entitled " Ueber den Melaphyr," and was published in 1856. From this time until 1860 he was engaged in geological surveys in southern Tyrol and in the Carpathian Moun- tains. His writings during this period dealt principally with the dolomites of the Tyrol, which he concluded were formed as coral reefs, and with precious metal deposits associated with Carpathian volcanics. In 1860, with the rank of a Legation secretary, he joined a Prussian expedition under the leadership of Count Eulenburg, to visit Siam, China, and Japan. He left the expedition in Siam and proceeded to Ceylon where he showed that laterite could be derived from the decomposition of many kinds of rocks. His travels took him to Java, the Celebes, the Philippines, Formosa, Japan, China, and Burma, and in 1862 he went to California. Six years were spent in Cali- fornia and Nevada, the results of the work being his memoirs on the Comstock Lode (1866) and his "Natural System of Volcanic Rocks" (1868). The next four years were spent in the exploration of China, gather- ing material for his monumental work, "China." He discovered the vast coal fields of Shantung, thereby emphasizing the economic importance of China, and he advised his government to select Kiao Chau as a naval base for the Far East. In 1872 he returned to Germany to work up the results of his travels, and three years later the University of Bonn elected him to the chair of geography, but permitted him to complete the first part of the work on China before assuming his professorship in 1879. From Bonn he went to Leipzig in 1883, and to Berlin in 1886. At the University of Berlin he taught until the end of his life. Richthofen's gieatest work, "China," appeared in five volumes, the first being issued in 1877, the last in 1912. In the first volume there is an elaborate history of China and a statement of his evidence con- cerning the origin of loess. He held that loess was an eolian deposit formed on broad, open steppes, contrary to the belief of Pumpelly that the loeiss of Mongolia was deposited in lakes. These conclusions have since been confirmed by the discover}' of similar deposits of loess in Central Europe, Germany, and the Mississippi Valley, asso- ciated with a steppe fauna. In 1907 Richthcfen's "Tagebiicher aus China," in two volumes, was published. SIR HENRY ROSCOE. 92S Petrology was the branch of geology in which von Richthofen was first interested, and in his work on melaphyre he endeavored to systematize the science by adapting a classification of rock-species in a scheme similar to that used in systematic zoology. His complete classification was published in his "Natural System of Volcanic Rocks." The faults of his plan to separate similar rocks because of different ages reflect on the conceptions of geology in Germany at the time rather than on his appreciation of the facts. Two other works deserve mention: his "Fiihrer fiir Foischungs- reisende" (1886), and his " Geognostische Beschreibung von Pre- dazzo" (1860). In the former he emphasized the desirability of pre- cision in geographical description; in the latter he maintained that the dolomites of the South Tyrol were coial reef deposits and that the associated beds were contemporary deep-sea or coral-sand deposits. He re-asserted his conclusions in 1874, after a study of living Malay- sian reefs. Sidney Powers, at the request of R. A. Daly. SIR HENRY ROSCOE (1833-1915) Foreign Honorary Member in Class I, Section 3, 1890. The Rt. Hon. Sir Henry Enfield Roscoe died suddenly at his resi- dence near London December 18th, 1915. Born January 7th, 1833, he had nearly completed his 83rd year. Up to within an hour of his death he was in good health and spirits. Then "heart failure" put an end to his interesting and useful life. He was sent to University College, London, in 1848. Graham was then the professor of chemistry there, and he was succeeded by Wil- liamson during Roscoe's student years. After receiving the degree of B. A. from the University of London he went to Heidelberg to continue his studies under Bunsen. Here he took part in the well- known researches on the chemical action of light. He continued work in this field for some years after receiving the degree of Doctor of Philosophy from Heidelberg in 1854, returning to Heidelberg in the summer vacations for this purpose. In Ostwald's reprint in his Collection of Scientific Classics, he says: "The Photochemical Re- 924 SIR HENRY ROSCOE. searches of Bunsen and Roscoe deserve the name of a classical in- vestigation as they not only have gathered together all points known hitherto on the subject, but by their wide and thorough experiments have laid the foundation for all future work on the subject. It cannot be doubted that these researches not only serve as a classical, but as the classical type for all future works on the subject of physical chemistry. Roscoe's residence in Heidelberg and his association with Bunsen and with Kirchhoff naturally led him to take special interest in spectrum analysis. He translated the book of Bunsen and Kirchhoff and also lectured extensively on this subject. In 1857 he was elected professor of chemistry in Owens College, Manchester, succeeding Franklin. The College had then been in existence only six years, and it was in a low state financially. In his "Life and Experiences," published in 1906, he says: "The institution was at that time nearly in a state of collapse, and this fact had im- pressed itself even on the professors. I was standing one evening preparing myself for my lecture by smoking a cigar at the back gate of the building, when a tramp accosted me and asked me if this was the Manchester Night Asylum. I replied that it was not, but that if he would call again in six months he might find lodging there! That this opinion as to the future of the college was also generally prevalent is shown by the fact that the tenancy of a house in Dover Street was actually refused to me when the landlord learned that I was a professor in that institution." Roscoe's principal work for chemistry was done while holding the professorship in Owens College, which he resigned in 1885 to enter Parliament, after a continuous service of 28 years. During this period the college was completely transformed and in 1880 a Royal Charter constituting the Victoria University was granted. Of this new uni- versity Owens College was a part, the other parts being the University College of Liverpool and the Yorkshire College of Leeds. Roscoe undoubtedly rendered his country a great service in im- proving the methods of teaching chemistry. In this work he was ably seconded by Carl Schorlemmer, a German, who was thoroughly imbued with the spirit of the German universities. He held the chair of Organic Chemistry and distinguished himself by his researches and by his literary work. Roscoe had great admiration for the German methods and he did what he could to give his country the benefit of these methods. Roscoe's contributions to chemistry are not numerous. Probably HENRY CLIFTON SORBY. 925 the most important is that on vanadium. He says: "This is cer- tainly the best piece of work I ever did and I do not know that I ever enjoyed anything of an intellectual kind more thoroughly." * * * "The subject aroused very general attention throughout the scientific world, and my view concerning the relationships of the metal was universally adopted." His "Lessons in Elementary Chemistry" was published in 1866. Up to 1906 the number of copies sold was 211,000. Translations appeared in German, Russian, Italian, Hungarian, Polish, Swedish, Japanese, and even in one of the Indian vernaculars, Urdu. His primer also had a wide sale. The well-known " Treatise on Chemistry " by Roscoe and Schorlemmer is unquestionably the best English treatise on chemistry and certainly one of the best in any language. Roscoe was elected to Parliament in 1885 and then resigned his professorship. Since that time his activities have not been displayed especially in the field of Chemistry. He continued, however, to exert a strong influence upon educational matters. After his retirement from Parliament he and his wife lived in what he called "the most beautiful and healthy spot in the whole of Surrey, namely, at Wood- cote Lodge." It was here that his life ended suddenly and ideally. Ira Remsen. HENRY CLIFTON SORBY (1826-1908) Foreign Honorary Member in Class II, Section 1, 1892. Henry Clifton Sorby, who, in an address sent him in 1907 by many leading petrographers, was called "The Father of Microscopical Petrography" was born in 1826 and died in 1908 in his eighty-second year. From the age of 20 his long life was devoted to scientific re- search, mainly in geology, petrography and mineralogy, unhampered by any professional duties and assisted by a moderate but sufficient competence. He prepared in 1849 what was probably the first transparent microscopic section of a rock, and in 1850 his first paper illustrating the new method by the study of sedimentary rocks was published. This attracted little attention, and it was eight years later that the classic memoir on which his fame as a pioneer principally 926 HENRY CLIFTON SORBY. rests was published in the Quarterly Journal Geological Society for 1858 (read Dec. 1857) under the title: "On the Microscopical Struc- ture of Crystals, indicating the Origin of Mineral sand Rocks" — a memoir which "though neglected, and even ridiculed at the time, laid the foundations of microscopical petrography — and, in doing so, widened enormously the sphere of influence of mineralogy from which it borrowed so much." In the introduction the author summarizes his paper as an attempt to prove that artificial and natural crystalline substances possess microscopic characteristics by which their origin from either aqueous solution or igneous melt can be deduced, and in some cases an approxi- mation made to their rate of formation and to the accompanying- temperature and pressure. The 47 pages of the memoir contain detailed descriptions of many microscopic preparations, illustra,ted by drawings, an account of the methods of preparation, and a dis- cussion of the theoretical application. The further development of the new method was at first by German mineralogists, beginning with Dr. Ferdinand Zirkel, with whom Sorby had travelled in 1862 soon becoming an enthusiastic disciple. Zirkel's " Mikroskopische Gesteinstudie" (1863) brought the subject into wide notice and its further development was rapid. Thus, while Dr. Sorby gave the start to the new science, he left its further development to others, devoting himself to many other problems of geological science, and to other sciences, so that almost 250 papers represented his scientific activity, as continued over sixty years, and until his death. He was elected Foreign Honorary Mem- ber of this Academy in 1892. A detailed account of his life and activities has been given by the late Professor J. W. Judd in the Geological Magazine for 1908 and in the Mineralogical Magazine for the same year, from which the present memoir is mainly an abstract. John E. Wolff. EDUARD STRASBURGER. 927 EDUARD STRASBURGER (1844-1912) Foreign Honorary Member in Class II, Section 2, 1892. Eduard Strasburger was born in Warsaw in 1844. In 1862-1864 he studied at the Sorbonne. He then went to Bonn, where he studied under the direction of Hermann Schacht, from whom he acquired the great manual dexterity which was necessary for the pre- paration of microscopical specimens before the era of the microtome. In Bonn he came under the influence of the great teacher Sachs. Subsequently he studied under N. Pringsheim at Jena and received the Doctor's degree in 1866. In 1868 he began his career as a teacher at the University of Warsaw. In the following year he was called to Jena, as extraordinarius, largely through the influence of Haeckel, of whom he was an enthusiastic follower. In 1871 he became ordi- narius at Jena. In Jena he married Alexandrine Wertheim of Warsaw. Here were born his two children, a daughter and a son. In 1880 he accepted a call to Bonn, where he remained until his death in May, 1912. Strasburger had to an unusual degree the power to arouse and inspire his pupils. As a rule his laboratory had a large quota of students, of whom the majority were foreigners. His custom was to visit each student on his way to the morning lecture and it was astonishing to see how quickly he would grasp each new situation and with a few words outline the next step to be taken. His suggestions and his example never failed to stimulate the student to do his best. His insight into character and his remarkable tact accounted in no small measure for his influence upon his pupils. His command of technique was unrivalled. He originated a large number of the methods used in cytology. When he began his work such methods were almost unknown: indeed it is commonly stated that he was the first to use material fLxed and hardened in alcohol. To the very end of his life he was busily engaged in improving cytologi- cal methods. His Botanisches Practicum is a veritable encyclo- paedia of technique and will always remain a model of its kind. It must not be supposed, however, that his skill in technique was his principal asset. He had a keen analytical mind which delighted in profound generalizations. He became interested in nearly all the great botanical problems of his day and to most of them he made contributions of lasting value. He followed with remarkable eager- 928 CHARLES HALLET WING. ness and enthusiasm the progress of discovery in other branches of science and from them he gleaned materials for the enrichment of his own field of labor. His writing was characterized by a lucid and attractive style and he was a lecturer of unusual skill. Owing to the great variety of subjects which he investigated it is quite impossible to give in a few words an adequate summary of his contributions. They have to do in the main with the fundamentals of morphology and especially with cytology. Beginning with the great problem of homologizing the reproductive structures of phan- erogams and cryptogams, he soon became deeply interested in the finer details of the mechanism of heredity; he investigated most thoroughly the structure and behavior of the chromosomes, and this led eventually to the study of a great variety of problems in the morphology and physiology of the cell. His contributions in this field are classical and constitute a large part of the foundation of our present conception of the cell and of the mechanism of heredity. W. J. V. OSTERHOUT. CHARLES HALLET WING (1836-1915) Fellow in Class I, Section 3, 1874. Charles Hallet Wing, son of Benjamin Franklin and Adeline (Hallet) Wing, was born in Boston, August 5, 1836, and died suddenly at his home in Boston on September 13, 1915. He was a graduate of the Lawrence Scientific School of Harvard, and later was ap- pointed Professor of Chemistry at Cornell University in 1870, re- signing that position in 1874 to accept the Professorship of Analytical Chemistry at the Massachusetts Institute of Technology. Under his direction the Kidder Chemical Laboratories were planned and built, in which were introduced many new features, making them models for their time. As a teacher he demanded thoroughness, high stand- ards of scholarship and strict attention to work, and he held the respect and confidence of his pupils. In 1884 Professor Wing resigned his position at the Institute, and soon after took up his residence at Ledger, Mitchell Co., N. C, where EDWARD STICKNEY WOOD. 929 among the mountaineers he devoted his entire energies to the better- ment of his neighbors. He built and furnished a school-house, and provided teachers. In connection with this building, equipment for manual training was provided, and much of the instruction in this branch was given by Professor Wing himself. He started the first free public library in North Carolina, the books for which were donated either by Professor Wing himself or through his personal solicitation. When his physical strength began to fail he donated the school and library to the county, and returned to Boston. He was a lover of music and a student of languages, and in the society of his friends his last years were passed busily and happily. Professor Wing was elected a Fellow of the Academy in 1874, and retained his fellowship until the time of his death. H. P. Talbot. EDWARD STICKNEY WOOD (1846-1905) Fellow in Class I, Section 3, 1879. Edward Stickney Wood was born on April 28, 1846, and died July 11, 1905. He was the son of Alfred Wood and Laura Stickney Wood of Cambridge. After graduating from the Cambridge High School, he entered Harvard in 1863 and graduated with the class of 1867. He then entered the Harvard Medical School and completed his studies there in 1870, although he did not at that time receive his degree in medicine. After a period in the Marine Hospital in Chelsea, he entered the Massachusetts General Hospital and served there for a year as surgical house pupil. A vacancy occurring in the chemical de- partment gave Dr. Wood his opportunity. Therefore after some time spent in Europe in the study of this branch of medical science he was appointed assistant professor and in 1876 he succeeded to the full professorship of Chemistry, which position he held until the time of his death. His career as a teacher was attended with great success, his lectures being eminently practicable and intelligible, developing the enthusiasm which always attends the work of a popular lecturer. His judgment and advice were held in high esteem by the members of the faculty and D30 JOHN HENRY WRIGHT. the President of the University. This was due to his cahii and sym- pathetic temperament and good common-sense. He occupied the position of leading medical expert in medico-legal cases all over New England and his record as a witness in many important trials won him universal regard in both the medical and legal profession. Dr. Wood also found time for many articles in current medical literature, including addresses to medical bodies and participation in the discussions of the medical societies. With Robert Amory he translated the work of Neubauer and Vogel on "Urinary Analysis" and revised that portion of Wharton and Stille on "Medical Juris- prudence" so far as his special line was concerned. In 1874 he was made a member of the commission which investigated the sanitary condition of the water of the Sudbury, Charles and other rivers. He also prepared in 1894 an important paper for the State Board of Health about "Arsenic in its relation to Domestic Life." He was a member of the American Pharmaceutical Association, the Massachusetts Medico Legal Society, the Boston Society for Medical Improvement and the Massachusetts Medical Society. He was also a member of the committee for the revision of the " Phar- macopoeia" in 1880 and chemist to the Massachusetts General Hospital. He was a man of the most sterling character and much beloved and respected by all who came in contact with him. J. Collins Warren. JOHN HENRY WRIGHT (1852-1908) Fellow in Class III, Section 2, 1893. John Henry Wright, Professor of Greek and Dean of the Graduate School of Arts and Sciences in Harvard University, died at Cambridge, Massachusetts, Nov. 25th, 1908. He was born at Urumyat, Persia, Feb. 4, 1852, the son of Rev. Austin Hazen Wright, a missionary in Persia from 1840 to 1865. He graduated from Dartmouth College in 1873, and became Asst. Professor of Ancient Languages (Greek and Latin) at what is now Ohio State University. In 1876 he re- JOHN HENRY WRIGHT. 931 ceived the degree of Master of Arts from Dartmouth, and after study- ing classical philology and Sanskrit at Leipzig for two years, he re- turned to Dartmouth as Associate Professor of Greek, a position which he held until 1886. He married, April 2, 1879, Mary, the daughter of President Eli Todd Tappan of Kenyon College. In 1886 he became Professor of Classical Philology and Dean of the Collegiate Board at Johns Hopkins University, whence he was called to Harvard in 1887. He went to Athens, Greece, as annual professor at the American School of Classical Studies in 1906-7. He became a Fellow of the American Academy of Arts and Sciences in 1893, was a member of the Council of the Archaeological Institute of America, and President of the American Philological ^Association in 1894. Besides being editor-in- chief of the American Journal of Archaeology from its re-organization in 1897 until 1906, he was an associate editor of the Classical Review, 1888-1906, and of its successor the Classical Quarterly from 1907. He was active in the interests of the New England Classical Associa- tion, presiding with wit and tact at its first meeting at Springfield in April, 1906. His many monographs and contributions to classical learning can- not all be listed here, but the following are the more important works by which he made himself known to scholars the world over: 1882. An address on The Place of Original Research in Collegiate Education. 1886. The College in the University and Classical Philology in the College. 1886. Translation of Collignon's Manual of Greek Archaeology. 1892. The Date of Cylon. Although not printed until after the dis- covery of Aristotle's work on the Constitution of Athens, it anticipated correctly the chronology given in that work. 1893. Herondaea. Valuable studies on the newly discovered papyrus of Herondas. 1894. Studies in Sophocles. 1902. Edited " Masterpieces of Greek Literature." 1904. Present Problems of the History of Classical Literature, an address delivered at the Congress of Arts and Sciences at the St. Louis Exposition. 1905. Editor of "A History of All Nations from the Earliest times," in 24 volumes. Based in part on the Allgemeine Weltgeschichte of T. Flathe and others. 1906. The origin of Plato's cave. In addition, his work as one of the editors of the classical section of the Twentieth Centurv Textbooks should be mentioned. 932 JOHN HENRY WRIGHT. These brief biographical details reflect but little of the eminent personal qualities which Professor Wright possessed, but they indicate the wide variety and extent of his training and experience, and explain the beneficent influence which he exerted on so many of the younger scholars of this country. The traits that immediately impressed a student on his first meeting with him were his generosity, sympathy, and learning. His kindness was unfailing, his courtesy never shaken. Possessing a keen sense of humor, he was merciful to the blunderer, ready to overlook the crudeness and awkwardness of the tyro, but equally firm in correcting the puerile and in rebuking the insincere. His own scholarship was fertile, whether expressed in his writings or in the work which he inspired in others. Broad in its range, it was deep in its thoroughness. Versed as he was in the technical minutiae of those branches of classical philology in which he was a specialist, he had the gift of imparting human interest and a literary quality to his exposition of scientific subjects. As a writer his style had charm, so that the study of a problem in Greek epigraphy, for example, became in his hands not only a work of scholarly importance, but also a matter of interest to a reader not trained in technicalities. He had a rare insight into the beauties of English, and his taste guided him surely in the interpretation of the subtleties and graces of Greek style. His sense of form expressed itself also in his love of Greek art, and he found congenial labor in the editorship of the Journal of Archaeology. Through his edition of Collignon's Manual and the courses which he offered in the University on the subject he became the pioneer in the teaching of classical archaeology in this country. Charles Burton Gulick. Class I. Elihu Thomson, Class III. George F. Moore. American Academy of Arts and Sciences OFFICERS AND COMMITTEES FOR 1916-17. PRESIDENT. Hexry p. Walcott. VICE -PRESIDENTS . Class II. William M. Davis, CORRESPONDING SECRETARY. Harry W. Tyler. RECORDING SECRETARY. Wm. Sturgis Bigelow. TREASURER. Henry H. Edes. LIBRARIAN. Arthur G. ^^'EBSTER. COUNCILLORS. Class II. John Collins Warren, Terms expire 1917. Alfred C. Lane, Terms expire 1918. Benjamin L. Robinson, Terms expire 1919. William M. Wheeler, Terms expire 1920. COMMITTEE OF FINANCE. John Trowbridge, George V. Leverett, RUMFORD COMMITTEE. Charles R. Cross, Chairman, Edward C. Pickering, Arthur G. Webster Class I. Desmond FitzGerald, Frederick S. Woods, Harvey N. Davis, Gregory P. Baxter, Henry P. Walcott, Class III. Mark A. DeW. Howe. Samuel Williston. Fred X. Robinson. Archibald C. Coolidge. Louis Bell, Walter L. Jennings, Arthur A. Noyes, Arthur A. Noyes, C. M. WARREN COMMITTEE. Henry P. Talbot, Chairman. Elihu Thomson, Theodore Lyman. Gregory P. Baxter, William H. Walker. Charles L. Jackson, Arthur D. Little, COMMITTEE OF PUBLICATION. Edward V. Huntington, of Class I, Chairman, Walter B. Cannon, of Class II, Albert A. Howard, of Class III. COMMITTEE ON THE LIBRARY. Arthur G. Webster, Chairman, Harry M. Goodwin, of Class I, Samuel Henshaw, of Class II, William C. Lane, of Class III. AUDITING COMMITTEE. George R. Agassiz, John E. Thayer. HOUSE COMMITTEE. Louis Derr, Hammond V. Hayes, Chairman, Wm. Sturgis Bigelow. COMMITTEE ON MEETINGS. The President, The Recording Secretary, , William M. Davis, Edwin B. Wilson, George F. Moore. LIST OF THE FELLOWS AND FOREIGN HONORARY MEMBERS. (Corrected to July 1, 1916.) FELLOWS.— 485. (Number limited to six himdred.) Class L — Mathematical and Physical Sciences. — 182. Section I. — Mathematics and Astronomy. — 40. George Russell Agassiz Boston Solon Irving Bailey Cambridge Edward Emerson Barnard Williams Bay, Wis. George David Birkhoff Cambridge Charles Leonard Bouton Cambridge Ernest William Brown New Haven, Conn. Sherburne Wesley Burnham Williams Bay, Wis. William Elwood Byerly Cambridge William Wallace Campbell Mt. Hamilton, Cal. Julian Lowell Coolidge Cambridge George Cary Comstock Madison, Wis. Leonard Eugene Dickson Chicago, 111. Philip Fox Evanston, 111. Fabian Franklin New York, N. Y. Edwin Brant Frost Williams Bay, Wis. Frank Lauren Hitchcock Boston Edward Vermilye Huntington Cambridge Dunham Jackson Cambridge Edward Skinner King Cambridge Carl Otto Lampland Flagstaff, Ariz. Percival Lowell Boston 936 FELLOWS. Emory McClintock New York, N. Y. Joel Hastings Metealf Winchester Clarence Lemuel Elisha Moore Boston Eliakim Hastings Moore Chicago, 111. Edward Charles Pickering Cambridge William Henry Pickering Cambridge Charles Lane Poor New York, N. Y. Roland George Dwight Richardson Providence, R. L Arthur Searle Cambridge George Mary Searle Washington, D. C. Vesto Melvin Slipher Flagstaff, Ariz. John Nelson Stockwell Cleveland, O. William Edward Story Worcester Henry Taber Worcester Harry Walter Tyler Boston Robert Wheeler Willson Cambridge Edwin Bidwell Wilson Brookline Frederick Shenstone Woods Newton Paul Sebastian Yendell Dorchester Class I., Section II. — Physics. — 55. Joseph Sweetman Ames Baltimore, Md. Carl Barus Providence, R. I. Louis Agricola Bauer Washington, D. C. Alexander Graham Bell Washington, D. C. Louis Bell Boston Clarence John Blake Boston Percy Williams Bridgman Cambridge Henry Andrews Bumstead New Haven, Conn. George Ashley Campbell New York, N. Y. Emory Leon Chaffee Belmont Daniel Frost Comstock Boston William David Coolidge Schenectady, N. Y. Henry Crew Evanston, 111. Charles Robert Cross Brookline Harvey Nathaniel Davis Cambridge Arthur Louis Day Washington, D. C. Louis Derr Brookline William Johnson Drisko Boston William Duane Boston Alexander Wilmer Duff Worcester FELLOWS. 937 Arthur Woolsey Ewell Worcester Harry Manley Goodwin Brookline George Ellery Hale Pasadena, Cal. Edwin Herbert Hall Cambridge Hammond Vinton Hayes Cambridge William Leslie Hooper Somerville John Charles Hubbard Worcester Charles Clifford Hutchins Brunswick, Me. James Edmund Ives Worcester WiUiam Wliite Jacques Boston Norton Adams Kent Cambridge Frank Arthur Laws Boston Henry Lefavour Boston Theodore Lyman Brookline Richard Cockburn Maclaurin Boston Thomas Corwin Mendenhall Ravenna, O. Ernest George Merritt Ithaca, N. Y. Albert Abraham Michelson Chicago, 111. Dayton Clarence Miller Cleveland, O. Robert Andrews Millikan Chicago, 111, Harry Wheeler Morse Los Angeles, Cal. Edward Leamington Nichols Ithaca, N. Y. Ernest Fox Nichols Hanover, N. H. Charles Ladd Norton Boston George Washington Pierce Cambridge Michael Idvorsky Pupin New York, N. Y. Wallace Clement Sabine Boston Frederick Albert Saunders Poughkeepsie, N. Y. John Stone Stone New York, N. Y, Maurice delvay Thompson Boston Elihu Thomson Swampscott John Trowbridge Cambridge Arthur Gordon Webster Worcester Charles Herbert Williams Milton Robert Williams Wood Baltimore, Md. Class I., Section III. — Chemistry. — 45. Wilder Dwight Bancroft Ithaca, N. Y. Gregory Paul Baxter Cambridge Marston Taylor Bogert New York, N. Y. 938 FELLOWS. Bertram Borden Boltwood New Haven, (Jonn. William Crowell Bray Berkeley, C'al. Russel Henry Chittenden New Haven, (. onn. Arthur Messinger Comey Chester, Pa. James Mason Crafts Boston Charles William Eliot Cambridge Henry Fay Boston George Shannon Forbes Cambridge Frank Austin Gooch New Haven, Conn. Lawrence Joseph Henderson Cambridge Eugene Waldemar Hilgard Berkeley, Cal. Charles Loring Jackson Cambridge Walter Louis Jennings Worcester Elmer Peter Kohler Cambridge Charles August Kraus Worcester Arthur Becket Lamb Cauibridge Gilbert Newton Lewis Berkeley, Cal. Warren Kendall Lewis Boston Arthur Dehon Little Brookline Charles Frederic Mabery Cleveland, O. Forris Jevvett Moore Boston George Dunning Moore Worcester Edward Williams Morley West Hartfon!, Conn. Harmon Northrop Morse Baltimore, Md. Samuel Parsons Mulliken Boston Charles Edward Munroe Washington, D. C. James Flack Norris Boston Arthur Amos Noyes Boston William Albert Noyes LTrbana, 111. Thomas Burr Osborne New Haven, Conn. Samuel Cate Prescott Brookline Ira Rcmsen Baltimore, Md. Robert Hallowell Richards Jamaica Plain Theodore William Richards Caml)ridge Martin Andre Rosanoff W^orcester Stephen Paschall Sharpies Cambridge Miles Standish Sherrill Brookline Alexander Smith New York, N. Y. Julius Oscar Stieglitz Chicago, 111. Henry Paul Talbot Newton William Hultz Walker Boston Willis Rodney Whitney Schenecta(?y, N. V. FELLOWS. 939 Class I., Section IV. — Technology and Engineering. — 42. Henry Larcom Abbot Cambridge Comfort Avery Adams Cambridge William Herbert Bixby Washington, D. C. Francis Tiffany Bowles Boston William Hubert Burr New Yoik, N. Y. Alfred Edgar Burton Boston John Joseph Carty New York, N. Y. Eliot Channing Clarke Boston Harry Ellsworth Clifford Newton Desmond FitzGerald Brookline John Ripley Freeman Providence, R. I. George Washington Goethals Culebra, Canal Zone John Hays Hammond New Yoik, N. Y, Rudolph Hering New Yoik, N. Y. Ira Nelson Hollis Cambridge Henry ^^arion Howe NewYoik, N. Y. Hector James Hughes Cambridge Alexander Crombie Humphreys New York, N. Y. Frederick Remsen Hutton New York, N. Y. Dugald Caleb Jackson Boston Lewis Jerome Johnson Cambridge Arthur Edwin Kennelly Cambridge Gaetano Lanza Philadelphia, Pa. William Roscoe Livermore Boston Lionel Simeon Marks Cambridge Edw ard Furber Miller Boston Hiram Francis Mills Lowell Charles Francis Park Boston William Barclay Parsons New Yoik, N. Y. Cecil Hobart Peabody Brookline Harold Pender Boston Edward Dyer Peters Dorchester Albert Sauveur Cambridge Peter Schwamb Arlington Henry Lloyd Smyth Cambridge Charles Milton Spofford Boston Frederic Pike Stearns Boston Charles Proteus Steinmetz Schenectady, N. Y. Geor;,'e Fillmore Swain Cambridge George Chandler W^hipple Cambridge 940 FELLOWS. Robert Simpson Woodward Washington, D. C. Joseph Ruggles Worcester Boston Class II. — Natural and Physiological Sciences. — 155. , Section I. — Geology, Mineralogy, and Physics of the Globe. — 43. Cleveland Abbe Washington, D. C. Wallace Walter Atwood Cambridge Joseph Barrell New Haven, Conn. George Hunt Barton Cambridge Thomas Chrowder Chamberlin Chicago, 111. William Bullock Clark Baltimore, Md. John Mason Clarke Albany, N. Y. Henry Helm Clayton Canton Herdman Fitzgerald Cleland Williamstown William Otis Crosby Jamaica Plain Reginald Aldworth Daly Cambridge Edward Salisbury Dana New Haven, Conn. Walter Gould Davis Cordova, Arg. William Morris Davis Cambridge Benjamin Kendall Emerson Amherst Grove Karl Gilbert Washington, D. C. Louis Caryl Graton Cambridge Ellsworth Huntington Milton Oliver Whipple Huntington Newport, R. I. Robert Tracy Jackson Boston Thomas Augustus Jaggar Honolulu, H. I. Douglas Wilson Johnson New York, N. Y. Alfred Church Lane Cambridge Andrew Cowper Lawson Berkeley, Cal. Charles Kenneth Leith Madison, Wis. Waldemar Lindgren Boston Charles Palache Cambridge John Elliott Pillsbury Washington, D. C. Raphael Pumpelly Newport, R. I. Robert Wilcox Sayles Cambridge Charles Schuchert New Haven, Conn. William Berryman Scott Princeton, N. J. Hervey Woodburn Shimer Boston Charles Richard Van Hise Madison, Wis. Charles Doolittle Walcott Washington, D. C. FELLOWS. 941 Robert DeCourcy Ward Cambridge Charles Hyde Warren Auburndale Herbert Percy Whitlock Albany, N. Y. Bailey Willis Washington, D. C. Samuel Wendell Williston . Chicago, 111. John Eliot Wolff Cambridge Jay Backus Woodworth Cambridge Frederick Eugene Wright Washington, D. C. Class II., Section II. — Botany. — 31. Oakes Ames North Easton Irving Widmer Bailey Cambridge Liberty Hyde Bailey Ithaca, N. Y. Douglas Houghton Campbell Stanford Univ., Cal. George Perkins Clinton New Haven, Conn. Frank Shipley Collins North Eastham John Merle Coulter Chicago, 111. Edward Murray East Jamaica Plain Alexander William Evans New Haven, Conn. William Gilson Farlow Cambridge Charles Edward Faxon Jamaica Plain Merritt Lyndon Fernald Cambridge George Lincoln Goodale Cambridge Robert Aimer Harper New York, N. Y. John George Jack Jamaica Plain Edward Charles Jeffrey Cambridge Fred Dayton Lambert Tufts College Burton Edward Livingston Baltimore, Md. George Richard Lyman Washington, D. C. Winthrop John Vanleuven Osterhout Cambridge Alfred Rehder Jamaica Plain Lincoln Ware Riddle Wellesley Benjamin Lincoln Robinson Cambridge Charles Sprague Sargent Brookline William Albert Setchell Berkeley, Cal. Arthur Bliss Seymour Cambridge Erwin Frink Smith Washington, D. C. John Donnell Smith Baltimore, Md. William Codman Sturgis Boston Roland Thaxter Cambridge William Trelease Urbana, 111. 942 FELLOWS. Class II., Section III. — Zoology and Physiology. — 50. Glover Morrill Allen Boston Joel Asaph Allen New York, N. Y. John Wallace Baird Worcester Thomas Barbour Boston Francis Gano Benedict Boston Henry Bryant Bigelow Concord Robert Payne Bigelow . . . .' Brookline William Brewster Cambridge Charles Thomas Brues Boston Hermon Carey Bumpus Tufts College Walter Bradford Cannon Cambridge William Ernest Castle Belmont Charles Value Chapin Providence, R. I. Samuel F'essenden Clarke Williamstown Edwin Grant Conklin Princeton, N. J. William Thomas Councilman Boston William Healey Dall Washington, D. C. Charles Benedict Davenport Cold Spring Harbor, N. Y. Gilman Arthur Drew Woods Hole Otto Knut Olof Folin . ., Brookline Alexander Forbes Milton Samuel Henshaw Cambridge Leland Ossian Howard Washington, D. C. Herbert Spencer Jennings Baltimore, Md. Charles Atwood Kofoid Berkeley, Cal. Frederic Thomas Lewis Waban Ralph Stayner Lillie Worcester Jacques Loeb New York, N. Y Franklin Paine Mall Baltimore, Md. Edward Lam-ens Mark Cambridge Ernest Gale Martin Cambridge Albert Davis Mead Providence, R. L Edward Sylvester Morse Salem Herbert Vincent Neal Tufts College Henry Fairfield Osborn New York, N. Y. George Howard Parker Cambridge John Charles Phillips Wenham James Jackson Putnam Boston Herbert Wilbur Rand Cambridge FELLOWS. 943 William Emerson Ritter La Jolla, Cal. William Thompson Sedgwick Boston Percy Goldthwait Stiles Newtonville John Eliot Thayer Lancaster Addison Emory Verrill New Haven, Conn. Arthur Wisswald Weysse Boston William Morton Wheeler Boston Harris Hawthorne Wilder Northampton Edmund Beecher Wilson New York, N. Y. Frederick Adams Woods Brookline Robert Mearns Yerkes Cambridge Class II., Section IV. — Medicine and Surgery. — 31. Edward Hickling Bradford Boston Henry Asbury Christian Boston Harvey Cushing Boston David Linn Edsall Boston Harold Clarence Ernst Jamaica Plain Simon Flexner New York, N. Y. William Stewart Halsted Baltimore, Md. Reid Hunt Brookline Abraham Jacobi New York, N. Y. Elliott Proctor Joslin Boston William Williams Keen Philadelphia, Pa. Frank Burr Mallory Brookline Samuel Jason Mixter Boston Edward Hall Nichols Boston Sir William Osier Oxford, Eng. Theophil Mitchell Prudd'en New York, N. Y. William Lambert Richardson Boston Milton Joseph Rosenau Boston Frederick Cheever Shattuck Boston Theobald Smith Princeton, N. J. Elmer Ernest Southard Boston Richard Pearson Strong Boston Ernest E Iward Tyz7,er Boston Frederick Herman Verhoeff Boston Henry Pickering Walcott Cambridge John Collins W^arren Boston William Henry Welch Baltimore, Md. 944 FELLOWS. Francis Henry Williams Boston Simeon Burt Wolbach Boston Horatio Curtis Wood Philadelphia, Pa. James Homer Wright Boston Class III. — Moral and Political Sciences. — 148. Section I. — Theology, Philosophy and Jurisprudence. — 40. Simeon Eben Baldwin New Haven, Conn. Joseph Henry Beale Cambridge Melville Madison Bigelow Cambridge Joseph Hodges Choate New York, N. Y. James De Normandie Roxbury Frederic Dodge Belmont Edward Staples Drown Cambridge William Harrison Dunbar Boston Timothy Dwight New Haven, Conn. William Wallace Fenn Cambridge Frederick Perry Fish Brookline George Angier Gordon Boston John Wilkes Hammond Cambridge Alfred Hemenway Boston William DeWitt Hyde Brunswick, Me. Marcus Perrin Knowlton Springfield William Lawrence Boston George Vasmer Leverett Boston William Caleb Loring Boston Nathan Matthews Boston Samuel Walker McCall Winchester Edward Caldwell Moore Cambridge George Herbert Palmer Cambridge George Wharton Pepper Philadelphia, Pa. John Winthrop Platner Cambridge Roscoe Pound Belmont Elihu Root New York, N. Y. James Hardy Ropes Cambridge Josiah Royce Cambridge Arthur Prentice Rugg Worcester Henry Newton Sheldon Boston Moorfield Storey Boston FELLOWS. 945 William Howard Taft New Haven, Conn. William Jewett Tucker Hanover, N. H. William Gushing Wait Medford Williston Walker New Haven, Conn. Eugene Wambaugh Cambridge Edward Henry Warren Boston Samuel Williston Belmont Woodrow Wilson Washington, D. C. Class III., Section II. — Philology and Ardiceology. — 45. Francis Greenleaf Allinson Providence, R. I. William Rosenzweig Arnold Cambridge Maurice Bloomfield Baltimore, Md. Franz Boas New York, N. Y. Charles Pickering Bowditch Jamaica Plain Franklin Carter Williamstown George Henry Chase Cambridge Roland Burrage Dixon Cambridge William Curtis Farabee Cambridge Jesse Walter Fewkes Washington, D. C. Jeremiah Denis Mathias Ford Cambridge Basil Lanneau Gildersleeve Baltimore, Md. Charles Hall Grandgent Cambridge Louis Herbert Gray Boston Charles Burton Gulick Cambridge William Arthur Heidel Middletown, Conn. Bert Hodge Hill Athens, Greece Edward Washburn Hopkins New Haven, Conn. Albert Andrew Howard Cambridge Ales Hrdlicka Washington, D. C. Carl Newell Jackson Cambridge Hans Carl Gunther von Jagemann Cambridge James Richard Jewett Cambridge Alfred Louis Kroeber Berkeley, Cal. Kirsopp Lake Cambridge Henry Roseman Lang New Haven, Conn. Charles Rockwell Lanman Cambridge David Gordon Lyon Cambridge Clifford Herschel Moore Cambridge George Foot Moore Cambridge 946 FELLOWS. Hanns Oertel New Haven, Conn. Charles Pomeroy Parker Cambriflge Bernadotte Perrin New Haven, Conn. Edward Kennard Rand CamUridge George Andrew Reisner Cambridge Edward Rol)inson .... New York, N. Y. Fred Norris Robinson C!!ani bridge Edward Stevens Sheldon Caml)ridge Herbert Weir Smyth Camliridge Franklin Bache Stephenson Claremont, Cal. Charles Cutler Torrey New Haven, ( "onn. Alfred Marston Tozzer Cambridge Andrew Dickson White Ithaca, N. Y. John Williams White Cambridge James Haughton Woods Cambridge Class III., Section III. — Political Economy and History. — 3i. Henry Adams Washington, D. C. Charles Jesse Bullock Cambridge Thomas Nixon Carver Cambridge John Bates Clark New York Archibald Cary Coolidge Boston Richard Henry Dana Cambridge Andrew McFarland Davis Cambridge Davis Rich Dewey Cambridge Edward Bangs Drew Cambridge Ephraim Etnerton Caml>ridge Henry Walcott Farnam New Haven, Conn. Irving Fisher New Haven, 'onn. Worthington Chauncey Ford Boston Edwin Francis Gay Cambridge Frank Johnson Goodnow Baltimore, Md. Arthur Twining Hadley New Haven, ( "onn. Albert Bushnell Hart Cambridge Charles Homer Haskins Cambridge Henry Cabot Lodge Nahant Al)bott Lawrence Lowell Cam' ridge Roger Bigelow Merriman Cambridge Sanuud Eliot Morison Boston William Bennett Munro Cambridge James Ford Rhodes • Boston FELLOWS. 947 William Mulligan Sloane New York, N. Y. Charles Card Smith Boston Henry Morse Stephens Berkeley, Cal. John Osborne Sumner Boston Frank William Taussig Cambridge William Roscoe Thayer Cambridge Frederick Jackson Turner Cambridge Thomas Franklin Waters Ipswich George Grafton Wilson Cambridge George Parker Winship Providence, R. I. Class III., Section IV. — Literature and the Fine Arts. — 29. George Pierce Baker Cambridge Arlo Bates Boston James Phinney Baxter Portland, Me. William Sturgis Bigelow Boston Le Baron Russell Briggs Cambridge Samuel McChord Crothers Cambridge Wilberforce Fames New York, N. Y. Henry Herbert Edes Cambridge Arthur Fairbanks Cambridge Arthur Foote Brookline Kuno Francke Cambridge Daniel Chester French Stockbridge Robert Grant Boston Henry Lee Higginson Boston James Kendall Hosmer Minneapolis, Minn. Mark Antony DeWolfe Howe Boston George Lyman Kittredge Cambridge William Coolidge Lane Cambridge Albert Matthews Boston William Allan Neilson Cambridge Bela Lyon Pratt Boston Herbert Putnam Washington, D. C. Denman Waldo Ross Cambridge John Singer Sargent London, Eng. EUery Sedgwick Boston Herbert Langford Warren Cambridge Barrett Wendell Boston Owen Wister Philadelphia, Pa. George Edward Woodberry Beverly 948 FOREIGN HONORARY MEMBERS. FOREIGN HONORARY MEMBERS.— 64. (Number limited to seventy- five). Class I. — Mathematical and Physical Sciences. — 22. Section I. — Mathematics and Astronomy. — 6. Arthur Auwers Berlin Johann Oskar Backlund Petrograd Felix Klein Gottingen Sir Joseph Norman Loekyer London Emile Picard Paris Charles Jean de la Vallee Poussin Louvain Class I., Section II. — Physics. — 9. Svante August Arrhenius . Stockholm Oliver Heaviside Torquay Sir Joseph Larmor Cambridge Hendrik Antoon Lorentz Leyden Max Planck Berlin Augusto Righi Bologna John William Strutt, Baron Rayleigh Witham Sir Ernest Rutherford Manchester Sir Joseph John Thomson . , . Cambridge Class I., Section III. — Chemistry. — 4. Adolf, Ritter von Baeyer Munich Emil Fischer Berlin Fritz Haber Berlin Wilhelm Ostwald Leipsic FOREIGN HONORARY MEMBERS. 94d Class I., Section IV. — Technology and Engineering. — 3. Heinrich Miiller-Breslau Berlin Vsevolod Jevgenjevic Timonoff Petrograd William Cawthorne Unwin London Class II. — Natural and Physiological Sciences. — 18. Section I. — Geology, Mineralogy, and Physics of the Globe. — 6. Waldemar Christofer Brogger Christiania Sir Archibald Geikie Haslemere, Surrey Viktor GoldschmiHt Heidelberg Julius Hann Vienna Albert Heim Zurich Johan Herman Lie Vogt Trondhjem Class II., Section II. — Botany. — 6. John Briquet Geneva Adolf Engler BerHn Wilhelm Pfeffer Leipsic Hermann, Graf zu Solms-Laubach Strassburg Ignatz Urban Berlin Eugene Warming Copenhagen Class II., Section III. — Zoology and Physiology. — 2. Sir Edwin Ray Lankester London Magnus Gustav Retzius Stockholm Class II., Section IV. — Medicine and Surgery. — 4.- Emil von Behring Marburg Sir Thomas Lauder Brunton, Bart London Angelo Celli Rome Adam Politzer Vienna 950 FOREIGN HONORARY MEMBERS. Class III. — Moral and Political Sciences. — 24. Section I. — Theology, Philosophy and Jurisprudence. — 4. Arthur James Balfour Prestonkirk Heinrich Brunner Berlin Albert Venn Dicey Oxford Sir Frederick Pollock, Bart London Section II. — Philology and Archaeology. — 8. Friedrich Delitzsch Berlin Hermann Dials . Berlin Wilhelm Dorpfeld Athens Henry Jackson Cambridge Hermann Georg Jacobi Bonn Sir Gaston Camille Charles Maspero Paris Alfred Percival Maudslay Hereford Eduard Seler Berlin Section III. — Political Economy and History. — 6. "Viscount Bryce London Adolf Harnack Berlin Alfred Marshall Cambridge John Morley, Viscount Morley of Blackburn London Sir George Otto Trevelyan, Bart London Pasquale Villari Florence Section IV. — Literature and the Fine Arts. — 6. Georg Brandes Copenhagen Thomas Hardy Dorchester Jean Adrien Aubin Jules Jusserand Paris Rudyard Kipling Burwash Sir Sidney Lee London Sir James Augustus Henry Murray Oxford STATUTES AND STANDING VOTES STATUTES Adopted November 8, 1911: amended May 8, 1912, January 8, and May 14, 1913, Apnl 14, 1915, Apnl 12, 1916. CHAPTER I The Corporate Seal Article 1, The Corporate Seal of the Academy shall be as here depicted : Article 2. The Recording Secretary shall have the custody of the Corporate Seal. See Chap. v. art. 3; chap. vi. art. 2. 952 STATUTES OF THE AMERICAN ACADEMY CHAPTER II Fellows and Foreign Honorary Members and Dues Article 1. The Academy consists of Fellows, who are either citizens or residents of the United States of America, and Foreign Honorary Members. They are arranged in three Classes, according to the Arts and Sciences in which they are severally proficient, and each Class is divided into four Sections, namely: Class I. The Mathematical and Physical Sciences Section 1. Mathematics and Astronomy Section 2. Physics Section 3. Chemistry Section 4. Technology and Engineering Class II. The Natural and Physiological Sciences Section 1. Geology, Mineralogy, and Physics of the Globe Section 2. Botany Section 3. Zoology and Physiology Section 4. Medicine and Surgery Class III. The Moral and Political Sciences Section 1. Theology, Philosophy, and Jurisprudence Section 2. Philology and Ai-chaeology Section 3. Political Economy and History Section 4. Literature and the Fine Arts Article 2. The number of Fellows shall not exceed Six hundred, of whom not more than Four hundred shall be residents of Massachu- setts, nor shall there be more than Two hundred in any one Class. Article 3. The number of Foreign Honorary Members shall not exceed Seventy-five. They shall be chosen from among citizens of foreign countries most eminent for their discoveries and attainments in any of the Classes above enumerated. There shall not be more than Twenty-five in any one Class. Article 4. If any person, after being notified of his election as Fellow, shall neglect for six months to accept in %^Titing and to pay his Admission Fee (unless he be absent from the Commonwealth at the time of his notification) his election shall be void; and if any Fellow resident within fifty miles of Boston shall neglect to pay his Annual Dues for six months after they are due, provided his attention shall have been called to this Article of the Statutes in the meantime, OF ARTS AND SCIENCES. 953 he shall cease to be a Fellow; but the Council may suspend the pro- visions of this Article for a reasonable time. "With the previous consent of the Council, the Treasurer may dis- pense (sub silentio) with the payment of the Admission Fee or of the Annual Dues or both whenever he shall deem it advisable. In the case of officers of the Army or Navy who are out of the Commonwealth on duty, payment of the Annual Dues may be waived during such absence if continued during the whole financial year and if notification of such expected absence be sent to the Treasiu-er. Upon similar notification to the Treasurer, similar exemption may be accorded to Fellows sub- ject to Annual Dues, who may temporarily remove their residence for at least two years to a place more than fifty miles from Boston. If any person elected a Foreign Honorary Member shall neglect for six months after being notified of his election to accept in writing, his election shall be void. See Chap. vii. art. 2. Article 5. Every Fellow hereafter elected shall pay an Admission Fee of Ten dollars. Every Fellow resident within fifty miles of Boston shall, and others may, pay such Annual Dues, not exceeding Fifteen dollars, as shall be voted by the Academy at each Annual Meeting, when they shall become due, except in the case of Fellows elected at the January meetings, who shall be obliged to pay but one half of such Annual Dues in the year in which they are elected; but any Fellow shall be exempt from the annual payment if, at any time after his admission, he shall pay into the treasury Two hundred dollars in addition to his previous payments. All Commutations of the Annual Dues shall be and remain perma- nently funded, the interest only to be used for current expenses. Any Fellow not previously subject to Annual Dues who takes up his residence within fifty miles of Boston, shall pay to the Treasm-er within three months thereafter Annual Dues for the current year, failing which his Fellowship shall cease; but the Council may suspend the provi- sions of this Ai'ticle for a reasonable time. Only Fellows who pay Annual Dues or have commuted them may hold office in the Academy or serve on the Standing Committees or vote at meetings. Article 6. Fellows who pay or have commuted the Annual Dues and Foreign Honorary Members shall be entitled to receive gratis one copy of all Publications of the Academy issued after their election. See Chap. x. art. 2. 954 STATUTES OF THE AMERICAN ACADEMY Article 7. Diplomas signed by the President and the Vice- President of the Class to which the member belongs, and countersigned by the Secretaries, shall be given to all the Fellows and Foreign Honorary Members. Article 8. If, in the opinion of a majority of the entire Council, any Fellow or Foreign Honorary Member shall have rendered himself unworthy of a place in the Academy, the Council shall recommend to the Academy the termination of his membership ; and if three fourths of the Fellows present, out of a total attendance of not less than fifty, at a Stated Meeting, or at a Special Meeting called for the purpose, shall adopt this recommendation, his name shall be stricken from the Roll. See Chap, ill.; chap. vi. art. 1; chap. ix. art. 1, 7; chap. x. art. 2. CHAPTER in Election of Fellows and Foreign Honorary Members Article 1. Elections of Fellows and Foreign Honorary Members shall be by ballot, and only at the Stated Meetings in January and May. Three fourths of the ballots cast, and not less than twenty, must be affirmative to effect an election. Article 2. Nominations to Fellowship or Foreign Honorary Membership in any Section must be signed by two Fellows in that Section. These nominations shall be sent to the Corresponding Secretary accompanied by statements of qualifications and brief biographical data, and shall be retained by him until the first of the following October or February, as the case may be. All nominations then in his hands shall be sent in printed form to every Fellow having the right to vote, with the names of the proposers in each case, and with a request to send to the Corresponding Secretary written com- ments on these names not later than the fifth of November or the fifth of March respectively. All the nominations, with the comments thereon, received up to the fifth of November or the fifth of March shall be referred at once to the appropriate Class Committees, which shall report their decisions to the Council at a special meeting to be called to consider nom- inations, not later than two days before the meeting of the Academy in December or April respectively. Notice shall be sent to every Fellow having the right to vote, not later than the fifteenth of September or January, of each year, calling OF ARTS AND SCIENCES. 955 attention to the fact that the limit of time for sending nominations to the Corresponding Secretary will expire on the first of the following month. Article 3. Not later then the fourth Wednesday of December and April, the Corresponding Secretary shall send in print to every Fellow having the right to vote all nominations that have been ap- proved by the Comicil, with a brief accomit of each nominee. See Chap, ii.; chap. vi. art. 1; chap. ix. art. 1. CHAPTER IV Officers Article 1. The Officers of the Academy shall be a President (who shall be Chairman of the Council), three Vice-Presidents (one from each Class), a Corresponding Secretary (who shall be Secretary of the Council), a Recording Secretary, a Treasurer, and a Librarian, all of whom shall be elected by ballot at the Annual Meeting, and shall hold their respective offices for one year, and until others are duly chosen and installed. There shall be also twelve Councillors, one from each Section of each Class. At each Annual Meeting three Councillors, one from each Class, shall be elected by ballot to serve for the full term of four years and until others are duly chosen and installed. The same Fellow shall not be eligible for two successive terms. The Councillors, with the other officers previously named, and the Chairman of the House Committee, ex officio, shall constitute the Council. See Chap. x. art. 1. Article 2. If any office shall become vacant during the year, the vacancy may be filled by the Council in its discretion for the unexpired term. Article 3. At the Stated Meeting in March, the President shall appoint a Nominating Committee of three Fellows having the right to vote, one from each Class. This Committee shall prepare a list of nominees for the several offices to be filled, and for the Standing Com- mittees, and file it with the Recording Secretary not later than four weeks before the Annual Meeting. See Chap. vi. art. 2. 956 STATUTES OF THE AMERICAN ACADEMY Article 4. Independent nominations for any office, if signed by at least twenty Fellows having the right to vote, and received by the Recording Secretary not less than ten days before the Annual Meet- ing, shall be inserted in the call therefor, and shall be mailed to all the Fellows having the right to vote. See Chap. vi. art. 2. Article 5. The Recording Secretary shall prepare for use in voting at the Annual Meeting a ballot containing the names of all persons duly nominated for office. CHAPTER V The President Article 1 . The President, or in his absence the senior Vice-Presi- dent present (seniority to be determined by length of continuous fellowship in the Academy), shall preside at all meetings of the Acad- emy. In the absence of all these officers, a Chairman of the meeting shall be chosen by ballot. Article 2. Unless otherwise ordered, all Committees which are not elected by ballot shall be appointed by the presiding officer. Article 3. Any deed or writing to which the Corporate Seal is to be affixed, except leases of real estate, shall be executed in the name of the Academy by the President or, in the event of his death, absence, or inability, by one of the Vice-Presidents, when thereto duly authorized. See Chap. ii. art. 7; chap. iv. art. 1, 3; chap. vi. art. 2; chap. vii. art. 1; chap. ix. art. 6; chap. x. art. 1; 2; chap. xi. art. 1. CHAPTER VI The Secretaries Article 1. The Corresponding Secretary shall conduct the corre- spondence of the Academy and of the Council, recording or making an entry of all letters written in its name, and preserving for the files all official papers which may be received. At each meeting of the Council he shall present the communications addressed to the Academy which have been received since the previous meeting, and at the next meeting of the Academy he shall present such as the Council may determine. OF ARTS AND SCIENCES. 957 He shall notify all persons who may be elected Fellows or Foreign Honorary Members, send to each a copy of the Statutes, and on their acceptance issue the proper Diploma. He shall also notify all meet- ings of the Council; and in case of the death, absence, or inability of the Recording Secretary he shall notify all meetings of the Academy. Under the direction of the Council, he shall keep a List of the Fellows and Foreign Honorary Members, arranged in their several Classes and Sections. It shall be printed annually and issued as of the first day of July. See Chap. ii. art. 7; chap. iii. art. 2, 3; chap. iv. art. 1; chap. ix. art. 6; chap. X. art. 1; chap. xi. art. 1. Article 2. The Recording Secretary shall have the custody of the Charter, Corporate Seal, Archives, Statute-Book, Journals, and all literary papers belonging to the Academy. Fellows borrowing such papers or documents shall receipt for them to their custodian. The Recording Secretary shall attend the meetings of the Academy and keep a faithful record of the proceedings with the names of the Fellows present; and after each meeting is duly opened, he shall read the record of the preceding meeting. He shall notify the meetings of the Academy to each Fellow by mail at least seven days beforehand, and in his discretion may also cause the meetings to be advertised; he shall apprise Officers and Commit- tees of their election or appointment, and inform the Treasurer of appropriations of money voted by the Academy. After all elections, he shall insert in the Records the names of the Fellows by whom the successful nominees were proposed. He shall send the Report of the Nominating Committee in print to every Fellow having the right to vote at least tln-ee weeks before the Annual Meeting. See Chap. iv. art. 3. In the absence of the President and of the Vice-Presidents he shall, if present, call the meeting to order, and preside until a Chairman is chosen. See Chap, i.; chap. ii. art. 7; chap. iv. art. 3, 4, 5; chap. ix. art. 6; chap. x. art. 1, 2; chap. xi. art. 1, 3. Article 3. The Secretaries, with the Chairman of the Committee of Publication, shall have authority to publish such of the records of the meetings of the Academy as may seem to them likely to promote its interests. 958 STATUTES OF THE AMERICAN ACADEMY CHAPTER VII The Treasurer and the Treasury Article 1. The Treasurer shall collect all money due or payable to the Academy, and all gifts and bequests made to it. He shall pay all bills due by the Academy, when approved by the proper officers, except those of the Treasurer's office, which may be paid without such ap- proval; in the name of the Academy he shall sign all leases of real estate; and, with the written consent of a member of the Committee on Finance, he shall make all transfers of stocks, bonds, and other securities belonging to the Academy, all of which shall be in his official custody. He shall keep a faithful account of all receipts and expenditures, submit his accounts annually to the Auditing Committee, and render them at the expiration of his term of office, or whenever required to do so by the Academy or the Council. He shall keep separate accounts of the income of the Rumford Fund, and of all other special Funds, and of the appropriation thereof, and render them annually. His accounts shall always be open to the inspection of the Council. Article 2. He shall report annually to the Council at its March meeting on the expected income of the various Funds and from all other sources during the ensuing financial year. He shall also report the names of all Fellows who may be then delinquent in the payment of their Annual Dues. Article 3. He shall give such security for the trust reposed in him as the Academy may require. Article 4. With the approval of a majority of the Committee on Finance, he may appoint an Assistant Treasurer to perform his du- ties, for whose acts, as such assistant, he shall be responsible; or, with like approval and responsibility, he may employ any Trust Company doing business in Boston as his agent for the same purpose, the com- pensation of such Assistant Treasurer or agent to be fixed by the Committee on Finance and paid from the funds of the Academy. Article 5. At the Annual Meeting he shall report in print all his official doings for the preceding year, stating the amount and condition OF ARTS AND SCIENCES. 959 of all the property of the Academy entrusted to him, and the character of the investments. Article 6. The Financial Year of the Academy shall begin with the first day of April. Article 7. No person or committee shall incur any debt or liability in the name of the Academy, unless in accordance with a previous vote and appropriation therefor by the Academy or the Council, or sell or otherwise dispose of any property of the Academy, except cash or invested funds, without the previous consent and ap- proval of the Council. See Chap. ii. art. 4, 5; chap. vi. art. 2; chap ix. art. 6; chap. x. art. 1, 2, 3; chap. xi. art. 1. CHAPTER VIII The Librarian and the Library Article 1. The Librarian shall have charge of the printed books, keep a correct catalogue thereof, and provide for their delivery from the Library. At the Annual Meeting, as Chairman of the Committee on the Li- brary, he shall make a Report on its condition. Article 2. In conjunction with the Committee on the Library he shall have authority to expend such sums as may be appropriated by the Academy for the purchase of books, periodicals, etc., and for de- fraying other necessary expenses connected with the Library. Article 3. All books procured from the income of the Rumford Fund or of other special Funds shall contain a book-plate expressing the fact. Article 4. Books taken from the Library shall be receipted for to the Librarian or his assistant. Article 5. Books shall be retiumed in good order, regard being had to necessary wear with good usage. If any book shall be lost or injured, the Fellow to whom it stands charged shall replace it by a new volume or bj'- a new set, if it belongs to a set, or pay the current price thereof to the Librarian, whereupon the remainder of the set, if any, 960 STATUTES OF THE AMERICAN ACADEMY shall be delivered to the Fellow so paying, unless such remainder be valuable by reason of association. Article 6. All books shall be returned to the Library for examina- tion at least one week before the Annual Meeting. Article 7. The Librarian shall have the custody of the Publica- tions of the Academy. With the advice and consent of the President, he may effect exchanges with other associations. See Chap. ii. art. 6; chap. x. art. 1, 2. CHAPTER IX The Council Article 1. The Council shall exercise a discreet supervision over all nominations and elections to membership, and in general supervise all the affairs of the Academy not explicitly reserved to the Academy as a whole or entrusted by it or by the Statutes to standing or special committees. It shall consider all nominations duly sent to it by any Class Com- mittee, and present to the Academy for action such of these nomina- tions as it may approve by a majority vote of the members present at a meeting, of whom not less than seven shall have voted in the affirmative. With the consent of the Fellow interested, it shall have power to make transfers between the several Sections of the same Class, report- ing its action to the Academy. See Chap. iii. art. 2, 3; chap. x. art. 1. Article 2. Seven members shall constitute a quorum. Article 3. It shall establish rules and regulations for the transac- tion of its business, and provide all printed and engraved blanks and books of record. Article 4. It shall act upon all resignations of officers, and all resignations and forfeitures of Fellowship; and cause the Statutes to be faithfully executed. It shall appoint all agents and subordinates not otherwise provided for by the Statutes, prescribe their duties, and fix their compensation. OF ARTS AND SCIENCES. 961 They shall hold their respective positions during the pleasure of the Council. Article 5. It may appoint, for terms not exceeding one year, and prescribe the functions of, such committees of its number, or of the Fellows of the Academy, as it may deem expedient, to facilitate the administration of the affairs of the Academy or to promote its interests. Article 6. At its March meeting it shall receive reports from the President, the Secretaries, the Treasurer, and the Standing Commit- tees, on the appropriations severally needed for the ensuing financial year. At the same meeting the Treasurer shall report on the expected income of the various Funds and from all other sources during the same year. A report from the Council shall be submitted to the Academy, for action, at the March meeting, recommending the appropriation which in the opinion of the Council should be made. On the recommendation of the Council, special appropriations may be made at any Stated Meeting of the Academy, or at a Special Meet- ing called for the purpose. See Chap. x. art. 3. Article 7. After the death of a Fellow or Foreign Honorary Mem- ber, it shall appoint a member of the Academy to prepare a Memoir for publication in the Proceedings. Article 8. It shall report at every meeting of the Academy such business as it may deem advisable to present. See Chap. ii. art. 4, 5, 8; chap. iv. art. 1, 2; chap. vi. art. 1; chap. vii. art. 1; chap. xi. art. 1, 4. CHAPTER X Standing Committees Article 1. The Class Committee of each Class shall consist of the Vice-President, who shall be chairman, and the four Councillors of the Class, together with such other officer or officers annually elected as may belong to the Class. It shall consider nominations to Fellowship in its own Class, and report in writing to the Council such as may receive at a Class Committee Meeting a majority of the votes cast, provided at least three shall have been in the affirmative. See Chap. iii. art, 2. 962 STATUTES OF THE AMERICAN ACADEMY Article 2. At the Annual Meeting the following Standing Com- mittees shall be elected by ballot to serve for the ensuing year: (i) The Committee on Finance, to consist of three Fellows, who, through the Treasurer, shall have full control and management of the funds and trusts of the Academy, with the power of investing the funds and of changing the investments thereof in their discretion. See Chap. iv. art. 3; chap. vii. art. 1,4; chap. ix. art. 6. (ii) The Rumford Committee, to consist of seven Fellows, who shall report to the Academy on all applications and claims for the Rumford Premium. It alone shall authorize the purchase of books publications and apparatus at the charge of the income from the Rumford Fund, and generally shall see to the proper execution of the trust. See Chap. iv. art. 3; chap. ix. art. 6. (iii) The Cyrus Moors Warren Committee, to consist of seven Fel- lows, who shall consider all applications for appropriations from the income of the Cyrus Moors Warren Fund, and generally shall see to the proper execution of the trust. See Chap. iv. art. 3; chap. ix. art. 6. (iv) The Committee of Publication, to consist of three Fellows, one from each Class, to whom all communications submitted to the Academy for publication shall be referred, and to whom the printing of the Proceedings and the Memoirs shall be entrusted. It shall fix the price at which the Publications shall be sold; but Fellows may be supplied at half price with volumes which may be needed to complete their sets, but which they are not entitled to receive gratis. Two hundred extra copies of each paper accepted for publication in the Proceedings or the Memoirs shall be placed at the disposal of the author without charge. See Chap. iv. art. 3; chap. vi. art. 1, 3; chap. ix. art. 6. (v) The Committee on the Library, to consist of the Librarian, ex officio, as Chairman, and three other Fellows, one from each Class, who shall examine the Library and make an annual report on its condition and management. See Chap. iv. art. 3; chap. viii. art. 1,2; chap. ix. art. 6. OF ARTS AND SCIENCES. 963 (vi) The Hoicse Committee, to consist of three Fellows, who shall have charge of all expenses connected with the House, including the general expenses of the Academy not specifically assigned to the care of other Committees or Officers. See Chap. iv. art. 1, 3; chap. ix. art. 6. (vii) The Committee on Meetings, to consist of the President, the Recording Secretary, and three other Fellows, who shall have charge of plans for meetings of the Academy. See Chap. iv. art. 3; chap. ix. art. 6. (viii) The Auditing Committee, to consist of two Fellows, who shall audit the accounts of the Treasurer, with power to employ an expert and to approve his bill. See Chap. iv. art. 3; chap. vii. art. 1; chap. ix. art. 6. Article 3. The Standing Committees shall report annually to the Council in March on the appropriations severally needed for the ensu- ing financial year; and all bills incurred on account of these Commit- tees, within the limits of the several appropriations made by the Academy, shall be approved by their respective Chairmen. In the absence of the Chairman of any Committee, bills may be approved by any member of the Committee whom he shall designate for the purpose. See Chap. vii. art. 1, 7; chap. ix. art. 6. CHAPTER XI Meetings, Communications, and Amendments Article 1. There shall be annually eight Stated Meetings of the Academy, namely, on the second Wednesday of October, November, December, January, February, March, April and May. Only at these meetings, or at adjournments thereof regularly notified, or at Special Meetings called for the purpose, shall appropriations of money be made or amendments of the Statutes or Standing Votes be effected. The Stated Meeting in May shall be the Annual Meeting of the Corporation. Special Meetings shall be called by either of the Secretaries at the request of the President, of a Vice-President, of the Council, or of ten 964 STATUTES OF THE AMERICAN ACADEMY Fellows having the right to vote; and notifications thereof shall state the purpose for which the meeting is called. A meeting for receiving and discussing literary or scientific com- munications may be held on the fourth Wednesday of each month, excepting July, August, and September; but no business shall be transacted at said meetings. Article 2. Twenty Fellows having the right to vote shall consti- tute a quorum for the transaction of business at Stated or Special Meetings. Fifteen Fellows shall be sufficient to constitute a meeting for literary or scientific communications and discussions. Article 3. Upon the request of the presiding officer or the Record- ing Secretary, any motion or resolution offered at any meeting shall be submitted in writing. Article 4. No report of any paper presented at a meeting of the Academy shall be published by any Fellow without the consent of the author; and no report shall in any case be published by any Fellow in a newspaper as an account of the proceedings of the Academy without the previous consent and approval of the Council. The Council, in its discretion, by a duly recorded vote, may delegate its authority in this regard to one or more of its members. Article 5. No Fellow shall introduce a guest at any meeting of the Academy until after the business has been transacted, and espe- cially until after the result of the balloting upon nominations has been declared. -< - Article 6. The Academy shall not express its judgment on literary or scientific memoirs or performances submitted to it, or included in its Publications. Article 7. All proposed Amendments of the Statutes shall be re- ferred to a committee, and on its report, at a subsequent Stated Meet- ing or at a Special Meeting called for the purpose, two thirds of the ballot cast, and not less than twenty, must be affirmative to effect enactment. Article 8. Standing Votes may be passed, amended, or rescinded at a Stated Meeting, or at a Special Meeting called for the purpose, by a vote of two thirds of the members present. They may be suspended by a unanimous vote. See Chap. ii. art. 5, 8; chap, iii.; chap. iv. art. 3, 4, 5; chap, v. art. 1 ; chap. vi. art. 1,2; chap. ix. art. 8. OF ARTS AND SCIENCES. 965 STANDING VOTES 1 . Communications of which notice has been given to either of the Secretaries shall take precedence of those not so notified. 2. Fellows may take from the Library six volumes at any one time, and may retain them for three months, and no longer. Upon special application, and for adequate reasons assigned, the Librarian may permit a larger number of volumes, not exceeding twelve, to be drawn from the Library for a limited period. 3. Works published in numbers, when unbound, shall not be taken from the Hall of the Academy without the leave of the Librarian. 4. There may be chosen by the Academy, under the same rules by which Fellows are now chosen, one hundred Resident Associates. Not more than forty Resident Associates shall be chosen in any one Class. The election of Resident Associates shall be for a term of three years with eligibility for reelection. Resident Associates shall be entitled to the same privileges as Fel- lows, in the use of the Academy building, may attend meetings and present papers, but they shall not have the right to vote. They shall pay no Admission Fee, and their Annual Dues shall be one-half that of Fellows residing within fifty miles of Boston. The Council and Committees of the Academy may ask one or more Resident Associates to act with them in an advisory or assistant ca- pacity. RUMFORD PREMIUM In conformity with the terms of the gift of Sir Benjamin Thompson, Count Rumford, of a certain Fund to the American Academy of Arts and Sciences, and with a decree of the Supreme Judicial Court of Massachusetts for carrying into effect the general charitable intent and purpose of Count Rumford, as expressed in his letter of gift, the Acad- emy is empowered to make from the income of the Rumford Fund, as it now exists, at any Annual Meeting, an award of a gold and a silver medal, being together of the intrinsic value of three hundred dollars. 966 STATUTES OF THE AMERICAN ACADEMY as a Premium to the author of any important discovery or useful improvement in light or heat, which shall have been made and pub- lished by printing, or in any way made known to the public, in any part of the continent of America, or any of the American Islands; preference always being given to such discoveries as, in the opinion of the Academy, shall tend most to promote the good of mankind ; and, if the Academy sees fit, to add to such medals, as a further Premium for such discovery and improvement, a sum of money not exceeding three hundred dollars. INDEX. Aeroplanes, The Construction of, 826; The Elementary Principles of Flight, 826; The Experimental Determination of the Dynamical Properties of, 826; The Man- oeuvering of, 826; The Mechan- ics of the Motion of, 826. Affel, H. A. See Kennelly, A. E., and Affel, H. A. Agaricia Fragilis Dana, On the Development of the Coral, 483. Allen, G. M., accepts Fellowship, 821. Ames, J. B., Notice of, 845. Amory, (Francis) Fund, 833. Angell, J, B., death of, 827. Ants, The Australian Honey, of the Genus Leptomyrmex Mayr, 253. Asbjornsen, P. C, portrait of, 823. Assessment, Annual, Amount of, 839. Atwood, W. W., accepts Fellowship, 821. Australian Honey-Ants of the Genus Leptomyrmex Mayr, 253. Baird, J. W., elected Fellow, 841. Barbour, Thomas, elected Fellow, 824; accepts Fellowship, 825. Barrell, Joseph, accepts Fellowship, 821. Barton, G. H., elected Fellow, 824; accepts Fellowship, 825. Baxter, J. P., accepts Fellowship, 822. Bell, Louis. See Verhoeff, F. H., Bell, Louis, and Walker, C. B. Bergen, J. Y., resigns Fellowship, 827. Bigelow, W. S., elected Recording Secretary, 822. Billings, J. S., Notice of, 847. Biographical Notices, List of, 843. Blake, S. F., Compositae new and transferred, chiefly Mexican, 513. Boissier, G., Notice of, 849. Bornet, J B. E., Notice of, 852. Borraginaceae, Certain, new or trans- ferred, 541. Boundary Conditions, Expansion Problems with Irregular, 381. Bowman, Isaiah, elected Fellow, 841. Bremer, J. L., elected Fellow, 841. Bridgman, P. W., Polymorphic Changes under Pressure of the Univalent Nitrates, 579; Poly- morphic Transformations of Solids under Pressure, 53. Brues, C. T., accepts Fellowship, 821 Brush, G. J., Notice of, 853. Bullard, A. J., On the Structure of Finite Continuous Groups, 822. Bumpus, H. C, accepts Fellowship, 821. Bunsen, R. W., portrait of, 823. Burrill, T. J., accepts Fellowship, 822; death of, 831; Notice of, 857. Butcher, S. H., Notice of, 858. By water, Ingram, death of, 822; No- tice of, 861. Canada, On the Life-History of Cera- tomyx acadiensis, a new species of Myxosporidia from the east- ern coast of, 549. Canals, Martian, Application of Im- proved Methods to the Photog- raphy of the, 842 Carty, J. J., accepts Fellowship, 821; Wireless Telephony, 822. Castle, W. E., The Inheritance of Size in Guinea Pig Crosses, 827. Ceratomyxa acadiensis. On the Life- History of, a new species of Myxosporidia from the eastern coast of Canada, 549. Chadwick. G. W., resigns Fellowship, 830. Chaffee, E. L., elected Fellow, 824; accepts Fellowship, 825. Chapman, H. L., Notice of, 862. Chitonomyces and Rickia, New or Critical Species of, 842. Clark, H. L., declines Fellowship, 821. Clark, J. B., elected Fellow, 824; accepts Fellowship, 825. Clark, W. B., elected Fellow, 841. 968 INDEX. Clarke, J. M., accepts Fellowship, 821. Coal, On the Microscopical Struc- ture of Vegetable Organisms found in, 823. Coolidge, W. D., accepts Fellowship, 821. . - Coral Agaricia Fragilis Dana, On the Developmentof the, 483. - Coral Reefs, The Glacial-Control Theory of, 155. Committees, Standing, elected, 840. Coinpositae new and transferred, chiefly Mexican, 513. Council, Report of, 831. Craven, Alfred, decUnes Fellowship, 823. Cross, C. R., Report of Rumford Committee, 835. Cryptogamic Laboratories of Har- vard University, Contributions from, 1. Daly, R. A., The Glacial-Control Theory of Coral Reefs, 155. Dana, R. H., accepts Fellowship, 821. Dana, R. H. (1815-82), portraits of, 823. Darwin, C. R., portraits of, 823. Darwin, G. H., Notice of, 863. Davenport, G. E., Notice of, 865. Davidson, G., Notice of, 866. Davis, B. M., elected Fellow, 824; accepts Fellowship, 825. Diaphragms, Telephone-Receiver, The Mechanics of, as derived from their Motional-Impedance Circles, 419. Dickson, L. E., accepts Fellowship. 821. Drew, E. B., accepts Fellowship. 821. Drown, E. S., accepts Fellowship, 821. Drown, T. M., Notice of, 868. Edes, Henry H., Report of Treasurer, 831. Endicott, M. T., lapse of Fellowship of, 823. Energy, Radiant, Pathological Ef- fects of, upon the Eye, 627. Expansion Problems with Irregular Boundary Conditions, 381. Eye, Pathological Effects of Radiant Energy upon the, 627. Feldspars, A Quantitative Study of Certain Perthitic, 125. Fellows deceased, (11) — J. B. Angell, 827. T. J. Burrill, 831. E. D. Leavitt, 827. J. U. Nef, 821. .F.W.Putnam, 821. A. H. Russell, 821. - :E. R. Thayer, 821. W. R. Ware, 821. William Watson, 821. J. C. White, 824. C. H. Wing, 821. • Fellows elected, (21) — J. W. Baird, 841. Thomas Barbour, 824. G. H. Barton, 824. Isaiah Bowman, 841. J. L. Bremer, 841. E. L. Chaffee, 824. J. B. Clark, 824. W. B. Clark, 841. B. M. Davis, 824 F. J. Goodnow, 824. L. H. Gray, 841. A. S.Hardy, 841. A. B.Hart, 841. Ellsworth Huntington, 841. C. K. Leith, 841. F. T. Lewis, 841. F. A. Saunders, 824. W. A. Setchell, 841. P. G. Stiles, 841. W. C. Sturgis, 841. J. O. Sumner, 824. Fellows elected, declining Fellow- ship, H. L. Clark, 821. Alfred Craven, 823. Fellows elected, allowing Fellowship to lapse, Endicott, M. T., 823. Fellows resigning Fellowship, J. Y. Bergen, 827. G. W. Chadwick, 830. Hugo Miinsterberg, 824. Fellows, List of, 845. Finite Continuous Groups, On the Structure of, 822. Fisher, G. P., Notice of, 870. Fission, Mitosis and Multiple, in Trichomonad Flagellates, 287. Fitz, R. H., Notice of, 871. Flagellates, Trichomonad, Mitosis and Multiple Fission in, 287. Flight, The Elementary Principles of, 826. INDEX. 969 Forbes, Alexander, accepts Fellow- ship, 821. Forbes, G. S., accepts Fellowship, 821. Foreign Honorary Members deceased (3)- Ingram By water, 822. Ludimar Hermann, 825. H. E. Roscoe, 824. Foreign Honorary Members elected (2) Thomas Hardy, 841. Alfred Marshall, 825. Foreign Honorary Members, List of, 858 Foster, Sir M., Notice of, 873. Fox, Philip, accepts Fellowship, 821. Fuller, M. W., Notice of, 875. Furness, H. H., Notice of, 879. General Fund, 831; Appropriations from the Income of, 822, 826, 827. Gibbs, W., Notice of, 881. Gill, Sir D., Notice of, 883. Glacial-Control Theory of Coral Reefs, 155. Gmelin, Leopold, portrait of, 823. Goodell, A. C, Notice of, 886. Goodnow, F. J., elected Fellow, 824; accepts Fellowship, 825. Gray, J. C, Notice of, 888. Gray, L. H., elected Fellow, 841. Gray Herbarium of Harvard Uni- versity, Contributions from, 513. Guinea Pig Crosses, The Inheritance of Size in, 827. Hale, G. E., Researches on the Nature of Sun-Spots, 842. Hardy, A. S., elected Fellow, 841. Hardy, Thomas, elected Foreign Honorary Member, 841. Hart, A. B., elected Fellow, 841. Harvard University. See Crypto- gamic Laboratories. Gray Her- barium. Jefferson Physical Laboratory. Zoological Labo- ratory. Hayes, H. V., Report of House Com- mittee, 838. Haynes, H. W., Notice of, 889. Hermann, Ludimar, death of, 825. Hill, G. W., Notice of, 890. Hitchcock, F. L., accepts Fellow- ship, 821. Holden, E. S., Notice of, 891. Hopkins, E. W., accepts Fellowship, 821. Hosmer, J. K., accepts Fellowship, 821. House Committee, Report of, 838. House Expenses, Appropriations for, 826. Hrdlicka, Ales, accepts Fellowship, 821. Hunsaker, J. C., The Experimental Determination of the Dynamical Properties of Aeroplanes, 826. Hunt, Reid, accepts Fellowship, 821. Huntington, Ellsworth, elected Fel- low, 841. Huntington, E. V., Report of Pub- lishing Committee, 838. Hyde, W. DeW., accepts Fellowship, 821. Indo-Malayan, New, Laboulbeniales, 1. Jackson, C. N., accepts Fellowship, 821. Jackson, Dunham, Expansion Prob- lems with Irregular Boundary Conditions, 381. Jefferson Physical Laboratory, Har- vard University, Contributions from, 53. Jeffrey, E. C, On the Microscopic Structure of Vegetable Organ- isms found in Coal. 823. Johnson, S. W., Notice of, 895. Jones, B. Q., The Manoeuvering of Aeroplanes, 826. Kaiserliche Academie der Wissen- schaften, Wien, Math.-nat. Classe, portrait of members of, 823. Kelvin, Lord, Notice of, 896. Kennelly, A E., and Affel, H. A., The Mechanics of Telephone- Receiver Diaphragms, as de- rived from their Motional-Im- pedance Circles, 419. Kilauea, On Cyclical Variation in Eruption at, 825. Kofoid, C. A., and Swezy, Olive, Mitosis and Multiple Fission in Trichomonad Flagellates, 287. Kraus, C. A., accepts Fellowship, 821. 970 INDEX. Laboulbeniales, New Indo-Malayan, 1. Lake, Kirsopp, accepts Fellowship, 821; The Monks of Mt. Athos, 825. Lang; H. R., accepts Fellowship, 82L Lawson, A. C, accepts Fellowship, 822 Lea, H.C., Notice of, 899. Leavitt, E. D., death of, 827. Leith, C. K., elected Fellow, 84L Leptomyrmex, Mayr, The Australian Honey-Ants of the Genus, 253. Lewis, F. T., elected Fellow, 84L Lewis, W. K., accepts Fellowship, 82L Library, Appropriations for, 826. Library Committee, Report of, 843. Little, A. D., elected member of Council,'822. Lockyer, Norman, accepts Foreign Honorary Membership, 82L Loening, G. C., The Construction of Aeroplanes, 826. Loring, W. C, accepts Fellowship, 821. Lowell, F. C, Notice of, 900. Lowell, Fercival, Application of Im- proved Methods to the Photog- raphy of the Martian Canals, 842. Macbride, J. F., Certain Borragina- ceae new or transferred, 54L Maitland, F. W., Notice of, 904. Marshall, Alfred, elected Foreign Honorary Member, 825; accepts Membership, 826. Martin, E. G., accepts Fellowship, 821. Massachusetts Charitable Eye and Ear Infirmary, Contribution from the Pathological Labora- tory of, 627. Mavor, J. W., On the Development of the Coral Agaricia Fragilis Dana, 483; On the Life-History of Ceratomyxa acadiensis, a new species of Myxosporidia from the eastern coast of Canada, 549. McCall, S. W., accepts Fellowship, 823. Mexican, Cornpositae new and trans- ferred, chiefly, 513. Mitosis and Multiple Fission in Trichomonad Flagellates, 287. Mohammedan Law, 825. Monkeys and Apes, Ideational Be- havior of, 823. Moore, G. F., Mohammedan Law, 825. Morison, S. E., accepts Fellowship, 821 ; transferred from Section 4 to Section 3, of Class III, 823. Mount Athos, The Monks of, 825. Mtinsterberg, Hugo, resigns Fellow- ship, 824. Museum of Comparative Zoology at Harvard College. See Zoologi- cal Laboratory. My.xosporidia, On the Life-History of Ceratomyxa acadiensis, a new species of, from the eastern coast of Canada, 549. Nash, B. H., Notice of, 906. Nef, J. U., deathof, 821. Nef, J. U., Notice of, 907. Newcomb, S., Notice of, 908. Nitrates, Univalent, Polymorphic changes under Pressure of the, 579. Noble, A., Notice of, 909. Nominating Committee, 826. Norris, J. F., resigns from Council, 821. Officers elected, 839; List of, 843. Ordway, J. M., Notice of, 911. Pathological Effects of Radiant Energy upon the Eye, 627. Pathological Laboratory of the Mass. Charitable Eye & Ear Infirmary, See Mass. Charitable Eye & Ear Infirmary, Pathological Labora- tory. Perrin, Bernadotte, accepts Fellow- .ship, 821. Phillips, J. C, accepts Fellowship, 821. Platner, J. W., accepts Fellowship, 821. Polymorphic Changes under Pressure of the Univalent Nitrates, 579. Polymorphic Transformations of Solids under Pressure, 53. Pressure of the Univalent Nitrates, Polymorphic Changes under 579. Pressure, Polymorphic Transforma- tions of Solids under, 53. Pringle, C. G., Notice of, 912. Publication, Appropriation for, 827. INDEX. 971 Publication Committee, Report of, 838. Publication Fund, 833; Appropria- tion from the Income of, 827. Putnam, C. P., Notice of, 916. Putnam, F. W., death of, 821; No- tice of, 920. Recording Secretary, W. S. Bigelow elected, 822. Records of Meetings, 821. von Richthofen, F., Notice of, 921. Rickia, New or Critical Species of Chitonomyces and, 842. Riddle, L. W., accepts Fellowship, 821. Robinson, B. L., New, reclassified, or otherwise noteworthy Sperma- tophytes, 527. Roscoe, Sir H. E., death of, 824; No- tice of, 923. Rumford Committee, Report of, 835. Rumford Fund, 832; Appropriations from the Income of, 827 ; Papers published by aid of, 53, 579, 627. Rumford Premium, 875. Russell, A. H., death of, 821. Rutherford, Ernest, accepts Foreign Honorary Membership, 821. Saunders, F. A., elected Fellow, 824; accepts Fellowship, 825. Sayles, R. W., accepts Fellowship, 821. Schuchert, Charles, accepts Fellow- ship, 821. Setchel, W. A., elected Fellow, 841. SherriU, M. S., accepts Fellowship, 821. Solids, Polymorphic Transformations of, under Pressure, 53. Sorby, H. C, Notice of, 925. Spermatophytes, New, reclassified, or otherwise noteworthy, 527. Standing Committees elected, 840; List of, 843. Standing Votes, 875. Statutes, 861, Amendment of, 830. Statutes, Committee appointed on Amendments of, 822; report of, 827 Stiles, P. G., elected Fellow, 841. Stockhardt, J. A., portrait of, 823. Strasburger, E., Notice of, 927. Sturgis, W. C, elected Fellow, 841. Sumner, J. O., elected Fellow, 824; accepts Fellowship, 825. Sun-Spots, Researches on the Nature of, 842. Swezy, Olive. See Kofoid, C. A., and Swezy, Olive. Talbot, H. P., Report of C. M. Warren Committee, 836; Telephone-Receiver Diaphragms, The Mechanics of, as derived from their Motional-Impedance Circles, 419. Thaxter, Roland, New Indo-Malayan Laboulbeniales, 1 ; New or Criti- cal Species of Chitonomyces and Rickia, 842. Thayer, E. R., death of, 821. Thayer, W. R., accepts Fellowship, 821. Treasurer, Report of, 831. University of California. See Zoologi- c al Laboratory. de la Valine Poussin, C. J., accepts Foreign Honorary Membership, 821. Vegetable Organisms found in Coal, On the Microscopical Structure of, 823. Verhoeff, F. H., Bell, Louis, and Walker, C . B . The Pathological Effects of Radiant Energy upon the Eye, 627. Vogt, J. H. L., accepts Foreign Hono- rary Membership, 821. Walker, C. B. See Verhoeff, F. H., Bell, Louis, and Walker, C- B. Ware, W. R., death of, 821. Warren, C. H., A Quantitative Study of Certain Perthitic Feldspars, 125. W^arren (C. M.) Committee, Report of, 836. Warren (C. M.) Fund, 832; Appro- priation from the Income of, 827. Watson, Wilham, death of, 821; acquisition of scientific books of, 823. Webster, A. G., The Elementary Principles of Flight, 826; Report of the Library Committee, 834. Weysse, A. W., accepts Fellowship, 821. 972 INDEX Wheeler, W. M., The AustraHan Wright, F. E., accepts Fellowship, Honey- Ants of the Genus Lepto- 821. myrmex Mayr, 253. Wright, J. H., Notice of, 930. White, J. C, death of, 824. Wyman, Jeffries, portrait of, 823. Willis, Bailey, accepts Fellowship, 821. Yerkes, R. M., accepts Fellowship, Williston, S. W., accepts Fellowship, 821; Ideational Behavior of 821. Monkeys and Apes, 823. WUson, Edwin B., The Mechanics of the Motion of Aeroplanes, 826. Zoological Laboratory of the Museum Wing, C. H., death of, 821; Notice of Comparative Zoology at Har- of, 928. vard College, Contributions from Wood, E. S., Notice of, 929. 483, 529. Wood, Harry, On Cyclical Variation Zoological Laboratory of the Uni- in Eruption at Kilanea, 825. versity of California, Contribu- Woods, F. A., accepts Fellowship, tions from, 287. 821.