I l.^ mm BIOLOGICAL LABORATORY. ived >ssion No. ^.^ v^ -<^ ^ by /h^*,t^z;^ ^^c2 Committee on Publication C. HART MERRIAM Chairman and Editor MARCUS BAKER F. A. LUCAS JOHN HYDE O. H. TITTMAN PROCEEDINGS Washington Academy of Sciences Vol. II 1900 WASHINGTON March-December igcxj AFFILIATED SOCIETIES The Anthropological Society of Washington The Biological Society of Washington The Chemical Society of Washington The Columbia Historical Society The Entomological Society of Washington The Geological Society of Washington The Medical Society of the District of Columbia The National Geographic Society The Philosophical Society of Washington 3 L 7 0 Prkss op The New Era Printing Company Lancaster, Pa. CONTENTS. Brochure i. Cardinal Principles of Science. W J McGee. Is- sued March 14, 1900. *' 2. Papers from the Harriman Alasl^a Expedition. — I. Descriptions of Twenty-six New Mammals from Alaska and British North America. C. Hart Merriam. Issued March 14, 1900. " 3. The Bats of the Genus Monophylhis. Gerrit S. Miller, Jr. Issued March 30, 1900. " 4. A New Shrew from Eastern Turkestan. Gerrit S. Miller, Jr. Issued March 30, 1900. " 5. A New Stony Meteorite from Allegan, Michigan, and a New Iron Meteorite from Mart, Texas. George P. Merrill and H. N. Stokes. Issued July 25, 1900. " 6. The Giant Squirrels of Burmah and the Malay Pe- ninsula. Gerrit S. Miller, Jr. Issued July 25, 1900. *' 7. Descriptions of Two New Squirrels from Trong, Lower Siam. Gerrit S. Miller, Jr. Issued July 25, 1900. " 8. Preliminary Revision of the European Redbacked Mice. Gerrit S. Miller, Jr. Issued July 26, 1900. " 9. Papers from the Harrimai] Alaska Expedition. — II. Harrimania maculosa^ a New Genus and Species of Enteropneusta from Alaska, with Special Re- gard to the Character of Its Notochord. Wm. E. RiTTER. Issued Aug. 20, 1900. " 10. Results of the Branner-Agassiz Expedition to Brazil. — I. The Decapod and Stomatopod Crustacea. Mary J. Rathbun. Issued Aug. 20, 1900. " II. Results of the Branner-Agassiz Expedition to Brazil. — II. The Isopod Crustacea. Harriet Rich- ardson. Issued Aug. 20, 1900. V Brochure i3. Results of the Branner-Agassiz Expedition to Brazil. — III. The Fishes. Charles H. Gilbert. Is- sued August 20, 1900. " 13. Results of the Branner-Agassiz Expedition to Brazil. — IV. Two Characteristic Geologic Sections on Northeast Coast of Brazil. John C. Branner. Issued August 20, 1900. " 14. Mammals Collected by Dr. W. L. Abbott on Islands ■ in South China Sea. Gerrit S. Miller, Jr Issued August 20, 1900. " 15. Papers from the Ilarriman Alaska Expedition. — III. Multiplication of Rays and Bilateral Symmetry in the 20-Rayed Star-fish, Pycnopodia heliaiithoides (Stimpson). Wm. E. Ritter and Gulielma R. Crocker. Issued Oct. 10, 1900. " 16. Papers from the Harriman Alaska Expedition. — IV. The Tree Willows of Alaska. Frederick V. Co- ville. Issued Oct. 10, 1900. " 17. Papers from the Harriman Alaska Expedition. — V. Notes on the Hepaticce Collected in Alaska. Alexander W. Evans. Issued Oct. 10, 1900. *' 18. Papers from the Harriman Alaska Expedition. — VI. The Bryozoa. Alice Robertson. Issued Oct. 10, 1900. " 19. Papers from the Harriman Alaska Expedition. — VII. The Tenthredinoidea. Trevor Kincaid. Issued Nov. 24, 1900. " 20. Papers from the Ilarriman Alaska Expedition. — VIII. The Metamorphoses of Some Alaska Coleoptera. Trevor Kincaid. Issued Nov. 24, 1900. *' 21. Papers from the Harriman Alaska Expcilition. — IX. Diptcra. D. W. CoqyiLLETT. Issued Dec. 7, 1900. " 22. Papers from the Ilarriman Alaska Expedition. — X. Neuropteroid Insects. Nathan Banks. Issued Dec. 20, 1900. " 23. Papers from the Harriman Alaska Expedition. — XI. Arachnida. Nathan Banks. Issued Dec. 20, 1900. vi Brochure 24. Papers from the Harriman Alaska Expedition. — XII. Lepidoptera. Harrison G. Dyar. Issued Dec. 20, 1900. *' 25. Papers from the Harriman Alaska Expedition. — XIII. Heteroptera. O. Heidemann. Issued Dec. 20, 1900. 4' 26. Papers from the Harriman Alaska Expedition. — XIV. The Sphegoidea and Vespoidea. Trevor Kin- CAiD. Issued Dec. 20, 1900. " 27. Papers from the Harriman Alaska Expedition. — XV. Orthoptera. A. N. Caudell. Issued Dec. 20, 1900. " 28. Papers from the Harriman Alaska Expedition. — XVI. Aphididse. Theo. Pergande. Issued Dec. 20, 1900. " 29. Papers from the Harriman Alaska Expedition. — XVII. Formicidie. Theo. Pergande. Issued Dec. 20, 1900. " 30. Papers from the Harriman Alaska Expedition. — XVIII. Coleoptera. E. A. Schwarz. Issued Dec. 20, 1900. " 31. Papers from the Harriman Alaska Expedition, — XIX. Psyllidae. E. A. Schwarz. Issued Dec. 20, 1900. "■ 32. A Contribution to the Study of the Insect Fauna of Human Excrement. L. O. Howard. Issued Dec. 28, 1900. " 33. A Theatrical Performance at Walpi. J. Walter Fewkes. Issued Dec. 28, 1900. " 34. A Collection of Small Mammals from Mt. Coffee, Liberia. Gerrit S. Miller, Jr. Issued Dec. 28, 1900. " 35. Fossil Land Shells of the John Day Region with Notes on Related Living Species. Robert E. C. Stearns. Issued Dec. 28, 1900. " 36. Preliminary Revision of the North American Red Foxes\ C. Hart Merriam. Issued Dec. 28, 1900. Vll ILLUSTRATIONS. PLATES. I. The Allegan Meteorite. II. The Allegan Meteorite. III. The Allegan Meteorite. IV. The Allegan Meteorite. V. The Allegan Meteorite. VI. The Mart Iron. VII. Harrimania maculosa gen. et sp. nov. VIII. Decapod and Stomatopod Crustaceans from Brazil. IX. New Fishes from Brazil. X. An exfoliated peak of quartz-monzonite at Sete de Setembro, Rio MucLiry. XL Deep cut near Bittencourt, on the Alagoas railway, exposing sedimentary beds resting against granite. XII. Granite hills near Gamelleira on the Alagoas railway, State of Alagoas, Brazil. XIII. Pycnopodia helianthoides. XIV. Pycnopodia helianthoides. XV. Salix amplifoUa sp. nov. XVI. Alaska Hepaticie : Gym?to?}iitri?au obtjisiitn and Anastro- phylluju reichardtii. XVII. " " Ccphalozia leucantha and Diplophyl- leia pi i cat a. XVIII. " " Diplophylleia plicata. XIX. Alaska Bryozoa. XX. Alaska Bryozoa. XXI. Alaska Bryozoa. XXII. Metamorphoses of Alaska Coleoptera. XXIII. Metamorphoses of Alaska Coleoptera. XXIV. Metamorphoses of Alaska Coleoptera. XXV. Metamorphoses of Alaska Coleoptera. XXVI. MctanKjrphoses of Alaska Coleoptera. XXVII. Alaska Neuropteroids. XXVIIL Alaska Neuropteroids. XXLX. Alaska Arachnids. XXX. I. Afusrina stal>ii/a>/s ; 2. Piophila casei. XXXI. I. mppclatcs Jlavipcs; 2. ^Drosophila ampelophaga ; 3. / 'sc/idop yrcllia cornicina. XXXII. Serpent screen. XXXin. Struggle of serpents witli ' Mudheatls.' XXXIV. Serpent elligies. viii XXXV. Fossil land shells from the John Day region. XXXVI. Molar teeth of Red Foxes. XXXVII. Molar teeth of Red Foxes. TEXT FIGURES. 1. Head of Corollana acuticauda. 2. Abdomen of Corollana acuticauda. 3. Bopyrus alphei. 4. Bopyrus alphei. 5 Coast of Brazil from Rio to Bahia. 6* Section along line of Bahia and Minas Railway, Brazil. 7. Plateau between Lagoa do Norte and Lagoa Manguaha, Brazil. 8. Map of Alagoas Railway, Brazil. 9 Section along main line of Alag6as Railway. 10. Islands in North China Sea, visited by Dr. Abbott. 11. Skulls of Mus tiomanicus, M. siantanicus, and M. ta??ibu- 1 2 Skulls of Sciurus ynimiculus and S. prevostii. 13'. '' " Sciurus notatus, S. klossi, and S. tenuirostris. 14. " " Tupaia malaicana. i^[ u u E?7iballomira anambensis and E. pe?iinsularis. 16. Skull of Pier opus lepidus. 17. Compsomyia macellaria. 18. Lucilia ccesar. * 19. Calliphora erytJirocephala. 20. Sarcophaga sarracenice. 21. Sarcophaga assidua. 22. Phorfnia terrcenovcB. 23. Musca domestica. 24. Morellia micans. 25. Muscina stabzdans. 26. Myospila meditabunda. • 27. Stomoxys calcitra?is. 28. Pseudopyrellia cornicina. 29. Ho?nalomyia brevis. 30. Limnophora arcuata. 31. Phorbia cinerella. 32. Sepsis violacea. 33. Nemopoda mi?tuta. 34. Limosina albipeiinis. 35. Limosina crass imana. 36. Sphcerocera subsultans. ix 37- ScatopJiaga fiircata. 38. Scatophaga fiircata. 39. Sciurus rufobrachiatiis liberictis and S. riifobrachiatus. 40. Skull of Dasymys rufidiis. 41. " " Arvicantkis planifro?zs and A. tinivitiatzis. 42. " " Alyosorex muricauda. 43. " " Pipistrelhis pipistrellus Vind P. mhitiscuhis. WASHINGTON ACADEMY OF SCIENCES OFFICERS FOR 1900 President Charles D. Walcott Vice-Presidents From the Anthropological Society W J McGee Biological Society F. V. Coville Chemical Society H. N. Stokes Columbia Historical Society John A. Kasson Entomological Society Theodore Gill Geographic Society A. Graham Bell Geological Society G. K. Gilbert Medical Society S. C. Busey Philosophical Society Geo. M. Sternberg secretary Treasurer Frank Baker Bernard R. Green Managers Class of igoi Class of 1902 Class of 1903 Marcus Baker L. O. Howard F. W. Clarke H. S. Pritchett J. W. Powell C. Hart Merriam G. M. Kober Carroll D. Wright Whitman Cross Standing Committees for the Year 1900 Committee on Rules Committee on Functions Carroll D. Wright F. W. Clarke Richard Rathbun C. L. Marlatt H. M. Paul C. E. Munroe Committee on Finance Committee on Publication Co?nmittee on Lectures Whitman Cross C. Hart Merriam W J McGee Bernard R. Green John Hyde Geo. M. Sternberg L. O. Howard Marcus Baker Willis L. Moore W. H. Ashmead O. H. Tittmann Geo. M. Kober F. W. Hodge F. A. Lucas H. N. Stokes W. H. Holmes Committee on Building- S. C. Busey H. W. Wiley Bernard R. Green L. O. Howard John M. Wilson A. Graham Bell Arnold Hague Whitman Cross E. M. Gallaudet John A. Kasson W J McGee Geo. M. Kober Geo. M. Sternberg H. M. Paul xi MEMBERS OF THE WASHINGTON ACADEMY OF SCIENCES December 31, 1900 Patrons Mrs. Gardiner G. Hubbard Thomas F. Walsh Mrs. Henry L. Higginson James W. Pinchot Gifford Pinchot Cleveland Abbe George N. Acker Henry Adams Samuel S. Adams Cyrus Adler W. II. Ashmead Vernon Bailey Frank Baker Marcus Bakeu L. A. Bauer George F. Becker A. Graham Bell I. Bermann F. H. BiGELOW Rogers Birnie J. Wesley Bovee J. H. Bryan S. M. Burnett S. C. BUSEY Frank K. Cameron M. R. Campbell V. K. Chesnut F. W. Clarke Thomas A. Claytor G. Wythe Cook O. F. Cook D. W. CfX^UILLETT Frederick V. Covii.l Whitman Cross N. H. Darton A. P. Davis David T. Day J. S. Diller H. G. Dyar William Eimbrck S. F. Emmons Resident Members b. w. evermann J. W. Fewkes A. K. Fisher Miss Alice C. Fletcher Robert Fletcher James M. Flint Henry D. Fry E. M. Gallaudet B. T. Galloway Henry Gannett G. K. Gilbert Theodore Gill R. U. Goode J. Howard Gore Henry S. Graves Bernard R. Green E. L. Greene J. G. Hagen Arnold Hague William Harkness R. A. Harris W. T. Harris C. W. Hayes R. T. Hill W. F. Hillebrand F. W. Hodge W. II. Holmes L L. O. Howard John Hyde Joseph Taber Johnson Geor(je W. Johnston W. W. Johnston SOFIE NORDHOFF JUNG John A. Kasson Arthur Keith James Kerr xii A. F. A. King C. H. A. Kleinschmidt F. H. Knowlton Geo. M. Kobeu Stephen J. KObel D. S. Lamb S. P. Langley A. Lindenkohl G. W. LiTTLEHALES F. A. Lucas Thos. E. McArdle W J McGee G. L. Magruder H. L. Marindin C. L. Marlatt C. F. Marvin Washington Matthews Geo. W. Melville C. Hart Merriam G. P. Merrill James C. Merrill John H. Metzerott W. M. Mew G. S. Miller, Jr. James Mooney Willis L. Moore A. T. Mosman C. E. Munroe E. W. Nelson F. H. Newell H. G. Ogden T. S. Palmer Edward W. Parker H. M. Paul A. C. Peale A. J. Pieters GiFFORD PiNCHOT J. W. Powell W. B. Powell J. F. Pratt H. S. Pritchett J. R. Procter G. R. Putnam Miss M. J. Rathbun Richard Rathbun Robert Reyburn Chas. W. Richardson C. W. Richmond Robert Ridgway Miles Rock J. N. Rose D. E. Salmon c. a. schott Charles Schuchert E. A. Schwarz E. A. de Schweinitz T. J. J. See A. R. Shands A. R. Spofford T. W. Stanton L. Stejneger Geo. M. Sternberg Mrs. Matilda C. Stevenson C. W. Stiles H. N. Stokes L S. Stone W. T. Swingle O. H. TiTTMANN C. H. Townsend F. W. True Frank Tweedy Chas. D. Walcott Lester F. Ward R. B. Warder C. K. Wead David White Milton Whitney H. W. Wiley Bailey Willis H. M. Wilson John M. Wilson A. F. Woods William C. Woodward Carroll D. Wright Walter Wyman Non-Resident Members I'^uANK Adams McGill University, Montreal, Canada. J. A. Ai-i.EX American Museum of Natural History, New York City. IvruAN Ai.le.v Andrews Johns Hopkins University, Baltimore, Md. Caul Barts Brown University, Providence, R. I. C. E. Bessev University of Nebraska, Lincoln, Neb. Major \V. H. Bixby U.S.EngineersOffice, Cincinnati, Ohio. Major Stanhope E. Blunt. .Rock Island Arsenal, 111. j. C. Branner Stanford University, Calif. William Bkkwstkr 145 Brattle St., Cambridge, Mass. EuNE.sT W. Brown Ilaverford College, Haverford, Penn. II. C. BuMPUS Brown University, Providence, R. I. Geo. C. Caldwell Cornell University, Ithaca, N. Y. W. W. Campbell Lick Observatory, Mt. Hamilton, Calif. J. McKeen Catteli Garrison-on-Hudson, New York. T. C. Chamberlin L^niversity of Chicago, Chicago, 111. C. P. Chandler Columbia University, New York Citv. \Vm. \i. Clark Johns Hopkins University, Baltimore, Md. P. S. Conner 215 West 9 St., Cincinnati, Ohio. John M. Coulter l^niversity of Chicago, Chicago, 111. .Major William Crozier Ordnance Office, War Department, Washington, D. C. E. S. Dana New Haven, Conn. (jKokgk Davidson 2221 Washington St., San Francisco, Calif. S.\.Mi i-.L \V. Dike Aulnirndale, Mass. (iEo. A. DoRSEV Field Columbian Museum, Chicago, 111. C. P>. Dudley Drawer 334, Altoona, Penn. j. K. Eastman Andover, N. H. Cakl H. I'vKiKNMANN Indiana University, Bloomington, Iiul. 1^- I • I'>LV University of Wisconsin, Madison, Wis. i'>. K. E.MEusoN Amherst, Mass. I.ivi\c;ston Fahhand Columbia University, Xcw York City. \\'ALri:u Faxon Museum Comparative Zoology, Cam- bridge, Mass. '<• 11. Vn/. uS Arlington St., Boston, Mass. S. .\. l-'oHHEs University of Illinois, Champaign, 111. \V. C. I'oKi) Pul)Iic Library, Boston, Mass. Simon II. (Im.k Cornell rnivcrsity, llliaca, \. ^■. IIknkv p.. (JAitONKi! 5.1 Stimsou Ave., Providence, R. I. xi\ Samuel Garman Museum Coinparalive Zooloj^y, Cam- bridge, Mass. C. H. GiLRERT Stanford University, Calif. Geo. L. Goooai.e Harvard University, Cambridge, Mass. E. R. L. Gould 2S1 4 Ave., New York City. Thomas Gray Terre Haute, Indiana. Wm. S. Halsted 1201 Eutaw St., Baltimore, Md. Edward Hartwell 5 Brimmer St., Boston, Mass. E. W. HiLGARD Berkeley, Calif. C. H. Hitchcock Hanover, N. H. J. A. Holmes Chapel Hill, N. C. Alpheus Hyatt Boston Society of Natural History, Bos- ton, Mass. J. P. Iddings. University of Chicago, Chicago, 111. Maj. James M. Ingalls War Department, Washington, D. C. Abraham Jacob: no West 34 St., New York City. E. G. Janeway 36 West 40 St., New York City. J. W. Jenks Cornell University, Ithaca, N. Y. Emory R. Johnson 75 Corcoran Bldg., Washington, D. C. David Starr Jordan Stanford University, Calif. James E. Keeler^ Lick Observatory, Mt. Hamilton, Calif. W. W. Keen 1729 Chestnut St., Philadelphia, Pa. Howard A. Kelley 14 18 Eutaw St., Baltimore, Md. Jas. F. Kemp Columbia University, New York City. David Kinley University of Illinois, Champaign, 111. Andrew C. Lawson University, Berkeley, Calif. Joseph LeConte University, Berkeley, Calif. J. H. Long 2421 Dearborn St., Chicago, 111. Gen. William Ludlow i744 G St., Washington, D. C. Alexander Macfarlane Lehigh University, South Bethlehem, Pa. J. W. Mallet University of Virginia, Charlottesville, Va. Richmond Mayo-Smith Columbia University, New York City, Clarence B. Moore 1321 Locust St., Philadelphia, Pa. E. W. MoRLEY Adelbert College, Cleveland, O, Hugo MOnsterburg 7 Ware St., Cambridge, Mass. S. N. D. North 70 Kilby St., Boston, Mass. Arthur A. Noyes Mass. Institute of Technology, Boston, Mass. C. C. Nutting University of Iowa, Iowa City, Iowa. Arnold Ortmann Princeton University, Princeton, N. J. ' Deceased. XV Henry F. Osborn Columbia University, New York City. William OsLER.. i West Franklin St., Baltimore, Md. Samuel L. Penfield Yale University, New Haven, Conn. Edward C. Pickering Harvard University Observatory, Cam- bridge, Mass. Benjamin Osgood Pierce — Jefferson Physical Laboratory, Cam- bridge, Mass. L. V. Pirssen Yale University, New Haven, Conn. T. Mitchell Prudden i6o West 59 St., New York City. F. W. Putnam Peabody Museum, Cambridge, Mass. Lieut. Col. C. W. Raymond. Philadelphia, Penn. Harry Fielding Reid Johns Hopkins University, Baltimore, Md. Ira Remsen Johns Hopkins University, Baltimore, Md. William E. Ritter University, Berkeley, Calif. J. D. Runkle Mass. Institute of Technology, Boston, Mass. Frank Russell. 25 Ware St., Cambridge, Mass. Rollin D. Salisbury University of Chicago, Chicago, 111. C. S. Sargent Arnold Arboretum, Jamaica Plain, Mass. W. B. Scott Princeton, N. J. E. R. Seligmann Columbia University, New York City. Nicholas Senn 532 Dearborn St., Chicago, 111. N. S. Shaler Harvard Universit}^, Cambridge, Mass. Eugene A. Smith University, Alabama. John B. Smith Agricultural Experiment Station, New Brunswick, N. J. R. E. C. Stearns 1025 East 18 St., Los Angeles, Calif. Simon vSterne 56 Beaver St., New York City. J.J. Stevenson University Heights, New York City. Ormond Stone Univ. of Virginia, Charlottesville, Va. R. H. Thukston Cornell University, Ithaca, N. Y. Chas. R. Van Hise I'uiversity of Wisconsin, Madison, Wis. F. P. Venable Cliapel Hill, N. C. Wm. H. Welch Johns Hopkins University, Baltimore, Md. R. p. Whitkikld American Museum of Natural History, New York City. Burt G. Wilder Cornell Ihiiversity, Ithaca, N. Y. W. F. WiLi.cox Cornell University, Ithaca, N. Y. H. S. Williams ^ ale University, New llavcu, Conn. Talcott Williams 916 Pine St., Philadelphia, Pa. Edmund B. Wilson Columbia University, New York City. J. C. Wilson 1437 Walnut St., Philadelphia, Pa. W. P. Wilson 233 South 4 St., Philadelphia, Pa. R. S. Woodward Cohnnbia University, New York City. Dean C. Worcester Manila, Philippine Islands. C. A. Young Princeton University, Princeton, N. J. Alexander Ziwet University of Michigan, Ann Arbor, Michigan. ABSTRACT OF PROCEEDINGS. January 17, 1900. Annual meeting. Election of officers. February 21, 1900. Reception to the American Institute of Mining Engineers in the Corcoran Gallery of Art. May 29, 1900. Business meeting. Amendment to by-laws pre- scribing that the annual meeting shall be held on the third Thursday in January in each year. November 9, 1900. Business meeting. Admission of additional members. Purchase of a lot for a building site. The following addresses have been delivered under the auspices of the Academy : February 13, 1900. Annual address of the President of the An- thropological Society, Mr. W J McGee, on ' The Cardinal Principles of Science.' December 8, 1900. Annual address of the President of the Philo- sophical Society, General George AI. Sternberg, M.D., on ' Malaria.' Also the following series of addresses for the exposition of Pho- tography as an Aid to Research : February (5, igoo. The History and Chemistry of Photography, Lyman J. Briggs. Photomicrography, Gen. Geo. M. Sternberg. Photography in Vegetable Physiology and Pathology, Jvl, B. Waite. February 75, igoo. Photography as an Aid in the Study of Plants, F. V. Covii,i-e. Photography in Animal Physiology and Pathology, Dr. D. S. Lamb. Photography as an Aid in Medicine and Surgery, Dr. W. C. Borden. Marc/i (5, igoo. Photography as an Aid in Zoologv, Dr. C. Hart Merriam. Outfits and Methods of the Bird Photographer, Frank M. Chapman. Photograpiiy as an Aid in Entomology, Dr. L. O. Howard. March 20, igoo. Photography as an Aid in Surveying, D. B. Waiwvright. Pliotograi)hic Surveying with the Ordinarv CamcMa, Prof. H. F. Knii.. xviii Geologic Research as Illustrated by Photography, G. K. Gilbert, N. H. Darton, and F. L. Ransome. April 3^ I goo. Photography applied to Sidereal Astronomy, Dr. H. S. Pritchett. Photography in Astrophysics, Prof. J. G. Hagen. Photography as used in Eclipse Observations, Prof. S. J. Brown. Photography in Physics, Prof. Frank H. Bigelow. April i8^ I goo. Photography in Anthropology, W. H. Holmes. Photography in Physical Anthropology, W J McGee. Photography in History and Literature, Dr. Cyrus Abler. May z, I goo. The Applications of the Kromskop, Frederick E. Ives. The Applications of the Phantascope, E. Francis Jenkins. Color Photography and Color Printing, N. H, Darton. The addresses of April i8, 1900, were followed by a conversazione in honor of the National Academy of Sciences; those of May i, 1900, were complimentary to the Congress of American Physicians and Surgeons. For the winter of 1900-1901 a series of addresses has been arranged on The Progress and Tendency of Science during the Nineteenth Cen- tury. The first of this series was given on December 11, 1900, by Professor F. W. Clarke, on ' The Progress and Tendency of Chem- istry.' PROCEEDINQS WASHINGTON ACADEMY OF SCIENCES Vol. II, pp. I-I2. March 14, 1900. CARDINAL PRINCIPLES OF SCIENCE.^ By W J McGee. Science may be defined as consciously organized knowledge. It rises far above that unconsciously integrated experience called instinct in animals and intuition in men ; it passes beyond that semi-conscious summation of experience called common-sense ; it even transcends that consciously integrated but only half- formulated experience called wisdom, which grows with the years and ends with the life of the sage ; yet it spans all these and other grades of actual knowledge, and seeks to reduce them to simple order. As the general definition implies, Science involves (i) methods of acquiring knowledge, (2) methods of formulating or organ- izing knowledge, and (3) the sum of knowledge (including knowledge of methods) acquired and formulated in accordance with experience. There are several branches of Science, each defined by the phenomena with which it directly deals ; yet the methods are alike, and all, from eldest-born Astronomy to youngest-born Anthropology, have contributed to the methods as well as to the sum of consciously organized knowledge. ^ Reduced to simple terms, the scientific methods of acquiring knowledge maybe defined as (i) observation, (2) generalization, and (3) inference ; though the three processes overlap to the ex- tent that practically there is no observation without some gener- alization, no generalization apart from observation and unaided » Address as President of the Anthropological Society of Washington ; deliv- ered before the Washington Academy of Sciences and the Affiliated Societies, February 13, 1900. Proc. Wash. Acad. Sci., March 1900. ^ MCGEE by some inference, and no inference completely dissevered from observation and generalization — indeed, generalization is little more than spontaneous expansion of simple observation, and inference little else than normal extension of generalization through the aid of memory. The source of scientific knowl- edge is experience, though some part of the experience maybe indirect, /. ^., that of other individuals and generations com- municated in ways implying measurable similarity in experi- ence. The methods of acquisition are shared not only by the several sciences but by all other lines of knowledge ; yet the proportion of intuitive (or unconsciously integrated) experi- ence called deductive knowledge decreases, while the ratio of consciously organized experience called inductive knowledge increases from pre-science to Science, and from branch to branch of the special sciences — Astronom}', Chemistr}^ Ph3'sics, Phytology, Zoology, Geology, Anthropology — in about the order of their historical development. It is in the methods of organizing knowledge that Science attains its most distinctive aspect ; and these methods are re- ducible to a few simple formulas, of which the greater part are expressed in phrases made familiar by long use in scientific speech and literature. Each of these formulas sums a vast body of experience, each was contributed b}^ one or more of the family of sciences, and all are sustained jointly by the several special sciences ; and since the}^ are not shared b}^ other branches of knowledge, they may be fairly held to represent the Cardi- nal Principles of Science. The origin of Science can be scried but dimly amid the shadows of antiquity, yet enough of the earliest traces may be caught to show that organized knowledge began with counting and grew into mathematics ; there are, indeed, unmistakable in- dications that primitive numbers were largely mystical or alma- cabalic' ; but the indications are equally clear that the counting whence the number-systems sprang expressed at least partly conscious recognition of the veritable existence and persistence ' As described in ' The Beginning of Mathematics,' American Anthropologist, New Series, Vol. I, 1S99, pp. 646-674, CARDINAL PRINCIPLES OF SCIENCE 3 (or the integrality) of the things counted — and this may fairly be deemed the inception of Science. Patient search among the lighter shadows of less remote antiquity discovers traditions and records of an epoch in which men living under the clear skies of arid regions were impressed with the apparent stability of the star-decked firmament, and gradually grouped the heavenly bodies into systems ; the early grouping was largely mystical or astrological, the primal constellations were mainly beast-gods translated to the heavens, and the crystal spheres invented to carry the greater luminaries and planets were but supernal fig- ments ; yet the crude system marked conscious effort to organ- ize experiences of consciously recognized things — and it is fair to date the definite conception of Science from this primitive Astronomy. The gradually multiplying records of olden times indicate that the mental processes fixed by counting and star- gazing were steadily extended to other things, which were slowly recognized as entities: Even before 450 B. C, Parmen- ides had held that substance may neither become nor perish, and about this early date Empedocles taught (of course in met- aphysical wise) that the ultimate particles of matter must be in- divisible, immutable, and indestructible ; while Democritus summed the philosophical opinion of his period (460 B. C. to 357 B. C.) in a proposition which may be rendered, " Out of nothing nothing arises ; nothing that is can be destroyed ; change is only combination or separation of atoms ; " and a like vievj^ was held by Epicurus and his disciple Lucretius dur- ing the century before the opening of our era. The progress of the notion lagged during the Dark Ages but revived with the Renaissance and the physical researches of Gassendi, Leibnitz, and especially Lavoisier (i 743-1 794), who showed experiment- ally that matter is neither lost nor gained in chemical change — and this experimental demonstration ma}^ justly be held to mark the full birth of Science as consciously organized knowledge. i\.t the same time the demonstration established that cardinal principle of Science commonly phrased The Indestructibility of Matter. Essentially, the integration expressed in this formula involved in the first place the extension of macroscopic obser- vation into microcosmic nature, and in the second place the ex- MCGE E tension of human faculty to pu-rposive pre-combinations of such sort as to afford prompt verification of inferences and confident demonstration of natural relations ; so that this first pillar of Science may be deemed the gift of Astronomy as to material, and of Chemistry as to structure and useful setting. The doctrine of the indestructibility of matter was at once the fruit of a philosophical renovation and the seed of an in- tellectual reformation ; for, as it ripened in mind after mind and was sown broadcast by the slow but persistent processes of the times, it generated new inquiry and experiment, which at the same time gave sustenance to many hungry minds and fur- nished fresh seed for yet other minds. Among the conse- quences of the intellectual quickening were certain exper- iments by Benjamin Thompson, or Count Rumford, which, when described in 1798, served to transfer the discussion of temperature from the domain of metaphysics to that of physical inquiry and to demonstrate that heat is a manifestation of mo- tion; Thompson was soon followed by Sir Humphry Davy and half a century later by Dr. J. P. Joule, who repeated and extended the experiments and fixed the mechanical equivalent of heat. Then came a brilliant coterie of physicists, Grove, Faraday, Helmholtz, Tyndall, and others, who verified the de- terminations by means of special experiments and by compar- ison with general human experience, established the inter- changeabiHty (or correlation) of forces, and gradually organized their growing knowledge in a system revealing another cardinal principle — the conservation of energy, or (preferably, in view of Thompson's terminology and of later researches) The Per- sistence of Motion. This new principle met a mental need, and found so many applications that it came to be regarded as the most important discovery of the centur}^. Originally the gift of Physics, it was soon extended into the realm of organic life by Liebig and others in Germany and by Carpenter in Britain, and only a few years later it was pushed into the realm of mental action by LeConte and Barker in America. For half a century the inconstructible and indestructible motion en- livening the universe was assumed to be constant only in the universe as a whole and constantly variable in the constituent CARDINAL PRINCIPLES OF SCIENCE 5 bodies ; this assumption served to pair fundamental notions of matter and motion, the one persisting'only in the particle and the other only in the sum of particles, in such ^manner as to satisfy the dualistic instinct expressed in most earlier philoso- phies ; but at the same time it introduced an indefinite, if not forever indeterminate, element, into the formula — for under the assumption the finality of the formula can never be ascer- tained until the universe is measured and weighed to its ut- termost limits. An alternative assumption, recently proposed by Powell, is that motion, like matter, is persistent in the ultimate particle ; this assumption has the merit of harmon- izing principles otherwise (apparently) discordant, and the special excellence of integrating general human experience ; but it still awaits that direct demonstration through laboratory experimentation, so prearranged as to test all possible infer- ences, which alone carries conviction to conservative minds. As Lavoisier's doctrine of indestructibility spread, the method of observation under voluntarily controlled conditions was ex- tended into new fields ; and even while Joule's mechanical ex- periments were still progressing at Manchester, Darwin was arranging a series of vital experiments at Down and comparing his results with the voluminous observations recorded by natur- alists in all parts of the world. Meantime Alfred Russell Wallace and Herbert Spencer were also seeking to organize anew the facts collected on the Linnaean plane, already become chaotic by reason of their very number ; and when the three thinkers independently generalized the teeming experiences of organic life in terms of sequence, the coincident opinions and the in- comparable thoroughness of Darwin's methods combined to force a new principle on an unwilling world of contemporary thought. At first the Darwinian doctrine was burdened by its own infantile feebleness — for there is no Minervan birth in Science — and bound by the swaddlings of scholasticism ; but support came from all sides, and it grew apace and soon be- came the sturdiest of that trinity of scientific principles recog- nized up to the sixth decade of the nineteenth century. For some years the doctrine was largely limited to the organic realm, and was deemed applicable solely to the evolution of animal MCGEE and vegetal species ; later it was extended (first intuitively and then purposively) to other realms of nature, and was found equally applicable to species of the sub-organic realm, such as minerals and mountains, and to those of the super-organic realm, such as knives and nations — when it became evident that the principle involves involution as well as that evolution so con- spicuous in the organic world. With this extension the doctrine becomes The Develo-pment of Sfecies^ and this phrase expresses the third cardinal principle of Science — the gift of biology, i. e.y of phytology and zoology jointly. Naturally the promulgation of three revolutionary doctrines within a century (1760-1860) produced an intellectual surfeit. The principle of indestructibility had lain germinant for two millenniums, and so found men's minds ready for the final plant- ing ; even the principle of persistence met a need of thought, and encountered little opposition ; but the principle of develop- ment found most mental fields already stocked with long-set growths and little prepared for the latest seeding. A conse- quence of the indifference or antipathy of the average mind was voluminous discussion, with the spontaneous separation of thinkers into opposing schools ; and in due course Science be- came a cult, and Thomas H. Huxley its most vigorous expo- nent. As the discussion proceeded, the principle of development was rounded out and applied to new problems ; and as the ap- plications multiplied, it was found to present many analogies with the principles of indestructibility and persistence. Mean- time a science of the earth was rapidly crystallizing about Sir Charles Lyell's idea of uniform process in earth-building — the ' uniformitarianism ' of that day of redundant doctrines — while under the masterly interpretation of Huxley and others, paleon- tology was resurrecting the record of early life on the earth ; and the several movements eventually joined in the formulation of a fourth principle, complementary to those of earlier decades and so harmonious with them as to facilitate the apprehension and adoption of the entire series. This fourth principle has ever smce been regarded as a (if not the) cardinal principle of science ; in Huxley's happy phrase it is The Uniformity of Nature. The principle must be credited largely to Geology CARDINAL PRINCIPLES OF SCIENCE 7 (including paleontology), but partly to the biotic and physical sciences jointly. At this end of the century, these four principles form the commonly accepted platform of Science : the indestructibility of matter, the persistence of motion, the development of species, and the uniformity of nature. There may be, indeed is, a question as to whether they constitute the entire platform of hu- man knowledge ; but in the minds of scientific men there is no question as to the verity of these principles so far as they go. True, the temperate scientist must admit the possibility that any or all of the principles may be erroneous ; but he does so, if at all, in full realization that the admission is tantamount to denial of the truth and trustworthiness of experience as a source of knowledge. The scientist cannot deny that the cosmos is shrinking toward nothingness, or growing by miraculous accre- tion toward greater magnificence ; he is very far from denying that the universe may be kept in motion by some extra-cosmic source of power ; he hardly ventures to dispute the transi- tional halting-place held by those who claim that at least the first life and the highest mind transcend natural development and demand a special supernatural explanation ; in certain cir- cumstances he is deterred by the tolerance of good breeding from denying that the world was fashioned from a lump of mud brought up by a muskrat from the bottom of the sea, as taught in sincere, albeit primitive, philosophy — he can only say that all such explanations of things are outside the range of experience, and hence beyond the domain of Science. It is also true that the unscientific thinker — whose name is Legion — enjoys the fullest freedom of rejection of any or all of the principles ; and it can only be said that thereby he keeps without the straight- laid fields of Science and within the broad and often attractive purviews of not-science. Yet a significant sign of the times is found in the fact that most men of civilized and enlightened lands are coming to appreciate the coin of experience above the dust of tradition, and are gradually entering, whether inten- tionall}' or not, into the ways of Science. Only a generation ago the very name of Science was the symbol of a cult to one class of thinkers and a juggler's gaud to another class ; to-day MCGEE Science is an actual part of the everyday life of all enlightened folk. So, at this end of the century, four principles of Science are held implicitly or proclaimed explicitly throughout the lead- ing nations ; and the adherents of the principles justify their confidence by the unimpeachable testimony of experience. On summing the principles forming the present platform of Science, it becomes manifest that they are interrelated in such wise as to form a harmonious series, each dealing with an as- pect of nature and the whole covering all nature in its more conspicuous aspects ; it becomes manifest, too, that the four principles are alike in two respects : in the first place the}^ are each and all integrations of experience along the lines to which they respectively pertain ; in the second place they each and all rest on the postulate that experience furnishes a key — and the only key — to nature. This correspondence would seem to indicate that the four principles might themselves be inte- grated in terms of their common attributes ; it also suggests that the principles may have been really integrated, albeit intuitively or subconsciously only, even before they were finally formulated. Thus, the principle of indestructibility was but a generalization of experiences of indestructibility, yet it could not have become clear to any mind devoid of the assumption (howsoever implicit or vague) that experience accurately reflects actualities; the principle of persistence was similarly a generalization of ex- periences, yet it could not have been grasped without the assumption that experiences of motion and duration are verit- able ; so, too, the principle of development involves the postu- late that the sequence pictured in the mind is the reflection of a real sequence in nature ; while the principle of uniformity de- mands the assumption that the mind of man is a faithful mirror of nature. Doubtless the essential postulate remained unformu- lated and half unrealized because of the preoccupation of the knowledge makers ; Lavoisier was too busy over chemical re- actions to give much thought to the mechanism of his own mind, Joule confined his apperception to extra-mental phenomena, Darwin remained naively negligent of his ownadmirable men- tations, and even Huxley deliberately dropped the veil of un- CARDINAL PRINCIPLES OF SCIENCE 9 tliinkableness between his conscious self and the mysteries of his subconsciousness ; yet these men made no observations, framed no generalizations, formed no inferences, without confident real- ization that experience is a reflex of nature — and it seems evi- dent that without this realization the signal advances in knowl- edge summed in the cardinal principles could not have been made. On seeking explanation of the constantly implied but never declared confidence reposed by the knowledge-makers alike in their own experiences and in their own mental operations, it is soon seen that the two bases of scientific procedure are closely akin ; for the evident source of confidence in experience is found in experience itself, while the source of confidence in mentality just as evidently arises in experience of mental work- ing. Now in every stage of culture the several faculties are manifestly coordinated, at least to the extent that bodily activi- ties affect thought and thinking measurably controls action ; hence the suggestion naturally arises that the largely intuitive development of confidence in external realities and in the power of the mind to grasp and interpret them was a normal sequence of the Renaissance following the Dark Ages, in which men turned to the most inspiring duties of their kind — the twin tasks of nature-conquest and nation-building ; and, if the suggestion be valid, it would seem to follow that the confidence in Nature and Mind expressed by successful effort was rather a forerunner than a mere attendant of those advances in knowledge marked by the framing of the cardinal principles. Fortunately the sug- gestion is supported by contemporary testimony. Nearly two centuries before Lavoisier, Francis Bacon laid a foundation for definite knowledge in the Novum Organum, the cornerstone of which appears in the initial aphorism of the first of the two books comprised in this memorable essay. Rendered into the more trenchant and vigorous language of the present, this aphorism reads : Man, as the minister and interpreter of nature, does and understands as much as his observations on the order of nature, either with regard to things or the mind, permit him, and neither knows nor is capable of more.^ 'Translation edited by Joseph Devej and published in Bohn's Philosophical Library under the title 'The Physical and Metaphysical Works of Lord Bacon.' London, 1889 (p. 3S3). lO MCGEE Divested of immaterial qualifications, and interpreted in the light of other propositions in the same book, this aphorism is a specific affirmation that man's mind is a mirror of Nature, re- flecting observations of Nature and nothing more — i. e., it is an explicit definition of the platform implicitly accepted by Lavoisier and Joule and Darwin and Huxley in their epoch- making researches. True, these men made little acknowledg- ment to Bacon, and at least one of them mildly decried his work ; yet the Novum Organum was one of the most notable treatises of its time, and must have been read widely and at least measur- ably assimilated into the common knowledge of succeeding cen- turies ; so it seems evident that the Baconian concept lay ger- minant, much as did the Greek concept of indestructibility, though for a lesser period — and hence that the framers of the four cardinal principles were debtors to Bacon none the less that the debt was not consciously recognized. Long after Bacon, and even after the biotic and geologic sciences bore fruit in cardinal principles, the Science of Man began to take shape. As in other sciences, the observations began with the rare, the remote, and the abnormal, and grad- ually passed toward the near and the normal ; and as the obser- vations became chaotic by reason of their number, they were from time to time integrated in such manner as to raise the sys- tem to successively higher and higher planes. A quarter-cen- tury ago the leaders of Anthropology were beginning to classify mankind in terms of mind-shaped activities and activital prod- ucts— arts and industries, social institutions, languages, and beliefs with their attendant ceremonies and symbols. One of the consequences of this classification was the discovery that the human activities and their products are so frequently alike in widely separated regions as to indicate that they must have arisen spontaneously and independently among unrelated peo- ples ; and twenty years ago Powell formulated this discovery into a law of ' activital similarities' — or a law of cultural coinci- dences, as it might now be called. Further observations supported the first discover}', and iinally gave warrant for the inference that the human mind, wheresoever placed, reflects its environment with striking fidelity. Brinton held that this par- CARDINAL PRINCIPLES OF SCIENCE I I allelism between Man and Nature expresses a unity of the human mind imposed by some extraneous power ; but the later researches indicate that the relation varies with the degree of cultural development — /. a laclirymal tubercle; audital bullas smaller and less inflated; incisive foramina larger and more broadly open anteriorly; zygomata more sharply elbowed at anterior base (as seen from above) ; jugal narrower. Contrasted with kadiacensis the anterior end of frontal is narrower and fails to develop a lachr^'mal tubercle; bulla' more inflated, blunter, and more rounded anteriorly; molars slightly larger. In young specimens (and one adult, apparently in left-over winter pelage) the fur is longer and softer, the upperparts are more buffy fulvous, and the white of the underparts is wasl^ed with yellowish buff. THE IIARRIMAN ALASKA EXPEDITION 23 Measitremcfits. — An adult female from type locality: total length 188; tail vertebne 43; hind foot 23. MIGROTLTS ABBREV'IATUS^ FISHER! subsp. nov. St. Matthew Island Vole. lype from St. Matthew Id., Bering Sea. No. 97976 $ ad. U. S. National Museimi, Biological Survey Collection. July 15, 1899. A. K. Fisher. Orig. no. 3189. Characters. — Similar to M. abbreviatus but yellow suffusion even more intense, particularly on belly; nasals and rostrum (above) de- cidedly longer; zygomata more strongly bowed outward; bulla; broader, less pointed anteriorly, and less flattened on outer side. ' Since the original description of Microtus abbreviatus was based on a poor alcoholic specimen and does not fairly present the characters of the species, I have drawn up the accompanying description from a series of adults collected by me at the type locality : MICROTUS ABBREVIATUS Miller. Hall Island Vole. Type from Hall Island, Bering Sea. Characters. — Size large; ears rather small ; tail exceedingly short, thick and densely covered with long hairs ; fur long, full and soft ; appearance lemming- like ; color deep yellowish gray unlike any other Microtus known to me, except the closely allied form from the adjacent St. Matthew Id. Color. — Upperparts uniform deep yellowish gray, intimately but inconspicu- ously mixed with black hairs, becoming light ochraceous buff or bright buff on the sides, and darkening to grayish fulvous on head; underparts whitish, strongly washed with yellowish buff; fore and hind feet whitish, washed with buffy ; tail bicolor, dusky above, overlaid with long buffy hairs, buffy below. Yoiittff ; upperparts yellov/ish or buffy brown ; underparts plumbeous washed with buffy ochraceous. Cranial characters. — Skull large (largest of the Bering Sea species except in- nuitus from St. Lawrence Id.) ; zygomata moderately bowed outward, their outer sides parallel ; jugal only slightly if at all expanded ; frontal ' pinched in ' between orbits, its muscular impressions uniting in a low ridge along median line; braincase large and broad; interparietal in adults pentagonal, the anterior border straight with a pointed projection on median line; the hinder border produced posteriorly so as to be broadly triangular, forming two sides ; bullae of moderate size, pointed anteriorly and flattened on outer side; incisive foramina shortlv open anteriorly, broadly constricted posteriorly; antorbital foramen without protruding anterior lip ; incisors not protruding as in innuitas : molars of medium or rather large size; ;«^ with only 2 completely closed tri- angles ; m^ with 5 closed triangles. Measurements. — Adult male: total length 170; tail vertebra; 29 ; hind foot 24. Average of 3 females: total length 156; tail vertebne 25 ; hind foot 23.2. Skull of male adult 1 979S1 ) : total length, 30.3 ; zygomatic breadth 18 ; mastoid breadth 14; nasals 8.8; molar series 7. 24 MERRIAM ATcasurcrueftts. — Type, J' ad. : total length 17S ; tail vertebrae 32 ; hind foot 24. Average of 5 females from type locality : total length 166; tail vertebrae 26.5 ; hind foot 23.5. Skull of type: basal length 31.5; zygomatic breadth 19; mastoid breadth 14; nasals 9.S; molar series y.3. MICROTUS MACFARLANI sp. nov. MacFarlane Vole. Type from Fort Anderson, north of Great Bear Lake, Arctic America. No. -jVjtV ^^- U- S- National Museum. R. MacFarlane. Orig. no. 3179. Characters. — Size medium or rather large; tail short. Similar externally to operariushwt (apparently) larger; tail decidedly shorter; skull larger; molars small. Color. — Upperparts uniform dull fulvous brown mixed with black hairs ; underparts whitish, washed with buffy ; hind foot brown with more or less whitish on toes; tail sharply bicolor, dusky above, soiled whitish below. Cranial c/iaracfcrs. — Skull similar to that of opcraritis but brain- case broader; zygomata with outer sides longer and parallel (not broadest in middle); bullae much more fully and roundly inflated ; nasals shorter ; mandible decidedly larger; incisors decidedly thicker (anteroposterior ly) ; molar series slightly if at all longer. Compared with yakutatoK^is the skull is shorter, flatter, and somewhat smaller ; with molars decidedly smaller. ]\Ieasi(rc>nc)its. — No measurements of fresh specimens are available. In the dry skin of the type the hind foot measures 1S.5 and the tail approximately 32. EVOTOMYS ORCA sp. nov. Orca Evotomys. Type from Orca, Prince William vSound, Alaska. No. 9S028 9 ad. U. S. National Museum, Biological Survey Collection. June 28, 1899. A. K. Fisher. Orig. no. 2139. Characters. — Size medium; coloration dark ; feet and tail dufeky. Color. — Dorsal area dark chestnut or hazel ; sides yellowish or buffy drab, intimately mixed with black hairs and darkest on rump ; face very dark, grizzled with buffy-gray and black; underparts deep buffy or buffy ochraceous, the plumbeous underfur showing through ; hind feet dusky; tail above dusky from base to tip, below buffy. Cranial characters. — Skull similar to that of dawsoni (with large subquadrate braincase and strongly developed postorbital crest of squamosal) but with decidedly smaller bulla; and decidedly larger in- cisors, particularly the upper ones. The underjaw is thicker and the angular process less flaring. THE HARRIMAN ALASKA EXPEDITION 2$ Measuremefi^s.— Type: total length 138; tail vertebrai 30; hind loot 20. Average of 6 specimens from type locality : total length 140; tail vertebra; 32; hind foot 20. JiemarJts.— The coloration varies somewhat, and the differences are probably seasonal. The type, which appears to be in fresh sum- mer pelage, is in the darkest stage. Other specimens, apparently in left-over winter pelage, or in the molt, are paler, but still have the characteristic dark feet and tail. In 6 out of the 7 specimens secured by us in Prince William Sound, the underparts are strongly suffused with buffy ochraceous; in one only they are whitish (no. 98035). DICROSTONYX UNALASCENSIS sp. nov. Unalaska Lemming. Type from Unalaska, Alaska. No. 99622 ad. U. S. National Museum, Biological Survey Collection. July 8, 1899. C. Hart Merriam. Found in owl pellets. C/iaracters.—S\ze rather large, equalling or exceeding B. Jmd- sotihts from Labrador; external characters unknown. Skull large and broad with subquadrate braincase, broadly spreading rounded zygomata, broadly sulcate frontal, and prominent peg-like postorbital processes of squamosal. Compared with D. hudsonius from Chimo, Ungava, the anterior angle of the zygoma is less expanded, jugal longer, angle of mandible broader and more strongly everted, bulla; much larger; upper in- cisors much broader; ;«i with 3 (instead of 2) closed triangles on inner side in addition to posterior (or heel) loop; m'^ with 2d loop on inner side normal and separated by full interspace from anterior crescent of w^, and with posterior (heel) loop on outer side com- plete. In Jmdsonms the 2d loop on inner side is narrowed or flat- tened and in contact with m^, and the posterior loop on outer side is absent or reduced to a small remnant. DICROSTONYX NELSONI sp. nov. White Tundra Lemming. Type from St. Michaels, Alaska. No. f|ff 9 yg. ad. Merriam Collection. Characters.— '^'xT.e rather small; a dark dorsal stripe present in summer pelage; winter pelage white. Color.— Type specimen in change from winter to summer pelage (date unknown) : Upperparts from halfway between nose and eyes to rump finely mixed chestnut and yellowish white (hairs with chest- nut tips and a broad subapical zone of yellowish or buffy) becoming 26 MERRIAM darker posteriorly ; rest of animal (including nose, cheek, feet and tail) yellowish white with indistinct fulvous markings as follows: a patch on each side posteriorly, one on pectoral region, reaching up on sides of neck to ears, and a ring around (but not reaching) base of tail ; a dark brownish median stripe reaching from middle of back to ring around tail. Cranial characters. — Skull similar to that of unalasce7isis in gen- eral form, form of zygoma and jugal, and in enamel pattern, but smaller, and differing from all known species in the bullae, which are very small and fiarrozv ; incisive foramina short; molars small and rather narrow, but not quite so narrow as in richardso7ii. j\Ieasure?7ients. — Type specimen (measured from alcohol by C. H. M.) : total length ii8; tail vertebrcE 13; hind foot 18. DICROSTONYX RICHARDSONI sp. nov. Churchill Lemming. Type from Fort Churchill, Hudson Bay. No. f yf f $ ad. Merriam Collection. July 1859. W. MacTavish. Characters. — Size large ; fur long and silky ; 3d nail of forefoot much larger than 4th ; b2illce broadly rounded and somewhat de- pressed. Colon. — Type specimen in winter pelage: silky white all over. Cranial characters. — Skull large ; zygomata very broad and squarely spreading; rostrum and nasals long, frontals rather narrow inter- orbitally; bullce depressed^ broadly and roundly inflated. Molar series narrow; enamel pattern as in nelsoni and unalascensis (not as in hudsonius'). This species resembles nelsoii but differs from all others in the small size and narrowness of the molar teeth, and is distinguishable at a glance from all known species by the broadly rounded and rather depressed bulla;. Measurements. — Type specimen (measured from alcohol by C. H. M.) : total length 143; tail vertebrie 14; hind foot 20. LEMMUS ALASCENSIS sp. nov. Point Barrow Lemming. Type from Point Barrow, Alaska. No. ||^^ 9 ad. Merriam Col- lection. Characters. — Size medium (slightly smaller than either lonmus or nigripes) ; ears smaller than in nigripes; feet and nose pale. Color. — Type specimen (in alcohol) from Point Barrow, and a skin (alcoholic) from St. Michaels: entire animal golden fulvous, darkest on head, where it is mixed with black hairs; brightest on sides, * where it is almost orange fulvous; palest on feet and around mouth. THE HARRIMAN ALASKA EXPEDITION 27 Cranial characters.— '&V\x\\ similar to that of L. Icmmus but slightly smaller; braincase smaller and less flattened; zygomata less widely spreading; angle of mandible much less everted; upper in- cisors heavier (broader). Compared with nigrlpes from the Pribilof Islands the differences are more marked : skull decidedly smaller and less massive ; nasals much shorter and smaller ; rostrum more slender ; frontals not elevated into tubercles anteriorly; bull^ larger (more broadly indated) ; angle of jaw much smaller and less everted. Molar series shorter. Measureme7its.—Tyy^^ specimen (measured from alcohol by C. H. M.) : total length 115 ; tail vertebra; 15 ; hind foot 18. LEMMUS YUKONENSIS sp. nov. Yukon Lemming. Typeh-om Charlie Creek, Yukon River, Alaska. No. 98849 ? ad. U. S. National Museum, Biological Survey Collection. Aug. 9, 1899. W. H. Osgood. Orig. no. 769. ^ Characters.— S\zc small, ears relatively large; general color dark anteriorly, with bright fulvous or rufous rump and flanks; audital bullce immense. Co/or.— Head, shoulders, and anterior half of back grayish brown, grizzled with black and yellowish fulvous, darkest on top of nose ; rump and hinder part of back rich rusty fulvous or orange fulvous; lower sides and belly golden fulvous ; deepest and brightest on flanks ; sides of nose grayish or grayish brown ; a band of golden fulvous on lower part of face from nose to below ear ; fore and hind feet dusky ; tail bicolor, dusky above, buffy below. Cranial characters.— '^\^yx\\ very different from any species known to me: braincase (seen from above) large, broad, ^nA produced pos- teriorly to cover the enormously enlarged bulla;; frontal narrow inter- orbitally, with a narrow median sulcus, but not ' pinched in' as in L, alascensis; rostrum small; zygomata squarely set but not widely spreading, the sides parallel; interparietal large, usually pentagonal; incisive foramina rather short ; bullce remarkably large and strongly inflated. Measureme7tts.—{Yxom well made dry skin:) total length 130; tail vertebrae 17; hind foot 19.5. ERETHIZON EPIXANTHUS MYOPS subsp. nov. Alaska Porcupine. Tvte from Portage Bay, Alaska Peninsula. No. 59140 9 old, U. S. * National Museum. September 1893. Chas. 11. Townsend. 28 MERRIAM Characters. — Similar to E. epixanthus but tail and hind feet shorter; body, flanks and sides of tail even yellower; face whiter from more abundant admixture of yellowish white hairs on sides of face and between eyes, so that the fore part of head appears almost grayish instead of sooty. Cranial characters. — Compared with typical epixanthus from California the skull as a whole is much shorter, broader, and more rounded (less angular) ; occipital and sagittal crests absent or much less developed ; zygomata bowed strongly outward [instead of sub- triangular], the outer side convex outward [not flat] ; braincase short, broad between posterior roots of zygomata, and tapering posteriorly so that the breadth between squamosals anteriorly is much greater than breadth of occiput [instead of subequal] ; outer wall of antorbital vacuity (seen from the side) nearly straight and vertical [instead of strongly concave]. Retnarks. — An immature female collected at Stepovak Bay, Alaska Peninsula, July 9, 1899, by Charles Palache is the yellowest porcupine I ever saw. It measured 630 mm in total length. LEPUS OTHUS sp. nov. Tundra Polar Hare. Type from St. Michaels, Alaska. Skull no. 158S3 U. S. National Museum. L. M. Turner. Characters. — Size very large (hind foot 174) ; winter pelage white all over except extreme tips of ears, which are black. Cranial characters. — The skull of Lepiis othus is so large, broad, and massive that it does not require comparison with the glacialis series from eastern Arctic America, but only with tschiiktshorjivi from the region about Plover Bay, Siberia. From tschuktshoriim it differs in the following particulars : f rontals slightly broader interoi*- bitally; nasals narrower and less flattened; supraoccipital shield nar- rower, much longer than broad [in tschiiktshorum decidedly broader than long] ; jugal decidedly narrower (less expanded vertically), longer posteriorly (protruding farther behind squamosal root of zygoma), and with muscular impression on outer face decidedly nar- nower anteriorly; basioccipital narrower; incisive vacuit}' smaller. Cranial measurements. — Type specimen: basal length 86.!^ ; zy- gomatic breadth antciiiorly 49; at widest point 52; postpalatal length 46.5 ; greatest breadth across postorbital processes 39; greatest breadth of nasals posteriorly 22; anteriorly 20. THE HARRIMAN ALASKA EXPEDITION 29 LEPUS POADROMUS sp. nov. Peninsula Arctic Hare. Type from Stepovak Bay, Alaska Peninsula. No. 98068 U. S. Na- tional Museum, Biological Survey Collection. July 8, 1899. C. Hart Merriam. (Shot by Charles Palache.) Orig. no. 2207 (Fisher catalogue). Characters. — Size smaller than Lepus othus from St. Michaels, but larger than the members of the glacialis series from eastern Arc- tic America; ears about as in L. ot/ius^i but hind feet and tail shorter; color in summer Y>^\i\^e orlzzled i^'ray\w\t\\ a fulvous tinge on head. Color. — Nose and head dull fulvous with a buffy ring around eye ; cheeks dull fulvous, grizzled with longer hairs tipped with black and buffy ; ears grizzled fulvous and black on anterior face, becoming white posteriorly, and narrowly edged with white except on basal third of anterior margin, which is grizzled fulvous and black like rest of an- terior face ; back, sides, and outer surface of hind legs gi"ay, grizzled with buffy and black ; forelegs grizzled grayish fulvous, becoming pale dull fulvous on dorsal surface of feet and mixed with white hairs about toes ; hind feet mainly white; tail hoary or grayish white, be- coming dusky along median line above ; chin and lips whitish ; throat with a broad grayish collar ; pectoral region, strip down belly, and inner sides of legs, whitish. Cranial characters. — Compared with its nearest known relative, L. othus from St. Michaels, the rostrum, nasals, braincase, palatal bridge, and incisive vacuity are narrower, the frontal breadth remain- ing essentially the same; nasals in particular much 7iarrower and more arched transversely; pterygoids broader; hamular processes more pro- duced and hooked ; angular process of underjaw with everted edge much broader posteriorly. Remarks. — Lepus poadrotnus, compared with a skin of tschuk- tshorum from Arikamchichi Island (No. 3384 U. S. Nat. Mus.) has the fulvous of the nose and head very much darker. Unfortunately no skins of L. othus in summer pelage are available for comparison. Measurc?7ie}its. — (From dry skin of type:) Hind foot 152; ear from notch 80. Cranial mcasurc7nents. — Basal length 82 ; zygomatic breadth an- teriorly 47.5; postpalatal length 43; greatest breadth across post- orbital processes 37; greatest breadth of nasals posteriorly 20; an- teriorly 16, LEPUS AMERICANUS DALLI subsp. nov. Dall's Varying Hare. Type (skull) from Nulato River, Alaska, no. 7579. U. S. National Museum. W. H. Dall. 30 MERRIAM Characters. — External characters unknown. Cranial characters. — Skull large; nasals long and slender; pos- terior part of mandible large. Compared with typical americanus from Ungava, Hudson Strait, the skull of dalli is very much larger (basal length 65 mm. or more instead of 61 or less) : rostrum, nasals, and incisive vacuity decidedly longer; jugal longer and somewhat broader (vertically) ; mandible very much larger and more expanded posteriorly ; molariform teeth larger. Remarks. — Superficially the skull of dalli resembles that of sub- species virginianus from Ossipee, New Hampshire, much more closely than typical americanus ; the agreement in size is better and both have rather long, slender nasals. But dalli has a much broader braincase, longer and heavier zygomata, broader jugals, and differs in other particulars. Compared with its geographical neighbor, L. macfarlani from Fort Anderson, it may be distinguished at a gltmce by its slender nasals and rostrum, longer zygomata, and larger mandible. Cranial fneasurements. — Type skull : Basal length 65 ; zygo- matic breadth anteriorly 37.5 ; at middle 38 ; postpalatal length 33.5 ; frontal breadth across middle of postorbital processes 21 ; breadth of nasals posteriorly 15; anteriorly 11. LEPUS AMERICANUS MACFARLANI subsp. nov. MacFarlane Varying Hare. Type (skull) from Fort Anderson, north of Great Bear Lake, Arctic America. No. 14467 U. S. National Museum. R. MacFarlane. Characters. — External characters unknown. Cranial characters. — Skull similar to that of typical americanus from Hudson Strait but ver}' much larger and heavier; nasals very broad (anteriorly as well as posteriorly); braincase less flattened; supraoccipital shield narrower and tapering posteriorly; mandible, molariform teeth, and lower incisors decidedly larger. Compared with dalli from Nulato River the nasals are very much broader and blunter; rostrum broader and more massive ; braincase higher (almost ridged along median line) and much less ikittcned ; supraoccipital shield narrower posteriorly (more tapering) ; mantlible somewhat smaller. Cranial measurements. — Basal length 67; zygomatic breadth an- teriorly 38.5; at middle 41 ; postpalatal length 34.5; frontal breadth across middle of postorbital processes 21; breadth ot" nasals pos- teriorly i7-5) anteriorly 14. PROCEEDINQS WASHINGTON ACADEMY OF SCIENCES Vol. II, pp. 31-38. M^^^" 30, 1900- THE BATS OF THE GENUS MONOPHYLLUS. By Gerrit S. Miller, Jr. Hitherto Mono-phyllus has been regarded as a monotypic genus, with M. redmant of Jamaica, and perhaps Cuba also, as its only representative. Material in the United States National Museum, though throwing no light on the question of the identity of the Cuban bat with that from Jamaica, shows that the genus contains at least three additional species, one from Porto Rico, one from Barbados and a third from an unknown locality. The four may be distinguished from each other as follows : Synopsis of the Species of Monophyllus. Size small, forearm about 36, greatest length of skull less than 20; second upper premolar with well developed postero-internal lobe. M. portoricensis. Size large, forearm about 40, greatest length of skull more than 20 ; sec- ond upper premolar with postero-internal lobe rudimentary or absent. Second lower premolar shorter than first and in contact with third ; no space between upper premolars M. flethodon. Second lower premolar longer than first and not in contact with third ; a space between upper premolars. Ventral surface of basioccipital (when viewed from under side) dipping forward at conspicuous angle (about 32°) with surface of presphenoid; bony palate narrow and conspicuously arched. M. clinedaphus. Ventral surface of basioccipital (when viewed from under side) dipping forward at inconspicuous angle (about 16°) with sur- face of presphenoid ; bony palate broad and slightly arched. M. redmani. Proc. Wash. Acad. Sci., April 1900. 3^ 32 MILLER MONOPHYLLUS REDMANI Leach. 1822. Monophylhis redmani Leach, Trans. Linn. Soc. London, XIII, p. 76. 1878. Mmophylhis redmani Dobson, Catal. Chiropt. Brit. Mus. p. 504. 1898. Monophylhis redmani H. Allen, Trans. Am. Philos, Soc. N. S., XIX, p. 248. Type locality. — Jamaica. Geographic distribution. — Jamaica. It is doubtful whether the specimens from Cuba recorded by Dobson are actually this species. General characters. — See synopsis. Ears. — The ears are short : laid forward they reach about 3 mm beyond eye. Anterior border strongly convex at base, then nearly straight but strongly oblique to narrowly rounded tip. Posterior bor- der slightly concave below tip, then straight to a little below level of anterior base ; below this abruptly convex. There is no antitragus. The posterior border terminates directly beneath anterior base, and about 7 mm. behind angle of mouth. A few rudimentary cross stri- ations are faintly indicated on inner side of conch near posterior bor- der. Two well developed ridges run outward along the meatus, one directly over the other. Anterior border of tragus straight, very slightly thickened above. Tip narrowly rounded. Posterior border straight from tip to two tooth-like projections, the lowermost of which is opposite anterior base, then distinctly convex. Muzzle and chin. — Horizontal portion of noseleaf oval, broader than high, indistinct along upper lip, distinct at side. Height of up- right portion about equal to width of horizontal portion, narrowing gradually to rounded tip the width of which is a little less than i mm. For a short distance below tip the sides are parallel. Whole leaf finely pubescent. Chin divided by a deep groove. Mefnbrafies. — The membranes are ample, moderately coarse in texture. Width of wing membrane at elbow about 30 mm. Width of propatagium at elbow, 15. Width of uropatagium at middle of femur, 8. Propatagium continued along forearm to include meta- carpal of thumb. The membranes ai-e practically naked throughout, though the fur extends in a narrow line on wings close to bod}^ and also over proximal third of forearm and on most of propatagium. On the latter it is very sparse and irregular. Eeet. — The feet are large, slightly more than half as long as tibia. Toes a little shorter than metacarpals, free from membrane. *They are armed with very strong claws. All five toes are essentially equal in length. Calcar distinct, but reduced to a mere rudiment, about 4 mm. long. THE BATS OF THE GENUS MONOPIIYI.LUS 33 Tail. — The tail is about as long as the foot, and a little less than half free from membrane. JFur and color. — The fur is loose in texture and very short (scarcely 5 mm. on middle of back). It is closely confined to body, scarcely reaching membranes except as stated above. Color of a specimen after five years' immersion in alcohol, rather pale Prouts brown above, Isabella color below, the hairs tipped with pale gray. These tips are inconspicuous above but very noticeable below. Skull. — The skull of Monophyllus redviani is slender but at the same time rather robustly built, in form almost precisely like that of the somewhat larger Glossonycteris geoffroyi. Rostrum nearly as long as braincase, its sides essentially parallel. It is narrowest just behind canines. A distinct swelling in lachrymal region over each orbit. Antorbital foramen directly above space between premolars, its upper and posterior margins swollen. Zygomata slender, distinctly expanded vertically near anterior and posterior base, the anterior ex- pansion upward, the posterior downward. Braincase slender, con- siderably longer than broad, the forehead rising above rostrum at angle of about 20°. Highest point of braincase over middle of mesen- cephalon. Audital bullae small, the area of each scarcely exceeding that of glenoid surface. That part of base of braincase formed by basioccipital and basisphenoid forms a plane dipping uninterruptedly forward (when skull is viewed with its dorsum down) from anterior lip of foramen magnum to well developed basisphenoid pits at an inconspicuous angle, about 16°, with surface of presphenoid. The entire ventral surface of the presphenoid lies in one plane, broken only by a narrow but well defined median longitudinal ridge which termi- nates posteriorly in a pointed projection lying ventral to the ridge separating basisphenoid pits. Measurements of an adult skull ( ^ ) from Kingston, Jamaica : Greatest length 22.4; basal length 21.4; basilar length 19.4; zygo- matic breadth 10; least interorbital breadth 4 ; mastoid breadth 9.8; greatest breadth of braincase 9 ; frontopalatal depth (at middle of molar series) 2.4; depth of braincase 8; maxillary toothrow (exclu- sive of incisors) 8.6; mandible 15.4; mandibular toothrow (exclu- sive of incisors) 9. Teeth. — The teeth are small relatively to the size of the skull, though not different in this respect from those of members of related genera. Both upper and lower toothrows are broken by a conspicu- ous gap in front of posterior premolar. Second upper premolar broader posteriorly than anteriorly, but without postero-internal lobe. 34 MILLER First lower premolar shorter than second or third, but relatively less narrow. Its posterior border is scarcely imbricated over anterior ex- tremity of second premolar. Measiire7}ients. — For measurements of Mojiophyllus red?natii see table, page 38. Specimens examined. — One, from Kingston, Jamaica. Remarks. — Monophyllus redtnani is readily distinguishable from the other members of the genus by its large size, heavy rostrum, and nearly flat base of braincase. The peculiarities of the tragus in the only specimen that I have examined may be due to the effect of strong alcohol. MONOPHYLLUS FORTORICENSIS sp. nov. Type. — Adult male (in alcohol), no. 86261 United States National Museum, collected in a cave near Bayamon, Porto Rico, January 18, 1899, by Paul Beckwith. General characters. — See synopsis. Ears. — The ears are essentially like those of J/, redmani. The tragus, however, is very conspicuously thickened along its anterior border. Inner surface of conch with scarcely a trace of cross striation. Muzzle a?id chin. — Noseleaf smaller than in JM. redtnani and less tapering above, the sides not parallel below tip. Groove in chin ap- parently deeper and narrower than in AI. redtJiani. jMembranes., feet and tail as in M. redmani. Fur and color. — Fur slightly longer than in M. redmani (about 6 mm. long on middle of back) but not different in quality or distribution. Color of a specimen (topotype) skinned after two months' immer- sion in formalin, seal brown above, broccoli brown below, the ex- treme tips of the hairs grayish. The light tips are conspicuous on the ventral surface, but scarcely noticeable above. Skull. — The skull of Monophyllus portoricensis is considerably smaller than that of AI. redmani., and as the reduction in size affects the rostrum relatively more than the braincase the general appearance of the skull is considerably altered. In fact at first glance the skull strongly suggests that of Glossophaga sorici/ia., though it is slightly smaller. The rostrum is very noticeably shorter than the braincase. Its sides are nearly parallel and scarcely divergent at base of canines. The braincase rises above rostrum at a slightly greater angle (about 25°) than in Af. redmani. Base of braincase exactly as in M. red- ma?ii^ except tliat median ridge of presphcnoitl is somewhat flattened posteriorly and perfectly continuous with ridge separating basisphe- noid pits. THE BATS OF THE GENUS RIONOPHVLLUS 35 ^leasurements of skull of tA'pe : Greatest length 19.6; basal length iS ; basilar length 16 ; zygomatic breadth S.S ; least interorbital breadth 4; mastoid breadth 8.8; greatest breadth of braincase 8.4; fronto- palatal depth at middle of molar series 2 ; depth of brain case 7.6; maxillary toothrow (exclusive of incisors) 7; mandible 12.8; man- dibular toothrow (exclusive of incisors) *]. 1 ccth. — Teeth as in J/. red??iani except that the gap in toothrow in front of posterior premolar is greatly reduced, and in the lower jaw nearly obliterated. Second upper premolar with well developed postero-internal lobe. First lower premolar slightly shorter than second or third, and conspicuously less narrow. Its main axis is set obliquely to that of the other premolars and its posterior border is noticeably imbricated over outer edge of second premolar. Measuretnents. — For measurements of Alonophyllus portoricensis see table, page 38. Specimens examined. — Six, all from the type locality. Rejnarks. — Monophylhis portoricensis is the smallest of the four species thus far known. It is further distinguished by the form of the posterior premolar. MONOPHYLLUS PLETHODON sp. nov. Type. — Adult male (in alcohol) no. 101530, United States National Museum, collected in St. Michaels Parish, Barbados, August i, 1899, by P. McDonough. Gejteral characters. — See synopsis. Ears. — The ears are essentially as in M. redmani though some- what broader relatively to their length. Inner surface without cross ridges. Tragus as in M. portoricensis. Muzzle and chin. — Noseleaf like that of j\f. portoricensis, but with upright portion even broader and more bluntly pointed. Groove in chin broader and shallower than in the Porto Rican species. Metnbranes., feet and tail as in ]\I. redma7ii. Fur and color. — Fur similar to that of AI. redmani in both quality and distribution. Color of type after five months' immersion in alcohol, broccoli brown slightly paler below. Skull. — The skull of the type specimen has been injured, though not sufficiently to obscure its essential characters. In size it closely agrees with that of M. clinedaphus., but in form it suggests that of M. portoricensis. Its peculiarities are in some respects carried even further than in the latter, as the rostrum though deeper is relatively narrower, the palate is more arched, and the canines actuall}' closer 36 MILLER together tlian in the smaller animal. Antorbital foramen over anterior extremity of second premolar and region of contact between it and first. Pterygoids imperfect, but evidently further apart than in M. redmani. Base of braincase as in M. redmant send j\I. portoricensis. Presphenoid injured, but apparently as in these species and not like that of JM. clinedaphus. Zygomata very lightly built, even more slender than in M. portoricensis. Mandible about as long as that of M. clinedaphus^ but more lightly built than in any of the know^n members of the genus ; while distinctly longer than that of J/, por- toricensis it is scarcely if any deeper. Measurements of skull of type : Greatest length 21.6; basal length 19.4; basilar length 17. S; zygomatic breadth 9.8; least interorbital breadth 4.6; mastoid breadth 9.6; greatest breadth of braincase 10; frontopalatal depth (at middle of molar series) 2.4; depth of braincase 6; maxillary toothrow (exclusive of incisors) 7.2; man- dible 14; postdental portion of mandible 5.8; mandibular toothrow (exclusive of incisors) ^.%. Teeth. — The teeth oi Monophyllus plethodon differ conspicuously from those of the other members of the genus in that they are so crowded that all trace of the characteristic space in front of posterior premolar of both jaws is obliterated. This tendency to crowding is also evident in the upper incisors, the innermost pair of which are almost in contact, while in the three other species they are separated by a noticeable space. Canines slender, relatively closer together than in other members of the genus, their shafts parallel or a little con- verge7it. Maxillary teeth otherwise essentially as in J/, redmani. Mandibular incisors very minute, but normal in form and position. First lower premolar distinctly longer than second, over which its posterior edge is imbricated, much as in Af. portorice?isis. The re- maining mandibular teeth do not differ conspicuously from those of the other ntembers of the genus. Measurements. — For measurements of Monophyllus plethodon see table, page 38. Speci?>iens examined. — One, the type. Remarks. — Monophyllus plethodon is immediately recognizable by its crowded teeth and slender rostrum. MONOPHYLLUS CLINEDAPHUS sp. nov. Type. — Adult male (in alcohol) no. •jV-fiTS' United States National Museum. No history. General characters. — See synopsis. THE BATS OF THE GENUS MONOPHYLI.US 37 Ear. — The ears exactly resemble those of M. redmani except that they are slightly broader and the inner surface of the conch is marked by seven distinct parallel cross ridges. j\Iuzzle and cJiin. — Noseleaf smaller than in M. redmani, the up- right portion less tapering at tip, its sides nowhere parallel. Groove in chin apparently narrower and deeper than in M. redmani. Membrane., feet and tail as in J/, redmani. JFur and color. — Fur similar to that of M. redmani in distribution and quality. Color after nearly forty years' immersion in alcohol, mars brown above, wood brown below. Skull. — The skull of Monophyllus clinedaphus is slightly smaller than that of M. red?nani.^ though in general appearance it is not no- ticeably different. In M. clinedaphus the rostrum is slightly narrower anteriorly, and the antorbital foramen is placed further forward, its anterior half lying over the first premolar. The striking differences between the two species are seen in the base of the braincase when the skull is held back down. In AI. clinedaphus the plane of the basioc- cipital and basisphenoid, instead of sloping forward at a faint angle as in the other species of the genus, pitches abruptly forward and down- ward at an angle of 32°. The resulting secondary changes in form of the adjacent parts are noticeable on comparison, though not easy to describe. Presphenoid with ventral surface in two planes, the change from the anterior more ventral level to the posterior more dorsal level taking place abruptly at a point about half way between tip of hamu- lars and line joining posterior molars. The median ridge extends as far back as the anterior (ventral) level only, leaving the posterior level essentially flat. Measurements of skull of type: Greatest length 31.8; basal length 19.4; basilar length 17.4; zygomatic breadth 9; least interorbital breadth 4 ; mastoid breadth 9 ; greatest breadth of braincase 9 ; fronto- palatal depth (at middle of molar series) 2.4; depth of l^raincase 8; maxillary toothrow (exclusive of incisors) 8; mandible 14; mandi- bular toothrow (exclusive of incisors) 8.4. Teeth. — The teeth are slightly smaller than in AI. redmani., but of the same form. AIeasure7nents . — For measurements of Afonophyllus clinedaphus see table, page 38. Specimens exa?ftined. — One, the type. Remarks. — This species is strongly characterized by the form of the presphenoid as well as by the more conspicuous peculiarities of the base of the braincase. It requires no special comparison with the other members of the jrenus. 38 MILLER a fa o en H Z m IS H 0 IH t-l •nAVOJD TOOJJ VO 00 M 0\0 (JnOnG Oo6 l-l M . ■;iSip pjtqx 00 t-»rON WVO Ttt^ 00 t~»I^t^t--t^t^t^ •liSip pUODSS 00 fO « ONOO w C3N cs rOtOfOPI N cOCS to •;iSip ;saia M CT^a^O^Oo6o6 P» M MM •nUB3JO^ O OVD VO r^ l^vD 00 ■^cOrOfOfOcOcOtO •JBDlBo) ^ t j fO •* Ti- •>d- ro Tj- CO w lOO^ M ON ON On 0\o6 on w •Biqix oovbvovovovo »or^ •i^x 't O 00 ONOO t^ ON ON o< M •q^SnaiiBjox CO lOO lOl^lONOO VDVOVONOVOVOVDVO •xas u V a 1 OOvOOnOOO'iO CO " 'O iO\D lOvD fO Tj- lONO VO VO \0 VO I-' VO CO CO 00 CO 00 o M 1 8 o vT- - - - '5 i 1 « V 1 .. The type was collected at L,eh. The National Museum contains a specimen taken by Dr. Abbott in a pine forest among the mountains of central Kashmir, at an elevation of about 10,000 feet. * In C. myoides the tubercles are apparently like those of C. russula, but the soles of the Kashmir specimen are injured. 39 40 MILLER the outer side are larger, and placed almost directly opposite those of inner side. Ears. — The ear is slightly smaller than that of C. myoides ; there- fore much smaller than in C. rjisstda. It is not peculiar in form. Fur. — The fur is dense and silky, that on middle of back about 7 mm. in length. The hairs throughout have a distinct gloss, produc- ing a strong contrast with the dull fur of Crocidiira inyoides. Color. — Entire dorsal surface of head and body wood brown, almost exactly matching the plate in Ridgway's Nomenclature of Colors. Tail wood brown, slightly paler below than above, its dorsal surface like back. Sides, legs, feet and entire ventral surface of head and body grayish-white. Fur of body slate-black through basal two-thirds or three-fourths, the dark portion crossed by several indistinct narrow pale bands which produce an effect like watered silk when the hairs are smoothly parted. Meas2irej)2e?its. — Total length 82 ; head and body 54 ; ^ tail verte- bras 28.5;^ hind foot 12 (11);^ ear from meatus 5.6; ear from crown 3 ; width of ear 6.S. Remarks.— In its peculiar color this species differs widely from described members of the genus and shows a striking parallelism with the small Mus pachycercus inhabiting the same region. What its cranial and dental characters will prove to be can only be conjectured, but there is every reason to suppose that they will essentially agree with those of Crocidura myoides and C. russula. ' From fresh specimen by collector. * Measurement of hind foot in parenthesis taken without claws. PROCEEDINQS OF THE WASHINGTON ACADEMY OF SCIENCES Vol. II, pp. 41-68. July 25, 1900. A NEW STONY METEORITE FROM ALLEGAN, MICHIGAN, AND A NEW IRON METEORITE FROM MART, TEXAS.^ By George P. Merrill and H. N. Stokes. CONTENTS : The Allegan Meteorite 41 History and Petrography 41 Chemistry , . 47 The Mart Iron 51 History and General Features ,....,.... 51 Chemistry 52 Supplementary Note by G. P. Merrill 53 THE ALLEGAN METEORITE. HISTORY AND PETROGRAPHY. A LITTLE after eight o'clock on the morning of July 10, 1899, there fell on what is locally known as Thomas Hill, on the Saugatuck Road, in Allegan, Michigan, a stony meteorite, the total weight of which cannot have been far from seventy pounds, although, unfortunately, it was badly shattered in strik- ing the ground, and its exact weight can never be known. •These meteorites have been the subject* of a preliminary notice by Dr. Merrill in Science for November 24, 1899, and the Allegan stone by Mr. H. L. Ward in the American Journal of Science for December, 1S99. The general and petrographic description are by G. P. Merrill, and the chemical examination is by Dr. H. N. Stokes. The chemical analyses v?ere made in the laboratory of the U. S. Geological Survey and are published here by permission of the Director. Proc. Wash. Acad. Sci., July, 1900. (41) 42 MERRILL AND STOKES The main mass of the stone (see PI. I, figs, i, 2 and 3) weigh- ing 62 y2 pounds, came into the possession of the National Museum, with an additional fragment weighing about lyi pounds. This, with a 4-pound fragment, sold to other parties, and many small pieces stated as var3ang from the size of a pea to that of a hickory nut, carried away by school children and others, would readily bring the total weight to the figure men- tioned. According to Mr. Walter Price, as quoted by H. L. Ward ^ the stone came from the northwest and passed within about forty feet of where he was working, striking the ground about ten rods beyond, in sand, and burying itself to the depth of about a foot and a half. The attention of the observers, it is stated, was first attracted b}^ a cannon-like report, followed by a rum- bling sound lasting about five minutes (?), which was followed, as the stone came nearer, by a hissing sound, compared to that of an engine blowing off steam. ^ When first seen in the air the stone had the appearance of a black ball about the size of a man's fist. As it passed the observer, it is stated, "there seemed to be a blue streak behind it, about six feet long, which tapered back to a sharp point." The stone was dug up about five minutes after striking and is stated to have been too hot for handling, necessitating removal with a shovel. "The sand was hot for about two feet round where it struck." Messrs. H. Ster» iS: Compan}-, of Allegan, from whom the National Museum obtained the main mass of the stone, furnished cor- roborative evidence. They state that the sand about the hole made by the meteor was quite warm an hour after the fall, and that the stone itself was still warm when placed in their shop window, some two and a half hours later. The general appearance of the stone is well shown in Plate I, figs. I, 2 and 3. From a study of the mass it would ap- pear that at the time of entering our atmosphere, and for most of its course, the point A (uppermost in fig. i) was in advance, 'Am. Jour. Sci., December, 1899. 2 It is well to note that there is no evidence to show that the report was ac- companied by a breaking up of the stone. But the one mass was seen to fall, and though this is somewhat angular in outline nothing indicates a fracturing after entering the earth's atmosphere and before striking the ground. ALLEGAN METEORITE AND MART IRON 43 but that the broad side (lower surface in the figure) was first to strike the Efround. These conclusions are based on the facts that the crust at the top A, where friction would have been greatest if this side had been foremost, is scarcely more than a millimeter in thickness, and, as shown by the microscope, is almost wholly glassy, enclosing only residual portions of un- fused silicates ; while on the other side it is from 2 to 3 mm. in thickness, blebby, and, as seen under the microscope, vesicular, and often crowded with minute silicate crystallites imperfectly secreted from the glassy base ^ (PL III, fig. 2). Further, the furrows on the outer surface of the crust, due to atmospheric friction, radiate in all directions from this uppermost point A. These furrows show somewhat indistinctly near B in fig. i. That the stone struck broad side down is shown by the grass stems and earth still adhering to this surface. A point of no inconsiderable interest in this connection lies in the fact that these grass leaves, which were wielded to the surface of the stone through impact, or were even driven into the frac- tures caused by the same, are not charred in the least, nor is there other evidence of heat than that furnished by the black crust already referred to. The grass leaves and earthy matter adhering to the surface of the stone are shown somewhat in- distinctly at C in PI. I, fig. I. To the unaided eye this stone shows on the broken surface a quite even granular structure of gray color and, on closer in- spection, abundant, beautifully spherulitic chondrules, averag- ing not more than one or two millimeters in diameter (PI. II). In two cases chondrules nearly 5 mm. in diameter w^ere ob- served. These are sometimes beautifully spherulitic, or again elongated and irregular in outline, and sometimes have pitted surfaces, such as are seen in compressed pebbles in conglomer- ates (see PI. IV, figs. 1-7). The majority of them are dark gray in color, but some are greenish white. They are com- posed of both olivine and enstatite, as will be noted later. Nu- merous brilliant metallic points of a silver-white color indicate the presence of disseminated iron. Viewed more closely the lA similar thickening of the crust at the rear (riickseite) was noted by Tschermak on the Gopalpur meteorite (Min. Mittheil., 1872, p. 96). 44 MERRILL AND STOKES stone is seen to be made up of the chondrules, iron, and dark gray silicate materials, imbedded in a light gray, ashy ground- mass. The stone is exceedingly friable, crumbling away readily be- tween the thumb and fingers. Indeed it is pronounced by Dr. Merrill to be without exception the most friable meteorite that has come to his attention. Naturally it is beautifully fresh and free from all oxidation products. Examined under the microscope, in thin section, the stone exhibits in a very marked degree the granular fragmental struc- ture which sometimes characterizes chondritic meteorites, as those of Gopalpur, San Miguel, etc., and which is regarded by Tschermak and some other authorities as indicative of a tuffa- ceous origin. Three types of chondrules are noted : first, the ordinary enstatite chondrules showing the eccentric, fan-shaped structure, too well known to need further description (PL V, figs. 3 and 4) ; second, those composed of olivines, sometimes quite idiomorphic, developed in a black glass (PL V, fig. 6) ; and, third, those which are apparently of enstatite but almost completely structureless (PL V, figs. 5, 7 and 9); these last form the greenish chondrules referred to above. There are also occasional olivine chondrules showing the barred or grate structure. As already noted, some of these chondrules are beautifully spherical and others are in the form of elongated blebs (PL IV, figs. I and 2). They are sharply differentiated in most cases from the groundmass and break away so readily as to make the preparation of satisfactory thin sections extremel}' difiicult. When isolated they often show one or more shallow concavities, such as might be formed by the pressure of one against another, but no two were found in such contact as to produce this result (PL IV, fig. I). Similar concavities have been noted by Tschermak and Makowsky in chondrules from the Tie- schitz (Moravia) meteorite.' Many of the enstatite chondrules are distinctly fragmental in outline (PL IV, figs. 3-7), and none of them show a holocrvstalline internal structure. 'Denkschriflen Kais. Akad. dcr Wissenschafteii, Math. Nat. Classc, B. 39, 1S79, P- 195- ALLEGAN METEORITE AND iMART IRON 45 Some of them, as separated mechanically from the groundmass, are simply blebs of enstatite slag with cavities resembling the steam cavities in terrestial rocks or in slags from smelting fur- naces (PI. IV, fig. 5). In a few cases small chondrules show little depressions on one side which are continued as holes into the interior for a third the diameter of the spherule. PI. IV, figs. 1-7, and PI. V, figs. i-6, show the forms and outlines of the chondrules as seen under the microscope in thin section, or isolated. It is to be noted that in the case of a fragment, such as is shown in PI. V, figs. 3, 4 and 5, no other portions of the same chondrule are recognizable in the slide, showing that the shattering took place prior to the consolidation of the stone. This same feature is emphasized by a study of the chondrules as picked out by hand, many of them being but fragments, with the broken surface covered by the grayish dust of the groundmass, showing that the fractures are old and not due to the breaking of the stone since it reached the earth. Some of the larger of these, showing surface fractures antedat- ing consolidation, are shown in PI. IV, figs. 2, 3, 4, 5, 6 and 7. The material obtained by passing the sand from the broken stone through a sieve of about 2 mm. mesh is made up in very large part of chondrules in all conditions, from mere frag- ments to nearly perfect spheres. The groundmass of the stone is a confused agglomerate of olivine and enstatite particles with interspersed metallic iron, iron sulphide, and chromic iron. In no case do the silicates oc- cur with perfect crystallographic outlines, nearly all, both olivine and enstatite, being of fragmental nature and of vary- ing size, ranging from particles a millimeter in diameter down to the finest dust. The iron has the usual form of blebs and extremely irregularly outlined patches serving as a cement, as shown in PI. V, figs. 6, 7, 8 and 9. By reflected light it shows up in strong contrast with the dull brassy yellow sul- phide. This last is in irregular form also, sometimes associated with the iron, sometimes quite isolated. So far as observed, it never occurs as rounded blebs enclosed in the iron, as some- times found in large masses of meteoric iron. On the other hand, the silicate minerals do thus occur. This is mentioned 46 MERRILL AND STOKES as having some bearing on the origin of the iron meteorites, the writer regarding them as residual masses of larger, coarse, granular forms from which the silicates have been lost through disintegration, perhaps before reaching the earth. Chromite in black specks is often associated with the sulphides, but does not in the section present good crystal outlines. The presence of alumina and alkalies, as indicated by Dr. Stokes's analyses, caused a careful search to be made for the presence of feldspar, but none was found, ^ though it is possible that sundry minute, clear and colorless, doubly refracting par- ticles may be thus referred. These never show twin structure, cleavage lines, nor other physical properties such as permit a definite determination. It is more probable, however, that these elements are accessory constituents of the enstatite. If such is the case, the stone, as shown by the analyses and micro- scopic investigation, is composed of nearly equal parts of highly ferriferous olivine and enstatite, the latter being low in mag- nesia in proportion as it is high in the accessory elements. Much of the interstitial material of the groundmass is so fine and dust-like that it is impossible to determine its mineral character in the thin section. After repeated trials the de- vice was adopted of taking some of the fragments, several grammes in weight, and after dusting them carefully with a camel's hair brush and blowing upon them to remove all external dust particles, placing them in a funnel upon a piece of silk bolting cloth and allowing a half liter or so of distilled water to trickle over them, drop by drop; the liquid was then evaporated in a porcelain dish and the resultant dust, which is believed to correctly represent the true ground- mass in an unaltered condition, was collected and submitted to a microscopic examination. It was found to be composed of beautifully fresh, sharply angular splinters, mainly of enstatite, though with some olivine and black glass (PI. Ill, fig. i). Cross sections of the thin portions of the crust (the highest 1 H. L. Ward (Am. Jour. Sci., Dec, 1899, p. 414) states that the stoue is feldspathic, and classes it with Meunier's montrejites. I cannot agree with him in this. It is essentially non feldspathic and belongs more nearly to Bre- zina's group 29, Kugelchen chondrite (C"). — G. P. M. point in fig. i, PI. I) show a black glass, interspersed with numerous residuary particles of unfused silicates, which passes down gradually into the unchanged granular stone. Sections of the thicker blebby glass from the lower surface show air vesicles and numerous crystallites imperfectly secreted from the glassy base and too small to be seen in the figure, together with residuary, unfused particles of the original minerals (PI. Ill, fig. 2). CHEMICAL COMPOSITION. As the nature and extent of the action of such solvents for nickel-iron as mercuric ammonium chloride on the troilite and the silicates of stony meteorites has not been satisfactorily de- termined, the method of magnetic separation was adopted. This, if thoroughly carried out, yields two portions, the one en- tirely free from metal, the other consisting of metal mixed with more or less silicate, troilite, and chromite. About 27 grams of the pulverized 'material was submitted to fractional separation by an electro-magnet, the more magnetic portions being finally separated by a weak magnet. Two frac- tions were thus obtained, the larger being entirely free from metal, but containing some material attracted by a strong mag- net. The relative proportions were : a. Portion free from metal, 72.05 percent. b. Portion consisting mainly of metal,. 27. 95 percent. Of each of these a complete analysis was made. By subtract- ing from the figures obtained for b the silica, oxides, and troilite the composition of the metal was deduced, while by combining them in proper proportion with the figures expressing the com- position of a the composition of the total stony part of the me- teorite was found. The usual separation of the stony part into portions soluble and insoluble in hydrochloric acid was also made, and of each a practically complete analysis was obtained. 48 MERRILL AND STOKES The analysis of the non-metallic portion, a, gave : SiOa 45- 60 TiOa .11 P2O5 .35 Al,03 3-04 Cr^Os .66 FeO II. II FeS 6.79 MnO .25 NiO trace CaO 2.26 MgO 28.82 K2O •32 NajO .92 LiaO faint trace ■ I above 110° .08 .26 100.58 No BaO or SrO could be detected. The .66 percent CrjOg corresponds to about 1,3 percent chromite. I have isolated a small amount of chromite in a state of imperfect purity by treat- ing a large quantity of the silicate repeatedly with hydrofluoric and sulphuric acids. .1147 gram of this gave : Cr^Oj AI2O3 FeO MgO TiO, 50.31 9.67 28.78' 2.76 1.20 The portion b^ containing all the metal, gave II. Fe 75-65 Cu .038 Ni 6.51 Co •55 SiO., 8.18 P.O5 .07 Al,03 1.32 Cr,03 .21 FeO 1.84 FeS .68 CaO •39 MgO 4-85 82.748 percent metal. 17.48 percent stony. 100.228 ' Total iron as FeO. ALLEGAN METEORITE AND MART IRON 49 The portion ^, therefore, contains 17.48 percent stony matter, but a comparison of the ratio of SiO^ and MgO shows that this has not the same composition as that of the main siHcate por- tion. While 50 percent of the Latter is soluble in hydrochloric acid, the former contains 39 percent soluble and 61 percent insoluble material, which may be due in part, at least, to the more magnetic properties of the enstatite. It is, therefore, er- roneous to assume that the silicate accompanying the metal has the same composition as the main portion. The above data give the means of calculating approximately the amount of FeO in the magnetic portion, while the troilite is found from a sulphur determination. The relatively greater amount of chromite in the magnetic portion is also noteworthy. For the separation of iron from nickel and cobalt, the am- monium sulphocyanate method of Zimmermann^ was used with satisfactory results. It is necessary to add a little more sodium carbonate than is sufficient just to destroy the red color of the ferric sulphocyanate and to heat to boiling, but even then the precipitation of the iron is not absolutely complete, a small portion always remaining in solution and being removed by am- monia after destroying the sulphocyanate by nitric acid, and concentrating. The bulk of the iron is free from nickel and cobalt after one repetition of the precipitation. Like all other methods for separating iron from nickel and cobalt, this is ap- proximative, but it avoids the tedious repetition of the acetate method and the precipitates filter well. The separation into a soluble and an insoluble portion was effected by repeated treatments with hot dilute hydrochloric acid and alternate digestion with caustic soda. The insoluble portion, after ignition, was 49.96 percent, and the analysis of this and of the solution gave : 'Ann. Chem. (Liebig), 199: 10. 50 MERRILL AND STOKES III. Soluble in HCl. SiO^ 17.26 TiO^ trace P2O5 •35' Ai,03 .67 Cr^Os .041 FeO 6.91 FeS 6.79 MnO .09 CaO •49 MgO 17.17 K2O .181 Na.O .08' Li,b .oiJ IV. Insoluble in HCl 28.17 .11 trace 2.41 .62 4.16 none .08 1.64 11-57 .14^ .842 none 221 50.04 49.86 From analysis II is deduced the percentage composition of the metal, as follows : Fe Cu Ni Co 91.42 .046 7.87 .66 99.996 From analyses I and II the composition of the total stony portion is as follows : SiO, TiO., 45-42 .10 P.O5 Al,03 Cr.Oa FeO •35 3^31 .69 11.02 FeS 6.57 MnO ■23 NiO trace CaO 2.24 MgO K2O Na.,0 28.60 -30 .86 Li,0 H.,0{ at 110° above 110° faiut trace .07 .24 ' By diflcrence. ''The figures for alkalies are not to be implicitly depended on, owing to the possible partial substitution of Na^jO for K^O during digestion with caustic soda. ALLEGAN METEORITE AND MART IRON 51 and, finally, the composition of the entire sample examined is: Fe Cu Ni Co SiOj TiO, P2O5 AI2O3 Cr,03 FeO FeS MnO NiO CaO MgO Na.O Li,0 rat 110° ^^^ I above 110° 2109 ^ .01 1 1. 81 I 34-95 1 .08 • 27 2.55 •53 8.47 5-05 .18 trace 1-73 21.99 • 23 .66 faint trace .06 •19 Metallic part 23.06 percent. ■ Stony part 76 94 percent. 100. in Specific gravity 3.905. An average of two determinations a picnometer flask at 27° C. by Dr. Merrill. THE MART IRON. HISTORY AND GENERAL FEATURES. The second meteorite to be described, wlriclr will be known as the Mart Iron, was found early in 1898, on the farm of H. T Vaughan, near Mart, in McLennan County, Texas. This iron weighed originally 15 V, pounds. From it a shce welh ng 456 grLs was cut for the collection of the Nattonal Mufeum the fron having been donated by the finder to the museum of Baylor University, at Waco, T^as. For the prtv.- lege of removing this slice we are mdebted to Mr O U Charlton, Curator of the museum. The or.g.na shape J^he u ;„ PI VT fifTs I and 2, was that of an irregu iron, as shown in rl. vi, ngs. i dii«a , , , , ,.,:tu ,ar oval, somewhat flattened at one side and rounded above, .t two large and deep pittings on the broader surface. The ong, nal dimensions were about 8.5 by 15 by 25.5 cm. It was 52 MERRILL AND STOKES seen to fall and had evidently lain in the soil some time, as the exterior was considerably oxidized and the troilite, which pre- sumably once occupied the pits, was completely eliminated. On cutting and etching the iron gave the surface shown in fig. I of the plate. The small dark points are due to troilite. Sundry cracks in the iron at various points on the etched surface, shown most plainly at the upper right in fig. i, are also filled with troilite. The blotches shown are due to the oxidation of the troilite in process of etching. Mr. Tassin, by whom the etching was done, calls attention to the perfection of the Widmannstatten figures, and particularly to the relief of the taenite bands. As shown by this etching, the iron belongs to the octahedral variety, and is of moderately coarse crystallization. Its general appearance is so similar to that of the Hamilton County (Texas) iron described by Howell ^ as to suggest that it may be a part of the same fall. The probability is still more evident when it is considered that the two localities are not over 50 miles apart in a straight line. The chemical evidence, as shown by a comparison of Mr. Eakins's analysis of the Hamilton iron with that of Dr. Stokes, is, however, not favorable to this view, though we believe the possible (if not probable) variation in composition in different parts of the same iron has not yet been fully worked out. Mart. Hamilton Co. Fe 89.68 86.54 Ni 9.20 12.77 Co 0-33 0.63 Cu 0.037 0.02 P 0.158 0.16 S 0.017 0.03 C O.II Chromite trace Fe^Oj t( 99.422 100.26 CHEMICAL COMPOSITION. The samples submitted were cut from the outer portion of the meteorite, including the oxidized crust; this was carefully re- ' Proc. Rochester Acad. Sci., Vol. I, 1890, pp. 87-89. *» SUPPLEMENTARY NOTE 53 moved by scraping and liling. There was a small quantity of rust in the cracks on the cut surface, but its amount was trivial. During the solution in aqua regia scales of schreibersite were observed. A few small black grains were left which showed crystal faces under the microscope, and which were identified as chromite by the usual reaction. A minute amount of color- less granular matter was also noted, the nature of which could not be determined. All determinations were made in a solution of the same por- tion of 3.8636 grams, the residue having been brought into so- lution and added. The analysis gave : 1,9318 grams .9659 gram Fe 89.68 Ni 9.20 Cu Co •33 P S Cr Total, 99-422 • .037 .158 .017 trace The separation of iron from nickel and cobalt was effected by the Zimmerman method (see preceding section on the Alle- gan meteorite). From the above data the composition of the iron may be expressed as follows : Nickel-iron (Fe, Ni, Cu, Co) 98.31 Schreibersite 1.06 Troilite .05 Chromite trace Fe^Os trace 99.42 SUPPLEMENTARY NOTE BY DR. MERRILL. The structure of the Allegan stone is such as to bring up prominently the question of the origin of meteorites in general. It is not, how^ever, my intention to go into or review the matter exhaustively, but rather to call attention to a few points here illustrated that have an important bearing on the subject. 54 MERRILL *^ As is well known, structures such as this stone possesses have been accepted by Tschermak and others as indicative of a tuffaceous origin — that is, they result not from the direct cool- ing of a molten magma, but from the agglomeration of already solidified particles, as is the case with volcanic tuffs. Others, of whom Dr. M. E. Wadsworth is a prominent example, regard them as the result of the hasty crystallization of an igneous magma. ^ That many meteoric stones result from the cooling and crystallization in place of igneous magmas is beyond ques- tion ; yet there is an almost equal certainty that others are of tuffaceous origin, though the nature of the evidence is not such as to be fully appreciable except by petrographers. In 1888 I described a meteorite from the San Emigdio range of California,- and announced my conclusion in favor of a tuf- faceous origin. Nevertheless, as the stone was badly weathered there has always been a question in my mind as to how much of the apparent fragmental appearance was due to w^eathering and how much was original. The absolutely fresh character of the Allegan stone, which is of a surprisingly similar nature, gives me an opportunity to reconsider the subject, and as it happens, to confirm the views first expressed. The general structure of the Allegan stone can, I believe, be accounted for only by regarding it as an agglomerate of chon- drules imbedded in a fragmental groundmass or matrix, the ma- terials for which were derived from the trituration of other chondrules. One fact which, in m}- mind, has always mitigated against the theory which would account for the peculiar structure of a meteorite of this type on the assumption of hasty crystallization, has been the complete absence of a glassy base in any but the chondritic portions. Obviously, if the stone is a product of cry- stallization in mass the chrondrules are products of the earliest crystallization, and should, judged by the standard of terrestrial petrography, be the most highly crystalline, while the base in which they are imbedded might be glassy or crvstalline, accord- 1 For a very clear exposition of Dr. Wadsworth's views see his Lithographic Studies, pp. 1 06- 1 17. -Proc. U. S. Nat. Museum, Vol, XI, iSSS, pp. 161-167. SUPPLEMENTARY NOTE 55 ing to conditions. In reality the reverse is the case, the chon- drules being more or less glassy, or at least imperfectly crys- talline, as in the barred and fan-shaped forms, while the groundmass of the rock is of crystalline particles, and of parti- cles of the chondrules themselves. So far as I have observed there is no true glassy base in meteorites of this type. That certain conditions of crystallization will give rise to the spherulitic forms of the enstatite is undoubted, but from a study of the crust of the Allegan stone it appears that when meteoric, material is fused and allowed to re-crystallize, even so rapidly as must have been the case in this crust, it is not spherulitic, but takes the form of crystallites in a glassy base, as among terrestrial rocks. It is evident that time is not the only factor that should be considered. The subject of the spherules in liparite has been pretty well worked out by Cross and Iddings,^ and while it is easy to con- ceive of the abrupt transition from a wholly or partly crystalline spherule to a glassy base, as sometimes seen in spherulites of obsidian, it will, in the present state of knowledge, puzzle any petrographer to account for an equally sharp transition from a glassy spherule (chondrule) to a base composed wholly of crys- talline particles, shown in many meteorites. Even could we account for such anomalies of crystallization as are above noted, the presence of plainly fragmental chondrules — chondrules which were fragments at the time of the final consolidation of the stone — remains to be explained. The forms shown in PL IV were all carefully picked from the rock. That they are original fragments, i. e., not due to fracturing in place, is shown by the dull and sometimes abraded character of the surface of fracture, and further by the fact that in no case was the re- mainder of the chondrule represented by one of these pieces found in the vicinity. Fig. 6 of this plate is one of the most striking illustrations of this nature, being that of a portion of an oval enstatite chondrule some 8 mm. in greatest diameter, im- bedded in a fine granular groundmass. The flat surface of fracture is dull and lusterless, and the fracture is, I believe, unquestionably an old one. Fig. 7 shows a side view of the 1 Bull. Phil. Soc. Washington, Vol. XI, 1891. 56 MERRILL same chondrule. In other cases, as in figs. 3 and 4 of the same plate, the fractures are old and show abraded surfaces. Figs. 2 and 5 are plainly those of fragments of elongated chondrules that have been broken across. Figs. 3, 4, 5 and 9, PI. V, are evidently sections of just such fragments, and in fig. 4 the splintering fracture along a cleavage plane of the enstatite (below in the figure) is plainly evident under the microscope. With reference to such forms as that shown in fig. 6, PI. V, one can assume that after the olivines had become imperfectly secreted the magma was resolved into spherical drops which cooled too rapidly for further crystallization, while in the en- statite forms crystallization may have been in some cases prior to the assumption of the globular form and in others subse- quent thereto. Such forms seem to lend support to the theory of Sorby ^ that "some at least of the constituent particles of meteorites were originally detached glassy globules, like fiery rain." It is possible to conceive that these chondrules, first as blebs of molten matter and then as consolidated particles, may have been triturated in the deep throat of some volcano. The spherical form, however, I do not regard as due to trituration, like volcanic lapilli, as formerly held by some writers, but rather to a previous molten condition. Be this as it may, consolidation must obviously have taken place before the mass was shot forth into space. The manner in which the metallic portions are wrapped about or even injected into the silicate particles and the chondrules (see figs. 6, 7, 8 and 9, PI. V) suggests the possible reduction of the iron — or at least a remelting in an atmosphere from which oxygen was largely excluded — after the stony portion assumed its present form. 1 Nature (London), Vol. 15, April 5, 1S77. PLATE I. Figs. I, 2, and 3. Allegan meteorite, as received at the National Museum. Fig. 3 on reduced scale and introduced to show outline of cross- section. (58) Proc. Wash. Acad. Sci. Vol ll Plate I. Jy xt-- The Allegan Meteorite. PLATE II. Broken surface of Allegan meteorite, magnified about five diameters. (60) The Allegan Meteorite. PLATE III. Fig. I. Fragmental particles washed out of Allegan meteorite. 2. Cross-section of thick crust from under surface of Allegan meteorite. The oval colorless areas are air vesicles ; the angular areas, residual particles of unfused silicates. Actual thickness of crust, 3 mm. (62) Proc. Wash. Acad. SCI. Vol.11 Plate III. The Allegan Meteorite. PLATE IV. Fig. I. Indented chondrule. Allegan meteorite. 2-7 inclusive. Fragmental chondrules showing old surfaces of fracture, indicating that the chondrules were fragments at the time of the consolidation of the stone in its present form. Actual sizes : fig. I, 2 mm.; figs. 2, 3 and 4, about 3 mm.; fig. 5, 4 mm. in length ; figs. 6 and 7, 8 mm. (64) Proc. Wash. Acad. Sci. Vol.11. Plate IV The Allegan Meteorite. PLATE V. Allegan meteorite. All greatly magftiified. Figs. I and 2. Clear colorless olivines in groundmass. 3. Radiating enstatite with clear olivines and iron (black). 4 and 5. Fragmental chondrules of nearly amorphous enstatite. 6. Chondrule composed of olivines imperfectly secreted from a black glass. The clear colorless portions in close contact represent olivines, the black areas, iron. 7. Imperfectly crystalline enstatite chondrule in contact with a frag- ment of a nearly amorphous one of the same mineral nature. 8. Iron, drawn to show the very irregular nature of the masses. 9. Nearly amorphous enstatite chondrule in contact with metallic iron. (66) Proc. Wash. Acad. Sci. Vol 1 1 PLATF.V The Allegan Meteorite. Kelio'-ypt Co.Bu5i«n. PLATE VI. Figs. I and 2. The Mart Iron. Proc. Wash. Acad. SCI. VoL.li. Plate VI The Mart Iron. PROCEKDINGS OF THE WASHINGTON ACADEMY OF SCIENCES Vol. II, pp. 69-77. July 25, 1900. THE GIANT SQi^IRRELS OF BURMAH AND THE MALAY PENINSULA. By Gerrit S. Miller, Jr. At least four giant squirrels occur in Burmah and the Malay- Peninsula.^ They are currently united under the name Raliifa ' hicolor ' notwithstanding the facts that their characters are well defined and that the Sciurus hicolor of Sparram was a Javan animal. The species may be distinguished as follows : Nasal bones scarcely or not at all expanded posteriorly ; teeth rela- tively heavy ; color of upperparts (in fresh pelage) black or blackish brown, sometimes a little yellowish tinged. Hind foot about 90 (85) ; outer surface of front legs entirely black; ears tufted Ratufa gigantea. Hind foot about 80 (75) ; outer surface of front legs partly yel- lowish brown; ears not tufted Ratiifa Dtela^topcpla. Nasal bones conspicuously expanded posteriorly; teeth relatively light ; color of upperparts (in fresh pelage) yellowish brown or grayish brown. Hind foot under So (75); mandible heavily built; underparts whitish in strong contrast with back Ratufa affinis. Hind foot over So (75) ; mandible lightly built; underparts dark yellowish brown in slight contrast with back.. i?a/z//«/j/'^o«o/a. 1 These are represented in the United States National Museum chiefly by material presented by Dr. W. L. Abbott. This paper is published here by per- mission of the Secretary of the Smithsonian Institution. Proc. Wash. Acad. Sci., July, 1900. (^) 70 MILLER RATUFA GIGANTEA (McClelland). 1839. Sciurtis gigantciis McClelland, Proc. Zool. Soc. London, vii, p. 150. 1892. Sciiirus bicolor Thomas, Ann. Mus. Civ. Storia Nat. Geneva, ser. 2, Vol. X, p. 20. Not of Sparram. Type locality. — Assam. Characters. — Size very large (total length about i m., hind foot about 90 mm.); ears distinctly tufted; general color black above, tawny yellow^ below, the latter extending down front legs to wrists, but strictly confined to inner surface ; skull with nasal bones narrowed posteriorly, and with audital bullte broad and low. Color. — Back, sides, top of head, ears, outer surface of legs, dorsal surface of feet, and entire tail glossy black. The color at first sight appears to be pure, but on close inspection it is seen that while the longer hairs are actually black the shorter hairs of the underfur are deep reddish brown, and that each is very faintly tipped with a minute yellowish speck. In certain lights the reddish under color becomes very apparent, while in others the yellowish tips of the shorter hairs produce a distinct variegation. The hairs of the tail are black at the tips only ; the greater portion of each is distinctly tinged with reddish brown. Cheeks, entire ventral surface of head and body, and inner side of front legs to wrists tawny yellow (very nearly the orange buff of Ridgway). This color extends down inner side of thighs, but on lower leg it soon becomes mixed with black, practically disappearing between knee and ankle. A black stripe about 30 mm. in length runs backward under eye from base of whiskers. It is separated by the yellow of cheek from both eye and lip. Two small black spots on chin. Skull. — The skull of Ratufa gigantea is very large (75 mm. in greatest length). Its general form is that characteristic of the genus. Outer border of nasals converging rapidly and uniformly to point about 6 mm. from posterior border, then nearly parallel. Front border irregularly convex, not double convex as in the other species. Palate strongly concave between premolars. Audital bullae broad and low, a line joining their summits (when skull is held upsidedown) barely clears the tips of the very prominent lateral processes of the basioc- cipital. Mandible heavily built, the ramus exceedingly robust in region between molars and incisor. Angular process broad and short. Teeth. — The teeth are heavily built but not peculiar in form. The crowns of the upper teeth in an adult male measure as follows: pre- molar 4x4; first molar 3.6 x 4; second molar 3.6 x 4.4; third molar 4 x 4.4. GIANT SQUIRRELS OF BURMAH AND MALAY PENINSULA 7 1 Measurements. — External measurements of an adult male : total length i03o; head and body 410; tail 610; pencil 100; hind foot 93 (88). Cranial measurements of some specimen : gi-eatest length 75 ; basal length 65; basilar length 61; palatal length 30; length of nasals 24.6; breadth of nasals anteriorly 14.6; breadth of nasals posteriorly 9 ; interorbital breadth 39 ; breadth between tips of postorbital proc- esses 43 ; zygomatic breadth 48 ; mastoid breadth 36 ; breadth of palate between premolars 1 1 ; depth of rostrum at posterior extremity of nasals 22; mandible 47; depth of mandible at posterior root of premolar 12.6; maxillary toothrow (alveoli) 15.4; breadth across both upper incisors together at rim of alveoli 8 ; mandibular toothrow (alveoli) 16. Specimens exa?fiined. — One, collected at Taho, Kareni, Burmah, by Fea. Re7narks. — In its large size, tufted ears, wholly black outer surface of front legs, and broad, low audital bullas Ratufa gigantea dif- fers conspicuously from the more southern representatives of the genus. RATUFA MELANOPEPLA sp. nov. Type. — Adult male (skin and skull) no. 83230 United States Na- tional Museum, collected in Trong, Lower Siam, February 27, 1S96, by Dr. W. L. Abbott. Characters. — Similar to Ratufa gigantea but smaller (total length about 800 mm., hind foot about 80 mm.) ; ears never tufted; tawny of inner side of front leg spreading conspicuously over outer surface be- tween wrist and elbow. Skull and teeth essentially as in R. gigantea. Color. — Fresh pelage : cheeks, throat, sides of neck, inner sur- face of legs, and entire ventral surface of body rich tawny yellow (in the type very nearly matching the ochraceous of Ridgway), the hairs mostly blackish at base. Remainder of pelage and entire tail glossy blue-black, the basal portion of the hairs tinged with dark reddish brown which 7iever appears on su7'face unless the fur is disarranged. Yellow of inner side of front leg spreading conspicuously over outer surface between elbow and wrist, where, like that on cheeks and side of neck, it becomes distinctly paler (generally somewhat yellower than the buff of Ridgway). A black spot on chin. No black stripe o?i cheek behind roots of xvhiskers. Bleached pelage: yellow por- tions of body essentially as in fresh coat ; black areas to a greater or less extent a dull indefinite, reddish brown (not far from the cinnamon 72 MILLER of Ridgway, but often even lighter) becoming slightly more yellow on flanks. Skull. — The skull is as in Rat?ifa giganteay though not so large. Audita! bullae narrower and lateral processes of basioccipital less de- veloped. Palate distinctly concave between premolars but less so than in R. gigajitea. Anterior margin of each nasal concave, so that the outline of the two together is distinctly double concave, the region of the median suture projecting as a conspicuous decurved point. Man- dible with angular process narrower and longer than in R. gigantea. Teeth. — The teeth are essentially as in Rattifa gigantea., though the crown of the middle upper molar appears to be less nearly square in outline. Measurc?Jients . — External measurements of t3'pe : total length Soo ; head and body 36S ; tail 431 ; pencil 80; hind foot So (74). Average of eight specimens from the type locality : total length 776 (762-S35) ; head and body 350 (305-3S1); tail 437 (406-462); hind foot 81.5 (79-85); hind foot without claws 73.7 (6S-76). Cranial measurements of type : greatest length 70 ; basal length 59 ; basilar length 53 ; palatal length 26.4 ; diastema 16; length of nasals 23.4; breadth of nasals anteriorly 13; breadth of nasals posteriorly 7 ; interorbital breadth 28 ; breadth between tips of postorbital proc- esses 41 ; zygomatic breadth 44; mastoid breadth 32.6; breadth of palate between premolars 10; depth of rostrum at posterior extremity of nasals 18.8; mandible 41.6; depth of mandible at posterior root of premolar 11; maxillary toothrow (alveoli) 14; breadth across both upper incisors together at rim of alveoli 8; mandibular toothrow (alveoli) 14.4. Specimens examined. — Twelve, all from the type locality. Remarks. — Ratufa melaiiopepla is readily distinguished from R. gigantea by its size, untufted ears, color of front legs, absence of faint annulation of underfur of back and absence of black sti'ipe on cheek. The cheek stripe is occasionally indicated, but apparently never well developed. The striking change in color due to bleaching of the black fur is closely paralleled by that recorded by Thomas' in Ratufa bjinguranensis. In one animal, however, the change is from black to cinnamon, and in the other from Prouts brown to " dirty whitish fawn." The fresh black coat first appears on the anterior half of tlie body, gradually spreading backward and encroaching on the area occupied by the brown. The abrupt contrasts of color thus brought about are frequently very striking. A yellowish brown 'Novitates Zoologica', II, p. 491, December, 1S95. GIANT SQUIRRELS OF BURMAII AND MALAY PENINSULA 73 suffusion occasionally appears on the nape in fresh pelage ; but the dif- ferent quality of the fur will at once distinguish such marking from the effect of sunburn ; while the hairs are never annulatcd as in the related species Ratiifa pyrsonota. This species is closely related to the Javan Ratufa bicolor^ but differs in the clear black of the upper- parts. RATUFA AFFINIS (Raffles). 1822. Sciurus affinis Raffles, Trans. Linn. Soc. London, xiii, p. 259. Type locality. — Singapore. Characters. — Size less than that of the other giant squirrels of the Malay Peninsula (total length about 700 mm., hind foot about 75 mm.); ears not tufted; general color grayish brown or yellowish brown above, the hairs without annulation, d7ill white below in ab- rtipt and striking contrast ; skull with nasal bones expanded pos- teriorly, and with deep, heavaly built mandible. Color. — In fresh pelage : back a fine grizzle of ecru drab, Isabella color and grayish white, the under fur entirely Isabella color, the longer hairs whitish at base and Isabella color at extreme tip but not annulated, the general effect intermediate between the drab and ecru drab of Ridgway. On shoulders and neck the ecru drab and grayish white predominate, giving way posteriorly to the Isabella color which also becomes more pronounced along middle of sides, to deepen on lower part of sides and outer surface of front legs to cinnamon, and on thighs to brownish russet. The hairs of the darker areas of outer side of thighs and front legs are obscurely annulated at tip. The cin- namon forms a distinct band on sides separating color of back (into which it fades) from whitish cream buff of belly. Top of head like shoulders and neck, but somewhat darker; cheeks griz- zled hair brown, this color continuous across chin in a band about 15 mm. wide; sides of muzzle dull whitish, the pale area extend- ing back a short distance along upper lip ; upper part of muzzle marked with seal brown. Ears seal brown externally, grizzled hair brown internally. Whiskers black. J^ect white or light grizzled biiff. Tail dull ochraceous buff above, marked with about fifteen faintly shaded cross bars of drab, and tinged with drab at tip. The hairs are all buffy white at extreme base, and this color appears irregu- larly and inconspicuously along median line. Lower surface of tail much like upper, but distinctly clouded with bister, and with the cross bars even less noticeable. The buffy white bases of the hairs form a distinct median longitudinal area about 25 mm. in width, divided 74 MILLER down the middle by the narrow, sharpl}' defined line formed by the short bister hairs with which the median ventral surface of the tail is clothed. Inner surface of legs concolor with belly but much tinged with the color of the svn-rounding parts. This is especially the case with the hind legs. In bleached pelage the drab areas of the fur are faded to a pale dirty buff, strikingly different from the color of the fresh coat, but the cinnamon lateral band and the hair brown cheeks remain unchanged. Skull. — Though the skull of Ratitfa affinis is much below the maximum size of members of its genus it differs from the others in minor details only. The nasal bones are distinctly expanded poste- riorly, that is the width of the two together along naso-frontal suture is considerably greater than it is at a point about 5 mm. further for- ward. This character alone is sufiicient to distinguish the species from Ratufa gigantea and R. ?nelanopcpla. Pterygoids short, the length of the interpterygoid fossa not twice as great as the width at middle, instead of much more than twice, as in the related species. Teeth. — The teeth are of the same form and relatively of the same size as in R. gigantea and R. melanopepla. Measic7-eme7tts. — External measurements of an adult male topotype (no. 86886): total length 711; head and body 330; tail vertebrae 381; pencil, 75; hind foot 77 (70). A second specimen (also a topotype, (no. 86887): total length 686; head and body 305 ; tail vertebrae 381 ; pencil, 90; hind foot 78 (73). Cranial measurements of adult male: greatest length 61; basal length 52; basilar length 49 ; palatal length 34 ; length of nasals 19. 8 ; breadth of nasals anteriorly 12 ; breadth of nasals posteriorly 10; least breadth of nasals 6.6 ; interorbital breadth 26 ; breadth between tips of postorbital processes 37 ; zygomatic breadth 39 ; mastoid breadth 29 ; brcadtli of palate between premolars 8.4 ; depth of rostrum at posterior extremity of nasals 17.4; mandible 37; depth of mandible at posterior root of premolar 10.4; least depth of mandible between incisors and premolar 8; maxillary toothrow (alveoli) 12.8; breadth across both upper incisors together at rim of alveoli 8; mandibular toothrow (alveoli) 14. Specimens exa7ni?icd. — Two, both from the type locality. Remarks. — In tlie absence of Javan and Sumatran material it ap- pears preferable to use the name a(ll)iis for this squirrel rather than attempt to apply to it one of the earlier names based on animals of the same general size and color. In a group so plastic as the genus Ratufa the chances that any one form occurs in both Ja\a and the GIANT SQUIRRELS OF BURIVIAH AND MALAY PENINSULA 75 Malay Peninsula arc infinitesimal ly small. It is worthy of note that the seasonal changes in color of this squirrel were described though not fully interpreted by Raflles. RATUFA PYRSONOTA sp. nov. Type. — Adult male (skin and skull) no. S34S3 United States Na- tional Museum, collected in Trong, Lower Siam, September 29, 1896, by Dr. W. L. Abbott. Characters. — General ajjpcarance much as in Ratufa affiitis., but size larger (essentially as in R. inelanopephi) and color throughout darker and iiiore rich, the underparts tavony yellowish brown scarcely contrasted with that of sides and back ; hairs of back noticeably an- nulated near tip ; skull with nasal bones as in Ratufa affinis^ but mandible slender and lightly built. Color. — Back, sides, outer surface of legs, and top of head uniform grizzled ochraceous, the general effect almost exactly matching No. 7 of Ridgway's Plate V. The individual hairs are all slate gray at base (this color does not appear at surface), those of the soft underfur light ochraceous at tip, those of the longer fur ochraceous at middle then dark tawny to tip, each, however, with a sharply defined subterminal band of light buff about 2 mm. in width. The color is remarkably uniform over entire dorsal surface, but is faintly paler across shoulders and darker in lumbar region. On sides, flanks and thighs it becomes more nearly tawny ochraceous and the speckling due to the light subter- minal bars nearly disappears. Outer side of thigh with conspicuous patch, 40—50 mm. in length, of whitish cream buff, lying mostly within the area occupied by the color of the side but bordering on the area of color of ventral surface. This pale area although not sharply defined is very noticeable, and none of the specimens are without it. Underparts of body and inner surface of legs clear ochraceous to base of hairs in very slight contrast with color of sides (such difference as •there is mainly due to absence of speckling on ventral surface), darker on throat, under side of neck, and middle of belly, paler on inner sur- face of legs. Cheeks and chin grizzled hair brown. Side of muzzle whitish ; top of muzzle seal brown. Ears clear seal brown externally, grizzled hair brown internally. I^eet clear seal brown or occasionally somewhat less dark, but never grizzled. Whiskers black. Tail very dark Prouts brown turning to rather light burnt umber at tip, the whitish bases of the hairs appearing inconspicuously at surface *above and much more noticeably below. Line of short hairs on 76 MILLER ventral surface of tail seal brown. Bleached pelage not known. In the type and one or two other specimens the last remnants are still visible in the form of some irregular patches of hair in the tail much lighter than that of the surrounding parts. Skull. — Skull nearly equal to that of Rahifa melanopepla in size, but in form of nasal bones agreeing with that of R. ajffinis. Ptery- goids long, the length of interpterygoid fossa much more than twice the breadth. Teeth. — While the teeth show no marked peculiarities in form they are actually as well as relatively smaller than in Ratufa affinis and R. incla?topepla, and the upper premolar is narrower. J\Ieasuremcnts. — External measurements of type : total length 756 ; head and body 356, tail vertebrae 400 ; pencil So ; hind foot 85 (77)' Average of seven specimens from the type locality: total length 749 (717-775) ; head and body 341 (330-356) ; tail vertebrae 408 (400-432); hind foot 81.4 (75-85); hind foot without claws 72.5 (69-77). Cranial measurements of type : greatest length 65 ; basal length 55 ; basilar length 51 ; palatal length 25.4; length of nasals 21 ; breadth of nasals anteriorly 12; breadth of nasals posteriorly 11.4; least breadth of nasals 8 ; interorbital breadth 26 ; breadth between tips of postorbital processes 38.6; zygomatic breadth 41; mastoid breadth 31; breadth of palate between premolars 10; depth of rostrum at posterior extremity of nasals 17.4; mandible 37; depth of mandible at posterior root of premolar 9.8 ; least depth of mandible between in- cisor and premolar 6.4; maxillary toothrow (alveoli) 13; breadth across both upper incisors at rim of alveoli 7 ; mandibular toothrow (alveoli) 1 1.6. Specimens examined. — Seven, all from the type locality. Rc7narks. — Ratufa pyrsonota differs so widely from R. a/fiiiis in both size and details of color that no close comparison between the two species is required. The line speckling of the upperparts, due to tlie minute pale subterminal bands of the longer hairs, is alone suilicient to distinguish the northern animal. Another noticeable character is the color of the feet. This is much darker than the leg in Ratufa pyrsofiota^ and conspicuously lighter than the leg in R. affuiis. From Ratufa vicla)iopcpla this species is readily separable by its smaller teeth and differently shaped nasal bones, aside from the totally dif- ferent color. Even bleached specimens of the black-backed animal show only the most superficial likness to R. pyrso)iota., a similarity that at once disappears when the details of coloration are examined. GIANT SQUIRRELS OF BURMAH AND MALAY PENINSULA 77 For instance, no amount of sunburn of the black-bucked species could produce the speckling characteristic of the other; and, however much alike the coloring of the underpart may appear, the hairs are always conspicuously blackish at base in Ratufa melanopepla^ while in R.pyrsonota they are uniform throughout, or possibly slightly paler proximally. JSfote. — Since this article has been in type I have received Mr. J. Lewis Bonhote's paper ' On the squirrels of the Ratufa {Sciurus') bicolor group' (Ann. and Mag. Nat. Hist., 7th ser., V, pp. 490-499. June, 1900). Three species are there recognized as occurring in the Malay peninsula : Ratufa gigantea^ R. bicolor^ and R. affims, the last divided into two races, R. ajffinis typica and R. affi}iis aurei- venter. Mr. Bonhote's Ratufa gigantea is the same as that of the present paper, and his Malayan R. bicolor is probably my R. melan- opepla. His two races of R. ' afjinis,^^ however, I cannot identify with either of the light colored species that I have examined. From Ratufa pyrsonota they are both separable by their pale feet, while from R. affinis they are equally well distinguished by the annulation of the hairs of the back. The Ratufa aureiventer^ with its yellowish orange belly, is evidently quite different from R. afiuis^ and should doubtless be recognized as a distinct species. The Ratufa afinis typica of Bonhote, though less strikingly distinct from R. affinis^ can hardly be the saine as Raffles's animal. In addition to the annu- lated fur of the back it appears to differ in the wholly yellowish brown head, in the absence of the cinnamon tints on sides, thighs, and front legs, and in the pure white underparts. This last discrepancy may be due to individual choice of descriptive terms; but the absence in the description of typicus of all reference to the dark brown cheeks and ears, and cinnamon at least of sides — these characters all equally conspicuous in both pelages of ajffinis — is scarcely to be thus explained. I am inclined to believe that the Ratufa of Johore is distinct from that of Singapore Island. Should this supposition prove to be correct the Johore animal will need a name, as typica Bonhote 1900 is preoccupied in the genus Rattifa by typiciis Sclater 1891 (Catal. Mamm. Ind. Mus. Calcutta, II, p. 7). In any event Mr. Bonhote's paper and mine supplement each other to the extent of increasing the number of mainland forhis of Ratufa ' bicolor^ from four to five or six. PROCEEDINGS OF THE WASHINGTON ACADEMY OF SCIENCES Vol. II, pp. 79-81. July 25, 1900. DESCRIPTIONS OF TWO NEW SQUIRRELS FROM TRONG, LOWER SIAM. By Gerrit S. Miller, Jr. Two of the squirrels collected by Dr. W. L. Abbott in Trong, Lower Siam, and presented to the United States National Museum, are at least subspecifically distinct from the forms of the same species occurring at Singapore. Neither appears to have hitherto received a name. This paper is published here by permission of the Secretary of the Smithsonian Institution. SCIURUS NOTATUS MINIATUS subsp. nov. Type. — Adult female (skin and skull) no. 84415 United States Na- tional Museum, Trong, Lower Siam, February 25, 1897. ^ Characters. — Similar to the form of Sciurtis notatus inhabiting Singapore Island, but red of underparts much darker, and terminal pencil and often the entire distal half or third of tail uniform red be- neath. Color. — Entire dorsal surface of head and body, a fine grizzle of black and yellowish wood brown, the individual hairs black with two or often three brown rings. The wood brown is (at the surface) everywhere in excess of the black. Cheeks, outer surface of legs, and upper surface of feet dull cinnamon, faintly grizzled. Underparts and inner side of legs rich clear rufous, the extreme base of hairs slaty. On sides the color of belly is separated from that of back by two longitudinal stripes extending from just behind axilla to flanks; 1 "Shot in heavy forest on the hills at about 1,000 feet ; apparently not com- mon in heavy forest." Collectors' note. Proc Wash. Acad. Sci., July, 1900. (79) 8o MILLER the lower blackish brown, the upper buff. Each is about lo mm. in width, though this is variable. The tail at extreme base, both above and below, is colored like back. On both surfaces the grizzle soon becomes much more coarse, giving way near middle below and at extreme lip above to clear rufous slightly darker than that of belly. Whiskers black. Skull and teeth. — The skull and teeth do not differ appreciably from those of the Singapore form. Measurements. — External measurements of type : total length 38 1 ; head and body 203; tail vertebrae 177; pencil 50; hind foot 50 (46). Average of ten specimens from the type locality: total length 391 (38 1 -41 3) ; head and body 205 (197-216) ; tail vertebrae 1S5 (165- 197); hind foot 49.2 (48-51.6); hind foot without claws 45.7 (45- Remarks. — The peculiarities of this race of the plantain squirrel were noted as long ago as 1S86 by Thomas, who says : " . . . . one is struck by the greater prevalence and greater intensity of the red colour of the belly in the northern Malay specimens as compai-ed with the southern ones .... red bellied specimens have in all cases red-tip- ped tails, while white and yellow-bellied ones have the tip annulated like the rest of the tail."^ This exactly states the differences between the races inhabiting Trong and Singapore, though the rich tawny ochraceous underparts of the latter would hardly lead one to call the animal " yellow-bellied." The extent of the clear red on the under surface of the tail is somewhat variable ; occasionally this color is con- fined to the pencil. SCIURUS TENUIS SURDUS subsp. nov. Type. — Adult male (skin and skull) no. S4412 United States Na- tional Museum, Trong, Lower Siam, February 3, 1897- Characters. — Similar to the typical form of Sciurus tenuis Hors- field from Singapore, but color paler and duller throughout ; flanks and shoulders with scarcely a trace of the ochraceous wash so notice- able in topotypes of S. tefiuis. Color. — Dorsal surface almost precisely as in Sc/urus notatus miniatus.1 except that the colors are more closely blended owing* to the greater extent of the light icings on the individual hairs and their less sharp demarkation from the black. Shoulders, flanks, and outer surface of legs just perceptibly tinged with dull ochraceous. Sides ' Proc. Zool. Soc. London, 1SS6, p. 77. TWO NEW SQUIRRELS FROM TRONG, LOWER SIAM 8 1 like back but slightly paler. Underparts and inner surface of legs whitish smoke gray faintly washed with buff on throat, the pale ventral area much narrower than the red in S. notatus. Tail coarsely grizzled black and ochraceous, the individual hairs ochraceous at base (2 mm.), then narrowly (i mm.) ringed with black, then again with ochraceous (3 mm.), then with black (8 mm.), and finally tipped (5 mm.) with pale buff ; pencil clear buff. On under side of tail the colors tend to a definite arrangement in longitudinal bands. Whiskers black. A light buff area behind each ear. Skull and teeth. — The skull and teeth are not distinguishable from those of typical 6". tenuis. Measure77ients. — External measurements of type : total length 244 ; head and body 130; tail vertebra 114; pencil 45 ; hind foot 37 (35). Average of five specimens from the type locality : total length 247 (235-257); head and body 137 (127-149); tail vertebrae 108 (i03- 127) ; hind foot 36.2 (35.4-37) ; hind foot without claws 33.7 (33-35). Rejnarks. — As compared with the race occurring at Singapore this squirrel is less differentiated than the Trong form of Sciurus notattis., and, curiously enough, toward dullness instead of richness of color. PROCEEDINGS OF THE WASHINGTON ACADEMY OF SCIENCES Vol. II, pp. S3-109. July 26, 1900. PRELIMINARY REVISION OF THE EUROPEAN REDBACKED MICE. By Gerrit S. Miller, Jr. It has been generally supposed that only three redbacked mice, Evotomys 7'utilHS, E. riifocamis^ and E. '' glai-eohts,' oc- cur in Europe, some writers assuming that even this small num- ber should be reduced. That this view is highly incorrect is shown by the material recently collected for the United States National Museum by Mr. J. Alden Loring, Miss Thora Stei- neger, and Mr. Robert T. Young, supplemented by series pre- sented and loaned by Mr. Oldfield Thomas and the specimens in my private collection, a total of about 300 specimens. While it is clear that the number of recognizable forms must be increased from three to ten or thereabouts, it is almost equally certain that still others remain to be discovered in the many unexplored portions of Europe. At present, therefore, I am able to give nothing more than a preliminary revision of the group, limited to the mice occurring west of Austria. Even in this small area much more collecting must be done before it will be possible to solve all of the systematic problems relating to this one genus. During the preparation of this paper, published here by per- mission of the Secretary of the Smithsonian Institution, I have received aid from Mr. Oldfield Thomas, Mr. G. E. H. Barrett- Hamilton, Dr. Robert Collett, and Mr. Vernon Bailey. NOMENCLATURE. While it is not necessary to enter into any further discussion of the generic names that have been applied to the redbacked Proc Wash. Acad. Sci., July, 1900. (S3) 84 MILLER mice ' the specific names based on European members of the genus Evototnys must be examined in detail. There are eleven of these. Bicolor. Myodes bicolor Fatio, Revue et Magasiii de zoologie, 2^ ser., XIV, p. 357, July, 1862. Although I have seen no topotypes of this annual I have little doubt that it is the boreal redbacked mouse of northern Switzerland, Evotoviys nageri Schinz. The type was taken in the Genthal, Berne, Switzerland, at an altitude of iSoo m. Fulvus. Arvicola fulvus Millet, Faune de Maine-et-Loire, II, p. 40, 1828. This name, as pointed out long ago by De Selys-Long- champs," applies to the redbacked mouse of France.^ It is, however, a mere misidentification of the Lennnus fulvjis of Etienne Geoffroy St. Hilaire* and the Arvicola fulvus of Desmarest, a species of Microtus (see De Selys-Longchamps, Faune Beige, I, pp. 34-35. 1842). Glareolus. Mus glareolus Schreber, Siiugthiere, IV, p. 680. The specific name glareolus was applied by Schreber to a short-eared mouse found by O. F. Midler among the beach vegetation of the Island of Laaland, Denmark.^ The animal is described in the text at the end of the account of Mus ceconomus^ and is figured on plate 190 B. The description is as follows: "It is cinnamon brown, darker on the middle of the back and whitish gray on the belly. The whiskers • For detailed treatmeut of the subject see North American Fauna, no. 12, July 23, 1896. Note on the name Anaptog07iia : Proc. Biolog. Soc. Washing- ton, XIII, p. 154, June 13, 1900. ^^^ Nota. Li' Arvicola /7ilviis de M. Millet, Faune-de-Maine-et-Loire, se rap- porte a VArv. rubidus \j^ Evotoviys hercyniciis rubidiis'\ et non a uotre Fulvus. Le meme erreur existe dans plusieurs musees d'Allemagne. " Mouogr. de Micromamm., |jp. 99. 1S39. 3 The original description is as follows : " Pelage roux-fauve en dessus ; d'uu gris roussatre sur les cot^s ; d'un blanc teint de jaunatre en dessous ; queue velue ; un peu plus courte que la moitid du corps, noiratre en dessus blanchatre en dessous ; pieds blanchatres ; oreillcs courtes, ovales, arrondies, tres-velues en dedans et en dehors. Longueur du corps et de la tete 3 ponces 9 lignes [95 mm.], de la queue 22 lignes [47mm.] ..." This copy was kindly furnished by Mr. Witmer Stone. 'Cat. JNIanim. Museum d'Hist. Nat., Paris, p. 1S7. 1S03. The original de- scription, a copy of which was recently made for me by Miss Thora Steineger, is as follows : "Queue plus courte que la moiti^ du corps, et plus velue que dans les pr(ic(!'dcus [?]; les poils sont tons cendres dans leur presque totality et fauves seulcment a l'extr(jmit^ ; les ongles sont jaunes, les oreilles presque euticremcnt cachd-s par les poils." 5 "... in dor lusel Laland zwischen dem Elyniiis arcnarius im Sande am Ufer der Ostsee. . ." EUROPEAN REDBACKED MICE 8$ also are brown." ^ The plate represents a misshapen animal with slender body, disproportionately large head, and no ears. In color it is light woodbrown reddening toward fawn. The name appears to have been left unused until 1S34, when Melchior applied it, not with- out considerable misgiving," to the redbacked mouse of Denmark. Melchior recognized the insufficiency of Schreber's description and plate (he suggests that the latter is a "bad drawing of a badly stuffed specimen ") and suspected that the redbacked mouse might be an un- described species. He preferred, however, to use the name glareolus provisionally for the animal until specimens could be obtained from the type locality.^ Subsequent authors have accepted Melchior's use of the name, but have lost sight of the difficulties that it involves. These difficulties are : that nothing in Schreber's description or plate points to the redbacked mouse, while both strongly suggest a young Microtiis agrestis\ that AIus glareolus inhabited the beach, a locality where Evotojnys is rarely seen, but where Microtus generally abounds ; and finally that if the name, as it undoubtedly must, be as- sumed to receive fixity of application to the redbacked mouse not earlier than 1834, it is certainly antedated by the Hyptid<^zis hercyniciis of Mehlis (1831) and possibly by the Le?nmus rubidus of Baillon (1834) also. Therefore, if the name glareolus is not transferred to the synonymy of Microtus agrestis^ its most proper disposition, it must be placed in that of Evotomys hercyiiicus. Hercynicus. Hypudccus hercyiiictis Mehlis, Isis, Jahrg. 1831, p. 876, 1 83 1, is the first tenable name for the small redbacked mouse of central Europe. The type locality is in the higher part of the Harz Mountains, Germany.* • " Sie ist zimmtbraun, mitten auf dem Riicken dunkler, und auf dem Bauclie weissgrau. Die Bartborsten sind ebenfalls braun." 2 " Den her beskrevne siellandske [Art] kan altsaa ei vsere Pallas's mus rntilus, og den maa altsaa enten vsere en aldeles ny Art, eller Schreber's mus glareolus, som manhidtil ikkun kiender af et Exemplar, fundet i Lolland af vor ber^mte Landsmand Conferentsraad O. Fr. Miiller ; dette sidste anseer Forf, for sandsynligt, og har derefter valgt Bencevnelsen da deu Schreberske Figur (Tab. 190), skiv^ndt den synes at vaere en slet tegning efter et slet udstoppet Exem- plar, dog har endeel Lighed med vores. Sagen vilde F^rst med Sikkerhed kunne oplyses, naar man var saa heldig at erholde et lollandsk Exemplar til Sammenligning, hvilket hidtil ikke har villet lykkes Forf." Den danske Stats og Norges Pattedyr, p. 119. 1834. ^Redbacked mice are now known to be common in Laaland ; but Microtus agrestis abounds there also, robbing the fact of the significance it would other- wise have. * " Sie ist in dem hohern Harze, z. B. am Bruchberge und in dessen nahern Umgebungen haufig. . . " 86 MILLER Nageri. Hypudceiis fiageri Schinz, Synopsis Mammalium, II, p. 237, 1845, is the first name based on the large, boreal redbacked mouse of the Alps. The type was collected on Mt. St. Gotthard near Andermatt.' Pratensis. Arvicola prate?zsts, F. Cuvier, Histoire Naturelle des Mammiferes, VII, Table Generale et Methodique, 1842 (Description and figure in same volume, livr. 68, March, 1834) is probably the redbacked mouse of northern France, Evotoinys hercynicus rubidus. Type locality, Abbeville, Somme, France. Riparius. Arvicola ripariiis Yarrell, Proc. Zool. Soc. London, 1832, p. 109, May 22, 1832, is the only name thus far based on the redbacked mouse of England. It is antedated by Arvicola ripariiis Ord 1825, a name applied to the common meadowmouse of the eastern United States (= Microttis pennsylvanicus) . No definite type localitv is given in the original description. Rubidus. '■'■ Leinmus riihidiis Baillon, Mem. Soc. Emul. Abbe- ville, 1834." This name I have been unable to verify. That it was originally applied to the redbacked mouse of the vicinity of Abbe- ville, Somme, France, was shown by DeSelys-Longchamps in 1S47.'' It is tenable for the dark race of Evotomys hercynicus occurring in Belgium and northern France. Rufescens. Arvicola rzcfescejis r>e Selys-Longchamps, Essai mon- ographique sur les Campagnols des Environs de Liege, p. 13, 1836, is a synonym of Evotomys hercyfiicus rubidus. That it was based on specimens taken at Waremme, Belgium, I have recently been in- formed by the author himself. Rufocanus. Hypudcezis rufocanus Sundvall, Oefversigt af Kongl. O Vetenskaps-Akademiens Forhandlingar, Arg. 3, p. 122, May 13, 1846. This is the first and only name for the European member of the sub- genus Craseotnys. Rutilus. Mus rntihis Pallas, Nov. Spec. Qiiadr. e Glir. Ord. p. 246, 1778, was based on a short-tailed arctic redbacked mouse from Siberia east of the Obi. No specimens from this region are available for comparison with the European animal to which the name rutilus is now applied. ' " Habitat in valle ursorum montis St. Gotthardi." 2 Revue zoologique par la Society Cuvierienne, lo"^ ann^e, p. 309. October, 1847. EUROPEAN REDBACKED MICE 87 THE GENUS EVOTOMYS AND ITS SUBGENERA. Genus EVOTOMYS Coues. 1874. Et'o/omys CouES, Proc. Acad. Nat. Sci. Philadelphia, ■ p. 186. Type, J/us rutiliis Pallas. 1899. Anaptogofiia Cope, Journ. Acad. Nat. Sci. Philadelphia, 2d ser., yi, p. 201 (part). Generic characters. — Teeth rooted as in Phenacoinys ; palate ter- minating in a thin-edged shelf between alveoli of posterior molars. Remarks. — The genus Evotomys as thus defined includes two well characterized subgenera, one represented by the aberrant Evototnys riifocajuis alone, th6 other containing the typical species. It is a boreal genus of circumpolar distribution, extending south to the transi- tion zone in America, and apparently somewhat farther in Europe. Subgenus EVOTOMYS Coues. Type. — Evotomys rtitilus Pallas. Characters. — Skull smooth and rounded, the postorbital processes obsolete ; teeth relatively much smaller than in JMicrotiis, the molar row considerably shorter than diastema; roots of molars fully formed relatively early in life ; root of posterior lower molar not forming an obvious swelling where it comes in contact with root of incisor. Remarks. — The subgenus Evotomys is a very compact and well characterized group. It contains all the known species of the genus, EvotoDiys I'ufocamis excepted, and its range is coincident with that of the genus. Subgenus CRASEOMYS subgen. nov. Type. — Evotomys rufocamis (Sundevall). Characters. — Skull as strongly angular as in Microtus., the postor- bital processes well developed ; teeth relatively as large as in Microtus., the molar row about equal to diastema ; roots of molars developed late in life ; root of posterior lower molar lying in a distinct capsule on lingual side of incisor root. Remarks. — The subgenus Craseomys shows a distinct approach to Microtus in the general character of the skull, and in the encapsuled posterior lower molar. The palate, on the other hand, and the molar roots, are typical of Evoto?nys. THE EUROPEAN SPECIES OF EVOTOMYS. Ten forms, species and subspecies, of redbacked mice are now known to inhabit Europe. Their characters are as follows : 88 MILLER Synopsis of the Redbacked Mice of Europe. Skull strongly angular as in Microtus^ the postorbital processes well developed ; teeth large, rooted late in life, the maxillary molar row about as long as diastema; posterior lower molar distinctly encap- suled ; red area of back very strongly contrasted with clear gray sides (subgenus Craseottiys) f Evotomys rufocanus Skull much less angular than in Microtus^ the postorbital processes obsolete or very small ; teeth small, rooted early in life, the maxillary molar row conspicuously shorter than diastema ; pos- terior lower molar not encapsuled; red area of back not very sharply contrasted with yellowish gray of sides (subgenus Evotomys'). Tail densely furred, forming about 30 percent of total length, its pencil at least one-fourth as long as caudal vertebrae. Evotomys rutilus Tail sparsely furred, forming about 33 percent of total length, its pencil never more than one-sixth as long as caudal verte- brcE. Greatest length of skull about 36 mm. ; hind foot 30 (iS) or more. Ratio of tail vertebrae to total length 35 to 3S ; skull broad and rounded, the zygomata light, generally con- vergent anteriorly and never abruptly flaring. Evotomys nageri Ratio of tail vertebrce to total length 32 to 36 ; skull nar- row and ridged, the zygomata heavy, abruptly flaring anteriorly and generally convergent poste- riorly. Nasals narrow (greatest width of both together dis- tinctly less than half length) ; ratio of tail verte- brae to total length generally 34 to 36. Evotomys norvegicus Nasals broad (greatest width of both together about half length) ; ratio of tail vertebrae to total length about 33 Evotomys vasconi^ Greatest length of skull about 34 mm. ; hind foot 20 (iS) or less Evotomys hercynicus EUROPEAN REDBACKED MICE 89 Skull relatively broad and audital bulhc relatively large ; dorsal red area (in winter pelage) ill defined and much intermixed with gray E. h. hclvcticus Skull relatively narrow and audital bullae relatively small ; dorsal red area never conspicuously mixed with gray. Summer pelage clear russet slightly darkened by chest- nut tipped hairs ; winter pelage more yellowish. E. h. hercynicus Summer pelage bister, mars brown, Vandyke brown, Prouts brown, mummy brown, etc., never clear russet ; winter pelage so far as known never strongly tinged with yellowish. Summer pelage with red dorsal area sharply defined and very narrow E. h. suecictis Summer pelage with red dorsal area broad, diffuse. Hind foot usually more than 18 (17) ; ratio of tail vertebras to total length about 36. E. h. rubidzcs Hind foot usually less than iS (17); ratio of tail vertebrsD to total length about 31. E. h. brittaniciis Subgenus CRASEOMYS Miller. EVOTOMYS RUFOCANUS (Sundevall). 1846. HypiidcEus ricfocanus Sundevall, Oefversigt af Kongl. Vetenskaps. Akademiens Forhandlingar, Arg. 3, p. 122, May 13, 1846. 1897. EvotoJiiys rufocamis'Qwi.^Y, Proc. Biolog. Soc. Washington, xi, p. 122. May 13, 1897. Type locality. — Lappmark, Sweden (specimens mentioned from Altawaara, Karesuando, Lule, and Pite). Faimal positio7t. — Upper and middle boreal zones. Geographic distribution. — Northern Europe, south in the mountains of Norway to Dovre.^ General characters. — Size large (hind foot, 20 (iS) or more; greatest length of skull, 26) ; fur long and dense ; dorsal stripe narrow- bright, and sharply contrasted with clear gray of sides. Color. — A specimen in fresh autvunnal pelage has a well defined dorsal stripe of a color intermediate between the hazel and cinnamon rufous of Ridgvvay,^ evenly but not conspicuously sprinkled with black-tipped hairs. This stripe begins between the eyes and extends 1 Collett, Nyt Mag. for Naturvidenskaberne, XXXVI, p. 2S0, March, 1S98. 2 Nomenclature of Colors, pi. IV, Nos. 12 and 16. 90 MILLER nearly to base o£ tail. Ears colored like dorsal stripe. Sides an indescribable grizzle of hair brown, whitish, black, and slate color, the general effect clear light gray. Whole underparts soiled buffy white, darkened by the slaty bases of the hairs, which show through irregularly on the surface, especially when the fur is disarranged. Cheeks, muzzle, and sides of head similar to sides of body, but slightly darker. Tail sharply bicolor, brownish above, dirty white below. Feet dirty white. In a half-grown specimen from Finmark, Norway, loaned by Mr. Robert Collett, the dorsal stripe is bister scarcely tinged with red except in the region between the ears. Here it is dull russet. Other- wise the coloring is essentially as in the adult. Sktill. — The skull of Evoto7nys 7-ufocanus has the general appear- ance of that of a medium sized Microtus with teeth slightly weaker than usual. Braincase relatively deeper than in Microtus agrestis from Upsala, Sweden, slightly longer than broad, squarely truncate posteriorly, its outline carried squarely forward by the large postor- bital processes. It is strongly marked by ridges for muscular attach- ment. Interorbital region narrow, with a deep longitudinal median furrow. Rostral protuberances ^ smaller than in the species of true Evoto7nys and placed close to root of zygomatic process of maxillary. Rostrum short, deep, and heavy. Palate and ventral aspect of skull much as in Evotomys iiorvegicus except for the relatively larger* teeth and audital bullae, the latter even larger than in Microtis agrestis. Measurements of an adult skull from Tanen, Postanger, Finmarken, Norway: Greatest length 26; basal length 26; basilar length 25; zygomatic breadth 15.4; mastoid breadth 13.6; interorbital con- striction 4; diastema S ; palatal length 13 ; occipital depth (in median line) 7.6; frontopalatal depth (at middle of molar series) 7S ; mandi- ble 17; maxillary toothrow (alveoli) 6.S ; mandibular toothrow (alveoli) 6. Teeth. — The teeth of Evotomys riifocamis as compared with those of other members of the genus * are chiefly remarkable for their large size, gi"eat strength, and sharply pointed salient angles. In robustness the teeth perhaps slightly surpass those of an adult Microtus agrestis from Upsala, Sweden, though this appearance may be due to the broader dentine spaces. It is especially noticeable in the mandilnilar 'See Proc. Acad. Nat. Sci., Philadelphia, 1S9S, p. 361. 2 Amoug the Europeau nicuibers of the subgenus Evoioviys^ I can find no tangible differential characters in the teeth. In the larger species the teeth are slightly larger than in the smaller ones ; but the proportions and the pattern of enamel folding remain the same. EUROPEAN REDBACKED MICE 9I teeth and in the second and thh-d upper mohirs. The triangles in the upper molars are closed throughout; those of the lower molars are occasionally and apparently quite irregularly open. In arrangement the triangles are essentially as in Evotomys norvegiciis. In the first lower molar, however, the anterior loop is longer, and deeply cut by a well developed reentrant angle on the inner side (scarcely indicated in E. tiorvegicus) which isolates a third closed triangle fully equal- ling the others of the same side in size. In the posterior upper molar the two external reentrant angles are approximately equal in depth, while in E. iiorvcgiciis the more anterior of the two is exceedingly shallow and often obsolete. Measiu'cments. — A well made skin from Qiiickjock, Sweden, measures: total length 150; tail vertebrae 40; hind foot with claws 19. 8; hind foot without claws 18. Specimens examined. — Five, from the following localities : Nor- way; Stabursnaes, Porsanger, i, Tanen, Finmarken, 2. Sweden; Qinckjock, i, Tornea, i. Rctnarks. — Evotomys riifocanus is so readily distinguishable by its subgeneric characters that it needs no special comparison with other members of the genus. Subgenus EVOTOMYS Coues. EVOTOMYS RUTILUS (Pallas). 1778. Mus ruiihis Pallas, Nov. Spec. Ouadr. e Glir Ord, p. 246. 1874. Evotomys rutilus QoMYS, Proc. Acad. Nat. Sci., Philadelphia, p. 187. 1899. A\jiaptogojiia\ rutila Cope, Journ. Acad. Nat. Sci., Philadelphia, 2d ser., XI, p. 201. Type locality. — Siberia, east of the Obi. Faunal position. — Evotomys rutilus is strictly confined to the up- per boreal zone. Geographic distribution. — Arctic Europe and Asia, south in Nor- way to Tromso.^ General characters. — Size small (total length about 125, hind foot 1 8, greatest length of skull 24) ; tail forming about 20 percent of total length, very densely haired, its pencil at least ^ as long as verte- brae ; dorsal stripe clear bright chestnut, sharply but not conspicuously contrasted with color of sides. Color. — An adult male from northern Sweden taken in December has the dorsal stripe clear bright chestnut, faintly darkened by a slight admixture of black-tipped hairs. Sides ochraceous buff. iCollett, Nyt Mag. for Naturvidenskaberne, XXXVI, p. 282. INIarch, 1898. 92 MILLER Belly dirty white or cream color, darkened by the plumbeous bases of the hairs which show through irregularly at the surface. Feet whit- ish. Ears like dorsal stripe. Tail sharply bicolor, brownish tinged with red above, dirty white below. Fur everywhere slaty plumbeous at base. Individual variation among adult skins is too inappreciable to need special comment. I have seen no immature specimens. Skull. — The skull of Evotomys riitilus is small, smooth, and rounded, in general appearance hardly distinguishable from that of E. hercyniciis. The zygomata are rather abruptly though not very widely flaring anteriorly, so that in the malar region the two arches are parallel with each other and with the main axis of the skull. Rostral protuberances small but well developed and standing out con- spicuously in front of zygomatic processes of maxillaries. Lateral outline of nasals nearly straight from base to tip. Posteriorly the nasals extend slightly beyond nasal branches of premaxillaries. They are squarely truncate. Fostorbital processes essentially obsolete. Braincase low and flat (the skull distinguishable by this character alone from that of any of the members of the hercynicus group), subrectangular in outline when viewed from above. Palate normal except that the foramina immediately in front of lateral bridges are so large and numerous that the bridges are generally somewhat en- croached upon and rendered imperfect. Audital bullae as in E. her- cynicics^ therefore slightly smaller than in Microtus agrestis and much smaller than in Evotomys rtifocaniis . Measurements of an adult skull from Lapland : Greatest length 24; basal length 32.4; basilar length 20.6; zygomatic breadth 13; mastoid breadth 11.8; interorbital constriction 4; nasals 7 ; incisive foramen 5; diastema 7.4; palatal length 10. S; occipital depth (in median line) 7; frontopalatal depth (at middle of molar series) 6; mandible 13. 8; maxillary toothrow (alveoli) 5; mandibular toothrow (alveoli), 5. Teeth. — The teeth of Evotoj)7ys riitilus present no characters of special importance. Relatively to the size of the skull they are pro- portioned as in E. hercynictis. I can see no tangible specific charac- ters in the enamel pattern. Measurements. — Two specimens from northern Sweden measure respectively : Tail vertebras 34 and 39 ; pencil 1 1 and 9 ; hind foot with claws iS and 1S.4; hind foot without claws 17. Specimens cxaynined. — Five, from the following localities : Norway : South Varanger, Finmark, 2. Sweden: Karesuando, i ; Lapland, 2. Rejuarks. — This arctic species is instantly recognizable by its small EUROPEAN REDBACKED MICE 93 size, bright colors, and very short densely haired tail. It is in no way closely related to any of the other European members of the genus. EVOTOMYS NORVEGICUS sp. nov. 1898. Microtus is^larcolus Collett, Nyt Mag. for Naturvidenskaberne, XXXVI, p. 278. March, 1898. (J>ioi Mus glareohts oi Schreh&r .) Type. — Adult 9 (skin and skull) no. S4674 United States National Museum, collected at Bergen, Norway, Alay 31, 1S9S, by Thora Steineger. Original number 20. Faiinal position. — This species is probably characteristic of the middle and lower boreal zones. GeograpJiic distribution. — The forested portions of Norway, north to the Saltdal (67° N.).^ General characters. — Size large, skull of adults about 36 mm. in greatest length, total length of adults 150-170, hind foot 20 or more; ratio of tail vertebra; to total tength 33-36; skull narrow and notice- ably ridged for muscular attachment, the zygomata heavy and abruptly flaring; color rather dull. Color. — Slimmer pelage : Dorsal stripe fairly well defined, extend- ing from forehead to rump. It is dark, dull ferruginous, slightly varied with light wood-brown and much darkened by a uniform sprinkling of black tipped hairs. Face, cheeks and sides light wood- brown fading to broccoli-brown on lower part of sides, especially pos- teriorly. Rump wood-brown tinged with red in median line and form- ing no noticeable contrast with color of back. Whole ventral surface pale drab-gray, irregularly tinged with yellowish along median line, the plumbeous bases of the hairs appearing irregularly at surface. Feet dull white. A dark shade at inner side of ankle. Sides of muz- zle slightly yellower than face. Ears nearly naked externally, covered with very short i-eddish hairs internally. Tail sharply bicolor, dark brown above, whitish below. Skull. — The skull of Evotoniys norvegicus is large, and, for a true Evotomys, rather strongly ridged and angled. In this respect it sur- passes all other European members of the genus, S. ( Craseomys") rufocanus excepted. The zygomata are heavily built, and so abruptly flaring anteriorly that the greatest zygomatic breadth is near anterior ex- tremity of molar ; back of this region the arches distinctly converge. Rostral protuberances well developed, placed close to base of zygo- mata. Postorbital processes distinct but very small. Braincase moder- ately high and rounded, distinctly rectangular in outline when viewed 1 Collett, Nyt Mag. for Naturvidenskaberne, XXXVI, p. 278. 94 MILLER from above. Lateral outline of nasals bowed inward near middle. The nasals are squarely truncate posteriori}^ a little in front of tips of nasal branches of premaxillae. Palate normal, the lateral bridges well de- veloped, and the foramina in front of bridges small. Audital bullae as in E. Jiercynicus. Measurements of an adult skull (the type) : Greatest length 36. 8 ; basal length 25; basilar length 23; zygomatic breadth 14.6; mastoid breadth 1 1. 6; interorbital constriction 4 ; nasals 8; incisive foramen 5.2; diastema 7.8; palatal length, 11. 8; occipital depth (in median line) 7.4; frontopalatal depth (at middle of molar series) 6.8; man- dible 16; maxillary toothrow (alveoli) 5.8; mandibular toothrow (alveoli) 5.4. Teeth. — The teeth of Evotoinys 7iorvegiciis present no tangible characters by which they may be distinguished from those of E. rzitilus and E. hercynicus. They are, of course, larger in propor- tion with the greater size of the skull. Aleasurements. — Average of nineteen specimens from the neighbor- hood of the type locality: total length 163; tail vertebras 56.3; hind foot with claws 20. 8. For details see table, page 106. Specimens examined. — Twenty-one from the following localities: Norway: Asker, i ; Bergen, 4; Granvin, 14, Opheim, 2. Rejuarks. — Evotomys norvegicus needs no close comparison with E. rutilus of the upper boreal zone or E. Jiercy^iicus of the transi- tion zone and upper austral zone. From both of these it is separated by its much larger size ; while from E. rutilus it is immediately distinguishable by its long, thinly haired tail. The only European species with which it is likely to be confused are the geographically remote E. 7iageri of the Alps and E. vasconice at present known from the Fyrenean foothills only. Both of these animals agree with E. norvegicus in size, but the former is readily separable from it by the long tail, pale sides, and broad smooth skull, while the latter differs in its shorter, less hairy tail, duller color, and heavier molar teeth. This species is the '• glareolus^ of authors dealing with the fauna of Norway exclusively. A form of E. hercynicus is, however, included under the name glareolus by those who refer to the Scandinavian fauna as a whole. EVOTOMYS NAGERI (Schinz). 1845. Hypudcpus iiagcri Schinz, Synopsis Mammahim, 11, p. 237. (Mt. St. Gotthard, Switzerland.) 1862. Myodes bicolor Fatio, Revue et Magasin de Zoologie, 2d ser., xiv, p. -57- J^ly- 1862. (Genthal, Berne, Switzerland.) EUROPEAN REDBACKED MICE 95 1867. HypudcFiis glarcolus Fatio, Les Campagnols du Bassin du Leman, p. 28 (part). Type locality. — Mount St. Gotthaid, Switzerland. JPaunal position. — This species is probably characteristic of tlie middle and lower boreal zones; at lower elevations it is replaced by E. hercyniciis hclveticjis. Geographic distribtition. — Evergreen forests of the Alps, except in the lower valleys; exact limits of range not known. GcJieral characters. — Size large, skull of adults about 26 mm. in greatest length, total length of adults 150-170, hind focTt 20 (18) or more; ratio of tail vertebrae to total length 35-3S ; skull broad and little ridged for muscular attachment, the zygomata weak and gradu- ally flaring; colors dull. Color. — Sjnnmer pelage : Dorsal stripe well defined, extending from forehead to rump. It is cinnamon rufous slightly varied with pale broccoli-brown and inconspicuously darkened by a sprinkling of black tipped hairs; the resulting color very nearly chestnut. Face, cheeks, and sides light hair-brown fading to smoky gray on lower part of sides. Rump light broccoli-brown tinged with red in median line and forming a noticeable though not very strong contrast with color of back. Whole ventral surface pale smoke-gray, faintly tinged with yellowish along median line, the plumbeous bases of the hairs appear- ing irregularly at surface. Feet dull white. A dark shade at inner side of ankle. Sides of muzzle pale Isabella color. Ears thinly haired, dull reddish. Tail bicolor, dark brown above, whitish below. Skull. — The skull of Evotomys nageri., while of the same size as that of E. norvegicus.f is readily distinguishable by its relatively broader braincase, more slender rostrum, shorter, broader interorbital region, and lighter, less abruptly flaring zygomata. The latter flare so gradu- ally that the greatest zygomatic breadth is at the posterior extremity of the jugals, from which point the arches converge forward. Rostral protuberances well developed, placed slightly further forward than in E. norvegictis. Postorbital processes nearly obsolete, though slightly more apparent than in E. rutihis. Nasals i^ounded posteriorly, ter- minating in line with nasal branches of premaxillie. The dorsal pro- file of the nasals is more rounded than in E. norvegictcs., and the con- striction near middle is less abrupt. Braincase broad and so rounded and slightly angular that the usual rectangular outline is much ob- scured. Audital bulUe more globular than in E. norvegicus., the dif- ference, however, appreciable on comparison only. Measurements of an adult skull (^) from Briinig, Switzerland: 96 MILLER Greatest length 26.4 ; basal length 24; basilar length 22 ; zygomatic breadth 14.6; mastoid breadth 12. S; interorbital constriction 4.4; nasals 8; incisive foramen 4.8; diastema 7; palatal length 12; occipi- tal depth (in median line) 7.2; frontopalatal depth (at middle of molar series) 6.8; mandible 15; maxillary toothrow (alveoli) 6; mandibular toothrow (alveoli) 6. Teeth. — Except that the incisors are slightly less robust the teeth of Evotomys nageri do not differ from those of E. Jiorvegicics. Measurements. — Average of thirty-six specimens from the neigh- borhood of the type locality: total length 156; tail vertebrtE 56.9; hind foot with claws 20.3 ; hind foot without claws 19.2. For details see table, page 105. Speci??iens exammed. — Eighty-two, from the following localities: Switzerland: Andermatt, 41 ; Briinig, 30; Meiringen, 11. Remarks. — Evotomys 7iageri is a much larger animal than its near geographical ally, E. hercyniciis helvetictis. From the large E. vascojii(t of the Pyrenees it is distinguishable by its relatively inuch longer tail, paler sides, heavier teeth and broader, less angular skull. For comparison with E. norvegicus see description of latter. EVOTOMYS VASCONI^ sp. nov. • Type. — Adult $ (skin and skull) no. S6994 United States National Museum, collected at Montrejeau, Hautes-Pyrenees, France, July 4, 1899, by Robert T. Young. Faunal position. — Evotoniys vasconicc is probably a boreal Pyre- nean species whose range extends downward in the cool woodlands to the foothills. Geographic distribution. — Although nothing is at present known of its geographic distribution this animal probably occurs throughout the isolated boreal region of Spain and southern France. General characters. — Closely related to Evotomys norvegicus., but with shorter, more thinly haired tail, duller colors, broader nasals, and heavier teeth. Color. — Summer pdage: Essentially as in E. Jiorvegicus., but dorsal stripe less well defined and less red — more nearly cinnamon- rufous than ferruginous. Sides and flanks slightly darker than in E. 7iorvegicus. Winter pelage not known. Skull. — The skull of Evotomys vasconicc is of essentially the same form and size as that of E. norvegicus. The zygomata are heavily built and abruptly flaring, therefore quite different from those of E. nageri. Nasals short and broad, the greatest width of both together EUROPEAN REDBACKED MICE 97 about one lialf length. Audital bulla) shorter and broader than in E. norvegicus. Measurements of an adult skull (the type) : Greatest length 25. S; basal length 23.6; basilar length 21.8; zygomatic breadth 14; mas- toid breadth 12: interorbital constriction 4.2; nasals 7.4; incisive foramen 5; diastema 7; palatal length 11.4; occipital depth (in median line) 7 ; f ronto-palatal depth (at middle of molar series) ().(>\ mandible 15.4; maxillary toothrow (alveoli) 6; mandibular toothrow (alveoli) 5.6. Teeth. — The teeth are more heavily built than in E. noi-veglctis., the difference most noticeable in the second and third molars. Meas7iremetits. — Average of tw^o specimens from the type locality : total length 160; tail vertebrae 52; hind foot w^ith clavv^s20.2; hind foot without claw^s 18.5. For details see table, page 106. Specimens examhzed. — Tw^o, both from the type locality. Remarks. — The discovery of Evotomys vasconice completes an illus- tration of the distribution of a boreal group of mammals in western Europe. Three isolated but closely related species of large European redbacked mice are now known, one each in the boreal zone of Nor- way, the Alps, and the Pyrenees.^ The small E. hercynicus occupies the lower zones throughout the region Intermediate between the ranges of E. fwrvegicus and the two other large species. That these are similarly Isolated from each other Is scarcely to be doubted. Evo- to7nys hercynicus has been traced southward to the lower Garonne, less than 150 miles from the type locality of E. vasconice, while there Is no apparent barrier to Its occurrence throughout the lowlands from the Department of Gironde to the Gulf of Lyons, thus effectually separating E. vasconice from E. nageri. EVOTOMYS HERCYNICUS (Mehlis). (Synonymy under subspecies.) Type locality. — Harz Mountains, Germany. Faunal position. — Evotomys hercynicus Is probably a character- istic species of the transition and upper austral zones. Geographic distribution. — Non-mountainous portions of Europe, from southern France and the valleys of northern Switzerland, north to ' Montrejeau, the type locality of E. vasconicz:, is probably in the Transition zone. Evotomys, however, can readily descend to this level at the base of a mountain chain by taking advantage of local conditions (see Miller, Science, NS., VIII, pp. 615-618. November 4, 1S9S.) Mr. Young's specimens were taken in heavy woods. pS MILLER Scotland, northern Sweden (Quickjock) and southern Finland (Tam- merfors). Limits of range not known. General characters. — Size small, skull of adult about 24 mm. in greatest leqgth, total length of adult animal 140-150 mm., hind foot 20 (18) or less; ratio of tail vertebrae to total length generally between 30 and 35. Color. — See descriptions of subspecies. Skull. — The skull of Evotomys hercynicus is readily distinguish- able from that of other European members of the subgenus Evotomys. From skulls of E. norvegicus., E. nageri and E. vasconice its small size immediately separates it; while from E. rutihis it differs in higher, more rounded braincase and less prominent^ rostral protuber- ances. On the whole the skull perhaps most closely resembles that of E. nageri., from which it is separated chiefly by smaller size and by the parallel zygomata. From the skull of E. norvegicus the differently shaped zygomatic arches will always sei*ve to distinguish it. Among themselves the skulls of the different races show no diagnostic char- acters, except in the case of E. hercynicus helvetictis. Teeth. — Except in their small size the teeth of Evoio7?iys hercynicus present no peculiarities. Measuretnents. — See descriptions of subspecies. Specimeits exanii7tecl,. — One hundred and ninety-one. For details see lists under subspecies. Remarks. — Evoto7nys hercy7iicus is separable from E. rutihis by its long, thinly haired tail and deeper, narrower skull, and from E. norvegicus ^ E. vasco77icc., and E. 7tageri by its much smaller size. It is a widely distributed, j^lastic species, its extensive, nearly im- broken range and great variability as compared with the other red- backed mice of Europe dependent on the physiographic conformation of the life zones which it inhabits. Five races are represented by the material thus far brought toirether. EVOTOMYS HERCYNICUS HELVETICUS subsp. nov. 1867. HypudcEics glarcoIusYxTio, Les Campagnols dii Bassin du Ldman, p. 28 (part). Type. — Adult male (skin and skull), in British Museum, col- lected at Montauban, Haute Savoie, France (near Geneva, Swit- zerland) at altitude of 900 m., November 8, 1S99, by Alphonse Robert. Original number, 22. ■ That is, less far removed from base of zygomata. EUROPEAN REDBACKED MICE 99 Geographic distribution. — Lowlands between the Alps and Jura. Ge?icral characters. — Skull broader and audital bullae larger than in any of the other known races of Evotomys hercynicus. Size rather large (hind foot 19 (i8), total length 145) ; tail forming 31-33 percent of total length. Winter pelage : dorsal red area diffuse, hazel, micch inter7nixed -with gray; belly slightly washed with buff; tail nearly black above. Color. — Summer pelage not known. Winter pelage : red dorsal area very diffuse and ill defined, extending from eyes to base of tail, but not widely spreading over sides. In color it is hazel, slightly darkened by black tipped hairs, and intermixed with broccoli-brown or ecru drab (the subterminal color bands) and grayish white. The brown and gray is'always noticeable and occasionally almost in excess of the red. Sides and cheeks broccoli-brown or ecru drab, incon- spicuously sprinkled with black. Face slightly darker than cheeks. Belly whitish gray washed with cream buff and darkened by the slaty underf ur. The buff wash is variable. It is nearly absent in two specimens among eighteen, and in a few it is considerably more heavy than usual, spreading noticeably over cheeks, face and sides. Tail very conspicuously bicolor, blackish above, buffy white beneath. Feet white. Ears like dorsal stripe. A tuft of whitish hairs springs from near posterior base of conch. This tuft forms a distinct spot in some specimens. Skull. — The skull of Evotomys hercynicus helveticus is slightly larger than that of the other races, the braincase is less elongated and the audital bulljE are larger. The differences are not easily described, but on comparison of series of skulls they at once become apparent. Measurements of skull of type : greatest length 24 ; basal length 22.2; basilar length 21; zygomatic breadth 13.6; mastoid breadth 1 1.8; greatest breadth of braincase above roots of zygomata 11.4; in- terorbital constriction 4 ; nasals 7 ; incisive foramen 5 ; diastema 6.B> ; palatal length 1 1 ; occipital depth (at front of basioccipital) 7 ; fronto- palatal depth (at middle of molar series) 6.2; mandible 15; maxil- lary toothrow (alveoli) 6; mandibular toothrow 5.6. Measurements. — Average of sixteen topotypes : total length 13S; tail vertebrae 44.6; hind foot with claws 19.2; hind foot without claws 17.9. For details see table, page 107. Specimens examined. — Twenty, from the following localities : France, Montauban (near Geneva, Switzerland) 17; Switzerland, Vand (near Lausanne) i (skull) ; St. Gallen, 2. Remarks. — This animal is so well differentiated from true Evotomys Proc. Wash. Acad. Sci., July, 1900. lOO • MILLER hercynicus in both cranial characters and color that I hesitate to give it a trinomial name. So far as known it is confined to the low country lying between the Alps and the Jura, where it is probably in- sulated. If intergradation with E. hercynicus hercynicus does occur it is to be looked for in the region about St. Gallen or in extreme southwestern Bavaria. The two specimens from St. Gallen though closely agreeing with the type in color show an approach to the usual narrow form of skull. At Mai'xheim, Bavaria, about 50 miles northeast of St. Gallen, E. h. hercynicus occurs in its most extreme phase. EVOTOMYS HERCYNICUS HERCYNICUS (Mehlis). 1831. HyfiudcBus hercynicus yi'EHiAS, Isis, 1831, p. 876., 1857. Arvicola glareolus Blasius, Fauna der Wirbelthiere Deutschlands, i, Saugethiere, p. 337 (part). Not Mies glareolus Schreber. Type locality. — Harz Mountains, Germany. Geographic distribution. — Dry interior region of Germany and western Russia. Limits of range not known. General characters. — Most brightly colored of the races of Evo- tomys hercynicus. Size rather large (hind foot 19 (iS), total length 140) ; tail forming 30-33 percent of total length. Red dorsal area in both pelages well defined and narrow, rufous, not intermixed with gray. Belly very slightly washed with buff. Sides and ruinp in summer pelage stro7tgly contrasted ivith back. Tail always brown above. Color. — Summer pelage: dorsal stripe narrow and well defined, not tending to spread over sides. It is rufous slightly varied with yel- lowish wood-brown, and rather thickly sprinkled with black-tipped hairs. Face, cheeks and sides pale yellowish wood-brown tinged with gray and fading to grayish white on belly. Rump like sides, strongly contrasted with dorsal stripe. Feet grayish white. A dark shade at ankle. Sides of muzzle essentially like cheeks. Ears thinly haired, like dorsal stripe in color. Tail sharply bicolor, brown above (darker at tip), soiled white below. Winter pelage: dorsal stripe slightly less sharply defined than in summer, the rufous paler and warmer, consid- erably varied with yellowish wood-brown, but very inconspicuously sprinkled with Ijlack-tippcd hairs. Face, cheeks, and sides more yel- lowish wood-brown than in summer, and scarcely tinged with gray. Rump slightly suffused with color of dorsal area, and tlierefore less contrasted with back than in summer. Feet pure white. Skull. — Measurements of an adult ( Miller collection. 2 Type. 3 Type. PROCEEDINGS OF THE WASHINGTON ACADEMY OF SCIENCES Vol. II, pp. 111-133. August 20, 1900. PAPERS FROiAI THE HARRIMAN ALASKA EXPEDITION. II. HARRIMANIA MACULOSA, A NEW GENUS AND SPECIES OF ENTEROPNEUSTA FROM ALASKA, WITH SPECIAL REGARD TO THE CHAR- ACTER OF ITS NOTOCHORD. By Wm. E. Ritter. . The preparation of my report on the Enteropneusta of the Pacific Coast of Nortli America, which is to form a part of the scientific results of the Harriman Alaska Expedition, has, owing to the unexpected richness of the material, proved a consider- ably greater task than I had anticipated. Since it has been im- possible to complete the work at as early a date as was orig- inally hoped, and since one of the five or six new species is of more than ordinary interest, it seems desirable that a brief ac- count of it should be published without delay. This species is the type of a new and very distinct genus v.hich I take great pleasure in dedicating to Mr. E. H. Harri- man. I do this not merely in acknowledgment of the fact that he was the patron of the expedition, but also in recognition of the way his own personal management contributed to its scientific success. In its superficial characters Harriniania maculosa resembles Balanoglossus kupj'eri v. Willemoes-Suhm and B. canadensis Spengel more closely than it does any other known species. It may be known by the diagnosis following. Proc. Wash. Acad. Sci., August, 1900. (m) 112 RITTER HARRIMANIA gen. nov. EXTERNAL CHARACTERS. Animal short in proportion to its thickness, not only as a whole, but in each of its three regions. Proboscis somewhat depressed, conical, but little longer than broad ; in preser\^ed specimens almost always contracted so as to become considerably shorter than broad. Collar short, always broader than long, usually in preserved specimens greatly so. Pharyngeo-genital region without genital pleuras ; not clearly set off from the abdominal region ; frequently much depressed, so as to be considerably broader than deep. Gonads beginning im. mediately behind the collar and extending either not at all or only a short way behind the pharynx, in the form of rather large, irregular lenticular masses. An inter-, as well as an extra-branchial series on each side. Maximum number of branchiEe about 40 pairs, these opening wholly on the dorsal aspect of the animal, in a well-defined ditch on each side ; both walls of each ditch composed chiefly of the extra- and inter-branchial gonad series, respectively (PI. VII, fig. 2). Length of pharynx about double that of proboscis. No external liver lobes ; no constriction between pharynx and abdomen. Abdomen nearly round in transverse section, relatively short, and tapering rather abruptly to its posterior extremity. The following measurements in millimeters are from an average adult living specimen : Length of pro- boscis 13 ; length of collar 4; length of pharynx 25 ; length of gonad series 27 ; length of abdomen 89 ; total, 131. Color. — General shade dark. Proboscis dark gray, mottled, par- ticularly on dorsal side, with many somewhat elongated, almost coal- black spots. Collar also dark and mottled, though less strikingly so than the proboscis. Gonads a yellowish white, but all the remaining portions of the pharyngeo-genital region dark, the dorsal side con- siderably more so than the ventral, the latter yellowish green; an- terior portion of abdomen dark grayish green ; posterior portion similar to it, but somewhat lighter. On the whole this is probably the darkest species of Enteropneusta known. ANATOMICAL CHARACTERS. Proboscis. — Circular muscle layer about equal in thickness to the nervous layer of the ectoderm. Longitudinal muscle fibers arranged in radiating plates as in Ptychodera. No central proboscis cavity, but in place of this a core of considerable size of very fine-fibered close- meshed connective tissue containing many small nuclei. In place of HARRIMANIA MACULOSA II3 the ventral proboscis septum, a plate of the connective tissue core wedges itself between two radial muscle plates, and reaches nearly to the ectoderm. Into this plate there extends to a point considerably beyond the anterior extremity of the proboscis complex a small but very distinct tube with clear-cut lumen and wall of well-defined cubical epitheloid cells. This tube is connected with the globus on its ventral side and toward the anterior end, and is probably a blood vessel, though the character of its wall as indicated above, contrasts sharply with the usual enteropneustic blood vessels. This structure is not con- nected with the notochord, hence is not homologous with the ' ver- miform process ' of Schizocardium. Glomus well developed, though the two halves not confluent excepting slightly antero-ventrally. The facts relative to the proboscis-coelom and its pores are as fol- lows : I . As stated above, unpaired central cavity wanting, this being replaced by the connective-tissue core. 2. The co^lom confined to the extreme base of the proboscis and consisting of two wholly separate lateral halves of equal size. 3. Posteriorly, each of these cavities bifurcate, the two legs or pouches of each being placed, the one dorsal to the other. 4. The ventral pouch on each side ends blindly, while the dorsal ones open to the outside, each through its own pore, so ^?i\. two proboscis pores are always present. 5. The proboscis pores open very close in front of, sometimes actually into, the anterior neuropore. Collar. — Dorsal nerve cord sometimes with and sometimes with- out a distinct anterior neuropore ; throughout its length numerous small, entirely closed cavities within its substance. ' Giant cells ' present. No dorsal crest or dorsal ' nerve roots.' Periha^mal diver- ticula not reaching forward quite to the anterior neuropore ; the septum between them complete to their extreme anterior end, con- taining only longitudinal muscle fibers. Dorsal blood vessels not situated within this septum, but a sinus bounded by the nerve cord dorsal ly, and by the walls of the periheemal diverticula ventro- laterally. Collar coelom extending far into the neck. Peripharyn- geal spaces apparently absent. Crura of nuchal skeleton reaching into the pharyngeal region aitd overlapping two or three pairs of branchial bars. Body of the skeleton also extending usually far back, reaching nearly to the middle of the collar. The skele- ton moderately strong, but the chondroid tissue not abundantly de- veloped. Notochord consisting of two distinct parts: the anterior pouch-shaped part corresponding to that found in all Enteropneusta, and 2i posterior gutter-shaped part directly continuous with the Jirst 114 RITTER vientioned part, and coextensive "with the esophagus, with which it is conitected. Collar funnels of the usual form, though very short; directed outward at nearly a right angle to the long axis of the body, instead of forward ; opening into the first gill pouch. (PL VII, figs. 3, 4, and 7.) Pharyngeo-genital region. — Separation of the eiiteric wall into pharyngeal and esophageal portions by deep longitudinal folds on each side. At the anterior end the pharynx occupies but little more than the dorsal half of the wall, while posteriorly the esophageal por- tion becomes much reduced. No synapticulae in branchial bars. Epibranchial ridge very prominent. Male gonads both extra- and in- ter-branchial, the extra-branchial far more voluminous, in the form of large, more or less globular lobes, the extra-branchial lobes much larger than the inter-branchial ones. Female gonads apparently extra- branchial only. Gonads retained in the body cavity, and not raised up into genital pleurse to overarch the branchial orifices. Abdo7ninal region. — No external liver lobes, but a single pair of broad laminar appendages attached to the dorsal side of the intestine immediately behind the pharynx, and for some distance in this region the intestinal wall much thickened, and the lumen correspondingly reduced in size as compared with the portions farther back. No in- testino-cutaneous pores. Ventral longitudinal muscle layer broad and but little thicker than the same layer in the dorsal and lateral regions. No circular muscle layer in this portion of the body. The trunk coelom narrows off on each side anteriorly and terminates immediately behind the posterior limit of the collar. The structure- less membrane lining the cavity becomes thickened and dense in this narrowed anterior end, and at the extreme tip there is a patch of this membrane on which is situated some peculiar epitheloid cells. These are high and slender, sometimes tailed, and ai'e set on both sides of the membrane, perpendicular to it, and close together. In some individuals this patch is thrust forward and outward in its middle so as to produce a definite though short and narrow pouch, into which the epitheloid cells of the inner surface project so as to divide the cavity of the pouch into narrow spaces. This peculiar structure seems to be 2:)resent in every specimen, but varies considerably in form and definiteness. It is usually in close re- lation to the collar funnels, though there seems to be no connection between them. I am unable to find anything comparable to them in any other species. The most plausible suggestion that I can make about them at present is that they may represent the peripharyngeal HARRIMANIA MACULOSA II5 spaces of other species. But to this suggestion T do not attach much importance. GENERAL REMARKS. From this brief diagnosis there is no doubt that Harrtmania mactdosa belongs to the family Balanoglossidae, if we accept Willey's proposal to separate the Enteropneusta into families, as I am disposed to do. Of the two genera hitherto known into which the species of this family naturally group themselves, Balanoglossus and Dolichoglosstis^ Harrimania is unquestion- ably closer of kin to Balanoglossus. From this, however, it is clearly set off by the -persistence of the esophageal notochord in the adult animal^ the invasion of the pharyngeal region by the skeleton crura, and probably by the peculiar structures at the anterior ends of the trunk coelom. This species was found at Prince William Sound and Kadiak, Alaska. At Prince William Sound it was collected at two points, Orca and Port Valdes. All the specimens from 'the latter place were taken by Dr. Wesley R. Coe, to whom I am indebted for well-preserved specimens, as I did not myself do any col- lecting at that point. It is an interesting and significant fact that this ani^nalis not a burroiver in the sense in which most Enteropneusta are. It lives under stones, where it often makes its way through the mud at the plane of contact between the stone and the earth. So far as my observations have gone it rarely plunges directly into the ground, as do other species. In collecting, one turns over the rocks and picks up the animals as he does various holo- thurians, etc., and does not need to dig in the mud for them. At Prince William Sound, where we found it in by far the greater abundance, it was collected at extreme low tide only. At Kadiak, however, the few specimens secured came from nearer high water mark than I have ever before seen Enterop- neusta living. These two facts relative to the animal's life ' Spengel suggested these terms as subgeneric designations for the short and long proboscised species, respectively, of the genus Balattoglossus as restricted by him. The knowledge obtained since Spengel wrote, added to the considera- tions which led him to suggest such a subdivision of the genus, convinces me of the wisdom of recognizing the two groups as distinct genera. Il6 RITTER habits, particularly its non-burrowing habit, may, it seems to me, be correlated with its primitiveness among Enteropneusta. The characteristic Balanoglossus odor is stronger and more persistent in this species than in any other with which I am ac- quainted. My alcoholic specimens still retain it (now nearly a year since they were taken) so strongly that in handling them it clings to one's hands for some time in spite of vigorous washing. I may mention here that between Dr. Coe and myself more than one hundred specimens were collected, so that I have had ample material to work with. I have sectioned about twenty individuals in various ways, so that every anatomical point has been confirmed by a goodly number of examinations, and all tell essentially the same story so far as the important facts are concerned. Perhaps in view of the great variation which Spengel, Willey, Hill, and others have found in the number and character of the proboscis pores in some other species, I ought to expressly state that I have given particular attention to this point in Harrimania maculosa, and in not a single instance have I failed to find both present and equally well developed. SPECIAL REMARKS CONCERNING THE NOTOCHORD. Although in the present communication I can neither describe in detail the structure of the animal nor discuss in full the theo- retical conclusions to which I am led by the study of this and the other species in my possession, a brief consideration of a few points beyond what has already been suggested must be given. And first, because most important, something more about the notochord — and at the outset a word concerning the use of the term as applied to the organ. It is well known that Bateson (1886) was the first to contend for the homology of the organ with the notochord of vertebrates ; and hence that it was he who first designated it by this name. Spengel (1893), as is also now well known to all zoologists who have interested them- selves in the problem of chordate phylogeny, takes strong grounds against Bateson's view, and consequently refuses to use the name notochord, but adopts the term ' Eicheldarm.' And recently Willey (1S99), whose utterances on all matters pertain- HARRIMANIA MACULOSA II7 ing to the protochordata are deserving of high regard, has ex- pressed the view that "What has been called the notochord in the Enteropneusta does not correspond with this definition [of it as used in the Chordata] except in its capacity of skeletal product of the gut wall." And he proposes for it the term ' stomochord,' though declaring at the same time that " this term involves no sacrifice of conviction whatever, since it leaves the question of the morphological relationships of the structure to which it refers quite open." From the evidence now at hand it appears to me there can be no serious doubt that we have in this organ the immediate genetic forerunner of the vertebrate noto- chord, and that it does not differ sufficiently cither in structure or relations from the essential nature of the vertebrate organ to zvarrant giving to it as a whole a different designation, I shall, consequently, follow Bateson and call it the notochord. The anterior pouch-like portion of the organ, the part corre- sponding to the entire organ as it exists in the adults of other species, I shall speak of as the nuchal notochord ; and the trough-like, posterior portion as the esophageal notochord. These terms I adopt as having reference not merely to the mor- phological relations of the respective parts — the nuchal noto- chord being situated chiefly within the neck or peduncle, while the esophageal notochord is in the esophagus — but also from the functional relations which I conceive to have been the prime factor in differentiating the organ into these two distinct parts. The nuchal notochord^ has in all probability been produced sec- ondarily from the original simple structure in coordination with the development of the proboscis and neck. This view has been held as a theory by various authors, but hitherto the only direct evidence in support of it has been the longitudinal con- striction on each side of the esophagus behind the notochordal evagination in the embryo, observed by Bateson (1884) and Morgan (1891). But the imperfect knowledge of the histor3'of these constructions has left their significance in doubt until now. Concerning the structure of the nuchal notochord, little need ' Willey 's term ' stomochord ' might well be retained for this part, but for the fact that a separate substantive applied to it in this way would imply that it is an independent organ, and this the facts do not warrant. Il8 RITTER be said here. The cavity in the anterior end is unusually vol- uminous, and in some places the walls become relatively thin, though everywhere they show distinctly the characteristic noto- chordal structure. The ventral coecum immediately in front of the body of the nuchal skeleton is well developed. The isth- mus is, as in all other species, much reduced, both in diam- eter and thickness of wall, as compared with the portion farther forward in the proboscis. It, however, maintains its size and distinctness in a greater degree than is usual. A few facts in connection with the portion of it immediately in front of the communication with the esophagus deserves particular atten- tion. A section from this locality is shown in PI. VII, fig. 5 . As maybe seen from this figure, the isthmus is here divided length- wise, sharply, into three portions, a median portion with a well-defined lumen {^in.n.no-c), and two lateral depressed, wing-like portions {l.iv. and r. w.), in which the lumen is nearly obliterated. It will be noted that the wing to the left in the figure is entirely separated from the median portion by a lamina of the nuchal skeleton. A few sections farther back the cut-off wing disappears entirely. The right wing under- goes the same cutting off and disappearance, though the sep- arating skeletal lamina comes in on this side a few sections be- hind the one here figured. Eighteen sections still farther back the body of the nuchal skeleton becomes divided lengthwise by a narrow slit, /. e.^ the skeletal crura arise, and the cubical celled epithelium forming the ventral wall of the notochord (PI. VII, fig. 5, V. cp.) reaches down through this slit and be- comes continuous with the dorsal gut epithelium, thus produc- ing an appearance in section like that shown in fig. 6. This lengthwise division of the isthmus is]|not a constant feature in all specimens, and the cutting off and disappearance of the wings I have found in the one case only. The condition is, however, none the less significant from its inconstancy. The thickening of the dorsal esophageal wall adjacent to the notochord and the vacuolation of the outer layer of its cells should be noticed here (PI. VII, fig. 5). From the level at which the wings disappear the lateral pro- jections of the isthmus, seen at .v and y (fig. 5), extend HARRIMANIA MACULOSA II9 rapidly and incline downward, until at the point of communi- cation with the esophagus the isthmus is a broad chamber, much depressed and applied ventrally close against the dorsal esoph- ageal wall (fig. 6). At this point a prominent ridge begins in the middle of the dorsal wall of the isthmus (w. r.). At its anterior end this ridge is sharply triangular in section with its apex pointing downward into the narrow slit between the closely approximated skeletal crura (fig. 7, in. r.). This is the point at which the nuchal notochord passes di- rectly over into the esophageal notochord. Fig. 7 presents a characteristic appearance of a section of the esophageal noto- chord a short distance behind its beginning. By the ridge [m. r.) mentioned above, it is divided into two well-marked lat- eral, pocket-like portions (/. ^, and r. p.). The ridge is com- posed of much elongated, closely appressed epithelial cells with their nuclei closely crowded in the apical part of the ridge. This ridge broadens and becomes relatively lower as it extends backward (fig. 4). The walls of the lateral pockets are composed of much vacuolated cells, with comparatively few nuclei confined to their inner ends. The histological similarity between the walls of these pockets and the nuchal notochord is perfect. In fact, as the description shows, the continuity from the one to the other is nowhere broken. The skeletal crura {sk.c.) mark the region at which the esophageal wall passes over into the notochordal pockets, as the figures show^, and as these pass backw^ard they diverge from each other and trend gradually ventrad (figs. 6, 7 and 4). From this disposition of the crura the esophageal notochord be- comes gradually broader, this broadening being effected chiefly by the broadening of the median dorsal ridge already described — the pockets retaining nearly their same size and character far back. As has been stated in the diagnosis, the skeletal crura reach into the pharyngeal region and overlap two or three branchial bars (fig. 3). As their posterior ends lie some distance ventral to the branchial apparatus, it follows that the first two or three branchial orifices are in the portion of the esophageal wall that is a direct backw^ard continuation of the esophageal notochord. In fact, one might say that the extreme posterior portion of the 120 RITTER notochordal pockets are pierced by the most anterior branchial orifices. But of course this is a region of transition. The skeletal crura gradually shade off into strands and thin laminae of skeletal substance ; the notochordal pockets are gradually lost, and the vacuolated notochordal tissue gradually disappears. Attention should be directed to the fact that the tissue of the median ridge of the notochord (w. r.) has more the character, histologically, of the general esophageal wall than have the pockets, and that this comes to constitute a larger and larger proportion of the entire esophageal notochord as we pass pos- teriorly. It is nevertheless true that the outermost layer of the ridge is to the very last devoid of nuclei, and is more highly vacuolated than is the epithelium in other parts of the esopha- gus. The middle portion of this ridge continues on through- out the pharynx as the c^ibranchial hand. This latter organ is unusually prominent in the present species, and its histological structure resembles closely the notochordal tissue where this reaches its best expression. Its cells are large and vacuolar, and its nuclei are few (fig. 8). Its most characteristic develop- ment is reached at some little distance behind the beginning of the pharynx, but I am inclined to believe that in this species a mid-dorsal hand of enteric zvall differentiated into chordoid tissue extends without interruption throughout the collar and pharyn- geal regions. It is impossible to discuss in the present paper the theoretical bearings of this proposition. Indeed, it may be as well that it is impossible. However, a somewhat fuller state- ment of the facts on which it rests is undoubtedly due. Spengel has pointed out^ that in the genus Balanoglossus the epibranchial band is a thickened strip of the mid-dorsal wall of the phar3-nx that is not encroached upon by the series of branchia?. It is, then, interposed between the dorsal portions of the two branchial series. In the other genera, notably Plychordera, on the other hand, the branchial series of the opposite sides come into actual contact dorsally, so that none of the enteric wall is retained between them. The former condition he regards ' See, for example, his description and figures of the structure in B. kupjferi and B. canadensis, and his diagrammatic text figures on page 545 of his mono- graph. HARRIMANIA INIACULOSA 121 as more primitive, and I should certainly agree with him in this view. The epibranchial ridge of Harrimania is of the same type as that of the species of Balanoglossiis above mentioned, and it is here, as I have already pointed out, that its histolog- ical structure is notochordal in character. Unfortunately Spen- gel was unable from scarcity of material to study exhaustively B. canadensis^ and he gives us no information about this struc- ture beyond the fact of its great breadth. Fig. 8 represents a section of the epibranchial band and the adjacent dorsal ends of a pair of septal bars between which it is included. The thickness of the epithelium, the complete absence of nuclei and of staining in the outer half of it, and the large, crowded, irregular, deeply-stained bodies among the nuclei in the inner half, make the band a striking object in a general view of the section of the animal as a whole. The wall of the gut ventral to the pharynx, not shown in the figure, is entirely different. In the first place it is scarcely more than half as thick. Struc- turally the nuclei are abundant to its very outermost stratum, and the long, narrow cell-bodies belonging to them can be seen stained, to some extent, through the entire thickness of the wall. Careful examination of sections of the band with higher mag- nification, and a comparison of these with sections of the nuchal notochord in the same individuals, convinces one of the close similarity of the two. What appear to be sharply defined fibers anastomosing with one another in a complicated way, with' a few small nuclei scattered here and there at the points of anas- tomosis, and with large vacant spaces among the fibers, are the characteristics which distinguish both tissues alike. The abrupt- ness with which the epithelium of the band passes over into the much thinner, fully nucleated, and non-vacuolar epithelium, lin- ing the inner surface of the branchial bar, will also be noted {e^, b. and br. e^.^ fig. 8). Having now presented in outline the facts relative to the pres- ence of an esophageal notochord in Harrimania maculosa, and also those which suggest that the epibranchial band in this species is to be interpreted as a continuation of the notochord into the pharynx, we may inquire whether there is any evidence afforded by other species of Enteropneusta of the presence, 122 RITTER either now or at some earlier time, of such an extension of this organ. In the first place, I present the results of my own observa- tions. In a new species of Dolichoglossiis which I am naming D. intermedms^ from the California coast, I find a condition of the supra-crural esophageal wall that I can interpret in no other way so satisfactorily as by supposing that it retains a strong rem- iniscence of the esophageal notochord. The lateral pockets and median ridge which so distinctly characterize the organ in Harrimania are absent. The supra-crural portion of the eso- phageal wall is, however, considerably broader than the thick- ness of the esophagus immediately, and for some distance, ven- tral to the crura. Furthermore, the wall is here thicker by a third or more than it is elsewhere in its circumference. But the most significant thing about it is its histological character. Fig. 9 represents a section of the dorsal half of the esophagus of this animal taken some distance behind the origin of the skeletal crura. The non-nucleated vacuolar condition of the cells in the supra-crural epithelium as compared with the infra- crural, is seen. I must say, furthermore, that the difference is in reality considerably more marked in some sections than in the one here figured. In some places the nuclei, excepting for a thin stratum at the inner surface of the epithelium, are almost wholly wanting, and the vacuolated condition is more pro- nounced. This structure is most distinct anteriorly, and gradu- ally disappears posteriorly, the epithelium taking on, by the time the extreme posterior limit of the collar is reached, th^ characters of the esophageal wall in general. There is very little or nothing in the epibranchial band of Dolichoglossiis in- termedins, so far as I am able to determine from the material at hand, to suggest its notochordal aflinities. Without doubt the esophageal notochord exists in Balano- glossus kuffferi. Spengel shows it clearly in PI. XV, figs. 29 and 44, of his monograph. The scant attention which he gives it compels me, however, to suppose that its characters, par- ticularly its histological ones, are much less clearly ex- pressed here than in Harrimania. I should expect that it is present in D. canadensis also, but Spengel's few and poorly HARRIMANIA MACULOSA 1 23 preserved fragmentary specimens of this species did not enable him to make an exhaustive study of it. As to the epibranchial band, Spengel has also shown this to have much the character in B, kti^fferi that it has in Harri- mayiia} Unfortunately, however, he tells us nothing about its histology. From the developmental side we have excellent evidence from both Bateson (1885) and Morgan (1891), of the presence of the esophageal notochord in ontogeny. Neither in- vestigator has given a full history of the organ ; in fact their references to it are rather meager. They are sufficient, never- theless, to leave no doubt about its existence. Bateson, for ex- ample, in summarizing the method of growth of the notochord, states the matter in these words : The growth of the notochord is due to : " i. A forward growth of the dorsal anterior portion of the archenteron. This is supplemented by, 2, a longitudinal constriction of the dorsal region of the pharynx, which gradu- ally travels backward, separating a hollow hypoblastic tube which remains open to the gut behind." Statement 2 is, of course, the one that particularly concerns us here. His further description and figures show that the skeletal crura are at the ' longitudinal constrictions' of which bespeaks. How far back these constrictions ' travel ' and what their relations are to the gill clefts he does not make clear. Morgan's description of the organ in the metamorphosing Toj'uarta studied by him is as follows : " The digestive-tract oesophagus in this section shows a tendency to division into two parts by lateral constrictions in its walls, forming a larger lower part and a smaller upper — the latter is directly contimiotis into the notochord farther for- ward. In the corners of the folds are two thickened chitin-like rods." (Page 424 ; italics are the author's.) His figures 58 and 59, PI. XXVni show it in characteristic form. In his theoret- ical discussion (p. 445) the author remarks: "It is certainly significant that the cavity of the notochord of Balanoglossiis is continuous with the strongly marked dorsal groove of the oeso- phagus." It would be of great interest to know more about the peculiar thick plate with rolled-up edges found on the dorsal side of the 1 PI. XVI, figs. 53, 54, and 55- 1 24 RITTER gut which Morgan shows in fig. 64, of a section taken from the extreme posterior portion of the collar region of his oldest young Balanoglossiis. The author mentions that the histological dif- ferentiation peculiar to the anterior pouch of the notochord, ex- tends to the ' strongly marked dorsal groove ' of the esophagus. There is no room for doubt, then, that both Bateson and Mor- gan have seen the esophageal notochord in the young Balano- glossus of both the species studied by them. And since these two investigators have carried out their studies more completely, both as to methods and details, and also as to stages of advance- ment in the ontogeny of the animal, than have any other zoolo- gists, we may confidently predict that fuller knowledge of the life history of other species will reveal its presence generally ^ at so7ne stage of development^ in the entire group of Enter- opncusta. REMARKS ON THE PRIMITIVENESS OF THE BALANOGLOSSID^. While, as already said, it is not my purpose to discuss the theoretical bearings of the facts presented in this paper, one point is so close at hand that I cannot ignore it entirely. The fact that the esophageal notochord is present, highly developed, in the adult of Harriniania^ while it exists only in the embryos of other species, and disappears wholly or almost so in the adult, is in itself strong evidence of the primitiveness of the species, and I conclude from this and other reasons that Harrimania stands at the very bottom among living species of the Enterop- neiista. But this view is in direct opposition to Willej^'s, who sees in the Ptychoderida; the most primitive forms, and in the Balanoglossidai the most modified, and hence the oldest, forms. After carefully considering the evidence on which Willey bases his views, the facts relating to the notochord left entirely aside, I must believe that his conclusions will not stand. That the species of animals of a homogeneous and well circumscribed group like the Enteropneusta, which are clearly the most highly differentiated and specialized, and are on the whole by far the largest in size, should be regarded as the most primitive ; while the simpler, smaller species are held to be the oldest and most modified, and this in face of the fact that all have essentially HARRIMANIA MACULOSA 1 25 the same habits of life, is a proposition that turns topsy-turvy our morphological and phylogenetic notions to such an extent that I imagine it can gain little favor among zoologists. Willey's arguments in support of his view are presented cate- gorically in what I suppose him to regard as the order of their importance in his paper on Ptychodera jiava} Stated in sub- stance only, they are as follows : (1) The gonads are diffuse in Ptychodera^ while they are disposed in masses with some sug- gestion of metameric order in some at least of the Balanoglos- sid«. (2) The gill slits of P. jlava and of some other species open directly to the exterior throughout their entire length, thus making the pharynx ' exposed ' as the author expresses it ; whereas the gill slits open into pockets which in turn communi- cate with the exterior by dorsal pores in the typical Balano- glossus. (3) The respiratory region of the enteron in Ptycho- dera is divided into a dorsal branchial portion, and a ventral esophageal portion, while in most of the Balanoglossidce there is no such division. (4) Ptychodera inhabits the littoral zone, while some of the Balanoglossidas, notably Glandiceps talaboti Marion, and G. abyssicola Spengel, are^deep-sea species. We will examine briefly these arguments one by one. The diffuse arrangement of the gonads, says the author, bears all the marks of an archaic type. In reply, I will quote another statement from the same paper (p. 174), and parallel it, slightly modified, with one of my own. In describing the gonads of P. jiava, he says: "They are not in the remotest degree ar- ranged one after the other, in a manner resembling a paired metameric series, as they are more or less in most other Enter- opneusta, but they are scattered in the most irregular way in the substance of the genital pleura." My parallel to this is as fol- lows, making the comparison between Am^hioxtis and Bdcllo- stoma instead of \)Q.\.\NQ.^r\. Balaiioglossiis and Ptychodera: They (the gonads of Bdellostoma^ are not in the remotest degree ar- ranged one after the other, in a manner resembling a paired metameric series, as they are in Amphwxtts, but are scattered in the most irregular way in the substance of the genital fold. So far as the evidence as to primitiveness is concerned, I believe ' Willey, 1897, p. 179. 126 RITTER this parallel is perfectly fair. Diffuseness of the gonads is no more a trustworthy index in this regard in the one case than in the other. But should it be objected that Amphwxus and Bdello- stoma are so much more remote from each other, taxonomically, than are Balanoglossus and Ptychodera, that the comparison is not altogether just, let us take another instance. In the genus Poly car fa among the simple ascidians, the gonads in some of the species are scattered widely over the mantle to which they are attached in what are known as polycarps ; whereas in As- ct'dia the gonad forms a somewhat lobulated, but yet a compact, single mass. In short the gonads are diffuse in the one case, and are not so in the other ; yet no one would conclude from this that Polycarfa is a more archaic genus than is Ascidia. I should not have expected that so excellent a morphologist as Willey would have staked so much on this point, familiar as he is with the wide differences to which the gonads are fre- quently subject, even within the range of rather circumscribed groups of animals. The truth of the matter is, as Willey himself has shown (see his more recent and much fuller memoir, 1899), that the gonads oiPtyckodera, like those of all other Enteropneusta, are w^ell de- fined, more or less globular masses, each with its own envelop- ing basement membrane, central cavity, and short duct opening directly or indirectly to the exterior. And the only way in which they are more diffuse in Ptychodera than in Balanoglos- sus is that these masses are much smaller, probably in correla- tion with the smaller size of the ova in the female, are much more numerous, and more closely crowded in the former than in the latter genus. On the whole, I am quite sure that so far as the gonads of Ptychodera indicate anything regarding the ancestry of the animal, they indicate it to be more highly modilied and hence presumably older than Balanoglossus. One thing among others pointing in this direction is the apparent fact that the gonadal cavities, which are unquestionably primitive, are more com- pletely obliterated in Ptychodera than in the Balanoglossida3. Concerning the exposed pharynx of Ptychodera flava and its congeners, I would say that while I am not at all confident HARRIMANIA MACULOSA 1 27 about its meaning, it seems to me that at least as strong a case can be made for the supposition as against it, that the condition is the result of secondary modification from the more tj^pical en- teropneustic arrangement. Willey's observations relative to the closing'over and mucous fusion of the genital pleurae at their free edges to produce a transient atrial chamber in this species, and his reflections upon the possible significance of the facts, are exceedingly interesting. Holding as he does, that the gen- ital pleurae are primitive structures, he is, of course, justified in assuminsf that in case there is a causal relation between them and the peculiarities of the pharynx, the two have developed pari ^assu, and hence that the one is as old as the other. On the other hand, granting such a relation between the two, if one holds as I do, following Spengel, that the pleurce are secondary acquisitions, then it might follow likewise that the peculiarity of the pharynx is secondary. But I doubt if there is yet suffi- cient evidence at hand to warrant a positive opinion one way or the other. I may add, for what it is worth, that in my new species DolichoglossiLS mtermedms, in which the outline of the body in the pharyngeal region is less broken by genital ridges and bran- chial grooves than in any other known Enteropneust, the gill orifices are more elongated transversely and more laterally directed than in any species except PtycJwdera of the jiava type. As to Willey's third point, that the absence of division of the anterior region of the alimentary canal into branchial and esoph- ageal portions in the Balanoglossidee " militates strongly against their being regarded as more primitive than P/y(:/z<932) Proc. Wash. Acad. Sci. Vol.11. Plate VII. sk. c ,<1. II. c d b V |» r^ , yr p. 1 p ,<^ "'V=f sk c. >-. ^. ''^ sk c. es. no-c.^ Ilavriinania maculosa. PROCEEDINQS WASHNGTON ACADEMY OF SCIENCES Vol. II, pp. 133-156. August 20, 1900. RESULTS OF THE BRANNER-AGASSIZ EXPEDI- TION TO BRAZIL. I. THE DECAPOD AND STOMATOPOD CRUSTACEA. By Mary J. Rathbun. During the summer of 1899, Dr. J. C. Branner visited Brazil for the purpose of studying the stone and coral reefs of the coast between Cape St. Roque and Rio de Janeiro. The ex- penses of the trip were borne chiefly by Professor Alexander Agassiz, of Harvard University. Mr. Arthur W. Greeley, of the San Diego State Normal School, California, accompanied the expedition as naturalist, and the biological collections were made chiefly by him, with such assistance as other members of the party could give from time to time. The collecting was all done between June 3 and August 8, 1899, between Natal, in the State of Rio Grande do Norte, and Maceio, State of Ala- goas. The decapod and stomatopod crustaceans collected number seventy species. Six of these were undescribed, and the known ranges of many other species have been extended. The types of new species are in the U. S. National Museum. Proc. Wash. Acad. Sci. August, 1900. (^33) 134 RATHBUN Order DECAPODA. Family OCYPODIDiE. OCYPODE ALBICANS Bosc. Ocypode albicans Bosc, Hist. Nat. Crust., I, 196, pi. iv, fig. i, 1802 (figure inaccurate). ^ Ocypode arenaria Say, Jour. Acad. Nat. Sci. Phila., i, 69, 1817. Ocypode arenaria Milne Edwards, Hist. Nat. Crust., 11, 44, pi. xix, figs. 13-14, 1837. Trai9ao, near Mamanguape River; one male. Maceio, Alagoas, on coral reef and sand beach; two males. UCA MARACOANI (Latreille). Ocypode Juaracoani'LkTV.Y.WA.^, Hist. Nat. Crust., vi, 46, 1803. Gelashnus maracoani Latreille, Diet. Hist. Nat., xii, 519, 1817 ; Milne Edwards, Ann. Sci. Nat. (3), xviii, 144 [108], pi. in, fig. i, 1852. Natal, Rio Grande do Norte; six males, one female. Mangroves, Rio Parahyba do Norte at Cabedello ; one male, one female. UCA MORDAX (Smith). Gelasimus mordaxSujTH, Trans. Conn. Acad. Sci., 11, 135, pi. 11, fig. 3 ; pi. IV, figs. 4, 4a, 1870. l/ca 7nordax Rathbun, Proc. U. S. Nat. Mus., xxii, 276, 1900. Pernambuco, on mangroves ; one male. UCA THAYERI sp. nov. Plate VIII, figs, i and 2. Carapace very broad in its anterior fourth, narrowing rapidly in its posterior three-fourths ; antero-lateral angles almost rectangular, blunt. Sometimes the carapace is narrower at the orbital angles than a little behind that point, caused by the curving inward, at the lateral angle, of one or both of the lateral margins. The dorsal furrows are deep, especially the cervical, and the transverse gastro-cardiac furrow ; less deep are the obliquely longitudinal branchial furrow and the post- orbital. vSurface finely and densely granulate through the lens, and tomentose, the hair retaining particles of mud. ' I have restored Bosc's name to this crab, as his description was made from specimens on the coast of Carolina, where no other species of the genus occurs. The fact that in the figure the artist has represented the eyes with stalks be- yond the cornete does not, I think, invalidate the species. Say was the first bi- nomial writer to use the specific name arenaria. THE DECAPOD AND STOMATOPOD CRUSTACEA 135 Front very narrow, not linear nor spatuliform, but subtriangular, at base or posteriorly less than one-fifth the width between the antero- lateral angles of the carapace, anteriorly truncate or nearly so ; sides oblique and almost straight. Superior orbital surface, or eyebrow, shallow, not varying much in length throughout its width ; margins finely granulate. Inferior margin of orbit with large truncate tuber- cles, increasing in size and distance apart towai'd the outer extremity. Large cheliped very heavy. Merus and carpus elongate, thick, rugose on the outer surface, and without armed margins. Outer sur- face of palm coarsely tuberculate on its upper half, the tubercles gradually becoming fine granules below ; upper and lower margins set off by deep grooves. Inner surface of palm with a ridge marked by a single line of large tubercles, leading obliquely upward from the lower mai"gin to the carpal cavity, where it turns at a little less than a right angle and is continued less than half way to the upper margin, or when continued further, the tubercles are obsolete. On the palm at the base of the dactylus are two tuberculate lines, the distal one very short ; both are slightly oblique to the lower margin. In full grown males the fingers are very long, the lower margin of the propodus sinuous, the pollex bent down for its distal third. The dactylus equals or overreaches the pollex. The prehensile tubercles are irreg- ular, but not sti-ikingly so. The dactylus is roughened at its base on the upper side, and has a short longitudinal groove on the outer side below the upper margin. The smaller cheliped is rather long, the fingers longer than the palm and somewhat gaping to the tips. The meral joints of the am- bulatory legs are dilated and very broad, especially those of the second and third pairs. TABLE OF DIMENSIONS. Locality. to a V h cd 0 rt -*-* rt 0 0 tu *. a . £ ort at . 0 0. 0 V ££? 4J 2 1 ■0 <«< 0 t a v Rio Parahyba do Norte. S 17.2 27.1 27.5 II.6 21 48.3 HS 35-4 do. V 13 19 17-5 9-5 15-5 do. s 15 23-4 22.5 10.5 17-3 42.5 12.7 31b Sexual and age variations. — Old males show a tendency to widen behind the antero-lateral angles ; this tendency is stronger on 136 RATHBUN the side of the large claw than on the side of the small one. In males which are young or have not reached their fullest development, the pollex of the large chela is straight, not bent down, and the lower margin of the propodus is convex, not sinuous. The meral joints of the ambulatory legs are wider in the female than in the male, as is the case in other species of the genus. Rang^. — This species was first taken by the Thayer Expedition on the coast of Brazil, at Rio Parahyba do Norte, Sao Matheos and Vic- toria. Specimens from these localities are in the Museum of Com- parative Zoology, and one male from the Rio Parahyba do Norte is in the U. S. National Museum. In 1884, the 'Albatross' collected at Jamaica one male, the claw of which was not reproduced at its last shedding. Types. — Eight males and one female were taken on the Branner- Agassiz Expedition, among the mangroves on the Rio Parahyba do Norte at Cabedello. Cat. No. 23753. Additio?ial locality. — Natal, Rio Grande do Norte; one male. UCA SPINICARPA Rathbun. Uca spinicarpa Rathbun, Amer. Nat., xxxiv, 586, 1900. Mamanguape stone reef; one male, soft shell. This specimen is too shapeless to be determined with certainty, but it appears to be U. spinicarpa^ which has been taken in the Gulf of Mexico on the coast of Alabama, Mississippi, Texas, and Tslexico. The species is distinguished by the truncate anterior margin of the front between the eyes, and the outline of the lateral margins, which are straight and subparallel in their anterior portion, and then curve abruptly inward and backward. The carpus of the large cheliped has a stout spine or tooth at the middle of its inner surface. The in- ner face of the palm has a prominent crest crowned by a single row of large tubercles extending obliquely upward from the lower margin to the cavity in which the carpus fits, thence it turns at a right angle and meets the upper margin ; there are two rows of tubercles at the base of the dactylus; the remainder of the surface is smooth or nearly so. UCA LEPTODACTYLA Rathbun. hca leptodactyla Rathhun, in Rankin, Ann. N. Y. Acad. Sci., xi, No. 12, 227, 1898, and synonymy. Mangroves, Rio Parahyba do Norte ; one male. THE DECAPOD AND STOMATOPOD CRUSTACEA 1 37 Family GECARCINIDiE. UCIDES CORDATUS (LinmBus). Cancer cordatus "Liti-iiJEVS, Amcen. Acad., vi, 414, 1763. Uca una Guerin, Icon. Regne Anim. Cuvier, pi. v, fig. 3, 9 . Uca cordxia White, List Crust. Brit. Mus., 31, 1847. Ucides cordatus Rathbun, Ann. Inst. Jamaica, i, 25, 1897, and synonymy. Mangroves, Rio Parahyba do Norte, Cabedello ; three females. Family GRAPSID^. GONIOPSIS CRUENTATUS (Latreille). Grapsus cruentatus hATK^i-Li,^, Hist. Nat. Crust., vi, 70, 1803. Grapsus {Goniopsis) crueiitatiis de Haan, Fauna Japon., 33, 1835. Goniograpsus criietitatus Dana, Crust. U. S. Expl. Exped., 342, 1852 ; atlas, pi. XXI, fig. 7, 1855. Rio Parahyba do Norte, Cabedello, on mangroves ; one male, one female. Pernambuco, on mangroves ; three males. PACHYGRAPSUS TRANSVERSUS Gibbes. PackygrapsustransverstisG\v.^^s, Proc. Amer. Assoc. Adv. Sci.,iii, 181, 1850. Goniograpsus innoiatus Dana, Proc. Acad. Nat. Sci. Phila., 249, 1851 ; Crust. U. S. Expl. Exped., 345, 1852 ; atlas, pi. xxi, fig. 9, 1855. Mamanguape stone reef; tw^o males. Rio Parahyba do Norte, Cabedello, on mangroves ; one male, one female. Rio Goyanna stone reef ; four males. Pernambuco stone reef ; one male. Stone reef at Boa Viagem, five miles south of Pernambuco ; one male. Coral reef, Maceio, Alagoas ; one male, one female. PACHYGRAPSUS GRACILIS (Saussure). Metopograps7is gracilis Saussure, Mem. Soc. Phys. Hist. Nat. Geneve, xiv, 443- pl- II. fig- 15, 1858. Pachygrapsus gracilis STIM.VSO-S, Ann. Lye. Nat. Hist. N. Y., x, 113, 1871. Rio Parahyba do Norte, on mangroves ; one male, one female. SESARMA (HOLOMETOPUS) RECTUM Randall. '.a recta Randall, Jour. Acad. Nat. Sci. Phila., viii, la ?nullcrii A. Milne Edwards, Nouv. Arch. Mus. 27, 1869. Pernambuco, on mangroves ; one male, one female. Sesarma recta Randall, Jour. Acad. Nat. Sci. Phila., viii, 123, 1839. Sesarma mullcrii A. Milne Edwards, Nouv. Arch. Mus. Hist. Nat. Paris, V, 27, 1869. 138 RATHBUN SESARMA (HOLOMETOPUS) MIERSII Rathbun. Sesarma {Holometopjis) micrsii Rathbun, Proc. Biol. Soc. Washington, xi, 91, 1897, and synonymy. Rio Parahyba do Norte, Cabedello, on mangroves; one male. ARATUS PISONII Milne Edwards. Sesarma pisonii Milne Edwards, Hist. Nat. Crust., 11, jS, pi. xix, figs. 4 and 5, 1837. AraftispisonzM.ii.NE Edwards, Ann. Sci. Nat. (3), xx, 187, 1853. Rio Parahyba do Norte, Cabedello, on mangroves; five males, nine females. Lagoa do Norte, Maceio, on mangroves ; three females. PLAGUSIA DEPRESSA (Fabricius). Cancer depressus Fabricius, Syst. Ent., 406, 1775. Plagusia depressiis Say, Jour. Acad. Nat. Sci. Phila., i, 100, 181 7. Mamanguape stone reef ; one female. Rio Goyanna stone reef; three females. Pernambuco stone reef ; one female. Color. — Brilliant; for the most part a white ground thickly covered with fine red dots ; also blotches of clear red.^ Family PILUMNID^. CARPILIUS CORALLINUS (Herbst). Cancer corallinus Herbst, Natur. Krabben u. Krebse, i, 133, pi. v, fig. 40, 1783. Carpilius corallinus \jEKCn, in Desmarest, Consid. sur les Crust., 104, 1825. Goyanna stone reef ; one male. CYCLOXANTHOPS DENTICULATUS (White). Xaniho denticulatus White, Ann. Mag. Nat. Hist. (2), 11, 285, 1848. A. Milne Edwards, Crust. R6g. Mex., 252, pi. xlv, fig. 2, 1879. Cycloxanthops denticulatus Rathbun, Ann. Inst. Jamaica, i, 14, 1897. Stone reef at Boa Viagem ; one female. Coral reef, Maceio, Alagoas ; thirteen males, seven females. Color. — Gay and not constant, varying from shades of light purplish red through bright red to orange red. 'Color notes have been made from specimens in formalin. THE DECAPOD AND STOMATOPOD CRUSTACEA I39 MENIPPE NODIFRONS Stimpson. Pseudocarcinus rwnphii Milne Edwards, Hist. Nat. Crust., i, 408, 1834. Not Cancer rumphii Fabricius. Menippc rumphiiDA^\, Crust. U. S. Expl. Exped., i, 179, 1852. A. Milne Edwards, Crust. Reg. Mex., 263, pi. xlviii, fig. 4, 1879. Menippe nodifrons Stimpson, Ann. Lye. Nat. Hist. N. Y., vii, 53, 1859. Mamanguape stone reef ; two young. Rio Goyanna stone reef ; one female. Pernambuco stone reef ; three young. Pernambuco stone reef at Ilha de Nogueira ; one young. Coral reef, Maceio, Alagoas; one male, two females. PILUMNUS RETICULATUS Stimpson. Pilumnus reticulatus STiMVSOti, Ann, Lye. Nat. Hist. N. Y., vii, 214, i860. Piluniniis tessellatus A. Milne Edwards, Crust. Reg. Mex., 295, pi. li, fig. 2, 1880. Pernambuco stone reef at Ilha de Nogueira; one male, one female. PILUMNUS ANDREWSII Rathbun. Pilumnus andrewsii Rathbun, Bull. Lab. Nat. Hist, State Univ. Iowa, iv, 266, pi. V, fig. 2, 1898. Coral reef, Maceio, Alagoas ; one male. In this specimen, which is a little smaller than the type, the frontal lobes are more sinuous than truncate, the lateral spines are less acu- minate, and the spinules of the carpal joints of the chielipeds less sharp. PILUMNUS DASYPODUS Kingsley. Pilumnus dasypodus Kingsley, Proe. Boston Soe. Nat, Hist,, xx, 155, 1879. Pilutnnus vinaceus A. Milne Edwards, Crust, R6g. Mex,, 283, pi, l, fig, 2, 1880, Pernambuco stone reef at Ilha de Nogueira ; three females. Stone reef at Boa Viagem ; three males, one female. Coral reef, Maceio, Alagoas ; one male, one female. LEPTODIUS FLORIDANUS (Gibbes). Chlorodius floridanus 0\^Y,Y.s, Proe. Amer. Assoc, Adv, Sci,, iii, 175, 1850. Leptodius floridanus A. Milne Edwards, Crust. Reg. Mex., 268, pi. xlix, fig. 2, 1880, Mamanguape stone reef ; two males, one female. Rio Goyanna stone reef ; three males, one female. Coral reef, Maceio, Alagoas; one male, three females. 140 RATHBUN EUPANOPEUS HERBSTII (Milne Edwards). Panopcits Iterbs/ii 'Mw.^Y, Edwards, Hist. Nat. Crust., i, 403, 1834. Bene- dict andRATHBUx, Proc. U. S. Nat. Mus., xiv, 358, pi. xix, figs, i and 2 ; pi. XXIII, figs. 10-12, 1891, and synonymy. Eiipanopetis herbstii Rath bun, Bull. Lab. Nat. Hist. State Univ. Iowa, iv, 273, 1898. Mamanguape stone reef ; one female. Rio Parahyba do Norte, Cabedello, on mangroves ; four males, five females. EUPANOPEUS OCCIDENT ALIS (Saussure). Panopeus occidenialis Skvssvr^, Rev. Mag. Zool. (2), ix, 502, 1S57. Bene- dict and Rathbun, Proc. U. S. Nat. Mus., xiv, 360, pi. xx, fig. 3 ; pi. xxili, fig. 14, 1 891, and synonymy. Eupanopfus occidentaiis Rathbun, Bull. Lab. Nat. Hist. State Univ. Iowa, IV, 273, 1898. Rio Parahyba do Norte, on mangroves; two males. Pernambuco stone reef ; one female. EUPANOPEUS AMERICANUS (Saussure). Panopeus mu eric a7i lis Saussure, Rev. Mag. Zool. (2), ix, 502, 1857 ; Mem. Soc. Phys, Geneve, xiv, 432, pi. i, fig. 8, 1857. Panopeus areolatiis Benedict and Rathbun, Proc. U. S. Nat. Mus., xiv, 361, pi. XXI, fig. 3, 1891. Eiipanopeits aviericamis Rathbun, Bull. Lab. Nat. Hist. State Univ. Iowa, IV, 273, 1898. Rio Parahyba do Norte, on mangroves ; one female. EUPANOPEUS BERAIUDENSIS (Benedict and Rathbun). Panopeus uiurdevtannii Benedict and Rathbun, Proc. U. S. Nat. Mus., XIV, 372, pi. XXIV, figs. 6 and 7, 1891 (not P-. laurde/nannit GimM-.s). Panopeus bermudensis Benedict and Rathbun, Proc. U. S. Nat. Mus., xiv, 376, pi. XX, fig. 2 ; pi. XXIV, figs. 14 and 15, 1891. Etipanopeus bertnudetisis Rathbun, Bull. Lab. Nat. Hist. State Univ. Iowa, IV, 273, 1898. Coral reef, Maceio, Alagoas ; one female. EUPANOPEUS HARTTII (Smith). Panopeus /iar/Zii Smith, Proc. Boston Soc. Nat. Hist., xii, 280, 1869 ; Trans. Conn. Acad. Sci., 11, pp. 5 and 34, pi. i, fig. 5, 1869. Eupanopeus harttii RAthhun, Bull. Lab. Nat. Hist. State Univ. Iowa, iv, 273, 1898. Coral reef, Maceio, Alagoas; two males. THE DECAPOD AND STOMATOPOD CRUSTACEA I4I EUPANOPEUS ABBREVIATUS (Stimpson). Xantho parviilus Milne Edwards, Hist. Nat. Crust., i, 395, 1834 ; not CiXHcer parvulus Fabricius. Panopt-us abbreviatus Stim.vso^, Ann. Lye. Nat. Hist. N. Y., vii, 211, i860. Eurypaiiopeus parvulus A. Milne Edwards, Crust. Reg. Me.x., 322, pi. Lix, fig. 5, 1880. Eurypanopeus abbreviatus A. Milne Edwards, Crust. Reg. Mex., 320, pi. LIX, fig. 3, 1880. Panope us parvulus Benedict and Rathhun, Proc. U. S. Nat. Mus., xiv, 369, pi. XXI, fig. I, pi. XXIII, figs. 2 and 3, 1891. Mamanguape stone reef ; five males, seven females. Rio Goyanna stone reef ; two males. Boa Viagem stone I'eef ; one male, one female. Maceio coral reef ; three males, one female. ERIPHIA GONAGRA (Fabricius). Cancer gonagra Fabricius, Sp. Ins., 505, 1781. Eriphia gonagra Milne Edwards, Hist. Nat. Crust., i, 426, pi. xvi, figs. 16 and 17, 1834. Mamanguape stone reef ; one male. Rio Goyanna stone reef; seven males, tw^o females. Pernambuco stone reef ; three males, two females. Boa Viagem stone reef; one female, soft shell. Maceio coral reef; eight males, three females. Color. — Variable, but always bright. Some are reddish brown or yellowish brown with darker patches ; spines and margins of front and orbits orange. Others show no red nor yellow, only browns and orange. Legs with light yellow ground, covered in great part with fine red dots. Claws with the tubercles of the upper half dark red or blue, of the lower half yellow ; fingers dark red. DOMECIA HISPIDA Eydoux and Souleyet. Domecia hispida Eydoux and Souleyet, Voy. Bonite, i. Crust., 235, 1842 ; atlas, pi. II, figs. 5-10 {Domecie herissee, on plate). Maceio coral reef ; one female. Family PORTUNIDiE. CALLINECTES DAN^ Smith. Lupa dicantha Dana, Crust. U. S. Expl. Exped., i, 272, 1852 ; atlas, pi. XVI, fig. 7, 1855. Callinectes dance SuiTH, Trans. Conn. Acad. Sci., 11, 7, 1869. Mamanguape stone reef; one young female. Rio Parahyba do Norte, Cabedello, on mangroves ; one male, one young female. 142 RATHBUN CALLINECTES MARGIN ATUS (A. Milne Edwards). Neptunus marginatus A. Milke Edwards, Arch. Mus. Hist. Nat. Paris, x, 318, pi. XXX, fig. 2, 1861. * Callinectes larvatus Ordway, Boston Jour. Nat. Hist., vii, 573, 1863. Caliifiecfes marginatus Rathbun, Proc. Biol. Soc. Washington, xi, 149, 1897, and synonymy. Mamanguape stone reef ; four young. Rio Goyanna stone reef ; one male. Rio Parahyba do Norte, on mangroves ; two young. Pernambuco stone reef; one male. Maceio coral reef; four males, one young female. Color. — Dull brown, with areas of bluish black ; claw brown above, blackish blue below and on the inner side ; last segments of swim- ming feet a brighter brown. AREN^US CRIBRARIUS (Lamarck). Portunus cribrarius Lamarck, Hist. Nat. Anim. sans Vert., v, 259, 1818. Arenaus cribrarius Dana, Crust. U. S. Expl. Exped., i, 290, 1852 ; atlas, pi. xviii, fig. 2, 1855. Mamanguape stone reef ; one male, two females. Family MAIID^. ACANTHONYX PETIVERII Milne Edwards. Acanthonyx petiverHyiws^^ Edwards, Hist. Nat. Crust., i, 343, 1834. Acanthonyx petiveri h.. Milne Edwards, Crust. Reg. Mex., 143, pi. xxvii, fig. 7, 1878, and synonymy. Mamanguape stone reef ; one female. Boa Viagem stone reef ; one female. MICROPHRYS BICORNUTUS (Latreille). Pisa bicornutct\jci.\xe-\\\Q, Encyc. Meth., Hist. Nat., Insectes, x, 141, 1825. Microphrys bicornutus A. Milne Edwards, Nouv. Arch. Mus. Hist. Nat. Paris, VIII, 247, 1872 ; Crust. Reg. Mex., 61, pi. xiv, figs. 2-4, 1875, ^^^ synonymy. Mamanguape stone reef ; two males. Pernambuco stone reef at Ilha de Nogueira ; two males, four fe- males. Maceio coral reef ; five males, five females. MITHRAX VERRUCOSUS Milne Edwards. Mithrax verrucosus '^llv.^v.Y.MWKYiXiS, Mag. Zool., 11, pi. iv, 1832. Mamanguape stone reef ; one young. Pernambuco stone reef at Ilha de Nogueira ; one young. Maceio coral reef ; two young. THE DECAPOD AND STOMATOPOD CRUSTACEA 1 43 MITHRAX BRAZILIENSIS Rathbun. Mithrax braziliensis Rathbun, Proc. U. S. Nat. Mus., xv, 268, pj. xxxvi, fig. 2, 1892. Maceio coral reef ; two males. Color. — A rich dark crimson. MITHRAX FORCEPS (A. Milne Edwards). Mithracuhis forceps A. Milne Edwards, Crust. Reg. Mex., 109, pi. xxni, AhUrfx forceps Mmns, Challenger Rept.. Zool., xvii, pp. 87 and 88, 1886. Rathbun, Proc. U. S. Nat. Mus., xv, 269, 1892, and synonymy. Natal, Rio Grande do Norte ; two males. Pernambuco stone reef at Ilha de Nogueira ; one male. Maceio coral reef ; four males, seven females. MITHRAX CORYPHE (Herbst). Cancer cor onaiiis Herbst, Natur. Krabben u. Krebse, i, 184, pi. xi, fig. 63, 1785 ; not C. coronatus Molina, 1782. Cancer coryphe Herbst, op. cit., iii. Heft 2, p. 8, 1801. Mithrax coronatus Miers, Challenger Rept., Zool., xvii, pp. 87, 89, 1886. Rathbun, Proc. U. S. Nat. Mus., xv, 272, 1892, and synonymy. Mithrax coryphe Rathbun, Ann, Inst. Jamaica, i, ii,__i897. Rio Goyanna stone reef ; two males. Pernambuco stone reef at Ilha de Nogueira ; three males. Boa Viagem stone reef ; two females. Maceio coral reef ; eleven males, four females. TELEOPHRYS CRISTULIPES Stimpson. Teleophrys cristulipes Stimpson, Ann. Lye. Nat. Hist. N. Y., vii, 190, pi. n, fig. 2, i860. Mithrax cristulipes Rathbun, Proc. U. S. Nat. Mus., xv, 273, 1892, and synonymy. Maceio coral reef ; five males, four females. Family DROMIID^. DROMIA ERYTHROPUS (George Edwards). Cancer marinus'chelis rubris Catesby, Nat. Hist. Carolina, Florida and Ba- hama Islands, 11. 37, pi. 37. I743- . Cancer erythropus George Edwards, Catalogue of Animals m Catesby s Nat. Hist, of Carolina, with the Linnaean names, 1 771. Dromia /a/6>r Milne Edwards, Hist. Nat. Crust., 11, 174. 1837. Dromia erythropus Rathbun, Ann. Inst. Jamaica, i, 39, 1897. Pernambuco stone reef at Ilha de Nogueira ; one young male. 144 RATHBUN Family PAGURID^. PETROCHIRUS INSIGNIS (Saussure). Pagurus insignis Saussure, Mem. Soc. Phys. Hist. Nat. Geneve, xiv, 453 [37], pi. Ill, fig. 20, 1858. Rio Goyanna stone reef ; one specimen. Maceio coral reef ; one specimen. Color. — Ambulatory legs with four transverse reddish bands. CALCINUS SULCATUS (Milne Edwards) . Pagurus sukatus Milne Edwards, Ann. Sci. Nat. (2), vi, 279, 1836. Ca/cinus su/ca/us STiM.FSO'ii, Proc. Acad. Nat. Sci. Phila., x, 234 [72], 1858. Mamanguape stone reef; one specimen. Pernambuco stone reef ; two specimens. Maceio coral reef ; two specimens. Color. — Gastric region green or greenish in the center, dark red at the sides. Eye stalks light brownish yellow in basal half, turning to red distally; band next to the cornea white ; cornea black. Claws dark brown, with red margins; fingers deep yellow, white near the tips. Second and third pairs of feet yellow, dactyli with a red band in the center, nails black. CLIBANARIUS VITTATUS (Bosc). Pagurtts vittatus Bosc, Hist. Nat. Crust., 11, 78, pi. xii, fig. i, 1802. Ciibanarius vittatus Stimpson, Proc. Acad. Nat. Sci. Phila., x, 235 [73], 1858. Mamanguape stone reef ; two specimens. Rio Parahyba do Norte, Cabedello, on mangroves; six specimens. CLIBANARIUS CUBENSIS (Saussure). ICancer sclopetxrius Herbst, Natur. Krabben u. Krebse, 11, 23, pi. xxiii, fig- 3. 1791- Pagurus cube?isis SAUSSURE, Mem. Soc. Phys. Hist. Nat. Geneve, xiv, 455 [39]. 1858. Ciibanarius sclopetarlus Stimpson, Proc. Acad. Nat. Sci. Phila., x, 235, 1858. Rathuun, Ann. Inst. Jamaica, i, 43, 1897. Rio Parahyba do Norte, Cabedello, on mangroves; six specimens. CLIBANARIUS ANTILLENSIS Stimpson. Ciibanarius antillensis Stimpson, Ann. Lye. Nat. Hist. N. Y. , vii, 85, 1859. Rio Goyanna stone reef; one specimen. Color. — Claws orange-red with a white spot at each spine. Am- bulatory legs with a white stripe through the middle of the outer and of the inner surface. THE DECAPOD AND STOMATOPOD CRUSTACEA I45 Family PORCELLANIDiE. PETROLISTHES ARMATUS (Gibbes). Porcellana armata GuMW.s, Proc. Amer. Assoc. Adv. Sci., ill, 190, 1850. Porcellana leporina Heller, Reise Novara, 78, pi. vi, fig. 7, 1865. Petrolisihcs armatiis Simvso'^, Proc. Acad. Nat. Sci. Phila., x, 227, 1858. Mamanguape stone reef ; eleven specimens. Rio Goyanna stone reef; six specimens. Maceio coral reef ; one specimen. Color. — Bright red and yellowish mottled. On the chelipeds the red strongly predominates, the tubercles being red, the interspaces yel- low. Propodal joints of ambulatory legs with two light transverse bands, one at the distal end, the other near the proximal. PETROLISTHES SEXSPINOSUS (Gibbes). Porcellana sexspinosa Gibbes, Proc. Amer. Assoc. Adv. Sci., iii, 190, 1850. Porcellana egregia Guerin, in La Sagra's Hist. Cuba, 2d part, vol. viii (Atlas), pi. II, fig. I, 1857. Petrolisthes sexspinosus Stimpson, Proc. Acad. Nat. Sci. Phila., x, 227, 1858. Mamanguape stone reef ; three specimens. Stone reef at Boa Viagem, Pernambuco ; three specimens. Coral reef, Maceio, Alagoas ; seven specimens. Color. — Dark crimson in transverse broken lines on a white ground. Similar oblique lines on chelipeds, showing a tendency to break up into spots; outer margin, and also the margin of the dactylus, with a row of crimson spots. Ambulatory legs with meral joints spotted ; carpal joints speckled with crimson and with a white stripe on the upper surface ; propodal joints crimson with two white bands, one at the distal end, the other on the proximal half. Dactyli crimson, with white toward the nail. PETROLISTHES SERRATUS Henderson. Petrolisthes serratus Henderson, Challenger Rept., Zool., xxvii, 107, pi. XI, fig. 2, 1888. The single specimen, a male, agrees with Henderson's description and figure, excepting that the inner border of the carpus of the right cheliped (the left is missing) has only four teeth, the small distal fifth tooth figured by Henderson being absent, and the dactylus is shorter than in the type, being 4.2 mm. long, while the inner margin of the propodus measures 6.4 mm. The dactylus has a large basal tooth, and a smaller tooth next it. Henderson says that the dactylus is " almost equal in length to the hand." His type was from off Bahia in 146 RATHBUN 7 to 20 fathoms. The specimen in hand is from the coral reef at Maceio, Alagoas ; like the type it is an adult male, but smaller ; the carapace is cracked, so that it cannot be accurately measured. The ground color is a deep crimson, covered with narrow and closely interlacing stripes of white, leaving small irregular interstices of crimson. The stripes are narrower and the interstices smaller near the margin of the carapace. Chelipeds similar in coloring. On the ambulatory legs the white stripes are broken, forming small irregular detached patches of white ; the propodal joints have three narrow transverse bands of white, two at the proximal end, and one near the distal end. P. serratus is closely allied to, if not identical with, P. amcctius (Guerin). PISOSOMA RIISEI Stimpson. Pisosoma rm<'/ Stimpson, Ann. Lye. Nat. Hist. N. Y., vii, 75, 1859. Pisosonia glabra Kingsley, Proc. Acad. Nat. Sci. Phila., xxxi, 406, pi. xiv, fig. 2, 1879 [1880]. Coral reef, Maceio, Alagoas ; two females. In these specimens the front is not the same : in one the upper margin is nearly straight, in the other more advanced at the middle. In the one cheliped present, the carpus is about as broad as long, op- posed to Kingsley's figure rather than to his description ; the inner margin is sinuous. The palm is subtriangular, wider at the base of the dactylus than its inner length. The outer marginal crest of the palm is continued two-thirds the length of the carpus, and on that seg- ment is more or less tuberculous or broken up by oblique rugte. Dactylus with a small basal tooth, which when the fingers ai-e closed, fits against a tooth on the propodus between the two fingers ; upper margin with a longitudinal groove on the basal half. Fingers gaping. The types of P. glabra Kingsley, with which these have been com- pared, are smaller, the fingers shut tight and their tips cross ; the in- ner margin of the carpus is more distinctly dentate. Di}ne7isio7is of female bearing eggs. — Length 4.5 mm. ; breadth 5.3 mm. Color. — A beautiful light crimson, with a small white spot on the lobe above the antenna and behind it on the margin a larger white spot; these two spots unite below the dorsal surface forming a white band extending to the cheliped. A small white dot on the branchial region adjoining the cardiac region. Chelipeds crimson, tips of fingers white. Ambulatory legs crimson with white bands, a narrow one at either end of the propodus and a wide one covering the proxi- mal half of the merus. THE DECAPOD AND STOMATOPOD CRUSTACEA I47 PISOSOMA GREELEYI sp. nov. PI. VIII, fig. 4- Male. Carapace slightly broader than long, subcircular. Front with a double margin, the upper margin arcuate, continuous with the line of the orbit ; the lower margin produced downwards in an acute median tooth and forming a rectangular preorbital tooth. Postorbi- tal tooth small and blunt ; behind this on the antero-lateral margin there is a shallow lobe followed by a notch at the cervical suture. Lateral margin marked by a narrow raised line which turns inward on the carapace at the posterior third of the branchial region. Proto- o-astric lobes large, separated by a shallow median sinus. Cervical sutures deep, Branchio-cardiac furrows also well marked. A slight transverse depression across the middle of the branchial region. Sur- face ornamented with scattered scabrous granules, posterior and pos- tero-lateral portions crossed by fine rugose lines. The upper surface of the merus of the chelipeds is crossed by fine transverse rugte ; inner margin with a blunt and granular prominence near the distal angle ; distal margin of lower surface with two or three spinules. Upper surface of carpus, propodus and dactylus covered with coarse granules, some of which are large and tuber- cular. The carpus at its widest portion is about as wide as long ; its inner margin is convex, and furnished with several small irregular granulated teeth. The outer margin of the propodus is convex, with a slight sinuosity toward the distal end of the palm. The palm is as wide as its length on the inner side. The fingers gape when closed ; the tip of the dactylus fits into a sinus on the inner edge of the pollex near the tip. Ambulatory legs smooth. Last three segments sparsely hairy. Dimensions.— M.ix\Q, length 5 mm.; width 5.2; length of propo- dus of cheliped 5.8; width of same 3. Female with eggs, length 5.2; width 6; length of propodus of cheliped 7.5; width of same 4.3 mm. Females are more swollen laterally than males. Color.— In formalin, a light brick red. Terminal half of fingers white. Type locality. — Coral Reef, Maceio, Alagoas ; three males, three females. Cat. No. 23754. Additional localities. — Mamanguape stone reef, one female with eggs; Pernambuco stone reef at Ilha de Nogueira, one male. 148 RATHBUN PACHYCHELES MONILIFER (Dana). Porcellana monilifera Dana, Crust. U. S. Expl. Exped., i, 413, 1852 ; pi, XXVI, fig. 3, 1855. Pachycheles vioniliferus Stimpson, Proc. Acad. Nat. Sci. Phila., x, 228 [66], 1858. Mamanguape stone reef ; one specimen. Pernambuco stone reef at Ilha de Nogueira ; five specimens. Stone reef at Boa Viagem, Pernambuco; nine specimens. Color. — Light yellowish red, deeper on the chelipeds. PORCELLANA ROSEA sp. nov. PL VIII, fig. 3. Length and breadth of carapace subequal. Carapace subtriangular. The front viewed from above has a well-marked median tooth, and between the middle and the orbit two sinuses. The orbit is very small and has a preorbital and a postorbital lobe. The lateral margin is in- terrupted by a slight notch at the cervical suture. Behind the suture there is a well-defined, coarsely granulated margin set off from the carapace by a furrow. This margin at a point far back on the branchial region turns obliquely inward and backward upon the cara- pace. The posterior margin has a deep median indentation. The carapace is very convex in an antero-posterior direction, slightly con- vex from side to side. A deep median furrow leads from the frontal tooth to the mesogastric region, the anterior portion of which is faintly outlined. On either side of the median furrow are two tubercular ele- vations. From the notch anterior to the preorbital lobe a furrow runs obliquely backward and joins the cervical suture, which is ver}^ deep. The cardiac region is sufficiently distinct. A shallow furrow, curving forward, crosses the branchial region transversely. The surface is covered with large punctae bearing short bristles, and the branchial region has scattered irregular tubercles. The merus of the cheliped has a blunt inner distal tooth; distal and outer margins of upper surface bordered by acute tubercles. Carpus with inner and distal margins thickened ; inner margin with a blunt tooth at its proximal third; outer margin tuberculous, a somewhat larger tubercle at the distal third ; upper surface very uneven, tuber- culous and granulous, three tubercles much larger than the others. The outer margin of the propodus is nearly straight for most of its length, and is marked by sharp granules; the inner margin forms a very prominent lobe behind the dactylus; the upper surface has four THE DECAPOD AND STOMATOPOD CRUSTACEA 1 49 blunt longitudinal crests, the innermost one being very short. Each finger has a similar crest ; the fingers meet along their prehensile edges. The lower surface of palm and fingers is also granulous. The second, third and fourth pairs of feet are granulous, the carpal joints each with a large tubercle on the outer side. Dimensions. — Of female with eggs, length and breadth 5.7 mm. ; of immature female (figured), length 4.3, breadth 4.6 mm. Localities. — Mamanguape stone reef, one male, soft shell ; one female, type. Cat. No. 23755. Pernjimbuco stone reef at Ilha de Nogueira ; one female, figured. Color. — In formalin, a light pink, without spots. This species approaches P. sociata Say, but in P. rosea the lat- eral marginal line extends much further back before turning inward ; and the ornamentation of carapace and chelipeds is quite different. MINYOCERUS ANGUSTUS (Dana). Porcellmia an^iista Dana, Crust. U. S. Expl. Exped., i, 423, 1852 ; pi. xxvi, fig. 12, 1855. Minyoccrus angiistus Stimpson, Proc. Acad. Nat. Sci. Phila., x, 229 [67], 1858. PorcelHiia stellicola Muller, Arch. f. Natur., xxviii, part i, 194, pi. vii, 1862. This species was described by Dana from a specimen from Rio de Janeiro, one half as large as the one in hand. Dana's type is not ex- tant. The individual in Dr. Branner's collection is an egg-bearing female. It does not accord in every respect with Dana's description and figure, but allowance has been made for difference in size and possible mutilation. The front between the orbits is two-spined instead of three; of these spines, the one on the left side is much shorter than the one on the right, and apparently was broken previous to the last moult ; it is possible that the median spine was also broken off and not renewed, there being only a very shallow lobe in its place. The outer orbital spines are advanced beyond the frontal sinuses. The eye is inter- mediate in length between its adjacent spines. The right cheliped only is present. The merus has an inner distal spine. The carpus has a strong spine on the inner margin just an- terior to the middle, and two smaller spines at the inner distal angle; the distal margin is spinulous. The palm increases gradually in width from the carpus to the fingers, which are less than one-third the length of the palm ; fingers stout and narrowly gaping. Proc. Wash. Acad. Sci. August, 1900. 150 RATHBUN Dimensions of feinale. — Length, measured from tip of longest frontal spine, 6,7 mm. ; width 4.8 mm. Locality. — jMangroves, Parahyba River. Colo?'. — White, with marblings of orange-brown ; a narrow median stripe of white extends from the posterior margin to a point in advance of the lateral spine. There is an irregular oblong patch of white on each branchial region, and in front of it a smaller transverse patch, and a third still smaller near the lateral spine. MiJller's figure is of a small individual, and shows the frontal teeth shorter, the carpus without spines, the fingers short as in our specimen ; the color marks in general resemble those of our specimen, but in the latter the light median stripe is not continued to the front. Family CALLIANASSIDJE. GLYPTURUS BRANNERI sp. nov. PL VIII, figs. 5-S. Dorsal suture very deep. Front with a short, acute, depressed ros- trum, and on either side above the insertion of the antennae, a shal- lower subacute projection. Eyestalks subtriangular, with convex upper surface, and contiguous nearly to their narrow truncate extremities, which ai'e obscurely dentate. Terminal joint of the peduncle of the antennulae about i| times as long as the penult joint. AntennuljE nearly as long as the carapace. The peduncle of the antenna is more slender than that of the an- tennula ; its penult joint reaches the middle of the terminal joint of the peduncle of the antennula. The flagellum is more than twice as long as that of the antennula. Outer maxillipeds similar to those of G. acanthochirus ; the ischium and merus are wider, the propodus somewhat longer and nar- rower, the dactyl us more slender than in that species. The outer sur- face is pubescent, the inner margin long-hairy. The larger cheliped is finely and inconspicuously dentate on its lower margin. The merus is twice as wide as the ischium. The carpus is nearly as wide as the manus, and twice as wide as long. The palm is nearly as wide as long. The fingers cross when closed. The dac- tylus has three teeth on the prehensile edge. Smaller cheliped with the carpus and manus less than half as wide as in the larger cheliped, and elongate. Both chelipeds have tufts of long hair on the margins. Inner caudal lamella as broad as long. Outer branch of the outer THE DECAPOD AND STOMATOPOD CRUSTACEA l^^i lamella completely overlapping and coalesced with the inner branch, and only slightly shorter. This species can be distinguished from G. acanthochirtis Stimpson and G. grandimanus (Gibbes, as Callianassa) by its non-spinous front and chelipeds and the subequal length of the two branches of the outer caudal lamella. Dimensions. — Length in median line, exclusive of antennte, 52.2 mm.; length of carapace 14.9; length of carpus and propodus to- gether, measured to end of pollex, 18.5 mm. Type.— A male, from Mamanguape stone reef. Cat. No. 23756. UPOGEBIA AFFINIS (Say). Gebiaaffinis Say, Jour. Acad. Nat. Sci. Phila., i, 241, 1818. Smith Vine- yard Sound Report, 549 [255], pi. n, fig. 7, 1874. Upogebia affinis STKBi&mG, Hist. Crust., 185, 1893, Mamanguape stone reef; six specimens. Parahyba River, mangroves ; one specimen. Maceio coral reef; three specimens. Family PANULIRIDiE. PANULIRUS ECHINATUS Smith. Panulirtcs echinatus Smith, Trans. Conn. Acad. Sci., 11, 20, 1869. PococK Jour. Linn. Soc. London, Zool., xx, 516, 1890. Pernambuco ; one male. Carapace, antenuc^, first two segments of abdomen and upper half of meral joints of legs pinkish red ; remainder of abdomen and legs dark blackish blue. No round spots as in P. guttatus iX'^\x€\\\€), which is green with yellow spots. The median spines of the carapace as well as all the prominent spines behind the anterior margin are smaller in echinatus than in guttatus ; there are no rudimentary spines on the antennulary segment, and its two large spines are less divergent and less ascending. The penult segments of antennae and antennulse are considerably shorter in echinattis. Family PEN^IDiE. PEN^US SETIFERUS (Linnaeus). Cancer setiferus Linn^us, Sys. Nat., ed. 12, i, pt. 2, 1054 1767 I^ncEusfluviatilis Say, Jour. Acad. Nat. Sci. Phila., i -36' 181 8 Pencsus setiferus ^l\\.^Y. Edwards, Hist. Nat. Crust., 11, ^i^, 1837. Maceio, on coral reef; one specimen. 152 RATHBUN PARAPEN^US GOODEI Smith. ParapencEits goodei Staitk, Proc. U. S. Nat. Mus., viii, 176, 1885. Maceio, on coral reef; one specimen. XIPHOPENEUS KROYERI (Heller). PencEus kroyeri Heller, SB. Acad. Wiss. Wien., xlv, i Abth., 425, pi. 11, fig. 51, 1862. Xiphopeneus hartlii Smith, Trans. Conn. Acad. Sci., n, 28, pi. i, fig. i, 1869. Xiphopefteus kroyeri Smith, Proc. U. S. Nat. Mus., viii, 188, 1885. Maceio, on coral reef; one specimen. Family ALPHEID^. ALPHEUS HETEROCHy^LIS Say. Alpheus heterochcElis Say, Jour. Acad. Nat. Sci. Phila., i, 243, 181 8. Alpheus heterochelis Herrick, Mem. Nat. Acad. Sci., v, 372, pi. 11 (colored), 1891. Mamanguape stone reef ; eleven specimens. Rio Parahyba do Norte, on mangroves ; three specimens. Rio Goyanna stone reef ; one specimen. Pernambuco stone reef at Ilha de Nogueira ; one specimen. Stone reef at Boa Viagem ; one specimen. Maceio coral reef ; nine specimens. ALPHEUS FORMOSUS Gibbes. Alphetis fornwsus Gi'B'B^s, Proc. Amer. Assoc. Adv. Sci., in, 196 (32), 1850. Alp/ieus poiyi GvtKi-a, in La Sagra's Hist. Cuba, 2nd part, vol. vii, p. xix ; vol. VIII, pi. II, fig. 10, 185^. Natal, Rio Grande do Norte; one specimen. Pernambuco stone reef at Ilha de Nogueira ; one specimen. Maceio coral reef ; six specimens. ALPHEUS CRISTULIFRONS Rathbun, nom. nov. Alpheus obesomanus V0ZOQ.Y., Jour. Linn. Soc. London, Zool., xx, 520, 1890 (nee Dana). Maceio coral reef ; tvv^o specimens. SYNALPHEUS MINUS (Say). Alpheus minus Say, Jour. Acad. Nat. Sci. Phila., i, 245, 1S18. Synalpheus 7ninus CouTiiiRE, Bull. Soc. Entom. France, 1898, No. 8, 190, fig. 4. Maceio coral reef; three specimens. THE DECATOD AND STOMATOPOD CRUSTACEA 1 53 Family HIPPOLYTID^. HIPPO LYSMATA WURDEMANNI (Gibbes). Hippolyte iviirdemaiini Gibbes, Proc. Amer. Assoc. Adv. Sci., in, 197 [33]. 1850. Hippolysviata wurdemanni Kingsley, Proc. Acad. Nat. Sci. Phila., xxx, 1878, 89 [i]. Mamanguape stone reef ; one specimen. HIPPOLYSMATA RHIZOPHOR^ sp. nov. PL VIII, fig. 9. Female. Carapace with a median carina on its anterior fifth. Median spines three, one on the carapace proper, and two on the ros- trum ; extremity of rostrum a slender spine not reaching to the end of the first antennular segment ; no inferior spines. An antennal spine present. Antennular peduncles elongate, the second segment twice as long as the third ; two slender flagella, the upper a little thicker than the lower in its basal portion, or for about 32 segments; the lower one twice as long as the carapace; the upper one shorter, at least in the single specimen in hand. Basal antennal spine short. Scale short, extending half the length of the third antennular segment, broad, extremity rounded ; outer dis- tal spine strong. Peduncle equal to the scale in length; flagellum two-thirds the length of the body. External maxillipeds extending beyond the antennular peduncle by the length of the dactylus and half the propodus. Propodus with a distal spine. Dactylus with a marginal row of spines in pairs, and two or three spines at the extremity. First pair of feet reaching a little beyond the propodus of the max- illiped. Carpus intermediate in length between the palm and the pro- podus. Palm with subparallel margins ; fingers two-thirds as long as palm. The carpus of the second pair of feet, when extended, reaches to the end of the maxilliped. The carpus is 24-jointed. Palm and fingers subequal in length. The sixth segment of the abdomen is one and one-half times as long as the fifth. The telson has four rather large spines forming a rectangle on the dorsal surface ; the extremity is subtruncate, with two long submedian spines and two shorter spines outside of these. The inner caudal lamina is oval- lanceolate and extends behind the telson 154 RATPIBUN to a distance equaling one-third the length of the telson ; the outer lamina is longer and broader than the inner, its distance behind the telson equaling fully one-half the length of the telson. Dijnejisions. — Female : Length from tip of rostrum to tip of telson 37 mm. ; length of carapace and rostrum 9 ; length of rostrum 2.5 mm. Type locality. — Rio Parahyba do Norte, on mangroves ; one ovi- gerous female. Cat. No. 23757. Family PAL^MONID^E. PAL^MON BRACHYLABIS sp. nov. PI. VIII, fig. 10. Female. Dorsal crest arising a little in front of the middle of the carapace, or about two-fifths the distance behind the orbits. Crest slightly ascending for its proximal half, strongly ascending for its distal half. Rostrum overreaching the antennal scale for its distal fourth. Upper margin of crest armed with 9 or 10 teeth ; the proxi- mal 7 or 8 are near together and equidistant ; the next one is about twice the distance from the others ; the last one is near the tip of the rostrum ; 2 or 3 of them are posterior to the orbit. The inferior margin is armed with 3 or 4 teeth, the distal one at some distance from the extremity. Terminal third of rostrum very slender. Upper flagellum of inner antennae with 10 segments united, 14 free. Carpus of second pair of feet reaching a little beyond the rostrum. Carpus as long as the palm and half the fingers. Palm two and a half times as long as the fingers, not swollen. Fingers not gaping. Di?nensio7is. — Female: Length of body 37.2 mm. ; length of cara- pace and rostrum 16.4; length of rostrum 9.4 mm. Type locality. — Mamanguape stone reef ; five specimens. Additio7ial locality. — Rio Parahyba do Noi'te, on mangroves; tsvo specimens. Cat. No. 2375S. The long upturned rostrum with ^ teeth, and the very short fingers of the second pair of feet readily distinguish this species. BITHYNIS ACANTHURUS (Wiegmann). Palamon acant/iurus Wirgmann, Arch. f. Natur., 11, part i, 150, 1836. PalcEvion forceps Milne Edwards, Hist. Nat. Crust., ir, 397, 1837. Palmnon forceps VQ-ayiAKTRiis, Arch, f. Natur., xxxv, part i, 28, pi. 11, fig. 4, 1869. Pernambuco, on mangroves; two specimens. Market at Maceio, Alagoas; said to have come from Lagoa do Norte, a large lake near Maceio. THE DECAPOD AND STOMATOPOD CRUSTACEA I55 UROCARIS LONGICAUDATA Stimpson. brocaris hmQ;icaiidata Stimpson, Proc. Acad. Nat. Sci. Phila., xii, 39 [108], "^1860. KiNGSLEY, Proc. Acad. Nat. Sci. Phila., xxxi, 424, 1879 [1880] • Off Jacuma, Parahyba, 15 feet; one specimen. Order STOMATOPODA. Family CHLORIDELLID^. GONODACTYLUS CHIRAGRA (Fabricius). Sqicilla chiragra Fabricius, Ent. Syst., 11, 513, 1793. Desmarest, Consid. Crust., 251, pi. XLiii, 1825. Gonodactylus chiragra Latreille, Encyc, Meth., Hist. Nat., Entom., x, 473. 1825. Off Jacuma, Parahyba, 15 feet; one specimen. Pernambuco stone reef at Ilha de Nogueira ; three specimens. Maceio coral reef; 31 specimens. PSEUDOSQUILLA OCULATA (Brulle). Squilla oculata Brull^ in Webb and Berthelot, Hist. Nat. Isles Canaries, II, part 2, p. 18, 1840. Pseudosqiiilla ociilatay[.\%vs, Ann. Mag. Nat. Hist. (5), v, no, pi. ill, figs. 3 and 4, 1880. Maceio coral reef ; two specimens. Besides the localities indicated by Bigelow,^ there are specimens in the National Museum from Honolulu and from Clarion Island, off the west coast of Mexico. ' Proc. U. S. Nat. Mus., XVII, 500, 1894. PLATE VIII. Fig. I. Uca ihayeri, $ , | nat. size. 2. Uca thayeri, $ , inner surface of large chela, | nat. size. 3. Porcellana rosea, 9 , 2f nat. size. 4. Pisosonia greeleyi, $, 2f nat. size. 5. Glypturus branneri, $ , front and antennae, if nat. size. 6. Glypturus branneri, $ , caudal extremity, i| nat. size. 7. Glypturus branneri, $ , outer maxilliped, 1 1 nat. size. 8. Glypturus brantieri, $ , large cheliped, i^ nat. size. 9. Hippolysmata rhizophorcs, 9 > ^i ^^t. size. 10. Palcemon brachylabis, 5 , \\ nat. size. (156) Proc. Wash. Acad. Sci. Vol.11. Plate VII Decapod and Stomatopod Crustaceans from Brazil. ^HOKnaco BALTIMORE:. PROCEEDINGS OF THE WASHINGTON ACADEMY OF SCIENCES Vol.. II, pp. 157-159- August 20, 1900. RESULTS OF THE BRANNER-AGASSIZ EXPEDI- TION TO BRAZIL. II. THE ISOPOD CRUSTACEA. By Harriet Richardson. Family CORALLANIDiE. CORALLANA ACUTICAUDA (Miers). Corallana acuticauda Miers, Proc. Zool. Soc. London, 1881, Pt. i, pp. 78-79, pi. VII, fig. 13. Hab. : Maceio coral reef, Alagoas (Branner-Agassiz Expedition) ; Hotspur Bank (S. Atlantic), lat. 17° 32' S., long. 35° 45' W. (Miers). The female of this species was described by Miers. Two speci- mens, a male and female, were collected by the Branner-Agassiz Ex- pedition at Maceio coral reef, Alagoas. The female is identical with Corallana acuticauda as figured and described by Miers. The de- scription of the male follows. Male. — Body convex. Last two thoracic seg- ments and postabdomen not hairy, as characteristic of female. Head transverse, with a prominent median point projecting between the basal joints of the first pair of antennae. On either side of the median point, the anterior margin of the head is deeply excavate for the reception of the much enlarged basal joints of the first pair of an- tenna. The antero-lateral angles of the head are acute, and extend anteriorly nearly as far as the median point. The eyes are large, black, and composed of many facets. At their base they are separated Free. Wash. Acad. Sci. August, 1900. (i57) Fig. J. Head oi Corallana acu- ticatida, $. I5S RICHARDSON by more than half the width of the head ; at their apex by more than a third. The first pair of antennae have the basal joints very much dilated ; they extend nearly to the posterior margin of the head. The second pair of antennae reach the posterior margin of the third thoracic segment . The mandibles are very prominent, the pair much overlapping, with the apical part forming a single cone. First five thoracic segments sub-equal, the last two being somewhat shorter. The first two epimera are not longer than the segments and are rounded posteriorly ; the last four are acute, and produced beyond the segments. The abdominal segments are tuberculate on their posterior margins, with a median longitudinal excavation, ornamented in the median line with low tubercles. The terminal segment is acute at the apex, triangulate, with a median longitudinal excavation. At the base of the segment are four tubercles in a transverse line, the two centre ones being largest. A double tubercle is situated lat- erally on either side. The uropoda have both Fig. 2. Abdomen branches equal in length and not longer than the of Corallana acuh- terminal segment. The inner branch is broad, ' and posteriorly acute ; the outer branch is narrow, and acutely produced. Both branches and also the terminal segment of the body are fringed with long hairs. The male of this species resembles the male of Corallana antillen- sis Hansen, especially in a comparison of the abdomens. They are, however, specifically distinct, as more careful examination has shown. Family BOPYRID^. BOPYRUS ALPHEI sp. nov. A specimen of Bopynis parasitic on Alphens heterochcclis Say, collected on mangroves on the Rio Parahyba do Norte is probabl}^ iden- tical with the Bopyrus mentioned by Fritz Miiller' as being found on a species of Alphens on the coast of Brazil. Fritz IMiiller's species has been recorded as Bopyrus alphei although no description or figure of it has ever been published. The body of the female is broadly oval, its greatest width being equal to its length, broadened anteriorly and narrowed posteriorly. ' Jenaische Zeitschrift, VI, p. 68,1871. THE ISOPOD CRUSTACEA 159 Head widening posteriorly, its antero-hitcral angles being produced into lobes, directed upward. Eyes absent. The thoracic segments have their lateral margins somewhat elevated and contiguous. The abdominal segments are not contiguous along the lateral margins but are separated by broad lateral incisions. The incubatory lamelloi are very wide apart, surrounding a broad open space, the marsupium, on the ventral side of the body. Fig. 3, Bopyrus alphei, 9 ■ Fig. 4. Bopyrus alphei, $ . Pleopoda simple, lamellar, placed on either side of the median line. Color, white. Male, oblong oval. Eyes wanting. Thoracic segments well de- fined. Abdominal segments imperfectly defined, except along the margin. Color white. Type locality. — Mangroves, Rio Parahyba do Norte, Brazil, in the branchial cavity of Alpheus heterochcBlis Say. Cat. No. 23759. PROCEEDINQS OF THE WASHINGTON ACADEMY OF SCIENCES Vol. II, pp. 161-184. August 20, 1900. RESULTS OF THE BRANNER-AGASSIZ EXPEDITION TO BRAZIL. III. THE FISHES. By Charles H. Gilbert. During his connection with the Branner-Agassiz Expedition to Brazil in the summer of 1899, Mr. A. W. Greeley made a small but interesting collection of fishes from that part of the coast lying between Mamanguape and Maceio. Most of the specimens were purchased from fishermen at Pernambuco and Maceio, but a few of the smaller kinds w^ere collected in the tide-pools of the reefs. Of the 85 species obtained, 4 are here described as new, and a number of little known forms are redescribed. The collection is deposited in the Zoological Museum of Stanford University. I take pleasure in expressing my indebtedness to Dr. J. C. Branner, who, as leader of the expedition, encouraged in every way the gathering of biological material. 1. Ginglymostoma cirratum (Gmelin). A female about 12 feet long, obtained at Maceio, was found to con- tain young, many of which were preserved. 2. Mustelus canis (Mitchill). Maceio. 3. Scoliodon terraenovae (Richardson). Maceio. 4. Sphyma zygaena (Linnaeus). Maceio. Proc. Wash. Acad. Sci., August, 1900. (161) 1 62 GILBERT 5. Felichthys marinus (Mitchill). Pernambuco ; Maceio. In two young specimens the pectoral and the dorsal spines are about equal, and are much shorter than the length of the head. The anal fin has 21 rays. The occipital buckler tapers rapidly backward, its posterior half being very narrow and of nearly uniform width. Eigenmann is probably right in referring F. bahiettsis to the syn- onymy of this species. 6. Selenaspis herzbergii (Bloch). Maceio. 7. Aspistor luniscutis (Cuvier & Valenciennes). Pernambuco ; Maceio. 8. Netuma grandicassis (Cuvier & Valenciennes). Maceio. 9. Tetragonopterus bahiensis Steindachner. Two specimens, '$>^ and 70 mm. long, from the stone reef at Mamanguape. T. bahietisis seems to differ from T. lacustris in the slightly larger head and slenderer body. Our specimens agree well with the original description, differing only in having one or two more scales along the lateral line, and in having but two rudimentary anal rays. The humeral and caudal spots are intensely black, with well-defined margins. The former is longitudinally oval, surrounded by a wide lighter area, the posterior portion of which is slightly darkened. The spot on caudal peduncle is widest opposite the base of the outer caudal rays and is continued as a narrow streak to the tips of the median rays. A faint darker band runs forward along the sides, and becomes fainter and narrower anteriorly. The head is contained 3| times in the total length to base of caudal ; depth 2\. Eye 3^ in head. Scales 34 or 35 ,^-, the longitudinal rows counted from origin of dorsal to base of outer ventral ray. The lateral line is complete. D. I, 10 or I, 11; A. II, 25 or II, 29. The maxillary reaches to below the anterior part of the eye. The distance from the tip of the snout to the end of the maxillary is less than that from the tip of the snout to the hinder edge of the pupil. By the aid of a lens, the maxillary in the larger specimen can be seen to have its entire edge finely crenulatcd. BRANNER-AGASSIZ EXPEDITION TO BRAZIL 163 10. Tetragonopterus unilineatus Gill. Two small specimens were obtained among the mangroves near Pernambuco. 11. Ahlia egmontis (Jordan). One specimen, 150 mm. long, from the coral reef near Maceio. It agrees in nearly all respects with the description of the types. The dorsal fin begins behind the origin of the anal a distance equal to the length of the gape of the mouth. The pectorals are broad and rather short, their length slightly less than the length of the snout. The length of the head equals one-fifth the distance from the snout to the anus. The head and trunk are a little shorter than the tail. The teeth are in a single series in each jaw, there being a few stronger ones anteriorly on the head of the vomer. No teeth on the shaft of vomer. The color is rather bright olive, the upper parts being densely dusted with black specks. The only differences of importance be- tween this specimen and the types as described consist in the smaller eye and smaller mouth in the Brazilian specimen. The eye is very small, about 3^ the length of the snout ; but the apparently larger eye in the types may be due to imperfect state of preservation. The smaller mouth is doubtless dependent upon immaturity, the types being 15 inches long. 12. Myrophis punctatus (Liitken). One specimen was obtained at the mouth of the Rio Goyanna. The head is contained 2^ times in the trunk, 5^ in the tail. 13. Lycodontis moringa (Cuvier). Mouth of the Rio Goyanna ; Pernambuco ; coral reef near Maceio. 14. Lycodontis funebris (Ranzani). Stone reef at Pernambuco. 15. Albula vulpes (Linnaeus). Pernambuco ; Maceio. 16. Stolephorus browni (Gmelin). Stone reef at Mamanguape. 17. Lycengraulis grossidens (Cuvier). Pernambuco ; one specimen. Head 4^ ; depth 4. D. II, 14; A. II, 25. Lateral line about 40. There are about 14 large teeth in a series along the middle of the lower jaw. These are rather widely separated, the middle of the in- terspace being occasionally occupied by a smaller tooth. Both toward the angle of the jaw and toward the symphysis these larger teeth are 164 GILBERT replaced by smaller ones continuing the series. The maxillary teeth are more uniform in size and about equal the smallest of those in the lower jaw. Along front and sides of jaw the teeth are slightly in- cur\'ed and are directed downward and backward ; laterally they be- come decidedly retrorse, pointing downward and forward. The anal begins under the middle of the dorsal. The origin of the dorsal is midway between the base of the caudal fin and the middle of the eye. The fins are somewhat mutilated and the scales have fallen. There seems to have been a wide silvery band along the sides, on a level with the eye. 18. Poecilia vivipara Bloch and Schneider. Four specimens from a fresh-water swamp, and from the man- groves, near Pernambuco. D. 6, 7 or 8 ; A. 7 or 8. L. L. 27. Basal half of dorsal dusky, with a narrow black convex cross-bar separating basal region from the distal area; the latter with a wide dusky mai-gin. In the largest specimen the coloration of the side is nearly uniform, without spots. Smaller specimens show a black spot on each side of back, in advance of the origin of the dorsal fin. In the youngest specimens these spots are very conspicuous. 19. Hyporhamphus roberti (Cuvier& Valenciennes). One specimen, 160 mm. long, from Pernambuco, seems to belong here rather than with H. unifasciatus^ as the body is very slender (depth 9 in length) and the beak is very long, its length beyond the premaxillaries being half longer than the distance from tip of upper jaw to opercular margin. Dorsal with 15 rays; anal with 16. 20. Hyporhamphus unifasciatus (Ranzani). Maceio. 21. Mugil curema Cuvier & Valenciennes. Maceio. 22. Mugil trichodon Poey. Stone reef at Mamanguape. 23. Sphyraena barracuda (Walbaum). Maceio. 24. Polydactylus virginicus (Linmuus). Pernambuco ; Ivlaceiu. 25. Myripristis jacobus Cuvier & Valenciennes. One specimen from a coral reef near Maceio. BRANNER-AGASSIZ EXPEDITION TO BRAZIL 1 65 In Jordan and Evermann's ' Fishes of North and Middle America' (Vol. I, p. 846), this species is described as having the head \ the length (to Ifase of caudal), the depth i. These measurements are copied from Giinther's Catalogue of Fishes, Vol. I, p. 19, no note being made of the fact that, as there given, they refer to ^ and ^ of the total length, including the caudal fin. In the specimen at hand the head is ^.and the depth | the length to base of caudal. 26. Holocentrus ascensionis (Osbeck). Mouth of the Rio Goyanna. 27. Upeneus maculatus (Bloch). Maceio. 28. Upeneus dentatus Gill. Maceio. 29. Upeneus caninus new species. One specimen, 200 mm. long, from Fernambuco. Characterized by the large scales, the outer canines in upper jaw, the long hog-like snout, the profusely branched tubes of the lateral line and the coloration. The profile rises in a straight or slightly incurved line from the tip of snout to front of orbit, thence in a well arched curve to front of dorsal, where the depth is greatest. The snout is very long. The distance from front of eye to tip of snout is greater by more than half the diameter of pupil than the distance from front of eye to end of opercular spine. The mouth is nearly horizontal, the maxillary greatly broadened at tip and hooked upward. The distance from tip of snout to the end of the maxillary is contained 2|- times in the length of the head. The teeth in the lower jaw are strong, bluntly conic, in a single series, none of them definitely canine-like. On the middle of one side of the jaw two of the teeth are larger than the others, but this is not true of the other side. The premaxillary teeth are in a single series, similar to those in the lower jaw, those anteriorly more widely spaced than those on the side of the jaw. In front of this series anteriorly are two strong canines on each side, the anterior of which is directed downward, the posterior one curved almost directly backward, paral- lel with the jaw. There are no teeth on the vomer or the palatines. There are three rows of scales on the cheeks, one on the interopercles, three on the opercles. The angle of the preopercle and a portion of the vertical limb have fine cross-ridges, which form minute crenulations at the free margin 1 66 GILBERT of the bone. Gill-rakers 6 -)- 21, this number including five anterior spinigerous rudiments. The longest gill-raker is contained i|- times in the diameter of the pupil. Eye 2^ in snout, 5 in head. The barbels extend to the anterior margin of the first median scale in front of the ventral base, their length 3^^ in length of fish (to base of caudal). Bony interorbital width 2-| in the snout. The dorsal spines are slender, rather high, the second and third about equal, half the length of the head. The ventrals reach the vent, and equal the length of the snout and eye. The pectorals are shorter, and equal snout and half eye. D. VIII-I, 8; A. 11,6; P. 16; V. 1,5. 32 scales in the course of the lateral line, 2i rows above it, 5^ rows below. The tubes of the lateral line are profusely branched, the branches covering the greater part of each scale, there being as many as 12 branches present along the middle of the course of the lateral line. In spirits almost uniform olivaceous, with a remaining trace of red on some of the scales. Two narrow yellow streaks, separated by an interspace slightly less than diameter of pupil, run horizontally from the opercular spine, converging gently backward to meet where they join lateral line on caudal peduncle. 30. Scomberomorus maculatus (Mitchill). Pernambuco ; Maceio. 31. Trichiurus lepturus (Linnaeus). Maceio. 32. Oligoplites saliens (Bloch). A single specimen from Maceio, 235 mm. long, is very close to O. saurus^ but differs from all specimens which I have seen of that species in the deeper body, the larger more oblique mouth, the nar- rower maxillary, the relative size and shape of the suborbital bones, and in having four instead of five detached spines in the first dorsal fin. As these are the characters said to distinguish O. saliens^ I place it provisionally in that species. Abundant material will be needed to determine the status of these two forms. The maxillary reaches the vertical from the posterior edge of the eye; its length is contained if times in the head. The mouth is con- siderably more oblique than in O. saurus, the maxillary is narrower at the tip, and does not curve downward so perceptibly. The man- dible is also slightly slenderer. The tip of the snout is thus higher, being above the upper edge of the pupil, while in O. saunis it is plainly below the upper edge of the pupil. The snout is shorter than the eye. BRANNER-AGASSIZ EXPEDITION TO BRAZIL 167 The lower suborbital bone is a third wider than the one above it, a deep re-entering angle between the two. Such a concavity in the posterior margin of the suborbitals is supposed to characterize a dis- tinct species or subspecies ( O. palomeia) from Lake Maracaibo. It is not probable that this character is of value. In our specimen, none of the suborbitals reach the preopercle. The dorsal spines are but four in number and the anterior rays of the soft dorsal and anal are more elevated than in O. saurus^ making the anterior profile of the fins decidedly falcate. The dorsal seems to have been yellow in life with a large black blotch on the anterior rays. The caudal was yellow, and traces of light yellow still persist on the anal fin. The depth of the body is 3! in the length (34 in O. saurus of equal length). 33. Caranx bartholomaei Cuvier & Valenciennes. Maceio. 34. Caranx hippos (LinnjEus) . Pernambuco. 35. Caranx latus Agassi z. Maceio. 36. Vomer spixii (Swainson). Maceio. Two specimens similar to those from Jamaica reported on by Jordan and Rutter,' and evidently agreeing with the figure and description by Agassiz and Spix which served as Swainson's basis for the species. Our specimens are 210 and 230 mm. long. The head is contained 2| times in the total length to base of caudal, the depth li times. The eye is contained 3f times in the head. The chord of the curved portion of the lateral line is contained i^ times in the straight portion (not " i^ the straight part," as given by Jordan and Rutter) . The pectoral fins are longer, their tips reaching to or nearly to the middle of the straight part of the lateral line. The origin of the anal fin is slightly behind the front of the second dorsal (well in advance of this point in V. setipinnis). The eye is much nearer the anterior profile of the head than the gill opening in V. setipinnis ; about equidistant between the two in V. spixii. 37. Chloroscombus chrysurus (Linnajus). Stone reef at Mamanguape ; coral reef near Maceio. iProc. Acad. Nat. Sci., Phila., 1897, loi. l68 GILBERT Four small specimens, the largest 66 mm. long, agree with C. chiysurtis in all respects in which this species is said to differ from C. ectemirus. Compared with two specimens of C. ecteniirus^ 85 mm. long, from Jamaica (collector, J. S. Robert), these Brazilian specimens show a much deeper body, with a much more strongly arched abdom- inal profile, which rises rapidly toward the snout and also along base of anal fin. The eye is also larger, 2\ in head, and the chord of the curved portion of lateral line is contained if times in the straight por- tion. In the young C. cctenurus referred to, the eye is 3^- in the head, and the chord of the curved portion of the lateral line is contained if times in the straight portion. D. VIII-I, 27; A. II-I, 26. Depth of body 2\ to 2\ in the length. Depth of caudal peduncle i|^ to i^ in its length (the latter measured from base of last dorsal ray to base of first caudal ray). The two species are probably valid, with their ranges overlapping. The young specimens of C ectenurus here referred to are not men- tioned among the types of this species. They are from the same locality as the types, and were sent in by the same collector. They are numbered 4966 in the register of the Zoological Museum of Stan- ford University. 38. Trachinotus falcatus (Linnasus). Maceio. 39. Apogon brasilianus new species. Type 61 mm. long. Mamanguape stone-reef, Brazil. June 23, 1899. A. W. Greeley, collector. Closely related to A. i?nberbis and A. dovil, but without trace of a black spot on caudal peduncle. This spot becomes obscure with age in A. dovii and probably also in A. iinberbis and may be wholly wanting in adults ; but it is always conspicuous in young specimens of the size here described. Total length 62 mm. Length to base of caudal 47 mm. Head (to end of opercular flap) 38 Snout .09 Interorbital width 09 Eye 13 Maxillary 20 Pectoral 24 Ventral 21 Highest (2d) dorsal spine 16 Highest soft dorsal ray 23 Distance from last dorsal spine to first soft ray 08 BRANNER-AGASSIZ EXPEDITION TO BRAZIL 169 Second aual spine 13 lyOugest aual ray 21 Longest ray of upper caudal lobe 32 Middle caudal ray 24 Depth 31 Depth of caudal peduncle 16 J- Length of caudal peduncle from base of last anal ray . . . .25 In outline closely resembling A. dovi'i, with snout acute as in that species, and the lower jaw included. There are broad bands of villi- form teeth in the jaws, and narrow bands (sometimes scarcely more than irregular single series) on the vomer and the palatines. The mouth is oblique, the maxillary reaching a vertical midway between pupil and hinder edge of orbit. The eye is large, i^ times the inter- orbital width, which is equal to the snout. The vertical limb of the preopercle is rigid, its posterior margin finely serrulate ; the horizontal limb is membranous, flexible, with entire margin. The gill-rakers are long and slender, the longest half the diameter of the orbit; there are 3 movable ones on the vertical and 10 or 11 on the horizontal limb, and in addition 3 anteriorly-placed immovable tubercles. The longest dorsal spine is contained i^ times in the longest soft ray. The ventrals reach the vent. The long narrow pectorals slightly overlap the front of the anal. In life, doubtless reddish, sparsely dusted with coarse black specks, which are somewhat more numerous posteriorly ; no indication of a black spot on caudal peduncle. A dark spot on opercle (black in the younger co-type) formed by the close juxtaposition of black specks. The vertical portion of the cheeks and the sides of snout are covered with coarse black specks. In the co-type these are concentrated into a small blotch behind the eye ; those on side of snout suggesting a band such as is often seen in the young of A. dovil. Fins translucent, the caudal and the anterior dorsal and anal rays black-edged. Scales ctenoid, 26 or 27 in the lateral line, 2\ series between the lateral line and the base of the dorsal ; 7 series between the lateral line and the base of the anal. This species may eventually prove to be identical with A. imberbis^ if the young of the latter are ever seen to lack the caudal spot and to possess an opercular spot. Such accounts as we have of the species do not indicate this. The American records of A. i77iberbis are all doubtful, and the species should be omitted from our lists until its occurrence in Amer- ican waters is verified. It seems improbable that this common iSIedi- 170 GILBERT terranean form should have been found at Newport, Rhode Island, and at Fernando da Noronha Island, Brazil. 40. Centropomus undecimalis (Bloch). Maceio. 41. Bodianus fulvus (Linnreus). Maceio. Four specimens, of which three represent the h-nber'' phase, one the brown colored '•punctahis' phase. The latter has a reddish tinge on the lower fins and on the lower half of the sides. It is exti'emely improbable that the red, yellow and brown forms {B. riiber,, B. ful- vus and B. fiuictatiis) represent more than color phases of a single species. 42. Epinephelus adscensionis (Osbeck). Maceio. 43. Promicrops guttatus (Linmeus). Maceio. 44. Alphestes afer (Bloch). Pernambuco. 45. Lutjanus apodus (Walbaum). Maceio. Four specimens, the longest 190 mm. These are darker in color than is usual in this species, three of the specimens being deep red- dish-brown on the head, body, and fins, with a vertical light streak on the marginal half of each scale. The fourth specimen is lighter, but appears faded. The outer ventral ray and the anterior margin of the anal are whitish. The spinous dorsal has a deep maroon terminal bar. A series of small round, or rarely oblong, blue spots crosses the middle of the preorbital below the eye, and extends to, and sometimes across, the middle of the opercle. There are sometimes one or more spots present, belonging to a second series, parallel to the first, and extending from the preopercular notch across the cheek to just above the angle of the mouth. One or more spots may also be present in a line behind the middle of the orbit. All of these spots are distinctly bordered by a blackish ring and do not look like the remains of streaks which were uninterrupted at an earlier age. In all structural details these specimens agree with L. apodiis^ hav- ing a deep body, large scales, an anchor-shaped patch of vomerine teeth, few gill-rakers, and low rounded fins. The caudal is shallowly forked, and the pectoral falcate, the latter reaching the vertical from the front of the anal. The scales above the lateral line run in series BRANNER-AGASSIZ EXrEDITION TO BRAZIL 171 parallel with the latter, but lose their orderly arrangement in char- acteristic fashion below the second dorsal. 46. Lutjanus analis (Cuvier & Valenciennes). Pernambuco. 47. Lutjanus synagris (Linnaeus). Coral reef near Maceio. The snout is marked by two parallel golden streaks, the lower of which curves around the lower border of the eye and extends back- ward across the opercle. A third streak traverses the middle of the cheek and opercle to just above the base of the pectoral. A fourth runs horizontally backward from the upper edge of the maxillary. The color agrees in other respects with current descriptions. 48. Ocyurus chrysurus (Bloch). Pernambuco ; Maceio. 49. Haemulon carbonarium Poey. Pernambuco. 50. Haemulon plumieri (Lacepede). Pernambuco ; coral reef near Maceio. 51. Bathystoma aurolineatum Cuvier & Valenciennes). Maceio. 52. Conodon nobilis (Linnteus). Pernambuco ; Maceio. 53. Brachydeuterus corvinaeformis (Steindachner). Pernambuco ; Maceio. 54. Eucinostomus harengulus Goode & Bean. Mouth of the Rio Goyanna ; Maceio. 55. Gerres rhombeus Cuvier & Valenciennes. Pernambuco ; Maceio. 56. Gerres lineatus (Humboldt) . Gerres brasilianns Cuvier & Valenciennes, Hist. Nat. Poiss., vi, 458. Gerres etnbryx Jordan & St arks, in Jordan and Evermann's Fishes of North and Middle America, p. 1379, 1898. Maceio. A single specimen, 170 mm. long, sustains the opinion expressed by Jordan and Evermann that the characters supposed to distinguish G. brasilia7tus from G. liizeatiis would not be permanent. This speci- men has the second dorsal spine decidedly longer than the third, and has 1 1 horizontal rows of scales between the vent and the lateral line, as in G. lineatus from the Pacific. No other characters have been 172 GILBERT pointed out as separating the two, so I use the oldest name for the spe- cies. Gerres ejnbryx is known from a single very large specimen, the size of which may well account for the slightly longer pectoral. There seems to be no other difference between the two. 57. Cynoscion jamaicensis (Vaillant & Bocourt). Maceio. Two specimens, which agree entirely with those reported on by Jordan and Rutter from Jamacia.^ The species is evidently distinct from C. obliqicatus^ having a greater depth, a larger eye, a longer snout, a shorter anal fin, and biconcave caudal. Its relationship with C. nothus is much closer. I have no specimens of the latter at hand for comparison, but from current de- scription can find nothing to distinguish C. jajnaicensis but a some- what longer head and deeper body. The anal may also be shorter. In C. notJnis^ the lateral line becomes straight under the seventh ray of the second dorsal, not under the seventh dorsal spine, as stated by Jordan and Eigenmann'^ and by Jordan and Evermann.' C acoupa is certainly different, as it has very different proportions and smal- ler fins. 58. Cynoscion virescens Cuvier & Valenciennes. A fine specimen, 50 mm. long, agrees perfectly with Steindachner's account of C. microps. Steindachner is correct in enumerating 55 enlarged scales in the lateral line. Our specimen has 57. It is not clear what can have led Jordan and Eigenmann* and later Jordan and Evermann* to state that the lateral line contains 80 pores. 59. Menticirrus martinicensis (Cuvier & Valenciennes). Maceio. 60. Eupomacentrus fuscus (Cuvier & Valenciennes). Mamanguape. Maceio ; Rio Goyanna. This is certainly identical with E. diencceus Jordan and Rutter. The latter is based on very dark specimens from Jamaica. The dark pigment thus involves the whole axil of the pectorals, appearing in a dark spot below as well as above the base of the fin. In one of our Brazilian specimens, the dark color almost reaches the lower edge of the axil. The ventral fins are subject to much variation in this species, 'See Proc. Acad. Nat. Sci. Phila., 1S97, 114. 2 Report U. vS. Com'r Fish and Fisheries (1SS6), 1889, 357. ^ Fishes of North and Middle America, p. 1406, 1S9S. * Review of the Sciitnidie of America and Europe, Report Com'r Fish and Fisheries for 1S86, p. 362. * Fishes of North and Middle America, p. 1415, 189S. BRANNER-AGASSIZ EXPEDITION TO BRAZIL 1 73 extending to the middle of the anal fin in some of our younger speci- mens. Whether the difference may be sexual is not evident. There is no apparent variation in the length of the axillary scale, which I find to be the same in one of the types of E. die7icceus and in the specimens of E. fuscus with which the authors compared it. The apparent dif- ference in the height of the vertical fins and in the forking of the cau- dal was due to the circumstance that the specimens of E. fusctis which the authors used for comparison had lost the tips of all the vertical fins, these being brittle owing to preservation in strong alcohol. Ten young specimens from the coral reef near Maceio vary greatly in general tint, some being uniformly blackish with the margins of the scales still darker, and all of the fins except the pectorals blackish. Others have the hinder half of the body much lighter, this being in one specimen distinctly yellowish. The spots are as given in current descriptions. 61. Abudefduf saxatilis (Linnaeus). Stone reef at Rio Goyanna, at Mamanguape, and at Pernambuco. 62. Iridic poeyi (Steindachner). Jidis crotaphus Cuv. & Val., Hist. Nat. Poiss., xiii, 395; not of Cuvier, Regne Animal. Iridio kirschii Jordan & Evermann, Fishes of North and Middle America, p. 1598, 1898. One specimen, 145 mm. long, from Pernambuco. The colors in alcohol (after prolonged immersion in formaldehyde) give doubtless but a faint clue to the colors in life. The sides of head and body are dull brownish violet, a line along middle of sides dividing into a darker dorsal and a lighter ventral half. Each scale on the back and sides has the basal half dusky (perhaps blue in life). There ai-e two darker spots between the lateral line and the middle of the sides, one just be- hind the head, another beyond the tip of the pectorals, the latter most conspicuous. These may be the lower ends of dark bars from the back, or they may not have existed in life. The dark spot behind the eye is conspicuous. Behind it, on the uppermost part of the opercle, is a lighter spot, probably brightly colored in life, with a narrow dark crescent above and below it. Below this is a diffuse darker shade on the opercle. A darker shade proceeds from the postocular spot in a curve toward the angle of the mouth, a second similar streak from the corner of the mouth backward across the cheek. The narrow curved streak on the base of the pectoral is very evident. The opercular flap and the axil of the pectoral are greenish. No trace is evident of the blue 174 GILBERT band described by Steindachner, and by Cuvier and Valenciennes, which is said to run obliquely from the opercle in front of the pectoral base to the ventral outline, nor of the second line described by Steindachner from the angle of the mouth to the base of the ventral fins. The ver- tical fins appear yellowish green and the caudal has no trace of the lines converging backward from the basal angles. The dorsal is mar- gined with blue or violet, but exhibits no distinguishable marks at the base of any of the rays except the last two. On each of these is a small black spot extending onto the adjacent part of the back. The anal is more obscurely greenish than the dorsal, and may have shown bluish or violet shades in life. The margin is distinctly violet or blue. A narrow line of the same color is visible on the anterior part of the fin, running horizontally nearer the base than the margin. Below this are traces of the blue spots margined with yellow, as described by Steindachner. This is evidently the /r^'^/Zc \_Julis\ ci'otaphiis oiQwv'xQX and Valen- ciennes and of authors generally, a species common in Brazil and the West Indies. Dr. Jordan has pointed out that in Cuvier's first use of the name crotap/ms it is a S3'nonym of /. radiatus^ and is hence not available for the present species, for which he therefore adopts Stein- dachner's later name poeyi. Subsequently Jordan and Evermann re- vise this judgment and point out certain alleged differences between the common Brazilian form, crotaphus, and Steindachner's description of /. poeyi. The former they now recognize as a distinct species under the name /. kh-schii. The most important difference between the two is stated to be in the size of the eye, which is nearly 3 in snout, 6\ in head in /. poeyi ; i-| in snout, 4|^ in head, in /. kirschii. In the specimen before me, which is 2 inches shorter than the type of /. poeyi^ the eye is 23- in snout, 6 in head, therein agreeing essentially with Steindachner's measurement. A more serious disagreement is in the depth of the body, stated by Steindachner to be 3^;} i^^ tJic total length. In my specimen the depth is 314- in the length to base of caudal. The striking coincidences in the description and the comparatively unim- portant differences lead me to adopt Dr. Jordan's earlier view, identi- fying /. poeyi with /. crotaphus (Cuvier & Valenciennes) and using the former as the earliest available name for the species. 63. Sparisoma frondosum (Cuvier). Pernambuco ; ISIaceio. Closely related to S. r^ibripinne. D. IX, 10; A. II, 9. Lat. L. 25, ^l • Head 4 in length to end of middle caudal rays; depth 3 1 in the BRANNER-AGASSIZ EXPEDITION TO BRAZIL 1 75 same. Teeth as in S. rubriplnne^ distinctly outlined and slightly pro- truding convexly, the marginal teeth in each jaw forming an irregular cutting edge ; no posterior canines. Subocular region and angle of preopercle covered with profusely branching canals. A series of five large scales below the eye, of which 3 are on the cheeks. The ex- posed portion of the eye 5^, the orbit 4I, in head. The tubes of the lateral line branch profusely and cover the scales. This branching occurs sometimes pinnately from a horizontal stem, sometimes palmately from the base of the scales. The primary branches are sparingly forked. The dorsal spines are slightly stiffer than in S. rubriplnne^ but are flexible. The longest equals \ length of head. The caudal is lunate. The pectorals are short, not reaching the vertical from the tips of the ventrals, i|- in head. Color uniform dark purplish, apparently including all the fins ex- cept the caudal, which appears more nearly grayish. The middle caudal rays have a light margin, and the outer ones are alternately light and dark, as in kS". rubi'ipinne. I cannot make out, however, that these are the beginnings of light and dark cross-bars which, in S. rubripinne^ traverse the fin. The species seems to differ from S. rubriplnne in the greater depth of the body, in the more steeply rising anterior profile, which de- scribes an even cvirve, in the shorter, blunter snout, which is but little more than! the head, in the somewhat more complexly branching tubes of the lateral line, and in the color. The preceding notes are based on an adult specimen, 306 mm. long. A young specimen, 145 mm. long, shows a general reddish- gray tint, but is much lighter and more variegated in coloration. The pectoral is distinctly orange-red at base after immersion for some time in formaldehyde and alcohol. The caudal is shallowly concave, with produced points. It shows a distinct lighter margin and is very ir- regularly cross-barred with light and with dark reddish-brown. The dorsal is variegated, as in S, rubriplnne. The tubes of the lateral line are typically branched. Two specimens, each 190 mm. long, are darker and more uniformly vinaceous than the young one noticed. They seem to have been some- what mottled with darker in life. All the fins are dark, of about the same tint as the body. The dorsal shows^some obscure darker mark- ings, without pattern, and the caudal has the irregular dark cross-bars most distinct near the upper and the lower margins. There is a very distinct narrow white border to the caudal. Other fins are uniform. 176 GILBERT In one specimen there is a broad white transverse bar below the chin as in S. rubripinne ; in the other this marking is not evident, and the bars on the caudal are indicated on the outer rays only. 64. Scarus croicensis (Bloch). Two specimens, 93 and 76 mm. long, from the coral reef near Maceio. These show the characteristic lengthwise dark streaks, but not the naiTow silvery lines along the sides of the belly. This may, however, be due to their preservation in formaldehyde, which dissolves silvery pigment. G^. Chaetodon ocellatus (Bloch). Mamangviape. 66. Chaetodon striatus Linnaeus. Pernambuco. 67. Teuthis bahianus (Castelnau). Mamanguape ; mouth of the Rio Goyanna ; Pernambuco. In one specimen from Pernambuco, 160 mm. long, there is a very short dorsal fin, with but 20 soft rays. In the same specimen the anal formula is III, 23. 68. Balistes vetula Linnaeus. Maceio. 69. Lagocephalus laevigatus (Linnaeus). Maceio. 70. Spheroides testudineus (Linnaeus). Pernambuco ; Maceio. 71. Spheroides greeleyi new species. Two specimens from the coral reef near Maceio, ii3 and 105 mm. long, the larger taken as the type of the species. Very close to S. spc?ig-leri, from which it differs principally in col- oration, as it entirely lacks the conspicuous series of black spots along the lower edge of the sides, which is so characteristic of S. spengleri and its near allies. The space between the bony orbital ridges is narrow, concave, its least width \\ in the exposed portion of the eye, 4 in the snout. Eye 5 in head; snout half head. Width of gill-slit equaling exposed por- tion of the eye. Upper caudal rays longest, equaling distance from tip of snout to posterior edge of pupil. Dorsal with 8 rays, its height contained 2| times in head. Anal with 7 rays, its height contained 3 times in head. Pectorals broad, the width of the base contained \\ times in the longest ray, which is 2^ in head. Pectoral rays 15. BRANNER-AGASSIZ EXPEDITION TO BRAZIL 1 77 Prickles cover the belly from the chin back to the anal fin and up on the sides as far as the base of the pectoral fins. They cover also the interorbital area and extend back as far as the origin of the dorsal fin, and send downward under the terminal portion of the pectoral fin a band connecting the prickly areas of the back and the belly. Along the line separating the ventral prickles from the smooth area of the sides is a series of conspicuous fleshy slips, all of which are white. A few smaller, dark-colored slips are scattered over the back and sides. Caudal peduncle and side of head without prickles. Belly and lower side of head and caudal peduncle white. Back and sides thickly spotted with blackish, the two areas sharply dis- tinguished along a lengthwise line running just below the pectoral fin. The spots grow somewhat larger along the lower part of the cheeks and the flanks, but none can be taken to represent the definite line of spots present in S. spe77glcri. The upper part of the back has its darker color broken up into polygonal areas by finely vermiculating light lines ; these faintly-defined areas are still further indented or broken up into very small spots by incursions of light lines. The caudal is dusky at the base and on the terminal half, but has no well- defined bars. Other fins are unmarked. 73. Prionotus punctatus (Bloch). One specimen, 210 mm. long, from Pernambuco. It is perhaps as well to follow Cuvier's identification of this species as the very doubtful Trigla punctata of Bloch, awaiting the discov- ery of a West Indian species more nearly agreeing in color with Bloch's figure. We have compared our specimen with the young individual from Bahia reported on by Jordan,^ and find substantial agreement except that in our adult the minute spine present in the young on the middle of the cheek has been lost, and the groove behind the eye is still less conspicuous. The latter is in fact no groove at all, merely a wider interval between transverse lines of granulation. The series of very fine serrations on the lateral margin of the snout terminates in a single stronger spine directed backward. A single spine is present halfway between the latter and the middle of the cheek. The center of radiation of the ridges on the cheek is without spine, a sharp low ridge beginning at that point and continuing on to the preopercular spine, which has a definite cusp at the base. The preopercular spine extends to or slightly beyond the subopercular margin. The nuchal ridge is obsolete on one side, very low on the ' Proc. U. S. Nat. Museum, 1S90, 328. Proc. Wash. Acad. Sci., August, 1900. 178 GILBERT other (these being higher and sharper, ending in spines in the young specimen). Spines otherwise as described. Mouth large, 2^ in head ; interorbital space narrow, deeply concave, 1^ in eye. Gill-rakers slender, 1+9 movable ones, the longest i di- ameter of orbit. The anal has 11 rays in both individuals (not 12, as given by Jordan and Evermann, Fishes of North America, p. 2169). The color in spirits is brownish above, whitish below, the upper parts appearing faintly blotched and with obscure round spots. Two dusky blotches on spinous dorsal. Soft dorsal with roundish dusky spots arranged in about 10 oblique series. Caudal with four ill-de- fined dark cross-bars. Pectorals blackish, the upper rays with a black blotch at the end of the basal third, and a fainter one at the end of the second third of their length. These are also present in the young specimen, and seem responsible for the statement that the pectorals are barred. The younger specimen shows a distinct blue line along_ the lower margin of the pectoral, as well as some lighter markings on the upper part of the fin, which do not appear in our adult. There are 53 pores in the lateral line, and about 98 vertical series of scales above the lateral line. In commenting on a specimen in the Paris Museum of Natural History, labeled Trigla puiictata^ "appar- ently in the handwriting of Valenciennes," Dr. Jordan states^ "pores in lateral line 85 to 90." Unless reference is made to the vertical series of scales rather than to the pores, the specimen cannot belong with this speci?s. 73. Cephalacanthus volitans (Linnaeus). Maceio. 74. Dormitator maculatus (Bloch). From among the mangroves at Pernambuco. 75. Guavina guavina (Cuvier & Valenciennes). Among the mangroves at Pernambuco. 76. Gobius soporator (Cuvier & Valenciennes). Mouth of Rio Goyanna ; Mamanguape. 77. Echeneis naucrates (Linnasus). Maceio. 78. Malacanthus plumieri (Bloch). Maceio. This species had been previously reported from Brazil by Cuvier and Valenciennes and by Castelnau. Our specimen answers well to the detailed description given by the first named authors. Jordan and 'Sec Proc. U. S. Nat. Museum, 1SS6, 545. BRANNER-AGASSIZ EXPEDITION TO BRAZIL 1 79 Evermann are in error in ascribing to this species but 49 articulated ravs in the dorsal fin. Cuvier and Valenciennes give the dorsal for- mula as VI, 55, Gunther gives it as VI, 59. Our specimen has the dorsal VI, 59, the anal 55. The bright colors which are said to characterize this species in the fresh state are mostly lost in spirits. Our specimen still shows traces of the alternating blue and yellow streaks before the eye. The dorsal and caudal are bright yellow, the dorsal dusky at base. The lower mai-gin of the caudal and the lower half of the upper lobe of the fin are black. The anal has a more reddish tinge, and the pectorals and ventrals are bluish or violet. 79. Malacoctenus delalandi (Cuvier & Valenciennes). Mouth of the Rio Goyanna. So. Labrisomus nuchipinnus (Qiioy & Gainard). Stone reef at Mamanguape; coral reef near Maceio. Jordan and Evermann are in error in stating that Labrisomus has no palatine teeth, and in the further statement that Z. nuchipinnis has the vomer provided "with a patch of smallish teeth." ^ The facts had been stated by GilP with substantial correctness and in consid- erable detail. I have examined numerous specimens from Brazil, two from Sta. Lucia Island, one from Jamaica, and two from the Canary Islands. The teeth are strong, conic, rather widely spaced, arranged typically in a single angulated series on the vomer and on the anterior portion of the palatines. The anterior vomerine tooth is often enlarged, but in some specimens this is scarcely noticeable. The palatine teeth are usually in small number, there being often not more than two present, sometimes as many as six in the single series. Oc- casionally an inner tooth may be present along the inner side of the palatine series, or there may be two or three, thus forming a double series. The inner teeth are, when present, entirely similar to those of the outer row. This variation has been found in specimens from Brazil and from Sta. Lucia. In the two specimens from the Canary Islands two equally well developed series are present on both palatines, suggesting a permanent condition which may properly be investigated. In the Canary Island specimens there is also present a number of vomerine teeth behind those In the outer row. In two specimens only from the western Atlantic, from Brazil and from Sta. Lucia, respectively, have I found a single vomerine tooth occupying this inner position. >See Fishes of North and Middle America, Vol. Ill, pp. 2361-2362, 1898. ^Proc. Acad. Nat. Sci. Phila., 1S60, p. 106. l8o GILBERT Z. 7i7ichipi7inis and L. xanti do not appear to differ in dentition. In six specimens o£ L. xanti from Mazatlan, Mexico, five have the palatine teeth in tv^^o distinct series, as in the Canary Island specimens of L. nuchipin7iis ; in the sixth no palatine teeth were present. In the specimens of L. xajiti examined, the vomer contained a single series in all but one individual, where a few teeth were present be- hind the outer row. The middle vomerine teeth are usually the largest. L. xanti seems, however, to be sufficiently distinguished by its slightly more elongate form, and by one constant difference in colora- tion. The conspicuous round, black, opercular spot, universally present in L. luicJiipinnis^ is never developed in L. xanti. Our specimens of L. 7iuchipinnis from Brazil and elsewhere, show that this spot is normally suiTounded by a white border, though this is sometimes wholly lacking. BRANNERELLA new genus (^Blenniidce'). Type, Bran7icrella brasilie7isis new species. Closely allied to Starksia Jordan and Evermann,^ differing only in the elongate detached first anal spine and in the absence of any well- defined notch behind the third dorsal spine. 8i. Brannerella brasiliensis new species. Type, a specimen 31 mm. long, from the coral reef near Maceio. D. XVIII, III, 7; A. I-I, 16. Lat. 1, 37 or 39. Form and general appearance much as in Aiic/ie7iopte7'zcs, but the head is larger than is usual in that genus. The mouth is large and very slightly oblique, the maxillary reaching to the vertical from the posterior margin of the orbit. The jaws are equal, the lips arc thickened, the upper jaw is protractile. The pre- maxillaries have an outer close-set series of cardiform teeth and an inner band of villiform teeth. The teeth on the mandible are similar, but those on the outer series are slightly coarser and more conical. A single series of small conical teeth on the vomer and the front of the palatine. Interorbital space ver}' narrow, concave. The posterior nostrils are each in a short tube, which bears a slender cirrus. A similar cirrus on the upper posterior part of the eye, and one on each side of the median line of tlie nape. No nuchal fringe of filaments. The series of pores on tlic head arc well developed. The branchiostegal rays are 'Fishes of North and Middle America, p. 2365, BRANNER-AGASSIZ EXPEDITION TO BRAZIL l8l 6 in number, the membranes broadly joined, free from the isthmus. The opercle is rounded, without ridges or spines. Scales large, cycloid, the lateral line running high anteriorly, de- scending to the middle of the flanks behind the pectoral fin, and trav- ersing 37 to 39 scales. Dorsal largely spinous, the last 7 rays articulated. A faint notch or none behind the third dorsal spine. A decided notch behind the eigh- teenth spine, the three following spines gradually increasing in length. Two anal spines, the first elongate, detached, running along the posterior margin of a broad membranous or fleshy expansion, which extends beyond its tip. The second spine is much shorter and is joined by membrane in the usual way to the first soft ray. Both dor- sal and anal fins are distinct from the caudal, the last dorsal mem- brane nearly touching the base of the first caudal ray. No hook on shoulder girdle. Color in spirits, dark brown, with some lighter rivulations and with occasional black spots, or blotches. A small black spot on the oper- cular membrane above the base of the pectoral. Back with ten dusky bars which are continued on to the basal portion of the dorsal fin, where each divides to form a pair of black blotches; 7 or 8 dark spots along the base of the anal. An irregular light-colored Y-shaped mark on the cheek ; two or three small round spots on the opercle, and three vertically placed on the basal portion of the pectoral fin. All of these light markings may have been brightly colored in life. Length to base of caudal 26>^ mm. Total length 31 mm. Head (to end of opercular flap) 34 of length. Tip of snout to end of maxillary 16 Snout 06^ Orbit ... 09>^ Interorbital width o2}4. Depth of body 2i>^ Least depth of caudal peduncle 09 Length of caudal peduncle from last anal ray . .06 Tip of snout to base of first dorsal ray ..... .30 Tip of snout to base of first anal ray 53 Tip of snout to base of first ventrals 28 Longest pectoral ray 23 Longest (inner) ventral ray 18 Highest (middle) dorsal spines 10 Shortest ( 1 8th) dorsal spines 06 (19th) dorsal spines 06 -f- (20th) dorsal spines 07^ (21st) dorsal spines 10 l82 GILBERT Highest dorsal ray 15 1st anal spine (including soft tip) 10^2 2nd anal spine o6}4 Highest anal ray 12,^2 Caudal 25 82. Blennius cristatus Linnaeus. Numerous specimens from the stone reef at the mouth of the Rio Goyanna, at Maiiianguape, and at Pernambuco. The following notes are principally corrections of statements in the description by Jordan and Everman.^ Depth of preorbital slightly more than | diameter of orbit. The interorbital space is slightly grooved, f the diameter of the eye. The posterior canine in the lower jaw is much longer than the front teeth, but is declined. The supraocular flap is divided from the base into 4 or 5 filaments. The filaments on the nape are mostly in a linear series surmounting the dermal crest, but occasionally some are scattered at the sides of the crest ; the number varies greatly, from 10 to 30 in our specimens. The dorsal formula is XII, 14 or 15. The anal, II, 16. No white spots are visible in this material, possibly because of its preservation in formaldehyde. 83. Rupiscartes atlanticus (Cuvier & Valenciennes). Coral reef near Maceio. The three specimens in the collection have a narrow light margin to the dorsal. In one specimen the light area of the pectoral is decidedly yellow. 84. Salariichthys textilis (Quoy & Gaimard). Mouth of Rio Goyanna ; stone reef at Mamanguape and at Per- nambuco. 85. Syacium micrurum Ranzani. Two adult males, 330 and 335 mm. long, and a young specimen, 130 mm. long, from Maceio. D. 88, 88, 89; A. 69, 70, 72. The adults have the interorbital space very wide, shallowly con- cave, the front of the upper eye in a vertical falling slightly behind the middle of the lower eye. Interorbital width ^ greater than the horizontal diameter of the lower eye; it equals the length of the snout and is contained 4^ times in the length of the head. The gill- rakers are short and broad, scarcely -^ the diameter of the pupil, 3 + 7 in number. (I find the same number in a specimen from Jamaica.) The two upper pectoral rays are elongate, equally so in one male, ' Fishes of North and Middle America, pp. 2382-2383. BRANNER-AGASSIZ EXPEDITION TO BRAZIL 1 83 the lower very little produced in the others. The longest ray extends beyond the middle of the anal, and is contained 2^ times in the length. The color is light grayish-brown, finely mottled and with a number of rather definitely placed round light spots with darker centers. The most conspicuous of these form a series of five or six parallel with the dorsal outline, and a similar series parallel with the ventral out- line. Others occupy the intermediate territory. A narrow curved dark bar extends from the upper eye obliquely across the interorbital space to the lower eye, bounded by a light line in front and behind ; another dark line and another light line may follow these. From the front of the upper eye, a narrow scaleless area runs downward and forward to the profile of the snout. This area is bounded above and below by narrow, dark lines, which were probably dark blue in life. The upper line ceases opposite the origin of the dorsal ; the lower ex- tends down along the profile to the tip of the snout, and is continued along the mandible to its joint. A dark streak — also probably blue in life — runs along the back at the base of the anterior ^ of the dorsal fin, on the colored side. This streak is formed of spots, one for each ray, the hinder ones more or less coalescent. The dorsal and anal fins are grayish, speckled with darker, a dark streak on the base of each eighth to tenth ray. Pectorals not barred. The blue streaks on the head have not been mentioned in any previous description of this species. They are doubtless developed only in adult males. I find a trace of them in a specimen from Jamaica (collector, Roberts). PLATE IX. Fig. I. Brannerella brasiliensis gen. and sp. nov. Type. Maceio, Brazil. 2. Upeneus caninus sp. nov. Type. Pernambuco, Brazil. 3. Apogon brasilianus sp. nov. Type. Mamanguape, Brazil. 4. Spheroides greeleyi sp. nov. Type. Maceio, Brazil. (1S4) Proc. Wash. Acad. Sci. Vol.11. Plate IX. New Pishes from Brazil. PROCEEDINGS OF THE WASHNGTON ACADEMY OF SCIENCES Vol. II, pp. 1S5-201. August 20, 1900. RESULTS OF THE BRANNER-AGASSIZ EXPEDITION TO BRAZIL. IV. TWO CHARACTERISTIC GEOLOGIC SECTIONS ON THE NORTHEAST COAST OF BRAZIL. By John C. Branner. ■» CONTENTS : Notes on Geology along the Baliia and Minas Railway 185 Notes on Geology along the Alagoas Railway 195 Conclusion 201 NOTES ON GEOLOGY ALONG THE BAHIA AND MINAS RAILWAY. In September, 1899, •"• "^^^e two trips over the Bahia and Minas Railway for the purpose of getting some idea' of the geology of the region it crosses. The following observations were made upon those trips, and, notwithstanding their obvious imperfections, are here published as a contribution to the geol- ogy of a little-known part of Brazil. Although railway geology is proverbially untrustworthy, this bit was done under circum- stances that redeem it somewhat. The Director of the rail- way,^ whose guest I was at the time, kindly had the train stopped wherever and whenever I wished, not to mention many other stops at stations, for inspection, and on account of tree trunks and telegraph poles fallen across the line. In order that the geology might be better seen the entire run was made ^ I am under obligations for many courtesies in connection with this trip to Dr. Joao Bley Filho, of Theophilo Ottoni, the able Director of the Bahia and Minas Railway. Proc. Wash. Acad. Sci., August, 1900. (185) i86 BRANNER very slowly and during daylight. The only other geological observations we have upon the region traversed by this railway are those made by Hartt, who went from Santa Clara to Theo- philo Ottoni in 1866, by the newly opened road across the country, some leagues south of the railway and parallel with it save near its upper end, where both the road and the railway are in the valley of the Rio Todos OS Santos.^ Hartt speaks of the crystalline rocks as gneiss veined with granite. The Bahia and Minas Railway leaves the sea coast at Ponta d'Areia, the terminal station, near the city of Caravellas, in the State of Bahia, and runs nearly due west to the city of Theophilo Ottoni (formerly called Philadelphia) in the State of Minas Geraes, a distance of 376 kilo- meters. The region traversed consists of three well - defined belts : a flat coastal plain, a belt of table lands, partly forest covered and partly campos, and a region of crystalline rocks. Along the first ten kilometers the line passes over a flat, sandy, coastal plain that rises less than three meters above high tide level. The highest point measured over this coastal belt is near the village of Barra de Caravellas, where it is 2.2 meters MAP SHOWINCTHE POSITION OF THE □AHIA E MlNAa RAILWAY eiV 9&aitnetit3. Fig. 5. Coast of Brazil from Rio to Bahia. ' Geology and Physical Geography of Brazil, by C. F. Hartt ; pp. 129-134. Boston, 1870. BRANNER-AGASSIZ EXPEDITION TO BRAZIL 187 above high tide level. It is possible that there is a little higher ground at the city of Caravellas. At Ponta d'Areia the highest ground is onl}^ one meter above high tide. Changes of level over the surface of the plain are usually either very gentle or very abrupt, the slopes being long, even, and imperceptible on one side and dropping off abruptly on the other. The soil of this coastal belt is all sandy. At the surface it is blackened by the decay of vegetation, but from two to six decimeters beneath the surface this sand is of various shades of yellow — sometimes almost orange colored. It is worthy of note in this connection that the great sand bars off the mouth of the Caravellas River have about the same yellow color. At many places where pits have been dug for wells in these sands of the coast belt, large quantities of marine shells have been thrown out, while at other places ants and other burrowing animals bring these shells or their fragments to the surface. At two places I found pieces of coral in the sand from such pits. One of these was a frag- ment of HeliastrcBa a-perta as big as one's fist, a coral growing abundantly on the great coral reefs off the coast. The other form was a small specimen of the little Astrangia solitaria. Oysters and other bivalves, worm tubes, and gasteropods are abundant in this sand. The molluscan remains found in the sands of this coast plain look very like the forms now living along the coast, but in the absence of collections of both the fossils and of the existing coast fauna no opinion can be ex- pressed in regard to the precise geologic age of the coast plain. It may be either Pliocene, Pleistocene, or recent. Beyond kilometer ten the railway passes gradually from the low sandy coast plain into open sandy campos. The loose sands along this portion of the line are white and have the ap- pearance of having been leached by acidulated waters. The country is nearly flat and only a little higher than the coast plain. The only rocks visible are soft black sandstones exposed in the trenches beside the railway track. Similar sandstones are found at many places along the northeast coast of Brazil, always at a low level, usually black to snuff-colored, and wherever their relation to the adjacent rocks is apparent, they rest unconformably against and upon the eroded Eocene (?) 1 88 BRANNER beds. At the city of Caravellas such sandstone is used occa- sionally in retaining walls. I was told that it was all brought in canoes from several miles up the estuarj^, but that it could be taken out only when the tide was low. No fossils have ever been found in these black sandstones. On the line of the rail- way the black sandstones are found here and there from kilo- meter 10 to about kilometer i8. At kilometer 19 the railwa}'" makes a three or four meter cut in a low ridge of light red sandy clay, which I take to be the approximate eastern margin of the Eocene (?) sedi- ments. From this point until after the Serra dos Aymores is passed the road passes over beds of more or less similar ma- terials. These beds are referred with some doubt to the Eocene upon evidence collected farther north in the States of Bahia, Sergipe, Alagoas, Pernambuco, etc. Along the line of the Bahia and Minas railway no fossiliferous beds have been found, but in general appearance and structural relations this series corresponds to the weathered portions of the Eocene ( ? ) . In this region the Eocene ( ?) beds are brown, yellow, pink, red, of various shades of gray, and mottled. They are mostly soft sandstones or sand clays. At a few places only do these sandstones form hard rocks : these are on the grade between Peruhype and Helvetia, about kilometer 68 ; in the vicinity of kilometer 136.5, near the Pasto de Godinho ; and about the caixa d'agua, at kilometer 141. 5, where the sandstone is from four to six meters thick. At man}^ places the upper parts of the beds contain irregular lumps of iron or of sand cemented with iron, about the size of two fists. These sedimentary beds are so nearly horizontal that no dip is apparent anywhere along the line of the road. The railway elevations, however, show that they have a gentle coastward dip. The Eocene (?) beds continue to and beyond the Serra dos Aymores, ending at kilo- meter 160. The Eocene (?) area forms extensive open campos or prairies, bare of trees, covered with a sparse and stunted vegetation, but in places dotted with small patches or ' islands ' or forests. Over other portions of the Eocene (?) belt are also some of the finest forests I have seen in any part of Brazil. The great forests begin about kilometer 73 and extend almost BRANNER-AGASSIZ EXPEDITION TO BRAZIL 189 to Theophilo Ottoni, kilometer 376. A large number of splendid rosewood and mahogany logs lie decaying along the line of the railway, while the forests contain vast quantities of these valuable woods. Peruhype station (kil. 65) is on Rio Peruhype, at an elevation of about only six meters above tide level, while the bottom of the river, where the railway crosses the stream a third of a kilometer below the station, is only half a meter above high tide level. The railway has, up to this point, made a wide detour to avoid the marshes about Caravellas, and after crossing the Eocene ( ?) hills at an elevation of sixty meters has come down almost to tide level again. The hills along the Rio Peruhype are about thirty meters high, flat on top, and slope down to the stream at an angle of about 35°. Beyond Peruhype station the railway climbs again to the top of the sedimentary plateau. At kilometer 67 a cut exposes well the usual horizontal, mot- tled and colored beds to a depth of about ten meters. At kilometer 69 the road is again in the open campos on top of the plateau, and from here the campos continues to about kilometer 73, where the railway enters the great forests. The top of this plateau about Helvetia station (kil. 72-73) is more or less rolling, but it is really a flat table-land crossed by gulches and narrow valleys that are visible only when one comes close to them. At kilometer 128 the railway descends into a small valley cut in the Eocene (?) beds. The beds exposed in the railway cuts are the usual yellow, red, and mottled clays ; some of them contain lumps hardened with iron, half a meter or more in diameter. At kilometer 132.5 the road reaches the bottom of a small narrow val- bt •fT*. •..■>, / ■'i' f ^7H %' ■'J; ;4 s n TT 4 11 / ^ ,'"^ >> \ ca . ;'.; ■^ V s" ^t to ca \' r n $^ 1^ N-a i i-S '""f i nJ 2 /^ I^ !3 cS V li pq "^Ia ii- 4i/,p 0 ^'% » (U ^ .9 ■' '^ m ^ a \l 13 T^ ' a .2 \i 2 (U ^•gr^ N fl:'.A vd m 0 A I'-'S "^'\P s . ^:«> >» •,;- ^ \ :^ _1 -Tr ^ ^ V?: /-•8 '.•i< .-?> '#. foj iLi _£ ^ •^ i!^">f'i o\ A^ * Vm^I ^ '^^if ^ R ii i< .* % 190 BRANNER ley at a place known as the Plasto de Godinho. Here the drainage has cut through the sedimentary beds and exposed beneath them a few small rounded bosses of the old under- lying crystalline rocks. These are exposed over a distance of only about one kilometer. The railway continues beyond this over the Eocene (?) rocks. The top of the grade in passing over the Serra dos Aymores is near kilometer 155.^ For several kilometers along the watershed of the Serra dos Aymores there are no rock exposures, and it is impossible to determine just where the sedimentary rocks come to an end. At kilometer 160, however, the crystalline rocks appear in the bottom of a gulch, and as boulders of sandstone are here mingled with blocks of quartz-monzonite, it seems probable that this place is close to the margin of the Eocene ( ?) sediments, and that the Serra dos Aymores, where it is crossed by the railway, is capped by the thin landward margin of these sediments. The country is covered with a dense forest, and only occasional glimpses can be had of the topography of the surrounding region. But these glimpses are suggestive. At one point (kil. 145) one sees to the right of the road an almost perpendicular and very sharp peak rising high above the surrounding forests. At another point another lofty peak is visible from the railway, also north of the line. These mountains all have the same rounded and exfoli- ated appearance so characteristic of peaks of crystalline rocks in Brazil. The Serra dos Aymores is the northward extension of the Serra do Mar along this part of the Brazilian coast. Seen from the ocean it stands out boldly as a great mountain chain, extending from about S. lat. 16° to about S. lat. 21° 30'. It is therefore a matter of interest to find the railway crossing this serra upon horizontal sediments. But the explanation seems to be that while these Eocene (?) beds abut against the main range of the Serra dos ' Unfortunately the elevations along the railway were not to be found when I was at the Director's office. In the ^Mcinoria descriptoya sohre a estrada de Ferro Bahia e Minas por M. de Teire e Ars^ollo,'' Rio de Janeiro, 18S3, I found the elevations up to kilometer 13S, while the elevation of Theophilo Ottoni is given in Hartt's book. Inasmuch as the road follows the river from where it first reaches the Mucurj-, the profile is represented as even from that place to Theophilo Ottoni. BRANNER-AGASSIZ EXPEDITION TO BRAZIL I9I Aymores, at this particular place the heart of the Serra is rep- resented by only a few isolated peaks that rose above the water as islands when the sediments were being deposited. The route followed by the railway is, therefore, a sort of gap in the range, where the ancient serra is represented only by these tall peaks. From kilometer i6o, where the railway descends the west slope of the Serra, the exposures of crystalline rocks become more and more abundant until, in the narrow valley from the watershed to the Rio Mucury, the Eocene (?) beds dis- appear entirely, and the hills on either side are all of crystalline rocks, which continue nearly to the city of Theophilo Ottoni. Professor H. W. Turner, of the U. S. Geological Survey, has kindly examined the rocks collected on this trip ; the one found at this place he calls qtiartz-monzojitte, adding that it is a rock that most petrographers would call a gabbro.^ At kilometer 171 the railway reaches the banks of the Rio Mucury. This stream throughout nearly its entire length flows over a bed of coarsely crystalline rocks, mostly of gabbro, or quartz-monzonite. The accompanying illustration (fig. 7) shows the round, exfoliated peak near the watering station, Sete de Setembro, near kilometer 172. This peak itself was not vis- ited, but the many exposures along the railway near its base are all of quartz-monzonite. The immediate valley of the Mucury is here hardly more than half a kilometer wide in the bottoms, and the hills rise on both sides to a height of a hundred meters or less. Through its entire length this valley remains narrow and its sides are steep and heavily wooded ; only near its head-waters do the valleys widen out somewhat. At kilometer 188 the railway crosses to the right side of the river, upon which it continues the 1 The following is Mr. Turner's description : F- 93- "Quartz-monzonite. Macroscopically a medium-grained, dark, slightly greenish granolite. "Microscopically composed of plagioclase >■ orthoclase > ^ quartz > ^ rhombic pyroxene p> biotite. There are present as accessories iron oxide, apatite and zenotime. The plagioclase, quartz, and pyroxene occur in comparatively small anhedrons, in part inclosed in large untwinned orthoclase grains which show minute interpositions, forming the ' faserige ' orthoclase of Becke, The biotite is deep reddish-brown in color, and frequently all the foils in the field extinguish at once, although disconnected in the section." 192 BRANNER rest of the way to Theophilo Ottoni. At Maj-rink station the rocks are quartz-monzo- nites and coarse-grained gneiss. In the bed of the stream there are good exposures showing the banding of the rocks to strike N. 25° E. From where the railwa}' enters the narrow Mu- cury valley at kilometer 172 to kilometer 345 the rocks are all either quartz-monzonites, gne- isses, or granites, the quartz-monzonites pre- dominating. I was un- able to learn anything of the relative ages of these cr\'stalline rocks. The bed of the Rio Mucury is everywhere of these massive rocks. Here and there hills of the bare, rounded rocks are exposed in the sides of the narrow river chan- nel, and enormous ex- foliated boulders are abundant throujrhout the area of crystalline rocks. There are some fine ex- amples of these boulders at Bias Fortes, kilometer 30S. In this lower por- tion of the crystalline BRANNER-AGASSIZ EXPEDITION TO BRAZIL 1 93 belt the rocks have a much fresher appearance than the}- have further inland. Specimens taken at kilometer 281 and at Francisco Sa station, kilometer 290, are described by Mr. Turner as quartz-monzonites.^ At Bias Fortes the rocks are quartz-monzonites again. ^ At kilometer 344 and at kilometer 363 bare peaks of exfoli- ated crystalline rock are visible on the right of the railway 100 to 150 meters above the valley. About kilometer 345 there is a marked geological change. The quartz-monzonites disap- pear altogether, the rocks are no longer coarse in texture and fresh in appearance, but are fine-grained gneisses, and pro- foundly decomposed. At kilometer 355 the change is still more marked, for from this point to Theophilo Ottoni the rocks are mostly mica and other schists, much faulted, wrinkled and cut by veins. All these rocks are so decomposed that one sel- dom sees a hard rock face in the many and deep railway cuts. "Within seventy-five meters of the terminal station at Theophilo Ottoni is a nearly perpendicular face ten meters high, cut in faulted and decomposed schists filled with small quartz veins. The city of Theophilo Ottoni is in a very hilly country, and the streets of the town follow narrow valleys between the hills. Here and there deep cuts have been made through the lower slopes of these hills for street grades or to make room for houses. Everywhere these cuts are in rock, decayed in place, and the earth is so soft that a knife can be thrust into it any- where. This earth, however, stands up in perpendicular w^alls like stone or brick walls through all kinds of tropical weather. ^ "F. 79. Quartz-monzonite. Macroscopically a coarse, dark, granitic rock, composed of large feldspars of an olive-green lint, quartz, and black mica. "Microscopically composed of orthoclase in large anhedrons, micropegma- tite, plagioclase, quartz, biotite, and brown amphibole. There are present as accessories apatite, iron-oxide, pyrrhotite, titanite, zircon, and muscovite. The plagioclase gives extinctions on 010 up to 16° and is probably chiefly basic andesine and labradorite. " 2 Mr. Turner says of them: "F, 83. Quartz monzonite. Microscopically composed of plagioclase > orthoclase and micropegmatite > quartz >■ rhom- bic p3'roxene. There are present as accessories iron oxide and apatite. The plagioclase gives extinctions on trace of 010 up to a maximum of 19°, and is probably chiefly basic andesine or labradorite. The rhombic pyroxene is slightly pleochroic, ab, faint reddish, c, light green." 194 BRANNER The soils of the surrounding country are of various shades of red and yellow. Before making the trip into Minas along this railway I found one dav in the vicinity of the terminal station at Ponta d'Areia, mingled with pieces of the quartz-monzonite so abundant along the upper Mucury, two fragments of blue limestone about the size of the two fists. These fragments contained the following fossils: Lingala umbonata Cox, Orthetes sp., pieces of cri- noids, one piece of a Pleurotomaria^ and several fragments of shells belonging to the Rhynchoncllacca. Most of these forms, so far as they can be identified, range from the Silurian to the upper coal measures, while the Lingula is found only in the lower and upper coal measures. I made diligent inquiry in regard to the origin of these pieces of limestone and was told that such rocks were common all along the Rio Mucury above Sete de Setembro. It was, therefore, largely in the hope of finding these fossiliferous limestones in place that the trip was made to Thephilo Ottoni. After what has been said of the geology of the region it is hardly necessary to state that these fossiliferous beds were not found, although the most careful lookout for them was kept up from one end of the road to the other, in the natural exposures, in the railway cuts, and in the ballast of the road bed. It may be added also that no other fragments of similar material were found in the stone paved streets and stone walls of the old city of Caravellas, or else- where about Ponta d'Areia or Barra de Caravellas. Of course it cannot be positively stated that these fossils do not come from the Mucury River, but I do not think now that they were found there originally. My opinion is that the fragments in w^hich they were found came to Ponta d 'Areia upon some vessel from some other port. I v/as told of limestone burned for lime west of Theophilo Ottoni, in the belt of decayed schists, but these rocks appear to be marbles. No fossils have been found in this Brazilian schist series, but it has generally been regarded as older Paleozoic, possibly Silurian. In regard to the age of the quartz-monzonites and the other fresher looking crystalline rocks nearer the coast, I have no facts upon which to base an opinion. BRANNER-AGASSIZ EXPEDITION TO BRAZIL 95 NOTES ON THE GEOLOGY ALONG THE ALAGOAS RAILWAY, STATE OF ALAGOAS. The part of the Alagoas Railway of greatest geologic interest is the section between Maceio and Albuquerque. This portion of the line was passed over three times in August, 1899. Mr. Fig. 8. Map of the Alagoas Railway, Brazil. H. Haynes, the able superintendent of the railway, to whom I am under obligations for many courtesies, kindly placed a loco- 196 BRANNER motive and special car at my disposal on one trip from Maceio to Assemblea, and had them stopped whenever the geology made it desirable. As a rule the geology of the region of metamorphic and crystalline rocks is complicated, and in notes made as these were but little else can be done than to indicate the margin of the area where the crystalline rocks are overlapped by the later sedimentary beds. The Lagoa do Norte is of interest in connection with the geography and later geology of the region. This lake, although large, is so shallow that only very small boats can navigate it. It has a depth of channel of only two or three meters along its south side. The water is brackish, and the tides are but little felt in the lake ; the rise and fall is only about 0.3 meter. This lake and Lagoa Manguaba, a similar body of water eight kilometers southwest of it, were valleys carved in the sediments of this region in Miocene Tertiary times, when the continent stood at a higher level. After these valleys were ex- cavated the region sank and the valleys were converted into bays. These ba3^s, in the course of time, had their mouths closed by the silts thrown back into them by the ocean, and still later they silted up as we now find them. Newly made land has already filled up the stream valle3^s emptying into these basins, and the land is constantly encroaching upon the lakes. They are rapidly silting up and must disappear soon, in the geological sense. Plate X, made from a photograph, is a view looking south across Lagoa do Norte from Fernao Velho, and showing the Eocene (?) plateau and its steep oceanward escarpment south of the town of Coqueiro Secco. The Maceio end of the railway is upon the low, flat, sandy lands upon which stands the business portion of the cit}^ known as Jaragua. Behind this low ground rises a line of flat-topped hills — the margin of a plateau — of parti-colored sediments sup- posed to be of Eocene Tertiary age. These hills follow the coast northeast for hundreds of miles, broken here and there by the valleys about the mouths of streams. From the lighthouse on top of the hills at Maceio this sedimentary plain appears to pROC. Wash Acad. Sci, Vol. Plate X. BRANNER-AGASSIZ EXPEDITION TO BRAZIL 1 97 be almost perfectly flat, but away to the north and northeast one sees three clusters of rounded hills rising above it and resem- blinor the granites that everywhere form the landward margin of these diments. These are the peaks of the Serra de Mariquita. Mr. Haynes informed me that a few years ago at the railway shops on the flat ground near the terminal station at Maceio a driven well w^as put down to the depth of 200 meters. The water obtained was brackish and the well was abandoned. No detailed record was kept of the strata passed through in sinking this well, but the materials are all soft and appear to be the newer beds that abut against the strata which form the high plateau of the upper city and the surrounding country. On leaving Maceio station the railway follows the strip of low ground between the north side of the Lagoa do Norte and the plateau. In some places the road bed has the steep hill rising on the north and the water of the lake on the south, while at others there is a wide strip of mangrove swamp be- tween the lake and the hills, and at still others there are strips of arable land on these flats. Wherever rocks are exposed along the north side of the railway between Maceio and Caihoeira they are the yellow, red, purple, gray, white, and mottled beds so characteristic of the weathered portions of the Eocene (?) along this part of the Brazilian coast. These Eocene ( ?) hills have slopes as high as 45°, and in places they are even steeper. At Bebedouro station (kil. 6) a stream from the north enters the lake through a flat-bottomed, steep- sided valley. At Fernao Velho (kil. 14) there are good exposures of the parti-colored beds near the station, and half a kilometer be- yond the station a cut shows the horizontality of these beds fairly well ; but they are all more less oxidized, even to the lowest ones exposed. At the Instituto Archeologico e Geographico Alagoano in Maceio, I was shown some fossil fishes said to have been found at the town of Fernao Velho. These fossils are in hard gray or cream-colored lumps or concretions of limestones. Lith- ologically the rocks bear a strong resemblance to the Ceara limestone that contains the fossil fishes, and very little to most 1 98 BRANNER of the rocks exposed in the railway cuts and about the town. I have some doubt about these fossils having come from Fernao Velho, and yet it is not impossible that in the alteration and oxidation of such beds, the more resisting lumps might be left in this form. This seems all the more probable, too, because I found fossil fishes on the coast at Riacho Doce, only twelve kilometers north of Maceio, in beds that stratigraphically seem to be the same as these at Fernao Velho. Shortly after passing Fernao Velho the railway leaves the lake shore and follows up the valley of Rio Mundahu. At several places heavy water-worn boulders are exposed in the railway cuts, sometimes rising as much as six or seven meters above the track. At Usina Leao (kil. 23-24) the granite boulders exposed in the cut by the reservoir are too large to have traveled far. A kilometer or more east of Utinga station horizontal gray shales are exposed beside the railway track. These shales have a general lithologic resemblance to the Eocene (?) shales of Riacho Doce. They were not closely examined for fossils, but they look promising. Between this exposure and the Utinga station (kil. 26.6) are large water-worn blocks of granite : similar blocks appear a hundred meters north of the station. These boulders have the appearance of being recent deposits, but they are probably either the basal conglomerate beds of the Eocene (?) series, or else loose materials lately washed from such a basal conglomerate. The material is too heavy to have been moved far from its origrinal source. About kilometer 29 the Rio Mundahu flows close to the foot of high hills known as the Serra de Ouro, and the roadbed of the railway has been cut in the steep face of the mountain close to the river. These cuts show almost throughout, and for a distance of two or three kilometers heavy beds of decayed granitic boulders. I was unable to determine certainly whether these boulder beds were formed by the Rio Mundahu, or were the decayed basal conglomerates encroached upon by the stream. Less than one kilometer south of the Cachoeira station these heavy conglomerates are exposed again in the railway cuts, with a thickness of about ten meters. At and immediately Proc. Wash, Acad Sci, Vol Plate XI. Deep cut near Bittencourt, on the Alagoas railway, exposing sedimentary beds resting against granite BRANNER-AGASSIZ EXPEDITION TO BRAZIL 199 north of Cachoeira station the road cuts decayed crystalline rocks (granite), and winds up a granite hill past the Cachoeira waterfalls of Rio - Mundahu to Albuquerque station. These falls are over granite rocks. From Albuquerque west along the Assemblea branch of the railway the line is upon granites or closely related crystalline rocks all the way to Assemblea. At Albu- querque the western branch line leaves Rio Mundahu and crosses the watershed into the drainage of Rio Satuba, then passes over a hiijh divide and descends into the basin of the Rio Parah3^ba. Between Albuquerque and Bittencourt there are several cuts in decayed granite and in places large exfoliated granite blocks. Bittencourt station is on a flat-looking granite plateau at the top of the watershed and has an elevation of 148.8 meters. The view toward the northwest from the station is remarkably tine. Immediately after passing this station the railway descends rapidly into the Para- hyba valley. On this grade are many cuts in red and yellow soft materials. Some of these cuts are twenty meters or more in depth, and |-v-%j y,\[ th the sides stand at remarkably high angles. Several of them half way down the grade exhibit well-marked stratification of the mate- rials. Plate XI shows the bedding in one of these deep cuts.^ Some of the beds contain unconsolidated, horizontal layers of white quartz boulders from 0.3 to i meter in thick- ness. These beds appear to rest upon and against granites, and are overlain by clays that exhibit but little or no evidence of stratification. In places these clays are as much as fifteen meters deep. These strati- fied beds appear to be a remnant of the same sediments as those exposed along the railway between Maceio and Cachoeira. ' This plate was kindly furnished me by Mr. Ambler, of the Alagoas Railway. 200 BRANNER At this Bittencourt cut they are somewhat higher than the basal beds near Cachoeira, but this due to the Bittencourt beds being higher than in the series. From Parahyba (or Capella) station (kil. 38+), looking north and northeast, one sees the mountains in the distance, while be- low and to the east of them the lower hills appear to be planed off so as to make an even sky-line. It looks as if this were an old base level of erosion, possibly formed at the time of the dep- osition of the sedimentary beds along this coast. The granites are well exposed in the Parahyba River at Capella, and at the railway station. The river in many places flows over a rocky bed of granite. Four or five kilometers below Gamelleira the granite in the river bed has a more or less sheeted appearance. At Gamelleira station, to the east of the line, is an exfoliated granite peak about 150 to 200 meters above the railway, or about 350 meters above tide level. The angle of the slope of this peak is from 50° to 85°. Between Gamelleira and Vicosa the railway passes through a gap in the main range of the Serra dos Dois Irmaos. Plate XII shows a view of this gap in the granite hills. Here the Rio Parahyba has a beautiful fall and a series of cataracts, wdth some fine examples of pot-holes in the granite bed of the stream. The channel is strewn in places with exfoliated granite slabs. Vigosa (or Assemblea), the terminal station of this branch line, is upon granites, ninety-seven kilometers from Maceio, and has an elevation of 214 meters. The surrounding hills are of granite and have an elevation of a hundred meters or more above the valley. These hills are mostly forest-covered, but one of them exposes a bare rock surface. From Albuquerque the main line of the Alagoas Railway con- tinues northward up the Rio Mundahi'i to Imperatriz (now called Uniao). This portion of the line is all upon the old crystalline rocks, and a brief reconnaissance of the region offers but little of variety or especial interest. At Bom Jardim (kil. 44.7) the valley of the Mundahii opens broadly, and the granite hills are only from fifty to sixty meters high. Near the station the rocks are decayed to a depth of about two meters. At Muricy station (kil. 64.3) the valley of the Mundahu is somewhat broader; the Proc Wash Acad. Sci. Vol. Plate XII. BRANNER-AGASSIZ EXPEDITION TO BRAZIL 20I rocks are gneiss and granite. The hills are about lOO meters high (184 above tide) and they all appear to rise to about this general level. This flatness of the hilltops continues to and beyond Nicho (kil. 71), and is apparent again as one looks southward and eastward from near Uniao. This seems to be the region of flat-topped hills seen from Capella on the Assem- blea branch of this same railway. About Nicho are many exposures of granite bosses, and up stream from the station there are many exposures of granite in the bed of the river. Granites and gneisses continue past Branquinha to Uniao, exposed on the slopes of the mountains, in the railway cuts, and in the bed of the river. At Uniao (kil. 88, elevation 106 meters) the Alagoas Railway connects with the government road running northward to Palmares and Pernambuco, and the crystalline rocks continue to its junction at Glycerio with the Sul de Pernambuco line. CONCLUSION. These two sections, although nearly a thousand kilometers apart, are essentially the same : a series of Mesozoic beds along the coast lapping back over old crystalline rocks. The question of the age of the sediments cannot be discussed in this place : it is too large a problem to be treated briefly. Without going into detail, it is enough to say that the evidence seems to point to these conclusions : I. The Bahia basin, formerly referred to the Cretaceous, is probably either Eocene, Tertiary, or Laramie. II. The parti-colored beds along the coast, formerly referred provisionally to the Tertiary, are the same as the Bahia Eocene. III. The sediments of the Alagoas section are of fresh-water origin, like those of Bahia. IV. No fossils have been found in the section along the Bahia and Minas Railway, but it seems probable that these beds are the southward continuation of the Bahia beds. Proc. Wash. Acad. Sci., August, 1900. PROCEEDINGS OF THE WASHINGTON ACADEMY OF SCIENCES Vol. II, pp. 203-246. August 20, 1900. MAMMALS COLLECTED BY DR. W. L. ABBOTT ON ISLANDS IN THE NORTH CHINA SEA. By Gerrit S. Miller, Jr. CONTENTS : Introduction 203 CORRECTION. Page 203. In title, and again in 3d line of 2d paragraph, for North read South. though numbering about 200 specimens, form only a part of the very extensive collections obtained, all of which have been presented to the United States National Museum. Leaving Singapore on June 29, 1899, accompanied by Mr. C. B. Kloss, Dr. Abbott spent about three months among the islands of the North China Sea. Mammals were collected on the following islands, the location of which is shown on the accompanying map (fig. i) : Linga (July 7th to 25th), Ste. Barbe (August ist and 2d), Tambelan group (August 13th to 15th), Anambas group (August 17th to September 28th), and Tioman (September 30th to October 4th). So far as it is represented by this collection the mammal fauna of the islands consists of local forms of the widely distributed and Proc. Wash. Acad. Sci., August, 1900. (203) PROCKEDINGS OF THE WASHINGTON ACADEMY OF SCIENCES Vol. II, pp. 203-246. August 20, 1900. MAMMALS COLLECTED BY DR. W. L. ABBOTT ON ISLANDS IN THE NORTH CHINA SEA. By Gerrit S. Miller, Jr. CONTENTS : Introduction 203 Systematic list of species 241 Notes on the fauna of the islands visited 242 INTRODUCTION. During the spring of 1899 Dr. W. L. Abbott fitted out a schooner at Singapore for the special purpose of exploring the less known coasts and islands of the East Indies. The results of the first cruise, so far as concerns the mammals, are embodied in the present paper, which is published here by permission of the Secretary of the Smithsonian Institution. The mammals, though numbering about 200 specimens, form only a part of the very extensive collections obtained, all of which have been presented to the United States National Museum. Leaving Singapore on June 29, 1899, accompanied by Mr., C. B. Kloss, Dr. Abbott spent about three months among the islands of the North China Sea. Mammals were collected on the following islands, the location of which is shown on the accompanying map (fig. i) : Linga (July 7th to 25th), Ste. Barbe (August ist and 2d), Tambelan group (August 13th to 15th), Anambas group (August 17th to September 28th), and Tioman (September 30th to October 4th). So far as it is represented by this collection the mammal fauna of the islands consists of local forms of the widely distributed and Proc. Wash. Acad. Sci., August, 1900. (203) 204 MILLER characteristic Malayan types. Each island and group of islands has its representatives of the common genera and species ; but 'C) Nalunas Anambas 't Fig. io. Islands visited by Dr. Abbott. in scarcely an instance is an insular race identical with that occurring on another island, unless of the same group, or on the mainland. SYSTEMATIC LIST OF SPECIES. MUS FLAVIVENTER sp. nov. Type. — Adult male (skin and skull). No. 101739 United States National Museum. From Puki Jimaja, Anambas Islands. Collected September 18, 1899. Characters. — Similar to AIus cremor/venter ^^UWcr^ from Trong, Lower Siam, but slightly larger and with entire pelage much more strongly suffused with yellow. J^ier. — The elements of the fur are as in J/, crcmoriventcr. Color. — General color above ochraceous, fading to orange-buff or buff-yellow on sides, the sides nearly clear, but back, shoulders, neck and head uniformly sprinkled with black-tipped hairs and bristles, which are nowhere in excess of the ochraceous. Cheeks clear orange buff. Muzzle broccoli-brown. Scarcely a trace of a dark eye ring. ' Proc. Biolog. Soc. Washington, XIII, p. 144, April 21, J900. MAMMALS COLLECTED BY DR. W. L. ABBOTT 205 Underparts and inner surface of legs pale straw-yellow to base of hairs, sharply defined from color of sides and extending to wrists and nearly to ankles. Feet yellowish white, faintly shaded with sepia. Tail and naked ears uniform dark brown throughout. Tail as in Mus cre?noriventer. Skull. — The skull closely resembles that of Afiis cronoriventer., differing only in slightly larger size and relatively smaller audital bullje, characters that may prove to be purely individual. Teeth as in Mus cremorivcnter. Measurements. — External measurements of type : total length 304 ; head and body 133; tail vertebrae 171; hind foot 28. 5 (26.5); ear from meatus 16; ear from crown 11. 5; width of ear 14. A second specimen, also a male: total length 285; head and body 133; tail vertebrae 152; hind foot 27 (25).^ Cranial measurements of type : greatest length 34 ; basal length 29 ; basilar length 26.8 ; palatal length 14.4; least width of palate be- tween anterior molars 3.8; diastema 9.4; length of incisive foramen 5.4; combined breadth of incisive foramina 3 ; length of nasals 13.6; combined breadth of nasals 3.8; zygomatic breadth 15.4; interorbital breadth 5.6; mastoid breadth 12.8; breadth of braincase over roots of zygomata 14; depth of braincase at front of basioccipital 10.4; frontopalatal depth at posterior extremity of nasals '].'$>\ least depth of rostrum immediately behind incisors 6.4 ; maxillary toothrow (alveoli) 5.8; width of first upper molar 1.6; mandible 17.4; mandi- bular toothrow (alveoli) 5.6. Specimens examined. — Two, from the type locality. Re?narks. — Mus Jlaviventer is a yellowish, insular form of M. O'etnoriventei'. It is not closely related to any of the species thus far recorded from the Natuna Islands, unless to the Mus ephippiutJi of Thomas. MUS ANAMB^ sp. nov. Type. — Adult female (skin and skull) . No. 1 01 737 United States National Museum. From Fulo Jimaja, Anambas Islands. Collected September 21, 1899. Characters. — Like Mus surifer Miller^ from Trong, Lower Siam, but with tail always much shorter than head and body. Coior. — Upper parts uniform tawny-ochraceous, heavily sprinkled with blackish brown on posterior half of back, less so on shoulders ^Measurement of hind foot in parenthesis is taken without claws. ''Proc. Biolog. Soc Washington, XIII, p. 148, April 21, 1900. 206 MILLER and head. Sides, flanks, cheeks and outer surface of legs clear tawny- ochraceous. Underparts soiled yellowish white to base of hairs. The white extends down inner side of front legs to wrists, but on hind legs, as in Miis szirrfer, it is usually interrupted between knee and heel. Feet dull white. Ears and dorsal surface of tail to terminal third or fourth dark brown. Underside of tail and whole of terminal third or fourth dull white. Underfur of back light slate color. Skull and teeth as in Mus surifer. Measureme7its. — Three adults measure as follows : No. 101737, 9 , Pulo Jimaja (type); total length 330; head and body 178; tail vertebrcB 152; hind foot 40 (39); ear from meatus 21; ear from crown 17; width of ear 16. No. 101738, ^, Pulo Jimaja; total length 330; head and body 1.78; tail vertebrae 152; hind foot 40 (39). No. 101701, 9 , Pulo Siantan; total length 343; head and body 184; tail vertebrcB 159; hind foot 41 (39). Specime7is exa?nifzed. — Five, all from the Anambas; two taken on Pulo Jimaja and three on Pulo Siantan. Remarks. — Though closely related to Afus surifer this species is readily distinguishable by its much shorter tail. MUS LINGENSIS sp. nov. Type. — Adult male (skin and skull). No. 101614 United States National Museum. From Linga Island. Collected July 15, 1899. Characters. — Closely related to Mus surifer and M. pcllax ; size and cranial characters as in the former, color more nearly like that of latter. White of inner side of hind leg broadly continuous with that of dorsal surface of foot. Bony palate narrower than in the allied species. Color. — Dorsal surface an indescribable grizzle of dark hair-brown, clay color, and drab, darkest on middle of back, lightest and most tinged with yellow on neck. Sides drab, slightly varied with hair- brown and clay color. Ventral surface of body and inner side of legs white tinged with cream color. The white of the hind leg is broadly continuous with that of dorsal surface of foot. A distinct blackish eye ring. Tail bicolor, whitish below and at tip, elsewhere dark brown. Ski4ll and teeth. — The skull closely resembles that of ]\/7(s surifer except in the form of the bony palate. This is more elongate than in either of the related mainland species or in ]\f. ana??ib(€. The ratio of width of palate between anterior molars to distance between pos- terior border of incisive foramen and anterior border of interpterygoid' space is 42 and 43 respectively in two specimens of Mus lingensisy MAMMALS COLLECTED BY DR. W. L. ABBOTT 207 while in two of M. surifcr it is 57 and 60, in the type of j\[. anatnbce 60, and in the t3'pe of AI. pellax 55. Teeth as in Mus surifcr. Meas7iretnents. — External measurements of type : total length 387 ; head and body 216; tail vertebrae 171 ; hind foot 42 (40.4) ; ear from meatus 21; ear from crown 16; width of ear 16. Average of five adults: total length 381 (362-400); head and body 215 (203-229); tail vertebrae 169 (159-1S4); hind foot 40.2 (38-43); hind foot without claws 38.7 (37~42)* Cranial measurements of type : greatest length 47.5 ; basal length 40.4; basilar length 37.5 ; palatal length 21.8; least width of palate between anterior molars 4.4 ; diastema 13.4; length of incisive fora- men 7 ; combined breadth of incisive foramina 4 ; length of nasal 19; combined breadth of nasals 5.4; zygomatic breadth 20; in- terorbital breadth 6.8; mastoid breadth 15.4; breadth of braincase over roots of zygomata 17 ; depth of braincase at front of basioccipital 1 1.4; frontopalatal depth at posterior extremity of nasals 11; least depth of rostrum immediately behind incisors 8.8 ; maxillary tooth- row (alveoli) 8; width of first upper molar 2.2; mandible 25.4; mandibular toothrow (alveoli) 7.2. Specime?zs examined. — Six, all taken on Linga Island, at an alti- tude of about 1000 feet. Remarks. — The dull color of the six specimens of j\Ius lingensis is evidently due in part to the abraded condition of the pelage in most of the individuals taken. After due allowance is made for this, how- ever, the differences separating the animal from Mus surifer are still very appreciable. MUS STREPITANS sp. nov. Type. — Adult female (skin and skull). No. 101697 United States National Museum. From Pulo Siantan, Anambas Island. Collected September 10, 1899. Characters, — Similar to Mus vociferansyVxW^x^ from Trong, Lower Siam, but color darker throughout, and ventral surface pale yellow in- stead of dull white. Skull as in the Bornean Mus sabatius. Color. — Back and sides ochraceous-buff, duller on back, clearer on sides, everywhere sprinkled with blackish brown. The brown pre- dominates on back and shoulders, gradually giving way on sides to the ochraceous-buff which, however, is nowhere without at least a faint sprinkling of brown. Cheeks and outer side of front legs nearly clear ochraceous-buff. Entire ventral surface and inner side of front legs ' Proc. Biolog. Soc. Washington, XIII, p. 138, April 21, 1900. 208 ' MILLER pale straw-yellow, brighter on chest, duller and more nearly ci'eam color posteriorly. The yellow of legs is continuous with dirty white of upper surface of feet. Metapodials heavily shaded with brown. Tail dull light brown, indistinctly paler beneath and throughout ter- minal third. Skull and teeth. — The skull differs from that oi Jlfus vociferans and resembles that of M. sabanus in the elevation of the region about posterior extremity of nasals. The character is one that is almost im- possible to describe, but which is readily appreciated on comparison of specimens. Teeth as in the related species. Measuroiients. — External measurements of type : total length 565; head and body 241 ; tail vertebra 324; hind foot 47 (43.6) ; ear from meatus 22; ear from crown 20; width of ear iS. Four speci- mens from Pulo Siantan average : total length 546 (509-5S4) ; head and body 235 (222-248); tail vertebrte 311 (29S-336) ; hind foot 46.4 (45-47); hind foot without claws 44. 1 (43-45). Three speci- mens from Pulo Jimaja average : total length 551 (547-559); head and body 233 (229-241) ; tail vertebrae 31S ; hind foot 45.5 (45-46) ; hind foot without claws 44 (43-45). Cranial measurements of type : greatest length 57 ; basal length 48 ; basilar length 45; palatal length 25.4; least width of palate between anterior molars 5; diastema 15.6; length of incisive foramen 8; com- bined breadth of incisive foramina 4; length of nasals 11.8; com- bined breadth of nasals 6 ; zygomatic breadth 25 ; interorbital breadth 9.6; mastoid breadth 19; breadth of braincase over roots of zygomata 20; depth of braincase at front of basioccipital 13.S; frontopalatal depth at posterior extremity of nasals 14 ; least depth of rostrum immediately behind incisors 10; maxillary toothrow (alveoli) 10; width of first upper molar 3; mandible 32.4; mandibular toothrow (alveoli) 9.6. Specimens cxa?nined. — Seven; four from Pulo Siantan and three from Pulo Jimaja. Remarks. — Its dark color immediately distinguishes this rat from JMus vociferatis and AI. sabanus. The seven specimens are remark- ably constant in all their characters, both external and cranial. None of them show a tendency to become markedly paler and yellower across the shoulders so conspicuous in Mus vociferans. Dr. Abbott notes that the animal is quite as noisy as its mainland representative. ?MUS VOCIFERANS Miller. A single much al)radcd specimen ( 9 No. 101756) taken on Tionian Island, October 3, 1S99, is probably referable to this species. Its MAMMALS COLLECTED BY DR. W. L. ARBOTT 209 tail is defective and the pelage is so worn that the inidcrfur appears at the surface in irregular patches. Under the circumstances any exact determination is impossible. In its whitish belly and in the form of the skull the animal agrees with Mzis vociferans and shows no approach toward the characters of M. strcpitans. MUS TIOMANICUS sp. nov. Type. — Adult male (skin and skull). No. 101763 United States National' Museum. From Tioman Island. Collected October 4, 1899. Characters. — In size, proportions, quality of fur, and cranial char- acters similar to the form of ]\Ius ' alexaizdrimts ' occurring in Trong, Lower Siara, but color throughout much darker. Color. — Dorsal surface raw umber heavily and finely grizzled with blackish brown, the fur faintly lustrous in certain lights. Sides and cheeks slightly paler than back and much less profusely sprinkled with dark hairs. Ventral surface dull white, the chin and throat tinged with cream-buff, the chest faintly washed with brown. Throughout the pelage except on chin and throat the hairs are light slate-gray at base. Outer surface of front legs drab-gray. Inner surface of all four legs like belly. Muzzle and dorsal surface of feet dull sepia. Tail uniform dark brown throughout. Skull and teeth. — In both cranial and dental characters this species is practically identical with the form of Miis alexandrinus found in Fig. II. a Mus tiomanictis. b M. siantaniciis. c M. tamhelanicus.' Natural size. Trong, Lower Siam. The skull as a whole is, however, slightly more elongate than in the mainland animal, the difference most noticeable 2IO MILLER in the region of the anterior zygomatic roots, and in the posterior por- tion of the braincase. The antero-posterior diameter of the inter- parietal is usually greater than in the mainland form ; though this character is not wholly constant it is sufficiently so to differentiate series of specimens. Measurements. — External measurements of the type: total length 362; head and body 171 ; tail vertebrae 191 ; hind foot 37 (35) ; ear from meatus 20.4; ear from crown 15.8; width of ear 15.6. Average of four adults from the type locality: total length 375 (363-381); head and body 1S4 (171-197); tail vertebrae 189' (184-197) ; hind foot 36.5 (34-38) ; hind foot without claws 34.5 (33-35)' Cranial measurements of type: greatest length 43 (41); 'basal length 37 (36) ; basilar length 34 (33.6) ; palatal length 20 (19.4) ; least width of palate between anterior molars 4.4 (4.4) ; diastema 1 1.6 (11.4) ; length of incisive foramen 7.8 (7-6) ; combined breadth of incisive foramina 3 (3) ; length of nasals 15.4 (14.4) ; combined breadth of nasals 4.4 (3) ; zygomatic breadth 19 (20) ; interorbital breadth 6 (6.4) ; mastoid breadth 16.4 (16.6) ; breadth of braincase over roots of zygomata 15.4(16.4); depth of braincase at front of basioccipital 11.6 (11); frontopalatal depth at posterior extremity of nasals 9.4 (9.6) ; least depth of rostrum immediately behind incisors 7.8 (7.8) ; maxillary toothrow (alveoli) 7 (7.6) ; width of first upper molar 2 (2) ; mandible 33.6 (22.4) ; mandibular toothrow (alveoli) 7(7)- _ Specimens examined. — Elev^en, all from the type locality. Remarks. — On comparison of the eleven specimens of Al7is tio- ma}iiciis with ten of the continental race of Mus ' alexandrinns ' from Trong the differences in color are seen to be remarkably con- stant. The slight cranial peculiarities of the insular form, while hardly sufiicient by themselves to separate the animal, are of un- doubted significance when considered in connection with the external characters. For comparison with the allied insular species see de- scription of Mus ta??ibelanicus, p. 213. MUS SIANTANICUS sp. nov. Type. — Adult male (skin and skull). No. 101705 United States National Museum. From Fulo Siantan, Anambas Islands. Collected September 11, 1S99. Characters. — Size, proportions, quality of fur, and color essentially ' Measurements in parenthesis are those of an adult male Mus ^ alexandri- nus' from Trong (No. 84395). MAMMALS COLLECTED BY DR. W. L. ABBOTT 211 as in the form of Mus '■ alexandrinus' occurring in Trong, Lower Siam, but skull with noticeably broader, less arched rostrum. Color. — Dorsal surface a fine grizzle of dull ochraceous-buff and brownish black, the hairs with a distinct gloss in certain lights. On shoulders and middle of back the brown predominates, and at the same time the ochraceous-buff is duller and paler than elsewhere. On sides, flanks and across lumbar region the dark hairs are less numer- ous and the ochraceous-buff is distinctly tinged with russet. Ventral surface of body and inner side of legs yellowish white, darker, and somewhat brownish tinged across chest. Fur of back pale slate-gray at base, that of ventral surface mostly unicolor to base, but irregularly tinged with gray beneath the surface. Feet light brown. Tail uni- form dark brown throughout. Skull and teeth. — The skull of Mus siantanicus is essentially like that of M. tiomanictis and j\I. ' alexandrinus ' except in the form of the rostrum. This differs conspicuously from that of the related spe- cies in its breadth, and in the flatness of the dorsal surface. The characters are most noticeable when the skull is viewed from directly in front. It is then seen that the nasal bones are very slightly de- pressed at the sides only, while in AIus tiomanictis and M. alexan- drinus they are uniformly arched from suture to periphery, the out- line of the two together forming nearly one-half of a circle. Viewed from the side, the dorsal profile of the rostrum appears nearly straight. Viewed from beneath, the sides are seen to be more nearly parallel than in the two related species. Teeth in no way peculiar. Measurements. — External measurements of type : total length 356 ; head and body 191 ; tail vertebrce 165 ; hind foot 35 (33) ; ear from meatus i7-^; e^r from crown 14. 8; width of ear 15. Average of five adults from the type locality; total length 365 (349-382) ; head and body 186 (17S-191); tail vertebras 179 (165-191); hind foot 35.6 (35-36) ; hind foot without claws 36.9 (33-35). Cranial measurements of type : greatest length 42 ; basal length 37; basilar length 35 ; palatal length 21 ; least width of palate be- tween anterior molars 4; diastema 12; length of incisive foramen 8; combined breadth of incisive foramina 3; length of nasals 15.4; com- bined breath of nasals 4.6; zygomatic breadth 18; interorbital breadth 6.4; mastoid breadth 15.6; breadth of braincase over roots of zygo- mata 16,4; depth of braincase at front of basioccipital 12; fronto- palatal depth at posterior extremity of nasals 9.4; least depth of rostrum immediately behind incisors 7 ; maxillary toothi'ow (alveoli) 7; width of first upper molar 2; mandible 23.6; mandibular tooth- row (alveoli) 7. 212 MILLER Specime7is examined. — Six, all from the type locality. Remarks. — Unlike the Tioman rat, JMus siantajiicus retains the external appearance of the mainland form, but differs widely in its cranial characters. In the latter it shows an approach to the much larger Tambelan representative of the group. MUS TAMBELANICUS sp. nov. Type. — Adult male (skin and skull). No. 101665 United States National Museum. From Big Tambelan Island. Collected August 10, 1899. Characters. — Much larger than ^Tus siantanicus oxXh^iowa of M. ' alexandrinus ' occurring in Trong, Lower Siam. Fur coarser than in the related species ; color about as in Mus tiomanicics^ or slightly more red, but ventral surface darker and more yellow. Skull with rostrum relatively even more heavily built than in J7}/5 siantaiiiciis^ but with dorsal surface less flattened. Color. — Back a fine grizzle of dull cinnamon and blackish brown, the two colors nearly equally mixed. Sides nearly clear cinnamon, brightening to russet on flanks. Ventral surface dull buff, faintly and irregularly marked with brownish, and very obscurely contrasted with color of sides. Feet sepia. Tail uniform dark brown throughout. Skull. — The skull most nearly resembles that of Miis siantanlcus but is much larger and the rostrum is relatively broader than in the Anambas species. Region between anterior zygomatic roots convex laterally^ not flat or concave as in the allied species. Dorsal profile of nasals when viewed from in front slightly convex, but much less so than in Mus tiom.anictis and the Trong form of M. ' alcxai/dri/n/s. Lateral margins of rostrum nearly parallel. Teeth larger tlian in the related species, but not disproportionately so. Measurements . — External measurements of type : head and body (tail imperfect) 216; hind foot 42 (39); ear from meatus 21; ear from crown 15; width of ear 16.4. Two adults from Pulo Bunoa measure: (No. 101646, 9) total length 425; head and body 203 ; tail vertebnc 222; hind foot 40 (37); and (No. 101642, 9) total length 375; head and body iSi ; tail vertebra? 191 ; hind foot 40 (37). An adult male (No. 101669) ^•"<^'''^ Pulo Wai measures: total length 393; head and body 191 ; tail vertebnc 203; hind foot 41 (38). Cranial measurements of type: greatest length 47; basal length 41; basilar length 38; palatal length 24; least width of palate be- tween anterior molars 4.6; diastema 13; length of incisive foramen S ; combined breadth of incisive foramina 3.8; length of nasals 17; com- MAMMALS COLLECTED BY DR. \V. L. ABBOTT 213 bineJ breadth of nasals 5.4; zygomatic breadth 22.4; interorbital breadth 7; mastoid breadth iS; breadth of braincase over roots of zygomata 17.6; depth of braincase at front of basioccipital 12.S; frontopalatal depth at posterior extremity of nasals 11.4; least depth of rostrum immediately behind incisors 8.8; maxillary toothrow (alveoli) 8; width of first upper molar 2.4; mandible 26.4; mandi- bular toothrow (alveoli) 7.4. Specimens examined. — Five, one adult from Big Tambelan Island, two from Pulo Bunoa, an adult from Pulo Wai, and an immature in- dividual from Pulo Gilla. Remarks. — ^Iiis tambelatziciis marks, so far as the species repre- sented by the j^resent collection are concerned, the culmination of a series beginning with the light colored, slender skulled animal of Trong. The first step away from this type is represented by AIus tiotnanicus^ in which the color is distinctly darkened, but in which the cranial charac- ters remain practically unchanged. In the representative of the group occurring on the Anambas the rostrum is noticeably altered in form, though the color is less affected than in the Tioman animal. Finally, in the Tambelan Islands a race has been developed carrying the cranial peculiarities even further than in that inhabiting the Anambas ; the changes in form of skull accompanied by marked increase in the gen- eral size of the animal, in the quality of the fur and in the color. The characters of each of these closely related species are remarkably constant in view of the reputation for variation enjoyed by ]\Ius ' alex- andrius.' The adult specimen of M. tambelamcus from Pulo Wai differs from the skins from Big Tambelan in less yellow ventral sur- face, but otherwise it is not peculiar. The specimen from Pulo Gilla is too immature to be comparable with the others. MUS OBSCURUS sp. nov. Type. — Adult male (skin and skull). No. 101764 United States National Museum. From Tioman Island. Collected October i, 1899. Characters. — Similar to Mtts coficolor Blyth, but larger, and with more coarsely spinous fur. Skull larger than that of ]\Ius coficolor., the dorsal profile less convex, the audital bullae larger and teeth rela- tively smaller. Color. — General appearance that of a much overgrown house- mouse. Dorsal surface a fine grizzle of seal brown and dull ochra- ccous-buff, the former predominating on crown of head and in himbar 214 MILLER region, the latter slightly in excess on rump, flanks and shoulders, as well as on sides of body and outer surface of legs. Ventral surface of body and inner side of legs pale drab, Irregularly and faintly washed with brownish yellow. Feet drab-gray. Tail uniform dark brown throughout. Fur of body everywhere slate-gray at base. Fur. — The fur is rather short, about 8 mm. in length on middle of back, and everywhere thickly and uniformly sprinkled with well de- veloped, slender, grooved bristles. On lower part of sides the bristles become even more slender, while on ventral surface they are at first sight not noticeably different from ordinary hairs. 7a//.^The tail is closely and uniformly annulated ; about i6 rings to the centimeter at middle. Though sharply distinguished from each other, the rings are very obscurely divided into scales. Tail uniformly clothed with very short hairs (their length mostly about double width of rings) which neither hide the annulation nor form a terminal pencil. Feet. — Soles 6-tuberculate, the sixth pad well developed. The surface of the sole appears (on soaking the foot of the dry specimen) to be distinctly and closely reticulate in the region between the pads, but perfectly smooth behind them. Skull. — As compared with a skull of ]\Tus concolor from Bhamo, Burmah (No. 1015 19 United States National Museum), that of JSf. obscurus differs noticeably in its larger size, more slender rostrum, and particularly in the less convex profile when viewed from the side. The antorbital foramina are larger than in the related species, and less contracted below. Audital bullie relatively as well as actually larger than in Mus concolor.^ and not very different in form. Teeth. — The teeth are essentially as in AIus concolor^ but actually smaller. Measure7)ients. — Total length 254; head and body 121 ; tail verte- bra; 133; hind foot 34 (22.6) ; ear from meatus 16; ear from crown 13 ; width of ear 13. Skull: greatest length 30 (28. 6) ; ^ basal length 26 (24); basilar length 23 (21.8); palatal length 13 (12); least width of palate be- tween anterior molars, 3.2 (2.8) ; diastema 8 (7) ; length of incisive foramen 5.4 (5); combined breadth of incisive foramina 2 (1.8); length of nasals 11 (10.2) ; combined breadth of nasals 3 (3) ; zygo- matic breadth 13.6 (13) ; interorbital breadth 4 (5) ; mastoid breadth 12 (12) ; breadth of braincase over rootsof zygomata 13 (12.6) ; depth ' Measurements in parenthesis are those of an adult female I\[us concolor from Bhamo, Burmah. MAMMALS COLLECTED BY DR. W. L. AHIiOTT 215 of braincasc at front of basioccipital 9 (S.4) ; frontopalatal deptli at posterior extremity of nasals 6.6 (6) ; least depth of rostrum im- mediately behind incisors 5 (4.8) maxillary toothrow (alveoli) 4.6 (5); width of first upper molar 1.6 (1.6); mandible 15.4 ( 14.6) ; mandibular toothrow (alveoli) 4.6 (5). Spcc/'/nc}is examined. — One, the type. 7^CMar/:s. — J/^/s obscurus will probably be found to occur on the mainland of the Malay Peninsula, where its range is likely to coincide witli that of ^[ks surifer and AIus vocifera7is. RATUFA ANAMB^ sp. nov. Type. — Adult male (skin and skull). No. 101725 United States National Museum. From Pulo Jimaja, Anambas Islands. Collected September 25, 1S99. Characters. — In color and in cranial characters similar to Ratufa melanopepla, but size nearly equal to that of Ratufa gigantea. Ears not tufted. Color. — Cheeks, throat, sides of neck, inner surface of legs, and entire ventral surface of body rich tawny yellow (ochraceous) paler on throat, cheeks, and sides of neck, richer on chest, the belly much darkened by appearance at surface of blackish bases of hairs. Re- mainder of pelage and entire tail glossy blue-black, with the usual reddish brown cast beneath the surface. Yellow of front leg spread- ing conspicuously over outer side between elbow and wrist. Skull and teeth. — Except for their distinctly larger size the skull and teeth scarcely differ from those of R. ?nelanopepla. Audital bullae narrower and more elevated above surface of basioccipital (when skull is held upside down) than in R. melanopepla, the peculiarity in form the same as that of the Tioman race, but more accentuated. Measurements. — External measurements of type : total length 825 ; head and body 374; tail vertebree 451 ; pencil 100; hind foot 83 (76). Average of six specimens from the type locality : total length 806 (781-S3S) ; head and body 370 (356-381); tail vertebra 436 (400- 470) ; hind foot 82.8 (82-84) ; hind foot without claws 75.8 (75-77). Cranial measurements of type: greatest length 72; basal length 60 ; basilar length 56 ; palatal length 28 ; length of nasals 25 ; breadth of nasals anteriorly 13; breadth of nasals posteriorly '$>.6\ interorbital breadth 26 ; breadth between tips of postorbital processes 40 ; zygo- matic breadth 45 ; mastoid breadth 33 ; breadth of palate between premolars 9; depth of rostrum at posterior extremity of nasals 21; mandible 42 ; depth of mandible at posterior root of premolar 1 1 .4 ; 2l6 MILLER least depth of mandible between incisor and premolar 10.2 ; maxillary toothrow (alveoli) 14; breadth across both upper incisors at rim of al- veoli 9.8; mandibular toothrow (alveoli) 14. S. Specime7is examined. — Six, all from the type locality. Remarks. — This is an insular race characterized, contrary to the general rule, by increase in size. Its peculiarities are well marked and constant. RATUFA TIOMANENSIS sp. nov. Type. — Adult male skin and (skull). No. 101751 United States National Museum. From Tioman Island. Collected October 4, 1S99. Characters. — Color exactly similar to that of Ratufa anambce and R. uielanopepla., but size smaller and tail relatively longer than in either. Skull like that of 7?. ^nclanopepla., but with more inflated audital bullae. Color. — See description of Ratufa anajnbce.^ which applies equally well to this animal. Skull. — In size and general form the skull of Ratufa tiomanensis so closely agrees with that of R. melanopepla that no special descrip- tion is necessary. The audital bullae are, however, slightly though constantly different in form. Though of about the same size as in R. i7ielanopepla the bullae are more inflated, so that when viewed from behind (the skull held upside down) they appear to rise higher and more abruptly above surface of basioccipital. This character, though slight, is readily appreciable when specimens are compared, and its constancy makes it of more importance than would at first sight be supposed. Measurements. — External measurements of t3'pe : total length 724; head and body 368; tail vertebra 356; pencil 90; hind foot 78 (73). Average of five specimens from the type locality: total length 740 (724-762) ; head and body 343 (305-368) ; tail vertebne 397 (356- 419) ; hind foot 76.4 (74-78) ; hind foot without claws 69.8 (68-72). Cranial measurements of type: greatest length 68.4 (70);' basal length 56.4 (59) ; basilar length 52 (53) ; palatal length 26 (26.4) ; diastema 15.4 (16) ; length of nasals 23 (23.4) ; breadth of nasals an- teriorl}' 12.6 (13); breadth of nasals posteriorly 7 (7) ; interorbital breadth 26 (28) ; breadth between tips of postorbital processes 39 (41); zygomatic breadth 43.4 (44); mastoid breadth 31 (32.6); breadth of palate ])etwcen premolars 10 (10); depth of rostrum at ' Measurements in pircnthesis are those of the type of Ratufa tiielanopepla. MAMMALS COLLECTED BY DR. W. L. AHJJOTT 217 posterior extremity of nasals iS. 6 (1S.8); mandible 41 (41.6) depth of mandible at posterior root of premolar 11.4 (11) ; maxillary tooth- row (alveoli) 13 (14) ; breadth across both upper incisors together at rim of alveoli 8.4 (8) ; mandibular toothrovv (alveoli) 14 (14.4). Specimens examined. — Five, all from the type locality. Remarks. — Ratufa tiomanensis is less differentiated from R. melanopepla than is the Anambas form. Nevertheless its characters are easily recognizable on comparison, while their constancy is very marked. FUNAMBULUS CASTANEUS sp. nov. Type. — Immature male (skin and skull). No. 101696 United States National Museum. From Pulo Siantan, Anambas Islands. Collected September 10, 1899. Characters. — Similar to the form of Funambulus insignis occur- ring in Trong, Lower Siam, but darker in color, and sides and flanks not grizzled ; skull with much larger audital bullae. Color. — Sides, flanks, and outer surface of legs clear chestnut, the individual hairs slate-gray at base, ferruginous at middle and black at tip. The color is brightest on flanks and palest on front legs and sides of head. Neck and middle of back grizzled gray faintly washed with tawny. Top of head similar but with tawny tinge more con- spicuous. In lumbar region and on rump the gray gives place grad- ually to color of flanks. The gray area is bounded on each side and divided in the middle by a black stripe about 3 mm, in breadth. These three black stripes begin a little in front of shoulders and extend to rump, where they disappear. The two strijDCs into which the gray area is thus divided are each about 5 mm. in width. Ventral surface of body and inner side of legs soiled yellowish white to base of hairs, the inner side of hind legs strongly washed with ferruginous. Feet a dull grizzle of black and tawny. Ears brownish internally, whitish externally. The hairs of the tail are black at base (3 mm.) then light, dull ferruginous (5 mm.), followed by black (lo mm.), the extreme tips whitish. On dorsal surface the ferruginous is nearly con- cealed by the black, but on ventral surface it forms a well-defined median area. Skzill. — The type specimen is so young that no accurate description of the cranial characters can be given. The audital bullae, however, are actually larger than in a fully adult specimen of Funatnbulus insignis from Trong. They are so large that when viewed from the side they conceal the tips of the pterygoids, which in F. i?zsignis are plainly visible. l8 MILLER Measurements. — External measurements of type : total length 241; head and body 152; tail vertebrae 89; pencil 30; hind foot 45 (40. Speciinen examined. — One, the type. Retnarks. — Although the only specimen of this squirrel is imma- ture (crown of last upper molar at rim of alveolus) its characters appear to be quite sufficient to separate the Anambas form from that of the mainland. It is probably more closely allied to the much brighter colored Fu7iajnbulus diversus (Thomas) of Borneo. SCIURUS MIMELLUS sp. nov. Type. — Adult male (skin and skull). No. loi 668 United States National Museum. From Pulo Wai, Tambelan Islands. Collected August 13, 1899. Characters. — A miniature of Sciurus prevostii (hind foot about 53 instead of 63, greatest length of skull 51 instead of 58). Color. — Entire tail, dorsal surface of head and body, and outer sur- face of hind legs glossy black to base of hairs. Underparts, feet, inner side of hind leg, and entire fore leg bright reddish chestnut (brighter than Ridgway's pi. IV, No. 9) fading to cinnamon-rufous on thighs and to light orange-rufous on outer side of upper arms. Black of back separated from red of belly by a lateral band of white beginning just behind axilla and extending down front of hind leg to knee and about 12 mm. wide at middle. Cheeks, chin and sides of neck grizzled bluish gray. Sides of muzzle whitish. Whiskers and ears black. The fur of the red area is everywhere red to base except on sides, close to white stripe. Here it is blackish beneath the sur- face. Skull. — Except for its smaller size the skull of Sciurus mimcllus exactly resembles that of S. prevostii. Measurements . — External measurements of t3^pe : total length 451 ; head and body 229; tail vertebras 222; pencil 60; hind foot 55 (51). A somewhat younger individual (female) from the type locality : total length 442 ; head and body 222; tail vertebrce 210; pencil 60; hind foot 54 (51). Cranial measurements of type: greatest length 51.6 (58'); basal length 43.6 (52); basilar length 40.6 (48.4) ; palatal length 22.8 (26) length of nasals 15 (19); breadth of nasals anteriorly 72 i^-^) breadth of nasals posteriorly 4.4 (5) ; interorbital breadth, 18.4 (24. 8) ' Measurements in parenthesis are those of an adult female Sciurus prevostii from Troug, Lower Siam. mam:mals collected by dr. w. l. abhott 219 breadth between tips of postorbital processes 27.4 (34) ; zygomatic breadth 3 1 (37) ; mastoid breadth 21.4 (25) ; breadth of palate between premolars 6.S (S.4) ; depth of rostrum at posterior extremity of nasals 12.S (13.6) ; mandible 31.4 (37) ; depth of mandible at posterior root of large premolars S (9); maxillary toothrow (alveoli) 10 (11); breadth across both upper incisors together at rim of alveoli 5 (6) ; mandibular toothrow (alveoli) 10 (11.4). Specimens exatnl?icd. — Two, both from the type locality. Remarks. — Sciunis tjiimelhis is distinguished from S. prcvostil, aside from its size, by the reduction in width of the white lateral band. In the mainland animal this band is so wide that on the hind leg it oc- cupies the entire outer side, reducing the black to a mere inconspicu- ous line between it and the red. In S. mimellus on the other hand the white occupies rather less space on the leg than the black, so that the latter color is more conspicuous than the white when the skin is A'iewed from above. Six adult specimens of Schiriis prevostii from Trong, Lower Siam, measure: total length, 533 (502-553); head and body 262 (24S-273) ; tail vertebrae 275 (254-286); hind foot 62.5 (60-65) ; hind foot without claws 58 (56-60). SCIURUS MIMICULUS sp. nov. Type. — Adult male (skin and skull). No. 101616 United States National ]Museum. From Ste. Barbe Island. Collected August i, 1899. Charactei's. — Similar to Sciurus mifnellus but even smaller (hind foot about 40, greatest length of skull about 48). Color, — In color Scizirzis mimiciihis exactly agrees with S. mimel- lus^ even to the narrowing of the white lateral band and consequent increase in the amount of black on the hind leg as compared with S. prevostii. Sktill. — The skull (fig. 12, a) is as in Sciurus mim,ellus, though much smaller. Audital bullae slightly less inflated than in the larger species. Measurements. — External measurements of type: total length 381 ; head and body 203; tail vertebrae 178; pencil 50; hind foot 42 (38). Average of nine specimens fi'om the type locality : total length 387 (375-406); head and body 198 (191-203); tail vertebrae 189 (17S- 203) ; hind foot 43.5 (42-49) ; hind foot without claws 42.4 (38-45). Cranial measurements of type : greatest length 48 ; basal length 41 ; basilar length 38; palatal length 20; length of nasals 14; breadth of nasals anteriorly 7 ; breadth of nasals posteriorly 5 ; interorbital 220 MILLER breadth 19; breadth between tips of postorbital processes 27; zygo- matic breadth 29 ; mastoid breadth 20 ; breath of palate between l^remolars 6; depth of rostrum at posterior extremity of nasals 11.8; mandible 28. S ; depth of mandible at posterior root of large premolar 7.2; maxillary toothrow (alveoli) 9; breadth across both upper in- cisors together at rim of alveoli 4.6; mandibular toothrow (alveoli) 9. Fig. 12. a Sciiinis »iimiculiis. b Sciurua frcx'osfii. Natural size. Speciyjiens examined. — Nine, all from the type locality. Remarks. — One of the specimens of this species ( $ No. 101617, August i) is in very abraded pelage, the longer hairs of the back be- ing for the most part worn away. Yet there is no appreciable change in color. The tail has become a dark, indefinite brown, but the re- mainder of the pelage retains its colors practically unaltered. MAMMALS COLLECTED BY DR. W. L. ABBOTT 221 SCIURUS TENUIS Horsfield. 1824. Sciunis tenuis Horsfield, Zoological Researches in Java and the neigh- boring Islands. (Pages not numbered.) Type from Singapore. Specimens of this sqtiirrel indistinguishable from topotypes were taken on the following islands : Linga; one adult female. Aiiambas; Pulo Siantan, nine specimens of both sexes ; Pulo Mobur, one adult male ; Pulo Jimaja, oiae adult female. Tioinan; one adult male. SCIURUS NOTATUS Boddaert. 1785. &-z«n« «(7/a/«5 Boddaert, " Elenchus Mammalium, p. 119." Four females collected on Linga Island, July 9-21, agree in all re- spects with the form of Sciurus notatiis occurring at Singapore. Whether this is true notatus cannot now be determined. SCIURUS TENUIROSTRIS sp. nov. Type. — Adult female (skin and skull). No. 1 01 753 United States National Museum. From Tioman Island. Collected September 30, 1899. Characters. — Externally similar to Sclicrits notatus from Singa- pore. Skull with rostral portion more slender and lightly built than in the continental forms. Color. — Entire dorsal surface of head and body a fine grizzle of black and yellowish wood-brown, the individual hairs black with two or often three brown rings. The wood-brown is (at the surface) everywhere in excess of the black. Cheeks, outer surface of legs and upper surface of feet dull cinnamon, faintly grizzled. Underparts and inner side of legs orange-rufous, brightest posteriorly (where it closely agrees with Ridgway's figure) , considerably paler on chest, throat and inner side of front legs. On sides the color of belly is separated from that of back by two longitudinal stripes extending from just behind axilla to flank, the lower black, the upper buff. Though the width of the stripes is somewhat variable, that of the black is usu- ally about lo mm., that of the buff slightly less. Dorsal surface of tail like back but less finely grizzled, each hair with five or six brown rings alternating with similar areas of black. Ventral surface even less finely grizzled and the brown distinctly tinged with cinnamon like that of flanks. Pencil in no way different from rest of tail. Whiskers black. 222 MILLER Skull. — The entire skull of Sciurus tenuirostris (fig. 13, c) is nar- rower than that of S. notatus irovn. Singapore (fig. 13, a), but the dif- ference is most strikingly apparent in the rostrum. The greatest width Fig. I- ;.^ notatus. b S. klossi. c S. tcnuiroilrii. Natural size* of both nasals together Is more than half of nasal length in S. notatus^ while in S. teriuirostris it is contained in nasal length nearly two and one-half times. Viewed from beneath the sides of the rostrum are seen to be much more contracted in S. tenuirostris., so that a distinct ridge extends from incisive foramina to henselion. The palate also is much narrower than in S. notatus. Teeth as in S. ?iotatus. Measureinents. — External measurements of type: total length 356; head and body 191 ; tail vertebrai 165 ; pencil 50; hind foot 43 (40.4). A second specimen ( 9 No. 101754) : total length 362 ; head and body 191 ; tail vertebras 171 ; pencil 55; hind foot 43 (41). Cranial measurements of type : greatest length 50 (51)' basal length ' Measuremeuts iu parenthesis are those of au adult female Sciurus notatus from Singapore. MAMMALS COLLECTED BV DR. W. L. ABBOTT 223 41.6 (43.6); basilar length 39 (40); palatal length 21.6 (23); dia- stema 1 1.6 (n.6) ; length of nasals 15.6 (15) ; breadth of nasals ante- riorly 6 (7-8) ; breadth of nasals posteriorly 3 (4) ; interorbital breadth 17.4 (18) ; breadth between tips of postorbital processes 26 (28); zygomatic breadth 28 (31); mastoid breadth 20.4 (22); breadth of palate between premolars 5.6 (6.4) ; depth of rostrum at posterior extremity of nasals 1 1. S (13); mandible 30 (29.6); depth of mandible at posterior root of large premolar 7 (8) ; maxillarj^ tooth- row (alveoli) — (9'6) ; breadth across both upper incisors together at rim of alveoli 4 (5) ; mandibular toothrow (alveoli) — (9). Speci7iiens examined. — Two, both from the type locality. Re??iarks. — The peculiarities of the skull characterize Sciuriis temiirostris as a strongly marked local form of the S. notatus group. SCIURUS ANAMBENSIS sp. nov. Type. — Adult female (skin and skull). No. 101686 United States National Museum. From Pulo Siantan, Anambas Islands. Collected September 12, 1899. Characters. — Like Sciurus notatus from Singapore but slightly smaller; color, particularly on underparts, paler. Skull similar to to that of 6". notatus (though in some respects resembling that of 6". temiirostris') but somewhat smaller and with distinctly narrower palate. Color. — Upperparts and tail as in Sciurus notatus and S. tenulros- trls^ but slightly paler, the difference easily perceptible on comparison but impossible to define in words. Underparts orange-rufous, but everywhere much paler than in Ridgway's figure and on the throat and inner side of front legs fading to ochraceous-buff. Outer surface of legs scarcely if at all tinged with cinnamon. Cheeks, chin, and sides of neck wood-brown. Otherwise as in S. notatus. Skull. — The skull of Sciurus anainbe7isis^ except for its slightly smaller size, closely resembles that of S. notatus from Singapore. The nasals are exactly as in this species, and do not in the least sug- gest those of 6". temiirostris. On the other hand the palate is nearly as narrow as in the Tioman species, while the rostrum when viewed from below has nearly the same form as in the latter. Audital bullee slightly larger than in S. notatus and S. te?iziirostris. When viewed from behind (the skull held upside down) they appear to rise higher and more abruptly above level of basioccipital than in the re- lated species. Teeth in no way peculiar. Measurejncnts. — External measurements of type : total length 369 ; 224 MILLER head and body 191 ; tail vertebrje 178 ; pencil 70; hind foot 44 (40.4). Average of five specimens from the type locality: total length 367 (355-375); head and body 194 (1S4-310); tail vertebrae 176 (171- 17S); hind foot 46.6 (44-49); hind foot without claws 43.6 (42- 45.4). Average of five specimens from Pulo Jimaja, Anambas : total length 368 (356-381) ; head and body 199 (197-203) ; tail verte- brae 169 (165-178) ; hindfoot 47.1 (45-48) ; hind foot without claws 44 (42-45). An adult female from Pulo Mata, Anambas: total length 368; head and body 203; tail vertebriE 165; hind foot 48 (44). An adult male from Pulo Riabu, Anambas: total length 362 ; head and body 191 ; tail vertebrae 171 ; hind foot 47 (43.6). Cranial measurements of type : greatest length 49.6; basal length 41.4; basilar length 39 ; palatal length 21.4 ; diastema 1 1 ; length of nasals 14.4; breadth of nasals anteriorly 7; breadth of nasals pos- teriorly 4.2; interorbital breadth 17.4; breadth between tips of post- orbital processes 26 ; zygomatic breadth 29 ; mastoid breadth 20 ; breadth of palate between premolars 6.2 ; depth of rostrum at posterior extremity of nasals 1 1 ; mandible 29 ; depth of mandible at posterior root of large premolar 7; maxillary toothrow (alveoli) 9; breadth across both upj^er incisors at rim of alveoli 4 ; mandibular toothrow (alveoli) 9. Specime?is examined. — Fifteen, from the following islands of the Anambas group : Pulo Siantan, 6 ; Pulo Jimaja, 6 ; Pulo Mata, 2 ; Pulo Riabu, i. Remarks. — The plantain squirrel of the Anambas Islands is readily distinguishable from other members of the Sciurus notatus group, though no one of its characters alone is very striking. Specimens from the different islands agree closely with each other, the one from Pulo Riabu excepted. This differs from the others in a marked re- duction in the size of the audital bullae, in this respect approaching the Tambelan species. In other respects this specimen is typical. SCIURUS ABBOTTII sp. nov. Type. — Adult female (skin and skull). No. 101662 United States National Museum. From Big Tambelan Island. Collected August 10, 1899. Characters. — vSimilar to Sciurus a7ia7}ibensis but skull with more slender rostrum and smaller, less inflated audital bullaj. Color. — The color of Sciicr?is abbottii agrees so closely with that of S. anambcnsis that no detailed description is necessary. In many of the specimens there is a distinct sprinkling of grayish hairs in tawny of underparts, especially in region bordering the dark lateral stripe. M.UIMALS COLLECTED BY DR. W. L. ABBOTT 225 Skull. — Skull much like that of Scinriis a)ia}nbensis in size and o-eneral form. The rostrum is more slender, particularly when viewed from below, and the audital bulla) are noticeably different in form. The bullae are so flattened that when viewed from behind (the skull held upside down) they appear to rise scarcely above level of tips of hamulars. Each is constricted near middle by a groove extending from just behind the lateral process of basioccipital over ventral ridge of bulla and disappearing on outer side. This groove, faintly indicated in the related species, is here so greatly developed as strongly to affect the shape of the bullae, particularly when viewed from the side. Measurements. — External measurements of type : total length 375 ; head and body 197; tail vertebrae 17S; pencil 60; hind foot 46 (41). Average of six specimens from the type locality : total length 373 (361-394); head and body 200 (197-203); tail vertebrae 177 (15S- 191) ; hind foot 45.2 (43-47) ; hind foot without claws 41.2 (40-44). Average of four specimens from Pulo Bunoa, Tambelans : total length 369 (362-3S2) ; head and body 197 (1S4-210) ; tail vertebra 171 (158-191); hind foot 44.7 (44-46.4); hind foot without claws 41.6 (40-43)- The skull is so nearly like that of Sciurus anambcnsis in size that it is unnecessary to give detailed cranial measurements. Greatest length 49; length of nasals 16; greatest breadth of both nasals together 6.6 ; width of palate between premolars 6. Specimens examined. — Fourteen ; ten from Big Tambelan Island, four from Pulo Bunoa. Remarks. — This is another well marked local form of the Sciurus notatus group. Specimens from Big Tambelan Island and Pulo Bunoa do not differ appreciably from each other. SCIURUS KLOSSIP sp. nov. Type. — Adult male (skin and skull). No. 10167S United States National Museum. From Kaju Ara or Saddle Island, Tambelan group. Collected August 15, 1S99. Characters. — A small, blue-bellied member of the Sciurus notatus group, somewhat closely resembling the Bornean S. cr^^/^i' (Thomas), but differing in much duller, more fulvous coloration. Tail with clear black pencil. Color. — General appearance of upperparts as in Sciurus anambe7i- sis and 6*. abbottii but light rings on hairs of tail broader and paler, in some specimens producing indistinct cross-bars, and pencil ahvays 1 At Dr. Abbott's request this species is named after Mr. C. B. Kloss. 2 26 MILLER clear blacky without grizzling. Outer surface of hind legs like back, that of front legs encroached upon by the bluish gray of inner sur- face. Feet slightly more yellow than back, and much more finely grizzled. Cheeks^ chin, muzzle and sides of 7ieck dull ochraceous^ in strong contrast with color of back. Pale lateral stripe dull ocJiraceous btiff"^ very slightly d iff eroit from color of cheeks. Dark lateral stripe of the usual breadth, but hardly as long as in the related red-bellied spe- cies. Ears intermediate in color between cheeks and back. N'o light area behind ear., the hairs in this region colored like those of cheeks, but without grizzling. Underparts and inner surface of legs light gray (very nearly Ridgway's PI. II, No. 9), slightly grizzled with ochra- ceous along median line and everywhere irregularly dulled by the ap- pearance at the surface of the slate gray bases of the hairs. Under surface of tail more ochraceous than upper surface, coarsely grizzled, and not showing as strong a tendency to faint cross-barring. Skull. — The skull of Sciurus klossii (fig. 13, 3) is essentially a miniature of that of S. notatus from Singapore. It is, however, broader relatively to its length. Teeth as in S. notatus., due allow- ance being made for difference in size. Measurements. — External measurements of type : total length 323 ; head and body 171 ; tail vertebrae 152; pencil 70; hind foot 41 (38). Average of six specimens from the type locality: total length 320 (310-330); head and body 176 (171-1S4) ; tail vertebrae 147 (140- 152) : hind foot'40 (38-41.6) ; hind foot without claws 37.1 (35-39)- Cranial measurements of type : greatest length 42 ; basal length 36 ; basilar length 33.4; palatal length 17.4; diastema 9.S; ^length of nasals 12; breadth of nasals anteriorly 6 ; breadth of nasals posteriorly 3; interorbital breadth 16; breadth between tips of postorbital proc- esses 22.6; zygomatic breadth 26.6 ; mastoid breadth 19; breadth of palate between premolars 6 ; depth of rostrum at posterior extremity of nasals 10; mandible 25; depth of mandible at posterior root of large premolar 6; maxillary toothrow (alveoli) S; breadth across both upper incisors at rim of alveoli 4; mandibular toothrow (alveoli) 7.6. Specimois examined. — Nine, all from the type locality. Remarks. — By its small size Sciurus klossii is separated from all the other blue-bellied members of the S. notatus group cxcei^t S. orestes. I have seen no specimens of this animal, but the original de- scription shows that it differs from the Saddle Island species in several important details of coloration, chief among which are the less fulvous cheeks,' white patch behind each ear, tind ' white "^ lateral stripe. 1 " Cheeks and sides of muzzle less fulvous than back, instead of more so." 2 "Sides with the usual lines of black and white." MAMMALS COLLECTED BY DR. W. L. ABBOTT 2 27 TRAGULUS NAPU (F. Cuvici). 1S22. Moschus napic F. Cuviek, Histoire Naturelle des Mammiferes, iv, livr. 11. November, 1822. (Sumatra.) One specimen (an old female with much worn teeth) of a large Traguhis was taken on Linga Island on July 23. It is probably identical with the T. napii of F. Cuvier, the type of which was col- lected in Sumatra. In color it is more yellow than the specimens described by Cuvier, though not approaching the red tints of the Tioman form. It measures as follows: head and body 533 (tail lacking) ; hind foo<" i30 (105) ; ear from meatus 34; ear from crown 30; width of ear 33. Skull: greatest length 114; basal length 106; basilar length 99; occipito-nasal length 104; length of nasals 34; breadth of nasals anteriorly 9 ; breadth of nasals posteriorly 1 1 ; me- dian length of bony palate (measured from tip of premaxillary) 72 ; width of palate between anterior premolars 1 7 ; width between pos- terior molars 21 ; diastema 15; zygomatic breadth 48 ; least interor- bital breadth 30 ; breadth between outer edges of paroccipital proc- esses 32; greatest length of audital bulla 23.4; greatest width of audital bulla 13.4; distance between anterior extremities of audital bullae 11; least distance between audital bullis ^.^^ depth from parietal to lower edge of audital bulla 38 ; depth from upper rim of orbit to alveolus of last molar 34.4 ; depth of rostrum at middle of diastema 18; mandible 90; maxillary toothrow (alveoli) 34; man- dibular toothrow (alveoli) 39.6. TRAGULUS RUFULUS sp. nov. Type. — Young adult' male (skin and skull). No. 101767 United States National Museum. From Tioman Island. Collected Septem- ber 30, 1S99. Characters. — Related to Traguhcs napii but smaller and more brightly colored ; middle permanent upper premolar with inner root nearly opposite posterior outer root, and whole form of crown propor- tionally altered. Color. — Upper parts bright tawny red, closely approaching the orange-rufous of Ridgway on back, rump, tail, and outer side of thighs, slightly paler on cheeks, neck and outer side of front legs, duller on crown and face. Sides fading to ochraceous-buff. Except on neck and legs the hairs of this tawny area are everywhere tipped with black, which produces a distinct but irregular clouding ^ Permanent dentition fully in place but practically unworn. 228 MILLER effect. Underparts (except throat) and inner side of legs to heel white. A dusky shade along median line of belly. Throat like neck, but with a narrow white median sti'ipe, and on each side a well devel- oped white lateral stripe. A white patch beneath each jaw. Tail white beneath and at tip. Ears and naked parts of muzzle, face and legs blackish in dried specimen. Hoofs horn color. Skull. — The skull is essentially similar to that of the specimen of T. 7iapit from Linga Island. Teeth. — Teeth as in Tragulus napii with the exception of the mid- dle upper premolar of the permanent dentition. In this the inner root is equidistant from each of the outer roots, with the posterior of which it Is nearly in line. The bases of the three roots therefore stand at the corners of a nearly equilatei'al triangle. The resulting form of the crown is quite different from that in the other large species of the genus. Measurements. — External measurements of type : total length 534 ; head and body 464 ; tail vertebrae 70 ; pencil 30 ; hind foot 1 30 (107) ; ear from meatus 31 ; ear from crown 27; width of ear 31. Cranial measurements of type: greatest length 104; basal length 97; basilar length 91 ; occipito-nasal length, 96; length of nasals 39; breadth of nasals anteriorly 9 ; breadth of nasals posteriorly 1 1 ; median length of bony palate (measured from tip of premaxillary) 64; width of palate between anterior premolars 14; width between posterior molars 15; diastema 9 ; zygomatic breadth 46 ; least inter- orbital breadth 36.4; breadth between outer edges of paroccipital processes 38; greatest length of audital bulla 20.4; greatest width of audital bulla 1 3 ; distance between anterior extremities of audital bullae 10; least distance between audital bullae 6.S ; depth from parietal to lower edge of audital bulla 38; depth from upper rim of orbit to alveolus of last molar 32 ; depth of rostrum at middle of dia- stema 18.6; mandible 81 ; maxillary toothrow (alveoli) 37 ; mandib- ular toothrow (alveoli) 42. Remarks. — This is an insular form of the napu characterized by its rich, bright, color, and by the peculiar form of the middle upper premolar. It is doubtless confined to Tioman Island. PARADOXURUS HERTvIAPHRODITUS (Schrebcr). 1778. Viverra /icrinaphrodifa Schreber, Saugethiere, ill, p. 426. (' Bar- bary.') One specimen, an adult female, with much worn teeth was secured on Tioman Island, October 3, 1899. Its measurements are as follows : MAMMALS COLLECTED BY DR. W. L. ABBOTT 229 total length 914; head and body 495 ; tail vertebrae 419; hind foot 76 (71). The fur is much paler in color than specimens from Trong, Lower Siam, and the stripes on the back show a strong tendency to break up into rows of spots. Skull of same size as in Trong speci- mens, but audital bullae slightly less inflated, and paroccipital proc- esses shorter and broader. Teeth too much worn to show any char- acters. TUPAIA TANA Raffles. 1S22. Tupaia taiia Raffles, Trans. Linn. See. London, xiii, p. 257. (Sumatra.) One specimen of this species, an adult male (No. 101597) was taken on Linga Island, July 16, 1899. Its measurements are: total length 324; head and body 191; tail vertebrae 133; pencil 25; hind foot 41(37)- TUPAIA BUNO^ sp. nov. Type. — Adult female (skin and skull). No. 101640 United States National Museum. From Pulo Bunoa, Tambelan Islands. Collected August 5, 1S99. Characters. — Externally similar to Tupaia tana but fulvous of un- derparts, and especially that of under side of tail paler. Skull with broader antorbital region and larger, more inflated audital bullae than in T. tana. Color. — Back behind shoulders glossy seal-brown, the shorter hairs narrowly ringed with orange-rufous subterminally and thus producing an inconspicuous grizzle better defined in some lights than in others. On sides and outer surface of legs the brown gradually gives way to the orange rufous, but nowhere entirely disappears. Head, nape, and shoulders a fine grizzle of sepia and yellowish white, the grizzled area narrowing posteriorly and extending back in median line slightly be- hind shoulders. It is traversed by a median seal-brown line about 3 mm. in width, extending nape to dark area of back. Immediately over each shoulder is a faintly defined oblique pale stripe. Entire ventral surface of body and inner side of legs ochraceous, fading to ochra- ceous-buff on throat and chin and becoming somewhat darker and more dull (nearly clay color) on inner side of hind legs. Tail ferru- ginous throughout, the dorsal surface strongly washed with seal brown, especially near base. Feet, ears, and muzzle dull brown. The fur of the body is everywhere slate gray at base, less noticeably so below than above. Skull. — The skull of Tupaia bunoce closely resembles that of T. tafza from Linga Island but is slightly larger and has a distinctly less 230 MILLER attenuate rostrum. The suborbital vacuity is wider than in the Linga animal. Audital bull^E larger and more inflated than in T. ta7ia and less distinctly triangular in outline. Teeth. — The teeth are as in Tupaia tana from Linga Island. First and second upper molars with small but distinct hypocone. Measurements. — External measurements of type : total length 363 ; head and body 210; tail vertebras 152; pencil 35 ; hind foot 45 (42). Average of three specimens from the type locality : total length 349 (330-362); head and body 197 (17S-210) ; tail vertebras 152; hind foot 44.4 (43.4-45) ; hind foot without claws 41.3 (40-42). Cranial measurements of type: greatest length 57.6 (54.6) ;^ basal length 53 (50.4) ; basilar length 50 (47) ; median palatal length 32 (30.6) ; length of nasals (joosterior extremity not clearly determinable) 21 (20.4) ; distance from lachrymal notch to a tip of premaxillary 28.4 (27.6) ; diastema 6 (5.6) ; width of palate at middle of diastema 5.2 (4.6) ; width of palate between anterior molars 9 (S.S) ; lachry- mal breadth 19.8 (19) ; breadth of both nasals together at middle 3.6 (3) ; breadth of rostrum at middle of diastema 6 (5.4) ; least in- terorbital breadth 15 (14.4); zygomatic breadth 25.4 (25); breadth of braincase above roots of zygomata 19.6 (19) ; occipital depth 12.S (13) ; depth from middle of parietal to lower surface of audital bulla 18 (18) ; depth of rostrum at base of nasals 6.6 (6) ; depth of rostrum at middle of diastema 4.6 (5); mandible 39 (36); maxillary tooth- row (behind diastema) 20 (19); mandibular toothrow (behind dia- stema) 17 (16.4). Specimens exa?mjzed. — Four : three from Pulo Bunoa, one from Big Tambelan Island. Remarks. — The specimen from Big Tambelan agrees in all respects with those from Bunoa. Its hind foot measures 44 (41) mm. Tail imperfect. In color Tupaia bicnocc is very constant, the four speci- mens showing practically no individual variation. The under side of the tail in Tupaia tana is a very deep ferruginous, approaching chest- nut, while the ventral surface of the body is rich tawny ochraceous. TUPAIA MALACCANA Anderson. 1879. Tupaia vialaccaiia Anderson, Anat. and Zool. Researches, p. 134. (' Malacca.') Three specimens from Linga Island agree in all essential characters with those given in the original description of this species, though the color of the underparts appears to l)e darker and that of the lower side ' Measurements in parenthesis are those of the specimen of Tupaia tana from Linga Island. MAMMALS COLLECTED BY DR. W. L. ABIiOTT 231 of the tail lighter than indicated by Anderson's account. Average measurements of the three individuals: total length 302 (298-305); Fig. 14. Tupaia malaccami. Natural size- head and body 13S (133-140); tail vertebrae 165; hind foot with claws 33.7 (33-34) ; hind foot without claws 31.5 (30-32.4). TUPAIA SORDIDA sp. nov. Type^ adult male (skin and skull). No. 1 01 747 United States National Museum. From Tioman Island. Collected October 2, 1899. Characters. — Very similar to Tupaia ferrugbiea but slightly smaller, and color of flanks and underparts darker and less yellow. Skull shorter and broader than in T. ferrugiuea^ but Avith distinctly narrower bony palate. Color. — The color is so closely similar to that of Tupaia ferruginea that no detailed description is necessary. Underparts dull brownish buff, not very different from lower side of tail, slightly more yellow along middle of throat. Flanks somewhat darker and browner than in T. ferrugi7iea. Skull. — The skull of Ttipaia sordida differs from that of T. fer- ruginea in slightly shorter rostrum, distinctly broader, less elongate braincase, broader frontal region and narrower palate ; this last char- acter in direct contradiction to the sreneral features of the skull. 232 MILLER Teeth. — The teeth do not differ appreciably from those of Tzipaia ferruginea. Measurements. — External measurements of type : total length 336 ; head and body 171 ; tail vertebrae 165 ; pencil 25 ; hind foot 41 (38. 8). External measurements of an adult female from the t3'pe localit}^ : total length 3 iS; head and body 178; tail vertebrae 140; pencil 20; hind foot 40 (37.6). Cranial measurements of type: greatest length 49.6 (49.4)^; basal length 46 (45.6) ; basilar length 43 (42.4) ; median palatal length 27 (25) ; length of nasals (posterior extremity not clearly determin- able) 15.4 ( — ) ; distance from lachrymal notch to tip of premaxillary 22 (19) ; diastema 4 (3.6) ; width of palate between anterior molars 8 (9.4) ; lachrymal breadth 19 (17-4) ; breadth of rostrum at middle of diastema 7 (6.8); least interorbital breadth 15.4 (14); zygomatic breadth 25.4 (25) ; breadth of braincase above roots of z3'gomata 19.4 (18.8); least distance from inion to rim of orbit 21 (23); occipital depth 12 (13); depth from middle of parietal to lower surface of audital bulla 17.2 (17) ; depth of rostrum at base of nasals 7 (7-8) ; depth of rostrum at middle of diastema 5 (5.4) ; mandible 33.4 (32.6); maxillary toothrow (behind diastema) 15 (17); mandibular toothrow (behind diastema) 13 (13.6). Specimens examined. — Two, both from the type locality. Rejuaf'ks. — Tiipaia sordida differs from its mainland representa- tive to about the same degree as the other insular members of the Tioman fauna. TUPAIA CHRYSOMALLA sp. nov. Type, adult female (skin and skull). No. 101710 United States National Museum. From Pulo Siantan, Anambas Islands, collected August 24, 1899. Characters. — Size, proportions, and cranial characters as in Tupaia ferruginea. Color as in T. ferruginea except that shoulder stripes are even more indistinct. The back is tinged with dull ferruginous and the whole tail is suffused with golden yellowish brown which forms a clear band on under surface. Color. — Summer pelage : back, sides, and outer surface of legs a fine grizzle of ferruginous and seal-brown, the former predominating, particularly in region between shoulders, the latter most conspicuous on legs. Head essentially like back but more finely grizzled and slightly tinged with gray. Underparts and inner side of legs dull 'Measurements in parenthesis are those of an adult male Tupaia fciTuginca from Trong, IvOwcr Siam. MAMMALS COLLECTED BY DR. W. L. ABBOTT 233 grizzled tawny-olive, bu££ tinged on chest. Feet and ears seal brown, faintly grizzled. Shoulder stripes ill defined, dull ochraceous. Tail glossy seal-brown above, irregularly lightened and grizzled, espe- cially through terminal half, by the appearance at surface of the light basal portion of the hairs. The basal portion of the hairs of the en- tire tail is a bright golden yellowish brown. Above this color is mostly hidden by the seal-brown tips, but beneath it forms a broad, clear area extending from base to and including pencil, and only a little darkened along extreme edge. Short hairs of under surface of tail less yellow than others, slightly grizzled with brown. Autumn pelage similar to summer coat but paler, the ferruginous of the back and sides replaced by ochraceous. The new coat first appears on shoulders and head, gradually extending backward and producing peculiar contrasts with the darker and brighter summer pelage. Skull. — The skull of Tupaia ckrysomalla so closely resembles that of specimens of T. ferruginea from Trong, Lower Siam, that I can see no characters to distinguish them. Teeth. — The teeth are exactly as in Tupaia ferruginea. Meastiremoits. — External measurements of type : total length 330; head and body 17S; tail vertebrae 152; pencil 30; hind foot 43 (40). Average of three specimens from Pulo Jimaja : total length 315 (305- 330); head and body iSo (171-191); tail vertebra 146 (140-152); hind foot 42.5 (42-43) ; hind foot without claws 41. Specimens examined. — Four : one from Pulo Siantan, the others from Pulo Jimaja. Re?}iarks. — Except for the differences due to pelage, the four speci- mens are very uniform in color. The yellow of the tail is occasionally darker than in the type, approaching tawny-ochraceous, but never tinged with red. Under surface of tail never grizzled as in T. pala- ivanensis. MAGADERMA SPASMA (Linnaeus). 1758. \_Vespertilio] spastna Linn/EUS, Systema Naturae I, loth ed., p. 32. (Asia.) Six specimens, two of which are skins, were collected on Pulo Bunoa, Tambelan Islands, August 5-7, and one on Big Tambelan Island. The color is faintly darker than in specimens from Trong, Lower Siam ; but in size as well as in cranial and dental characters the insular bats show no peculiarities. HIPPOSIDEROS BARBENSIS sp. nov. Type. — Adult male (skin and skull). No. 101625 United States National Museum. From Ste. Barbe Island. Collected August i, 1S99. 234 MILLER Characters. — Like specimens of Hipposideros larvat2isixovi\ Trong, Lower Siam, but smaller in size and duller in color. Color. — Back dark hair-brown, the hairs yellowish drab-gray through basal two-thirds, the extreme tips wood-brown. On shoulders and head the hair-brown disappears, giving place on the former to the drab-gray and on the latter to the wood-brown. Entire underparts rather light and dull wood-brown, the extreme bases of the hairs tinged with slaty. Ears and membranes blackish. Immature speci- mens are less yellowish than the adults, but are otherwise similar. Skull. — The skull of Hipposideros barbensis is throughout slightly smaller than that of mainland specimens of H. larvatus. Its form is peculiar in the shortness and relative breadth of the braincase ; but in all structural details it closely agrees with that of the Siamese ani- mal. Interpterygoid space shorter and relatively broader than in H. larvatus. Teeth. — The teeth are slightly smaller than those of Hipposideros larvatus., but in form they show no peculiarities. In one specimen there is on the right mandible a minute premolar wedged between the two normally present. Aleasurements. — For external measurements see table, page 240. Cranial measurements of type: greatest length 21 (23);^ basal length 18.2 (20); basilar length 16.2 (17.8); zygomatic breadth 12.4 (13); least interorbital breadth 3 (3.4); mastoid breadth 10.6 (11) ; greatest length of braincase 12 (14) ; greatest breadth of brain- case above roots of zygomata '$).6 (9.6) ; frontopalatal depth (at mid- dle of molar series) 4 (4) ; depth of braincase 6.4 (6.4) ; maxillary toothrow (exclusive of incisor) 8.2 (9) ; mandible 14.4 (15) ; mandi- bular toothrow (exclusive of incisors) 9 (9.8) . Speci?nens cxaiuincd. — Five, all from the tj'^pe locality. Remarks. — By its dull color and small size this species is readily separable from its mainland representative, though in all details of structure and form the two animals are scarcely distinguished. .? RHINOLOPHUS ROUXII Temminck. 1835. Rhinolophus roiixii Temmixck, Monographies de Mammalogie, 11 (Huitieme Monographie), p. 30 b. Type locality, Calcutta. 1892. RhinolopJnis ajfinis rouxii Thomas, Ann. Mus. Civ. Stor. Nat. Geneva, ser. 2, x, p. 923. A bat in formalin, collected on Pulo Siantan, Anambas Islands, during September is apparently referable to the species recorded from ' Measurements in parenthesis are those of an adult female Hipposideros larvatus from Trong, Lower Siam. MAMMALS COLLECTED BY DR. W. L. ABBOTT 235 Burmah by Thomas, under the name Rhinolophiis affinis rouxii. In size it agrees with the measurements of the Burman specimens; while the ridge beneath the free margin of the noseleaf is much more de- veloped than in Rhinolophus affinis^ and might readily be described as a supplemental leaflet. Without further material it is impossible to decide whether the island species is identical with that from the peninsula, or whether either is the same as the animal described by Temminck. That all three are closely allied there can be no doubt. RHINOLOPHUS MINUTUS sp. nov. Type. — Adult male (in alcohol). No. 101715 United States National Museum, Pulo Siantan, Anambas Islands, September, 1899. Chat-acters, — Similar to Rhinolophus minor Horsfield, but with shorter ear and tibia. First lower premolar in contact with third. Color. — General color broccoli-brown, somewhat lighter below than above, the hairs everywhere with glossy pale tips, these most notice- able on underparts. Ears and membranes blackish brown. Skull a?zd teeth. — The skull is smaller than that of a specimen of Rhinolophus minor from the mainland of India, and the brain- case is more narrow, but otherwise no important differences are apparent. Teeth as in R. minor., but first and third lower premolars strongly in contact, the second, therefore, wholly external to the tooth row. JMeasu.re7nents. — For external measurements of Rhinolophus min- utus see table, page 240. Cranial measurements: greatest length 15 ; basal length 13.4; basi- lar length ii.S; zygomatic breadth 7.8; least interorbital breadth 2; mastoid breadth 7.8; greatest length of braincase 9.6; greatest breadth of braincase above roots of zygomata 6.8 ; frontopalatal depth (at middle of molar series) 3 ; depth of braincase 5 ; maxillary tooth- row (exclusive of incisors) 6.2; mandible 10,2; mandibular toothrow (exclusive of incisor) 6.4. Specijnens examined. — Two (both in alcohol) : one from Pulo Sian- tan, the other without definite locality. Remarks. — Although this species is closely related to Rhinolophus minor its differently proportioned leg and foot appear to present suf- ficient reason for its recognition as a local form. The measurements of the type as given by Dobson show that Rhinolophus niinor from Java (the type locality) agrees in proportions with specimens from the mainland. 236 MILLER EMBALLOXURA ANAMBEXSIS sp. nov. Type, adult female (in alcohol). Xo. 101716 United States National Museum, Pulo Mobur, Anambas Islands, August 26, 1S99. Characters. — Similar to Emballoimra peninsular is Miller^ from Trong, Lower Siam, and E. discolor Peters from the Philippine Is- lands, but with fur nowhere whitish at base. Skull more slender than that of E. peninsularis. Color. — General color above sepia, the hairs darkening nearl}* to seal-brown through terminal fourth, and becoming broccoli-brown at base. Beneath broccoli-brown tinged with drab, the extreme bases of the hairs slightly paler. Skull. — The skull of Emballomira amanbensis (fig. 15, (?) differs from that of E. penitisnlaris (fig. 15, b') in its much more slender Fig. 15. a Etnballoriiira atiamltc/isis. b E. peninsularis. Twice natural size. general form. This is particularly noticeable in the braincase, but is visible in the rostral portion also. Viewed from the side and from below the differences are less apparent. TeetJi. — The teeth, though similar in form to those of Emballomira peninsular is .1 are smaller throughout, particularly the molars of the lower jaw. j\lcasure?nefits. — For external measurements of Emballomira an- ambensis^ see table, page 240. Cranial measurements of type: greatest length 15; basal length 13.2; basilar length 11.4; greatest antorbital breadth 6; least interor- ' Proc. Acad. Nat. Sci., Philadelphia, 1S9S, p. 324. MAMMALS COLLIX'TED 15Y DR. W. L. ABHOTT 237 bital breadth 3.4; zygomatic breadth S.S ; mastoid breadth 7.4; great- est length of braincase 9.6 ; gi^eatest width of braincase above roots of zygomata 7; depth of braincase 5.6; frontopalatal depth (at middle of molar series) i.S; maxillary toothrow (exclusive of incisors) 5.6; mandible 10.3; mandibular toothrow (exclusive of incisors) 5.8. Specimens examined. — Four (in alcohol) : two each from Pulo Mobur and Pulo Siantan, Anambas Islands. Remarks. — JSmballonura anajnbensis is a member of the group of large species containing E. discolor and E. peninsularis. From both of these it differs in the absence of the whitish bases of the hairs, while from the latter it is further distinguished by the form of the skull. Together with E. peninsularis it differs from E. semicazidata^ the type of which is still extant in the collection of the United States National Museum, in the nearly parallel sides of the tragus. I^EROPUS VAMPYRUS Linn^Eus. 1758, [^Vespertilio] va^npyrus Li'SNMVS, Systema Naturee, i, loth ed., p. 31. (Asia.) Eight specimens from Linga Island, July 19, 20, and 21. These appear to differ little if any from a skin secured in Java by Mr. D. G. Fairchild. PTEROPUS LEPIDUS sp. nov. Type. — Adult female (skin and skull). No. 101670 United States National Museum. From Kaju Ara, or Saddle Island, Tambelan Islands. Collected August 15, 1S99. Characters. — Similar to Pteropus condorettsis Peters from Pulo Condor, off the coast of Lower Cochin China, but quite different in color ; the shining light gray back sharply contrasted with the dark reddish brown or blackish brown shoulders and neck. Fur. — The fur of the back is silky in texture and closely appressed ; the individual hairs about 12 mm. in length. On rump it becomes slightly woolly, but much less so than on other parts of the body. It _ extends down legs to knee, beyond which a narrow line runs down inner side of tibia to middle. Fur of shoulders, head, neck, and un- derside of body loose and somewhat woolly in texture, not at all ap- pressed, and therefore appearing much longer than that of back, though the individual hairs are only about 14 mm. in length. It ex- tends on lower side of legs to knees, and on humerus nearly to elbow, also sparsely over lower side of propatagium and on wing membrane along basal half of forearm. Upper surface of membranes naked except for a sprinkling of hairs along forearm. 238 MILLER Color. — Back a mixture of peai'l-gray and pale broccoli-brown, the latter predominating posteriorly, the former anterioi'ly and along edge of wing membranes. The hairs have a distinct silky gloss. Shoulders, neck, chest, and upper side of head chestnut clouded irregularly with seal-brown and fading to light russet on forehead. Cheeks, region behind ears, and whole under side of head and throat seal brown faintly tinged with chestnut, and irregularly sprinkled with glistening Fig. iC. Ptcropiis lepidus. (Tvpe). Natural size. silvery hairs. Underparts of body mostly broccoli-brown, irregularly shaded with seal-brown at sides of chest and \\-\l\\ an undefined yel- lowish brown on middle of chest. Ears, membranes, and muzzle blackish. The four individuals show marked variation in color, chiefly through extension of the seal Inown areas at the expense of MAMMALS COLLECTED BY DR. W. L. ABBOTT 230 the chestnut. In one specimen the chestnut is ahnost hidden by the darker brown, though still visible beneath the surface when the fur is disarranged. Another skin, probably melanistic, is nearly uniform seal-brown throughout, though both chestnut of shoulders and gray of back are still faintly indicated. Skjill and teeth. — The skull and teeth of Pteropus Icpidtis (fig. 16) closely agree with that of Pterop2is condorensis as figured by Mat- schie.^ The palate appears to be narrower and the teeth less robust. ]\feasure??iefits. — For external measurements of Pteropus lepidzts^ see table, page 240. Cranial measurements of type : greatest length 64 ; basal length 60; basilar length 58; median palatal length 36; palatal breadth (between anterior molars) 11.4; zygomatic breadth 33.6; least inter- orbital breadth in front of postorbital processes 8 ; least interorbital breadth behind postorbital processes 7 ; breadth between tips of post- orbital processes 21.6 ; greatest breadth of braincase above roots of zygomata 22.8; greatest depth of braincase 17.8; occipital depth 1 1.8; depth of rostrum at middle of diastema 8 ; mandible 50; maxil- lary toothrow (exclusive of incisors) 25 ; mandibular toothrow (ex- clusive of incisors) 27.6; crown of third upper premolar 4.4 x 3; crown of first upper molar 5.6 x 3; crown of first lower molar 4.4 X 2.4. Specijnens cxa?mned. — Four : one from Saddle Island, the rest from Big Tambelan. Remarks. — This light bright colored species is evidently a near relative of the sombre Pteropus condornesis. It apparently needs no comparison with other species. SEMNOPITHECUS MAURUS (Schreber). 1775. Simia maura Schreber, Saugthiere, i, pi. xxii B. One specimen (adult female, No. 101601) taken on Linga Island, July 19, 1S99. MACACUS ' CYNOMOLGUS ' Auct. Two males of this species were taken on Linga Island, July 23, 1899, and another on Tioman Island, October 4. The Linga speci- mens agree in all respects with those previously collected by Dr. Abbott on the mainland. That from Tioman Island is smaller, but the peculiar condition of the skull shows it to be a dwarfed and ab- normal individual. ' Die Fledermiiuse der Berliner Museums fiir Naturkunde, pi. S, fig. 2. 240 MILLER < < g S u a o ;^ w H O o (I. Cfi H < pq o JB3 JO mpiAi r) N VO •nAVOiD raoaj aea li-i ■ ro >0 0 >i^ 0 t^ t^ t^ u^vo i^vo M M M N •JtSip pnoD3S VO t^ t^ ro ONVO 0 rorOfOO 0> 0 0 •;tSip 1^0 0 M 'f 1-1 M 0\ N t^ t^VO VO vO lOVO •nUBaJOji rcro -i-VO ■^ ro trjOO VD JBDIBD 0 0 « fO CO •^ rO "!i- ■* 10 •;ooj 00 -^ OD -t^ 0 ro« CO t^vD VO -"J- •* Tj- Tl- •BiqiX ^0 t^l>.vO 00 to !>• 10 1-1 w 1-1 \r^\o irjvD •IiBX ^'^ 1 1 1 1 coco ^ 1 1 1 1 •q;Sn3l CO t^ LCVO 00 000 Th M Ovvo M VO lovo M r< 1-1 c« f< M N c< ^)«X3 ^ 0+00+0+