THE BIOLOGICAL BULLETIN PUBLISHED BY THE MARINE BIOLOGICAL LABORATORY Editorial Board GARY N. CALKINS, Columbia University E. G. CONKLIN, Princeton University FRANK R. LlLLffi, University of Chicago E. N. HARVEY, Princeton University CARL R. MOORE, University of Chicago SELIG HECHT, Columbia University GEORGE T. MOORE, Missouri Botanical Garden LEIGH HOADLEY, Harvard University T. H. MORGAN, California Institute of Technology M. H. JACOBS, University of Pennsylvania G. H. PARKER, Harvard University H. S. JENNINGS, Johns Hopkins University W. M. WHEELER, Harvard University E. E. JUST, Howard University EDMUND B. WILSON, Columbia University ALFRED C. REDFIELD, Harvard University Managing Editor VOLUME LXIII AUGUST TO DECEMBER, 1932 Printed and Issued by LANCASTER PRESS, Inc. PRINCE & LEMON STS. LANCASTER, PA. 11 THE BIOLOGICAL BULLETIN is issued six times a year. Single numbers, $1.75. Subscription per volume (3 numbers), $4.50. Subscriptions and other matter should be addressed to the Biological Bulletin, Prince and Lemon Streets, Lancaster, Pa. Agent for Great Britain: Wheldon & Wesley, Limited, 2, 3 and 4 Arthur Street, New Oxford Street, London, W.C. 2. Communications relative to manuscripts should be sent to the Managing Editor, Marine Biological Laboratory, Woods Hole, Mass., between May 1 and October 1 and to the Institute of Biology, Divinity Avenue, Cambridge, Mass., during the re- mainder of the year. Entered October 10, 1902, at Lancaster, Pa., as second-class matter under Act of Congress of July 16, 1894. LANCASTER PRESS, INC. LANCASTER, PA. CONTENTS No. 1. AUGUST, 1932 PAGE THIRTY-FOURTH REPORT OF THE MARINE BIOLOGICAL LABORATORY 1 DAWSON, ALDEN B. The Reaction of the Erythrocytes of Vertebrates, Especially Fishes, to Vital Dyes 48 POWERS, PHILIP B. A. Cyclotrichiuni nieunieri Sp. Nov. (Protozoa, Ciliata) ; Cause of Red Water in the Gulf of Maine 74 ANDERSON, BERTIL GOTTFRID The Number of Pre-aclult Instars, Growth, Relative Growth, and Variation in Daphnia magna 81 DRAPER, JOHN W., AND DAYTON J. EDWARDS Some Effects of High Pressure on Developing Marine Forms . 99 PERKINS, EARLE B., AND BENJAMIN KROPP The Crustacean Eye Hormone as a Vertebrate Melanophore Activator 108 RICHARDS, OSCAR W., AND G. WELLFORD TAYLOR ' Mitogenetic Rays " -A Critique of the Yeast-Detector Method 113 CLAUSEN, H. J. Rate of Regeneration of Partly Histolyzed Anuran Tail Skin . 129 TYLER, ALBERT The Polarity of the Egg of Urechis caupo 145 No. 2. OCTOBER, 1932 FRY, HENRY J. Studies of the Mitotic Figure. I. Chaetopterus : central body structure at metaphase, first cleavage, after picro-acetic fixation 149 SONNEBORN, T. M. Experimental Production of Chains and its Genetic Conse- quences in the Ciliate Protozoan, Colpidium campylum (Stokes) 187 TYLER, ALBERT Chromosomes of Artificially Activated Eggs of Urechis 212 TYLER, ALBERT Production of Cleavage by Suppression of the Polar Bodies in Artificially Activated Eggs of Urechis 218 iii IV CONTENTS JACOBS, M. H., AND ARTHUR K. PARPART Osmotic Properties of the Erythrocyte. V. The rate of hemol- ysis in hypotonic solutions of electrolytes 224 BODINE, JOSEPH HALL, AND TITUS C. EVANS Hibernation and Diapause. Physiological changes during hi- bernation and diapause in the Mud-dauber Wasp, Sceliphron caementarium (Hymenoptera) 235 COOK, S. E. The Respiratory Gas Exchange in Termopsis nevadensis 246 BOWEN, EDITH S. Further Studies of the Aggregating Behavior of Ameiurus melas , 258 HERRICK, EARL H. Mechanism of Movement of Epidermis, especially its Melano- phores, in Wound Healing, and Behavior of Skin Grafts in Frog Tadpoles 271 LACKEY, JAMES P. Oxygen Deficiency and Sewage Protozoa: with Descriptions of Some New Species 287 WHITING, P. W. Modification of Traits in Mosaics from Binucleate Eggs of Habrobracon 296 WELSH, JOHN H. Temperature and Light as Factors Influencing the Rate of Swimming of Larvae of the Mussel Crab, Pinnotheres macu- latus Say 310 KEYS, ANCEL B., AND J. B. BATEMAN Branchial Responses to Adrenaline and to Pitressin in the Eel . 327 No. 3. DECEMBER, 1932 GOLDSCHMIDT, RlCHARD The Fourth Reynold A. Spaeth Memorial Lecture. Genetics and Development 337 DARLINGTON, C. D. The Origin and Behavior of Chiasmata. V. Chorthippus elegans 357 VI. Hyacinthus amethystinus 368 SPEICHER, B. R. Dominance of Two Kidney Allelomorphs in Habrobracon ju- glandis (Ash.) 372 BERRILL, N. J. The Mosaic Development of the Ascidian Egg 381 CONTENTS v NEWBY, W. W. The Early Embryology of the Echiuroid, Urechis 387 WINSOR, CHARLES P. AND AGNES A. Polyvitelline Eggs and Double Monsters in the Pond Snail Lymmea columella Say 400 HELFF, O. M. Studies on Amphibian Metamorphosis. X. Hydrogen-ion con- centration of the blood of anuran larvae during involution 405 COE, W. R. Sexual Phases in the American Oyster (Ostrea virginica) .... 419 FERRIS, JOSEPHINE CAROLYN A Comparison of the Life Histories of Mictic and Amictic Fe- males in the Rotifer, Hydatina senta 442 SHAPIRO, HERBERT The Rate of Oviposition in the Fruit Fly, Drosophila 456 ADDISON, WILLIAM H. F., AND MAURICE N. RICHTER A Note on the Thyroid Gland of the Swordfish (Xiphias glad- ius, L.) 472 ORIAS, OSCAR Influence of Hypophysectomy on the Pancreatic Diabetes of Dogfish : 477 SMITH, GEORGE MILTON Melanophores Induced by X-Ray Compared with those Ex- isting in Patterns as Seen in Carassius auratus 484 DAWSON, ALDEN B. Intracellular Crystallization of Hemoglobin in the Erythrocytes of the Northern Pipefish, Syngnathus fuscus 492 BARNES, T. CUNLIFFE Salt Requirements and Space Orientation of the Littoral Iso- pod Ligia in Bermuda 496 COOK, S. F., AND K. G. SCOTT The Relation between Absorption and Elimination of Water by Termopsis angusticollis 505 Vol. LXIII, No. 1 August, 1932 THE BIOLOGICAL BULLETIN PUBLISHED BY THE MARINE BIOLOGICAL LABORATORY THE MARINE BIOLOGICAL LABORATORY THIRTY-FOURTH REPORT FOR THE YEAR 1931- FORTY-FOURTH YEAR I. TRUSTEES AND EXECUTIVE COMMITTEE (AS OF AUGUST 11, 1931) 1 LIBRARY COMMITTEE II. ACT OF INCORPORATION 3 III. BY-LAWS OF THE CORPORATION 3 IV. REPORT OF THE TREASURER 5 V. REPORT OF THE LIBRARIAN 9 VI. REPORT OF THE DIRECTOR 11 Statement 11 Addenda : 1. The Staff, 1931 16 2. Investigators and Students, 1931 18 3. Tabular View of Attendance 29 4. Subscribing and Cooperating Institutions, 1931 ... 30 5. Evening Lectures, 1931 31 6. Shorter Scientific Papers, 1931 32 7. Members of the Corporation 37 I. TRUSTEES EX OFFICIO FRANK R. LILLIE, President of the Corporation, University of Chicago. MERKEL H. JACOBS, Director, University of Pennsylvania. LAWRASON RIGGS, JR., Treasurer, 25 Broad Street, New York City. CHARLES PACKARD, Clerk of the Corporation, Columbia University. GARY N. CALKINS, Secretary of the Board of Trustees, Columbia University. EMERITUS CORNELIA M. CLAPP, Mount Holyoke College. C. R. CRANE, New York City. H. H. DONALDSON, Wistar Institute of Anatomy and Biology. OILMAN A. DREW, Eagle Lake, Florida. WILLIAM PATTEN, Dartmouth College. W. B. SCOTT, Princeton University. E. B. WILSON, Columbia University. 1 1 2 MARINE BIOLOGICAL LABORATORY TO SERVE UNTIL 1935 H. C. BUMPUS, Brown University. GARY N. CALKINS, Columbia University. W. C. CURTIS, University of Missouri. B. M. DUGGAR, University of Wisconsin. L. V. HEILBRUNN, University of Pennsylvania. W. J. V. OSTERHOUT, Rockefeller Institute for Medical Research. WILLIAM M. WHEELER, Harvard University. LORANDE L. WOODRUFF, Yale University. TO SERVE UNTIL 1934 E. R. CLARK, University of Pennsylvania. E. G. CON KLIN, Princeton University. OTTO C. GLASER, Amherst College. Ross G. HARRISON, Yale University. E. N. HARVEY, Princeton University. H. S. JENNINGS, Johns Hopkins University. F. P. KNOWLTON, Syracuse University. M. M. METCALF, Johns Hopkins University. TO SERVE UNTIL 1933 H. C. BRADLEY, University of Wisconsin. H. B. GOODRICH, Wesleyan University. I. F. LEWIS, University of Virginia. R. S. LILLIE, University of Chicago. C. E. McCLUNG, University of Pennsylvania. T. H. MORGAN, California Institute of Technology. A. C. REDFIELD, Harvard University. D. H. TENNENT, Bryn Mawr College. TO SERVE UNTIL 1932 R. CHAMBERS, Washington Square College, New York University. W. E. CARREY, Vanderbilt University Medical School. CASWELL GRAVE, Washington University. M. J. GREENMAN, Wistar Institute of Anatomy and Biology. R. A. HARPER, Columbia University. A. P. MATHEWS. University of Cincinnati. G. H. PARKER, Harvard University. C. R. STOCKARD, Cornell University Medical College. EXECUTIVE COMMITTEE OF THE BOARD OF TRUSTEES FRANK R. LILLIE, Ex Off. Chairman. MERKEL H. JACOBS, Ex. Off. LAWRASON RIGGS, JR., Ex. Off. W. C. CURTIS, to serve until 1932. A. C. REDFIELD. to serve until 1932. E. G. CONKLIN, to serve until 1933. CHARLES PACKARD, to serve until 1933. ACT OF INCORPORATION THE LIBRARY COMMITTEE C. E. McCLUNG, Chairman. ROBERT A. BUDINGTON. E. E. JUST. M. M. METCALF. A. H. STURTEVANT. II. ACT OF INCORPORATION No. 3170 COMMONWEALTH OF MASSACHUSETTS Be It Known, That whereas Alpheus Hyatt, William Sanford Stevens, William T. Sedgwick, Edward G. Gardiner, Susan Minns, Charles Sedg- wick Minot, Samuel Wells, William G. Farlow, Anna D. Phillips and B. H. Van Vleck have associated themselves with the intention of forming a Corporation under the name of the Marine Biological Laboratory, for the purpose of establishing and maintaining a laboratory or station for scien- tific study and investigation, and a school for instruction in biology and natural history, and have complied with the provisions of the statutes of this Commonwealth in such case made and provided, as appears from the cer- tificate of the President, Treasurer, and Trustees of said Corporation, duly approved by the Commissioner of Corporations, and recorded in this office ; Now, therefore, I, HENRY B. PIERCE, Secretary of the Commonwealth of Massachusetts, do hereby certify that said A. Hyatt, W. S. Stevens, W. T. Sedgwick, E. G. Gardiner, S. Minns, C. S. Minot, S. Wells, W. G. Farlow, A. D. Phillips, and B. H. Van Vleck, their associates and suc- cessors, are legally organized and established as, and are hereby made, an existing Corporation, under the name of the MARINE BIOLOGICAL LABORATORY, with the powers, rights, and privileges, and subject to the limitations, duties, and restrictions, which by law appertain thereto. Witness my official signature hereunto subscribed, and the seal of the Commonwealth of Massachusetts hereunto affixed, this twentieth day of March, in the year of our Lord One Thousand Eight Hundred and Eighty- Eight. [SEAL] HENRY B. PIERCE, Secretary of the Commonwealth. III. BY-LAWS OF THE CORPORATION OF THE MARINE BIOLOGICAL LABORATORY I. The annual meeting of the members shall be held on the second Tuesday in August, at the Laboratory, in Woods Hole, Mass., at 11.30 A.M., daylight saving time, in each year, and at such meeting the members shall choose by ballot a Treasurer and a Clerk to serve one year, and eight Trustees to serve four years. There shall be thirty-two Trustees thus chosen divided into four classes, each to serve four years, and in addition there shall be two groups of Trustees as follows: (a) Trustees ex officio, who shall be the MARINE BIOLOGICAL LABORATORY President of the Corporation, the Director of the Laboratory, the Associate Director, the Treasurer and the Clerk; (b) Trustees Emeritus, who shall be elected from the Trustees by the Corporation. Any regular Trustee who has attained the age of seventy years shall continue to serve as Trustee until the next annual meeting of the Corporation, whereupon his office as regular Trustee shall become vacant and be filled by election by the Cor- poration and he shall become eligible for election as Trustee Emeritus for life. The Trustees ex officio and Emeritus shall have all rights of the Trustees except that Trustees Emeritus shall not have the right to vote. The Trustees and officers shall hold their respective offices until their successors are chosen and have qualified in their stead. II. Special meetings of the members may be called by the Trustees to be held in Boston or in Woods Hole at such time and place as may be designated. III. The Clerk shall give notice of meetings of the members by pub- lication in some daily newspaper published in Boston at least fifteen days before such meeting, and in case of a special meeting the notice shall state the purpose for which it is called. IV. Twenty-five members shall constitute a quorum at any meeting. V. The Trustees shall have the control and management of the affairs of the Corporation ; they shall present a report of its condition at every annual meeting; they shall elect one of their number President of the Cor- poration who shall also be Chairman of the Board of Trustees ; they shall appoint a Director of the Laboratory; and they may choose such other officers and agents as they may think best; they may fix the compensation and define the duties of all the officers and agents; and may remove them, or any of them, except those chosen by the members, at any time ; they may fill vacancies occurring in any manner in their own number or in any of the offices. They shall from time to time elect members to the Corporation upon such terms and conditions as they may think best. VI. Meetings of the Trustees shall be called by the President, or by any two Trustees, and the Secretary shall give notice thereof by written or printed notice sent to each Trustee by mail, postpaid. Seven Trustees shall constitute a quorum for the transaction of business. The Board of Trustees shall have power to choose an Executive Committee from their own number, and to delegate to such Committee such of their own powers as they may deem expedient. VII. The accounts of the Treasurer shall be audited annually by a certified public accountant. VIII. The consent of every Trustee shall be necessary to dissolution of the Marine Biological Laboratory. In case of dissolution, the property shall be disposed of in such manner and upon such terms as shall be de- termined by the affirmative vote of two-thirds of the Board of Trustees. IX. These By-laws may be altered at any meeting of the Trustees, pro- vided that the notice of such meeting shall state that an alteration of the By-laws will be acted upon. X. Any member in good standing may vote at any meeting, either in person or by proxy duly executed. REPORT OF THE TREASURER IV. THE REPORT OF THE TREASURER To THE TRUSTEES OF THE MARINE BIOLOGICAL LABORATORY : Gentlemen: Herewith is submitted my report as Treasurer of the Marine Biological Laboratory for the year 1931. The accounts have been audited by Seamans, Stetson and Tuttle, certified public accountants. A copy of their report is on file at the Laboratory and is open to inspection by members of the Corporation. At the end of the year 1931, the book value of the General Endow- ment Fund in the hands of the Central Hanover Bank and Trust Com- pany (of New York) as Trustee was $908,895 in securities and $53.77 in cash. The book value of the Library Fund was $198,605 in securities, and $340 in cash. The Reserve Fund consisted of securities of the book value of $20,- 868.75 and cash of $909.75. The Retirement Fund consisted of securities of the book value of $18,896.07, invested in mortgages. There has been little change in the other minor funds. At this date (March 28, 1932) there is no default in any of the securities held in the above-mentioned funds. The land, buildings, equipment, and library, excluding the Devil's Lane and Gansett properties, represented an investment of $1,642,- 665.24, less depreciation of $286,404.20, or a net amount of $1,356,- 261.04. Current income exceeded expenses including depreciation by $2,304.35. Over $21,000 was expended from current funds on buildings, equip- ment, and on books, the greater part having been expended for books. At the end of the year, the Laboratory owed $2,349.65 in accounts payable, and $27,000 on bond and mortgage, and had over $30,000 in its bank accounts. Following is the Balance Sheet as of December 31, 1931, and the condensed statement of income and outgo for the year, also the surplus account. s •*"' |LU I L I B -n». 6 MARINE BIOLOGICAL LABORATORY EXHIBIT A MARINE BIOLOGICAL LABORATORY BALANCE SHEET, DECEMBER 31, 1931 Assets Endowment Assets and Equities : Securities and Cash in Hands of Central Hanover Bank & Trust Company (of New York), Trustee— Schedules I-a and I-b $1,107,893.77 Securities and Cash — Minor Funds — Schedule II 10,411.70 $1,118,305.4) Plant Assets : Land— Schedule IV $ 97,103.05 Buildings— Schedule IV 1,207,554.14 Equipment— Schedule IV 162,965.48 Library— Schedule IV 175,042.57 $1,642,665.24 Less Reserve for Depreciation 286,404.20 $1,356,261.04 Securities and Cash in Reserve Fund 21,778.50 Cash in Dormitory Building Fund 818.96 $1,378,858.50 Current Assets : Cash $ 30,872.07 Accounts — Receivable 17,998.23 Inventories : Supply Department $ 36,327.34 Biological Bulletin 9,077.28 45,404.62 Investments : Devil's Lane Property $ 39,301.81 Gansett Property 1.00 Stock in General Biological Supply House, Inc 12,700.00 Retirement Fund Assets 18,896.07 70,898.88 Prepaid Insurance 3,902.12 Items in Suspense (Net) 292.12 $ 169,368.04 $2,666,532.01 Liabilities Endowment Funds : General Endowment Funds— Schedule III $1,107,893.77 Minor Funds— Schedule III 10,411.70 $1,118,305.47 Plant Funds : Donations and Gifts— Schedule III $1,029,372.61 Other Investments in Plant from Gifts and Cur- rent Funds . 347,485.89 $1,376,858.50 Mortgage, Danchakoff Estate 2,000.00 $1,378,858.50 REPORT OF THE TREASURER Current Liabilities and Surplus : Mortgage, Devil's Lane Property $ 25,000.00 Accounts — Payable 2,349.65 Woods Hole Oceanographic Institution : Amount received for Purchase of Books for their Library $ 5,000.00 Less Expenditures 4,327.63 672.37 $ 28,022.02 Current Surplus— Exhibit C 141,346.02 $ 169,368.04 $2,666,532.01 EXHIBIT B MARINE BIOLOGICAL LABORATORY INCOME AND EXPENSE, YEAR ENDED DECEMBER 31, 1931 Total Net Expense Income Expense Income Income : General Endowment Fund ... $ 47,293.49 $ 47,293.49 Library Fund 10,434.77 10,434.77 Gifts 445.29 445.29 Instruction 8,110.79 9,425.00 1,314.21 Research 4,063.36 18,085.00 14,021.64 Evening Lectures 167.58 167.58 Biological Bulletin and Mem- bership Dues 9,060.65 9,976.43 915.78 Supply Department — Schedule V 44,834.46 57,979.24 13,144.78 Mess— Schedule VI 30,664.25 34,360.65 3,696.40 Dormitories — Schedule VII 31,739.30 13,500.86 18,238.44 (Interest and Depreciation charged to above 3 Depart- ments. See Schedules V, VI, and VII) 35,453.91 35,453.91 Dividends, General Biological Supply House, Inc 2,032.00 2,032.00 Rent, Danchakoff Cottages .... 505.17 1,039.00 533.83 Rent, Microscopes 469.00 469.00 Rent, Garage, Railway, etc. ... 427.60 427.60 Rent, Newman Cottage 92.94 150.00 57.06 Rent, Janitor's House 118.83 390.00 271.17 Sales of Duplicate Library Sets 238.70 238.70 Interest on Bank Balances 213.76 213.76 Sundry Items 71.82 71.82 Maintenance of Plant : New Laboratory Expense .. 18,057.74 18,057.74 Chemical and Special Appa- ratus 9,869.52 9,869.52 Maintenance, Buildings and Grounds 9,147.32 9,147.32 MARINE BIOLOGICAL LABORATORY Library Department Expenses 8,982.70 8.982.70 Carpenter Department Ex- penses 1,602.02 1,602.02 Truck Expenses 942.02 942.02 Sundry Expenses 282.91 282.91 Workmen's Compensation In- surance 554.02 554.02 General Expenses : Administration Expenses 15,907.89 15,907.89 Endowment Fund Trustee . . . 968.50 968.50 Interest on Loans 100.00 100.00 Bad Debts 631.24 631.24 Naples Zoological Station . . . 250.00 250.00 Mosquito Fund Contribution . 100.00 100.00 Reserve for Depreciation 39,778.56 39,778.56 Museum Expenses 3,150.40 3,150.40 Excess of Income over Expenses carried to Current Surplus — Exhibit C . 2,304.35 2,304.35 $206,532.61 $206,532.61 $131,035.21 $131,035.21 EXHIBIT C MARINE BIOLOGICAL LABORATORY, CURRENT SURPLUS ACCOUNT YEAR ENDED DECEMBER 31, 1931 Balance, January 1, 1931 $119,401.09 Add: Reserve for Depreciation charged to Plant Funds 39,778.56 Excess of Income over Expenses for Year as shown in Exhibit B 2,304.35 Excess of Gansett Property Receipts over Cost of Property and Development Expenses, etc 1,885.95 Income of Retirement Fund $ 734.47 Less Pensions Paid . 720.00 14.47 $163,384.42 Deduct : Payments from Current Funds during Year for Plant Assets as shown in Schedule IV, Buildings $ 200.11 Equipment 3,919.20 Library Books, etc 17,635.22 $21,754.53 Income and Reserve Fund for 1931 credited to Current Surplus and now transferred to Plant Funds 283.87 22,038.40 Balance, December 31, 1931— Exhibit A $141,346.02 Respectfully submitted, LAWRASON RIGGS, JR., Treasurer. REPORT OF THE LIBRARIAN V. THE REPORT OF THE LIBRARIAN The expenditures of the Library remain the same as last year, $24,000 for the Marine Biological Laboratory, and $2,123.21 for the Oceanographic Institution. The $5,000 appropriated by the last-named in March, 1930, will be expended before the end of their fiscal year of February 29, 1932, for necessary books and back sets. A separate rendering of this account is kept on file. The only item in the regular $24,000 necessary to mention is that for current serials, which has, during 1930 and 1931, overreached the $5.000 assigned in 1929, by $72.94 and $631.07 respectively. In 1932, $6,000 is allowed for cur- rent serials. This increase was, however, anticipated in 1929 and the definite plan made to enlarge the subscriptions each year by a sum taken from the $8,200 assigned in 1929 for back sets. Until the back sets of 450 different serials are purchased which are very necessary to the Li- brary before it is on a par with the finest libraries of the country, it is the judgment of the Librarian that the current expenditure on serials should not greatly exceed $6,000, thus leaving for a number of years the $7,000 that will, if wisely spent in completing back sets, increase the usefulness and monetary value of the Library much beyond the intrinsic sum spent. The full list of over 1,000 serial sets imperfect in the back holdings or in current receipts has been carefully scrutinized during the winter, 1931-32, and definite decision has been made, based on various different and combined reasons, to borrow or to leave unfilled about half of the number. Since every serial title in the Library, regardless of the method of acquisition and of its value, has been automatically recorded when deficient, the weeding out of half of these seems reasonable when choice was made without regard to any definite number to be retained or discarded, but the decisions are subject, of course, to future revision on the part of the investigators. Before the date of the summer report an estimate will have been made of the definite sum of money that will be necessary in order that the Library shall record in its catalogue 1,319 complete serial publications as against 869 perfectly complete now. A sum of $1,000 should always be available under this item, however, since new current serials will come on our list entailing the purchase in most cases of previously issued volumes. A statement of the holdings of the Library shows the following: 1,080 currently received serials; 33,780 volumes; and 69,851 reprints. A detail of the current serials, 391 subscriptions, 481 exchanges, 208 gifts, shows an increase over last year in current paid subscriptions by 45, of which 14 were for the Woods Hole Oceanographic Institution, and 39 new exchanges. The low total increase from the year 1930 to 10 MARINE BIOLOGICAL LABORATORY 1931 from 1,060 to 1,080 is due to a rigid elimination of gifts and ex- changes not regularly received, and will be enhanced another year when a selection will have been made of those important to us. As a tempo- rary measure, a great number of incomplete and irregularly received United States Government publications have been entirely segregated from the serials, pending necessary decision as to their value for us. The number of bound volumes added to the Library in 1931 was 1,923; of these 1,576 were serials, 48 sets having been filled in, seven of these for the Oceanographic, 33 partially filled in, two of these for the Oceanographic, and 347 were books. The unusual number of 156 books was purchased by us, using, in addition to the budgeted $300, a sum of $500 from the " back sets " money, and 95 were purchased for the Oceanographic Institution. In calling attention to the 33,780 vol- umes in the Library at the present time, it will not be out of place to record the fact that the accessioned number of volumes is a misleading figure as to the actual number of serial volumes the Library contains. For the sake of economy in binding and in order to coordinate thickness and height in volumes and make them look well as they stand on the shelves, we have during the past ten years increased rather than dimin- ished the number in which we bind two or more volumes in one, thus reducing consistently the accessioned number below the actual. The shelf space occupied by the volumes indicates, indeed, a figure much more nearly 40,000 than the correct 33,780. The estimated capacity of the shelving space in the Library as given by the architect at 100,000 is correct for the small size volumes of our serial sets. Since one of the five floors of stacks must be reserved for reprints, the inference may be made that the space for serials and books will be completely exhausted only when the number 33,780 volumes, bound and accessioned in ac- cordance with our present method, is doubled. The reprints added to the files this year were 5,620; 2,745 of these were catalogued from the gift of Professor Metcalf and the Sidney I. Smith collection, and 2,875 total the year's current receipts. The ma- jority of current reprints are from the authors, but the " Collected Papers " of 63 different laboratories are regularly received. Of these sets of reprints, 27 are complete and 36 are incomplete in the back files If these were counted as serial publications our list of current serial receipts would be 1,143 instead of 1,080. Of the reprints 1,828 are bound. Three hundred and seventy-five of these volumes contain 10-40 separates. While the gifts of books from publishers are many fewer than some years ago when special effort was made to secure presentation copies direct from the publisher, the number from authors was exceptionally REPORT OF THE DIRECTOR 11 high. This was due to Professor Baitsell and to Mr. Ware Cattell, who very generously presented to the Library the hooks reviewed during the summer in the " Collecting Net." A book plate records the gift as a joint one from the author, publisher, reviewer, and the " Collecting Net." The books presented to the Library this year are briefly enu- merated below. These gifts are acknowledged with pleasure, and with very warm thanks. "Collecting Net" 19 Authors 14 P. Blakiston's Son and Co 4 Bruce Publishing Co 1 Chicago University Press 1 Harvard University Press 1 John Wiley & Sons, Inc 1 Alfred A. Knopf 1 Lea & Febiger 1 Macmillan Co 2 W. B. Saunders & Co 1 Dr. Henry McE. Knower 3 Dr. C. A. Cheever 1 National Academy of Sciences 1 Eastman Kodak Company 1 VI. THE REPORT OF THE DIRECTOR To THE TRUSTEES OF THE MARINE BIOLOGICAL LABORATORY : Gentlemen: I beg to submit herewith a report of the forty- fourth session of the Marine Biological Laboratory for the year 1931. 1. Attendance. The attendance for 1931 was unexpectedly large, the total number of investigators for the season being 362 as compared with 337 in 1930, the next-highest year. The tabular view of attendance on page 29 shows in detail the relation of the past season to the four immediately preceding it. Particularly noteworthy is the large number of institutions represented in 1931, this being 137 as compared with 126 in 1930 and 123 in 1929. Of considerable importance to the Laboratory from a practical standpoint is the peak attendance. It is this, rather than the total at- tendance, which determines the degree of crowding of the Laboratory, the Mess, and the available living accommodations. Following the high peak of 286 in 1928, a measure of relief was afforded for a time by a new arrangement of the courses, which tended to spread the attendance over a longer period; and, for several years, a further growth of the total attendance occurred without evidences of undue crowding. Dur- ing the past season, however, the peak attendance suddenly increased by 43 over that for 1930 and surpassed by 20 even the previous high 12 MARINE BIOLOGICAL LABORATORY record of 1928. Since it seems impracticable at present for tbe Lab- oratory to increase the available research and living accommodations, the attention of investigators is again invited to the advantages of utiliz- ing to a greater extent than at present the early and the late parts of the season. The following tabulation of attendance on selected dates for the past five years gives a clear view of the general situation. 1927 1928 1929 1930 1931 30 . 7 15 June 10 50 64 55 50 51 ft 20 114 140 139 153 153 U 30 212 240 197 208 217 July 10 247 281 238 253 258 a 20 247 282 242 250 273 a 30 245 272 249 253 281 August 10 234 250 256 254 302 i< 20 208 226 243 245 280 it 30 168 183 220 204 239 September 10 110 112 157 122 136 li 20 50 43 59 44 69 it 30 . 12 14 14 8 14 2. The Report of the Treasurer. This report shows a slight in- crease in the total assets of the Laboratory over the preceding year, the figures being $2,666,532.01 for 1931 as compared with $2,660,559.11 for 1930. Though the income for 1931 was less than that for 1930 ($206,532.61 as compared with $210,110.86), a reduction of expenses from $213,878.11 to $204,228.36 permitted for the first time in six years the appearance of an excess of income over expenses, including depre- ciation charges. While the Laboratory was not greatly affected during 1931 by the existing financial depression, the indications at the time of the preparation of this report are that the year 1932 will show a serious decrease in the subscriptions received from cooperating and subscrib- ing institutions and somewhat smaller decreases in the income from the Supply Department and from the permanent endowment fund. It is hoped, however, that by means of all practicable economies it may be possible to preserve the very gratifying condition of financial soundness that has characterized the Laboratory for so many years, even though the exceptionally favorable showing for 1931 may not again be equalled for some time. 3. The Report of the Librarian. The growth of the Library in 1931 has continued at approximately the same rate as that for the past six years. Its development since 1925 may perhaps best be shown by means of the following figures taken in part from the reports of previous years. REPORT OF THE DIRECTOR 13 Particularly noteworthy is the fact that 869 sets of serial publications are now complete; these include most of the sets in common use. 1925 1926 1927 1928 1929 1930 1931 Serials received cur- rently 500 628 764 874 985 1060 1080 Total number of bound volumes 15000 18200 22800 26500 28300 31500 33800 Reprints 25000 38000 43000 51000 59000 64000 70000 4. Publications. During the early years of the Laboratory, a record was kept of all the published scientific work that issued from it. With its continued growth, however, the difficulty of keeping such a record became very great and the practice was discontinued about twenty years ago. The development in recent years of a trained library staff accus- tomed to the collection and the cataloguing of reprints has again made it possible, not merely to keep a record of all publications, but to bind and index the papers themselves in such a way that they may be readily available both for the use of investigators and of other persons inter- ested in the scientific accomplishments of the Laboratory. Through the kind cooperation of the investigators concerned, approximately 100 papers, based on work done at the Laboratory and published in 1930, are now ready for binding. It is planned to prepare similar sets for each subsequent year ; and it is to be expected that the value of these sets, both historical and scientific, will in time become very great. The thanks of the Laboratory are due to all who have so generously con- tributed the necessary reprints of their papers. 5. Lectures and Scientific Meetings. Twelve evening lectures were delivered during 1931, including the special Reynold A. Spaeth Me- morial Lecture by Professor Ross G. Harrison. In addition, there were held 12 other meetings, at which 63 shorter scientific papers were pre- sented and discussed. A successful innovation in 1931 was the special scientific session held on September 3 and devoted exclusively to work completed at the Lab- oratory during the current season. Twenty-two papers conveniently grouped by subjects were presented at this session, which occupied the greater part of the day. The remainder of the day was devoted to the inspection of demonstrations of work in progress at the Laboratory. The titles of the various lectures and shorter papers for 1931, which give a very representative cross-section of the work of the Laboratory for that year, are listed on pages 31 to 36. 6. Supply Department and Museum. The period of management of the Supply Department by the General Biological Supply House 14 MARINE BIOLOGICAL LABORATORY having terminated in August, 1931, Mr. James Mclnnis, who for the past year had acted as Resident Manager, was placed in full charge of the Department. Under his management the Supply Department has continued, in spite of generally unfavorable business conditions, to pro- vide an important part of the revenue needed for the running of the Laboratory, besides filling with efficiency its primary function of sup- plying living material to investigators. During the past year very gratifying progress has been made in the development of a working museum by its Curator, Mr. George M. Gray. A large number of representatives of the local fauna, properly pre- served and labeled, are now available for the use of investigators who wish to identify material of their own or whose work in other ways requires museum facilities. In addition to the preservation of material, particular attention has been given during the year to the accumulation and tabulation of data on the geographical and seasonal distribution of the local forms. Mr. Gray's long experience with the fauna of the Woods Hole region gives him exceptional qualifications for carrying out this very important part of the work of the Museum. 7 . Meeting of the Corporation. At the annual meeting of the Cor- poration, held on August 11, the report of a special committee appointed a year previously to consider changes in the method of nominating trustees was considered, and the following recommendations of the Committee were adopted : 1. " The Corporation affirms its position that instruction is a fundamental part of the work of the Laboratory and hence this work should be adequately represented upon the Board of Trustees." 2. " That the Committee of the Corporation for nomination of Trustees consist of five members, of whom not less than two shall be Trustee members and not less than two shall be non-Trustee members of the Corporation." 3. " That on or about July first of each year, the Clerk shall send a cir- cular letter to each member of the Corporation giving the names of the Nominating Committee and stating that this Committee desires suggestions regarding nominations." 4. " That the Nominating Committee shall post the list of nominations at least one week in advance of the annual meeting of the Corporation." A recommendation that no trustee shall be eligible for re-election until one year after the expiration of the term for which he was elected was discussed at length, but the motion to accept it was lost. The following new trustees were elected by the Corporation: L. V. Heilbrunn (Class of 1935), H. B. Goodrich (Class of 1933). Following the request of Dr. Gary N. Calkins, who for 19 years had ably served the Laboratory as Clerk of the Corporation, that his name be not presented to the Corporation for re-election, the Nominating Com- REPORT OF THE DIRECTOR 15 mittee selected in his place Dr. Charles Packard, of Columbia University, who was duly elected to the position, thereby becoming at the same time a Trustee ex-officio. Dr. Calkins, who on his retirement from office ceased to be a Trustee ex-officio, was elected to fill one of the regular positions in the Class of 1935. A resolution of appreciation of the long and valued services of Dr. Calkins as its Clerk was adopted by a rising vote of the Corporation. 8. Changes in tlic B \-latvs. At the annual meeting of the Board of Trustees two changes in the By-laws were approved. Those parts of the By-laws affected by the changes in question now read, as amended : 1. " The annual meeting of the members shall be held on the second Tuesday in August, at the Laboratory, in Woods Hole, Mass., at 11 :30 A.M., daylight saving time, in each year, and at such meeting the members shall choose by ballot a Treasurer and a Clerk to serve one year, and eight Trus- tees to serve four years.'' 2. " Trustees ex efficio and Emeritus shall have all rights of the Trustees except that Trustees Emeritus shall not have the right to vote." 9. Gifts. Appreciative acknowledgment is made of the assistance of the Committee on the Effects of Radiation of the National Research Council and of various persons, who through the efforts of this Com- mittee became interested in the needs of the Laboratory, in connection with investigations requiring the use of X-rays and other types of radiation. Without this assistance much of the work accomplished in this field during the past year would have been impossible. The thanks of the Laboratory are also due to Mrs. J. C. Hemmeter for the gift of scientific apparatus formerly belonging to her husband, the late Dr. John C. Hemmeter, and to Mr. Ware Cattell for the continuation in 1931 of the Collecting Net Scholarships and for the gift to the Library of a considerable number of books. There are appended as parts of this report : 1. The Staff, 1931. 2. Investigators and Students, 1931. 3. A Tabular View of Attendance, 1927-1931. 4. Subscribing and Cooperating Institutions, 1931. 5. Evening Lectures, 1931. 6. Shorter Scientific Papers, 1931. 7. Members of the Corporation, August, 1931. Respectfully submitted, M. H. JACOBS, Director. 16 MARINE BIOLOGICAL LABORATORY 1. THE STAFF, 1931 MERKEL H. JACOBS, Director, Professor of General Physiology, University of Pennsylvania. Associate Director: — ZOOLOGY I. INVESTIGATION GARY N. CALKINS, Professor of Protozoology, Columbia University. E. G. CONKLIN, Professor of Zoology, Princeton University. CASWELL GRAVE, Professor of Zoology, Washington University. H. S. JENNINGS, Professor of Zoology, Johns Hopkins University. FRANK R. LILLIE, Professor of Embryology, University of Chicago. C. E. McCLUNG, Professor of Zoology, University of Pennsylvania. S. O. MAST, Professor of Zoology, Johns Hopkins University. T. H. MORGAN, Director of the Biological Laboratory, California Institute of Technology. G. H. PARKER, Professor of Zoology, Harvard University. E. B. WILSON, Professor of Zoology, Columbia University. LORANDE L. WOODRUFF, Professor of Protozoology, Yale University. II. INSTRUCTION J. A. DAWSON, Assistant Professor of Biology, College of the City of New York. T. H. BISSONNETTE, Professor of Biology, Trinity College. E. C. COLE, Associate Professor of Biology, Williams College. O. E. NELSEN, Instructor in Zoology, University of Pennsylvania. A. W. POLLISTER, Instructor in Zoology, Columbia University. L. P. SAYLES, Instructor in Biology, College of the City of New York. A. E. SEVERINGHAUS, Assistant Professor of Anatomy, College of Phy- sicians and Surgeons, Columbia University. JUNIOR INSTRUCTORS B. R. COONFIELD, Professor of Biology, Southwestern College. C. E. HADLEY, Assistant Professor of Biology, New Jersey State Teachers College at Montclair. PROTOZOOLOGY I. INVESTIGATION (See Zoology) II. INSTRUCTION GARY N. CALKINS, Professor of Protozoology, Columbia University. RACHEL BOWLING, Instructor in Zoology, Columbia University. W. BYERS UNGER, Assistant Professor of Zoology, Dartmouth College. REPORT OF THE DIRECTOR 17 EMBRYOLOGY I. INVESTIGATION (See Zoology) II. INSTRUCTION HUBERT B. GOODRICH, Professor of Biology, Wesleyan University. BENJAMIN H. GRAVE, Professor of Biology, De Pauw University. LEIGH HOADLEY, Professor of Zoology, Harvard University. CHARLES PACKARD, Assistant Professor of Zoology, Institute of Cancer Research, Columbia University. HAROLD H. PLOUGH, Professor of Biology, Amherst College. PHYSIOLOGY I. INVESTIGATION HAROLD C. BRADLEY, Professor of Physiological Chemistry, University of Wisconsin. WALTER E. CARREY, Professor of Physiology, Vanderbilt University Med- ical School. RALPH S. LILLIE, Professor of General Physiology, University of Chicago. ALBERT P. MATHEWS, Professor of Biochemistry, University of Cincinnati. II. INSTRUCTION Teaching Staff WILLIAM R. AMRERSON, Professor of Physiology, University of Tennessee. PHILIP BARD, Assistant Professor of Physiology, Harvard Medical School. RALPH W. GERARD, Assistant Professor of Physiology, University of Chi- cago. LAURENCE IRVING, Associate Professor of Physiology, University of Toronto. LEONOR MICHAELIS, Member of the Rockefeller Institute, New York City. MARGARET SUM WALT, Assistant Professor of Physiology, Woman's Medical College of Pennsylvania. Special Lecturers EDWIN J. COHN, Associate Professor of Physical Chemistry, Harvard Uni- versity. HENRY J. FRY, Associate Professor of Biology, Washington Square College, New York University. E. NEWTON HARVEY, Professor of Physiology, Princeton University. SELIG HECHT, Professor of Biophysics, Columbia University. MERKEL H. JACOBS, Professor of General Physiology, University of Penn- sylvania. BALDUIN LUCRE, Associate Professor of Pathology, University of Pennsyl- vania. BOTANY I. INVESTIGATION B. M. DUGGAR, Professor of Physiological and Economic Botany, University of Wisconsin. C. E. ALLEN, Professor of Botany, University of Wisconsin. S. C. BROOKS, Professor of Zoology, University of California. IVEY F. LEWIS, Professor of Biology, University of Virginia. WM. J. ROBBINS, Professor of Botany, University of Missouri. 2 18 MARINE BIOLOGICAL LABORATORY II. INSTRUCTION WILLIAM RANDOLPH TAYLOR, Professor of Botany, University of Penn- sylvania. HANNAH T. CROASDALE, Biological Abstracts, University of Pennsylvania. JAMES P. POOLE, Professor of Evolution, Dartmouth College. LIBRARY PRISCILLA B. MONTGOMERY (MRS. THOMAS H. MONTGOMERY, JR.), Li- brarian. DEBORAH LAWRENCE, Secretary. HESTER ANN BRADBURY, HAZEL BLANCHARD, MARY A. ROHAN, Assistants. CHEMICAL SUPPLIES OSCAR W. RICHARDS, Instructor in Biology. Yale University. SCIENTIFIC APPARATUS AND TECHNICAL SUPPLIES SAMUEL E. POND, Assistant Professor of Physiology, Schools of Medicine and Dentistry, University of Pennsylvania, in charge. A. R. APGAR, Photographer. LESTER F. Boss, Mechanician. J. D. GRAHAM, Glassblower. P. H. LILJESTRAND, Assistant. MUSEUM GEORGE M. GRAY, Curator. SUPPLY DEPARTMENT JAMES MC!NNIS, Manager. WALTER KAHLER, Collector. A. M. HILTON, Collector. GEOFFREY LEHY, Collector. MILTON B. GRAY, Collector. A. W. LEATHERS, Shipping. BOATS JOHN J. VEEDER, Captain. E. M. LEWIS, Chief Engineer. F. M. MACNAUGHT, Business Manager. HERBERT A. HILTON, Superintendent of Buildings and Grounds. THOMAS LARKIN, Superintendent of Mechanical Department. WILLIAM HEMENWAY, Carpenter. 2. INVESTIGATORS AND STUDENTS, 1931 Independent Investigators ADAMS, A. ELIZABETH, Professor of Zoology, Mount Holyoke College. ADDISON, WILLIAM H. F., Professor of Normal Histology and Embryology, Uni- versity of Pennsylvania. ALLEE, W. C., Professor of Zoology, University of Chicago. AMBERSON, WILLIAM R., Professor of Physiology, University of Tennessee. ANDERSON, RUBERT S., Research Associate, Princeton University. ARMSTRONG, PHILIP B., Assistant Professor of Anatomy, Cornell University Medi- cal College. REPORT OF THE DIRECTOR 19 ASTROM, I. ELISABETH, Class Assistant, University of Toronto. AUSTIN, MARY L., Instructor in Zoology, Wellesley College. BAILEY, PERCY L., JR., Instructor in Physiology, College of the City of New York. BAITSELL, GEORGE A., Professor of Biology, Yale University. BAKWIN, HARRY, Assistant Clinical Professor, New York University. BAKWIN, RUTH MORRIS, Instructor in Pediatrics, New York University. BALL, ERIC G., Instructor in Physical Chemistry, Johns Hopkins University Medi- cal School. BALLARD, WILLIAM W., Instructor in Zoology, Dartmouth College. BARD, PHILIP, Assistant Professor of Physiology, Harvard Medical School. BARRON, E. S. GUZMAN, Research Associate, University of Chicago. BARTH, L. G., National Research Council Fellow, University of Chicago. BEAMS, H. W., Assistant Professor of Zoology, State University of Iowa. BELKIN, MORRIS, Instructor in Biology, Washington Square College, New York University. BEUTNER, R., Professor of Pharmacology, University of Louisville, School of Med- icine. BISSONNETTE, THOMAS H., Professor of Biology, Trinity College. BODANSKY, OSCAR, Instructor in Pediatrics, New York University and Bellevue Hospital. BORODIN, D. N., 621 West 42d Street, New York City. BOWLING, RACHEL, Instructor in Zoology, Columbia University. BRADLEY, H. C., Professor of Physiological Chemistry, University of Wisconsin. BRIDGES, CALVIN B., Research Assistant in Genetics, Carnegie Institution of Wash- ington. BRINLEY, FLOYD J., Assistant Professor of Zoology, North Dakota State College. BRONFENBRENNER, J., Professor of Bacteriology and Public Health, Washington University Medical School. BUDINGTON, ROBERT A., Professor of Zoology, Oberlin College. CALKINS, GARY N., Professor of Protozoology, Columbia University. CANNAN, ROBERT K., Professor of Chemistry, New York University and Bellevue Hospital Medical College. CARPENTER, RUSSELL L., Instructor in Anatomy, College of Physicians and Sur- geons, Columbia University. CARVER, GAIL L., Professor of Biology, Alercer University. CASTLE, WILLIAM A., Instructor in Biology, Brown University. CATTELL, WARE, New York University. CHAMBERS, ROBERT, Research Professor and Chairman of Department of Biology, Washington Square College, New York University. CHEEVER, CLARENCE A., Member, Boston Society of Natural History. CHENEY, RALPH H., Chairman, Biology Department, Long Island University. CHIDESTER, F. E., Professor of Zoology, West Virginia University. CHRISTIE, JESSE R., Associate Nematologist, United States Department of Agri- culture. CLARK, ELEANOR LINTON, Research Assistant in Anatomy, University of Pennsyl- vania. CLARK, ELIOT R., Director of Department of Anatomy, University of Pennsylvania. CLOWES, G. H. A., Director, Lilly Research Laboratories. COBB, N. A., Principal Nematologist, United States Department of Agriculture. COE, WESLEY R., Professor of Biology, Yale University. COLE, ELBERT C., Associate Professor of Biology, Williams College. COLE, KENNETH S., Assistant Professor of Physiology, College of Physicians and Surgeons, Columbia University. CONKLIN, EDWIN G., Professor of Biology, Princeton University. COONFIELD, B. R., Instructor in Zoology, Brooklyn College of the City of New York. 20 MARINE BIOLOGICAL LABORATORY COPELAND, MANTON, Professor of Biology, Bowdoin College. COWDKY, E. V., Professor of Cytology, Washington University. COWLES, R. P., Professor of Zoology, Johns Hopkins University. CURTIS, W. C., Professor of Zoology, University of Missouri. CURWEN, ALICE O., Instructor in Histology and Embryology, Woman's Medical College of Pennsylvania. DANKS, W. B. C., Government Officer of Kenya Colony. DARRAH, WM. C., Fellow in Paleobotany, Carnegie Museum, University of Pitts- burgh. DAWSON, ALDEN B., Associate Professor of Zoology, Harvard University. DAWSON, J. A., Assistant Professor of Biology, College of the City of New York. DODDS, GIDEON S., Professor of Histology and Embryology, West Virginia Uni- versity. DOLLEY, WILLIAM L., JR., Professor of Biology, University of Buffalo. DONALDSON, HENRY H., Member, Wistar Institute. DuBois, EUGENE F., Professor of Medicine, Cornell University Medical College. DUNBAR, FRANCIS F., Graduate Assistant in Zoology, Columbia University. EDWARDS, DAYTON J., Associate Professor of Physiology, Cornell University Medi- cal College. EINARSON, LARUS, Research Fellow of the Rockefeller Foundation, Harvard Uni- versity Medical School. FAILLA, G., Physicist, Memorial Hospital, New York City. FAVILLI, GIOVANNI, First Assistant in the Institute of General Pathology, Royal University, Florence, Italy. FOGG, LLOYD C., Instructor in Biology, Washington Square College, New York University. FRASER, DORIS A., Assistant in Anatomy, University of Pennsylvania. FRENCH, CHARLES S., Graduate Student and Assistant in General Physiology, Harvard University. FRY, HENRY J., Professor of Biology, Washington Square College, New York University. FURTH, JACOB, Associate in Pathology, The Henry Phipps Institute. GARREY, W. E., Professor of Physiology, Vanderbilt University School of Medi- cine. GAYET, RENE, Directeur Adjoint Laboratoire de Physiologic Pathologique, College de France. GEIMAN, QUENTIN M., Director of Science Department, Swarthmore Preparatory School. GELFAN, SAMUEL, Assistant Professor of Physiology and Pharmacology, Univer- sity of Alberta. GERARD, R. W., Associate Professor of Physiology, University of Chicago. GILSON, LEWIS E., Instructor in Biochemistry, University of Cincinnati. GOLDFORB, A. J., Professor of Biology, College of the City of New York. GOODRICH, HUBERT B., Professor of Biology, Wesleyan University. GRAVE, B. H., Professor of Zoology, DePauw University. GRAVE, CASWELL, Professor of Zoology, Washington University. GREEN, ARDA A., Research Fellow in Physical Chemistry, Harvard University Medical School. GREENWOOD, ALAN W., Lecturer, Institute of Genetics, University of Edinburgh. GRUNDFEST, HARRY, National Research Council Fellow, Johnson Foundation, Uni- versity of Pennsylvania. HADLEY, CHARLES E., Assistant Professor of Biology, New Jersey State Teachers College. HAHNERT, WILLIAM F., National Research Fellow, The Johns Hopkins University. HAM, ARTHUR W., Instructor in Cytology, Washington University School of Med- icine. REPORT OF THE DIRECTOR 21 HAMBURGER, RUDOLF T., Assistant in the Medical Clinic, University of Groningen, Holland. HARNLY, MORRIS H., Assistant Professor, New York University. HARTLINE, H. K., Fellow, Johnson Foundation, University of Pennsylvania. HARVEY, ETHEL BROWNE, Assistant in Biology, Washington Square College, New York University. HARVEY, E. NEWTON, Professor of Physiology, Princeton University. HAYDEN, MARGARET A., Assistant Professor of Zoology, Wellesley College. HAYNES, FLORENCE W., Harvard University Medical School. HAYWOOD, CHARLOTTE, Associate Professor of Physiology, Mount Holyoke College. HEILBRUNN, L. V., Associate Professor of Zoology, University of Pennsylvania. HELWIG, EDWIN R., Instructor, University of Pennsylvania. HENDERSON, JEAN T., Lecturer, McGill University. HENSHAW, PAUL S., Biophysicist, Memorial Hospital, New York City. HILL, SAMUEL E., Assistant in Physiology, Rockefeller Institute. HOADLEY, LEIGH, Professor of Zoology, Harvard University. HODGE, CHARLES, JR., Instructor in Zoology, University of Pennsylvania. HODGE, RUTH M. PATRICK, University of Virginia, Charlottesville, Virginia. HOGUE, MARY JANE, Instructor in Anatomy, Medical School, University of Penn- sylvania. HOOK, SABRA J., Instructor in Biology, University of Rochester. HOPPE, ELLA N., Research Assistant, Division of Laboratories and Research, New York State Department of Health. HORNING, E. S., University of Sydney. HOWARD, EVELYN, University of Pennsylvania. HOWE, H. E., Editor, Industrial and Engineering Chemistry. HUETTNER, ALFRED F., Associate Professor, Washington Square College, New York University. IMAI, TAKEO, Assistant in Biology, Tohoku Imperial University, Sendai, Japan. IRVING, LAURENCE, Associate Professor of Physiology, University of Toronto. IRWIN, MARIAN, Associate, Rockefeller Institute. JACOBS, M. H., Professor of General Physiology, University of Pennsylvania. JOHLIN, J. M., Associate Professor of Biochemistry, Vanderbilt University, School of Medicine. JOHNSON, DUNCAN S., Professor of Botany, Johns Hopkins University. JOHNSON, H. HERBERT, Instructor, College of the City of New York. JOHNSTON, ROBERT L., Head of Research Division, Cleveland Clinic Foundation. KAUFMANN, BERWIND P., Professor of Botany, University of Alabama. KEEFE, REV. A. M., Rector, St. Norbert College. KEIL, ELSA M., Instructor in Zoology, New Jersey College for Women. KILLE, FRANK R., Associate Professor of Biology, Birmingham-Southern College. KINDRED, JAMES E., Associate Professor of Histology and Embryology, University of Virginia. KING, ROBERT L., Associate Professor, State University of Iowa. KIRBY-SMITH, HENRY T., Instructor in Anatomy, University of Pennsylvania. KNOWER, HENRY McE., Associate Professor of Anatomy, Albany Medical College. KNOWLTON, FRANK P., Professor of Physiology, Syracuse University, College of Medicine. KOSTIR, WENCEL J., Assistant Professor of Zoology, Ohio State University. LACKEY, JAMES B., Professor of Biology, Southwestern. LEVINE, PHILIP, Associate, Rockefeller Institute. LEWIS, IVEY F., Professor of Botany, University of Virginia. LILLIE, FRANK R., Chairman of the Department of Zoology, University of Chicago. LILLIE, RALPH S., Professor of General Physiology, University of Chicago. LOEBEL, ROBERT O., Fellow, Cornell University Medical College. LUCAS, ALFRED M., Assistant Professor of Cytology, Washington University Med- ical School. MARINE BIOLOGICAL LABORATORY LUCAS, MIRIAM SCOTT, Instructor in Cytology, Washington University Medical School. LUCRE, BALDUIN, Associate Professor of Pathology, University of Pennsylvania. LUND, E. J., Professor of Physiology, University of Texas. LYNCH, RUTH STOCKING, Instructor in Genetics, Johns Hopkins University. McCLUNG, C. E., Director, Zoological Laboratory, University of Pennsylvania. McGLONE, BARTGIS, Instructor in Physiology, University of Pennsylvania. McGouN, RALPH C., JR., Instructor in Biology, Amherst College. MCGREGOR, JAMES H., Professor of Zoology, Columbia University. AIARGOLIN, SYDNEY, Graduate Student, Columbia University. MATHEWS, ALBERT P., Carnegie Professor of Biochemistry, University of Cincin- nati. MAVOR, JAMES W., Professor of Biology, Union College. METZ, CHARLES W., Professor, Johns Hopkins University. MICHAELIS, L., Member, Rockefeller Institute for Medical Research. MILLER, HELEN M., National Research Council Fellow, Johns Hopkins University. MITCHELL, PHILIP H., Professor of Physiology, Brown University. MORGAN, ANN H., Professor of Zoology, Mount Holyoke College. MORGAN, LILIAN V., California Institute of Technology. MORGAN, T. H., Professor of Biology, California Institute of Technology. MORGULIS, SERGIUS, Professor of Biochemistry, University of Nebraska. MORRILL, CHARLES V., Associate Professor of Anatomy, Cornell University Med- ical College. MORRIS, HELEN S., Graduate Student, Columbia University. NABRIT, S. MILTON, Professor of Biology, Morehouse College. NAVEZ, ALBERT E., Lecturer in General Physiology, Harvard University. NELSEN, OLIN E., Instructor in Zoology, University of Pennsylvania. NICHOLAS, WARREN W., X-Ray Physicist, National Bureau of Standards. NONIDEZ, JOSE F., Assistant Professor of Anatomy, Cornell University Medical College. ORR, PAUL R., Instructor, University of Pennsylvania. OSTERHOUT, W. J. V., Member, Rockefeller Institute for Medical Research. PACKARD, CHARLES, Assistant Professor of Zoology, Columbia University. PAPENFUSS, GEORGE F., Student Assistant, Johns Hopkins University. PARKER, G. H., Professor of Zoology, Harvard University. PARMENTER, CHARLES L., Associate Professor of Zoology, University of Pennsyl- vania. PARPART, ARTHUR K., Instructor in Physiology, University of Pennsylvania. PATCH, ESTHER M., Windsor, Vermont. PAYNE, FERNANDUS, Head of Department of Zoology and Dean of Graduate School, Indiana University. PAYNE, NELLIE M., University of Pennsylvania. PEEBLES, FLORENCE, Professor of Biology, California Christian College. PINNEY, MARY E., Professor of Zoology, Milwaukee-Downer College. PLOUGH, HAROLD H., Professor of Biology, Amherst College. POLLISTER, ARTHUR W., Instructor in Zoology, Columbia University. POLLISTER, PRISCILLA FREW, Graduate Student, Columbia University. POND, SAMUEL E., Assistant Professor of Physiology, University of Pennsylvania. POOLE, JAMES P., Professor of Evolution, Dartmouth College. RAFFEL, DANIEL, Fellow, Johns Hopkins University. REDFIELD, HELEN, California Institute of Technology. REESE, ALBERT M., Head of Department of Zoology, West Virginia University. DE RENYI, GEORGE S., Associate Professor of Anatomy, University of Pennsylvania. REZNIKOFF, PAUL, Instructor in Aledicine, Cornell University Medical College. RICHARDS, ALFRED N., Professor of Pharmacology, University of Pennsylvania. RICHARDS, OSCAR W., Instructor in Biology, Yale University. REPORT OF THE DIRECTOR 23 RIJLANT, PIERRE, Professor of Human Physiology, University of Brussels. RISLEY, PAUL L., Instructor in Zoology, University of Michigan. ROOT, WALTER S., Assistant Professor of Physiology, College of Medicine, Syra- cuse University. RUGH, ROBERTS, Instructor in Zoology, Hunter College. SAYLES, LEONARD P., Instructor in Biology, College of the City of New York. SCHAUFFLER, WILLIAM G., Princeton, New Jersey. SCHMIDT, IDA T. GENTIIER, Research Fellow, Children's Hospital, Cincinnati, Ohio. SCHMIDT, LEON H., Research Fellow, University of Cincinnati. SCHRADER, FRANZ, Professor of Zoology, Columbia University. SCHRADER, SALLY HUGHES, Columbia University. SCHULTZ, JACK, Investigator, Carnegie Institute of Washington. SCOTT, SISTER FLORENCE MARIE, Assistant Professor, Seton Hill College. SELLMEYER, BERNARD L., Head of Department and Professor of Biology, Loyola University. SEVERINGHAUS, AURA E., Assistant Professor of Anatomy, Columbia University. SICKLES, GRACE, Assistant Bacteriologist, New York State Department of Health. SLIFER, ELEANOR H., National Research Council Fellow in Zoology, State Univer- sity of Iowa. SNOOK, THEODORE, Instructor in Histology and Embryology, Cornell University. SONNEBORN, TRACY M., Research Associate, Johns Hopkins University. SPEIDEL, CARL C., Associate Professor of Anatomy, University of Virginia. SPEMANN, HANS. Travelling Professor, Rockefeller Foundation. STEINBACH, HENRY BURR, Instructor in Zoology, University of Pennsylvania. STEWART, DOROTHY R., Assistant Professor of Biology, Skidmore College. STOCKARD, CHARLES R., Professor of Anatomy, Cornell University Medical College. STREET, SIBYL, 29 Jewett Place, Utica, New York. STRONG, OLIVER S., Professor of Neurology and Neuro-Histology, Columbia Uni- versity. SUMWALT, MARGARET, Assistant Professor of Physiology, Woman's Medical Col- lege of Pennsylvania. TAFT, CHARLES H., JR., Associate Professor of Pharmacology, University of Texas, Medical School. TASHIRO, SHIRO, Professor of Biochemistry, University of Cincinnati. TAYLOR, WM. RANDOLPH, Professor of Botany, University of Michigan. TITTLER, IRVING A., Assistant in Zoology, Columbia University. TITUS, CHARLES P., East Orange, New Jersey. TOHYAMA, Guzo, Assistant Professor, Tokyo Imperial University. TURNER, JOHN P., Instructor in Zoology, University of Minnesota. TYLER, ALBERT, Instructor in Embryology, California Institute of Technology. UNGER, W. BYERS, Assistant Professor of Zoology, Dartmouth College. VAN SLYKE, E., Instructor, University of Pittsburgh. VICARI, EMILIA M., Research Associate in Anatomy, Cornell University Medical College. WALKER, RUTH L, Instructor in Botany, University of Wisconsin. WARREN, HOWARD C., Professor of Psychology, Princeton University. WHEDON, ARTHUR D., Professor of Zoology and Head of Department of Zoology and Physiology, North Dakota Agricultural College. WTHITAKER, D. M., Assistant Professor of Zoology, Columbia University. WHITING, ANNA R., Professor of Biology, Head of Department, Pennsylvania Col- lege for Women. WHITING, P. W., Associate Professor of Zoology, University of Pittsburgh. WIEMAN, H. L., Professor of Zoology, University of Cincinnati. WILSON, EDMUND B., Professor Emeritus, Columbia University. WITSCHI, EMIL, Professor of Zoology, State University of Iowa 24 MARINE BIOLOGICAL LABORATORY WODEHOUSE, ROGER P., Director of the Biological Laboratory, The Arlington Chemical Co. WOLF, E. ALFRED, Assistant Professor of Zoology, University of Pittsburgh. WOODWARD, ALVALYN E., Assistant Professor, University of Michigan. YOUNG, ROGER ARLINER, Assistant Professor of Zoology, Howard University. Beginning Investigators ADAMS, EDGAR M., Laboratory Assistant, University of Cincinnati. APGAR, GRACE M., University of Pennsylvania. ATLAS, MEYER, Assistant in Embryology, Columbia University. AURINGER, JACK, Graduate Student, Detroit City College. BECK, LYLE V., Graduate Assistant in Physiology, New York University. BENKERT, JOSEPH M., Science Director, Ambridge High School. BENKERT, LYSBETH HAMILTON, Instructor, Pennsylvania College for Women. BOSTIAN, C. H., Assistant Professor of Zoology, North Carolina State College. BRADWAY, WINNEFRED, Assistant in Biology, Washington Square College, New York University. BUCHHEIT, J. ROBERT, Graduate Assistant in Zoology, University of Illinois. CABLE, RAYMOND M., Teaching Assistant, New York University. CARLSON, J. GORDON, Demonstrator in Biology, Bryn Mawr College. CHEN, T. T., Instructor and Graduate Student, University of Pennsylvania. CHOR, HERMAN, Research Fellow, Washington University Medical School. CLINE, ELSIE, Graduate Student, Johns Hopkins University. COHEN, BERNARD M., Student, Johns Hopkins University. CORSON, SAMUEL A., University of Pennsylvania. COSTELLO, DONALD P., Instructor in Zoology, University of Pennsylvania. CROASDALE, HANNAH T., Scientific Staff of Biological Abstracts, University of Pennsylvania. DAN, KATSUMA, University of Pennsylvania. DAUGHERTY, KATHRYN, Research Assistant, University of Pennsylvania. DAVIS, J. F., Instructor in Zoology, University of Pennsylvania. DEARING, WILLIAM H., Student, University of Pennsylvania Medical School. EASTLICK, HERBERT L., Part Time Assistant, Washington University. FIELD, MADELEINE E., Research Fellow in Physiology, Harvard School of Public Health. GILMORE, KATHRYN, Graduate Student, University of Pittsburgh. GRAY, NINA E., Assistant in Zoology, University of Wisconsin. GREEN, DAVID E., Assistant, Washington Square College, New York University. HAMBURGH, MORTON, Johns Hopkins University Medical School. HEISS, ELIZABETH, Instructor in Histology, Purdue University. HILL, EDGAR S., University of Cincinnati. HITSCHLER, WILLIAM J., Research Worker, University of Pennsylvania. ISHII, KENJIRO, Graduate Student, University of Cincinnati. JACKSON, JOHN R., Graduate Assistant in Botany, University of Missouri. JONES, EDGAR P., Instructor, University of Pittsburgh. KATZ, JACOB D., Graduate Student, New York University. KELTCH, ANNA K., Lilly Research Laboratories. KIDDER, GEORGE W., Graduate Student, Columbia University. KINNEY, ELIZABETH T., Assistant in Zoology, Barnard College. KRAJNIK, BOHUMIL, Research Assistant, University at Bratislava, Czechoslovakia. LAMBERT, ELIZABETH F., Technician, Harvard University Medical School. LUNDSTROM, HELEN M., Graduate Student, University of Pennsylvania. McQuESTEN, BARBARA, 14 Still Street, Brookline, Massachusetts. MACHELLA, THOMAS E., Assistant Instructor, University of Pennsylvania. MANN, D. R., Graduate Assistant, Duke University. MARSHAK, ALFRED, Student, Bussey Institution. REPORT OF THE DIRECTOR 25 MARSLAND, DOUGLAS A., Assistant Professor of Biology, Washington Square Uni- versity, New York University. MEDARIS, DON, Student, DePauw University. MELTZER, ADOLPH, Student, Cornell University Medical College. MENDELSON, EMANUEL S., Research Technician, University of Pennsylvania. MICHAELIS, EVA, 1185 Park Avenue, New York City. MORRIS, SAMUEL, Instructor in Zoology, University of Pennsylvania. NICOLL, PAUL A., Part Time Assistant, Washington University. OLTMANN, CLARA, Student, Columbia University. ORMSBY, ANDREW A., 14209 Winthrop Avenue, Detroit, Michigan. PARKS, MARK E., Assistant in Biology, New York University. PIERCE, MADELENE E., Graduate Student, Radcliffe College. POWERS, PHIL B., Instructor in Zoology, University of Pennsylvania. ROBINSON, ELLIS J., Graduate Assistant, New York University. ROOT, CLINTON W., Assistant in Biochemistry, Princeton University. SCHECHTER, VICTOR, College of the City of New York. SCHUETT, J. F., Assistant in Zoology, University of Chicago. SCHWEITZER, MORTON D., Assistant in Zoology, Columbia University. SHAPIRO, HERBERT, Assistant in Zoology, Columbia University. SHAW, C. RUTH, Faculty Member, Kent State College. SHORE, AGNES, Instructor in Chemistry, Bellevue Medical College, New York University. SICHEL, FERDINAND J., Assistant Instructor, Washington Square College, New York University. SOUTHWICK, WALTER E., Student, Harvard University. SPEICHER, B. R., Graduate Assistant, University of Pittsburgh. STANCATI, MILTON F., Graduate Assistant, University of Pittsburgh. STURDIVANT, H. P., Instructor, Columbia University. TORVIK, MAGNIIILD, Graduate Assistant, University of Pittsburgh. TOWNSEND, GRACE, Fellow, University of Chicago. TWYEFFORT, Louis H., Instructor in Biology, Princeton University. WALD, GEORGE, Research Assistant, Columbia University. WALKER, PAUL A., Student, Bowdoin College. WF.INER, ELEANOR, Graduate Student, University of Pennsylvania. WEISMAN, MAXWELL N., Fellow, College of the City of New York. WELTY, CARL, Assistant Professor of Biology, Parsons College. WILDE, MARY H., Graduate Assistant in Botany, New Jersey College for Women. WILLIAMS, MARY E., 37 Addington Road, Brookline, Massachusetts. WINSOR, AGNES A., Volunteer Associate in Biology, School of Hygiene, Johns Hopkins University. WINSOR, CHARLES P., Associate in Biology, School of Hygiene and Public Health, Johns Hopkins University. Research Assistants ALDERMAN, EVANGELINE, Graduate Assistant, Wellesley College. ARMSTRONG, ELEANOR F., Cornell University Medical College. ASHKENAZ, DAVID M., Assistant in Biology, New York University. BAILEY, SARAH W., Student, Radcliffe College. BOYD, MILFORD J., Assistant Graduate Student, University of Cincinnati. BUTT, CHARLES, Research Assistant, Princeton University. CARABELLI, A. ALBERT, Medical Student, University of Pennsylvania. COLDWATER, K. B., Instructor in Zoology, University of Missouri. COLE, ROBERT, 45 King Street, Oberlin, Ohio. DAVIS, JAMES E., Laskar Fellow and Research Assistant in Medicine, University of Chicago. 26 MARINE BIOLOGICAL LABORATORY DuBois, ANNE MARIE, Research Assistant, Carnegie Institution of Washington. ERLANGER, MARGARET, Harvard University Medical School. EYRE, SARAH W., Research Assistant, Long Island University. FOWLER, J. R., Research Assistant, University of Chicago. FRANCIS, DOROTHY S., Research Assistant in Biophysics, Memorial Hospital, New York City. FRIEDHEIM, ERNST A. H., Rockefeller Institute. GRAHAM, CLARENCE H., National Research Fellow, Johnson Foundation, Univer- sity of Pennsylvania. GRAUBARD, MARC A., National Research Council Fellow. HARNLY, MARIE L., Assistant in Biology, New York University. HARRYMAN, ILENE, Lilly Research Laboratories. HENSHAW, CHRISTINE T., Assistant Biophysicist, Memorial Hospital, New York City. HILSMAN, HELEN, Assistant, University of Pittsburgh. HOLT, HELEN, Assistant, Washington Square College, New York University. HOMES, MARCEL V. L., Assistant in Botany, University of Brussels. LARRABEE, MARTIN G., Student, Harvard College. PARKS, ELIZABETH K., Instructor in Histology and Embryology, Boston Univer- sity Medical School. PARPART, ETHEL R., 230 Nassau Street, Princeton, New Jersey. REMPE, ALOIS E., Washington University Medical School. SCHLUGER, JACK, Research Assistant, New York University. SCHMUCK, M. LOUISE, Research Assistant, Carnegie Institution. SCOTT, ALLAN C, Graduate Assistant in Zoology, University of Pittsburgh. SELL, JAMES P., Student, Oberlin College. SMITH, HELEN BERENICE, Research Assistant, Carnegie Institution of Washington. SMITH, M. DOREEN, Research Assistant, University of Toronto. SMITH, SUZANNE G., Research Assistant; University of Missouri. STOKES, JULIA C., Research Assistant, Washington University. TANG, PEI-SUNG, Research Fellow, Harvard University. TOCKER, ALBERT M., Research Assistant, Washington University Medical School. TYLER, BETTY S., California Institute of Technology. VAN ALSTYNE, MARGARET A., Assistant, Harvard University Medical School. WADE, LUCILLE W., Lilly Research Laboratories. WILSON, HILDEGARD N., Teaching Fellow, New York University and Bellevue Hospital Medical College. YOUNG, SAUL B., Technician, Rockefeller Institute. Students BOTANY ALTLAND, CLAIR S., American University. ANDREW, BARBARA R., Assistant Instructor in Botany, University of Alabama. BRYAN, HILAH F., Smith College. HUNT, WILLIAM, Student, Southwestern. JACKSON, JOHN R., Graduate Assistant in Botany, University of Missouri. MOORE, CAROLINE, Student, University of Pennsylvania. PERRY, LILY M., Graduate Student, Washington University. EMBRYOLOGY ALDERMAN, EVANGELINE, Graduate Assistant, Wellesley College. ALEXANDER, LLOYD E., Instructor, Fisk University. ALEXANDERSON, AMELIE M., Student, Bryn Mawr College. REPORT OF THE DIRECTOR 27 BOONE, ELEANOR S., Instructor in Zoology, Mills College. BUCHHEIT, J. ROBERT, Assistant in Zoology, University of Illinois. BUCK, MARGARET A., Student, University of Maine. CABLE, RAYMOND M., University Fellow, New York University. CARLSON, J. GORDON, Graduate Student, University of Pennsylvania. CHASE, HYMAN Y., Graduate Student, Howard University. CHEN, HSIN T., Student, Harvard University. COULTER, EDITH A., Student, Goucher College. DENNY, MARTHA, Radcliffe College. DERRICKSON, MARY B., Graduate Assistant, Syracuse University. DICK, GEORGE A., Professor of Animal Industry, University of Pennsylvania. KASTON, BENJAMIN J., Graduate Student, Yale University. MAGRUDER, SAMUEL R., Graduate Assistant in Zoology, University of Cincinnati. NEWCOMER, A. VIRGINIA, Radcliffe College. OPPENHEIMER, JANE M., Student, Bryn Mawr College. PLYLER, PHYLLIS V., Goucher College. PRICE, JOHN B., Graduate Student, Stanford University. SAWYER, ELIZABETH L., 29 Elm Street, Bangor, Maine. SELL, JAMES P., Oberlin College. SHEA, MARGARET M., Graduate Assistant, Wellesley College. SMITH, WILSON F., JR., Student, Cornell University Medical College. TOWNSEND, GRACE, Fellow, University of Chicago. WALKER, PAUL A., Bowdoin College. WEED, MILTON R.. Wesleyan University. WOODRUFF, BETH H., Graduate Assistant, Western Reserve University. WOODSIDE, GILBERT L., Assistant in Embryology Laboratory, DePauw University. PHYSIOLOGY BARNEY, RAYMOND L., Professor of Biology, Middlebury College. BECK, LYLE V., Graduate Assistant, New York University. FISHER, KENNETH C., Student, Acadia University. GAETJENS, LAURA C., Elmira College. GREEN, DAVID E., Teaching Fellow, New York University. HEISS, ELIZABETH M., Instructor in Biology, Purdue University. LUNDSTROM, HELEN M., Graduate Student, University of Pennsylvania. McQuESTEN, BARBARA, 14 Still Street, Brookline, Massachusetts. MICHAELIS, EVA, 1185 Park Avenue, New York City, New York. MOORE, ELINOR, Student, University of Pennsylvania. MORGAN, ISABEL M., Student, Stanford University. NOLL, CLARENCE L, Laboratory Assistant in Chemistry, Trinity College. PROSSER, C. LADD, Graduate s'tudent and Assistant, Johns Hopkins University^ REID, MARION A., Instructor in Physiology, Boston University School of Medicine. SCHERP, HENRY \V., Assistant, Rockefeller Institute. SWEETMAN, HARVEY L., Assistant Professor of Entomology, Massachusetts State College. WILLARD, WILLIAM R., Student, Yale School of Medicine. PROTOZOOLOGY ADELL, JAMES C., Assistant, Teachers College, Columbia University. AURINGER, JACK, College of City of Detroit. BROWN, REBECCA R., Columbia University. BURR, EDITH ROGERS, Columbia University. CARPENTER, HELENA JANE, Ohio Wesleyan University. DEE, M. BARBARA, Assistant, Jamaica Plain High School. 28 MARINE BIOLOGICAL LABORATORY ERICSON, ALMA L., Critic Teacher in Biology, Hunter College Model School. ESKRIDGE, LYDIA C, Technical Assistant, Johns Hopkins University. FENTON, FRANCES E., Teacher of General Science, Connecticut College. HENDERSON, LILLIAN O., Instructor in Biology, H. Sophie Newcomb College. HUTCHINGS, Lois M., Teacher of High School Biology, Newark, New Jersey. ICKES, MARGUERITE, Teacher of Biology, Lincoln High School, Cleveland, Ohio. JAMES, MIRIAM E., Science Teacher in High School, Gloucester, Massachusetts. ORMSBY, ANDREW A., College of City of Detroit. SPERRY, HELEN O., Teacher of Biology, Cathedral School of St. Alary. STEWART, PAUL A., University of Rochester. WATKINS, EVELYN G., Student, Vassar College. INVERTEBRATE ZOOLOGY AGUAGO, CARLOS S., Assistant Professor of Biology, University of Havana. ANTHONY, ELIZABETH S., Graduate Student, Brown University. BACHRACH, JOSEPHINE E., Student, Vassar College. BAKER, E. G. STANLEY, Student Assistant, DePauw University. BARRON, DONALD H., Graduate Student, Yale University. BELCHER, JANE C., Colby College. BREWSTER, JAMES R., Biological Production, University Film Foundation. CHASE, HYMAN Y., Graduate Student, Howard University. CHEN, HSIN T., Student, Harvard University. CHINN, MARY PRISCILLA, Goucher College. CLARK, ADELE F., Student, Tufts College. CLARK, JEAN M., Student, Wilson College. CLAUSEN, RALPH G., Instructor, Union College. COHEN, BERNARD M., Assistant in Zoology, Johns Hopkins University. COREY, H. IRENE, Research Assistant and Secretary, University of Pennsylvania. CROLY, JOHN T., Student, Dartmouth College. DIMICK, HELEN, Student, Wellesley College. DREW, R. W., Student, Wesleyan University. DRUGG, HELEN, Teacher, University of Vermont. EASTLICK, HERBERT L., Graduate Assistant in Zoology, Washington University. ELLIS, LOLA M., Assistant in Biology Department, Southwestern. FISH, HAROLD S., Student Assistant, Harvard University. FORHAN, LAURA Jo, Student, University of Montana. FUCHS, BARRETT, Student Assistant in Biology, American University. GERSTELL, RICHARD, Student, Dartmouth College. GLIDDEN, DOROTHY P., Smith College. GODWIN, MELVIN, Student, DePauw University. HEGNER, ISABEL, Student, Radcliffe College. HETRICK, LAWRENCE A., JR., American University. HOWARD, JOHN W., Student, Hamilton College. HUSSEY, KATHLEEN L., Fellow in Zoology, Ohio Wesleyan University. JEFFERSON, MARGARET D., Graduate Student, University of Pennsylvania. JOHNSON, ARLENE C., Student, Wheaton College. JONES, E. R., Lecturer in Biology, Dalhousie University. KILGORE, BYRON, JR., Butler University. KRAMER, THEODORE C., Research Assistant in Biology, Western Reserve Univer sity. LANGSTROTH, MURIEL A., Dalhousie University. MANN, DONALD R., Investigator, Duke University. METZNER, JEROME, Fellow, College of the City of New York. MOMENT, GAIRDNER B., Graduate Student, Yale University. REPORT OF THE DIRECTOR 29 NICOLL, PAUL A., Graduate Assistant, Washington University. POMERAT, CHARLES M., Assistant in Biology, Clark University. RAYE, WILLIAM H., JR.. Amherst College. ROSENBAUM, LOUISE, Laboratory Assistant, University of Pennsylvania. ROUNTREE, KATHERINE E., Instructor in Biology, Wesleyan College. SANDERS, ROSALTHA, Student, Yale University. SMITH, OSGOOD R., Hamilton College. SOLBERG, ARCHIE N., Instructor in Zoology, North Dakota Agricultural College. STEWART, PAUL A., University of Rochester. THOMAS, THURLO B., Graduate Assistant, Oberlin College. WARTERS, MARY, Assistant Professor of Zoology, Centenary College. WESTKAEMPER, SISTER REMBERTA, Head of Department of Biology, College of St. Benedict. WILLARD, WILLIAM R., Student, Yale School of Medicine. WISMER, VIRGINIA, Assistant, University of Pennsylvania. YOUNG, GEORGE D., Student, Acadia University. 3. TABULAR VIEW OF ATTENDANCE 1927 1928 1929 1930 1931 INVESTIGATORS— Total 294 323 329 337 362 Independent 209 217 234 217 236 Under Instruction 57 81 71 87 83 Research Assistants 28 25 24 33 43 STUDENTS— Total 141 133 125 136 125 Zoology 57 57 53 56 55 Protozoology 17 16 15 14 17 Embryology 32 29 28 27 29 Physiology 19 15 17 23 17 Botany 16 16 12 16 TOTAL ATTENDANCE 435 456 454 473 487 Less Persons registered as both Students and Investigators 1 10 14 434 454 444 459 467 INSTITUTIONS REPRESENTED — Total Ill 111 123 126 By Investigators 89 80 96 95 102 By Students .... 63 06 64 SCHOOLS AND ACADEMIES REPRESENTED By Investigators By Students 4 1 1 FOREIGN INSTITUTIONS REPRESENTED By Investigators 15 13 30 8 By Students 8 8 3 2 1 30 MARINE BIOLOGICAL LABORATORY 4. SUBSCRIBING AND COOPERATING INSTITUTIONS, 1931 Acadia University American University, Washington, D. C. Amherst College Barnard College Bowdoin College Brown University Bryn Mawr College Butler College C. R. B. Educational Foundation California Institute of Technology Carnegie Institution, Cold Spring Harbor Carnegie Institution of Washington College of St. Benedict Columbia University Cornell University Cornell University Medical College Dalhousie University Dartmouth College DePauw University Duke University Elmira College General Education Board Goucher College Hamilton College Harvard University Harvard University Medical School Howard University Hunter College Indiana University Industrial & Engineering Chemistry, of the American Chemical Society Johns Hopkins University Johns Hopkins University Medical School Eli Lilly & Co. Long Island University Loyola University Memorial Hospital of New York City Morehouse College Mount Holyoke College National Research Council New York State Department of Health New York University Oberlin College Pennsylvania College for Women Princeton University RadclifTe College Rockefeller Foundation Rockefeller Institute for Medical Research Rutgers University Seton Hill College Smith College Sophie Newcomb College Southwestern Syracuse University Tufts College Union College United States Department of Agri- culture University of Buffalo University of Chicago University of Chicago Medical School University of Cincinnati University of Illinois University of Iowa University of Minnesota University of Missouri University of Pennsylvania University of Pennsylvania Medical School University of Pittsburgh University of Rochester University of Texas University of Virginia University of Wisconsin Vanderbilt University Medical School Vassar College Washington University Washington University Medical School Wellesley College Wesleyan University Western Reserve University West Virginia University Wheaton College Wilson College Wistar Institute of Anatomy and Bi- ology Yale University REPORT OF THE DIRECTOR 31 SCHOLARSHIP TABLES Lucretia Crocker Scholarships for Teachers in Boston. Scholarship of $100 supported by a friend of the Laboratory since 1898. The Edwin S. Linton Memorial Endowment of Washington and Jefferson College, given by Edwin Linton and Margaret Brownson Linton in memory of their son, member of the class of 1913, who gave his life in France during the World War. The endowment amounts to $2,500, and the income therefrom is to be used to encourage study and research at the Marine Biological Laboratory, Woods Hole, Massachusetts. The Bio Club Scholarship of the College of the City of New York. 5. EVENING LECTURES, 1931 Tuesday, June 23 DR. LEONOR MICHAELIS " Theory of the Heme Pigments as Oxygen Carriers and as Oxidation Catalysts." Friday, July 3 DR. E. B. WILSON " The Central Bodies." Illustrated by Original Photomicrographs. Friday, July 10 DR. G. H. PARKER " Humoral Agents in Nervous Activi- ties with Special Reference to Chro- matophores." Friday, July 17 DR. ELIOT R. CLARK " The Microscopic Study of Cells and Tissues in the Living Mammal." Friday, July 24 DR. T. H. MORGAN " The Marine Laboratories of the World and Their Work." Friday, July 31 DR H SPEMANN " Experiments on the Amphibian Egg." Friday, August 7 THE REYNOLD A. SPAETH MEMO- RIAL LECTURE, delivered by DR. Ross G. HARRISON " Problems and Methods of Experi- mental Embryology." Friday, August 14 DR.' F. L. HISAW " The Corpus Luteum and Anterior Lobe Hormones and Their Physical Inter-relationships." Friday, August 21 DR. CHARLES R. STOCKARD " An Experimental Dog Farm for the Study of Form and Type." Wednesday, August 26 DR. E. D. CONGDON " Some Impressions, Racial and Cul- tural, of the Siamese." MARINE BIOLOGICAL LABORATORY Friday, August 28 DR. J. H. MCGREGOR " Motion Pictures Taken in the Bel- gian Congo and the Cameroon by the African Expedition (1929-30) of Columbia University and the American Museum of Natural His- tory." Friday, September 4 DR. BRADLEY M. PATTON " Micro-moving Pictures Applied to the Study of the Living Embryo.'' SPECIAL LECTURES AND MOTION PICTURES Saturday, July 18 MOTION PICTURES " Modern Studies of Sulphur." " Monel Metal." Thursday, July 23 DR. A. E. NAVEZ " Geotropism in Plants." MR. LEONARD CRASKE " The Art and Uses of Color Photog- raphy." Thursday, August 20 MOTION PICTURES " Cleaving Eggs of Echinarachnius and Arbacia." (Dr. Henry J. Fry) ' The Redistribution of Granules in Centrifuged Arbacia Eggs." (Dr. E. N. Harvey) " Arterio-venous Anastomoses." (Dr. E. R. Clark) " The Function of the Intestinal Villi." (Dr. F. Verzar — comments by Dr. W. E. Garrey) 6. SHORTER SCIENTIFIC PAPERS, 1931 Tuesday, June 30 DR. G. S. DODDS " Osteoclasts and Chondroclasts." DR. A. W. POLLISTER " The Architecture of the Liver Cells of Amphiuma." DR. G. H. PARKER " Passage of Sperms and Eggs through the Mammalian Oviduct." Tuesday, July 7 DR. A. C. REDFIELD " Effect of Hydrogen Ion Concentra- tion and Salt Concentration on the Oxygen Dissociation Constant of Hemocyanin.'' DR. LAURENCE IRVING '' The Carbon Dioxide Dissociation Curve of Living Mammalian Mus- cle." DR. E. N. HARVEY " Photo-electric Records of Animal Luminescence.1' Tuesday, July 14 DR. H. H. PLOUGH " Some Observations on Self Sterility in Stvela." REPORT OF THE DIRECTOR DR. R. CHAMBERS " Evidence of a Direct Action of the Nucleus on the Cytoplasm in Tissue Cultures." DR. A. F. HUETTNER •' Genetic Continuity of the Central Bodies." Tuesday, July 21 DR. S. MORGULIS " The Chemistry of Bone Ash." DR. J. M. JOHLIN " The Ejiolization of Gelatin by Neu- tral Salts." DR. E. S. GUZMAN BARRON " Oxidations Produced by Gonococci." DR. SHIRO TASHIRO AND MR. L. H. SCHMIDT " Bile Salts." Monday, July 27 DR. J. P. TURNER " The Fibrillar System in Euplotes." DR. DANIEL RAFFEL " Types of Variation Produced by Conjugation in Paramecium aure- lia." DR. RUTH S. LYNCH " Effects of Conjugation in a Number of Clones of Paramecium aurelia." DR. T. M. SONNEBORN " Crossing Diverse Clones of Parame- cium aurelia." Wednesday, July 29 UNDER THE AUSPICES OF THE SOCI- ETY OF CELLULAR BIOLOGY DR. BALDUIN LUCKE " The Mechanism of Bacteriotropin Action." DR. M. H. JACOBS AND DR. A. K. PARPART "Is the Permeability of the Erythro- cyte to Water Decreased by Nar- cotics? " DR. L. V. HEILBRUNN " The Action of the Common Cations on the Protoplasmic Viscosity of Amoeba." DR. R. CHAMBERS " The Formation of Ice Crystals in the Protoplasm of Various Cells." Tuesday, August 4 DR. E. F. DuBois " Surface Temperature and the Radia- tion of Heat from the Human Body." DR. PIERRE RIJLANT " Oscillographic Study of the Cardiac Ganglion of Limulus polyphemus." DR. D. M. WHITAKER '' The Change in Rate of Oxygen Con- sumption at Fertilization of the Eggs of Chaetopterus, Cumingia, Nereis, Arbacia and Fucus." DR. R. W. GERARD " Phosphocreatin in Nerve in Rela- tion to Activity." Tuesday, August 11 DR. W. H. F. ADDISON '' Aquatic Mammals — A Description of a Special Cell Type in the Cere- bellum." 3 34 MARINE BIOLOGICAL LABORATORY DR. C. C. SPEIDEL " Living Nerve Sprouts." DR. J. E. KINDRED " Histologic Effects of Ligation on the Vasa of the Spleen of the Albino Rat." DR. G. S. DERENYI " The Effect of Radium Irradiation upon the Ovaries of the Albino Rat." Tuesday, August 18 DR. HELEN B. SMITH " Genetic Studies on Selective Segre- gation of Chromosomes in Sciara." DR. P. W. WHITING " Local and Correlative Gene Effects in Mosaics of Habrobracon." DR. C. B. BRIDGES " Specific Modifiers in Drosophila melanogaster." Tuesday, August 25 DR. PAUL S. HENSHAW " Recovery from X-ray Effects as Ob- served in Arbacia Eggs." MR. WARE CATTELL '' The Reaction of the Fundulus Ovum to the Direct Electric Current." DR. E. A. WOLF AND DR. H. H. COLLINS " The Effect of Ultra-violet Radiation upon the Color Pattern of Triturus." DR. G. H. PARKER " The Discharge of Nematocysts." DR. DMITRY N. BORODIN " Biological Spectrum and M-rays." (Motion pictures) Tuesday, September 1 DR. ALBERT TYLER "Artificial Parthenogenesis in the Eggs of the Pacific Coast Echiu- roid, Urechis caupo." DR. PAUL S. GALTSOFF " Specificity of Sexual Reactions in the Genus Ostrea." DR. K. B. COLDWATER " The Effect of Sulphydryl Com- pounds upon Regenerative Growth." DR. N. A. COBB " The Use of Live Nemas in Zoologi- cal Courses in Schools and Col- leges." Thursday, September 3 DR. CHARLOTTE HAYWOOD AND DR. WALTER S. ROOT " The Cleavage Rate of the Arbacia Egg in the Presence of Carbon Di- oxide and Bicarbonate." DR. ARTHUR K. PARPART AND DR. M. H. JACOBS " The Action of Acetic Acid and Its Sodium Salt on the Cleavage of Ar- bacia Eggs." MR. K. DAN " Cataphoretic Studies of Marine Eggs." DR. KENNETH COLE " Surface Forces of the Arbacia Egg.'' DR. E. N. HARVEY '' The Tension at the Surface of Ar- bacia Eggs, Determined by Centrif- ugal Force." REPORT OF THE DIRECTOR DR. ETHEL BROWNE HARVEY " Development of Arbacia Half-Eggs Produced by Centrifugal Force." DR. BALDUIN LUCRE "Osmotic Properties of 'Fragments' of Arbacia Eggs Obtained by Cen- trifugal Force." DR. M. H. JACOBS AND DR. DOROTHY R. STEWART "A Method for the Quantitative Measurement of Cell Permeability/' DR. DOROTHY R. STEWART AND DR. M. H. JACOBS " The Effect of Fertilization on the Permeability of the Arbacia Egg to Ethylene Glycol." DR. MIRIAM SCOTT LUCAS "' Recent Observations upon a Type of Fission Undescribed for Ciliates." DR. E. C. COLE '' Selective Intra-Vitam Staining of Specific Elements in the Integument of the Squid.'' DR. C. C. SPEIDEL " Types of Nerve Regeneration, as Revealed by Prolonged Observation of Individual Fibers in Living Frog Tadpoles." DR. H. H. JOHNSON " Centrioles and Other Cytoplasmic Bodies in Living Cells of Gryllids." DR. P. W. WHITING " Genetic Results in Habrobracon Bearing on Maturation and Ferti- lization." MR. L. V. BECK AND MR. D. E. GREEN '' Oxidation-reduction Potentials of Cytolyzed and Intact Echinoderm Eggs.'" DR. ERIC G. BALL ," Hemolysis of Fish Erythrocytes by an Impurity in Sodium Chloride." DR. G. H. A. CLOWES, DR. I. H. PAGE AND MR. H. A. SHONLE ''On the Contrasting Cytolytic Ef- fects E-xerted by Soaps of the Type of Sodium Ricinoleate and Sodium Oleate at Different H Ion Concen- trations and the Relation of These Effects to the Oil-water Interfacial Tensions Exerted by the Soaps in Question." Miss ANNA KELTCH, Miss ILENE HARRYMAN AND DR. G. H. A. CLOWES " Influence of H Ion Concentration on the Anesthetic Value of a Series of General and Local Anesthetics and Hypnotics." MR. S. A. CORSON '' The Action of Acid and Alkali on the Protoplasmic Viscosity of Amoeba dubia." 36 MARINE BIOLOGICAL LABORATORY MR. H. B. STEINBACH " The Effect of Salts on the Injury Current of Scallop Muscle." DR. ROBERT CHAMBERS AND MR. D. A. MARSLAND " The Action of the Common Salts on the Protoplasm of the Echinoderm KO-O- " ^ &*»• MR. MORRIS BELKIN " Capping of Oils on Protoplasmic Surfaces." DEMONSTRATIONS MR. DAVID M. ASHKENAZ " The Effect of Sodium and Calcium Chlorides on Changes in Penetrabil- ity of Neutral Red." DR. ERIC G. BALL " Hemolysis of Fish Erythrocytes by an Impurity in Sodium Chloride." MR. MORRIS BELKIN " The Capping Phenomenon in Amoeba dubia." DR. C. B. BRIDGES " Apparatus and Designs for Raising Drosophila." DR. E. R. CLARK, MRS. E, L. CLARK, DR. H. T. KIRBY-SMITH AND DR. W. J. HITSCHLER " Living Tissues as Seen in Trans- parent Chambers Introduced into the Rabbit's Ear." DR. E. C. COLE " Selective Intra-Vitam Staining of Specific Elements in the Integument of the Squid." DR. KENNETH COLE •' An Egg Crusher." DR. E. N. HARVEY AND DR. E. B. HARVEY " Arbacia Half -cells (Fertilized and Unfertilized) Produced by Centrif- ugal Force." DR. H. H. JOHNSON '' Centrioles and Other Cytoplasmic Bodies in Living Cells of Gryllids." DR. MIRIAM SCOTT LUCAS " Demonstration of Fission of Cyatho- didinium piriforme." DR. C. W. METZ '' Demonstration of Chromosomes of Sciara." DR. A. E. NAVEZ " Cardiac Frequency of Anomya as a Function of Temperature." DR. NELLIE M. PAYNE " The Effect of Temperature upon the Duration of ' Death Feigning.' ' MR. F. J. M. SICHEL " Apparatus for Studying Tension in Isolated Muscle Cells." DR. C. C. SPEIDEL '' Nerve Sprouts, Sheath Cells and Myelin Segments in Living Frog Tadpoles." DR. ANNA R. WHITING. Miss MAGNHILD M. TORVIK, MRS. LYS- BETH H. BENKERT AND Miss KATH- RYN A. GILMORE " Exhibit of Mutants and Mosaics in Habrobracon." REPORT OF THE DIRECTOR 37 7. MEMBERS OF THE CORPORATION 1. LIFE MEMBERS ALLIS, MR. E. P., JR., Palais Carnoles, Menton, France. ANDREWS, MRS. GWENDOLEN FOULKE, Baltimore, Md. BILLINGS, MR. R. C, 66 Franklin St., Boston, Mass. CONKLIN, PROF. EDWIN G., Princeton University, Princeton, N. J. COOLIDGE, MR. C. A., Ames Building, Boston, Mass. CRANE, MR. C. R., New York City. EVANS, MRS. GLENDOWER, 12 Otis Place, Boston, Mass. FAY, Miss S. B., 88 Mt. Vernon St., Boston, Mass. FOOT, Miss KATHERINE, Care of Morgan Harjes Cie, Paris, France. GARDINER, MRS. E. G., Woods Hole, Mass. JACKSON, Miss M. C., 88 Marlboro St., Boston, Mass. JACKSON, MR. CHAS. C., 24 Congress St., Boston, Mass. KIDDER, MR. NATHANIEL T., Milton, Mass. KING, MR. CHAS. A. LEE, MRS. FREDERIC S., 279 Madison Ave., New York City, N. Y. LOWELL, MR. A. LAWRENCE, 17 Quincy St., Cambridge, Mass. McMuRRicii, PROF. J. P., University of Toronto, Toronto, Canada. MEANS, DR. JAMES HOWARD, 15 Chestnut St., Boston, Mass. MERRIMAN, MRS. DANIEL, 73 Bay State Road, Boston, Mass. MINNS, Miss SUSAN, 14 Louisburg Square, Boston, Mass. MORGAN, MR. J. PIERPONT, JR., Wall and Broad Sts., New York City, N. Y. MORGAN, PROF. T. H., Director of Biological Laboratory, California Institute of Technology, Pasadena, Calif. MORGAN, MRS. T. H., Pasadena, Calif. NOYES, Miss EVA J. OSBORN, PROF. HENRY F., American Museum of Natural History, New York, N. Y. PHILLIPS, MRS. JOHN C., Windy Knob, Wenham, Mass. PORTER, DR. H. C., University of Pennsylvania, Philadelphia, Pa. SEARS, DR. HENRY F., 86 Beacon St., Boston, Mass. SHEDD, MR. E. A. THORNDIKE, DR. EDWARD L., Teachers College, Columbia University, New York City, N. Y. TRELEASE, PROF. WILLIAM, University of Illinois, Urbana, 111. WARE, Miss MARY L., 41 Brimmer St., Boston, Mass. WILLIAMS, MRS. ANNA P., 505 Beacon St., Boston, Mass. WILSON, DR. E. B., Columbia University, New York City, N. Y. 38 MARINE BIOLOGICAL LABORATORY 2. REGULAR MEMBERS, 1931 ADAMS, DR. A. ELIZABETH, Mount Holyoke College, South Hadley, Mass. ADDISON, DR. W. H. F., University of Pennsylvania Medical School, Philadelphia, Pa. ADOLPH, DR. EDWARD F., University of Rochester, School of Medicine and Dentistry, Rochester, N. Y. ALLEE, DR. W. C., University of Chicago, Chicago, 111. ALLEN, PROF. CHARLES E., University of Wisconsin, Madison, Wis. ALLEN, PROF. EZRA, New York Homeopathic Medical College, New York City, N. Y. ALLYN, DR. HARRIET M., Mount Holyoke College, South Hadley, Mass. AMBERSON, DR. WILLIAM R., University of Tennessee, Memphis, Tenn. ANDERSON, DR. E. G., California Institute of Technology, Pasadena, Calif. AUSTIN, DR. MARY L., Wellesley College, Wellesley, Mass. BAITSELL, DR. GEORGE A., Yale University, New Haven, Conn. BAKER, DR. E. H., 5312 Hyde Park Boulevard, Hyde Park Station, Chicago, 111. BALDWIN, DR. F. M., University of Southern California, Los Angeles, Calif. BARD, PROF. PHILIP, Harvard University Medical School, Boston, Mass. BECKWITH, DR. CORA J., Yassar College, Poughkeepsie, N. Y. BEHRE, DR. ELINOR H., Louisiana State University, Baton Rouge, La. BENNITT, DR. RUDOLF, University of Missouri, Columbia, Mo. BIGELOW, DR. H. B., Museum of Comparative Zoology, Cambridge, Mass. BIGELOW, PROF. R. P., Massachusetts Institute of Technology, Cam- bridge, Mass. BINFORD, PROF. RAYMOND, Guilford College, Guilford College, N. C. BISSONNETTE, DR. T. H., Trinity College, Hartford, Conn. BLANCHARD, PROF. KENNETH C., New York University, Washington Square College, New York City, N. Y. BODINE, DR. J. H., University of Iowa, Iowa City, la. BORING, DR. ALICE M., Yenching University, Peking, China. BOWLING, Miss RACHEL, Columbia University, New York City, N. Y. Box, Miss CORA M., University of Cincinnati, Cincinnati, O. BRADLEY, PROF. HAROLD C., University of Wisconsin, Madison, Wis. BRAILEY, Miss MIRIAM E., 800 Broadway, Baltimore, Md. BRIDGES, DR. CALVIN B., California Institute of Technology, Pasadena, Calif. REPORT OF THE DIRECTOR 39 BRONK, DR. D. W., University of Pennsylvania, Philadelphia, Pa. BROOKS, DR. S. C., University of California, Berkeley, Calif. BUCKINGHAM, Miss EDITH N., Sudbury, Mass. BUDINGTON, PROF. R. A., Oberlin College, Oberlin, O. BULLINGTON, DR. W. E., Randolph-Macon College, Ashland, Va. BUMPUS, PROF. H. C., 76 Carlton Road, Waban, Mass. BYRNES, DR. ESTHER F., 1803 North Camac Street, Philadelphia, Pa. CALKINS, PROF. GARY N., Columbia University, New York City, N. Y. CALVERT, PROF. PHILIP P., University of Pennsylvania, Philadelphia, Pa. CANNAN, PROF. R. K., University and Bellevue Hospital Medical Col- lege, New York City, N. Y. CARLSON, PROF. A. J., University of Chicago, Chicago, 111. CAROTHERS, DR. E. ELEANOR, University of Pennsylvania, Philadelphia, Pa. CARROLL, PROF. MITCHEL, Franklin and Marshall College, Lancaster, Pa. CARVER, PROF. GAIL L., Mercer University, Macon, Ga. CATTELL, DR. McKEEN, Cornell University Medical College, New York City, N. Y. CATTELL, PROF. J. MC!VEEN, Garrison-on-Hudson, N. Y. CATTELL, MR. WARE, Garrison-on-Hudson, N. Y. CHAMBERS, DR. ROBERT, Washington Square College, New York Uni- versity, Washington Square, New York City, N. Y. CHARLTON, DR. HARRY H., University of Missouri, Columbia, Mo. CHATTON, DR. EDOUARD, University of Strasbourg, Strasbourg, France. CHIDESTER, PROF. F. E., West Virginia University, Morgantown, W. Va. CHILD, PROF. C. M., University of Chicago, Chicago, 111. CLAPP, PROF. CORNELIA M., Montague, Mass. CLARK, PROF. E. R., University of Pennsylvania, Philadelphia, Pa. CLELAND, PROF. RALPH E., Goucher College, Baltimore, Md. CLOWES, PROF. G. H. A., Eli Lilly & Co.. Indianapolis, Ind. COE, PROF. W. R., Yale University, New Haven, Conn. COHN, DR. EDWIN J., 183 Brattle Street, Cambridge, Mass. COLE, DR. ELBERT C., Williams College, Williamstown, Mass. COLE, DR. LEON J., College of Agriculture, Madison, Wis. COLLETT, DR. MARY E., Western Reserve University, Cleveland, O. COLLEY, MRS. MARY W., 36 Argyle Place, Rockville Centre, Long Is- land, N. Y. COLTON, PROF. H. S., Box 127, Flagstaff, Ariz. CONNOLLY, DR. C. J., Catholic University, Washington, D. C. 40 MARINE BIOLOGICAL LABORATORY COPELAND. PROF. MANTON, Bowdoin College, Brunswick, Me. COWDRY, DR. E. V., Washington University, St. Louis, Mo. CRAMPTON, PROF. H. E., Barnard College, Columbia University, New York City, N. Y. CRANE, MRS. C. R., Woods Hole, Mass. CURTIS, DR. MAYNIE R., Crocker Laboratory, Columbia University, New York City, N. Y. CURTIS, PROF. W. C., University of Missouri, Columbia, Mo. DAVIS, DR. ALICE R., Castle Point, Hoboken, N. J. DAVIS, DR. DONALD W., College of William and Mary, Williamsburg, Va. DAWSON, DR. A. B., Harvard University, Cambridge, Mass. DAWSON, DR. J. A., College of the City of New York, New York City, N. Y. DEDERER, DR. PAULINE H., Connecticut College, New London. Conn. DELLINGER, DR. S. C., University of Arkansas, Fayetteville, Ark. DODDS, PROF. G. S., Medical School, University of West Virginia, Mor- gantown, W. Va. DOLLEY, PROF. WILLIAM L., University of Buffalo, Buffalo, N. Y. DONALDSON, PROF. H. H., Wistar Institute of Anatomy and Biology, Philadelphia, Pa. DONALDSON, DR. JOHN C., University of Pittsburgh, School of Medi- cine, Pittsburgh, Pa. DREW, PROF. OILMAN A., Eagle Lake, Fla. DuBois, DR. EUGENE F., Cornell University Medical College, New York City, N. Y. DUGGAR, DR. BENJAMIN M., University of Wisconsin, Madison, Wis. DUNGAY, DR. NEIL S., Carleton College, Northfield, Minn. DUNN, DR. L. C., Columbia University, New York City, N. Y. EDWARDS, DR. D. J., Cornell University Medical College, New York City, N. Y. ELLIS, DR. F. W., Monson, Mass. FARNUM, DR. LOUISE W., Hsiang-Ya Hospital, Changsha, Hunan, China. FAURE-FREMIET, PROF. EMMANUEL, College de France, Paris, France. FENN, DR. W. O., Rochester University, School of Medicine, Rochester, N. Y. FIELD, Miss HAZEL E., Occidental College, Los Angeles, Calif. FORBES, DR. ALEXANDER, Harvard University Medical School, Boston, Mass. FRY, DR. HENRY J., Washington Square College, New York City, N. Y. GAGE, PROF. S. H., Cornell University, Ithaca, N. Y. REPORT OF THE DIRECTOR 41 CARREY, PROF. W. E., Vanderbilt University Medical School, Nashville, Tenn. GATES, DR. F. L., 31 Fayerweather Street, Cambridge, Mass. GATES, PROF. R. RUGGLES, University of London, London, England. GEISER, DR. S. W., Southern Methodist University, Dallas, Tex. GERARD, PROF. R. W., University of Chicago, Chicago, 111. GLASER, PROF. O. C, Amherst College, Amherst, Mass. GOLDFORB, PROF. A. J., College of the City of New York, New York City, N. Y. GOODRICH, PROF. H. B., Wesleyan University, Middletown, Conn. GRAHAM, DR. J. Y., LTniversity of Alabama, University, Ala. GRAVE, PROF. B. H., DePauw University, Greencastle, Ind. GRAVE, PROF. CASWELL, Washington University, St. Louis, Mo. GRAY, PROF. IRVING E., Duke University, Durham, N. C. GREENMAN, PROF. M. J., Wistar Institute of Anatomy and Biology, Philadelphia, Pa. GREGORY, DR. LOUISE H., Barnard College, Columbia University, New York City, N. Y. GUTHRIE, DR. MARY J., University of Missouri, Columbia, Mo. GUYER, PROF. M. F., University of Wisconsin, Madison, Wis. HAGUE, DR. FLORENCE, Sweet Briar College, Sweet Briar, Va. HALL, PROF. FRANK G., Duke University, Durham, N. C. HANCE, DR. ROBERT T., University of Pittsburgh, Pittsburgh, Pa. HARGITT, PROF. GEORGE T., Duke University, Durham, N. C. HARMAN, DR. MARY T., Kansas State Agricultural College, Manhattan, Kans. HARPER, PROF. R. A., Columbia University, New York City, N. Y. HARRISON, PROF. Ross G., Yale University, New Haven, Conn. HARVEY, MRS. E. N., Princeton, N. J. HARVEY, PROF. E. N., Princeton University, Princeton, N. J. HAYDEN, DR. MARGARET A., Wellesley College, Wellesley, Mass. HAYWOOD, DR. CHARLOTTE, Mount Holyoke College, South Hadley, Mass. HAZEN, DR. T. E., Barnard College, Columbia University, New York City, N. Y. HEATH, PROF. HAROLD, Pacific Grove, Calif. HECHT, DR. SELIG, Columbia University, New York City, N. Y. HEGNER, PROF. R. W., Johns Hopkins University, Baltimore, Md. HEILBRUNN, DR. L. V., University of Pennsylvania, Philadelphia, Pa. HESS, PROF. WALTER N., Hamilton College, Clinton, N. Y. HINRICHS, DR. MARIE A., 1824 Blue Island Avenue, Chicago, 111. HISAW, DR. F. L., University of Wisconsin, Madison, Wis. 42 MARINE BIOLOGICAL LABORATORY HOADLEY, DR. LEIGH, Harvard University, Cambridge, Mass. HOGUE, DR. MARY J., 503 N. High Street, West Chester, Pa. HOLMES, PROF. S. J., University of California, Berkeley, Calif. HOOKER, PROF. DAVENPORT, University of Pittsburgh, Pittsburgh, Pa. HOPKINS, DR. HOYT S., New York University, College of Dentistry, New York City, N. Y. HOWARD, DR. HARVEY J., Washington University, St. Louis, Mo. HOWE, DR. H. E., 2702 36th Street, N.W., Washington, D. C. HOYT, DR. WILLIAM D., Washington and Lee University, Lexington, Va. HUMPHREY, MR. R. R., University of Buffalo, School of Medicine, Buffalo, N. Y. HYMAN, DR. LIBBIE H., University of Chicago, Chicago, 111. INMAN, PROF. ONDESS L., Antioch College, Yellow Springs, O. IRVING, PROF. LAURENCE, University of Toronto, Toronto, Canada. IRWIN, DR. MARIAN, Rockefeller Institute, New York City, N. Y. JACKSON, PROF. C. M., University of Minnesota, Minneapolis, Minn. JACOBS, PROF. MERKEL H., University of Pennsylvania, Philadelphia, Pa. JENNINGS, PROF. H. S.. Johns Hopkins University, Baltimore, Md. JEWETT, PROF. J. R., Harvard University, Cambridge. Mass. JOHNSON, PROF. GEORGE E., State Agricultural College, Manhattan, Kans. JONES, PROF. LYNDS, Oberlin College, Oberlin, O. JORDAN, PROF. E. O., University of Chicago, Chicago, 111. JUST, PROF. E. E., Howard University, Washington, D. C. KAUFMANN, PROF. B. P., University of Alabama, University, Ala. KEEFE, REV. ANSELM M., St. Norbert College, West Depere, Wis. KENNEDY, DR. HARRIS, Readville, Mass. KINDRED, DR. J. E., University of Virginia, Charlottesville, Va. KING, DR. HELEN D., Wistar Institute of Anatomy and Biology, Phila- delphia, Pa. KING, DR. ROBERT L., State University of Iowa, Iowa City, la. KINGSBURY, PROF. B. F., Cornell University, Ithaca, N. Y. KNAPKE, REV. BEDE, St. Bernard's College, St. Bernard, Ala. KNOWER, PROF. H. McE., Albany Medical College, Albany, N. Y. KNOWLTON, PROF. F. P., Syracuse University, Syracuse, N. Y. KOSTIR, DR. W. J., Ohio State University, Columbus, O. KRIBS, DR. HERBERT, 202A Copley Road, Upper Darby, Pa. KUYK, DR. MARGARET P., Westbrook Ave., Richmond, Va. LANCEFIELD, DR. D. E., Columbia University, New York City, N. Y. LANGE, DR. MATHILDE M., Wheaton College, Norton, Mass. REPORT OF THE DIRECTOR 43 LEE, PROF. F. S., College of Physicians and Surgeons, New York City, N. Y. LEWIS, PROF. I. F., University of Virginia, Charlottesville, Va. LEWIS, PROF. W. H., Johns Hopkins University, Baltimore, Md. LILLIE, PROF. FRANK R., University of Chicago, Chicago, 111. LILLIE, PROF. RALPH S., University of Chicago, Chicago, 111. LIXTON, PROF. EDWIN, University of Pennsylvania, Philadelphia, Pa. LOEB, PROF. LEO, Washington University Medical School, St. Louis, Mo. LOEB, MRS. LEO, 812 Boland Place, St. Louis, Mo. LOWTHER, MRS. FLORENCE DEL., Barnard College, Columbia Univer- sity, New York City, N. Y. LUCKE, PROF. BALDUIN, University of Pennsylvania, Philadelphia, Pa. LUND, DR. E. J., University of Texas, Austin, Tex. LUSCOMBE, MR. \Y. O., Woods Hole, Mass. LYNCH, DR. CLARA J., Rockefeller Institute, New York City, N. Y. LYNCH, DR. RUTH STOCKING, Johns Flopkins University, Baltimore, Md. LYON, PROF. E. P., University of Minnesota, Minneapolis, Minn. MACDOUGALL, DR. MARY S., Agnes Scott College, Decatur, Ga. McCLUNG, PROF. C. E., University of Pennsylvania, Philadelphia, Pa. McGEE, DR. ANITA NEWCOMB, Box 363, Southern Pines, N. C. MCGREGOR, DR. J. H., Columbia University, New York City, N. Y. McNAiR, DR. G. T., 1624 Alabama Street, Lawrence, Kans. MACKLIN, DR. CHARLES C., School of Medicine, University of Western Ontario, London, Canada. MALONE, PROF. E. F., University of Cincinnati, Cincinnati, O. MANWELL, DR. REGINALD D., Syracuse University, Syracuse, N. Y. MARTIN, PROF. E. A., College of the City of New York, New York City, N. Y. MAST, PROF. S. O., Johns Hopkins University, Baltimore, Md. MATHEWS, PROF. A. P., University of Cincinnati, Cincinnati, O. MAYOR, PROF. JAMES W., Union College, Schenectady, N. Y. MEDES, DR. GRACE, University of Minnesota, Minneapolis, Minn. MEIGS, DR. E. B., Dairy Division Experiment Station, Beltsville, Md. MEIGS, MRS. E. B., 1736 M Street, N.W., Washington, D. C. METCALF, PROF. M. M., Johns Hopkins University, Baltimore, Md. METZ, PROF. CHARLES W., Johns Hopkins University, Baltimore, Md. MICHAELIS, DR. LEONOR, Rockefeller Institute, New York City, N. Y. MILLER, DR. HELEN M., Yale University, New Haven, Conn. MINER, DR. ROY W., American Museum of Natural History, New York City, N. Y. 44 MARINE BIOLOGICAL LABORATORY MITCHELL, DR. PHILIP H., Brown University, Providence, R. I. MOORE, DR. CARL R., University of Chicago, Chicago, 111. MOORE, PROF. GEORGE T., Missouri Botanical Garden, St. Louis, Mo. MOORE, PROF. J. PERCY, University of Pennsylvania, Philadelphia, Pa. MORGULIS, DR. SERGIUS, University of Nebraska, Lincoln, Nebr. MORRILL, PROF. A. D., Hamilton College, Clinton, N. Y. MORRILL, PROF. C. V., Cornell University Medical College, New York City, N. Y. MULLER, DR. H. J., University of Texas, Austin, Tex. NABOURS, DR. R. K., Kansas State Agricultural College, Manhattan, Kans. NEAL, PROF. H. V., Tufts College, Tufts College, Mass. NEWMAN, PROF. H. H., University of Chicago, Chicago, 111. NICHOLS, DR. M. LOUISE, Dreycott Apartments, Haverford, Pa. NOBLE, DR. GLADWYN K., American Museum of Natural History, New York City, N. Y. NONIDEZ, DR. JOSE F., Cornell University Medical College, New York City, N. Y.' OKKELBERG, DR. PETER, University of Michigan, Ann Arbor, Mich. OSBURN, PROF. R. C., Ohio State University, Columbus, O. OSTERHOUT, PROF. W. T- V., Rockefeller Institute, New York City, N. Y. PACKARD, DR. CHARLES, Columbia University, Institute of Cancer Re- search, 1145 Amsterdam Ave., New York City, N. Y. PAGE, DR. IRVINE H., Rockefeller Institute, New York City, N. Y. PAPANICOLAOU, DR. GEORGE N., Cornell University Medical College, New York City, N. Y. PAPPENHEIMER, DR. A. M., Columbia University, New York City, N. Y. PARKER, PROF. G. H., Harvard University, Cambridge, Mass. PATON, PROF. STEWART, Princeton University, Princeton, N. J. PATTEN, DR. BRADLEY M., Western Reserve University, Cleveland, O. PATTEN, PROF. WILLIAM, Dartmouth College, Hanover, N. H. PAYNE, PROF. F., University of Indiana, Bloomington, Ind. PEARL, PROF. RAYMOND, Institute for Biological Research, 1901 East Madison Street, Baltimore, Md. PEARSE, PROF. A. S., Duke University, Durham, N. C. PEEBLES, PROF. FLORENCE, California Christian College, Los Angeles, Calif. PHILLIPS, DR. E. F., Cornell University, Ithaca, N. Y. PHILLIPS, DR. RUTH L., Western College, Oxford, O. PIKE, PROF. FRANK H., 437 West 59th Street, New York City, N. Y. REPORT OF THE DIRECTOR 45 PINNEY, DR. MARY E., Milwaukee-Downer College, Milwaukee, Wis. PLOUGH, PROF. HAROLD H., Amherst College, Amherst, Mass. POLLISTER, DR. A. W., Columbia University, New York City, N. Y. POND, DR. SAMUEL E., University of Pennsylvania, School of Medicine, Philadelphia, Pa. PRATT, DR. FREDERICK H., Boston University, School of Medicine, Bos- ton, Mass. RAFFEL, DR. DANIEL, Osborn Zoological Laboratory, Yale University, New Haven, Conn. RAND, DR. HERBERT W., Harvard University, Cambridge, Mass. RANKIN, PROF. W. M., Princeton University, Princeton, N. J. REDFIELD, DR. ALFRED C.. Harvard University, Cambridge, Mass. REESE, PROF. ALBERT M., West Virginia University, Morgantown, W. Va. REINKE, DR. E. E., Vanderbilt University, Nashville, Tenn. REZNIKOFF, DR. PAUL, Cornell University Medical College, New York City, N. Y. RHODES, PROF. ROBERT C., Emory University, Atlanta, Ga. RICE, PROF. EDWARD L., Ohio Wesleyan University, Delaware, O. RICHARDS, PROF. A., University of Oklahoma, Norman, Okla. RICHARDS, DR. O. W., Osborn Zoological Laboratory, Yale University, New Haven, Conn. RIGGS, MR. LAWRASON, JR., 25 Broad Street, New York City, N. Y. ROBERTSON, PROF. W. R. B., 1803 Anderson Street, Manhattan, Kans. ROGERS, PROF. CHARLES G., Oberlin College, Oberlin, O. ROMER, DR. ALFRED S., University of Chicago, Chicago, 111. ROOT, DR. W. S., Syracuse Medical School, Syracuse, N. Y. SANDS. Miss ADELAIDE G.. 562 King Street, Port Chester, N. Y. SCHRADER, DR. FRANZ, Department of Zoology, Columbia University, New York City, N. Y. SCHRAMM, PROF. J. R., University of Pennsylvania, Philadelphia, Pa. SCOTT, DR. ERNEST L., Columbia University, New York City, N. Y. SCOTT, PROF. G. G., College of the City of New York, New York City, N. Y. SCOTT, PROF. JOHN \V., University of Wyoming, Laramie, Wyo. SCOTT, PROF. WILLIAM B., 7 Cleveland Lane, Princeton, N. J. SHULL, PROF. A. FRANKLIN, University of Michigan, Ann Arbor, Mich. SHUMWAY, DR. WALDO, University of Illinois, Urbana, 111. SIVICKIS, DR. P. B., Pasto deze 130, Kaunas, Lithuania. SNOW, DR. LAETITIA M., Wellesley College, Wellesley, Mass. SNYDER, PROF. CHARLES D., Johns Hopkins University Medical School, Baltimore, Md. 46 MARINE BIOLOGICAL LABORATORY SOLLMAN, DR. TORALD, Western Reserve University, Cleveland, O. SONNEBORN, DR. T. M., Johns Hopkins University, Baltimore, Md. SPEIDEL, DR. CARL C., University of Virginia, University, Va. SPENCER, PROF. II. J., 24 West 10th Street, New York City, N. Y. STARK, DR. MARY B., New York Homeopathic Medical College and Flower Hospital, New York City, N. Y. STEWART, DR. DOROTHY R., Skidmore College, Saratoga Springs, N. Y. STOCK ARD, PROF. C. R., Cornell University Medical College, New York City, N. Y. STOKEY, DR. ALMA G., Mount Holyoke College, South Hadley, Mass. STRONG, PROF. O. S., College of Physicians and Surgeons, 630 West 168th Street, New York City, N. Y. STUNKARD, DR. HORACE W., New York University, University Heights, N. Y. STURTEVANT, DR. ALFRED H., California Institute of Technology, Pasa- dena, Calif. SUM WALT, DR. MARGARET, University of Pennsylvania, Medical School, Philadelphia, Pa. SWETT, DR. FRANCIS H., Duke University Medical School, Durham, N. C. TASHIRO, DR. SHIRO, Medical College, University of Cincinnati, Cin- cinnati, O. TAYLOR, WILLIAM R., University of Michigan, Ann Arbor, Mich. TENNENT, PROF. D. H., Bryn Mawr College, Bryn Mawr, Pa. THATCHER, MR. LLOYD E., Spring Hill, Tenn. TRACY, PROF. HENRY C., University of Kansas, Lawrence, Kans. TREADWELL, PROF. A. L., Vassar College, Poughkeepsie, N. Y. TURNER, PROF. C. L., Northwestern University, Evanston, 111. TYLER, DR. ALBERT, California Institute of Technology, Pasadena, Calif. UHLEMEYER, Miss BERTHA, Washington University, St. Louis, Mo. UHLENHUTH, DR. EDUARD, University of Maryland, School of Medi- cine, Baltimore, Md. UNGER, DR. W. BYERS, Dartmouth College, Hanover, N. H. VAN DER HEYDE, DR. H. C., Galeria, Corse, France. VISSCHER, DR. J. PAUL, Western Reserve University, Cleveland, O. WAITE, PROF. F. C., Western Reserve University Medical School, Cleveland, O. WALLACE, DR. LOUISE B., Spelman College, Atlanta, Ga. WARD, PROF. HENRY B., University of Illinois, Urbana, 111. WARREN, DR. HERBERT S., Department of Biology, Temple University, Philadelphia, Pa. WARREN, PROF. HOWARD C., Princeton University, Princeton, N. J. REPORT OF THE DIRECTOR 47 WENRICH, DR. D. H., University of Pennsylvania, Philadelphia, Pa. WHEDON, DR. A. D., North Dakota Agricultural College, Fargo, N. D. WHEELER, PROF. W. M., Museum of Comparative Zoology, Cambridge, Mass. WHERRY, DR. W. B., Cincinnati Hospital. Cincinnati, O. WHITAKER, DR. DOUGLAS M., Stanford University, Stanford Univer- sity, Calif. WHITE, DR. E. GRACE, Wilson College, Chambersburg, Pa. WHITING, DR. PHINEAS W., University of Pittsburgh, Pittsburgh, Pa. WHITNEY, DR. DAVID D., University of Nebraska, Lincoln, Nebr. WIEMAN, PROF. H. L., University of Cincinnati, Cincinnati, O. WILLIER, DR. B. H., University of Chicago, Chicago, 111. WILSON, PROF. H. V., University of North Carolina, Chapel Hill, N. C. WILSON, DR. J. W., Brown University, Providence, R. I. WITSCHI, PROF. EMIL. University of Iowa, Iowa City, la. WOGLOM, PROF. WILLIAM H., Columbia University, New York City, N. Y. WOODRUFF, PROF. L. L., Yale University, New Haven, Conn. WOODWARD, DR. ALVALYN E., Zoology Department, University of Michigan, Ann Arbor, Mich. YOUNG, DR. B. P., Cornell University, Ithaca, N. Y. YOUNG, DR. D. B., University of Maine, Orono, Me. ZELENY, DR. CHARLES, University of Illinois, Urbana, 111. THE REACTION OF THE ERYTHROCYTES OF VERTE- BRATES, ESPECIALLY FISHES, TO VITAL DYES ALDEN B. DAWSON (From tlic Zoological Laboratories, Harvard University, and the Marine Biological Laboratory, Woods Hole, Massachusetts) In the erythrocytes of most vertebrates, but especially of fishes and amphibians, discrete granules are characteristically present. In the urodele, Ncctunis, bipolar clusters of such granules are regularly found in mature red blood cells. They are visible in fresh preparations, demonstrable as basophilic bodies with Wright or Giemsa staining, are blackened with osmic acid and silver salts, and are stained with iron haematoxylin after Helly fixation. Accordingly, there is no doubt that they are preexistent structures and are not induced in supravital prep- arations by the action of the dyes. However, secondary granules may also appear in such cells ; the concentration of the dye, age of the preparation, brilliancy of illumina- tion, and increase in temperature being effective as formative factors, influencing the rate and manner of their appearance (Dawson, 1928, 1929, 1930). Moreover, with higher concentrations of dye, the red cells may also exhibit elaborate patterns of reticulation. The genetic relation of both types of granules, preexisting and induced, to the reticu- lated substance is an unsettled question. Morphologically there is no distinction between the two types and both are frequently enclosed in the reticular filaments. The reticulation pattern is apparently derived, through a reaction with the vital dye, from the basophilic substance which occurs diffusely in erythrocytes, and secondary granules are re- garded by some as the same substance in a different form. The literature on this subject is voluminous and many different views are advanced regarding the significance of the protoplasmic constituents reacting with vital dyes. First of all, the ability to react with a given vital dye may not be a specific characteristic of a single substance within the cytoplasm. Again, in many instances, the materials reacting with vital dyes are not always readily demonstrated by other technical meth- ods. Nectnnis erythrocytes appear to be exceptionally favorable in this respect. Moreover, the amount of material reacting with vital dyes appears to decrease gradually as the erythrocytes differentiate, but the acquisition of hemoglobin in a given species proceeds only up to a cer- tain stage at which the cell is said to be mature. Since the erythrocytes in different vertebrates do not attain at maturity the same relative degree 48 REACTION OF ERYTHROCYTES TO VITAL DYES 49 of differentiation, the picture in supravital preparations is subject to great variation ; the individual changes which occur in the maturation of the red blood cells may apparently proceed at different rates and to different degrees in different species. Accordingly, the relative con- centration of hemoglobin within the erythrocytes of different species appears to be to some extent independent of the amount and distribution of material reacting with vital dyes. These are only some of the factors which complicate the picture and make generalizations almost impossible. Coupled with these there is also the lack of complete information about the stages of maturation of these cells in many species. In other cases, too, findings by one method have not been adequately checked and con- firmed by other technics. Various theories of the origin and nature of the vitally-stained bodies have been advanced. Several earlier workers suggested a nuclear or nucleolar origin (Giglio-Tos, 1896; Jolly, 1903; Sabrazes et Muratet, 1908), while others have apparently confused them with centrosomes (von Apathy, 1897; Bremer, 1895; Dehler, 1895; Eisen, 1897-1899; Golgi, 1920) when demonstrated by non-vital staining methods such as silver nitrate impregnation or staining with iron haematoxylin. They were also identified as intracellular parasites, but this view was soon dropped (Sabrazes et Muratet, 1900). At present the vital granules are generally regarded as of cytoplasmic origin. Giglio-Tos (1896) suggested that they represented hemoglobin- forming substances and Yoffey (1929) is inclined to favor this view. Several modern investigators (Ferrari, 1930; Villa, 1930; Knoll, 1931) have advanced evidence in favor of an intranuclear origin of hemoglobin, and this interpretation accordingly would support the old hypothesis of Giglio-Tos that vital granules arise as nuclear emissions and are con- cerned in the elaboration of hemoglobin. On the basis of the reaction of these bodies with silver salts and osmic acid, they have been regarded as possibly homologous with the Golgi apparatus — dictyosomes (Bhat- tacharya and Brambell, 1925; Dawson, 1928, 1930; Dornesco and Steo- poe, 1930a, I930b; Urtubey, 1927). Nittis (1930) described a surface granule which in his opinion might be interpreted as the point of separa- tion of two daughter cells or as a part of a trophospongial system. Nit- tis' view seems untenable, as the single granule is intracellular and may frequently exhibit Brownian motion (Dawson, 1931). Many tend to class all vital granules as artifacts (Chlopin, 1927; Weidenreich, 1903), failing to distinguish sharply between preexistent and induced bodies (Beams, 1930). The number and distribution of vital granules in the erythrocytes of vertebrates vary during the differentiation of the cell. Usually they are 50 A. B. DAWSON not present in very young cells and are reduced or even disappear in old cells. In the intermediate stages they are frequently numerous and con- spicuous. In the mature cells of different vertebrates in which they still persist, vital granules frequently occur in rather definite numbers and have a characteristic position within the cell. The more common types of distribution are single unipolar; multiple, clustered unipolar; multiple, clustered bipolar; multiple perinuclear; and multiple scattered or dif- fuse (Dawson and Charipper, 1929; Dawson, 1930, 1931). The number and distribution of vital granules in the erythrocytes of cyclostomes and fishes vary considerably. They were apparently recog- nized as early as 1896 by Giglio-Tos, who observed them in the blood cells of the lamprey and called attention to the striking Brownian move- ments they exhibited. A few years later Sabrazes and Muratet (1900) described " corpuscles mobiles " within the erythrocytes of Hippo- campus. They were irregularly distributed, with usually five to ten granules within a cell. In the same year these investigators extended their observations to several more fishes. In Torpedo oculata the granules were numerous and distributed diffusely around the nucleus. The cells usually con- tained as many as forty granules although some had as few as three, four, and five. All granules exhibited Brownian movement. They were unequal in size ; some were coupled and some elongated or com- pressed. In Raia pastinaca the bodies were frequently oval and rela- tively large but were not so numerous as in Torpedo. The blood cells of Synynathns typhlc apparently did not contain any granules. In Petromizon marinus and Alosa finta the granules were present but not numerous. In adult Anyitilla vulyaris they were absent. In 1902, however, Sabrazes and Muratet found opportunity to examine the blood of some young eels, 6 to 7 cm. long. In these, vital granules with limited Brownian movement were present in about twenty per cent of the erythrocytes. They varied greatly in size, with usually one to three present in any one cell. In Torpedo maruiorata (Sabrazes et Muratet, 1908) the number and distribution of vital granules is very similar to Torpedo oculata. Jokl (1925) made an intensive study of the erythrocytes of Raia elavata and Raia batis and found in the main that the red blood cells were very like those of Torpedo in the number, size, and distribution of the vital granules. Further observations on the elasmobranchs were made by Lewis and Lewis (1926), who described numerous diffusely scattered perinuclear bodies in the dogfish and skate (Raia erinacca). These workers also described one to three neutral red granules in the erythrocytes of the sculpin, but illustrated the erythrocytes of the hake REACTION OF ERYTHROCYTES TO VITAL DYES 51 and cunner as lacking any granular inclusions. Stolz (1928) found that the erythrocytes of Cvprinns carpio contained numerous granules diffusely distributed about the nucleus. Yoffey (1929) confirmed the observations of Jokl (1925) on the red blood cells of Raia clavata and Raid bails, and noted numerous fine basophilic granules in the mature cells of Trlgla gurnardus. Dornesco and Steopoe (1930o) found that the erythrocytes of the dogfish are typical of the elasmobranchs in pos- sessing numerous, scattered, actively motile granules. They (1930&) also investigated the blood of several marine teleosts, Syngnathus acus, Blcnnlus f>ho!is, Solca Tidgarls, Plcuroncctcs platcssa, Gobhts paganellus, Coitus buballs, Labnts inclops, Onos iniistclla and Xcrophis luuibrici- fonnis. In all of these they found that the erythrocytes contained uni- formly one granule, usually located eccentrically near one pole of the nucleus. The conditions in the erythrocytes of Anichtnts ncbitlosus are the same as in the marine teleosts (Dawson. 1931). In the erythrocytes of the fishes studied the distribution of vital granules is not so variable as in the Amphibia, being limited chiefly to two types, single unipolar and multiple perinuclear. restricted or diffuse. The characteristic unipolar and bipolar clusters appearing in the red blood cells of many urodeles are not found. It has seemed advisable to extend the observations on the blood of fishes, with a view to obtaining more information regarding the relationships between the preexistent and induced granules and the reticular substance as revealed by vital dyes. MATERIAL AND METHODS During the summer of 1931 the blood of seventeen different species of fish, taken in the vicinity of Woods Hole, was examined. Supravital staining was carried out by the dry dye-film method, using neutral red alone, neutral red in combination with Janus green, and brilliant cresyl blue. The concentration of neutral red, 1 : 1250, previously used on amphibian blood proved satisfactory in the case of fishes, but more Janus green had to be added to obtain a clear view of the chondriosomes. Films of brilliant cresyl blue were made with a saturated solution in absolute alcohol. Permanent preparations of supravitally stained blood were made by the method of Scott ( 1928) . Smears stained by Wright's method were also prepared. In all cases the blood was freshly drawn from the heart of living fish. It is essential that neutral red alone be used to check neutral red- Janus green B preparations, since the toxicity of Janus green may cause injuries which result in more rapid induction of bodies stainable with neutral red. On the other hand, Janus green B cannot safely be used alone, since bodies regularly stained with neutral red may take up Janus 52 A. B. DAWSON green and confuse the chondriosome picture. However, when the dyes are used in combination, the neutral red displays a greater affinity for the bodies which are regularly stained by it and the Janus green reaction is accordingly confined entirely to the chondriosomes. The brilliant cresyl blue was applied at such high concentrations that the nuclei were stained and the reticular patterns produced almost immediately, and there did not appear to be any progressive induction of formed bodies (granules) while the preparations were being studied for reticulation patterns. DESCRIPTION Before proceeding to the description of the reactions of the erythro- cytes of the different fishes to vital dyes, certain general features of these reactions will be discussed. In most fishes there are sufficient immature red cells, in varying stages of differentiation, that the pictures presented after supravital staining lack uniformity. This makes the problem of interpretation more difficult, since the differences in hemo- globin concentration in some cases are scarcely perceptible. Differences in the degree of maturity of such cells are most strikingly demonstrated with brilliant cresyl blue and the reticulation patterns furnish an excel- lent index of cell age. The presence of young red cells in the blood is readily confirmed in smears differentially stained by Wright's method, where varying degrees of basophilia and polychromasia are clearly dem- onstrated. As the erythrocytes mature the amount of material reacting with brilliant cresyl blue gradually decreases and the patterns of reticulation also undergo changes in form and distribution. In very young cells the entire cytoplasm is filled with fine, densely massed granules with little evidence of real reticulation. In succeeding stages of differentia- tion the pattern assumes a true reticular form. At first the meshes are small and distributed throughout the whole cell. Later the meshes are more open and the reticulation does not extend completely to the pe- riphery of the cell, assuming the form of a perinuclear wreath. As the cells grow older, the meshes are still wider and the filaments eventually are partially interrupted, forming an open, fragmented wreath. As the reticular substance is further reduced, the filaments of the net are more frequently interrupted and the wreath-like form as well as the reticular appearance is lost. The more or less separated filaments then appear radially arranged about the nucleus, with very fine fragments inter- spersed between them. With the further disappearance of reticular substance the pattern is reduced to scattered, irregular, short filaments and fine, dust-like granules. In final stages the filamentous form may practically disappear, leaving a variable number of fine granules. REACTION OF ERYTHROCYTES TO VITAL DYES It is difficult to determine whether the reticular substance, which is demonstrated with concentrations of the vital dye sufficiently high to stain the nuclei immediately, is of the same constitution as the relatively large granules which may he induced in similar blood cells by a pro- longed exposure to the same dye in a lesser concentration, which will not stain the nucleus or produce reticulation. At present, it seems impos- sible to decide this question. On the other hand, it is frequently pos- sible to distinguish clearly between preexistent vital granules and the induction patterns, since in many cases the former may be seen in fresh untreated cells or may be demonstrated in fixed cells by a variety of •• J J reliable technical methods. With high concentrations of brilliant cresyl blue, nuclei, nucleoli (when present), preexistent and induced granules, and reticular sub- stance are all stained. The nuclei are pale blue, nucleoli deep blue, all granules deep blue-purple and the reticular substance a red-violet. The nucleoli. varying in number from one to three, are large and conspicuous in the erythroblasts but become progressively smaller as the cells mature. They may be distinguished as minute bodies in cells of the " wreath " stage but seldom persist in more mature forms. The difference in the color reaction of the granules and the reticular substance may or may not be significant. The granules are sharply limited, dense, and highly refractile, while the reticular substance has an irregular appearance as if formed by the aggregation of very fine particles. The difference in physical state accordingly may explain the different shades presented by the two types of bodies. The form and distribution of chondriosomes within the erythrocytes of the fishes vary in the different stages of maturity, but are nevertheless quite characteristic for the type of cell. In the younger stages they are quite numerous, granular, and diffusely scattered throughout the cyto- plasm. In later stages the number is greatly reduced and they are more closely aggregated about the nucleus. Their form, too, may be changed, many appearing as long, tortuous bodies. In fully mature cells the chondriosomes are usually entirely of the filamentous type, although occasional granular forms are seen. Some of these are not true gran- ules but represent filaments oriented lengthwise between the two flat- tened surfaces of the erythrocyte and may be observed to shift in posi- tion as the cell is modified by the injurious effects of the Janus green. It is difficult to obtain sharp images of chondriosomes in the fresh cell, and relatively high concentrations of dye must be used. As the staining progresses the chondriosomes are first seen as hazy blue-green bodies, and at this time the nuclear outline is also indistinct. Shortly both the nuclear membrane and chondriosomes are distinctly seen and 54 A. B. DAWSON the cytoplasm of the cell appears clearer. While these changes are tak- ing place striking movements of the mitochondria may be observed. Filaments which were closely applied to the nucleus may frequently swing out to lie at right angles to the nuclear surface or may move completely away, and others which were seen on end may assume a posi- tion parallel with the flattened surface of the cell. For purposes of description the fishes studied at Woods Hole will be divided into two major groups, based on the distribution of the pri- mary or preexistent granules as demonstrated by low concentrations of neutral red. In the first group the vitally stained granules are usually single, with occasionally one or more accessory bodies, and are located eccentrically near one pole of the nucleus. In the second group the granules are numerous and may either have a definite perinuclear ar- rangement or be scattered more or less diffusely throughout the cyto- plasm. The reaction of the primary granules to Wright's stain might also be used as a basis of classification, since in some fishes the granules uniformly give a basophilic reaction ; in others this reaction is limited to a varying number of cells ; while in still others the granules always re- main unstained. The three second groups would form natural sub- divisions of the major Group I, which is based primarily on the distribu- tion of granules, but in Group II this would not hold true since in no instances do the perinuclear or diffusely arranged granules react uni- formly with Wright's stain. Group I Toadfish, Opsaiuis tan (Linnaeus). With neutral red the primary granules ordinarily appear as single, unipolar bodies, although one or more accessory granules are often encountered (Fig. 1, b and c). The accessory granules are usually small, but in some cases all the granules are of equal size, resembling the variations described for Ameiunis (Dawson, 1931). After long exposure to neutral red the granules are increased in number and tend to form clusters about the primary bodies. Later, other secondary granules appear irregularly throughout the cell. The mitochondria are relatively few, varying in number from three to six. They usually appear as wavy filaments but are frequently di- lated to encapsulate a large spherical refractile body. These bodies are readily seen after the Janus green has been reduced. They do not take up neutral red and their significance is not known. Very little reticular substance is demonstrated by brilliant cresyl blue. The filaments are short and are usurJly radially arranged about the nucleus. Some end in contact with the cell membrane. Primary granules and a variable number of secondary granules are stained with REACTION OF ERYTHROCYTES TO VITAL DYES 55 this dye. In the cells with the least reticulation practically no induction of granules has taken place (Fig. 1, d). The cells with the most reticu- lation contain many more granules (Fig. 1, a}. All the cells appear relatively mature. An occasional cell in the " wreath " stage with a clear border was encountered, but there were no younger stages. In smears stained by Wright's method the primary granules appear as distinct basophilic bodies. No basophilic erythrocytes are present and only a few cells exhibit polychromasia. Tautog, Tautoga onitis (Linnaeus). The neutral red patterns are quite similar to those of the toadfish (Fig. 4, a and fr) . A number of erythrocytes possess granular mitochondria, but in the majority of the cells they are filamentous. The reticular substance is scanty in most cells (Fig. 4, d), but a few "wreath" stages were seen (Fig. 4, r). Occasional small cells (erythroblasts) with a dense reticulation and a conspicuous nucleolus are present. With Wright's stain the primary granules are stained blue and a few erythrocytes show basophilia and polychromasia. The number of immature red cells is, however, very small. Gunner, Tautogolabnts adspcrsus (Walbaum). The erythrocytes of this fish very closely resemble in their staining reactions those of the two preceding forms (Fig. 6, a and b). Fewer immature cells were noted than in the tautog, and in general the reticulation patterns are very meager (Fig. 6, c and d). The primary granules also appear as basophilic bodies with Wright's stain. Sea Bass, Centropristes si rial us (Linnaeus). The primary neutral red bodies are the same as in the erythrocytes of the preceding fishes (Fig. 3, a). Induction of secondary granules occurs rather freely (Fig. 3, b), but even when the secondary staining effect has appeared there are a number of apparently mature cells which do not show any reaction to the dye. The mitochondria are relatively few, the fila- mentous form predominating. With brilliant cresyl blue many cells fail to show any reticular pat- terns but contain only granules such as are seen after moderate induc- tion with neutral red. Other cells contain a variable number of reticular filaments (Fig. 3, c), but the "wreath" stage was observed in only a few cells (Fig. 3, d). Occasional young cells with a fine, dense reticu- Itim and conspicuous nucleoli were encountered. In the sea bass virtually all the erythrocytes stain orthochromatically with the eosin in Wright's stain. Polychromasia is rarely seen and occasional round basophilic erythroblasts are present. The primary granules demonstrated so readily with neutral red are not usually stained in the smears. Some of the larger granules appear as distinct baso- 56 A. B. DAWSON philic bodies, other smaller granules are barely distinguishable, and many cells appear not to have any granular inclusions. Sea Robin, Prionoins caroliinis (Linna?us). The erythrocytes of the sea robin in their staining reactions are very like those of the sea bass (Fig. 2). The mitochondria, however, usually appear as short rods and granules rather than filaments. The reticular patterns are very sparse, being represented by scattered particles radially aligned about the nucleus (Fig. 2, d). All the cells appear to be mature. This conclusion is supported by the staining reactions of the erythrocytes with Wright's stain. No granular inclusions could be distinguished in the stained smears. Scup, Stenotomus chrysops (Linnaeus). The neutral red patterns and mitochondria present no unusual features (Fig. 7, a and c}. The reticular patterns are quite variable, although in the majority of cells the filaments are reduced to a minimum (Fig. 7, b) . Young cells with a dense reticulum and conspicuous nucleoli are frequently seen and stages with fragmented open-meshed " wreaths " are quite numerous (Fig. 7, d). The presence of such immature cells is confirmed by an examination of stained smears, but the preexistent or primary granules were not demonstrable as basophilic bodies. Butterfish, Poronotus triacanthus (Peck). The primary granule is characteristically present in neutral red preparations (Fig. 8, b} and in- duction of new granules proceeds very slowly. The chondriosomes are usually granular. No long filamentous forms were observed. The reticular patterns and numbers of immature cells are about the same as in the scup (Fig. 8, a, c, and d'). Variegated minnow, Cyprinodon variegatus Lacepede. The pri- mary neutral red bodies usually appear as single or double bodies and induction of secondary bodies occurs slowly. The mitochondria are predominantly of the granular type and are relatively numerous (Fig. 11, a and b). In the majority of the cells the reticular material is very scanty (Fig. 11, r and d), but in a few cells complete perinuclear ;' wreaths " of reticular material were seen. Occasional younger cells, possessing large nucleoli and dense reticulations, were also noted. The primary neutral red bodies of mature cells did not give a basophilic re- action with Wright's stain, but in many polychromatic cells they could be distinctly seen as basophilic granules. Alummichogs, Fiiudiilus licicroclitits (Linnaeus) and F. inajalls (Walbaum). The behavior of the erythrocytes is essentially alike in these two species. The primary neutral red bodies are characteristically present (Fig. 12, b and r). The reticular substance in general is scanty, appearing as radially arranged filaments (Fig. 12, d). A few REACTION OF ERYTHROCYTES TO VITAL DYES cells in the " wreath " stage and an occasional young cell with a large nucleolus were seen. The primary grannies were not demonstrated by Wright's method. Common eel, Angnilla rostrata (Le Sneur). The neutral red bodies are very minute, almost at the limit of visibility (Fig. 5, a and a). However, they gradually increase in size on exposure to the dye and secondary granules slowly appear (Fig. 5, b} . The mitochondria are few in number and chiefly of the long, sinuous type. Brilliant cresyl blue reveals the presence of large numbers of immature erythrocytes, the reticular substance varying in distribution from complete " wreaths " to scattered isolated filaments and granules (Fig. 5, r). In stained smears many erythrocytes show varying degrees of cytoplasmic baso- philia. The primary granules frequently appear as basophilic bodies in cells which stain orthochromatically in eosin or show a slight poly- chromasia. However, they were not distinguishable in more basophilic cells. Sand dab, Hippoglossoidcs platessoides (Fabricius). The primary vital granules are readily demonstrated as single or double granules (Fig. 9, o). With longer exposure to the dye secondary granules, usually grouped in clusters, quickly appear (Fig. 9, r and rf). Chondri- osomes of both filamentous and granular types are present. Practically all the cells are mature, showing a very sparse reticulation (Fig. 9, b) . An occasional cell with an incomplete open-meshed " wreath " was seen. A few basophilic erythrocytes were demonstrated with Wright's stain. The primary granules were frequently seen as basophilic bodies in eryth- rocytes which exhibit a slight polychromasia but could not be distin- guished in mature cells. Group II In the erythrocytes of the fishes of this group the primary granules are not limited to a single or double unipolar body but are relatively numerous and have a perinuclear distribution. Menhaden, Brcvoortia t\rannus (Latrobe). In the menhaden the granules are very small, frequently barely visible at a magnification of nine hundred diameters. The granules are either arranged in a single definite line about the nucleus or scattered somewhat irregularly throughout the cytoplasm (Fig. 10, a and b}. The irregular distribu- tion of granules appears to be characteristic of the oldest cells. On longer staining with neutral red they increase in both size and number. Mitochondria of both the filamentous and granular types are present in restricted numbers. All stages of maturating erythrocytes were encountered, but the majority of the cells were fully differentiated. The reticular patterns 58 A. B. DAWSON vary all the way from a dense compact granular mass filling the entire cell to the mature condition in which only fine particles and scattered, short filaments are present (Fig. 10, c and d). In stained smears eryth- rocytes in the different stages of differentiation are clearly shown and the numerous perinuclear primary granules are seen distinctly as baso- philic bodies in both polychromatic and mature cells. Alewife, Poinolobiis pseudoharengus (Wilson). The erythrocytes of the alewife are essentially like those of the menhaden in all their reactions to the dyes, but more immature cells are present (Fig. 14). Many of the younger cells contain large, clear, refractile globules which appeared as vacuoles in stained smears. Mackerel, Scomber scombrus Linnaeus. The erythrocytes of the mackerel differ only slightly from those of the menhaden and alewife (Fig. 13). The primary neutral red patterns are alike. Mitochondria of the filamentous type predominate. More immature cells are present than in the alewife and the reticulation patterns are accordingly very variable. The red cells react rapidly with neutral red and many sec- ondary granules may develop. The primary neutral red granules are demonstrated as basophilic bodies by Wright's stain, but much more care must be taken in the differentiation of the stain. Smooth dogfish, Mustelus canis (Mitchill) and spotted skate, Raja dlaphanes Mitchill. The vital staining reactions of the erythrocytes of both these forms have been described by several investigators. The granules appearing after exposure to neutral red are very large and numerous (Figs. 15 and 17). The mitochondria are usually long and filamentous. Induction of new granules occurs rapidly. There is a relatively high proportion of immature red cells, and the reticulation patterns are very variable. In old cells the reticular substance is greatly reduced, appearing as scattered particles and short filaments. With brilliant cresyl blue both vital granules and reticulation pat- terns are demonstrated as in the teleosts, but on long standing the dye in the granules disappears while the reticular substance remains bril- liantly stained. The reduction of the dye in the granules appears to be characteristic only of the elasmobranchs and has been previously reported by Jokl (1925). Although the granules are large and readily seen in fresh unstained preparations, they cannot be demonstrated in smears stained by Wright's method and do not give a basophilic reaction. THE RELATIVE DEGREE OF DIFFERENTIATION OF THE MATURE ' ERYTHROCYTES OF VERTEBRATES In the running description of the findings in the erythrocytes of the different species of fish, the staining reactions described related particu- REACTION OF ERYTHROCYTES TO VITAL DYES 59 larly to the predominating cells, presumably the mature ones. During the course of the study it became increasingly obvious that no granules, either preexistent or induced, were present in young cells exhibiting a complete, dense reticulation pattern. In somewhat older stages the primary granules were readily demonstrated, but induction of new gran- ules either did not occur or proceeded very slowly. This secondary reaction varied greatly in different species, but the general trend was the same in all. In more mature stages induction usually occurred more readily, but in some old, perhaps senile, cells, which gave practi- cally no reticulation reaction with brilliant cresyl blue, the secondary response to vital dyes again decreased. The terms " primary " or " preexistent," and " secondary " or " in- duced " are used here without reservation to designate the granules which are under discussion, since the cumulative evidence gathered from studies on fishes and amphibians indicates that such a distinction is valid. The granules which stain so readily with neutral red are fre- quently seen as refractile bodies in fresh preparations, and in dry-fixed smears are frequently demonstrated as basophilic inclusions. In most cases the characteristic form and location of these elements make their identification easy when they are rendered visible by other than supra- vital methods. Dornesco and Steopoe (1930a, 1930/?) have also suc- cessfully blackened these primary bodies with silver nitrate methods (Da Fano and Cajal) in both teleosts and elasmobranchs. Their dem- onstration by a silver method in the elasmobranch erythrocyte is signifi- cant, since in these cells the granules fail to give a basophilic reaction in carefully stained smears. The reality of the secondary granules cannot be disputed since they can be observed as they arise within the cells. Accordingly we have to deal with three distinct morphological entities, at least as they are demonstrated by supravital methods. The question of their chemical constitution and relationships is a baffling one and apparently little further progress can be made in this regard until better methods of study are devised. Nevertheless, the characteristic reactions to vital dyes may be legiti- mately used as evidences of the progressive changes which occur within the differentiating erythrocyte, and such reactions therefore constitute useful criteria for determining the degree of differentiation reached by the mature erythrocytes of a given species. The criteria of maturity in one species are not necessarily valid in every detail for erythrocytes of another species, since the pictures obtained by vital dyes may vary in different groups of animals. Maturity of erythrocytes in general can best be defined as the stage at which the cell acquires its maximum 60 A. B. DAWSON concentration of hemoglobin, and in normal animals cells of this type should predominate in the circulation. The degree of concentration of hemoglobin in mature cells in a given species is relatively constant and is closely correlated with the acquisition of characteristic staining reac- tions, but similarity of staining patterns in different species does not by any means indicate that the different erythrocytes have acquired the same concentration of hemoglobin. In other words, the erythrocytes of the different vertebrates are mature at varying levels of differentiation, the latter being measured by such staining reactions as are usually regarded as evidences of immaturity in the most highly differentiated red cells. During the differentiation of the vertebrate erythroblast a striking series of changes occurs. Some are readily demonstrated in fixed prep- arations while others are adequately revealed only by supravital staining. Most of these changes are common to the erythrocytes of all vertebrates, but in the mammals an extreme degree of specialization is encountered. During differentiation the nuclear-cytoplasmic ratio changes greatly, the nucleoli undergo a gradual involution and may disappear, the distribu- tion of chromatin in the nucleus is modified, and the basophilia of the cytoplasm is gradually lost and replaced by the eosinophilia of the hemoglobin. Also the volume of mitochondrial substance is progres- sively reduced, and frequently the form of the individual elements is changed. The gradual reduction of the reticular substance, definitely correlated with a decreasing basophilia of the cytoplasm, is the most striking fea- ture of the maturing erythrocytes when seen in supravital preparations. Less conspicuous changes involve the appearance and behavior of both types of granular inclusions, preexistent and induced. Primary gran- ules are usually absent from very young cells and disappear in a later but somewhat variable stage of differentiation. When they first appear they do not give a basophilic reaction in stained smears, but later in many instances they are characteristically basophilic. In mature cells of some animals the basophilic reaction is subsequently lost and the bodies are again demonstrated only by supravital methods. Moreover, in some vertebrates the primary granules may entirely disappear while in others they persist and assume characteristic patterns of distribution. Other granules of similar morphology and behavior may appear in cells that have stood in preparations for some time. Apparently induction of granules can occur only after some degree of differentiation of the erythrocyte has been attained, but the ability to respond in this manner may persist until full maturity is reached, and in many instances in- duced granules may appear even after the primary granules are no longer in evidence. In the mammals, however, the nucleus and all the cytoplasmic inclusions eventually disappear completely. REACTION OF ERYTHROCYTES TO VITAL DYES 61 In the different vertebrates each of these changes in the maturing erythrocyte may take place at a different rate and to a varying degree. Accordingly, with the exception of the mammals, it is not always easy to decide which cells represent at maturity the more fully differentiated stage. In all vertebrates below the mammals the cells are character- istically nucleated with a few striking exceptions in the Amphibia (Emmel, 1924). In all of these animals the basophilia is eventually replaced by eosinophilia. The nucleoli may disappear; the nuclear- cytoplasmic ratio undergoes considerable changes, the nucleus becoming condensed and acquiring a characteristic chromatic pattern. That is, in ordinary stained smears the mature nucleated erythrocytes of vertebrates appear essentially alike except that they vary greatly in size. However, with the more delicate methods of supravital staining, quite striking differences are brought out. Accordingly, the patterns of granulation and reticulation afford the best criteria of the degree of differentiation of these cells, and of these two criteria the degree of persistent reticula- tion is probably the better since the reticular material is continuously present in the cell while the granules have a variable and much more complicated history. Still, the latter criterion cannot be entirely dis- regarded in comparing the nucleated red cells of the vertebrates. Adequate data for the comparison of the erythrocytes of vertebrates are available for fishes and amphibians, but the blood of reptiles and birds has not been studied so extensively by means of vital dyes. In order to secure first-hand information concerning the conditions in the reptiles and birds, supravital preparations of blood from the painted turtle, horned toad, fence lizard (Sceloporus undulatus), and the com- mon fowl were studied. In the fishes (both elasmobranchs and teleosts) and birds little retic- ular substance is present in the mature cells (Fig. 18, a, c, and rf), being represented mostly by scattered filaments and granular fragments. My observations on the fowl differ from those of Doan, Cunningham, and Sabin (1925). who report that the final stages of reticulation consist in a few discrete bodies stainable with vital dyes. These bodies are readily demonstrated by low concentrations of either neutral red or brilliant cresyl blue (Fig. 18. b), but when brilliant cresyl blue in sufficiently high concentration to stain the nuclear reticulum is used, scattered fragments and short filaments are also uniformly present in the cytoplasm. Bril- liant cresyl blue is apparently reduced to some extent in the cell and the demonstration of persisting fragments of reticulum is possible only when the dye is present in considerable excess. In the amphibians and reptiles, on the other hand, a definite, more or less complete reticular pattern can be demonstrated in the mature erythrocytes, but in most 62 A. B. DAWSON instances the amphibian erythrocytes contain the greater amount of reticular substance (Figs. 16, 19, 20, and 22). If the degree of per- sistence of reticulation is regarded as evidence of the degree of differ- entiation attained by the mature erythrocyte, it must be concluded that the red blood cells of fishes and birds are relatively more highly differ- entiated than those of the amphibians and reptiles. The amount of material demonstrable as so-called primary vital granules in the mature erythrocytes of the several classes of vertebrates cannot be readily correlated with the degree of persistent reticular sub- stance, but the history of the vital granules indicates that their presence in red blood cells is in some degree a measure of relative maturity. It has been already pointed out that the reticular substance is at a maximum in the young cells and gradually decreases as the erythrocytes mature, while in general the granular substance is not present at all until the cells are partially mature. It is practically impossible to make any significant generalizations regarding the vital granules. The erythrocytes of each species must be considered separately if any accurate conclusions are to be drawn re- garding relative maturity of the cells. Tt has been held by many that the granular material is but a phase of the filamentous reticulum. In very young cells the reticular substance is definitely granular before acquiring the filamentous form. The granular form is doubtless de- pendent on its concentration within the cell and is the result of the agglutinating or precipitating effect of the vital dye. The primary granules are, however, definite, discrete bodies, frequently distinguish- able in fresh unstained cells, and may often be demonstrated in fixed material. They are apparently associated in some way with cell metabo- lism, and when the erythrocytes are exposed to a penetrating dye it accumulates first of all in these preformed structures. Whether the secondarv or induced bodies are derived from reticular material is less j easy to determine. But certain lines of evidence would appear to indicate that they are the result of a specific reaction of the cytoplasm to an excess of dye and do not represent local, sharply delimited accumu- lations of reticular material (Chlopin, 1927). The history of the behavior of these three elements in the maturing erythrocytes further suggests that they are separate entities. In elasmo- branchs the primary granules of mature cells are large and numerous when the reticular substance is reduced to the same degree as in the teleosts. In the amphibians the history is also variable, granules being either present or absent, depending on the species (Arrigoni, 1908; Beams. 1930; Dawson and Charipper, 1929; DeRoo and Ufford, 1930; Goda, 1929; Hibbard, 1928; Jordan, 1925; Stolz, 1928), while the REACTION OF ERYTHROCYTES TO VITAL DYES 63 reticular material persists as a fairly complete, open network. In Tri- tnnis I'iridescens the erythrocytes rarely contain any granules either primary or secondary, but as Nigrelli (1929) has shown, the reticulation in mature cells is relatively abundant (Fig. 16). Other similar ex- amples could be cited. In the reptiles the reticulation is quite abundant in mature cells but the primary granules are single in the alligator, horned toad, and fence lizard, and clustered at one pole in the painted turtle. The granules in these several species can also be demonstrated as basophilic bodies in dried smears stained by Wright's method (author's observations), and the question of their being induced bodies cannot be raised. In the common fowl and pigeon, Doan, Cunningham, and Sabin (1925) regard the single vital granule as a vestige of the reticular substance. Forkner (1929) also states that in the domestic fowl the cytoplasm of the eryth- rocytes contains, after staining with neutral red, from none to several small, reddish brown bodies which are usually near the nucleus but often move about and may be far out near the cell border. These bodies are not demonstrated in smears stained by Wright's method but, as pointed out earlier, they have probably been demonstrated by other methods and mistaken for centrosomes. Their close resemblance to similar struc- tures in teleosts and reptiles which can be shown to be primary bodies lends strength to the view that they are preexistent in the fowl erythro- cytes. and are not produced by a reaction of the dye with remnants of the reticular substance. In view of the evidence accumulated from a study of the nucleated erythrocytes of vertebrates, the author is in- clined to accept the conclusion of Michels (1931) that the reticular substance in reticulocytes has no genetic relation to the vital granules of the mature red cells, but would disagree with his acceptance of the view that vital granules are surface structures, either precipitates of the stain or stained precipitates of the plasma. After a study of the irregu- lar behavior of the vital granules in nucleated erythrocytes, it does not seem surprising that such granular cytoplasmic inclusions do persist even after the nucleus, chondriosomes, and reticular substance have dis- appeared from the mammalian cell. The appearance of primary gran- ules in the cytoplasm is apparently a constant phenomenon in the matu- ration of the erythrocytes. It is only in the mammalian erythroplastid that they uniformly completely disappear and in this instance they mark the acme of erythrocyte differentiation. They, however, are not always the last vestige of immaturity to disappear, since in some higher urodeles and several anurans they disappear while the reticular substance is still present in relatively large amounts. The appearance of secondary or induced granules in erythrocytes 64 A. B. DAWSON following exposure to vital dyes apparently has only a very limited rela- tion to the degree of differentiation attained by such cells. It is true that young cells with a high concentration of reticular substance react slowly and to a very limited degree to dyes, but the amount of reaction obtained in more mature cells is also very variable and seems not to be directly determined by the degree of differentiation attained. Rather the reaction appears to be species specific and to depend on the perme- ability of the cell and other factors inherent in the cytoplasm of the given species. Also the degree of the reaction with the cytoplasm can- not be correlated with the presence of a certain amount of reticular substance, since as much induction may occur in mature erythrocytes' of fishes with a minimum of reticulation as in those of amphibians and reptiles where the reticular substance persists in greater amounts. The induction phenomenon in cells containing primary granules is usually confined at first to the areas of the cell containing such preexistent bodies. In the early stages of induction the primary granules them- selves become enlarged and new granules then appear in their immedi- ate vicinity. Later new bodies may form irregularly throughout the cytoplasm. The phenomenon of induction in erythrocytes appears to be closely related to the " krinome " reaction to vital dyes described by Chlopin (1927) for many other cells of the animal body. Freely penetrating stains such as the basic dyes appear to accumulate within the cytoplasm of the red blood cells and appear first in the preformed bodies when they are present. Later, as more dye is accumulated, it is segregated by some reaction of the cytoplasm into newly formed structures. In some erythrocytes this accumulation and subsequent segregation of dye within the cytoplasm appears to continue progressively and to surprising limits, while in other erythrocytes the reaction proceeds only to a mini- mum extent. As has been already noted, many of the external factors influencing this reaction are known, but the factors within the cytoplasm, which are apparently of utmost importance, are unknown. The re- action, as in the anurans (Beams, 1930), proceeds as well in mature cells without preformed bodies as in cells in which preformed bodies are numerous and conspicuous. The final loss of the ability to react, as in the most highly differentiated mammalian erythrocytes, would seem to indicate that the degree of differentiation attained by the cell in some way determined or limited the cytoplasmic reaction to the dye. Such a conclusion, however, is rendered more or less untenable by the irregular behavior of the nucleated erythrocytes of other vertebrates, whose degree of differentiation at maturity can be estimated by the degree of per- sistence of reticulation. In such cells, except in very young stages, REACTION OF ERYTHROCYTES TO VITAL DYES 65 there is no significant correlation between the relative degree of differ- entiation attained and the degree of neo-formation of vitally stained bodies. Before concluding this discussion of the relative degree of differenti- ation of the mature vertebrate erythrocytes, one other morphological feature of maturing erytlirocytes should be mentioned, although the observations on it are not all comprehensive. It has already been noted that with relatively high concentrations of brilliant cresyl blue the nucle- oli in immature cells of the fishes appeared as deep blue bodies in a light blue, apparently homogeneous nucleus, the nuclear reticulum not being shown with such concentrations of the dye. The nucleoli (plasmo- somes) are relatively large in young cells and vary in number, but there are seldom more than three in any cell. They grow smaller as the erythroblasts differentiate and usually appear as single, small, spherical bodies. After the stage at which the reticular substance appears as an open-meshed, almost complete network the nucleoli are more rarely seen, and in the mature cells containing scattered fragments of reticulum they are usually absent. In the catfish, however, nucleoli are uniformly present in the mature erythrocytes. Nucleoli are not readily demonstrated in the mature cells of the Amphibia by brilliant cresyl blue, but they may be stained if concentra- tions of dye sufficiently high to bring out the chromatin reticulum of the nuclei are used. The dye, however, must not be intense enough to stain the chromatin a dark blue or the nucleoli are obscured (Fig. 16). In such preparations of the erythrocytes of Ncctunts and Tritnnis the nucleoli, numbering from one to four, appear as dark blue bodies lying between the coarse bars of light blue chromatin. They are somewhat irregular in form and are not so sharply outlined as in the fishes. At first it seemed doubtful if these bodies were nucleoli, but a comparison of the mature cells with younger stages in the circulation appears to establish their plasmosomal nature. In the reptiles studied (horned toad, fence lizard, and painted turtle) for this feature of the mature erythrocyte, the nucleolus is uniformly found as a single spherical body in all mature erythrocytes, and is a striking feature of all properly stained preparations (Figs. 19, 20, and 22). Immature cells were rarely encountered in the blood of these ani- mals and no comparisons with the nucleoli of younger erythrocytes were made. In the blood of the fowl no nucleoli could be distinguished in mature cells ; but in an occasional immature cell, still in the stage with a more or less complete reticular net, small single spherical nucleoli were observed. This method of demonstrating nucleoli, supravitally with brilliant 5 66 A. B. DAWSON cresyl blue, in red blood cells does not appear to have been previously employed. It seems to be a delicate method and to give clear pictures of nucleoli even when they cannot be easily demonstrated in fixed and stained preparations. In many amphibian erythrocytes brilliant cresyl blue also stains the achromatic contents of the nucleus a reddish violet, while the chromatin is a light blue and the nucleoli deep blue. In such cells the nucleoli appeared to be imbedded in the achromatic substance. These observations on the persistence of the nucleoli are correlated in a very satisfactory manner with the findings on the degree of per- sistence of reticular material and afford additional evidence that the mature erythrocytes of fishes and birds are relatively more highly differ- entiated than those of amphibians and reptiles. The degree of differentiation of erythrocytes of the several classes of vertebrates as determined by the criteria of persistent reticulation and presence of nucleoli also seems to correlate fairly well with the size of these cells. The Amphibia as a class have the largest erythrocytes (Fig. 21). Reptiles have blood cells next in size and fishes come next, then birds and mammals. It is difficult, however, to conceive that cell size could directly influence the degree of persistent reticulation or the persistence of nucleoli. The presence of primary granules might pos- sibly be dependent to some extent on this factor, since the products of metabolism might be less readily eliminated from larger cells and tem- porary accumulations be segregated in the cytoplasm in the form of granules. SUMMARY The reactions of the mature erythrocytes of seventeen species of fishes to the vital dyes neutral red, Janus green B, and brilliant cresyl blue, have been studied. In most teleosts the primary vital granules are readily demonstrated by neutral red and consist of one or two small granules eccentrically placed near one pole of the nucleus, but in the menhaden, alewife, and mackerel the primary granules are most nu- merous and are either arranged in a single definite line about the nucleus or scattered irregularly throughout the cytoplasm. In the elasmo- branchs the granules are large, numerous, and scattered. In a majority of the teleosts the primary granules may be demonstrated as basophilic bodies in dry films stained by Wright's method and are also frequently seen in fresh unstained preparations. Secondary or induced granules may also appear in the cytoplasm of cells exposed for long periods to the dye. The degree of induction of new bodies does not appear to depend entirely on external factors but is determined to a large extent by factors inherent in the cytoplasm of the REACTION OF ERYTHROCYTES TO VITAL DYES 67 given species. In general the mitochondria are filamentous, reduced in number, and lie in close contact with the surface of the nucleus. The reticular substance in all mature erythrocytes of the fishes is greatly reduced and appears either as short irregular filaments or as minute granular remnants. It is best demonstrated with brilliant cresyl blue. An attempt is made to compare the relative degree of differentiation attained by the mature erythrocytes of the several classes of vertebrates. The following criteria have been considered : changes in nuclear-cyto- plasmic ratio; chromatin distribution in the nucleus; involution of the nucleoli ; loss of basophilia ; changes in the form, distribution, and vol- ume of mitochondrial substance; reduction of reticular substance; amount of primary granules ; and degree of induction of secondary granules. Of these criteria the degree of persistence of reticulation has been found to be the most consistent, and on this basis the several classes of vertebrates are arranged in the following ascending order of relative differentiation attained by their erythrocytes at maturity : am- phibians, reptiles, fishes, birds, and mammals. This arrangement is also supported by the behavior of the nucleoli, which persist in the erythro- cytes of amphibians and reptiles but are not usually demonstrated in the mature cells of fishes and birds. The history of the primary and secondary granules is less regular and consequently less useful for measuring the relative differentiation attained by the cells of different classes of vertebrates. However, within a given class of vertebrates it is concluded that the presence of a large number of primary granules or the rapid induction of new granules in mature cells may be regarded as supplementary evidence of a lesser degree of differentiation. The presence of primary granules or the degree of induction of new granules, however, cannot always be cor- related with the degree of persistence of reticulation. It is also concluded on the basis of this survey of the vertebrate erythrocyte that primary granules, secondary granules, and patterns of reticulation as revealed by vital dyes, must be regarded as three separate entities which are not genetically related. LITERATURE CITED ARRIGONI, C., 1908. Ueber die Metamorphose des Kernes der menschlichen Ery- throblasten und iiber die Natur der chromatophilen Substanz der Erythro- zyten. Folia Hacin., 6: 444. BEAMS, H. W., 1930. The So-called ' Segregation Apparatus ' of the Erythrocyte of Frog and Necturus Blood. Anat. Rcc., 47: 341 (abstract). BHATTACHARYA, D. R., AND F. W. R. BRAMBELL, 1925. The Golgi Body in the Erythrocytes of the Sauropsida. Quart. Jour. Micros. Sci., 69: 357. 68 A. B. DAWSON BREMER, L., 1895. -Ueber das Paranuclearkurperchen cler gekernten Erythrocyten, nebst Bemerkungen uber den Bau der Erythrocyten in Allgemeinen. Arch. f. mikr. Anat., 45: 433. BREMER, L., 1895. Die Indentitat des Paranuclearkorperchens der gekernten Erythrocyten mit dem Centrosom. Arch. f. mikr. Anat.. 46: 618. CHLOPIN, N. G., 1927. Experimentelle Untersuchungen uber die sekretorischen Prozesse im Zytoplasma. I. Uber die Reaktion der Gewebselemente auf intravitale Neutralrotfarbung. Arch. f. expcr. Zellforscli., 4: 462. DAWSON, A. B., 1928. The ' Segregation Apparatus ' of the Amphibian Erythro- cyte and Its Possible Relation to the Golgi Apparatus. Anat. Rec., 39: 137. DAWSON, A. B., 1929. A Further Study of the Reaction bf the Amphibian Ery- throcyte to Vital Dyes, Osmic Acid, and Silver Salts, with Special Ref- erence to Basophilia and Reticulation. Anat. Rcc., 42: 281. DAWSON, A. B., 1930. Chondriome and Vacuome of the Differentiating Erythro- cyte of Necturus and Their Relation to the So-called Golgi Substance of Erythrocytes. Anat. Rcc., 46: 281. DAWSON, A. B., 1931. Supravital Studies on the Erythrocyte of the Catfish (Ameiurus nebulosus, Lesueur), with Special Reference to the Nittis stigma. Anat. Rcc.. 49: 121. DAWSON, A. B., AND H. A. CHARIPI'KR, 1929. A Comparative Study of the Amount and Distribution of the Neutral-Red Bodies in the Erythrocytes of Urodeles. Anat. Rcc., 43: 299. DEHLER, A., 1895. Beitrag zur Kenntnis des feineren Baues der roten Blut- korperchen beim Huhnerembryo. Arch. j. mikr. Anat.. 46: 414. DfiRoo, G. I., AND E. II. UFFORD. 1930. An Investigation of the Staining Reactions of Erythrocytes of the Leopard Frog to Nile-blue Sulphate, with Special Reference to the Nature of the Segregation Apparatus and Golgi Sub- stance. Anat. Rcc., 46: 297. DOAN, C. A., R. S. CUNNINGHAM, AND F. R. SABIN, 1925. Experimental Studies on the Origin and Maturation of Avian and Mammalian Red Blood-Cells. Contributions to Embryology, Vol. 16, Pub. Carnegie lust., No. 361, pp. 163-227. DORNESCO, G. TH., ET J. STEOPOE, 1930o. L'appareil de Golgi des globules rouges des Teleosteens marins. Comf>l. rend. Soc. Dial., 105: 288. DORNESCO, G. TH., ET J. STEOPOE, 1930/>. L'appareil de Golgi des hematics des Selaciens. Com fit. rend. Soc. Biol., 105: 446. EISEN, G., 1897-1899. Plasmocytes ; the Survival of the Centrosomes and Archo- plasm of the Nucleated Erythrocytes, as Free and Independent Elements in the Blood of Batrachoseps attenuatus Esch. Proc. Calif. Acad. Sci., Ser. 3, Zool., Vol. 1. EISEN, G., 1899. On the Blood Plates of the Human Blood with Notes on the Erythrocytes of Amphiuma and Necturus. Jour. Morph., 15: 635. EMMEL, V. E., 1924. Studies on the Non-nucleated Elements of the Blood. II. The occurrence and genesis of non-nucleated erythrocytes or erythro- plastids in vertebrates other than mammals. Am. Jour. Anat., 33: 347. FERRARI, R., 1930. Sur la formation de I'hemoglobine dans le noyau des erythro- blastes de grenouille. Arch. Ital. Biol., 84: 44. FORKNER, C. E., 1929. Blood and Bone Marrow Cells of the Domestic Fowl. Jour. E.\-pcr. Mcil.. 50: 121. GiGLio-Tos, E., 1896. Sur les cellules du sang de la Lamproie. Arch. Ital. Biol., 26: 93. GODA, T., 1929. Cytoplasmic Inclusions of Amphibian Cells with Special Refer- ence to Melanin. Jour. Fac. Sci., Imp. Unii'., Tok\o, Sec. IV, Zool., Vol. 2, pp. 51-122. REACTION OF ERYTHROCYTES TO VITAL DYES 69 GOLGI, C, 1920. II centrosoma dei globuli rossi del sangue circolante dcll'uomo e di altri animali. Hocin. Arch. Ital., Anno 1, pp. 333-359. HIBBARD, H., 1928. Contribution a 1'etude de I'ovogenese, de la fecondation, et de 1'histogenese chez Discoglossus pictus Otth. Arch, dc Biof., 38: 251. JOKL, A., 1925. Uber vitalfarbbare Erythrozytengranulationen (,, Substantia meta- chromatico-granularis ") beim Rochen, nebst weiteren Bemerkungen uber das Blut dieser Tiere. Zcitschr. f. mikr.-anat. 1'orsch.. 2: 461. JOLLY, J., 1903. Origine nucleaire des paranuclei des globules sanguins du Triton. Conif>t. rend. I'Assoc. d. Anat.. 5: 115. JORDAN, H. E., 1925. A Study of the Blood of the Leopard Frog, by the Method of Supravital Staining Combined with the Injection of India Ink into the Dorsal Lymph Sac, with Special Reference to the Genetic Relationships among Leucocytes. Am. Jour. Ajiat., 35: 105. KNOLL, W., 1931. Weitere Beitriige iiber die Entstehung des Hamoglobins im Erythroblastcnkern. Folia Haeni.. 44: 310. LEWIS, M. R., AND W. H. LEWIS, 1926. Transformation of Mononuclear Blood- cells into Macrophages, Epithelioid Cells and Giant Cells in Hanging- Drop Blood Cultures from Lower Vertebrates. Contribution to Embry- ology, Vol. 18, Pub. Carnegie lust., No. 363, pp. 95-120. MICHELS, N. A., 1931. The Erythrocyte. A critical review of its normal and pathological morphology and physiology with data on the normal red cell count and technic. Haematologica, Vol. 2. NIGRELLI, R. F., 1929. Atypical Erythrocytes and Erythroplastids in the Blood of Triturus viridescens. Anat. Rcc., 43: 257. NITTIS, S., 1930. A Surface Structure (?) in Normal Nucleated Erythrocytes. Anat. Rcc., 46: 365. SABRAZES, J., ET L. MURATET, 1900. Granulations mobiles des globules rouges de 1'hippocampe. Act. Soc. Linn. Bordeaux, 55: Ixv. SABRAZES, J., ET L. MURATET, 1900. Corpuscles mobiles endoglobulaires de 1'hippo- campe. Com fit. rend. Soc. Biol., 52: 365. SABRAZES, J., ET L. MURATET, 1900. Granulations mobiles dans les globules rouges de certains poissons. Coinpt. rend. Soc. Biol., 52: 415. SABRAZES, J., ET L. MURATET, 1902. Granulations endoglobulaires des globules rouges des anguilles jeunes. Act. Soc. Linn. Bordeaux, 57: cix. SABRAZES, J., ET L. MURATET, 1908. Observations sur le sang de la torpille (Tor- pedo marmorata Risso). Act. Soc. Linn. Bordeaux, 62: cxiii. SABRAZES, J., ET L. MURATET, 1908. Le sang de 1'axolotl. Granulations du cyto- plasme : origine nucleolaire. Folia Hacni., 6: 171. SCOTT, G. H., 1928. A Method for Making Permanent Preparations of Supra- vitally Stained Blood Cells. Anat. Rcc.. 38: 233. STOLZ, R., 1928. Le granulazioni basofile degli eritrociti nei vertebrati inferibri. Atti Soc. Ital. Sci. Nat. Mns. Civ. (Milano), 67: 93. STOLZ, R., 1928. Ematopoiesi normale e sperimentale nei pesci teleostei. Haema- tologica, 9: 419. URTUBEY, L., 1927. Sobre la fijacion del azul de metileno por el vacuoma en los eritrocitos de " Pleurodeles waltlii." Arch, dc Card, y Heniat., 8: 390. VILLA, L., 1930. Sull' origine dell' emoglobina. Arch. d. Fisiol., 28: 233. VON APATHY, ST., 1897. Protokollauszug der am 2. April 1897 abgehaltenen naturwissenschaftlichen Fachsitzung der medizinisch naturwissenschaft- lichen Sektion. Silzysl>cr. d. nicd.-natiiru'. Sektion des Siebenburg. Mu- seumsvereins. Jahrg. 22 II, naturw. Abt. WEIDENREICH, F., 1903. Die roten Blutkorperchen. I. Erg. der Anat. Enhi'ickl., 13: 1. YOFFEY, J. M., 1929. A Contribution to the Study of the Comparative Histology and Physiology of the Spleen, with Reference Chiefly to its Cellular Con- stituents. I. In fishes. Jour. Anat., 63: 314. 70 A. B. DAWSON EXPLANATION OF PLATES All drawings are from dry preparations and were outlined at the same magni- fication by means of a camera lucida. The details were filled in free-hand from sketches of the fresh cells. Erythrocytes stained with brilliant cresyl blue to demonstrate patterns of reticulation are shown with solid nuclei. All others, with the exception of Fig. 21, were stained either with neutral red and Janus green B or neutral red alone. The mitochondria are usually filamentous but some granular forms are present. Ordinarily they may be distinguished by their juxta-nuclear position. PLATE I Explanation of Figures 1. Erythrocytes of the toadfish showing, (a) induction of granules and reticu- lation; (/; and c) primary granules and mitochondria; (d) primary granules and reticulation. 2. Erythrocytes of the sea robin showing, (a and b) the primary granules and mitochondria; (c) induction of granules after twenty minutes ; (d) reticulation. 3. Erythrocytes of the sea bass showing, (a) primary granules and mito- chondria; (/>) induction of granules after twenty minutes; (c) reticulation in a mature cell; (d) reticulation in an immature cell which possesses a nucleolus. 4. Erythrocytes of the tautog showing, (a and b) primary granules and mito- chondria; (c) reticulation in an immature cell; (d) reticulation in a mature cell. 5. Erythrocytes of the eel showing, (a and d) primary granules and mito- chondria; (b) induction of granules after twenty minutes; (c) reticulation in a mature cell. 6. Erythrocytes of the cunner showing, (a and b) primary granules and mito- chondria; (c and d) reticulation in mature cells. 7. Erythrocytes of the scup showing, (a and c) primary granules and mito- chondria; (b) reticulation in a mature cell; (d) reticulation in an immature cell. 8. Erythrocytes of the butterfish showing, (a) reticulation in a mature cell; (b) primary granules and mitochondria; (c and d) reticulation in immature cells. 9. Erythrocytes of the sand dab showing, (a) primary granules and mito- chondria; (b) granules and reticulation; (c and d) induction of granules after twenty minutes. 10. Erythrocytes of the menhaden showing, (a) perinuclear primary granules and mitochondria; (b) granules after twenty minutes; (c) reticulation in a mature cell; (d) reticulation in an immature cell. 11. Erythrocytes of Cypritiodon showing, (a and b) primary granules and granular mitochondria; (c and d) reticulation in mature cells. 12. Erythrocytes of Fniidulns majalis showing, (a) induction of granules after twenty minutes; (b and c) primary granules and mitochondria; (d) reticulation in a mature cell. 13. Erythrocytes of the mackerel showing, (a) reticulation in an immature cell; (b) perinuclear primary granules and mitochondria; (c) induction of gran- ules after twenty minutes ; (d) reticulation in a mature cell. 14. Erythrocytes of the alewife showing, (a) induction of granules after twenty minutes; (b) reticulation in a mature cell; (c) primary granules and mitochondria; (d) reticulation in an immature cell. REACTION OF ERYTHROCYTES TO VITAL DYES 71 /" • * 6 - ••'• 3 x- J ' ,' V < ^' 8 • '•;• : ) -i "~" •\V» «N <-\ :•• < ... 11 b 13 10 r ? :•'••- c « « • • *L, 01- 13 14 A. B. DAWSON PLATE II Explanation of Figures 15. Erythrocytes of the smooth dogfish showing, (a) the primary granules and mitochondria; (£>) reticulation in a mature cell; (r) reticulation in an imma- ture cell. 16. An erythrocyte of Tritnnis virldcsccns showing reticulation of a mature cell. Note the complete absence of granules. 17. Erythrocytes of the spotted skate showing, (a, b, and c ) primary granules and mitochondria ; (d) reticulation in a mature cell. 18. Erythrocytes of the domestic fowl showing, (a, r, and d) reticulation of mature cells; (/') primary granules and mitochondria. 19. Erythrocytes of the fence lizard showing, (a) primary granules and mito- chondria; (/>) induction of granules after twenty minutes; (c and d) reticulation and nucleoli in mature cells. 20. Erythrocytes of the horned toad showing, («) primary granule and mito- chondria; (b and c) reticulation and nucleoli in mature cells. 21. An erythrocyte of Aniphiuma means from a smear stained by Wright's method, showing the primary granules as clusters of basophilic bodies. 22. Erythrocytes of the painted turtle showing, (a) the primary granules and mitochondria; (b and c) reticulation and nucleoli in mature cells. REACTION OF ERYTHROCYTES TO VITAL DYES 73 '?/• •'.' 17 16 21 CYCLOTRICHIUM MEUNIERI SP. NOV. (PROTOZOA, CILIATA) ; CAUSE OF RED WATER IN THE GULF OF MAINE PHILIP B. A. POWERS ZOOLOGICAL LABORATORY, UNIVERSITY OF PENNSYLVANIA I. INTRODUCTION During the warmest days of August, 1931, there were noticed in Frenchman Bay x on several occasions, great patches and lanes of red water. This has heen known to occur in previous years but rarely so early as the fifth of August. These patches of red water were caused by swarms of small red animals, which appeared in layers from one to three meters beneath the water level. In spots these animals would be more crowded, thus varying the density of the patch as a whole and making the color vary from a brick to a blood red. Between the action of the wind and tide these patches would be streaked out in great lanes and irregular areas. On some days these patches would be more numerous than on others. As to the origin and ultimate fate of these areas of red water, the writer has no information. A microscopic examination of a drop of sea water containing these organisms revealed hundreds of small red animals which moved rapidly in a zig-zag fashion reminding one of the characteristic movements of Haltcria. On preliminary examination their appearance suggested that of trochophore larva1 ; however, their incredible numbers and small size suggested that they were protozoa. Since the living animals died and disintegrated within one or two minutes upon exposure in a drop of water on a glass slip, no definite conclusion could be reached as to their morphology from a study of the living material. Further study indicated that these animals were swarms of a ciliate belonging to the genus C\dotrichium Meunier. I have named it Cyclo- trichiuui nicuiiicri sp. nov. The writer wishes to express his appreciation to Mr. William Procter of Bar Harbor, Me., in whose laboratory this study was initiated; and to Dr. D. H. Wenrich for his many helpful suggestions and criticism of the manuscript. 1 Frenchman Bay is that part of the Gulf of Maine which separates the north- west side of Mt. Desert Island from the mainland. 74 CYCLOTRICHIUM MEUNIERI SP. NOV. 75 II. TECHNIQUE Quart jars of this red water were collected and allowed to stand in a cold place for about 20 minutes, in which time the organisms settled to the bottom. With a long pipette this sediment was collected and spurted into flasks half filled with warm (40° C.) fixative. Both Schaudinn's and Bouin's fixatives were used. After 30 minutes' fixa- tion the animals were removed from the fixative with the aid of a centri- fuge, washed and stored in 70 per cent alcohol. At the close of the summer, vials containing this fixed material were brought back to the University, where the slides were made and the study completed. By mixing, on a cover-glass, a drop of this alcoholic sediment of fixed organisms with an equal amount of Mayer's egg albumen, spread- ing it carefully so that it would dry slightly, and finally flooding the cover-glass with absolute alcohol, these minute organisms were fastened to the cover-glass to facilitate their easy manipulation during staining. After 15 minutes in absolute alcohol the cover-slips were gradually transferred to water. Part of the material was stained with iron hsematoxylin, Mayer's hiemalum, or Delafield's haematoxylin ; and part by Feuigen's nucleal reaction. It was found necessary to section some of this material, and in order to handle these small animals in paraffin a mixture of the alcoholic sediment and Mayer's egg albumen was placed in a concavity cut in a small block of preserved (70 per cent alcohol) liver, coagulated with absolute alcohol ; and this block, with embedded protozoa, was then handled as a piece of tissue. By infiltrating in 67° paraffin and treating the block with ice water before cutting, sections of 4 ^ could be cut with ease. The same stains were used for the sectioned material as for the whole mounts, with the exception that counterstains of orange G or cosine were used to demonstrate the cilia. III. OBSERVATIONS Cyclotrlchlnni nicunicri (Fig. 1), is almost oval, with the anterior end blunted and the posterior region slightly conoid. Size. — This ciliate is relatively small. Twenty-five specimens se- lected at random from fixed material averaged 33 p, (25-42 ^) in length and 22 fj. (18-34^) in width through the anterior region, whose di- ameter is somewhat greater than that of the posterior portion. Cytostoinc. — The cytostome could not be definitely located ; however, most of the specimens show, at the larger end, a depression (Figs. 1 and 2) slightly funnel-shaped and leading into the interior without any well-marked structures. This location of the cytostome agrees in gen- eral with that described by Meunier (1910) for C. cyclokaryon; and 76 PHILIP B. A. POWERS with descriptions given of the cytostomal regions of a number of species of Cyclotrickiuvn described by Faure-Fremiet (1924). Ciliary Band. — A broad band of cilia is found about the middle. The. fine, closely-set cilia are from 5 to 6 /x long and are found only in the depressed mid-region or ciliary band. Besides being depressed and ciliated, this band is further indicated by 52-60 striations which run parallel to the longitudinal axis. Each striation is made up of 5 to 9 granules; anteriorly the granules form a definite ring (Figs. 1 and 3, A. G. R.}, while posteriorly they become slightly larger (Fig. 1 ). With iron haematoxyiin these granules stain black, as do the metaplasmic granules of the endoplasm. With Delafield's haematoxyiin or Mayer's hsemalum the metaplasmic granules fail to stain while the granules of the ciliary band are definitely outlined, thus indicating that the two types of granules are not the same. Because of the close association EXPLANATION OF PLATE Cyclotrichium mcnnicri sp. nov. All drawings have been made at a magnifica- tion of 2900 diameters and reduced about two-fifths in printing. The animals have been fixed with Bouin's fluid, stained in Heidenhain's hsematoxylin, with the excep- tion of those shown in Figs. 7 and 8 which were treated for the Feulgen reaction. Abbreviations A.G.R., anterior granular ring E.G., basal granules C., cilia C.B., ciliary band C.Bu., ciliary bundles C.C., compound cilia Ch., chromatin C.P., chromatophore platelets Cy., cytostome E., endoplasm with inclusions E.V., extra-nuclear vesicle L.S., longitudinal striations M., macronucleus ;;;.. micronucleus Me., metaplastids P.. pellicle P.B., pyrenoid body PL, plastin S.S., shrinkage space or artifact Explanation of Figures 1. Side view of Cyclotrichium mcnnicri. Chromatophore platelets and pyre- noid bodies do not stain in the whole mounts. 2. Sagittal section showing arrangement of chromatophore platelets, pyrenoid bodies, etc. 3. Sagittal section through region of ciliary band showing the arrangement of the basal granules. 4. A diagram of the possible arrangement of cilia in compound units to explain structures as shown in Fig. 6. 5. Cross section through anterior region. 6. Cross section through region of ciliary band. Cilia are finer and more nu- merous than could be shown in drawing. 7. Early telophase in dividing individual ; six chromosomes are seen in each daughter. 8. Nuclear complexes showing variations, (a) micronucleus and two extra- nuclear vesicles ; ( /> ) micro- and macronucleus and two extra-nuclear vesicles ; (c) a stage somewhat later than that shown in Fig. 7. Macronucleus is seen in only one of the daughter cells. CYCLOTRICHIUM MEUNIERI SP. NOV. PLATE I 77 AGR.- ' i" - ""•:,'( LS L ca A /•:"?"»"; T-r •-•-.; Pl^:^J:?^ ^JJrt:::;^ -CC -- \ -B.G -£ -PB 4. -cp 7 ,' :. % C. P PB; -PB. 78 PHILIP B. A. POWERS of these granules of the striations with the cilia, I have designated them as basal granules. The uneven movement of the living animal suggests the presence of cirri or membranelles ; however, since the sectioned material showed only these exceedingly fine and numerous cilia and their large associated (compound?) basal granules (Fig. 3, B. G.}, it would seem that we have in the living condition a system of compound cilia which at death separate into their component units (Fig. 6, C.). A diagrammatic in- terpretation of these compound cilia is shown in Fig. 4. Further evi- dence for this arrangement is presented by the staining reaction. With iron hsematoxylin the points of insertion of the cilia stain after the manner of a clump of fibrils (Fig. 6, C., Bit.}, suggesting definite bunches or bundles of cilia. Ectoplasm. — Beneath a thin, smooth pellicle, completely shielding the endoplasm, is a series of irregularly concave chromatophore plate- lets (Fig. 2, C. P.), each with an associated pyrenoid body (Fig. 2, P. B.} ; the whole being inclosed by a large vacuole. The red color of C. iiicunieri is doubtless due to the presence of a haematochromatous substance localized in the chromatophore platelets, which are possibly of an amylaceous nature. These platelets stain deeply with iron haema- toxylin but not at all with Mayer's haemalum. That these platelets have a definite body is demonstrated by the fact that they are often torn from their place due to the impact of the knife during sectioning. In the region of the ciliary band the platelets are missing while the pyrenoid bodies with their vacuoles remain (Fig. 2). Endoplasm. — After staining sections with iron haematoxylin the endoplasm is crowded with many darkly staining bodies. These are designated metaplasmic granules, for they are doubtless associated with the metabolic processes. Because of the abundance of these meta- plasmic granules, the nuclear apparatus could lie demonstrated success- fully only by the aid of the Feulgen technique. A typical ciliate nuclear complex is present. The macronucleus is slightly irregular and demonstrates a definite core of plastin (Fig. Sb, M.} surrounded by a layer of granular chromatin. The micronucleus is small, sometimes vesiculated (Fig. Sb, ;;/). Besides the macro- and micronucleus there are found one or more bodies, irregular in shape and staining but slightly with Feulgen's reagent, which are designated as extra-nuclear vesicles (Fig. 8, E. V.}. These extra-nuclear vesicles are always present and seem to be either the formative or degenerative stage in the development of the macronucleus. A number of dividing individuals were studied, and in one (Fig. 7) showing an early telophase stage in the division of the micronucleus six CYCLOTRICHIUM MEUNIERI SP. NOV. chromosomes could be counted in each daughter nucleus. A somewhat later stage than this is shown in Fig. Sc ; in this case only one daughter individual received a definitive macronucleus, the other had but the extra- nuclear vesicles. Occurrence. — Since single individuals of C. inciinicri have been found in sea water taken either from the storage tanks of the laboratory, or among the material from plankton hauls ; it seems reasonable to look upon this species as a member of the protozoan fauna of this region. Its sudden appearance in swarms among the surface plankton of the bay must be correlated with the periodic enrichment of the water by the nitrogen-bearing algae whose numbers increase during periods of warmth and excessive sunshine ; all of which factors tend to make areas of the bay excellent culture media for these red water ciliates. In the present instance th^summer had been warmer than usual, par- ticularly during the last of ^fcjy and the first week of August. IV. DISCUSSION Many organisms are known, under favorable conditions, to multiply in such numbers as to discolor great bodies of water. Martin and Nel- son (1929) review this subject and give instances of red water occurring in Delaware Bay due to the swarming of Amphidinium fusiforme. Kofoid and Swezy (1921) record the occurrence of swarms of Gony- aitla.v polyhcdra as being the most frequent cause of red water along the Pacific coast. I have placed the ciliate causing red water in the Gulf of Maine in the genus Cyclotrichium Meunier because it seems to resemble very closely a ciliate of rare occurrence in the plankton hauls from Barents Sea. Meunier (1910) established this genus for C. cyclokaryon but figured a number of ciliates as Cyclotrichium Sp? because their poor preservation would not permit further classification. Those ciliates were included in the genus Cyclotrichium which had a ciliary band or belt in an equatorial depression which divided the body into two sub- spherical halves. Making allowance for the poor fixation of his mate- rial, the organisms which he described under the name of Cyclotrichium Sp? may be likened to those found off the coast of Maine which I have named C. incunicri. The Maine fishermen recognize this red water as a source of food for the herring sardine and it is said that when the red water is present in the herring's intestines they become unfit for sale. Mackerel are also reported by the fishermen as sometimes containing " red feed " in their digestive tracts. It would seem that " red water " and " red feed " are due to two different organisms. Inquiring into this condition fur- 80 PHILIP B. A. POWERS ther, the writer corresponded with Dr. A. G. Huntsman, Director of the Atlantic Biological Station at St. Andrews, mentioning these occur- rences. Dr. Huntsman replied as follows : ' In the St. Andrews region we have never seen the water coloured crimson by the form you mention. In warm summers the warmest strip near the center of the upper end of Passamaquoddy Bay sometimes be- comes decidedly reddish and this has been found due to large numbers of different kinds of Tintinnoids which have been studied by Professor J. N. Gowanlock, although his report has not yet been published. In the St. Andrews region the term ' red feed ' is given to copepods as they occur in the stomach of herring, particularly the young herring or sar- dine. The most abundant form is Calanus and at times the swarms of this species give a reddish cast to small areas of the water." To my knowledge Cydotrichium meuiueri is the first holotrichous ciliate to be associated with the appearance of red water in the ocean. V. SUMMARY 1. Red water in Frenchman Bay is caused by the swarming of a small red ciliate, Cydotrichium ineunieri sp. nov. 2. This organism is ovoid, about 33 /x long, and has in a wide de- pression about its middle a band of fine cilia. It has been suggested that compound cilia may be present in the living animals. The endo- plasm is well shielded by a peripheral series of chromatophore platelets, each with an associated pyrenoid bod}-. 3. Six chromosomes were observed in each of the daughter micro- nuclei in an early telophase of division. LITERATURE CITED FAURE-FREMIET, E., 1924. Contribution a la Connaissance des Infusoires Plank- toniques. Paris. KOFOID, C. A., AND OLIVE SWEZY, 1921. The Free-Living Unarmored Dinoflagel- lata. Memoirs of the University of California, Volume 5, p. 44. KUDO, R. R., 1931. Handbook of Protozoology. Chas. C. Thomas, Pub. MARTIN, G. W., AND THURLOW C. NELSON, 1929. Swarming of Dinoflagellates in Delaware Bay, New Jersey. Bot. Gaz., 88 (2) : 218-224. MEUNIER, A., 1910. Microplankton des mers de Barents et de Kara. Due D'Orleans. Campagne Arctique de 1907. Vol. I, p. 168. Vol. II, PI. xvii, Fig. 7; PI. xx, Fig. 15, 16, 17; PI. xxiii, Fig. 9. THE NUMBER OF PRE-ADULT INSTARS, GROWTH, RELATIVE GROWTH, AND VARIATION IN DAPHNIA MAGNA BERTIL GOTTFRID ANDERSON 1 BIOLOGICAL LABORATORY, WESTERN RESERVE UNIVERSITY Until recently growth and variation in Crustacea have been studied by means of preserved materials. These materials may consist of a single collection taken in a certain locality at a specific time or several collections made in different localities at different times. Often conclu- sions are drawn from an isolated collection. Such studies necessarily neglect the past history of the individual. Any conclusions reached by such methods regarding growth and variation can only be regarded as tentative and subject to verification by experiments with individually reared animals. As a result of such studies on Coronis, Brooks (1886) was led to the statement: "' . . . the length of the larva increases uniformly at each moult by one fourth of its length before the moult." Later Fowler (1909) designated this numerical relation " Brooks' Law." He also called the fixed fractional increase the " growth factor." That " Brooks' Law " does not hold for Cladocera has been shown by Rammner (1930) after a series of studies on individually reared animals from several genera. Gurney (1929) doubts that the above relations exist for copepods. Other arthropods as well do not follow " Brooks' Law." Calvert (1929) found that larval Odonata grow quite irregularly. Such studies with Cladocera have perhaps led to the conclusion that individuals of a species pass through a definite number of pre-adult instars. When a graph of a population is made wherein the number of individuals of a size class is plotted against size, i.e., total length, a series of size modes is secured (Fig. 1). These size modes are taken as repre- sentative of the growth stages of the organism. Usually females of a given species which are of a specific size or larger bear eggs in their brood chambers. The number of size modes between the embryonic stages and the mode of the smallest egg-bearing females is taken as the number of pre-adult instars for the species. 1 A part of the experimental work related in this paper was done at the Zoolog- ical Laboratory of the State University of Iowa. 6 81 82 B. G. ANDERSON Inspection of graphs of size distributions such as in Fig. 1 shows that the size groups represented hy the modes are not entirely distinct. The size groups in the lower size ranges are usually quite distinct but not decidedly disjoined from each other. Those in the higher size ranges 420 FIG. 1. Distribution of a population of Chydonis sphccrlcus according to length (after Werner). tend to run together and are not so easily identified. Since the size groups are not entirely distinct over the pre-adult range and the past history of each individual is not known, the conclusion that all indi- FIG. 2. Diagram showing methods of making measurements. T — total length, longest dimension of animal exclusive of spine. C — carapace length, longest di- mension of carapace exclusive of spine. H — height, the shortest distance between two lines tangent to the carapace, as illustrated, and parallel to the line of T. This measure of height is affected but little by the number of young in the brood chamber. vicluals of a species have a definite number of pre-adult instars is not necessarily valid. The writer finds that the number of pre-adult instars for Daphnia niagna is rather variable. Calvert (1929) reported that the number of larval instars for certain species of Odonata is not constant. Dover (1931) reported that the number of moults in Orgyia turbata decreases GROWTH AND VARIATION IN DAPHNIA MAGNA 83 when fed on Crotolaria. Singh-Pruthi (1925, etc.) found that the number of moults for mealworms can be varied. Schubert (1929) ex- amined Ceriodaphnia rcticnlata from two ponds. The material from the one pond was collected early in July, 1926. He believed that all C. rcticnlata in this collection were primiparous in the fourth instar. The material from the second pond was taken in mid-September of the same year. His findings regarding the latter were that 60 per cent of the individuals were primiparous in the third instar and that 16 per cent were still immature in the fourth. He concluded that at least two varieties of C. rcticnlata exist in these ponds. Apparently the num- ber of pre-adult instars is variable in this species. Agar (1930) re- ports that Siiiioccphalus gibbosns and Daplinia carinata have three and four pre-adult instars respectively but adds that a small percentage of the individuals become mature in one less than the average. The aim of this paper is to present data on growth of individually reared Daplmia inagna and further to consider the application of the equation y - - bx* to relative growth and variation in Cladocera. MATERIALS AND METHODS Female DapJuria inayna Straus, of several clones were employed. One clone was used for the major portion of the work. To determine whether or not the observed results were characteristic only of the one clone, some six others were tested. One of the latter was secured from Dr. A. M. Banta. All of the other clones, including the one first men- tioned, were derived from ephippial eggs which may be traced to Banta's stocks. Individual females were isolated within six hours of their release from the brood chambers of the mothers. These were measured and placed in separate vials containing thirty to thirty-five cubic centimeters of a culture medium. Banta's manure-soil medium (Banta, 1921) and oatmeal and wheat modifications were employed. More uniform re- sults were obtained from the regular manure-soil medium than from modifications, as the results of one series of experiments indicate. Where a modified culture medium was used, the animals were placed in old manure-soil medium and a drop of an oatmeal mixture or wheat infusion was added daily. Water was added to replace the fluid lost by evaporation. The oatmeal mixture was prepared by cooking rolled oats for a half hour in about twice the amount of water ordinarily used in making a porridge. The mixture was then strained through gauze and kept in a refrigerator until used. The wheat infusion was prepared by 84 B. G. ANDERSON TABLE I Mean total lengths of female Daphnia magna of one clone during each of the pre-adult instars and the first adult for different classes of individuals. The classes are based on the number of pre-adult instars and the nature of the culture medium. MEDIUM MANURE- SOIL OATMEAL MODIFICATION MEAN TOTAL LENGTH IN MM. Number of Pre-adult Instars 5 5 6 7 8 Number of Cases 19 8 29 39 5 First Instar 0.85 ±0.01 0.80 ±0.01 0.84 ±0.01 0.83 ±0.01 0.77 ±0.02 Second Instar 1.09 ±0.02 1.02 ±0.01 1.00 ±0.01 0.94 ±0.01 0.86 ±0.02 Third Instar 1.39 ±0.02 1.33 ±0.03 1.17 ±0.01 1.05 ±0.01 0.99 ± 0.03 Fourth Instar 1.77 ±0.02 1.70 ±0.03 1.36 ±0.02 1.19 ±0.01 1.12 ±0.04 Fifth Instar 2.13 ±0.02 2.06 ± 0.04 1.64 ±0.02 1.38 ±0.01 1.22 ±0.05 Sixth Instar Adult 2.60 ±0.02 Adult 2.56 ± 0.03 1.99 ±0.02 1.69 ±0.01 1.55 ±0.10 Seventh Instar Adult 2.49 ± 0.02 2. 08 ±0.03 1.87 ±0.09 Eighth Instar Adult 2.56 ±0.03 2.20 ±0.04 Ninth Instar Adult 2.52 ±0.03 Number of Young in First Brood 6.4 ± 0.3 6.4 ±0.6 7.3 ±0.2 8.6 ±0.3 5.8 ±0.8 Number of Young in Second Brood 9.5 ± 0.5 9.4 ±0.6 7.4 ±0.3 7.1 ±0.2 5.2 ± 1.0 GROWTH AND VARIATION IN DAPHNIA MAGNA boiling wheat in water, decanting the fluid, and allowing this to stand open for several days before using. When the manure-soil medium was used exclusively, one-third of the fluid was removed semiweekly from each vial and replaced by fresh medium. At the time of isolation of the individuals and after every moult each individual was placed in a watch glass together with a few drops of the culture medium. Just enough of a saturated solution of chloretone was added to bring about cessation of movement. The chloretone did not appear to have any detrimental effects. By means of an ocular microm- eter the measurements as illustrated in Figure 2 were made. The three measurements consisted of total length exclusive of the spine, carapace length, and height. Camera lucida outline drawings were made during each instar for each of three animals. In addition to these measure- ments, note was taken as to the presence of eggs in the brood chamber. Within twenty-four hours after their release the young were removed from the vials and counted. Size and shape change only at the time of moulting in this species. Repeated measurements of any dimension during a single instar always gave the same results. Agar (1930) pointed out this fact for Daphnia carinata and Simocephalus gibbosus. Such being the case, a single set of measurements suffice for each animal during any one instar. All the experiments were carried out at room temperature (18°- 23° C). NUMBER OF PRK- ADULT INSTARS AND GROWTH The results of observations on the number of pre-adult instars, i.e., the number of instars elapsing between the time of release of the indi- vidual female from the brood chamber of her mother and the appearance of eggs in its own brood chamber, and the mean total length of the ani- mals during each are summarized in Table I. The data were segregated on the basis of the number of pre-adult instars and the culture medium used. The number of pre-adult instars varied for the animals reared in the oatmeal modification. For those reared on the standard manure- soil medium the number was constant. Twenty individuals only were used in the latter series, one of which died before reaching maturity. The number of pre-adult instars above five may perhaps be explained on the basis of food deficiency. Great numbers of large ciliates were found in the modified medium. These probably reduced the number of smaller microorganisms that ordinarily would have been available as food to the Daphnia. The ciliates themselves were too large to be con- sumed. Conditions similar to this arose at a time when great numbers of the annelid Acolosouia were found in the culture medium. At times 86 B. G. ANDERSON a culture medium seems to stimulate the adult females to produce great numbers of young, but this same lot of medium does not seem sufficient for the development of the young. ffl/TJ. I I I I I I I I I I I I I I I I ' I I I I ' I \ ' ' I I I ' ' ' I I I I I I I I I I I I I I I I tnsfar /4 /6 I I I I I I l I I I I I I I /O /2 FIG. 3. Growth curves of different classes of individuals reared on the oat- meal modification; A — those animals with five pre-adult instars (7 individuals), B — six (18 individuals), C — seven (30 individuals), D— eight (5 individuals). T — total length ; C — carapace length ; H — height. The arrow indicates the first adult instar. The animals used in the experiment with the modified medium were all released on the same day. The time taken to reach sexual maturity varied from six to ten days. The number of days corresponded approx- imately to the number of pre-adult instars in each case. The animals were all kept under identical conditions and the culture medium was from the same lot. The experiment using the standard manure-soil medium was begun several weeks later. GROWTH AXD VARIATION IN DAPHNIA MAGNA 87 Table I brings out an interesting relation. The mean size for each group during the instar when eggs first appear in the brood chamber varies from 2.49 mm. to 2.60 millimeters. Apparently Daphnia magna must attain a certain size before becoming sexually mature. Singh- TII /nsTar 6 /O /4 Fin. 4. Growth curves of a group of animals reared on unmodified manure- soil medium. These had five pre-adult instars (7 individuals). T — total length; C — carapace length ; H — height. The arrow indicates the first adult instar. Pruthi (1925, etc.) found that metamorphosis of mealworms took place only after the larvae were full-grown. Rammner (1930^) has found that eggs were produced after the fourth instar in Daplinia inagna. After observation on well over a thousand individually reared females, the writer has never observed less than five. L. A. Brown has observed five pre-adult instars for this TABLE II Number of Individuals Primiparous During the Sixth to Ninth Instar Clone Sixth Instar Seventh Instar Eighth Instar Ninth Instar oo B. . 4 10 4 9 0 C 8 3 1 0 1 D 7 4 0 0 0 E. . 9 1 1 0 0 F. . 6 11 7 0 0 G * 10 4 4 0 1 =c Those living beyond the 10th instar without bearing eggs. * Clone secured from Dr. A. M. Banta. 88 B. G. ANDERSON species (unpublished data). Rammner's smaller number may be due to some factor in the culture medium or perhaps to a genetic difference in his animals. Figures 3 and 4 are growth curves constructed from data secured from animals included in Table I. The data used were taken only 3 2 3 I I I I I I I I I I I I I I I I I I I I I I I i i i i T i i i i i i i i i i i i i i i i 24 S 6 /O /2 /4 /nsfar 6 8 /O /2 14 16 FIG. 5. Growth curves of three broodmates of clone B. Animal A had five pre-adult instars; B, six; and C, seven. T — total length; C — carapace length; H— height. Arrow indicates first adult instar. from those animals which were still living during the last instar recorded on the graphs. All such are included. The inflection in the curves always occurs at a point which corresponds to the time of sexual maturity. All the individuals used in the above experiments were of the same GROWTH AND VARIATION IX DAPHXIA MAGXA 89 clone. To determine whether or not the variation in the number of pre-adult instars was characteristic only of this clone, another series of experiments was performed using six other clones. Five of these were raised from ephippial eggs. A sixth clone was secured from Dr. A. M. Banta. Manure-soil medium and the wheat modification were em- ployed in these tests. Table II gives the results. Examination of this mm. VJ7T FIG. 6. Outline drawings of a single animal during each of the first thirteen instars. The animal is the one for which growth curves are shown in Figure 5 A. The arabic numerals designate pre-adult instars ; the Roman numerals — adult instars. table shows that the variability in the number of pre-adult instars is characteristic of all clones used. Variation occurred to about the same extent in both media. Two of the animals included in the above experiments lived for at least thirteen instars without bearing young. Ordinarily the ovary in the adult assumes a green color, especially toward the end of the instar. Microscopic examination did at no time reveal the development of the green color in the ovaries of either animal. In one of them growth did not seem retarded. Her total length during the thirteenth instar equaled 90 B. G. ANDERSON the average total length, during the same instar, of those primiparous in the sixth. Individual growth curves for three brood mates of clone B are shown in Figure 5. Each of these as designated A, B, and C, bore its first clutch of eggs in the sixth, seventh, and eighth instars respectively. These were raised under the same conditions and at the same time, using the same lot of unmodified manure-soil medium. The measure- TABLE III Growth ratios for various classes of animals according to the nature of the culture medium, number of pre-adult instars, and clone CLj SiSS* M6 M7 M8 M9 R6 B6 B7 B8 1-2 1.27 1.16 1.13 1.11 1.26 1.24 1.19 1.25 2-3 1.29 1.17 1.10 1.15 1.29 1.28 1.21 1.23 3- 4 1.27 1.14 1.17 1.14 1.27 1.19 1.23 1.24 4-5 1.25 1.21 1.14 1.09 1.22 1.23 1.23 1.21 5-6 ...... 1.24f 1.21 1.22 1.27 1.22f 1.19f 1.18 1.07 6-7 1.10 1.26t 1.24 1.21 1.11 1.02 .15f 1.00 7-8 1 07 1.09 1.23f 1.18 1.07 1.10 .04 l.OSf 8-9 . .... 1 05 1 06 1.08 i.ist 1.05 1.11 .07 1.05 9-10 1.04 1.07 1.06 1.05 1.05 1.07 .02 1.15 10-11 1 05 1.04 1 05 1.02 1.05 09 1.12 11-12. 1.03 1.04 1.05 1.02 1.01 1.07 1.03 12-13 .... 1.03 1.05 1.08 1.02 1.02 1.01 13-14 1.03 Number of Animals 7 18 30 5 7 1 1 1 * Classes — M6, M7, M8, and M9 from clone A reared on modified medium and primiparous in the sixth, seventh, eighth, and ninth instars respectively. R6 from clone A and reared on the regular manure-soil medium, primiparous in the sixth instar. B6, B7, and B8 from clone B, primiparous in the sixth, seventh, and eighth instars respectively. This table is based on the same data as are Figs. 3, 4, and 5. The ratios are based on total length. t Denotes instar during which animals were primiparous. merits for these were made by means of camera-lucida drawings. Fig- ure 6 is a reproduction of the series of camera-lucida drawings for animal A. The irregularities in the number of pre-adult instars and in the growth of Daphnia inacjna as brought out above indicate that no law of growth such as " Brooks' Law " is valid under all circumstances. Ta- ble III brings the case out more clearly. This table gives the values of the ratio of the total length during each instar for each group to that GROWTH AND VARIATION IN DAPHNIA MAGNA 91 of the previous instar. These values correspond to the " Wachstums- quotienten " of Rammner (1930) and others. For those cases in which the individuals were primiparous in the sixth instar the value is fairly 3 4 FIG. 7. Lo// plots of the relations between carapace length and total length C/T, height and total length H/T, and height and carapace length H/C during each instar for different classes of animals reared on the oatmeal modification — the same for which growth curves are shown in Figure 3. The lines were drawn ac- cording to the calculated values of the constants given in Table IV. The breaks in the relations are coincident with sexual maturity. A — animals with five pre- adult instars (7 individuals), B — six (18 individuals), C — seven (30 individuals), and D — eight (5 individuals). constant during the pre-adult stages. In all other instances the value is quite variable. After sexual maturity the value decreases consider- ably and approaches unity in old age. The validity of " Brooks' Law >: with regard to Daplniia magna may therefore be considered to depend on the conditions of the individual and of the environment. 92 B. G. ANDERSON RELATIVE GROWTH AND VARIATION The foregoing portion of this paper has brought out the wide irregu- larities in the general growth rate of individual Daphnia inogna. In spite of these irregularities the relative growth of parts is quite constant. If the logarithm of the carapace length be plotted against the logarithm of the total length for each, two straight lines may be drawn — one through the points for the pre-adult instars and the other for the adult instars. The lines so drawn differ slightly in slope and in position (Figs. 7, 8. and 9). Such log log plots of height against total length mm 4 2 /.S W .8 .6 .5 .4 C .5 .6 .8 /.O mm. /.S 3 FIG. 8. Ltni /<)— 3 ~; r/2 a J bi i 5 r^. CO O "^ ^'^ T— i ^H ^ — i •^ ro CARAPACE LENGTH (x) HEIGHT (y) 4J "3 T3 < Q t^-^J — OOrOOOI^- r^ t>* t — o f^- ^^ t~^ ^~ OOOOOOOO M -tr^roOCO-f1^"^ O^H^HI~~JOOOO •c rt OJ (H (X £> ^H r\i r^i *— i O "5 ^O "^ OO OO OO OO CO t—_ t—_ 1^; OOOOOOC3O M ONOOOOsOt^Or^r<0 o o p p p p p p TOTAL LENGTH (x) HEIGHT (y) 3 •o < £ l~- ui O '-i 10 -f ^1 O OOOOOOOO OOOOOOOO ^ i-^t^irj-^LOOO'O O p — ; "-; p p ~ p 4J •o cj! O PH j3 r^rviO-lfMO'^"^"-^ OOOOOOOO OOOOOOOO ^ rsju->O-tr^.oOooO~H(^ir^ oopppppp X VI V) _rt 0 ISISss^s ' cn ns >o c b -^ i i C W !> 3 •" V ^ .2 tn i- cn "O c ^ o3 9J cn "C "C Q, n a a In o 94 B. G. ANDERSON An interesting feature coming out of this study is that relative growth changes at sexual maturity. This may be seen on examination of any one of the log log graphs. The change is much more distinct in clone A (Figs. 7 and 8). The change in clone B is less discernible .4 FIG. 9. Log log plots of the relations between carapace length and total length C/T, height and total length H/T, and height and carapace length H/C during each instar for three individual broodmates of clone B — the same for which growth curves are shown in Figure 5. The lines were drawn according to the calculated values of the constants given in Table IV. The breaks in the relations are coinci- dent with sexual maturity. Animal A had five pre-adult instars, B six, and C seven. (Fig. 9). Table IV gives the values of the constants b and k for both clones. These were computed by the method of averages. Since all points on any one graph represent the same number of cases, the cal- GROWTH AND VARIATION IN DAPHNIA MAGNA culated values of the constants are practically as accurate as those se- cured by a more elaborate method. Huxley (1927) has suggested that heterogony in Main begins at sexual maturity. He also noted that this is probably true for male Gaininanis. Examination of Figure 1 in the same report shows a change in relative growth of the large chela as against the rest of the body in Uca pugna.v. Huxley makes no mention as to its significance. Robb (1929) has pointed out that a change takes place in the relative growth of various organs and suggests that another factor becomes in- volved. These changes are apparently coincident with sexual maturity. Since the value of k approximates unity for all above relations, both for pre-adult and adult instars of Daphnia luagna, and the value of b for each relation varies only over a small range, the animal does not change in proportions, to any great extent, during growth. Wesenberg- Lund (1926) writes that variations in D. inagna of Denmark are insig- nificant. This condition is therefore as expected. In a great many Cladocera pronounced variations occur. Rammner (1927) has discussed several methods for their study. As far as the writer is aware, no one has considered applying the equation y-= in such studies. Masses of data on variation have been collected by various workers on several species (Werner, 1924; Rammner, 1926; Heberer, 1928). The data as presented by these workers do not allow of the proper manipulation for testing the validity of the equation for the purpose. Were the original data available, a satisfactory test might be made. Rammner's (1928; data on individually reared Chydonis sphccricus and Pleuroxus trigonellns seem quite irregular. Woltereck's (1925) data on individually reared Daphnia cucullata seem satisfactory for the pre- adult instars. In log log plots of the head length against the body length a break in the relations appears in about the middle of the pre- adult instar range and another at sexual maturity. The first change seems likely to affect only the value of the constant b in the equation while k remains approximately the same. The second change would probably affect the values of both constants. The values of the con- stants for the adult instars cannot be determined, since the data includes measurements for only one and sometimes two adult instars. Different varieties give different values for the constants. The general trend of the relations is the same. 96 B. G. ANDERSON The advantage of applying the equation y = b.v7" to data on variation in Cladocera is that the direction of variation may be expressed numerically in the constants b and k. The data employed could be taken from preserved material such as that used by Werner and others. Huxley and his coworkers have used this method in the study of heterogeny in Crustacea with apparent success. Collections from several localities, as well as those made in the same locality at different seasons, might thus be readily analysed and compared. The author wishes to express his appreciation to Drs. J. H. Bodine of the State University of Iowa, L. A. Brown of George Washington University, and A. H. Hersh of Western Reserve University for their many helpful suggestions and criticisms, and especially to Dr. Hersh for suggesting the method of treatment of relative growth. SUMMARY Observations as to the number of pre-adult instars have been made on over 200 individually reared female Daphnia niagna of seven clones. In all cases measurements of the total length during each instar were taken. For well over a hundred individuals measurements of carapace length and height were taken in addition to those of total length. The number of pre-adult instars varied upward from five. This variation was found in all clones tested. Growth curves have been constructed for various groups based on the number of pre-adult instars. The inflection in the growth curve of any dimension in any group coincides with the time of sexual maturity. " Brooks' Law " holds only for those groups which were primiparous during the sixth instar and then only approximately. Relative growth in the dimensions studied, .r and y. may be expressed by the equation A1 = - b.vk. Relative growth changes at sexual maturity, i.e., a change occurs in the values of the constants b and k. The above equation may perhaps be used to advantage in the study of variation in other Cladocera. By the use of this equation the direc- tion of variation can be given numerical values. Comparisons of differ- GROWTH AND VARIATION IN DAPHNIA MAGNA 97 ent races and varieties, both seasonal and geographic, may be readily made on this basis. BIBLIOGRAPHY AGAR, W. E., 1930. A Statistical Study of Regeneration in Two Species of Crus- tacea. Brit. Jour. E.vpcr. Biol.. 7: 349. BANTA, A. M.. 1921. A Convenient Culture Medium for Daphnids. Science, N. S., 53: 557. BROOKS, W. K., 1886. Report on the Stomatopoda collected by H. M. S. Chal- lenger during the years 1873-1876. Report of the Scientific Results of the voyage of H. M. S. Challenger during the years 1873-1876. Zoologv, Vol. 16, Part 45. CALVERT, P. P., 1929. Different Rates of Growth among Animals with Special Reference to the Odonata. Proc. Am. Phil. Soc., 68: 227. DOVER, C, 1931. Effects of Inadequate Feeding on Insect Metamorphosis. Na- titre, London, 128: 303. FOWLER, G. H., 1909. Biscayan Plankton Collected during a Cruise of H. M. S. ' Research,' 1900. Part XII. The Ostracoda. Trans. Linn. Soc., Lon- don, 2d Ser., Zool.. 10: 219. GURNEY. R., 1929. Dimorphism and Rate of Growth in Copepoda. /;;/. Rev. Hydrob. Hydrogr., 21: 189. HEBERER, G., 1928. Uber eine population von Daphnia ccphalata King aus Flores. Zool. Anz.. Supplementband 3. pp. 70-78. HERSH, A. H., 1928. Organic Correlation and its Modification in the Bar Series of Drosophila. Jour. E.rpcr. Zool.. 50: 239. HERSH, A. H., 1931. Facet Number and Genetic Growth Constants in Bar-eyed Stocks of Drosophila. Jour. E.rper. Zool., 60: 213. HUXLEY, J. S., 1924. Constant Differential Growth-ratios and their Significance. Nature, London, 114: 895. HUXLEY, J. S., 1927. Further Work on Heterogonic Growth. Biol. Zentralbl., 47: 151. RAMMNER, W., 1926. Formanalytische Untersuchungen an Bosminen. Int. Rev. Hydrob. Hydrogr.. 15: 89. RAMMNER, W., 1927. Die beschreibende und die bildliche Darstellung der Formanderung bei Cladoceren. Int. Rev. Hydrob. Hydrogr., 17: 115. RAMMNER, W., 1928. Zur morphogenese und biologie von Chydorus sphsericus und Pleuroxus trigonellus (Ergebnisse aus einzelzuchten). Zeitschr. Morph. Okol., 12: 283. RAMMNER, W., 1930a. Uber die Gultigkeit des Brooksschen Wachstumsgesetzes bei den Cladoceren. Arch. cntv.'. incch., 121: 111. RAMMNER, W., 1930k. Uber milieubedingte Missbildungen bei Daphnia pulex und Daphnia magna. Int. Rev. Hydrob. Hydrogr., 24: 1. ROBB, R. C., 1929. On the Nature of Hereditary Size Limitation. II. The growth of parts in relation to the whole. Brit. Jour. E.rper. Biol., 6: 311. SCHUBERT, A., 1929. Uber die (postembryonale) Formentwicklung bei zwei Lokalrassen von Ceriodaphnia reticulata Jurine. Int. Rev. Hydrob. Hy- drogr., 22: 111. SINGH-PRUTHI, H., 1925o. Studies on Insect Metamorphosis. III. Influence of starvation. Brit. Jour. E.rper. Biol., 3: 1. SINGH-PRUTHI, H., 1925/>. Moulting of Insects. Nature, London, 116: 938. SINGH-PRUTHI, H., 1931. Effects of Inadequate Feeding on Insect Metamorphosis. Nature, London, 128: 869. 98 B. G. ANDERSON WERNER, F., 1924. Variationsanalytische Untersuchungen an Chydoriden. Ver- such einer quantitativen Morphologic der Cladoceren-Schale. Zeitschr. Morph. Okol., 2: 58. WESENBERG-LUND, C., 1926. Contributions to the Biology and Morphology of the Genus Daphnia, with Some Remarks on Heredity. Kc/l. Danskc Vidcnsk. Selsk. Skriftcr. naturiv. mathcni. Afd., 8 Raekke, 11: 89. WOLTERECK, R., 1925. Notizen zur Biotypenbildung bei Cladoceren. I. Experi- mentelle Untersuchung der Ceresio-Daphnien. Int. Rcr. H\drob. H\- drogr., 14: 121. SOME EFFECTS OF HIGH PRESSURE ON DEVELOPING MARINE FORMS JOHN W. DRAPER AND DAYTON J. EDWARDS (From flic Department of Physiology, Cornell University Medical College, New York City, and the Marine Biological Laboratory, Woods Hole, Mass.) The action of hydrostatic pressure, helow certain limits, in producing an increase in the contractility of cardiac and skeletal muscle, raises a question of prime interest as to the nature of the effect. Does an in- crease in the absolute pressure of the environment of a tissue, with the resulting increase in tissue density, give rise to a general stimulation of the fundamental processes in the cells? Several lines of approach have been considered lor an answer to this question, but the aim in the present experiments will be to deal only with the influence of pressure on certain of the fundamental processes in the early development of marine eggs. The experimental observations relate mainly to two features : first, the effect of maintained compression on the rate of development of fer- tilized eggs, particularly in the early stages ; and second, the effects pro- duced by pressure on the rate of the heart in embryos at the stage when pulsations are just beginning and then later when the rhythm is fully established. The observations were made in an apparatus having essentially the same construction as that previously described (Edwards and Cattell, 1928). The fertilized eggs of Fundiilus were placed in glass vials filled with sea water and covered with thin rubber membrane to prevent their escape. One lot of the eggs was then placed in the compression cham- ber, which was also filled with sea water, and another lot, the control sample, was placed in a second chamber similar to the first. Iji this manner the factors of temperature, respiration and amount of agitation were kept constant, and the experimental sample differed only in having pressure applied for known intervals. In a few experiments the technic was modified to permit observation of the eggs during the period when they were under the action of pres- sure. The essential things for this purpose consisted of a heavy glass window mounted in one end of the compression chamber, which made it possible to view with low magnification objects placed immediately beneath it, and a depression slide fixed in position close to the inside sur- face of the window to contain the eggs. A small mirror backing the slide and a " Pointolite " lamp, which directed a beam of light through the window and against the mirror, permitted fair illumination for a microscopic examination of the eggs in the depression of the slide. All 99 100 JOHN W. DRAPER AND DAYTON J. EDWARDS of the observations on the change of heart rate under pressure were made by using this adaptation of the pressure apparatus. An observation of general interest arising from these experiments on compression of the developing egg is the extremely slight change that occurs in the egg structures when subjected to comparatively high pressures. The method has permitted observations on the diameter of the egg, the size and state of aggregation of the fat globules, the size and position of the blastodisc, and the size of the finer blood vessels in parts of the embryo. A close study of these different parts of the de- veloping egg, made during and immediately after the onset of pressure of 110 atmospheres, reveals no significant changes. Observations on advanced embryos within the egg, made while pressures of 1500 pounds per square inch were applied, may reveal nothing more than a few quick TABLE I The effect of pressure on the cleavage of Fundidus eggs. The time intervals in the fourth column signify the delay in the development of eggs subjected to pressure as compared with those maintained under control conditions. The observations were made at the cleavage stage shown in the fifth column. Experi- ment No. Pressure Duration of compression Delay of pressure eggs Stage of cleavage Ib. min. min. 1 1500 140 15 second 2 1700 60 15 first 13 second 3 1500 108 20 third 4 1950 120 10 third 5 1950 100 15 first 15 second movements of the embryo not unlike those shown at indifferent intervals by material of this kind. In order to test the effects of pressure on the rate of cell division it was necessary to have some standard for comparing the eggs subjected to pressure with those maintained under control conditions. In our initial experiments the completion of the membrane between two daugh- ter cells was taken arbitrarily as an end-point. With this criterion the data contained in Table I show that the eggs subjected to pressure are delayed about fifteen minutes in reaching a given stage of development. The exact time, however, at which a batch of developing eggs reached a given stage in division was often difficult to determine. Additional ob- servations, therefore, were taken (1) by making counts on the control EFFECTS OF PRESSURE ON MARINE FORMS 101 and the experimental samples to determine the predominant stage at a given time; and (2) by removing samples of 25 or more eggs which were placed in a fixing solution for later examination. The results by these methods of study are set forth in detail in Tables II and III and the evidence they bring forth lends support definitely to the conclusion that compression retards cell division. The eggs subjected to pressure have been carefully watched for any abnormalities that might occur. A number of monsters have been found of the types showing gross distortions of the body, changes of the cardiovascular system, and tendencies towards anophthalmus. The per- centage of abnormal specimens, however, was not large and there was no dominant type of dysmorphism. The abnormalities of the eye were usu- TABLE II The effect of pressure on the cleavage of Fundulus eggs. The data contained in the fourth and fifth columns represent simultaneous observations on samples of eggs subjected to pressure (pressure sample) and samples maintained for the control. The observations were made on living material. Experi- ment No. Pressure Duration of compression Pressure sample Control sample Ib. min. division stage division stage 1 1300 68 1-cell 2-cell 2 1700 112 2-cell 4-cell 3 1650 60 1 and 2-cell 2 and 4-cell 4 1650 30 1-cell 2-cell 5 1575 180 2-cell 4-cell 6 1560 130 1-cell 2-cell 7 1525 140 1-cell 2-cell 8 1300 70 1-cell 2-cell 9 1500 108 2-cell 4-cell 10 1950 120 4-cell 8-cell ally those showing one located craniad to the other but not exactly in «/ o j the midline, while the deformities of the cardiovascular system appeared usually as an asymmetrical development of the vessels. The failure of the blood vascular system to develop symmetrically causes the heart to be drawn to one side of the pericardial cavity. The Fundulus embryo is a favorable object for the study of the rate of the heart, since the pulsations of this organ may be observed almost at the time automaticity starts and may be followed until a completely functioning circulatory system is established. As heart automaticity is a fundamental property in development, with a fairly definite time of appearance, the study of the influence of pressure on this phenomenon presented features of unusual interest. The results of 17 experiments 102 JOHN W. DRAPER AND DAYTON J. EDWARDS are summarized in Table IV. These data show that a pressure of 1200 pounds produces a reduction in the heart rate within two minutes, amounting to an average decrease of 9.9 per cent from the control rate. The decline in rate continues, however, somewhat more slowly than the initial change, so that at the end of a 10-minute period of compression the average reduction was only 16.6 per cent below the control value. When pressure was applied, the pulsations of the heart showed a defi- nite reduction in the rate within a half minute of the onset. Although in some instances a more or less gradual decline prevails under pressure, the more common type of change appears to be a fairly abrupt slowing. The action of different amounts of pressure was tested, within the range of 400 to 1200 pounds, in an attempt to determine if critical points TABLE III The effect of pressure on the cleavage of Fundulus eggs. The data contained in the fourth and fifth columns represent counts made on samples of eggs subjected to pressure (pressure sample) and samples maintained for the control. The observa- tions were made on fixed material. Pressure sample Control sample Duration Experi- Pressure of com- ment pression 2-cell 4-cell 8-cell 2-cell 4-cell 8-cell stage stage stage stage stage stage Ib. min. 1 1500 135 32 1 0 0 35 0 2 1500 135 30 0 0 0 27 0 3 1500 190 20 0 0 0 20 0 4 1500 175 0 28 4 0 0 32 5 1500 140 13 0 0 0 12 0 Total count 96 29 4 0 94 32 exist in the pressure effect. The results support the view that the pres- sure effect becomes progressively greater with the higher grades of compression on the heart. With pressures ranging from 1200 to 1900 pounds we have been able to suppress the automaticity of the heart to the extent of not being able to observe under low magnification any indication of a contractile response. We do not overlook the fact that localized fine pulsations of a few fibers may have persisted in these preparations, so small indeed as to have been beyond our range of identification, but the essential fact is that automaticity was practically stopped. On release of pressure these hearts which have been held quiescent for several minutes immediately show signs of activity and EFFECTS OF PRESSURE ON MARINE FORMS 103 eventually recover, thereby confirming our previous observations (Ed- wards and Cattell. 1928) that the pressure effect is freely reversible in nature. In establishing a pressure standstill of an embryo heart several changes have been observed to occur, such as arrhythmia, partial and complete heart block, and fibrillary motion of the auricles terminating in a cessation of activity in the region of the sinus. With the reestab- lishment of activity after release of pressure, sometimes recovery ap- peared almost simultaneously in sinus, auricle, and ventricle, while in TABLE IV Effect of Pressure on the Heart Rate in Fundulus Embryo Experiment Age Control heart rate Pressure Heart rate dur- ing compression Heart rate after pressure release Per cent .A in heart rate during com- pression 2 min. 10 min. 2 min. 5 min. 2 min. 10 min. hr. beats per min. Ib. beats per min. beats per min. beats per min. beats per min. 2 120 68 1200 65 65 60 60 7.3 7.3 3 79 62 1200 58 52 53 — 6.4 16.2 4 96 55 1200 47 43 53 50 9.1 21.9 5a 119 53 1200 48 48 55 54 9.4 9.4 5b 120 55 1200 50 48 — — 9.1 12.8 6 76 55 1200 50 50 — — • 9.1 9.1 7 121 50 1200 44 46 54 — 12.0 8.0 8 192 62 1200 53 53 62 63 14.5 14.5 9 172 57 1200 54 53 56 — . 5.2 7.0 lOa 216 65 1200 60 59 67 66 7.7 9.2 lOb 217 66 1200 60 — 67 67 9.0 . — . 11 150 73 1200 67 66 72 — . 8.2 9.6 17 74 72 1200 69 63 69 72 4.1 12.6 19 26 60 1200 56 50 53 60 6.6 11.7 20 120 60 1200 52 53 64 60 13.3 11.7 21 144 116 1200 96 100 106 103 17.2 13.8 22 144 112 1200 88 84 99 104 21.4 25.1 Average 9.9 16.6 other instances the return followed the reverse order of the disappear- ance with some degree of arrhythmia preceding the dominance of the normal type. The remarkable effect of pressure in reducing and abolishing the rhythmicity of the embryo heart raises a question as to the nature of the action. A factor that immediately occurs to one as a possible cause for the restraining action of pressure is a direct excitation of inhibitory nerve fibers supplying the heart. Experiments designed to throw light 104 JOHN W. DRAPER AND DAYTON J. EDWARDS on this suggestion were carried out in the following way. Embryos ranging in age from 14 to 20 days were carefully dissected and the heart completely isolated from the surrounding tissue. The preparations were kept immersed in a 40/60 dilution of sea water and tap water to which was added 5 per cent of glucose. These hearts are not easy to handle on account of the very small size, but with care it was possible to mount them on a depression slide and to place them in the compression chamber where their activity could be followed through the window with a low power objective. The results of these observations confirm in all essentials those ob- tained with the entire embryos. Moreover, the same type of changes was present, as, for example, the arrhythmia, the different degrees of block in conduction, and the localized areas of rhythmicity. When these results are considered in conjunction with those on whole embryos taken at an early stage before nerve connections are established, they furnish additional support to the view that the depressing action of pressure on rhythmicity is not through an effect on the inhibitory nerve mechanism. DISCUSSION In the agent pressure we have an instrument by which the contractil- ity of cardiac and skeletal muscle tissue may be greatly stepped-up, but this action, remarkable as it is for these tissues, appears to be a peculiar effect on the contractile mechanism of these types of muscle. In the present experiments we have evidence that such fundamental biological properties as cell division and automaticity of the embryonic heart be- come restrained by pressure — a fact that presents equally difficult ques- tions to answer as those given by the augmentation phenomenon in cer- tain types of contractile tissue. While the types of chemical change underlying the processes of cell division and heart rhythmicity obviously are complex and the factors common to both cannot be set down, yet the fact is not without interest that some forms of chemical reactions are known to be influenced by pressure. Rothmund (1896) found that the acid inversion of cane sugar, a first order catalytic reaction, is decreased in velocity about five per cent when subjected to a pressure of 500 atmospheres, and on the- oretical grounds there is reason to conclude, according to Jones (1915), that the velocity of second order reactions is influenced in direct pro- portion to the pressure. Attention was called in the early part of this paper to the observation that the cell constituents gave no gross signs of change under pressure. EFFECTS OF PRESSURE ON MARINE FORMS 105 With albumin, however, an addition of energy to the material produces a tendency to coagulation, as shown by Fermau and Pauli (1915) on irradiating with radium salts, and by Bridgman (1914) with high pres- sures : therefore it is probable that some alteration takes place in the egg substance even at the comparatively low pressures used in our experi- ments, but its character is such as to be not easily recognizable. The observations of Heilbrunn (1920. 1921) show in a definite way how great are the changes in viscosity of egg protoplasm during the process of division and they emphasize also the manner in which mitosis may be inhibited by factors that tend to modify the viscosity changes. The pressure effect in retarding cell division may be, therefore, largely one of altering the viscous properties of the cell constituents. Perhaps the one factor upon which attention first becomes focused in the attempt to account for the decrease in division rate of eggs under pressure is a possible change in the oxygen supply. This feature re- ceived thoughtful attention at the outset of our experiments, and we believe that the details of procedure employed provide adequate controls and successfully rule out a change in gas tension as a factor in our results. Our early observations on the effect of pressure gave evidence to support the view that the rhythmicity of the isolated heart is accelerated when a compression up to 60 atmospheres (882 pounds per square inch) is made to act upon this organ. The procedure in these initial experi- ments, however, of subjecting the preparation to pressure for only brief intervals and of taking records of the heart cycles immediately following the onset of pressure gave rise to an incomplete conception of the action of this agent. A review of our original tracings and evidence from many additional observations made since show quite conclusively that pressure does not induce an acceleration of the rate of the heart except as a temporary event. The present results, therefore, showing always depression of the rate under pressure seem at first discordant with previous findings, but full consideration of the facts indicates quite clearly that we are dealing with the secondary action of pressure on rhythmicity in contrast to the initial temporary effect that occurred during the first eight to ten cycles in a rhythmic preparation following the onset of compression. The effect of pressure in slowing the heart rate raises a question con- cerning the mechanism of this action. While a mechanical factor may have contributed some part in this type of response, it is evident that such an effect does not arise from an increase in the resistance to the 106 JOHN W. DRAPER AND DAYTON J. EDWARDS circulation by pressure narrowing the peripheral vessels, since careful observations of the size of certain small vascular channels reveal no detectable alterations in their caliber with the application of pressure. The property of automaticity of the heart, on the other hand, is influ- enced in a marked degree by changes in the ionic relations of its en- vironment, and the observations of Bogojawlensky and Tammann (1898) on the conductivity-pressure coefficient of some conducting sys- tems gives proof of the changes in the mobility of ions and of the alterations in the ionization of electrolytes by pressure. The precise changes in ionic equilibrium that may have operated in producing a retardation of the heart rate under compression cannot be given from the data at hand, but that a factor of this nature plays a part is strongly suggested by the type of effect that pressure gives in slowing the embrvonic heart. •/ SUMMARY Hydrostatic pressure of as much as 1900 pounds per square inch applied to living eggs of the Fnndulus causes no evident changes in the cell constituents. The rate of cell division of I: undid us eggs was slowed by maintain- ing them under pressures of 1300 to 1900 pounds for periods ranging from 30 minutes to three hours. In the pressure samples a few in- stances appeared of abnormal forms of development. The automaticity of the heart in young embryos becomes slowed by pressure and apparently may be abolished if the compression is main- tained but returns within a few minutes following release of the pres- sure. In ten experiments an average decrease of 9.9 per cent in the heart rhythm under pressure was present within an interval of two min- utes. The embryonic heart under pressure develops arrhythmia, types of local block, and isolated fibrillary activity. The action of pressure in slowing the heart rate does not depend on the presence of inhibitory nerves to this organ, since compression pro- duces a slower rhythm of (a) hearts from older embryos that have been dissected free of all extrinsic nerve connections; and (b) hearts of young embryos that have no intrinsic nerves developed. The bearing of viscosity and ionic changes produced by pressure are considered in relation to the decrease in the rate of cell division and the automaticitv of the embrvonic heart. EFFECTS OF PRESSURE ON MARINE FORMS 107 BIBLIOGRAPHY BOGOJAWLENSKY, A., AND G. TAMMANY, 1898. Zcitsclir. f. pltys. Chcui.. 27: 457. BRIDGMAN, P. W., 1914. Jour. Biol. Client., 19: 511. EDWARDS, D. J., AND McKEEN CATTELL, 1928. Am. Join: Physiol.. 84: 472. FERMAU, A.. AND W. PAULI. 1915. Biochcm. Zcitsclu:, 70: 426. HEILBRL-NX, L. V., 1920. Biol. Bull., 39: 307. HEILBRUNN, L. V., 1921. Join: E.rpcr. Zo'ol., 34: 417. JONES, H. C, 1915. Elements of Physical Chemistry, 4th ed. New York. ROTHMUND, V., 1896. Zcitsclu: }. pltys. Chcm.. 20: 168. THE CRUSTACEAN EYE HORMONE AS A VERTEBRATE MELANOPHORE ACTIVATOR EARLE B. PERKINS AND BENJAMIN KROPP ZOOLOGICAL LABORATORY, RUTGERS UNIVERSITY There is produced in the eye stalks of crustaceans a hormone which, carried in the blood stream, is effective in inducing chromatophore changes. This has been demonstrated repeatedly by Perkins (1928) and Perkins and Snook (1931) for Palccmonctcs, and by Roller (1928) for Crangon. The active material acts in every case as a contractor, — that is, an extract from eye stalks, when introduced into the circulation of blinded or black-adapted shrimps, causes maximal contraction of the red and yellow chromatophores. Roller (1928) also reported the oc- currence of the reciprocal phenomenon, namely, expansion of chromato- phores of white-adapted shrimps induced by extracts from the rostral region of Crangon. Thus far this has failed of experimental confirma- tion (Perkins and Snook, 1931). In 1929, Rropp reported the presence in the eyes of black-adapted tadpoles of a substance effective in producing expansion of melanophores in white-adapted Fund id us and tadpoles of Rana clainitaiis. In this case the use of extracts from eyes of white-adapted tadpoles gave no results on black-adapted animals. Here, as in the case of the inverte- brates, evidence pointed to the presence in the eyes of a substance which, under proper conditions, induced effects on chromatophores. To test the interspecificity of the chromatophore activator found in the crustacean eye stalk, and thus further to establish its hormone na- ture, was the object of the experiments here reported. Extracts from the eye stalks of Pahrinoiictes I'lilgarls were made in 0.7 per cent NaCl solution. Thirty eye stalks were macerated in 5 cc. salt solution, boiled, centrifugecl, and the clear extract decanted off and cooled. Tadpoles of Rana clainiians were placed in white and black dishes until their skin melanophores were respectively nearly maximally contracted and ex- panded. To the unaided eye the white-adapted tadpoles appeared light yellow-green, the black-adapted almost black. Each individual then received 0.2 cc. of the extract in the dorsal lymph sinus. White-adapted controls received 0.2 cc. 0.7 per cent NaCl solution. Animals adapted to a black background, which received 0.2 cc. of the eye extract, showed no obvious melanophore movements as a result of the injection, nor did 108 CRUSTACEAN EYE HORMONE 109 the controls at any time show marked changes in pigmentation. The animals adapted to a white background, however, did show pronounced effects. Within five minutes of the injection these tadpoles began to darken over most of the dorsal surface, even though they remained on a white background. Maximal darkening was reached after thirteen minutes. The tadpoles at this time showed a very dark band across the FIG. 1. Left, black-adapted tadpole injected with 0.2 cc. Palcrmonctcs eye ex- tract. Center, white-adapted tadpole injected with 0.2 cc. Palccinonctcs eye extract: melanophores expanded following injection. Right, white-adapted tadpole injected with 0.2 cc. 0.7 per cent NaCl : no expansion of melanophores. whole dorsal surface, extending forward to a line connecting the pos- terior margins of the eyes, and backwards to the anterior root of the tail (Fig. 1). The snout region remained characteristically in the original light condition when the injection was made into the dorsal lymph spaces. We assumed that since the skin between the eyes and over most of the head adheres closely to the cranium, the lack of response of the 110 E. B. PERKINS AND B. KROPP head melanophores was due solely to the presence of this mechanical bar- rier and the consequent inability of the material to come into contact with the skin in this region. This assumption proved correct, for injec- tion of the extract directly under the head skin produced the darkening reaction over the head. The tail did not show the effect, although a suggestion of it could be obtained by injecting the extract directly into FIG. 2. Left, eye-extract injected white-adapted tadpole. Center, pallor ap- pearing after effect of injecting eye extract had worn off. Sharp line of demarca- tion between area of punctate melanophores and stellate melanophores of tail clearly shown. Right, normal black-adapted tadpole. the tail musculature. This is only an apparent anomaly and by no means real, as will be shown below. Microscopic examination of the skin of the living animal at this time revealed the expected condition — a dense network of expanded melano- phores all over the darkened area. Fifteen minutes after the injection the darkening began to disappear until, thirty minutes after maximum darkening, the animal was once more in its original light state, the ef- CRUSTACEAN EYE HORMONE 111 fects of the extract apparently having been dissipated. The process of returning pallor did not. however, stop at this point. It continued until the entire dorsal surface of the animal became lighter than it was before the extract was injected. This after-pallor may persist for twenty-four hours or more. It is confined to the dorsum and the root of the tail, the greater mass of the tail being non-reactive and appearing dark in con- trast to the pale dorsal surface (Fig. 2). Microscopic examination of the living animal's skin showed, as expected, sharply contracted, punctate dermal melanophores in the trunk, while those of the tail were stellate. The region of demarcation between trunk and tail was strikingly sharp and in different animals was surprisingly similar in contour. The darker area of the tail formed a V, pointing cephalad, at a point just posterior to the union of tail and trunk. Laterally the line ran pos- teriorly and ventrally until near the ventral raphe of the tail, where it continued laterally but anteriorly and was lost in the lightly pigmented skin of the venter. In some animals the line followed superficially the course of the spinal nerves in the tail, while in others there was a more irregular course. In addition to eye stalks of Palfcnionctcs rulgaris we used those of the blue crab, Callineclcs sapidus. The active substance was present here also, and the results described above hold true in every detail for tadpoles treated with extracts of eye stalks of Callinectes. The only differences observed were that with Palffiuonctcs extracts the onset of the darkening reaction was more rapid, and the degree of maximum darkening attained was somewhat greater. Since hormones are carried by the blood stream to all parts of the organism, it appears irregular at first sight that the tadpole tail should not respond as did the dorsal surface of the trunk. The tadpole tail has often been an object for experimentation on color change and pig- mentary reactions, but results have usually been inconclusive or contra- dictory (Kropp, 1927). It is indeed not strange that this is so, for the tail of the tadpole, especially in older specimens and when the hind legs have appeared, is in a state of physiological flux. The disorganization of the pigmentary system may be correlated with circulatory changes and the general resorption process of this organ. The fact that the tail is a poorly reacting system is probably due to this physiological instabil- ity. In our experiments the reacting substance was carried by lymph and to some extent by the blood stream. For practical reasons we did not inject intravenously at this time. However, we venture the predic- tion that intravenous injection will not greatly alter the reaction of the tail as described above. The injected substance has opposite effects in the shrimp and in tad- 112 E. B. PERKINS AND B. KROPP poles, — producing contraction of chromatophores in the former, expan- sion in the latter. This is a further example of a hormone producing different results depending on the reacting system or type of chromato- phore concerned. The chromatophore activator produced by the crustacean eye stalk shows all the characteristics of a true hormone now that we may include its wide interspecificity. This has recently been reported also by Koller and Meyer (1930) and by Meyer (1931), who injected eye extracts and rostral-region extracts of Crangon into the fishes Gobiits and Plcuron- cctcs, producing chromatophore movements. \Yith the eye-stalk extract they report the production of pallor in these fishes, — the opposite of the effect we obtained on frog tadpoles. The fact that an invertebrate hormone is effective on a vertebrate system seems highly significant. As yet very few endocrine processes are known among the invertebrates — the crustacean eye hormone being by far the most spectacular. The attention that has been given to the vertebrate endocrine system has disclosed a variety of humoral effects, J -/ but it is by no means certain that this important coordinating mechanism is confined to the vertebrates alone. It is possible that every tissue possesses the ability to affect other tissues by means of substances lib- erated into the circulating media, not only in the vertebrates but in the invertebrates as well. The occurrence of such a substance in crusta- ceans which is effective in fishes and amphibians may indicate an evolu- tionary precursor of the more highly developed vertebrate endocrine system. CITATIONS KOLLER, G., 1928. Zcitscln: rcn/l. Physio!.. 8: 601. ROLLER, G., AND E. MEYER. 1930. B'wl Zcntralbl.. 50: 759. KROPP, B., 1927. Join: E.rpcr. Zool. 49: 289. KROPP, B., 1929. Pror. Nat. Acad. Sci.. 15: 693. MEYER, E., 1931. Zool. Jahrb.. Abt. Allg. Zool. it. Physiol.. 49: 231. PERKINS, E. B., 1928. Jour. E.vpcr. Zool.. 50: 71. PERKINS, E. B., AND THEODORE SNOOK, 1931. Proc. Nat. Acad. Sci., 17: 282. " MITOGENETK; RAYS "—A CRITIQUE OF THE YEAST- DETECTOR METHOD OSCAR W. RICHARDS AND G. WELLFORD TAYLOR1 (From the Osborn Zoological Laboratory, Yale University, the Physiological Laboratory, Princeton University, and the Marine Biological Laboratory, Woods Hole) The interest in the theory of mitogenetic radiation has brought forth a considerable literature. Schrieber and Luntz (1931) recently list 49 cases indicating the existence of these rays and 26 cases where negative results were obtained when yeast was used as a detector for the radia- tion. These rays of short wave length (2000-3400 A) are supposed to be given off by a wide variety of tissues and to possess the property of accelerating the division of other cells placed in their path. Further, the budding of yeast is generally considered by the proponents of this theory to be especially sensitive to these rays and to be one of the best detectors of this radiation. The literature has been reviewed recently by Hollaender and Schoeffel (1931) and by Taylor and Harvey (1931). Baron (1930) reports mutuoinduction or the acceleration of the bud- ding of yeast by their own radiation. The husbandry of the yeast in many of the published papers on the effects of such radiation is not satisfactory to one familiar with the growth of yeast, and the following minimal essentials are stated to estab- lish criteria for the evaluation of experimental data on the theory of mitogenetic radiation. Violation of these well-established fundamentals of yeast culture accounts for much of the conflicting opinion and inade- quate literature on this subject. First, the yeast used as a detector for this radiation must be a pure strain obtained from an isolated, single cell. Single cell isolation studies by Wallace and Tanner (1928) have shown the amount of variation of different cells from the same pure species of yeast. Second, both the rate of the growth and the yield of yeast should be reported as well as the percentage of cells with buds. An increase in the population, over a period of time, is a better index of proliferation than the percentage 1 Mr. Taylor expresses his appreciation to the staff of the Woods Hole Labora- tory of the U. S. Bureau of Fisheries for facilities placed at his disposal during the summer of 1931 when these experiments were performed. 8 113 114 O. W. RICHARDS AND G. W. TAYLOR of buds. Third, the mortality of cells must be determined at the same time. A high mortality would vitiate such experiments and evidence should be produced that there has been no selective killing as normally occurs during the latter part of the growth of a population when the culture medium is not maintained effectively constant (Richards, 1928a, b, 1932a). Fourth, the relative and absolute errors of sampling the populations and of the counting, etc., must be measured and stated. Fifth, the yeast should be maintained at the proper temperature for the species studied and the temperature used stated (e.g., not just as room temperature, Streline, 1929) . Sac char omyces ccrevisice shows irregu- larity of division with elongate cells at 30.0° and injury at higher tem- peratures.2 Sixth, the medium should be maintained effectively con- stant, except for the radiation of the experimental cultures, so that the rate of growth of the yeast otherwise is constant. With many of the culture fluids this limits the length of the experiment with respect to the amount of seeding and the volume of the medium to less than 40 hours unless the medium is changed during the experiment. For a given amount of medium a definitely limited yeast crop will be obtained when the food is exhausted and the waste products accumulate in suffi- 'cient concentration to check the growth. With large seedings this crop is obtained in less time than with smaller seedings, even though the rate of growth (percentage and rate of budding) may be the same in both cases. When very small amounts of culture medium, single drops, are used the succession of the known changes which determine the growth of the population 3 would influence and obscure the effects of radiation. This is especially true with experiments on mutuoinduction when the cell densities are proportionally greater. The proponents of the theory of mitogenetic rays have implied that yeast is unsatisfactory as a detector when it is growing at a constant rate in an effectively constant environment because it is then growing as rapidly as possible. They presume that there is a latent period in cell division which limits the rate of its recurrence. This objection to con- stant conditions does not affect the experiments to be described because with the same culture conditions the addition of a growth stimulant, such as inosite (Richards, 1932&), in a concentration of 1 : 100,000 will significantly increase the rate of growth. It is also to be remembered that the budding of most of the common species of yeast is different from that of cultures of tissues from multicellular organisms in that the same mother cell may form simultaneously two or more buds. Baron ./ ^ 2 Borodin (1930) neglects this fact, which probably accounts for much of the variation in his experiments (cf. Richards, 1928c). 3 Cf. Richards (19322 G3 G4 Go G3Gt G3 G3 G4 G» 2 18 26 20 17 17 27 -31 -16 + 18 ± 0 -26 - 4 4 24 28 29 14 20 16 -14 + 19 +45 -30 -80 - 1 Series 6 19 14 25 14 20 20 +36 - 2 + 25 -30 +25 -30 109 8 14 14 21 23 15 17 db 0 + 25 +40 +53 +24 +35 22 13 12 13 13 12 17 + 9 -10 + 8 + 8 -24 -24 2 19 18.5 18.5 17 20 20 + 5 -16 -11 -22 -11 — 22 4 18.5 20.5 17.5 16 16.5 22 - 1 -12 + 6 + 3 -11 -28 Series 6 16.5 12.5 12.5 15 14 26 +36 -31 -11 + 7 -51 -44 110 9 11.5 13 13.0 16.5 13 16.5 -12 ± 0 ± 0 + 19 -15 ± 0 24 7 10.5 12.5 13.5 14 9 -33 + 13 -11 - 4 +39 +50 3 28 27 35.5 32 34 28.5 + 3 ± 0 + 4 - 6 + 25 + 12 5 21 26 20.5 16 18 23 -19 -15 + 14 -11 -11 -30 Series 8 18 22 21.5 17 23 21 -18 - 6 - 7 -26 + 2 -19 112 12 25.5 30.5 35.5 37 38.5 33 -16 + 1 - 8 - 4 + 8 + 12 24 23.5 24 21.5 20.5 20 25.5 2 - 8 + 8 + 3 -16 -20 3 35 29 30.5 29.5 34 26 + 27 ± 0 -10 -14 +40 + 12 Series 7 22 27 25 23 25 23 -23 ± 0 db 0 - 8 + 9 ± 0 113 11 24.5 26.5 22 18.5 23 24.5 - 8 -16 - 4 -20 -10 -25 24 28.5 28.5 39 42 39.5 42.5 ± 0 - 1 -13 + 6 - 8 1 * Calculated from the average percentage of budding with the two quartz tube and the two glass tube populations. QiG3 was calculated from the first quartz tube percentage and from the first glass tube percentage, etc., for all four combinations of the data. culture of luminous bacteria, from a single cell of Vibrio phosphorescens, for the generation of mitogenetic rays. The results are reported in Table I, Series 109 and 110, The percentage of induction, 7, was cal- culated from the formula given by Gurwitsch (1925) : r i 118 o. W. RICHARDS AND G. W. TAYLOR where Q is the percentage of budding in the quartz tubes and G the percentage of budding in the control or glass tubes. In the special con- trol of a glass and a quartz tube shielded from any rays an apparent positive induction occurred four times out of ten possibilities (Table I, columns Q0 and G0), two of which were within the significant range, according to Gurwitsch (1925), of 30-120 per cent even though no bacteria were available to generate mitogenetic rays. The rest of the observations were negative or zero. When the percentages of the two quartz and two glass tubes placed in the bacterial suspension were averaged respectively and the averages used to calculate the induction (columns Ql Q2, G.A G4), only three positive cases in ten experiments were found, and in none of these was the induction great enough to be significant. It is further desirable to calculate the induction from the four com- binations of the percentages of the first glass control tube population (G3) and the first and second quartz (Ql and Q2) tubes separately and likewise with the second glass control tube (G4) and the two quartz tube percentages of budding. Examination of the results tabulated in the last four columns of Table I shows only 15 positive cases of the 40 computed percentages and of these only 7 are of sufficient magnitude to be significant even though all had been exposed to an actively growing culture of bacteria which was present to generate mitogenetic radiation. Another species of yeast, Sac char oinyccs clUpsoidcus, has been sug- gested as being possibly more sensitive to mitogenetic rays than S. cere- visicr, so we obtained a transfer of culture No. 4116 from the American Type Culture Collection and repeated the above experiment using this species. This time we used malt extract medium, which is a very suit- able culture fluid for this species, and the results of this experiment are given as Series 112 in Table I. Distinctly less induction was observed with a preponderance of negative cases. It is noteworthy that one sig- nificant positive case of 36 per cent was obtained in the special control without radiation, since this shows the amount of variation occurring with very carefully controlled conditions. The only explanation for these consistent negative results is the possibility that luminous bacteria, or at least the species here used, do not emit mitogenetic radiation. In order to show that this bacterium is not unique in this respect, the above experiments were repeated using Phytoinonas tuiiicfacicus (Am. Type Coll. Xo. 4452), as this species has been reported by other experimenters to be a good generator of mitogenetic rays. The observations are recorded in Table I as Series 113 and only one significant positive case of induction was observed out of eight observations. MITOGENETIC RAYS 119 During most of the period of the experiments the growth of the yeast is logarithmic. Should there be any acceleration of the multipli- cation of the yeast due to stimulation by mitogenetic radiation, it should show not only as an increase in the rate of growth but also as an in- creased yield in the quartz tubes over that obtained in the glass tubes. A slight acceleration, possibly less than would be noticed in the per- centages of budding, would be distinctly noticeable by the end of the period of growth. To test this the ratios of the number of cells in the quartz tubes to the number in the glass tubes at the end of each series were calculated. For the control which was shielded from the bacteria, the average of the ratios was 0.91. showing that there were slightly more FIGURE 2 FIG. 2. Arrangement for exposing tubes to radiation from onion roots. When in use, the beaker was partly filled with water to prevent drying of the onion roots. cells present in the glass than in the quartz tubes. The average of the same ratios for the tubes suspended in the bacterial culture was 0.99. Since twice as many counts were averaged than with the special control, more of the variation is cancelled, which accounts for the index being nearer unity. The maximum yield is less in the quartz tubes than in the glass tubes, which again fails to indicate any stimulation of budding attributable to a mitogenetic radiation. When the logarithm of the number of cells is plotted against time, the slope of the resulting curve is the relative rate of multiplication of the cells. This graph is linear during all of these experiments except the last observation of five cultures, because the seeding was from ac- 120 O. W. RICHARDS AND G. W. TAYLOR tively growing cultures. These curves were plotted for each tube of the four series and the slope of the curve measured with a tangentmeter (Richards and Roope, 1930). An index was obtained by dividing the tangent of the curve from a quartz tube by that from the corresponding glass tube. The average of these ratios for the control with no bacteria TABLE II Percentage of Budding and "Induction" Series Time Budding in Quartz Budding in Glass Induction hours per cent per cent per cent 117 2 21 24.5 -14 4 23 28 -18 6 27.5 28 - 2 8 33.5 33 + 2 11 22.5 26.5 -15 23 23 23 - 8 118 2 22 33 -33 41 28.5 32.5 -13 81 35 30.5 + 15 121 36.5 31.5 + 16 26 27 28 - 4 119 2 37 34.5 + 7 5 37 38.5 - 4 7 35 45 -22 12.5 35 40 -12.5 24 37.5 35.5 + 4 Summary 5+ 31% 11- 69% Ratios Quartz/Glass Series Growth rate Yield Detector Generator 117 1.07 1.11 5. ellip. Onion 118 .96 1.05 S. cerev. Onion 119 1.00 .84 S. ellip. Onion was 0.97 and for the tubes in the bacterial suspension 0.86. This shows that the rate of growth was slightly greater for the yeast populations in glass tubes, which might be due to some material in the glass dissolving into the medium, which causes a very slight increase of growth. This could not be enough to impair the value of the glass as a control as the differences are but slightly more than the probable error of the differ- MITOGENETIC RAYS 121 ences, and the sign of the difference was not the same in the several experiments. This is not clue to inhibition, because some of the experi- ments gave greater growth in quartz than in glass. These two addi- tional _ criteria also indicate that there is no influence passing through quartz but not through glass to stimulate the multiplication of yeast. Three series of experiments were made using onion roots as the generator of mitogenetic rays and using both species of yeast as de- u o CM E HH CO W U o C) -50 -30 -20 -4, cr < ui 3 _i (A & O 119 • o S. ELLIR 117 * A « 1 18 • D S. CEREV. 4 I 8 I 12 J 16 20 I I 24- 26 I I HOURS FIG. 3. .4. The growth curve of Series 117-119. B. Hypothetical growth curve that would be obtained with a 10 per cent stimulation of cell multiplication (cf. text). tectors. The arrangement of the experiments is shown in Figure 2. Since the yeast suspension in the tubes covered a considerable area, all of the roots were left on the onion and the positions of the tubes were reversed at each count to insure comparable conditions. The results are given in Table II. The negative cases again predominate and no essential differences are noted in the ratios of the rates of growth or in the yield between the experimental and the control cultures. 122 Q. W. RICHARDS AND G. W. TAYLOR The actual counts of the growth of cells are given in Fig. 3, curve A. Each point is an average of several counts for a single tube. Because only one tube was used in each experiment, all of the variation is shown in the figure. The growth for the first twelve hours is logarithmic, but the final observation of both Series 117 and 118 was beyond the period of constant growth rate. The effects of a lag period are slight. Care- ful examination of the points shows no consistent change in the growth that can be interpreted as an acceleration or an acceleration followed by a retardation. The average growth of the onion roots in the moist at- mosphere during the experiments was 15 per cent. The concluding experiments were designed to test whether yeast would stimulate other yeast when the two populations were separated by quartz. The arrangement of the dishes is shown in Figure 4. Two cubic centimeters of the yeast suspension were used for both the gen- erator and detector populations. This amount filled the space in the lower container up to the quartz partition and is about the smallest amount that allows counts without considerable errors in sampling. RTZ. FIGURE 4- FIG. 4. Cross-section showing arrangement of a quartz dish within a Petri dish. The detector yeast was placed at A and the generator yeast at B. The observations are given in Table III. Two duplicate experiments were made in each case. The induction percentages for the averages of both duplicates give an equal number of positive and negative cases in none of which are the differences great enough to be significant. When all combinations of the experiments are calculated, there is again an excess of negative cases. The average figures suggest that there may be some increased budding in the quartz dishes at two to four hours after seeding. These differences are not of sufficient magnitude to be significant and in the case of Series 119, the positive deviation is due to variation in only one of the two populations. Ill There is noticed a considerable variation between the different cul- tures even though conditions were made as nearly uniform as possible. The extent of this natural variation was measured in several cases, two of which are given in Table IV showing the percentages of buds counted and the probable error of the percentages. This was obtained by keep- MITOGENETIC RAYS 123 ing a separate record of the counts in each of the 20 unit volumes counted and the usual statistical formula applied. With divergent per- centages (Case 1) the induction varies from 2 per cent to 35 per cent for values within the range (mean ± its probable error) wherein the TABLE III Percentage of "Induction" Percentage of Series Time induction Av. Glass Qid OiG2 QiGi Q2G2 Av. Quartz hrs. H7a 2 + 2 ± 0 + 4 ± 0 + 14 4 + 27 +39 +45 + 9 + 14 6 ± 0 -13 + 11 - 6 + 20 8 -26 -26 -26 -26 -26 11 -17 -11 -14 -16 -19 23 -20 -15 -19 -19 -27 118a 2 -12 -21 -11 -13 - 2 4 + 28 + 25 +21 + 27 +33 8 -13 - 6 -17 - 7 -18 12 + 3 -19 -14 + 19 + 28 26 + 11 + 7 + 7 + 15 + 15 119a 2 - 4 + 27 - 1 - 4 - 8 5 + 14 -21 -10 - 4 + 8 7 -12 i - 7 - 9 -14 12.5 + 12 + 15 +30 -11 + 2 24 - 2 -11 - 7 + 4 + 9 Summary 8+ 50% 8- 50% Summary 25+ 38% 2±0 4% 37- 58% Initial concentration of yeast (cells in 1/250 mm.3) Series Generator Detector H7a 70.8 5. ellipsoideus 11.0 5. ellipsoideus llSa 47.1 S. cerevisiae 6.4 S. cerevisiae 119a 39.0 S. cerevisiae 5.1 S. ellipsoideus occurrence of another observation is expected to be equally probable. This result is important, as percentages of induction greater than 30 per cent are significant according to Gurwitsch (1925). Where the values approach each other more closely (Case 2), the induction calcu- lated varies from - - 2 per cent to -j- 33 per cent ; or from a negative 124 O. W. RICHARDS AND G. W. TAYLOR result (less budding in quartz than in glass) to a significantly positive result. This clearly shows that the index of induction obtained from this formula is invalidated by the normal variation of the biological ma- terial and is, therefore, not a satisfactory criterion for establishing the existence of mitogenetic radiation. Gurwitsch (1931) has recently indicated that continued radiation may produce an inhibitory effect on cell multiplication. The above ex- periments do not support this hypothesis because in them the direction of variation is independent of time. Further, it would be improbable that the inhibition would be such as to give, in all cases, the same or TABLE IV Effects of Variation on ''Induction" * PERCENTAGE OF BUDS COUNTED PERCENTAGE OF "INDUCTION" Quartz tubes Glass tubes Case 1 30.5 ± 2.1 26.0 ± 1.8 30.5 26.0 + 17 30.5 + 2.1 == 32.6 26 - 1.8 = 24.2 +35 30.5 - 2.1 == 28.4 26 + 1.8 = 27.8 + 2 32.6 27.8 + 17 28.4 24.2 + 1" Case 2 35.9 ± 2.4 31.5 ± 2.7 35.9 31.5 + 14 38.3 28.8 +33 33.5 34.2 - 2 38.3 34.2 + 12 33.5 28.8 + 16 * In each case the mean and its probable error are given. The "induction" is calculated with the Gurwitsch formula (cf. text) from the mean values and for the combinations of the mean values both plus and minus the probable errors of the re- spective means. slightly less variation than the amount of control growth when the con- centrations of the generator and detector populations varied over the limits used in our experiments (cf. Table III). Baron (1930) reports mutuoinduction in yeast populations when he used hanging drops with different densities of cells and rounded up or spread out in form. He starts with cells having no buds and reports the percentage of cells which still have no buds at each observation. The difference in the percentages indicated increased budding when the cells were in close proximity, which increase passed through a maximum and then decreased. A few negative cases were found and explained MITOGENETIC RAYS 125 by having too densely seeded original populations. Baron attributes the shortening of the lag period to premature stimulation of budding by the radiation of the cells on themselves, or by the radiation of other cells, depending on the arrangement of the experiments. Three other explanations are possible besides the one that he gives. Eijkman (1912) has shown that the order of spore germination is logarithmic. The recovery from previously unfavorable environmental conditions which stopped budding is also probably logarithmic, so that the numbers would pass through a maximum as the logarithmic order was decreased by the increasingly unfavorable conditions of the small, single drop environments. Another possibility is that the presence of more cells may reduce the lag period by some chemical effect of the aggregation, of the nature discussed by Allee (1930). This is less likely, since Peskett (1927) was unable to demonstrate any allelocata- lytic effects on the growth of yeast. Possible effects of a chemical con- ditioning of the medium should be tested for when large concentrations of cells are present, before a decreased lag period can be attributed to the effect of autoradiation. Baron did not report any counts of the cells present so a more com- plete analysis of this experiment is not possible, nor is it clear that he worked with homogeneous material derived from a single cell isolation. It is possible that his denser suspensions used up the food in the drops more rapidly and that and the resulting increase of toxic waste products brought about the diminution of budding which he observed. Slator (1921) has stated that a considerable concentration of cells will fail sometimes to come out of the lag period and increase their number. A combination of these effects could give Baron's results, and it is be- lieved his observations can be accepted only when he measures the changes in the environments to demonstrate that the environmental changes have no effect that might be attributed to radiation, or better, maintains the environments effectively constant except for the radiation of the experimental cultures. The changes in the environment as cells grow when the medium is not renewed have been measured for larger environments, but not as yet for the smaller environments. When the environment is restricted to a single drop, these changes must regulate the growth of the yeast more effectively than with larger environments. It is probable that failure to avoid the effects of such changes accounts for the difficulties that have prevented direct comparison of experi- mental findings with hanging drop and agar block cultures. Clark (1922) varied the seeding from five cells to eight million cells per cubic centimeter with a strain of 5. cerevisia different from that used in our experiments and found no significant differences in the rate 126 O. W. RICHARDS AND G. W. TAYLOR of growth of his populations that could be attributed to a mutuoinduction of the cells upon each other. Richards (1932a) found that the rate of growth, with the same strain of yeast, was the same when the initial seeding was varied from fifteen thousand to one and one-half million yeast cells per cubic centimeter of medium. Streline (1929) finds an initial stimulation of growth due to mito- genetic radiation which rapidly disappears. He emphasizes the need for counts of the number of cells present and for liquid culture medium. Streline uses Nadsonia julvescens, which has a different form of bud- ding from the yeast reported in this paper, and in reporting his experi- ments counts two small buds as one cell. This prevents further analysis of his data, and without knowledge of the mortality of his cells it is not possible to analyze the interesting disappearance of the stimulation. He, too, reports some negative cases. IV When there is no killing of the cells in the population, any stimula- tion of cell proliferation would give greater differences between the size of the control and the experimental populations. For instance, had there been a 10 per cent increase of cells in the experiments reported in Fig. 3, curve A, during the first two hours that had continued, the in- crease that would have occurred is shown by the broken line, Fig. 3, curve B. Had there been an initial stimulation that had not continued, then the growth curve would have been a line above and parallel to the growth curve of the control. The figure demonstrates that no such stimulation occurred in our cultures. In our experiments the environment was maintained effectively constant, except for the possible radiation of the experimental cultures, and the actual variation in both the number of cells present and the number of buds and in the death rate was measured. From these data the rate of growth and the percentage of budding were calculated. These carefully controlled experiments, using S. ccrcvisioz and ellip- soideus for detectors, and onion root, two species of bacteria — V . phos- phorescent and P. twnefaciens — and the two above-mentioned yeast as generators, show no stimulation of the multiplication of yeast by mito- genetic radiation that might pass through quartz but not through glass. It is not possible to compare our experiments closely with those ob- tained by other investigators because the latter have not given their com- plete data. Only when the complete evidence, as has been indicated in Sections I and III of this paper, is given together with a definite accel- eration of the multiplication of yeast by mitogenetic radiation, can a MITOGENETIC RAYS 127 • theory of mitogenetic radiation be established. The formula for the percentage of induction proposed by Gurwitsch is inadequate and un- satisfactory because it too greatly exaggerates the normal variation of the growth of yeast.4 SUMMARY Yeasts (Saccharomyces ccrwisice and ellipsoideus) grown in a liquid medium which was maintained effectively constant in quartz and in glass containers, were exposed to supposedly potent sources of mitogenetic radiation : bacteria ( Vibrio phosphor esc ens and Phytomonas tumefaci- ens}, onion roots, and the two above-mentioned species of yeast. In 58 per cent of the experiments, negative results were obtained, while in 42 per cent there wras a slight positive variation which did not exceed normal control variations. No effects of a mitogenetic radiation could be detected in the exposed yeast. In all cases the percentage range of variation of the experimental cultures was within the range of normal variation. It is further shown that the formula used by the Gurwitsch school in calculating the induction effect is unsatisfactory because it exaggerates the variations found in the normal cultures until they appear as induc- tion or inhibition effects. The necessity of an effectively constant environment which does not restrict the potentially unlimited growth of the yeast is stressed. A constant environment avoids the difficulties of measurement and evalua- tion of the unfavorable effects of changing environments in experiments designed to test the effect of mitogenetic radiation or other stimulants of yeast growth. Thus certain conditions and criteria are established which, it is be- lieved, the proponents of the theory of mitogenetic radiation cannot legitimately disregard in their attempt to establish the existence of such a radiation when yeast cultures are used as detectors. BIBLIOGRAPHY ALLEE, W. C, 1930. Animal Aggregations. Chicago. BARON, M., 1930. Analyse der mitogenetischen Induktion und deren Bedeutung in der Biologic der Hefe. Plant a, 10: 28. BORODIN, D. N., 1930. Energy Emanation During Cell Division Processes (M- rays). Plant Physlol, 5: 119. BORODIN, D. N., 1931. Biological Spectrum and M-rays. Collecting Net, 6: 274. CLARK, N. A., 1922. The Rate of Formation and the Yield of Yeast in Wort. Jour. Phys. Chcm., 26: 42. EIKMAN, C., 1912-13. Verslagen k. Akad. Wetensch., Amsterdam, Wis-en Natmirk., 21 : 507. (From Rahn.) 4 If the conclusions of Rahn (1932) are correct, this normal variation never can be avoided completely. 128 O. W. RICHARDS AND G. W. TAYLOR GURWITSCH, A., 1925. Methodik der mitogenetischen Strahlenforschung. Abdcr- haldcn's Handb., Abt. 5, T. 2, H. 13: 1401. GURWITSCH, A., 1931. Die fundamentalen Gesetze der mitogenetischen Erregung. Arch. Exper. Zellforsch., 11: 1. HOLLAENDER, A., AND E. ScHOEFFEL, 1931. Mitogeiietic Rays. Quart. Rev. Biol., 6: 215. PESKETT, G. L., 1927. Studies on the Growth of Yeast. III. Biochem. Jour., 21: 104. RAHN, O., 1932. A Chemical Explanation of the Variability of the Growth Rate. Jour. Gen. Physiol., 15: 257. RICHARDS, O. W., 1928a. Potentially Unlimited Multiplication of Yeast with Constant Environment, and the Limiting of Growth by Changing En- vironment. Jour. Gen. Physiol., 11: 525. RICHARDS, O. W., 1928b. Changes in Sizes of Yeast Cells during Multiplication. Dot. Gas., 86: 93. RICHARDS, O. W., 1928c. The Rate of the Multiplication of Yeast at Different Temperatures. Jour. Phys. Client., 32: 1865. RICHARDS, O. W., 1928d. The Growth of the Yeast Saccharomyces cerevisiae. I. The growth curve, its mathematical analysis, and the effect of temperature on the yeast growth. Ann. Bot., 42: 271. RICHARDS, O. W., 1932a. The Second Cycle and Subsequent Growth of a Popula- tion of Yeast. Arch. f. Protist. In press. RICHARDS, O. W., 19326. The Increased Growth of a Population of Yeast with Inosite. Proc. Soc. Exper. Biol. Med., 29: 627. RICHARDS, O. W., AND P. M. ROOPE, 1930. A Tangent Meter for Graphical Dif- ferentiation. Sci., 71: 290. SCHRIEBER, H., AND A. LUNTZ, 1931. Mitogenetische St'rahlung. Tabulae Biol. Period., 1 : 285. SLATOR, A., 1921. Yeast Crops and the Factors which Determine Them. Jour. Chem. Soc., 119: 115. STRELINE, G. S., 1929. [In Russian.] Vestn. Rcntgenol, 7: 191. TAYLOR, G. W., AND E. N. HARVEY, 1931. The Theory of Mitogenetic Radiation. Biol. Bull., 61 : 280. WALLACE, G. L, AND F. W. TANNER. 1928. Studies on the " Bios Question." Centrlbl. f. Bakt. u. s. «•., 76: 1. RATE OF REGENERATION OF PARTLY HISTOLYZED ANURAN TAIL SKIN * H. J. CLAUSEN DEPARTMENT OF BIOLOGY, UNIVERSITY COLLEGE, NEW YORK UNIVERSITY INTRODUCTION Recent studies on anuran integument and muscle transplantations (Helff, 1926; Lincleman, 19290 ; Helff and Clausen, 1929; and Clausen, 1930) all suggest the probability that the causal histolytic influence of the various tissues of the tail may be due to specific substances in the blood stream or a general lowering of the pH of the blood during meta- morphosis. It has been suggested (Lindeman. 1929/7) that this influ- ence is essential, not only as an initiatory agent, but must also be present continuously during tail resorption. This was demonstrated by the transplantation of histolyzing anuran tail-skin to normal individuals, resulting in the normal reconstitution of the graft in question. The various intermediate regenerative stages were not described, however. Part of the present paper is devoted, therefore, to a description of the normal histological regenerative processes, which occur in tail integu- ment following partial histolysis. In a previous paper (Clausen, 1930) a susceptibility gradient to histolysis was demonstrated for various regions of tail-skin. In this connection it was clearly shown that integument derived from anterior regions of the tail undergoes a more rapid rate of histolysis than is true of integument derived from more posterior regions. It was, conse- quently, thought of interest to determine whether or not any differences existed between the rates of regeneration of tail-skin derived from vari- ous anterior-posterior levels of the tail. Work on amphibian tail re- generation by several investigators seems to have given rise to rather diverse results and interpretations. Morgan (1906). working with the salamander Dicinyctylits riridesccns, and Ellis (1908, 1909), on the anuran Rana clainitaiis, seem to agree in a general way that the rate of regeneration of new portions of the tail is directly proportional to the distance the tail is cut off from the tip. In other words, the more distal the cut, the slower the regeneration. Conversely, Comes (1928), work- ing with the anuran Discoglossus pictus, states that regeneration of the 1 This paper, together with Parts I and II of this series of studies, was sub- mitted to the Graduate School of New York University in partial fulfillment of the requirements for the degree of Doctor of Philosophy, April 1, 1932. 9 129 130 H. J. CLAUSEN tail fin diminishes anteriorly and with the age of the animal. Speidel (1929), however, using several types of anurans and salamanders, re- ports no differences in the rate of resorptive or regenerative processes of the tail fin at anterior, middle, or posterior levels following thyroid treatment. Since the investigations of Morgan and Ellis on tail regen- eration concern regeneration of all tissues of the tail, and the work of Comes and Speidel concerns the rate of tail fin regeneration in situ, it seemed of interest to the writer to inquire into the possible regenerative rate differences of only one type of tail tissue as might be demonstrated by transplantation to foreign regions. By so doing, it was thought possible to demonstrate whether a difference in regenerative rate was due to the association of the tissue with a definite level of the tail and adjacent tissues or whether the particular tissue in question possesses a specific rate of regeneration typical of integument of that region. As far as the writer is aware, the present paper constitutes the first attempt to explain the above points, especially as regards the rate of regenera- tion of histolyzing tail-skin taken from various levels of the tail. The general methods of attack which were followed in the present investigation were briefly as follows: The integument from four levels (anterior-posterior) of the tail of normal larvae was transplanted, auto- plastically, to the back. Following this procedure, artificial metamor- phosis was induced and the engrafted integument removed at certain stages of histolysis and transplanted, homeoplastically, to the back or tail of normal individuals. Subsequent macroscopic and microscopic observations were made of the various transplants at certain definite intervals as regeneration progressed. The following series of trans- plantations, which will be described in detail in the respective sections, were performed : Series 1. Autoplastic skin transplantations from four regions of the tail to the back. (One hundred and twenty cases of four trans- plants each.) Series 2. Skin transplants (previously transplanted in Series 1 and in various stages of histolysis) transplanted homeoplastically to the tail or back. (Two hundred and twenty-five cases of two trans- plants each.) Series 3. Homeoplastic skin transplantations of four regions of the tail of normal larvae to the tail or back of normal larvae. (Thirty- two cases, control series.) This investigation was undertaken at the suggestion of Dr. O. M. Helff, to whom the writer wishes to express his appreciation for his helpful suggestions, kindly advice, and criticism during the course of the work. REGENERATION ANURAN TAIL SKIN 131 » MATERIAL AND METHODS The stock used for all operations was Rana pipicns larvae obtained from ponds in the vicinity of Cold Spring Harbor, Long Island, during the months of June and July, 1931. The animals were taken to the laboratory and placed in a large tank supplied with fresh running water. It was found advisable to let the stock remain in the large tank from four to five days under laboratory conditions before selecting individuals to be used for experimental purposes. By doing this, only those indi- viduals which survived the new environmental conditions would be se- lected. These animals ranged from 70 to 90 mm. in body length with hind limbs between 3 and 12 mm. in length. All experimental indi- viduals were normal larvae in all respects and remained as such until induction of artificial metamorphosis following the autoplastic opera- tions. All operated animals were kept in individual aquaria and main- tained under constant laboratory conditions. With the exception of a small number, it was found necessary to anaesthetize all animals used for transplantation purposes. The animals were anaesthetized in a 0.05 per cent aqueous solution of chloretone for five to ten minutes, depending on individual differences and the tempera- ture of the solution. These anaesthetized animals remained inactive sufficiently long to carry out operative procedures and insure adhesion of the transplants and their cut edges. In the first series of experiments, autoplastic transplantations were made consisting of the removal of tail-skin from regions 1 and 3 or from regions 2 and 4 (text figure A). These portions of integument were then carefully shaped into rectangles measuring approximately four millimeters in width and seven millimeters in length. Two similar- sized and -shaped rectangular portions of integument from the medial region of the back were next removed and the tail-skin transplants from regions 1 and 3 or 2 and 4 were then transplanted to the denuded areas of the back. As regards the orientation of the grafts, the linear se- quence of transplants number 1 and 3 or 2 and 4 were alternated in 50 per cent of the cases. In other words, graft number 1 or 2 was grafted to the more anterior wound area, leaving graft number 3 or 4 to be trans- planted to the remaining posterior wound area ; while in other cases 132 H. J. CLAUSEN graft number 3 or 4 was placed more anteriorly. After the above trans- plantations were made, the animal was placed in a Petri dish containing a small amount of water. This procedure allowed the grafted areas to be exposed to the air and hastened the adherence of the cut edges. Following this, the animals were placed in individual aquaria containing water which was kept at a temperature ranging from 18° to 20° C. After five days, the time allowed for sufficient healing of the graft to take place, precocious metamorphosis was induced by the feeding of desiccated thyroid. During larval transformation, daily macroscopical observations were made of the above individual transplants to note the reduction in area of the grafts. When the grafted integument had been reduced in area approximately eighteen per cent, a number of grafts originally derived from tail regions 1, 2, 3, and 4 were removed and transplanted, homeo- plastically, to normal larvae. The same procedure was carried out for a like number of grafts when the reduction in area had reached approx- imately 50 and 80 per cent. In making the homeoplastic transplanta- tions, the histolyzing graft was removed and placed on previously de- nuded areas of the back or lateral regions of the tail of normal larvae. Following these transplantations, the normal larvae were again placed in individual aquaria and kept as such under constant laboratory conditions. At intervals of 8, 11, 14, 17, 20, 23, and 26 days, following the homeo- plastic operations, representative individuals possessing the transplanted integument of each of the four tail regions which had previously been reduced in area by 18, 50, or 80 per cent, were removed from their indi- vidual aquaria and preserved in toto in Benin's Picro-Formol fluid for subsequent sectioning and histological study of the grafts. A second series of homeoplastic transplantations was also made. The technique consisted of removing rectangular portions of tail-skin from regions 1, 2, 3, and 4 of normal larvae. These normal skin trans- plants were then placed on denuded areas of the back or tail of other normal larvae. Individuals of this series were also preserved for histo- logical study of the transplants at 8, 11, 14, 17, 20, 23, and 26-day inter- vals following transplantation. This series was designed as a control to make certain that histolysis does not occur in larval skin following homeoplastic transplantation to normal larvae. RESULTS The Process of Degeneration and Regeneration in Tail Integument 1. The normal integument. — The normal tail integument in the par- ticular species studied (Rana pipiens) presents a macroscopical picture differing in appearance from integument of body regions. It appears REGENERATION ANURAN TAIL SKIN 133 I much more translucent and is also decidedly more delicate in texture, being injured and torn more readily than integument of the side or back. It is also attached more rigidly to the underlying musculature as com- pared with hack-skin. The pigmented areas are regularly distributed over the entire tail region and appear to be less intense in depth of shade as compared with similar areas of the back. As to histological structure, tail-skin is composed of an epidermis, a corium, and a layer of subcutaneous connective tissue. The epidermis is composed of several layers of cells, the outermost of which consists of a single layer of small, flattened cells forming the cuticle. The more basal cells of the epidermis are for the most part cuboidal in shape with round or oval-shaped nuclei. The epidermal cells basal to the cuticle are usually arranged into two or three irregular layers. The corium, which is separable into two definite layers in most regions of the body, differs markedly in this respect in tail integument. The outer, com- paratively loose layer (stratum spongiosum), so typical of side and back integument, is almost obliterated in tail integument. This layer is repre- sented, however, by a thin layer of pigment cells lying directly beneath the epidermis. The lower layer of the corium (stratum compactum) is composed of dense connective tissue, the fibers running in a wave-like course parallel to the surface. The characteristic much convoluted ap- pearance, typical of this layer in body regions, is absent in tail integu- ment. This layer is, moreover, much thinner as compared with the condition found in back integument. These corionic modifications, typical of tail-skin, are responsible for the latter's relative thinness and more delicate structure as compared with integument of body regions. The subcutaneous connective tissue forms a loose layer underneath the stratum compactum. This layer is quite vascular, containing large numbers of blood cells, chiefly lymphocytes. It is also considerably thinner than the subcutaneous connective tissue layer beneath back integument. 2. The degenerative process. — The same histolytic characteristics are exhibited for all four anterior-posterior transplants of tail integument when the degree of resorption, as measured by reduction in surface area, is the same. Therefore, in the explanations to follow, the process of degeneration will be described only as observed in tail-skin from region number 1 (text figure A). When the transplant had been reduced in surface area 18 per cent, no apparent macroscopical differences as compared to normal integument of the tail could be noted. The pig- mentation remained normal throughout this period of resorption (Fig. 3). However, when the transplants had been reduced 50 and 80 per cent, respectively, a decided change in depth of shade could be noted. 134 H. J. CLAUSEN In other words, the greater the degree of histological disintegration, the darker in color the transplant became (Figs. 1 and 2). This increase in depth of coloration proceeded from the margins of the transplant toward the center as histolysis progressed. In examining histological sections of the tail-skin transplants at various stages of area reduction, the first signs of histolysis are con- cerned with the structure of the stratum compactum. When the trans- plant had been reduced in surface area 18 per cent, the stratum com- pactum had lost its wavy appearance ; the fibers becoming more or less dissociated so that a very disorganized picture is presented (Fig. 5). With this dissociation, lymphocytes make their appearance, apparently migrating from the subcutaneous tissue into the compactum layer. These lymphocytes are particularly abundant at the edges of the trans- plant, where fusion has occurred with the surrounding integument of the back. At this stage of transformation the epidermis appears to be thickening, especially near the central region of the graft. When the engrafted integument had undergone resorptive changes amounting to 50 per cent of its original surface area, the stratum com- pactum was entirely obliterated leaving only scattered pigment cells, lymphocytes, subcutaneous connective tissue, and the epidermal layers. An interesting histological feature of the graft in this stage of histolysis is the increased thickness of the epidermal portion, which appears to be several layers thicker at this time (Fig. 6). The basal layer of this PLATE I Explanation of Figures FIGS. 1-5. BI, adjacent back integument; P, pigment masses; TST, tail-skin transplant; E, epidermis; SS, stratum spongiosum ; SC, stratum compactum; CT, subcutaneous connective tissue; L, lymphocytes; TT, boundaries of transplant; BV , blood vessel. FIG. 1. Alacroscopic appearance of tail-skin graft histolyzed 50 per cent its original surface area. FIG. 2. Macroscopic appearance of tail-skin transplant histolyzed 80 per cent its original surface area. FIG. 3. Macroscopical picture of normal tail-skin transplant as it appears five days following autoplastic transplantation. FIG. 4. Histological section through autoplastic tail-skin transplant and ad- jacent back-skin at a time when the graft had histolyzed 80 per cent (surface area reduction). The stratum compactum is obliterated. The epidermal layer is be- ginning to dissociate and lymphocytes are migrating into this particular region. There is also a slight condensation of pigment cells basal to the epidermal portion of the transplant. FIG. 5. Histological section through autoplastic tail-skin graft and adjacent back integument when the transplant had histolyzed 18 per cent. The stratum compactum has lost its wavy appearance and is becoming more or less dissociated. Subsequent lymphocytic invasion is shown following this dissociation. The epi- dermal portion of the graft appears normal at this time. REGENERATION ANURAN TAIL SKIN 135 1 P TST PLATE I •••it* BI 2 BI TST V te: i,'^ },''<'' « •: =%* • ' -;.r'%v m P TST BI m E > ;.>;'"- '^'""/"-^"V^ 136 H. J. CLAUSEN thickened epidermis contains cells which have hypertrophied and as- sumed a more columnar shape. In sections of tail-skin grafts, where the original surface area had been reduced approximately 80 per cent (Fig. 4), the previously thickened epidermis, as illustrated in Figure 6, is now very much thinner. Sections show a separation and basal mi- gration of groups of these epidermal cells. With this epidermal dis- similation there is a gradual invasion inward of all layers of the sur- rounding back integument. This invasion probably places the graft under a certain degree of peripheral pressure. The lymphocytic cells are now found scattered throughout all the various layers of the graft, including the epidermis. The cuticle is still intact at this stage and ap- parently is the last layer to exhibit a histolytic change. The above histological and macroscopical findings show that the integumentary grafts exhibit histolytic processes which are identical with those typical of histolyzing tail integument, in situ, during normal metamorphic tail atrophy. In this connection, it may be stated that a more complete study of the histological processes which take place dur- ing histolysis was not attempted ; only those sections taken at definite periods, when surface area had been reduced 18, 50, and 80 per cent, having been studied. It may be stated here, however, that the micro- scopical pictures of tail-skin histolysis as thus observed compare favor- ably with those of HeliT (1926) and of Lindeman (1929a). PLATE II Explanation of Figures FIGS. 6-10. Histological sections through back integument including homeo- plastic tail-skin transplants originally derived from the most anterior regions of the tail. The sketches illustrate successive steps in the histological regeneration of the transplants which had previously undergone partial histolysis following autoplastic transplantation. E, epidermis ; SS, stratum spongiosum ; SC, stratum compactum ; CT, subcutaneous connective tissue; P, pigment masses; TT, boundaries of the transplant; L, lymphocytes; BV , blood vessel. FIG. 6. Section through autoplastic tail-skin graft and adjacent back integu- ment showing condition of transplant when histolyzed 50 per cent (surface area reduction). The stratum compactum is nearly obliterated. The epidermis is in- creased in thickness with its more basal cells somewhat hypertrophied. FIG. 7. Eight days following homeoplastic transplantation. The first signs of regeneration as shown by a condensation of the epidermis. FIG. 8. Eleven days following homeoplastic transplantation. Basal cells of the epidermis breaking off and separating from the main portion of that layer. FIG. 9. Fourteen days following homeoplastic transplantation. Epidermis still dissociating in the basal layers. Evidence of a stratum compactum layer, its fibers being more or less loosely associated. FIG. 10. Seventeen days following homeoplastic transplantation. Histologi- cally, the transplant appears as normal tail integument. REGENERATION ANURAN TAIL SKIN 137 PLATE II T P V «)«" *' }• •- .• <£\f;:"~ -E So j^ -sc >• >rj — CT T ^___ T ^gSggS*i£jfite8s]K^t£g%&^. "^^^^-fe- • ' ' T v : *.:;• . | \f- -.•- " " • ^V® V ^ m - 7 '" • / ^ .. • ,. e — ss .V X f BV 'J' — CT ^g^^s^^ r-s-^P- T P-' - - . j>* ,.-> SC CT 9 sgSS5KSSSX»5Bei- -«T-- ^*?> ^c|-^; -^"f— E •„ " * ,- "/ - ^-- - 1" : T ' " " — ss j. - ~ --—sc 'tcnis eggs in metaphase are fixed with Boveri's reagent there is a definite relation between central body structure and 4 It is difficult to select concise terms that describe clearly the differences in the coarseness of rays and in their shape. For example, the use of "very coarse" in contrast to '' medium coarse " is somewhat clumsy, but it is necessary to express two degrees of coarseness as contrasted with the condition described by the term " delicate." The terms " rippled " and " undulating " were selected because the former suggests the idea of short sharp curves, in contrast to the long gentle curves suggested by the latter. 5 Even among cells in metaphase which are alike in the general structure of rays and central bodies there is considerable variation in the size of the center It is practically impossible to secure accurate measurements in most cases, because the central body is demarked from the ray area either vaguely or not at all. In each type, therefore, the cell selected for illustration has a central body of about the average size for its class, based on attempted measurements of about twenty figures of that type. There is also much variation in the shape of the centers : some are round, others quite elongate, others intermediate. The cells selected for illustration as representative for each type have central bodies that are intermediate in shape, being but slightly elongate. 156 HENRY J. FRY ray structure : all asters with rippled rays have disrupted centers that are either empty or scattered ; all those with undulating rays have even centers that are homogeneously rilled. There are no exceptions to this relationship. Asters with slightly rippled rays, which are intermediate between these two classes, show all types of centers. CLASSES OF CENTRAL BODIES Empty Scattered Even CLASSES OF RAYS Rippled : 62 Asters 66 Asters Slig-htly Rippled 74- Asters 34 Asters 3O Afters Undulating 57 'Art* •ers CHART 1. THE RELATION BETWEEN RAY STRUCTURE AND CEN- TRAL BODY STRUCTURE UNDER OPTIMUM CONDITIONS Central bodies were studied in metaphase first-cleavage figures in Chaetopterus eggs, after using Boveri's picro-acetic fixation, and the standard slide-making procedure. Result: Central body structure is related to ray structure : centers are " disrupted ", i.e., empty or containing scattered material, if rays are rippled in shape; centers are "even", i.e., homogeneously filled and stained like the ray area, if rays are undulating. Centrioles are not demonstrated. The slightly rippled rays are obviously intermediate between rippled and undulating rays. Since typically rippled rays are always associ- ated with disrupted centers (empty or scattered) and typically un- dulating ones with evenly filled centers, the members of this inter- mediate, or slightly rippled group will hereafter be included with either the rippled or the undulating class, depending upon which group the aster most resembles. Thus the four classes of asters illustrated in Chart 1, in which the centers are empty or scattered and rays either CHAETOPTERUS CENTRAL BODIES 157 rippled or slightly rippled (Figs. 2-5) will hereafter be regarded as but minor variations of a single major type, and represented in later charts by an illustration showing a disrupted (scattered) center and rippled rays (Fig. 4). Similarly, the two classes with even centers and rays either slightly rippled or undulating (Figs. 6 and 7) will hereafter be regarded as variations of another major type, represented in later charts by an illustration showing an even center and undulating rays (Fig. 7). These two major types of asters are but two of twelve that occur in this investigation, all of them illustrated in Chart 7 (p. 177). The type with disrupted centers is designated as IB and that with even centers as 2B. The basis of classification is discussed on page 176. Very rarely, i.e., in 16 of the 243 asters studied, one or more granules, like the smaller ones present in the cytoplasm, occur in the even type of center. But since they vary in size, location, and staining capacity, they could not be interpreted as centrioles. These random granules are not the structures Mead illustrated ; the kind of central body he described was not produced in this experiment, in which Boveri's picro-acetic reagent was used in the usual way. Experiment 2. The Effects of Uncontrolled Factors in the Handling of Different Egg-sets Under Optimum Conditions The purpose of this experiment was to learn whether the percent- ages of disrupted and even centers (Types IB and 2B~] occurring in one set of eggs handled under optimum conditions are the same as those occurring in other egg-sets run under supposedly similar conditions but on different days. Table I shows counts of these two types of centers in five different sets, each sample including from 33 to 67 eggs. These data are arranged in the order of increasing percentages of the dis- rupted type. When Sets 1 and 2, 2 and 3, 3 and 4, and 4 and 5 are compared, there are only minor differences in the percentages of the two central body types present, and these could be explained by the relatively large errors always involved in reporting small samples. But when Sets 1 and 3. 2 and 4, and 3 and 5 are compared, the differences are large enough to suggest some cause beyond error of sampling. And when Sets 1 and 4, 1 and 5, and 2 and 5 are contrasted it is obvious that the discrepancies are too great to be explained by errors of sampling alone ; they must be due either to differences between the living eggs of the various sets prior to fixation, or to uncontrolled modifications of tech- nique. When these relations are analyzed statistically, by determining the value of P according to Pierson's method (1924, p. Ixx) the ex- 158 HENRY J. FRY istence of factors which cause differences, other than errors due to small samples, is convincingly demonstrated.6 The approximate values of P for this material are shown in Table I. Three problems present themselves : ( 1 ) to secure more data con- cerning the relation between central body structure and ray structure ; (2) to explain the cause of the difference in the relative numbers of disrupted and even centers present in different sets after picro-acetic fixation under optimum conditions; and (3) to ascertain what varia- tion of the picro-acetic technique Mead used to demonstrate typical centrioles. The experiments which follow were planned in an attempt to secure information on these points. GROUP B. TECHNIQUE MODIFIED: SLIDE-MAKING PROCEDURE VARIED; BOVERI'S PlCRO-AcETIC FIXATION Experiment 3. TJic Effects of Varying tJie Depth of Stain (Heiden- liain's Hffmatoxylin*) It is conceivable that a center containing considerable irregularly scattered material might, if lightly stained, be listed as belonging to the disrupted (scattered) type, while the same center, if darkly stained, might appear to be evenly filled and would then be counted as an even type. In that event, the differences in the relative numbers of dis- rupted and even centers in the five sets of the previous experiment might be due to differences in depth of stain. In that experiment the eggs were stained a deep blue color with Heidenhain's hsematoxylin. Table II repeats the data of that experi- ment at the left, and also shows, at the right, the numbers of disrupted and even centers which occurred when eggs of the same five sets were stained a very pale blue. In each set all experimental conditions were identical, in so far as they could be controlled, except the depth of stain. 6 Pierson's method determines whether a given sample of objects (the 67 eggs of Set 1) having a given number of one class (40 disrupted centers) and a given number of a second class (27 even centers) does or does not belong to the same population as another sample (the 55 eggs of Set 2) with different numbers of the same classes (36 disrupted centers and 19 even centers). These relations' are ex- pressed in terms of P. Thus the value of P for Sets 1 and 2, as well as for Sets 2 and 3, is 0.4, meaning that there are 4 chances out of 10 that each pair belongs to the same population, and indicating that the differences between them are not sig- nificant. At the other extreme, however, the value of P for Sets 1 and 4 is 0.02, and for Sets 1 and 5 it is 0.001, showing that the chances are, respectively, one out of 50 and one out of 1000, that each pair belongs to the same population. The differences between these sets are therefore significant. CHAETOPTERUS CENTRAL BODIES 159 TABLE I The effects of uncontrolled factors in the handling of different egg-sets under optimum conditions. Central bodies were studied in metaphase first-cleavage fig- ures in Chaetopterus eggs, after using Boveri's picro-acetic fixation, and the stand- ard slide-making procedure. Result: Some factor or factors are involved which modify significantly the numbers of the two central body types in the different egg-sets. Egg- Set Symbol Number of Cells Studied Distribution of Astral Types in Each Set P 1 B (Disrupted) 2 B (Even) 1 67 40 60% 27 40' r 01 1 O.f .01 0.4 0 0.4 2 55 36 65% 19 35% 3 48 34 71% 14 29% 0.1 0 0.1 _^ \ 4 33 26 79% 7 21% 5 39 35 90% 4 10% It is apparent, both by inspection and by the statistical determination of the value of P, that the depth of stain with Heidenhain's haema- toxylin does not modify significantly the percentages of the two central body types occurring in any one egg-set after Boveri's picro-acetic fixa- tion. This is especially obvious when comparing the totals.7 7 This fact does not indicate that various depths of stain would be equally ineffective in modifying the appearance of other central body types, to be de- scribed later. 160 HENRY J. FRY TABLE II The effects of varying the depth of stain (Heldenhain's hcematoxylin) . Cen- tral bodies were studied in metaphase first-cleavage figures in Chaetopterus eggs, after using Boveri's picro-acetic fixation. The slide-making procedure was stand- ard except that egg-sets reported at the left in the table were stained darkly and those at the right were stained lightly. Result: Differences in depth of stain do not significantly modify the numbers of the two central body types occurring in the different egg-sets. EGG-SET SYMBOL EGGS STAINED DARKLY P EGGS STAINED LIGHTLY Number of Cells Studied Distribution of Astral Types in Each Set Number of Cells Studied Distribution of Astral Types in Each Set 1 B (Disrupted) 2 B (Even) 1 B (Disrupted) 2 B (Even) 1 67 40 60% 27 40% *-0.8-.. 36 21 58% 15 42% 2 55 36 65% 19 35% — 0.9-* 42 27 64% 15 36% 3 48 34 71% 14 29% ^0.6-* 44 33 7CP7 '•J 7o 11 25% 4 33 26 79% 7 21% ^0.7-+ 48 39 81% 9 19% 5 39 35 90'; 4 10% *-0.1-» 73 61 84% 12 16% TOTALS 242 171 n< • i 71 29% 243 181 74% 62 26% Experiment 4. The Effects of Other Modifications of the Slide-making Procedure Wide variations as to the speed with which the eggs are passed through the alcohols and xylol, the temperature at which they are em- bedded, and the thickness at which the ribbon is sectioned, do not change the appearance of the central bodies when Boveri's picro-acetic reagent and Heidenhain's haematoxylin are used. The effects of modifying one other phase of the slide-making proced- ure were also studied — accidental overheating of the ribbon when it is warmed to bring about its expansion. Different samples of ribbon were variously treated during this operation : in one the ribbon was ex- panded in the usual manner, remaining opaque without melting through- out the operation ; another was so overheated that the ribbon melted CHAETOPTERUS CENTRAL BODIES 161 and became transparent for a moment ; in a third the same situation was produced but maintained for a longer time ; finally, one slide was so extremely overheated that the ribbon not only melted but bubbled and gave off vapors. In the latter case a number of the eggs have a peculiar glassy appearance in the region of the spindle, the inner zone of the astral rays, and the central body area (Fig. 8). This type, pro- duced by faculty technique, is readily recognized and could not be con- fused with the types occurring under usual conditions. ^^i^^^^^m £ .-•; ;; ' W- • V,-j''; /.-'T-V^.v,- -"<-': -* SvMllisiiiif FIG. 8. The glassy type of astral center caused by overheating the ribbon when expanding it. Central bodies were studied in metaphase first-cleavage figures in Chactoptcnis eggs, after using Boveri's picro-acetic fixation. The slide-making procedure was standard, except that the ribbon was overheated when expanding it. Result: The central body area, the inner part of the ray region, and the spindle have a glassy appearance. It can therefore be concluded that when Chactopterus eggs are fixed with Boveri's reagent, the inevitable slight variations in the slide-making procedure do not modify central body structure. While this conclu- sion applies only to the disrupted and even types, it is probable that it also holds true for the other types to be described. GROUP C. TECHNIQUE MODIFIED: PICRO-ACETIC FIXATION VARIED; REGULAR SLIDE-MAKING PROCEDURE Experiment 5. The Effccis of Varying Simultaneously the Amounts of Picric and Acetic Acids When eggs are added to the fixative in a vial, even the densest egg suspension includes some sea water, which slightly dilutes the reagent. The present experiment deals with the effects of a series of dilutions, beginning with one part of distilled water and 99 parts of reagent, and ending with 99 parts of water and one part of reagent. To test-tubes 162 HENRY J. FRY containing 25 cc. of each dilution were added 0.5 cc. of eggs. The amount of sea water included with the eggs was so small, compared with the 25 cc. of diluted fixative, that the effects of any further dilution could safely be ignored. The eggs fixed at 99 per cent, 90 per cent, 75 per cent, 50 per cent, 10 per cent and 1 per cent strength were from one egg-set; those fixed at 30 per cent, 20 per cent, 15 per cent, and 5 per cent were from another egg-set. Chart 2 shows the percentages of the central body types occurring after fixation with each dilution. When the reagent is from 99 per cent to 50 per cent strength the disrupted and even types previously described (IB and 25) are the only ones present, their numbers vary- ing at the different dilutions. At 30 per cent strength two new types appear, one having a vaguely delimited centrosome stained darker than the ray area, but without a centriole (4A), and one with a similar centrosome but containing one or two centrioles (5 A). Between 30 per cent and 5 per cent strength these are the major classes. At 5 per cent strength a new type of aster makes its appearance, similar to the one with an even center so frequently described except that it has very delicate rays (2C). At 1 per cent strength the asters are very small and vague and their centers are undifferentiated from the ray area. Cytasters are also present,8 and the chromosomes are abnormal. The relation between ray structure and central body structure is again obvious, and additional evidence on this point is supplied by the new types. (1) If rays are fixed vaguely, the central area of the aster is entirely undifferentiated from the peripheral part (2D). (2) If rays are distinct and rippled the center is disrupted (15). (3) If rays are distinct but undulating, the astral center is filled, its structure vary- ing with the coarseness of the rays, as follows: When rays are either delicate or medium coarse, the center is a homogeneous centrosome, about 5 X 6/. ii^i - |l ;--,V-"f M'1^-1^^"; • '.'"J.'.Vv1 '.*' .V " v 99 1 34 32 94% 2 6% 30 1O 34 29 65% ^ 15% 75 25 46 3J 67% 15 33% 50 50 82 80 96% 2 2% 30 7O 59 5 13% 5 13% 72 37% 17 43% 20 80 52, 6 .2'/0 6 12°/0 19 36% 21 40% 15 55 41 I 2% 4 10% /o 25% 25 61% / 2% 10 90 41 6 73% 41 87% 5 95 4? 19 40% IT 36% 11 24% 1 39 40 40 200% CHART 2. THE EFFECTS OF VARYING SIMULTANEOUSLY THE AMOUNTS OF PICRIC AND ACETIC ACIDS Central bodies were studied in metaphase first-cleavage figures in Chaetopterus eggs-, after fixation with Boveri's reagent diluted to various degrees with distilled water. The slide-making procedure was standard. Results: (1) Central body structure is related to ray structure, depending upon the coarseness and shape of the latter. (2) Mead (1898) demonstrated centrioles by diluting the reagent. (3) Uncontrolled dilution effects, resulting from the manner in which eggs are added to vials under usual conditions (Exp. 2) are probably responsible for the variations in numbers of the central body types in different egg-sets. 164 HENRY J. FRY fessor Mead, suggesting that possibly a comparable dilution might have been responsible for the centriole phenomena he found, and hence for the writer's failure to repeat the work with a full-strength reagent. Professor Mead replied that dilution did occur in his experiments, that the results in different preparations were not uniform, and that in many eggs the centers were disrupted. His failure to mention this dilution in his paper is undoubtedly due to the fact that he attached no significance to it — a situation which occurs almost universally in the case of many supposedly minor variations in different steps of the cytological tech- nique. It is worth noting, however, that diluting the reagent is the only modification of the many employed in this investigation which demon- strates centrioles. Where Mead used sea water to dilute the fixative, distilled water was used in the present work. It is probable that both kinds of dilu- tions produce similar results, for centrioles can be demonstrated after dilution with either. It is possible, however, that certain differences may exist. Varied dilution of the reagent explains the occurrence of disrupted and even types of central bodies side by side on the same slide, as well as the variations in their relative numbers in various egg-sets supposedly fixed and run up in the optimum manner (Table I, p. 159). Uncon- trolled dilution effects, differing from vial to vial and modifying the central area in asters, are brought about by slight differences in the manner in which eggs happen to be added to the reagent in the vial. If eggs are squirted quickly from the pipette into the vial containing the reagent the mixing is practically instantaneous, but the eggs are handled rather violently. Furthermore, the dish containing the living eggs may become contaminated if the tip of the pipette, which must be held close to the vial when the eggs are added, is splashed with minute droplets of the reagent. For these reasons, eggs are usually allowed to drop gently out of the pipette. Under such conditions the eggs and any sea water accompanying them are more or less segregated near the top of the vial for the first few seconds, the reagent mixing with them from below. It will be shown later (Experiment 9) that fixation occurs within one second ; hence the eggs added to the vial are exposed during the first second to various dilutions of the fixative, depending upon whether they happen to be at the top or at the bottom of the mixture of eggs and sea water just added to the reagent. The data of Chart 2 show that with few exceptions fixation with a full-strength reagent produces disrupted centers (IB), and that progres- sively greater dilutions of the reagent result in filled centers (2B, 4A, and 5./). In the preceding experiment (Table I) such effects were in- CHAETOPTERUS CENTRAL BODIES 165 advertently produced by the manner in which the eggs of the five egg-sets were added to their vials. In Set 1 a considerable percentage of the eggs undoubtedly received a somewhat dilute fixation, since 40 per cent of them have even centers; but those of Set 5 were added in such a manner that only 10 per cent of them were so fixed. The effect of varied dilutions also explains why, under usual con- ditions of fixation, about 70 per cent of the centers are disrupted, the remaining ones being even (Table II). Since Chaetoptcrus eggs fre- quently stick to the bottom of the vial, the writer has heretofore used a narrow vial (9 mm. wide, inside diameter, and 4.5 cm. high), because in it the eggs pile up on top of each other, and few of them touch the bottom. Prior to fixation these vials contained from 4 to 4.5 cc. of reagent; at the time of fixation from 0.5 to 1.0 cc. of eggs were added. The strength of the reagent thus varied from 80 per cent to 90 per cent, the average being about 85 per cent. In the five sets fixed under these usual conditions, 71 per cent of the eggs have disrupted centers and 29 per cent even ones — a result which is comparable to that of the set fixed at 75 per cent strength in the dilution experiment (Chart 2), where the percentages are 67 and 33. In this dilution experiment the major differences in types and num- bers of central bodies occurring at any given strength of the reagent are without doubt primarily the result of the varied dilutions. Within each set, however, the numbers and proportions of central body types were probably also modified by the exact way in which the eggs were added to the reagent. Until the results of this experiment were known the writer had not realized what radical effects can be produced by dilution. Had the eggs been expelled suddenly from the pipette into the fixative instead of being dropped in gently, not only in the present experiment but in the others as well, it is probable that in many cases the numbers of types would have been reduced and their relative percentages some- what altered.9 The fact that dilution of Boveri's picro-acetic reagent causes such marked differences in the structure of rays and central bodies does not mean that this would be true of all cells fixed with any dilute reagent. However, it is probable that the cells of various species are susceptible to dilution effects in various fixatives. For example, although Echhi- arachnius eggs ordinarily show large pleuricorpuscular central bodies 9 The existence of such an uncontrolled cause of variation in central body types makes it useless to study large egg samples in each group for the sake of reducing the error inherent in small samples. For this reason the number of eggs studied at each variation is seldom more than 50, the average being 36. But although uncontrolled dilution effects may modify the percentages of any type occurring in the eggs of any vial, they do not alter the relation between central body structure and ray structure, which always holds true, without exception. 11 166 HENRY J. FRY when fixed with various full strength reagents, including Benin's, this fluid diluted with 90 parts water produces a minute centriole surrounded by a homogeneous centrosome. The second paper on Chaetopterus eggs will demonstrate that the effects of diluting certain chemically diverse reagents are similar to those of picro-acetic dilutions, although the de- tails differ in each case. Whether dilution effects have played an unsuspected role in central body studies of other species remains to be ascertained by future in- vestigations. It may be that in many eggs extensive dilution of the reagent does not modify astral structure. One egg, however, that may be affected in this way is that of Ascaris (Fogg, 1931 and many earlier papers) ; its chorion is so impermeable that it can live in certain strong reagents for hours. \Yhen the one usually used, a chloroform-alcohol- acetic mixture, does finally penetrate, it is quite possible that only a very dilute amount of the reagent as a whole, or of one of its components, actually brings about the fixation. This can be determined only if a technique can be developed whereby Ascaris eggs can be subjected to instantaneous fixation with full-strength reagents. Another egg where dilution effects probably occur is that of Dro- sopliila (\Yilson and Huettner, 1931). It too is impermeable to the reagent. Therefore it is pricked at one end to permit the fixative to enter, and thirty to sixty seconds are required for it to traverse the egg. This may explain the variations found in different mitotic figures in the same stage. In different eggs the variations may be determined by the size of the puncture, which governs the rate at which the reagent enters ; but similar effects also occur in different parts of the same egg, probably depending on various dilution effects at different distances from the point of puncture. Here again it would be interesting to compare the results produced by a technique permitting instantaneous, full strength fixation, with those of the present method. The fact that a certain configuration is produced by a given dilution of a reagent but is not shown when that fixative is used full .strength, argues neither for nor against the validity of that coagulation product. Such differences, produced by different dilutions of the same fixative, may have neither more nor less significance than differences produced by chemically diverse reagents. Nevertheless, if certain configurations are produced only by a diluted reagent, that fact must be kept in mind. The manner in which significant variations are produced in Chactop- tcnts eggs by diluting the reagent suggests that certain precautions should be followed generally in fixing eggs. The amount of reagent should be large enough to prevent any appreciable dilution by the addi- tion of any fluid accompanying the eggs. Furthermore, the eggs should CHAETOPTERUS CENTRAL BODIES 167 be added in such a way as to effect instantaneous mixing, in order to prevent inadvertent dilution effects during the first second, when the eggs are fixed. Experiment 6. Tlie Effects of J'aryini/ in Turn flic Amounts of Picric and Acetic Acids In the preceding experiment the amounts of both acids in the reagent were progressively and simultaneously reduced. The present experi- ment was carried out in order to study the effect on central body struc- ture when each acid in turn is kept constant while the other is varied. Four egg-sets, run on different days, were used in this experiment. Each illustration in Chart 3 shows the astral configuration resulting from fixation with a different combination of picric and acetic acids. The figures, grouped in four series, are arranged in the order of per- centage by weight of both acids present in each case. In series A, shown by the top horizontal row of figures, the picric acid, used as a saturated solution, was kept constant at about 1.2 per cent, and the acetic varied from 0.007 per cent to 20 per cent, the picric here being three times as strong as in Boveri's picro-acetic formula. In Series B, shown by the second horizontal row of figures, the saturated picric acid solution was diluted with two volumes of water (a mixture designated on the chart as " % picric "). This is the same concentration used in Boveri's reagent, which is 0.4 per cent by weight. In Series C, illus- trated by the vertical line of figures near the center of the chart, the acetic acid was kept constant at 1 per cent, its concentration in Boveri's formula, while the picric was varied from 0.01 per cent to 1.2 per cent.10 In Series D, shown by the oblique row of figures in the lower left-hand corner of the chart, both acids were varied, beginning with picric at 0.4 per cent when acetic is at 1.0 per cent, and ending with picric at 0.003 per cent and acetic at 0.01 per cent. This last series, which was reported in the previous experiment, Chart 2, is included here for pur- poses of comparison with the other three series. Chart 3 also shows the central body type present after fixation with a 1 per cent solution of acetic acid containing no picric acid (Fig. 43), as well as that occur- ring after fixation with " % picric," containing no acetic acid (Fig. 44). Table III shows the various formulae employed. In calculating the 111 In Chart 3, the picric acid of Series A and B is spoken of as being con- stant, but this is only relatively true. In Series A, for example, while the acetic acid was varied from 0.007 per cent to 20 per cent, the picric was varied only from 1.2 per cent to 1.1 per cent and thus was relatively constant. The same situation applies to Series B. In Series C the acetic acid actually was kept constant, since 1 cc. of it was added to 99 cc. of different dilutions of picric acid solution. 168 HENRY J. FRY SERIES A Sat. Picric Constant Acetic Varied ri/3 Picric "Constant Acetic Varied SERIES D Doth Picric and Acetic Varied. . r i Molariiy'- — VSew J/ioo i/5o V2° V10 GLACIAL ACETIC ACID CONCENTRATION Per Cent f by Weight:-'' CHART 3. THE EFFECTS OF VARYING IN TURN Central bodies were studied in metaphase first-cleavage figures in Choetopterus eggs, after fixation with various modifications of the picro-acetic reagent. In some the amount of picric acid was kept constant while the acetic acid was varied (Sets A and B) ; in others the reverse was done (Set C). The data of the preceding CHAETOPTERUS CENTRAL BODIES 169 SER.1ES C Picric Varied, Acetic constant Picric" luilhout Acetic 1% Acetic without Picric THE AMOUNTS OF PICRIC AND ACETIC ACIDS dilution experiment are included for comparison (Set D). The slide-making procedure was standard. Result: Only under a very narrow set of conditions are centrioles demonstrated and rays fixed in an undulating and very coarse con- figuration. 170 HENRY J. FRY percentage by weight in each formula, 1 cc. of glacial acetic acid was regarded as weighing 1 gram, since its specific gravity is 1.05. The effect upon the solubility of the picric acid (1.22 grams are soluble in 100 cc. of water at 20° C.) caused by changes in room temperature, as well as the possible effect of the presence of acetic acid in the mixture, was ignored. Extreme accuracy in making up the formulae is mean- ingless, since slight unknown dilutions were brought about in each case by the addition of small amounts of sea water with the eggs at the time of fixation. The molarities of the two reagents are also shown ; they too are only approximately correct. TABLE III Formula: of the plcro-acctic reagents used in flic experiments reported in Charts 2 and 3 Series A SATURATED PIC- RIC CONSTANT: ACETIC VARIED Series B "1/3" PICRIC CONSTANT: ACETIC VARIED Series C PICRIC VARIED 1% ACETIC CONSTANT Series D BOTH PICRIC AND ACETIC VARIED Formulae of Sat. Picric Solutions Picric Pir Picro- Sat. Picric Acetic Fig. Diluted with 2 Acetic Fig. i 1(_~ ric Sols. Acet- ic Fig. Acetic Re- Water Fig. Parts Pts. Pts. agent Water Sat. Water Picric cc. cc. cc. cc. cc. cc. cc. cc. 99.99 .007 16 99.99 .007 25 100 0 99 1 20 99 1 29 99.94 .06 17 99.94 .06 26 50 50 99 1 24 75 25 36 99.75 .25 18 99.75 .25 27 33 67 99 1 29 50 50 37 99.5 .5 19 99.5 .5 28 12 88 99 1 33 30 70 38 99. 1. 20 99. 1. 29 6 94 99 1 34 20 80 39 97.5 2.5 21 97.5 2.5 30 1.5 98.5 99 1 35 10 90 40 95. 5. 22 95. 5. 31 5 95 41 80. 20. 23 80. 20. 32 1 99 42 In order to accommodate the drawings, the abscissa of the chart, which shows the percentage of acetic acid, is drawn to various scales in its different parts; i.e., 1 cc. = - % inch; 0.1 cc. = % inch; 0.01 cc. = Y$ inch; 0.001 cc. = %o inch. The ordinate showing the per- centages of picric acid has been similarly adjusted. In Series A, B, and C, where there is only slight variation in struc- ture at each modification of the reagent, the illustration on the chart represents the type which is most abundant, based on a study of about 20 eggs in each case. In Series D, where the variation is considerable, large numbers of eggs were studied, but again only the major class occurring after fixation with each dilution is shown, since all of the types occurring are illustrated in Chart 2. CHAETOPTERUS CENTRAL BODIES 171 Six new astral types occur in this experiment: (1) one having a disrupted center and delicate rippled rays ( 1C, Figs. 20-22, 28, and 30-32) ; (2} one with a dense center, stained more darkly than the ray area but not delimited from it, with medium coarse, undulating rays (35, Fig. 34) ; (3) a similar aster with delicate rays (3C, Fig. 43) ; (4) one with a slightly demarked center, darker than the ray region, with medium coarse, undulating rays (4B , Fig. 35) ; (5) one with a disrupted center accompanied by vague or doubtful rays (ID, Figs. 17-19, 23, 26, and 27) ; and (6) a similar center accompanied by a ray area that is entirely non-radial (IE, Figs. 16, 25, and 44). The results of this experiment are as follows: (1) A typical cen- triole and its accompanying very coarse undulating rays can be demon- strated only when the percentage by weight of the picric acid is between about 0.01 per cent and 0.2 per cent, and the acetic is at the same time between about 0.05 per cent and 0.4 per cent. Other types of central bodies and ray configurations also occur within this range. (2) If the concentration of the acetic acid is extended to about 2.0 per cent and the picric is held within the range just mentioned, the centers, which are accompanied by medium coarse, undulating rays, are darker than the ray area and are either undemarked from it, i.e., dense (3B, Fig. 34), or slightly demarked from it (4B, Fig. 35). (3) If the con- centration of the picric acid is extended beyond about 0.2 per cent, and that of the acetic beyond about 2.0 per cent, the centers are disrupted and the rays rippled, being either coarse (IB}, delicate (1C), vague (ID), or absent (1£), depending upon the formula used. (4) If both picric and acetic acids are very dilute, the picric being less than about 0.005 per cent and the acetic less than about 0.01 per cent, the aster shows only a vague radial organization (2D, Fig. 42). This experiment establishes the fact that the demonstration of a centriole, which is always accompanied by very coarse undulating rays (SA), occurs only when the picric and acetic acids are simultaneously within a certain narrow range of concentration. Experiment 7. The Effects of Varying the Hydrogen Ion Concentra- tion of Boveri's Reagent Chart 4 shows the effects of the addition of varying amounts of normal sodium hydroxide solution to Boveri's picro-acetic fixative. The pH reported in each case " was modified to a slight but unknown de- gree by the addition of the small amount of sea water which always accompanies the eggs at the time of fixation. In this experiment, how- ever, the egg samples were kept unusually small, so as to change the 11 The hydrogen ion concentration was determined electrometrically by Mr. Delafield Dubois, who used an apparatus with glass electrodes. Amounts of normal sodium hydroxide solution listed in Chart 4 were added to 40 cc. of Boveri's reagent. 172 HENRY J. FRY pli as little as possible. All eggs used in this experiment were from the same egg-set. The two common central body types previously described occur after using the reagent at its natural pH, 2.2. When the pH is increased to 3.9 the rays are largely suppressed, but the centers remain the same (ID and 2D). When the pH is 4.7, 5.3, or 5.7, rays are undulating and delicate, while the centers are either even (2C) or dense (3C). In all sets except the one in which the pH was unmodified the mitotic figures are very small, the chromosomes poorly fixed, and the cytoplasm, except in the region of the mitotic figure, crowded with large, darkly- staining granules. A definite relation between ray structure and central body structure is again shown. E M PH CPE1 .EN1 DAI C.C. of Normal NaOH added to Boi'erii Reagent u~ fAL 'A Dumber of cells studied DIS1 Type IB XIBUT ION 01 AT EA ID F A5TK .CH pH 2D AL T^ 2C fPES 3C 1 1 it-Ril ^sj& n • || 2-2. 0 2O 14 6 3.9 2 19 15 4 4-7 4 25 11 6 5-3 to 23 7 73 5-1 7 14 J4 CHART 4. THE EFFECTS OF VARYING THE HYDROGEN ION CON- CENTRATION OF BOVERFS REAGENT Central bodies were studied in metaphase first-cleavage figures in Chaetopterus eggs, after fixation in Boveri's reagent at various hydrogen ion concentrations. The slide-making procedure was standard. Result: Central body structure is related to ray structure ; changes occur in both as the pH is varied. l:..\-pcrlnicni 8. TJic Effects of Varying the Temperature of Boveri's Reagent Eggs from a single egg-set were fixed in two samples of Boveri's picro-acetic fluid, one chilled to 1°C, the other heated to 95 °C. These temperatures were of course slightly modified when the eggs and the ever-present sea-water were added, as the latter were at 21° C. CHAETOPTERUS CENTRAL BODIES 173 When the reagent is at 1°C, the same two types (IB and 25) occur in about the same proportions as when the eggs are fixed under ordinary conditions at room temperature. But at 95° C. all asters have vague rippled rays and disrupted centers (ID). In other words, chill- ing the fixative produces no deviation from the usual condition, but heating it causes changes (Chart 5). The relation between ray struc- ture and central body structure is once more apparent. ILvpcr'uneni 9. The Effects of 1 'drying the Duration of Fixation with B overt s Reagent Eggs were fixed in Boveri's picro-acetic reagent for periods of time varying from one second to six months, as shown in Chart 6. This EXPERIMENTAL DATA Temperature of Boveri'5 Reagent at time of Fix- ation 95"C Number of cells studied DISTRIBUTION of ASTRAL TYPES at each TEMPERATURE Type IB FigSl 19 61% 12. 39%, ID JO 10O7, CHART 5. THE EFFECTS OF VARYING THE TEMPERATURE OF BOVERI'S REAGENT Central bodies were studied in metaphase first-cleavage figures in Chaetopterus eggs, after fixation in Boveri's reagent at 1° C. and at 95° C. The slide-making procedure was standard. Result: Central body structure is related to ray struc- ture. At low temperatures the central bodies and ray configurations are the same as those occurring at ordinary temperatures ; at high temperatures the situation is modified. experiment shows the speed at which the reagent penetrates and fixes the eggs and also the effects of duration of fixation upon astral struc- ture. The experimental set-up employed when fixing eggs for periods as brief as one or several seconds was as follows : in order to effect in- stantaneous mixing, two drops of a thick egg suspension were suddenly squirted into 2 cc. of the reagent, which was kept in motion in a small 174 HENRY J. FRY Stender dish. This proportion of eggs to reagent weakens the fixative so slightly that any effects of dilution can safely be ignored. After the desired interval of fixation, one or more seconds, the entire mixture was poured into 400 cc. of 70 per cent alcohol, the fluid always used to wash material fixed with a picro-acetic reagent. The addition of 2 cc. of eggs and fixative to 400 cc. of alcohol dilutes the fixative so effectively as to stop its further action. However, to make sure that the alcohol itself, plus the limited amount of reagent added with the eggs, did not modify the coagulation product, a control experiment was run, in which eggs were fixed in 70 per cent alcohol containing 0.5 per cent picro-acetic reagent. The asters in these eggs show only vague rays, and central areas entirely undifferentiated from the peripheral part (2D, Fig. 57). Hence the washing in alcohol can be dismissed as a factor in producing the types of astral structure occurring in the experiment. The set-up for fixing eggs for longer periods of time was the usual one.12 This experiment shows, first, that complete fixation of the eggs in full strength picro-acetic reagent occurs within one second ; (the speed of fixation in dilute reagents is not known). In the second place, it shows that different lengths of exposure to the reagent modify the types of central bodies present. These differences, at the various periods of fixation, are as follows : Fixation time 1 , 3, or 6 seconds: All asters have medium coarse un- dulating rays; 28 of the 34 centers studied are even (2£>) ; the remain- ing 6 are dense (3/?). Fixation 'time 30 seconds: 35 per cent of the asters have disrupted centers and medium coarse rippled rays (IB), and 65 per cent have even centers and medium coarse undulating rays (2/j). These are the same types which occur after the usual periods of fixation for hours, but the proportions in which they occur are intermediate between those of eggs fixed but a few seconds, when there are no disrupted centers, and those occurring under usual conditions, when there is a high per- centage of disrupted centers. Fixation time 10 minutes or fifteen Iwitrs: 90 per cent of the asters have disrupted centers (IB), and 10 per cent have even ones (27? ). Fixation time thirteen days: only 28 per cent of these eggs are of the disrupted type. 72 per cent having evenly filled centers. The per- centages of both types occurring in this set are again intermediate, this 12 The samples of eggs at the three shortest intervals are intentionally small, because of the effort to keep the egg mass as small as possible. The equally small samples reported for the sets fixed at longer intervals are explained by the fact that this particular batch was fixed a little too early, due to a miscalculation, and but few of the eggs had reached metaphase. However, in spite of the small size of the samples and the error necessarily involved, the major facts are clear. CHAETOPTERUS CENTRAL BODIES 175 EXPERIME DAT Duration of fixation in Boveri's Keagent :NTAL A Number cf Cells studied DISTR1BT AT EACIr Type IB JTION OF [ DURATK 2B ' ASTRA1 3N °S FIX >3B , TYPE5 ;ATION 2D H mm $S**^ - -\ i y V *^ '' tet^l^Mi PfflfeS^-? l^p§l 1 second. 15 11 73% 4 27% 3 seconds. 10 9 90K I 10% 6 seconds. 9 6 697. 1 117, 30 seconds. 20 7 357. 13 657. 10 minutes. 20 Id 90% 2 1O7. 15 hours. 20 1& 30't, 2 10'Z 13 dayj. 18 5 267, 13 72% 6 months. 21 21 1007, CONTROL:- Egfs fixed in TO'Z Alcohol contain in <£- o.57. Boverrs R_eagent . 20 2(9 200% CHART 6. THE EFFECTS OF VARYING THE DURATION OF FIXA- TION WITH BOVERI'S REAGENT Central bodies were studied in metaohase first-cleavage figures in Chactoptcrus eggs, after fixation in Boveri's reagent for various periods of time. The slide- making procedure was standard. Result: Central body structure is related to ray structure. Fixation for a few seconds produces only even central bodies and undulating rays ; fixation for hours produces in most cases disrupted centers and rippled rays ; fixation for months produces the same result as fixation for seconds. 176 HENRY J. FRY time between those of eggs fixed for hours and those fixed for months. Fixation time six months: all asters have even centers and undulat- ing rays (25). In brief, this experiment shows that fixation for a few seconds produces only asters with filled centers, even or dense, and undulating rays ; fixation for minutes and hours produces a majority of asters with disrupted centers and rippled rays and a minority with filled centers and undulating rays ; and fixation for months returns the aster to the same condition as that produced by fixation for a few seconds. The major result established in the previous experiments is reaffirmed here : ray structure and central body structure are related. In all of the experiments reported in this paper except this one, the eggs were left in the reagent from six to fifteen hours. This period of fixation falls within the limits of the groups fixed in this experiment at ten minutes and fifteen hours, during which time no modification of central body structure occurred as a result of the length of exposure to the reagent. Hence it can be assumed that they were not affected by differences in length of exposure to the fixative. IV. DISCUSSION All of the astral types previously described are arranged in Chart 7, where they are classified on a three-fold basis : (A) According to differences in structure there are five classes of central bodies : Number and Name Structure 1. Disrupted. Not filled. Entirely empty or containing scat- tered material. 2. Even. Filled. Not demarked from ray area. Stained like ray area. 3. Dense. Filled. Not demarked from ray area. Stained darker than ray area. 4. Slightly demarked, with- Filled. Slightly demarked from ray area, out centriole. Stained darker than ray area. 5. Slightly demarked, with Filled. Slightly demarked from ray area, single or double centriole. Stained darker than ray area. (B) In shape, the rays take two forms: first, rippled or serpentine, and second, undulating or almost straight.13 Members of the inter- mediate, or slightly rippled class are included with the other two (p. 156). (C) On the basis of coarseness there are five classes of rays: (1) very coarse, (2) medium coarse, (3) delicate, (4) vague, and (5) absent. 13 See footnote 4, p. 155. CHAETOPTERUS CENTRAL BODIES 177 CLASSES OF CENTRAL BODIES 1 Disrupted Empty or .Scattered 2 Even ., probably scusibilis). The present work is in part an independent corroboration of the work of the Chattons, because the author was unacquainted with their work until the essential points in which the two investigations agree had already been established. In the present work, though not in the work of the Chattons, experi- mentally produced chains gave rise to a new racial type, — double ani- mals. As a result of this, the question in the foreground of interest 1 During the performance of most of the experimental work, the author was Fellow of the National Research Council. The author wishes to acknowledge his indebtedness to Professor H. S. Jennings for many important suggestions, particularly concerning preparation of the manuscript; and to Dr. Ruth S. Lynch for critical reading of the manuscript. 2 Identified through the courtesy of Professor Edouard Chatton. 187 188 T. M. SONNEBORN in this paper is : What are the genetic consequences of the experimental production of chains? In the treatment of this question, the following matters will be taken up: (1) the experimental determination of chain- formation; (2) the mode of origin of chains and their subsequent de- velopment into races of doubles; (3) the persistence and stability of the races of doubles ; (4) the existence of diverse biotypes among dou- bles and their descendants; (5) an examination of the nature of the processes involved in these changes of hereditary characteristics in vegetative reproduction. II. MATERIALS AND METHODS The methods employed for the cultivation of large mass cultures of Colpidiuin were given in an earlier paper (Sonneborn, 1930a).3 The basic fluid was an infusion of 1.5 grams rye grains boiled for ten min- utes in 100 cc. spring water, filtered and let stand for 24 hours to '' ripen." Two hundred cc. of this infusion plus five boiled rye grains were put into a finger bowl 10 cm. in diameter and 4 cm. deep. This was inoculated with 10 to 15 cc. of an old culture of C. cainpylum; within 24 hours it would develop into a flourishing culture and remain so for several days. When the colpidia began to get smaller and paler, usually in two to six days, subcultures were made in the same way. Only large stock cultures were maintained in this way. Smaller mass cultures were kept in square salt cellars and in Colum- bia dishes. These were begun with several drops of ripe fluid (without any rye grains) and ten to thirty colpidia; each day more ripe fluid was added until after three to five days the dish was full (about thirty drops). Subcultures were then made in the same way. In special experiments designed to investigate carefully the factors determining chain-formation, Columbia dish cultures were made in a different way, following, in many respects, the bacteriological technique employed by Raffel (1930) in the cultivation of Paramecium. The standard rye infusion, as soon as filtered, was distributed in cotton- plugged test tubes, autoclaved and stored until needed. When needed, a tube was opened over a Bunsen flame and inoculated by means of a platinum needle with a pure culture of either Achrornobacter sp., prob- ably candicans, or Micrococcus sp., probably scnsibilis, grown on beef- agar slants. The inoculated culture fluid was then pipetted into Colum- bia dishes, inside of Petri dishes. These, and the cotton-plugged pi- pettes, inside of jars, had all been heated for one hour at 150° C. in a hot-air sterilizer. Each pipette was used only once and the top of the Petri dish containing the Columbia dish was raised only enough to admit 3 In that paper, the species was incorrectly called Colpidhim striatum. GENETICS OF CHAIN FORMATION IN COLPIDIUM 189 the pipette with culture fluid and colpidia. The colpidia had previously heen washed according to the method of Parpart (1928). After such a culture was set up, it was opened only once for purposes of subculture. In all the rest of the work, the colpidia were cultured in isolation on ground glass slides containing two concavities. Twelve of these slides were placed on a glass plate raised on glass supports in an inverted 9-inch Petri dish sealed with water at the bottom. The " ripe " rye infusion was used as cultivation medium, one drop to each concavity. Each day one Colpidinui was placed in such a drop; 24 hours later the drop was again observed, records of reproduction and other matters of interest were made, and one of the colpidia was transferred to a fresh drop of ripe culture fluid. Each day a similar procedure was followed. A number of other details of procedure in the isolation cultures are important: (1) The possibility of perpetuating, by daily transfers, dif- ferences in bacterial flora between different lines was avoided by col- lecting a small amount of fluid from each 24-hour-old culture drop and using these small drops for inoculating the culture fluid for the next day. Such cross-inoculations were performed daily in some experiments, less frequently in others. (2) The possibility of repeating daily systematic differences in the treatment of the lines compared was avoided by the following methods: (a) within each moist chamber containing 24 lines, each type of animal in a particular set of comparisons was represented by the same number of lines; (b) within each moist chamber the lines were distributed according to a plan whereby the lines of the same type or race were separated from each other and whereby the arrangement in no two moist chambers was the same; (c) the order in which the moist chambers were transferred was systematically changed from day to day. (3) The possibility of personal bias influencing the results was avoided by assigning to each line for daily identification a name that gave no indication of its genetic history. The key to these names was not con- sulted until the end of the experiment. Many of the experiments were conducted at room temperature of 20° -25° C., some were run in a constant temperature chamber which only rarely exceeded the range 22°-23° C. The animals used in all the work here reported were descended from one individual and were never known to conjugate, even when efforts were made to make them do so. III. THE DETERMINATION OF CHAIN-FORMATION As already mentioned, fidouard and Mine. Chatton were able to in- duce the formation of chains in Colpidium campylum and Colpidium colpoda by feeding these species a particular strain of Bacillus coli. Other species of bacteria, other strains of Bacillus coli, and this strain 190 T. M. SONNEBORN grown on other than a vegetable base did not induce chain- formation in Colpidium. These investigators thus demonstrated that the formation of chains in Colpidium is determined by a particular strain of bacteria grown on a vegetable base. This result of the Chattons was confirmed by the present investiga- tion, as will now be set forth. The rye infusion in which chains of colpidia had arisen was plated out on agar (with the help of Dr. Raffel) and the different kinds of bacteria were separately cultivated. Only two kinds of colonies could be distinguished : one yellow and rapidly growing, the other white and slowly growing. These two species were identified for me through the courtesy of Professor William W. Ford of the Department of Bacteriology, School of Hygiene and Public Health, the Johns Hopkins University. The yellow species was identi- fied as belonging to the genus Micrococcus (Bergey) and corresponding closely to the species M. scnsibilis; the white species was identified as belonging to the genus Achromoboctcr (Bergey) and corresponding closely to the species A. candicaus. A number of experiments indicated an increase in the frequency of chain-formation when Micrococcus predominated in the food supply. In a few cultures some chains were formed in fluid in which Achromo- bactcr predominated ; in these cultures, however, Micrococcus had not been excluded. In all other cultures in which chains were formed, Micrococcus had definitely been inoculated into the culture fluid. Critical experiments, using the rigorous bacteriological methods de- scribed above in Section II, were performed to discover the relation of these two species of bacteria to chain-formation. On July 29, eight normal colpidia were washed, according to Parpart's (1928) method, in two steps : five washings were followed by a lapse of five hours, after which five more washings were made. Four of the eight colpidia were washed in autoclaved rye fluid inoculated with Achroinobactcr ; the other four in similar fluid inoculated with Micrococcus. By July 31, each set of four had multiplied to form eighteen colpidia. The animals in Micrococcus fluid were used to establish nine cultures, each consisting of two colpidia in sixteen drops of autoclaved rye fluid inoculated with Micrococcus. The animals in Achroinobactcr were used to establish nine similar cultures in Achroinobactcr fluid. Counts of the number of singles and the number of chains in each of these eighteen cultures were made on August 3. As appears in Table I, not one chain was formed among 46,716 colpidia produced in the pure Achromobacter fluid, but 51 chains were formed among 11,912 colpidia produced in the pure Micrococcus fluid. The production of chains is clearly deter- GENETICS OF CHAIN FORMATION IN COLPIDIUM 191 w j P3 •^ "S o M-, <-, O o 3 NO o O cs _- „ ~s «— I ^H '-O -H o H SO ON O 00 o O ON 10 OO ^^ O T— 1 OO IO CO o ^ o O -? O 0 00 OO r-\ -t 1-1 o o o so so IO ^ -h •5 E 3 C rt 1 _C cu 'be rt n o W S -J S « S < "^ H •« 13 "-3 "2 5 Q X 03 •a u cu O. Q &r f~* ^ o c*3 -t ^H — <^ **^ ^O i e>> ID 1s r^ P 1 « OO tt % § 2 -t 3 -t q g >,•""* { l -f o ^ ( 3 1 i—) 00 -t ro CO w 00 "3 ' OO -+ OO cs <* "* « "* i (N _, o o o 0 C 3 ro ** CO ^ |7 0, 3 ~-f o 1 CU ~H c — o OO * o ON 0 cs Cxi Q O tn CU be C '35 en CU be c '35 doubles ' doubles "So c '35 en — be c '35 doubles : doubles 5 w OH u CU o en en o o (U u ^2 u O en o cu E 3 C 03 cu C 03 CU £ 3 E 3 c 03 cu C 03 cu s 3 ^ S ^ Z z ^ ^ Z r^ CU CU c c o o U U GENETICS OF CHAIN FORMATION IN COLPIDIUM 199 vation here employed. The doubles fail to double their number at the fissions which yield one double and two singles ; but the singles double their number at every fission and the proportion present is continually being increased by the transformation of some doubles into singles. Hence, in the sampling method of culture renewal here employed, even- tually too small a proportion of doubles will be present to find a place in the sample. Thus the rate of transformation of a series of cultures would depend on the frequency with which doubles produce singles. This frequency was not ascertained, so that the relative importance of this factor remains unknown. Other factors, however, become important as soon as some singles have been produced. The subsequent changes in proportion of the two types present must then depend partly on their relative rates of multi- plication and partly on their relative rates of mortality. The fission rates of the singles and doubles of clones 3 and 7 are given in Table II. In clone 3, the singles reproduced 0.1 to 0.3 fissions per day more rapidly than the doubles in all periods except the first two. In these two periods there was practically no difference (0.01 fission per day) between the means for the two groups. However, there were fewer (ten or less) lines in each group during these periods than in any of the later ones; hence, the comparisons during these periods are corre- spondingly less valuable than during the other periods in which the dif- ference between the two groups was well marked. In clone 7, there can be no doubt that the singles reproduced more rapidly than the dou- bles. The difference is well marked in all periods for which compari- sons are available, even in those in which but few lines were compared. It can be said, therefore, that certainly in clone 7, and very probably in clone 3, the singles reproduced more rapidly than the doubles. The mortality rates for these groups are given in Table III. With a negligible exception in clone 3, during the period July 4—7, the mortal- ity of singles in both clones 3 and 7 is consistently greater than the mor- tality of the corresponding doubles. In both clones the mortality rate for the total of all periods is over twice as great among the singles as among the corresponding doubles. Therefore, in the series of mass cultures of doubles, the gradual change of the proportions of doubles and singles present was not brought about by differential mortality, — indeed was retarded by it ; but was due to the continual production of singles by doubles and to the faster reproduction of the singles thus produced. That the death of all doubles or the transformation of them all into singles played no role in their extinction is demonstrated by the following account of isolation cultures of doubles. 200 T. M. SONNEBORN On April 22, 1930, 48 lines of doubles were isolated in concavities on culture slides. Among these lines eight clones were represented by six lines each. Each line was cultivated in isolation until June 2, 1930. During these 42 days of culture, every clone continued to maintain it- self as doubles ; no clone died out or transformed completely into singles. TABLE III Comparison of mortality rates of doubles and singles derived from them PERIOD June 17-21 June 27- July 1 July 4-7 July 8-11 July 12-16 TOTALS AND MEANS Clone 3 Singles No. line-days 50 60 192 188 230 720 No. died 0 1 0 6 9 16 No. deaths per 100 line-days 0.0 1.7 0.0 3.2 3.9 2.2 Doubles No. deaths per 100 line-days 0.0 0.0 0.05 2.1 0.9 1.0 No. died 0 0 1 4 2 7 Xo. line-days 50 60 191 190 234 725 Clone 7 Singles No. line-days — 52 186 179 218 635 No. died — 9 8 8 16 41 No. deaths per 100 line-days — 17.3 4.3 4.5 7.3 6.5 Doubles No. deaths per 100 line-days — 3.3 1.1 4.4 3.1 2.9 No. died — 2 2 8 7 19 No. line-days — 60 188 183 225 656 Singles appeared occasionally along with the doubles, especially during the early history of the lines ; but during the last 18 days of cultivation no singles were produced in any of the lines. This change in the fre- quency with which singles were produced was probably a consequence of the method of selection followed each day, for each day the line was perpetuated by the most perfectly doubled individual present (that is, GENETICS OF CHAIN FORMATION IN COLPIDIUM 201 by the one showing the least development of an anterior cleft). The period (42 days) during which these isolation cultures were maintained as doubles was as long as the period (41 days) during which the mass cultures had transformed completely to singles. Further, these same clones of doubles were maintained from April 5 until October 15. Part of this time they were in isolation culture, part of the time in mass cul- ture ; in the latter renewals were made by selecting chiefly doubles. As the usual rate of reproduction was three fissions per day, about 582 gen- erations must have passed while the colpidia remained double. At the end of this time the cultures were discontinued, but there was no reason to suppose that the doubles could not have been maintained indefinitely. The stability of the doubles was further demonstrated by their main- tenance of organization through encystment. The one time encystment was observed in four years of close attention to Colpidiuin campyluin, it FIGS. 24 and 25. Both from notebook sketches and measurements made with ocular micrometer. Fig. 24, cyst 164 M x 85 M containing three double colpidia, two small and one large. Fig. 25, one of these doubles immediately after ex- cystment. occurred in a line of double animals, during isolation culture, 34 days (more than 100 generations) after this line originated from a multiple monster. The cyst (Fig. 24) was discovered less than 24 hours after it formed; in it there were three vigorously moving doubles. Of these three, one was large, two small. Within half an hour, the larger animal divided into two. The cyst remained in this condition one more day. On the third day, five animals were present in the cyst; on the fourth day, six animals. On the fifth day, the cyst was opened with a fine glass needle to allow the encysted animals (of which one is shown in Fig. 25) to emerge. Each of these was a double animal; four of them were used to initiate separate lines of descent. Records showed no dif- ferences among the emerged animals and their descendants or between the descendants of encysted and non-encysted members of the same clone, so that more detailed studies on this matter were not made. However, the important point is this : in spite of the reorganizations known to occur in cysts of ciliates, the encysted colpidium did not re- organize as a normal single, but emerged from as it had entered the cyst. — with the double organization. 202 T. M. SONNEBORN VI. DIVERSITY OF BIOTYPE AMONG DOUBLES AND AMONG THE SINGLES PRODUCED BY THEM The origin of diverse biotypes during vegetative reproduction has been the subject of numerous investigations (see review by Jennings, 1929). To these must be added the present one on Colpidhnu, in which it has been shown (Section IV) that biotypes of doubles originate from a clone of singles under the influence of a special environmental condi- tion (diet including Micrococcus). Further (see Section V), some individuals in these biotypes of doubles give rise to new biotypes of singles. We take up in this section the question of whether there arise other biotypic diversities among the doubles and the singles pro- duced by them. Of this question there are several aspects: (1) Do bio- typic diversities exist among the different clones of doubles? (2) Do TABLE IV Comparison oj fission rates of doubles of clones 3 and her of divisions per line per day. Rates given in nutn- PERIOD May 1-7 Mav 8-14 June 4-11 June 17-21 June 27- July 1 July 4-7 July 8-11 Julv 12-16 Weighted mean for 8 periods Clone 3 No. lines 6 6 6 10 12 47 44 45 Fission rate 3.05 3.05 2.79 2.74 2.90 2.72 3.31 3.33 3.064 Clone 7 Fission rate 2.95 2.73 2.52 2.58 2.64 2.48 3.11 3.12 2.835 No. lines 6 6 6 10 9 46 40 39 Clone 3 minus Clone 7 0.10 0.32 0.27 0.16 0.26 0.24 0.20 0.21 0.229 biotypic diversities exist among singles derived from different clones of doubles? (3) Do biotypic diversities exist among singles produced independently from the same clone of doubles? (4) Do biotypic di- versities exist between singles produced by doubles and singles not de- scended from doubles? (5) Do biotypic diversities exist among differ- ent lines of descent within a clone? These questions will now be taken up in the order mentioned. (1) Biotypic dircrsitics among different clones of doubles. — Eight clones of doubles were compared in (a) the rate of decrease in propor- tion of doubles present in series of mass cultures; (b) the rate of multi- plication; and (c) the rate of mortality. GENETICS OF CHAIN FORMATION IN COLPIDIUM 203 (c) Striking differences in the rate of decrease in proportion of dou- bles present appeared among the clones cultivated in series of mass cul- tures, as described in Section V. In the series fed Achromobacter , at the end of the period of observation, no doubles remained in any of the eight clones, except in clone 2. In this clone, however, the vast majority of colpidia present were still doubles. Clearly, the rate of decrease in proportion of doubles was less in clone 2 than in any other clone. This was apparent in both types of food cultures and at all stages in the series. For example, after three renewals of the cultures, all colpidia of clone 2 were still double when fed Micrococcus and 95 per cent were double when fed Achromobacter. But in clone 3 only 60 per cent of the colpidia were double in Micrococcus and only 5 per cent in Achromo- bacter. Between these two extremes, in clone 8, 88 per cent of the colpidia were doubles in Micrococcus, 60 per cent in Achromobacter; and in clone 1, 72 per cent were doubles in Micrococcus, 40 per cent in Achromobacter. The different clones thus manifested at least four different rates of transformation and these differences between the clones were the same in both types of culture fluid. (b} An extensive comparison was made of the rates of multiplica- tion in clones 3 and 7. Their mean fission rates for eight periods of from four to seven days each are given in Table IV. In all periods clone 3 multiplied more rapidly than clone 7. At the end of the third period, the slowest line of clone 3 was selected to give rise to all later members of the clone; at the same time, the fastest line in clone 7 was selected to give rise to all later members of this clone. As appears in the table, this radical adverse selection in both clones changed neither the direction nor the magnitude of the difference between the two clones. Clone 3, for the eight periods, had a mean rate of 3.064 fissions per line per day ; during the same time, clone 7 had a mean rate of 2.835 fissions per line per day. The different periods gave results very similar to the general mean : in five of the eight periods the excess of clone 3 over clone 7 was between 0.20 and 0.27 fission per line per day; in two pe- riods it was below this range (0.10 and 0.16 fission) and in one period above it (0.32 fission). There can be no doubt of the uniform heredi- tary difference between clones 3 and 7 in fission rate. (r ) The rates of mortality were also extensively compared in these two clones (see Table V). In seven periods, extending over a period of 86 days and including records for 54 of these days, the mortalitv rates were 0.78 deaths per 100 line-days in clone 3 and 2.40 deaths per 100 line-days in clone 7. In no period is the mortality rate of clone 3 higher than that of clone 7. The difference in mortality rates of these 204 T. M. SONNEBORN two clones appears constantly through all parts of the experiment and is thus a biotypic diversity. Among the clones of doubles here compared, clones 2 and 8 arose at different times from one multiple monster and the other six clones arose at different times from another multiple monster. Thus, differences in the rate of decrease in proportion of doubles present in series of mass cultures existed between clones derived from different multiple mon- sters : clones 2 and 8 as compared with clones 1 and 3 ; but similar differ- ences also existed between clones derived from the same multiple mon- ster : the rate in clone 2 differing from that in clone 8, and the rate in TABLE V Comparison of mortality rates of doubles of clones 3 and 7 PERIOD April 22- May 14 June 4-11 June 17-21 June 27- July 1 July 4-7 July 8-11 July 12-16 TOTALS AND MEANS No. line-days 120 48 50 60 191 190 234 893 No. died 0 0 0 0 1 4 2 7 t '1 ~i No. deaths per 100 line-days 0.0 0.0 0.0 0.0 0.05 2.1 0.9 0.78 No. deaths per 100 line-days 1.7 0.0 0.0 3.3 1.10 4.4 3.1 2.40 Clone 7 No. deaths 2 0 0 2 2 8 7 21 No. line-days 120 48 50 60 188 183 225 874 clone 1 differing from that in clone 3. Likewise, differences between clones of doubles (clones 3 and 7) derived from the same multiple monster were found in rate of fission and in rate of mortality. (2) Biotypic diversities among singles derived from different clones of doubles. — Clones 3 and 7 of doubles have just been shown to differ in rate of fission and in rate of mortality. Do the singles produced by these two clones of doubles differ in the same way? One single from each of these two clones of doubles was permitted to give rise to a num- ber of lines and the mean fission rates of these two groups of singles were compared in four periods. The differences found were small and not constant, so that no significance may be attached to them. In mor- tality rate, however, the situation was different. As appears in Table VI, in all four periods the rate of mortality is greater — usually very GENETICS OF CHAIN FORMATION IN COLPIDIUM 205 much greater — among the singles of clone 7 than among the singles of clone 3. The rate for the total time is 2.39 deaths per 100 line-days in clone 3 and 6.33 deaths per 100 line-days in clone 7. The rate for clone 7 is thus 2.65 times as great as that for clone 3. In connection with this difference, it is of interest to note (see Table V) that the doubles of clone 7 had a mortality rate 3.08 times as great as the doubles of clone 3. TABLE VI Comparison of mortality rates of stocks of shn/lcs produced by one single from a clone 3 double and one sinc/lc from a clone 7 double PERIOD June 4-11 June 27- July 1 July 4-7 July 8-11 July 12-16 TOTAL No. of line-days 42 60 192 188 230 712 Singles from ^i__rt -j No. of deaths 1 1 0 6 9 17 clone 6 double No. of deaths per 100 line- days 2.4 1.7 0 3.2 3.9 2.39 No. of line-days 44 52 186 179 218 679 Singles from No. of deaths 2 9 8 8 16 43 clone 7 double No. of deaths per 100 line- days 4.5 17.3 4.3 4.5 7.3 6.33 Rate of clone 7 singles minus rate of clone 3 singles 2.1 15.6 4.3 1.3 3.4 3.94 (3) Biotypic diversities among singles produced independently from the same done of doubles. — As set forth above, singles were produced two at a time from the anterior half of a double animal. One arises on the right side, one on the left (see Fig. 20). Do the two singles of such a pair differ? That there might be a difference in symmetry was suggested to me by Mr. Donald Costello; but I was unable to detect it. However, it was clear that the two singles of a pair differed sometimes, but by no means always, in other respects. These differences occurred when the anterior cleft of the parental double was much shifted from the usual median position. Frequently the narrower part was sep- arated off as a single one division or more before the broader part. The single produced from the narrower part was invariably narrow, pale, and short as compared with the single produced from the broader part or with the singles ordinarily produced from doubles. The fates 206 T. M. SONNEBORN of these two different types of singles were frequently observed. The difference in their fates is illustrated by an experiment on 48 lines ob- served from May 16-19, 1930. In this group, of the seven singles de- rived from the narrower parts of unequally cleft doubles, the progeny of four died. Among the 41 lines not derived from narrow parts of unequally cleft doubles, the progeny of only two died. Thus 57 per cent of the one group died as compared with only 5 per cent of the other group. Many other observations confirmed the results in these groups, so that there was no doubt of the very much greater mortality among TABLE VII Comparison of mortality rates of a group of singles descended from one single produced by a double of clone 7 and one single not descended from doubles. PERIOD June 27- July 1 July 4-7 July 8-11 July 12-16 TOTAL Number of line-days 479 384 376 473 1712 Singles not descended Number of deaths 5 3 10 8 26 f doubles Number of deaths per 100 line- days 1.0 0.8 2.7 1.7 1.52 Number of line-days 52 186 179 218 635 Singles from Number of deaths 9 8 8 16 41 1 7 double Number of deaths per 100 line- days 17.3 4.3 4.5 7.3 6.46 Singles of clone 7 minus singles not descended from doubles 16.3 3.5 1.8 5.6 4.94 the descendants of singles derived from the narrower parts of unequally cleft doubles. Whether the singles produced from the ordinary medially cleft doubles at one time were diverse from those produced from a double of the same clone at another time was not investigated. (4) Biotypic diversities betzvcen singles deseended from doubles and singles not descended from doubles. — Comparisons were made between singles descended from doubles and singles not descended from doubles in rate of fission and in rate of mortality. The differences found in rate of fission were small and not consistent, so that no significance may be attached to them. In rate of mortality, however, the differences were clear (see Table VII). The total difference between the singles not de- GENETICS OF CHAIN FORMATION IN COLPIDIUM 207 scended from doubles and a group of singles descended from one single produced by a double of clone 7 is very great : the mortality rate of the latter group is 4.25 times as great as that of the former. This differ- ence is clearly manifested in every period and demonstrates a biotypic difference in rate of mortality between these two groups of singles. (5) Biotypic diversities within a clone. — As already set forth, the most striking differentiation into biotypes that occurs within a clone of doubles is its splitting into biotypes of singles and doubles. This oc- curred repeatedly in all clones of doubles studied. Furthermore, the biotypes of singles so formed within a clone of doubles were not all alike: some were normal singles, others were narrow, pale, and highly inviable. In addition to these biotypic differences, there were indica- tions of other biotypic differences within a clone, among the doubles themselves. In the isolation cultures of doubles maintained from April 22 until June 2, the frequency with which singles were produced changed strikingly. This was probably brought to light by the practice of selecting daily from among the individuals produced during the previous 24 hours in each line, the individual which showed least development of an anterior cleft. In each line this individual was used to perpetuate the line and the remaining individuals were discarded. In spite of this method of selection, cleft individuals continued to appear and give rise to singles during the early history of these isolation lines. During April 22-30, singles were produced in 31 of the 48 lines of doubles under cultivation. In some lines they appeared more than once. The product of the number of lines in which they appeared by the number of days on which they appeared gives a measure of their frequency of production. This product was 63 line-days, yielding an interval of 6.86 line-days between successive productions of singles. These figures are in striking contrast to those obtained for the same lines during the period May 16-June 2. Not one single was produced during these 864 line-days. In order to bring to light such a great change in the fre- quency with which doubles produced singles within the same lines of descent, there would have to be genetic differences in frequency of single production among the doubles of each clone. Such a conclusion seems required by the evidence. Attempts to isolate by selection biotypic differences in rate of fis- sion within clones of doubles, within clones of singles, and within clones of singles descended from doubles were all fruitless. The coefficient of variation of fission rate of a clone of doubles (8.38 per cent) was greater than the corresponding coefficients for singles descended from this clone of doubles (6.46 per cent) and for a clone of singles not descended 208 T. M. SONNEBORN from doubles (6.66 per cent). But this greater variability of the clone of doubles was probably not an index of the existence of biotypic di- versities in fission rate within this clone, because extreme selection did not result in the isolation of sub-clones with diverse fission rates. VII. DISCUSSION The racial effects of environmental conditions found here in Col- pidium bear a striking parallel in many respects to the relations previ- ously found in the rhabdocoel turbellarian, Stenostomum incaudatum (Sonneborn, 1930/0- As in Colpidium, so in Stenostomum, special environmental conditions induced abnormalities in reproduction result- ing in the formation of irregular monsters. These, likewise, gave rise to double animals that differed from each other and produced races of doubles differing in the same way. Further, in both Colpidium and Stenostomum the races of doubles maintained their character after re- moval from the environment that induced their formation, except that singles were formed whenever a cleft of sufficient extent occurred in the growing region perpendicular to the plane of fission. In both, singles gave rise to races of singles of higher viability than the doubles from which they arose. The degree of similarity between a protozoan and a flatworm in the effects of environmental conditions on their hereditary characteristics is particularly striking in contrast to the very different results of most similar work on higher organisms. What is the basis of this difference? It seems to be the method of reproduction. In sexual reproduction, change of hereditary characters depends largely on changes in the na- ture or in the composition of the chromatin. Environmental conditions of special penetrability are required to get at this material, so that but few environmental conditions are effective in altering hereditary char- acteristics. In asexual reproduction, on the other hand, change of hereditary characters may be brought about without in the least affecting the nature or composition of the chromatin ; changes in the composition of components of a larger order are also capable of self-perpetuation — that is, are heritable. Examples of this are the rearrangement of parts in homopolar doubles in Stenostomum and Colpidium. That changes in the chromatin were not involved in these examples was demonstrated by the fact that when individuals were produced from cleft parts of doubles, these were always singles and gave rise to biotypes of singles. The type of hereditary change involved in the production of biotypes of doubles in Stenostomum and Colpidium is similar to the type involved in the production of stocks of Drosophila in which the two X-chromo- GENETICS OF CHAIN FORMATION IN COLPIDIUM 209 somes are united or in which translocations, inversions, or reduplications have occurred. All such examples illustrate hereditary changes not due to changes in the nature of the germinal material, but due to changes in the number of units or arrangement of units in the germinal material. It is remarkable that very diverse environmental conditions acting on such diverse organisms as Colpidhun and Stcnostouiuin should result in similar stable types. It may be that this is another example of the stability of whole multiples, as in polyploids ; and that of all the terato- logical consequences of diverse original stimuli, the whole multiples that result are particularly of the viability requisite for survival and self- perpetuation. VIII. SUMMARY In a clone of Colpidiuui campyhun (Stokes), a small proportion (less than 1.2 per cent) of the individuals formed chains when culti- vated in a rye infusion inoculated with the bacterium Micrococcns sp. (probably scusibilis} , but not when the infusion was inoculated with Achromobactcr sp. (probably candicans) . Other factors, one of which possibly was the concentration of colpidia in the culture fluid, affected the proportion of chains formed when the colpidia were fed Micrococcns. Chains thus produced went through a series of developments including the formation of heteropolar doubles and multiple monsters, and cul- minating often in the formation of homopolar doubles of a self-perpetu- ating, relatively stable sort. Similar biotypes of doubles were also formed once as a result of a " pseudo-conjugation." In clones of homo- polar doubles, singles sometimes arose by ordinary transverse fission across a double with a deep median anterior cleft. Consequently, mass cultures begun with doubles eventually contained singles also. As the two types multiplied side by side, the relative proportion of singles gradually increased. When the cultures were regularly renewed by taking a sample of the old culture to start a new one, eventually, after many such renewals, doubles entirely disappeared from the cultures, leaving only singles. The change occurred in both Achronwbacter and Micrococcns fluid, but more rapidly in the former ; this was probably due to the more rapid reproduction in that fluid. The change in the mass cultures was not due to dying out of doubles or to the transforma- tion of all of them into singles. In isolation culture, lines of doubles were maintained as long as the period required for doubles to disappear entirely from mass cultures. Furthermore, when doubles were delib- erately salvaged at each renewal of culture, they were maintained in cultivation, partly in isolation, partly in mass, for 194 days, during which about 582 generations passed. The disappearance of doubles 210 T. M. SONNEBORN from mass cultures in 41 days or less must therefore have been due to other factors than the inability of doubles to live and reproduce their kind. One of these factors was a differential rate of fission : the singles produced by the doubles of one clone multiplied 0.373 fissions per line per day more than the doubles of this clone. On the other hand, differ- ential mortality counteracted this to some extent, for the mortality rate of singles was higher than the mortality rate of the doubles that pro- duced them. The gradual disappearance of doubles in series of mass cultures was therefore due partly to their repeated production of singles and partly to their lower fission rate. The persistence of doubles during nearly 600 generations, by the end of which time no evidence of inability to maintain themselves had yet appeared, indicates that the type could have maintained itself indefinitely, even when the bacterium that led to its formation was absent or present in but very small quantities. Fur- ther evidence of stability of organization was the passage of a line of doubles through encystment without loss of the double organization. The question of whether diverse biotypes existed among the experi- mentally produced doubles and their descendants was extensively investi- gated. (1) Different clones of doubles differed (a) in the rate at which doubles disappeared from series of mass cultures, (b) in rate of multi- plication, (c) in rate of mortality. (2) Singles derived from diverse clones of doubles differed in rate of mortality to about the same extent as the clones of doubles from which they had been derived. (3) There were two very different kinds of singles derived from the same clone of doubles : the usual kind and a rarer kind formed by transverse fission across an unequally cleft double. The singles formed from the nar- rower anterior part of these were narrower, paler, and shorter than ordinary singles and had a much higher rate of mortality. (4) Singles not descended from doubles had a lower rate of mortality than singles descended from doubles. (5) Within a clone of doubles there were genetic differences in the frequency with which singles were produced, for long-continued selection within lines of doubles brought to light very great changes in this frequency. Attempts to isolate by selection biotypic diversities in rate of fission within clones of doubles and of singles failed, although the coefficient of variation of fission rate was higher for doubles than for singles. The general picture of the genetic consequences of environmental action in the ciliate protozoan, Colpidinui caiupylmii, is strikingly sim- ilar to the picture in the rhabdocoel turbellarian, Stcnostoiuuin incau- datnni. The similarity in these and the difference of both from the genetic effects of environmental action in higher organisms were ascribed to the method of reproduction. In asexual reproduction hereditary GENETICS OF CHAIN FORMATION IN COLPIDIUM 211 changes may arise without altering the nature of the chromatin ; they may be due simply to changes in the number of units or arrangement of units in the self -perpetuating parts. The changes induced in Colpidnun and Stcnostountm were of this sort and, in this respect, resemble trans- locations, inversions, and reduplications in Drosophila. IX. BIBLIOGRAPHY CHATTON, M. EDOUARD ET MME., 1925a. L'action des facteurs externes stir les Infusoires. Le detcrminisme de la formation des chaines (dystomie) chez les Colpidium. C outfit, rend. Acad. Sci., 180: 1225. CHATTON, M. EDOUARD ET MME., 1925/7. L'action des facteurs externes sur les Infusoires. Le determinisme de la formation des chaines (dystomie) chez les Colpidium. Rei: Suissc dc Zool., 32: 99. JENNINGS, H. S., 1929. Genetics of the Protozoa. Bibliografihia Gen., 5: 105. PARPART, A. K., 1928. The Bacteriological Sterilization of Paramecium. Biol. Bull, 55: 113. RAFFEL, D., 1930. The Effect of Conjugation within a Clone of Paramecium aure- lia. Biol. Bull., 58: 293. SONNEBORN, T. M., 1930(7. Genetic Studies on Stenostomum incaudatum (nov. spec.). I. The nature and origin of differences among individuals formed during vegetative reproduction. Join: E.rfiei: Zoo!., 57: 57. SONNEBORN, T. M., 1930/>. Genetic Studies on Stenostomum incaudatum. II. The effects of lead acetate on the hereditary constitution. Join: Exper. Zool, 57: 409. CHROMOSOMES OF ARTIFICIALLY ACTIVATED EGGS OF URECHIS ALBERT TYLER (Prom the William G. Kcrckhoff Laboratories of the Biological Sciences, Cali- fornia Institute of Technology, Pasadena, California and the William G. Kerckhoff Marine Laboratory, Corona del Mar, California) The eggs of Urcchis that cleave and develop as a result of activation by dilute sea water have been previously shown (Tyler, 193 la) to be those which extrude no polar bodies. It would appear then that the embryos produced by such eggs might be tetraploid, diploid, or haploid, depending upon the behaviour of the chromosomes during the first two nuclear divisions. A cytological investigation of such eggs shows that the embryos are diploid in chromosome number, and that only one maturation division occurs. The preparations were made by a method used by Karl Belaf and similar to that described by him (1928). It consists of joining and later separating two cover-slips, one containing a drop of eggs and the other a drop of fixing fluid. The eggs are flattened to any desired extent and stick to the cover-slips, which can be handled in the same manner as slides containing sectioned material. Two types of eggs are produced as a result of activation with dilute sea water (Tyler, 1931a). In one type the initial behaviour is identical with that of the normally fertilized eggs, two polar bodies are produced but none of the eggs divide. .In the other type the initial behaviour is quite different from that of normally fertilized eggs; no polar bodies are produced but practically all the eggs of this type divide and develop. In making the cytological preparations of the eggs of the first type, use was made of the fact, previously reported (Tyler, 1931&), that an inverse relation exists between the total percentage of activation and the percentage of cleavage. Thus treatments giving 100 per cent activation produce only eggs of the first type which do not divide. For prepara- tions of the type which does not extrude polar bodies, the eggs had to be isolated from dishes containing also unactivated eggs and activated eggs of the first type. These eggs can be readily distinguished at an early stage and can be removed for cytological preparations before the time at which the first polar body appears in the eggs of the first type. The counts of chromosome number were generally made from polar views of anaphase groups inasmuch as precociously divided chromosomes in metaphase might cause difficulty. 212 CHROMOSOMES IN PARTHENOGENETIC URECHIS EGGS 213 THE EGGS THAT PRODUCE POLAR BODIES The behaviour of the chromosomes in the maturation division of the artificially activated eggs that extrude two polar bodies is identical with that of the normally fertilized eggs. The normal diploid number in Urechis is most probably thirty-six chromosomes and the haploid number eighteen. The variability in the chromosome numbers shown in the tables is undoubtedly due to errors in counting. The artificially activated eggs (last section of Table I) show the haploid number of TABLE I FIRST CLEAVAGE OF -NORMAL FERTILIZED EGGS FIRST POLAR DIVISION OF NORMAL FERTILIZED EGGS FIRST POLAR DIVISION OF PARTHENO- GENETIC EGGS Chromosome Number Number of Groups Chromosome Number Number of Groups Chromosome Number Number of Groups 33 4 15 3 14 2 34 6 16 2 16 5 35 3 17 8 17 10 36 7 18 8 18 6 37 2 19 1 19 1 chromosomes at the first maturation division. The second polar division also occurs normally and the egg is left with the haploid number of chromosomes, which form a nucleus and move into the center of the egg. A large monaster then forms at about fifteen minutes after the extrusion of the second polar body and the chromosomes distribute themselves irregularly about the astral rays. The monaster disappears and a vesicular nucleus is formed about ten minutes later. The monaster then reappears about twenty minutes later and a larger and variable number of chromosomes are seen. The monaster may dis- appear and reappear a third time. This behaviour is essentially similar to that described by Herlant (1918) in the sea-urchin for eggs activated by butyric acid. The failure of the eggs of this type to divide appears then to be due to the failure to form an amphiaster. THE EGGS THAT DIVIDE In the eggs of this type the germinal vesicle breaks down and tetrad chromosomes appear at about twenty to twenty-five minutes after treat- ment. At this time the first polar spindle appears in the control eggs. But no spindle is seen in the eggs of this type and at about ten to twenty minutes later the tetrads each form a small vesicular karyomere. The 14 214 ALBERT TYLER nucleolus generally persists as such throughout this time and about seventeen or eighteen karyomeres may be seen distributed throughout the egg. Later a single large nucleus is formed apparently by the fusion of the karyomeres. The nucleolus remains intact and is seen within the nucleus. This nucleus is generally about two-thirds of the size of the original germinal vesicle and has a granular appearance similar to that of the cytoplasm. The eggs remain in this condition for about an hour, after which the first cleavage spindle appears. TABLE II Parthenogenctlc Eggs FIRST CLEAVAGE SECOND CLEAVAGE » THIRD CLEAVAGE Chromosome Number Number of Groups Chromosome Number Number of Groups Chromosome Number Number of Groups 13 1 32 2 32 5 15 4 33 2 33 7 16 8 34 3 34 6 17 5 35 8 35 4 18 5 36 4 36 5 38 1 37 38 40 2 2 1 45 1 13 2 46 1 14 2 48 2 15 2 50 1 16 4 53 2 18 5 54 1 20 1 An attempt was made to determine whether any division of the chromosomes occurred prior to the first cleavage, and whether fusion of the egg nucleus with a submerged polar body nucleus such as de- scribed by Morris (1917) in Cumingia, occurred in Urechis. The evidence shows that such behaviour does not occur in Urechis. It is possible to determine this point with some certainty in Urechis inasmuch as the eggs which are to divide remain indented until just before the first cleavage. The indented eggs were therefore preserved at close intervals up to that time. No division figures or fusion of two nuclei were observed. Furthermore, the chromosomes on the first cleavage spindle have the appearance of tetrads and are eighteen in number. In anaphase they open out as typical dyads. CHROMOSOMES IN PARTHENOGENETIC URECHIS EGGS 215 It appears then that the first cleavage spindle is identical with the first polar spindle as far as the chromosomes are concerned, and the first division may be considered a maturation division. At the second cleavage the diploid number of chromosomes is usually seen. This is the case with the twenty anaphase chromosome groups of the five eggs listed in Table II. The chromosomes of these groups occur in more or less closely associated pairs. This means that the chromosomes had divided previous to this division. In other words the second cleavage is a mitotic rather than maturation and the cliploid number of chromosomes is retained. In two eggs listed in Table II eight anaphase groups gave chromosome counts approximating fifty- four, TABLE III Parthenogentic Eggs FOURTH CLEAVAGE EMBRYOS Chromosome Number Number of Groups Chromosome Number Number of Groups 31 2 28 2 32 6 29 1 33 8 30 4 34 7 32 6 35 6 33 3 36 10 34 7 37 4 35 2 38 3 36 3 39 1 38 2 40 1 the triploid number. The origin of such chromosome groups was not determined, and they were not encountered in the slides of the later stages. In the anaphase of the third cleavage the diploid number is again usually seen. Two eggs were obtained, however, in which the chromo- somes were of the haploid number. These must have arisen by the occurrence of both maturation divisions in the first two cleavages of the egg. The haploid number was not obtained again in the later stages of other eggs studied. The chromosome counts at the fourth cleavage of the egg, Table III. again approximated thirty-six, the diploid number. For the embryos, normal top-swimmers were isolated and preserved in the usual manner. Counts on fifteen pairs of anaphase groups (Table III) showed that the diploid number is present. 216 ALBERT TYLER The failure of the artificially activated egg that extrudes polar bodies to divide may be attributed to its retaining only the haploid number of chromosomes or to the possession of only the inner central body of the second polar spindle which is incapable of forming an amphiaster. The former is an unlikely assumption inasmuch as some haploid cleavage has been obtained in Urechis. But if the interpretation is based upon the behaviour of the central bodies,1 it is difficult to see why the cleavage of those eggs that produce no polar bodies should go beyond the four-cell stage. At this stage the centrosomes should be equivalent to the three that would have gone into the polar bodies and the one that remains in the egg. It might appear then that when the polar body central bodies come to lie within the egg cytoplasm they are capable of forming amphiasters. It may be pointed out in this connection that the first polar body in Urechis may or may not divide. Correspondingly in the artificially activated eggs, one of the cells of the two-cell stage often fails to divide. Similarly in the four-cell stage one of the cells often fails to divide corresponding to the egg cell that receives the inner central body of the second polar spindle. This again may be interpreted to mean that the first polar spindle is used for the first division. How- ever, in a large number of cases all four cells divide, and since no acces- sory asters have been observed in these eggs, it appears that the inner central body of the second polar spindle has regained the ability to form an amphiaster. SUMMARY 1. The embryos resulting from the artificial activation of Urechis eggs are diploid in chromosome number. 2. The diploid number is apparently obtained by the utilization of the first polar spindle for the first cleavage and the substitution of a mitotic division for the second maturation division. BIBLIOGRAPHY BELAR, K., 1928. Peterfi's Methodik der Wissenschaftlichen Biologic. Vol. I, p. 779. HERLANT, M., 1918. Comment agit la solution hypertonique dans la partheno genese experimentale (Methode de Loeb). I. Origine et signification des asters accessoires. Arch, de zool. r.r/rr. ct gen., 57: 511. MORRIS, M., 1917. A Cytological Study of Artificial Parthenogenesis in Cumingia. Jour. Expcr. Zool., 22: 1. 1 This discussion is based on the assumption of the genetic continuity of the centrosomes to which there no longer appears to be serious objection. CHROMOSOMES IN PARTHENOGENETIC URECHIS EGGS 217 TYLER, A., 1931a. The Production of Normal Embryos by Artificial Partheno- genesis in the Echiuroid, Urechis. Biol. Bull., 60: 187. TYLER, A., 1931&. The Relation Between Cleavage and Total Activation in Ar- tificially Activated Eggs of Urechis. Biol. Bull.. 61: 45. PRODUCTION OF CLEAVAGE BY SUPPRESSION OF THE POLAR BODIES IN ARTIFICIALLY ACTIVATED EGGS OF URECHIS ALBERT TYLER (From the William G. Kcrckhoff Laboratories of the Biological Sciences, Cali- fornia Institute of Technology, Pasadena, California, and the William G. Kerckhoff Marine Laboratory, Corona del Mar, California) It was suggested in an earlier paper (Tyler, 193 Ib) that those eggs which do not ordinarily divide as a result of artificial activation could be made to do so by suppression of the polar bodies. This was based on the fact that upon activation by means of dilute sea water only those eggs divide that extrude no polar bodies. The other type of egg produced by artificial activation behaves very much like the normally fertilized egg in its initial reactions to the treatment, extrudes two polar bodies, but does not divide. The results of the experiments reported here show that the polar bodies can be suppressed in such eggs by means of a second treatment with dilute sea water and that the eggs then divide. When Urechis eggs are treated with dilute sea water the percentage of the activated eggs that divide after various lengths of exposure bears an inverse relation to the total percentage of activation (Tyler, 1931&). Thus exposures resulting in 100 per cent activation give no cleavage, and the eggs are all of the type that extrudes both polar bodies. This simplifies the task of re-treating such eggs, since no unactivated eggs and no eggs of the type that divides are present in the dishes. The percentage of activation can be determined at about ten minutes after the initial treatment, and since the first polar body appears at thirty minutes at room temperature there is ample time for the second treat- ment. The first attempts at polar body suppression were made by means of anesthetics such as ether, phenyl urethane, and chloretone in various concentrations. Low temperature was later tried, as was also hypertonic sea water. These agents gave variable results, and in general although the polar bodies were suppressed while the eggs remained under treat- ment they often appeared later when the eggs were removed to normal sea water. Dilute sea water was then tried and this was found to be quite an effective agent for suppressing the polar bodies and producing cleavage. 218 CLEAVAGE BY SUPPRESSION OF POLAR BODIES 219 The concentrations used were 50 and 55 per cent sea water. Higher concentrations generally failed to suppress the polar bodies and lower concentrations appeared to injure the eggs. THE SECOND TREATMENT WITH DILUTE SEA WATER In these experiments the eggs were first treated for various lengths of time with 30 or 40 per cent sea water. The length of exposure resulting in 100 per cent activation is known fairly well from previous experiments, and so treatments ranging about the optimum time were used. The dishes were then examined to determine which actually TABLE I Rc-Treatuieiil with 55 per cent sea u'atcr. Eggs first treated for 2 minutes with 30 per cent sea water gave 100 per cent activation and all eggs later showed two polar bodies and no cleavage. First polar body out at 30 minutes ; second at 40 minutes, p.b. = polar body. Time after first treatment Length of second treatment Cleaved Uncleaved 0 p.b. 1 p.b. 2 p.b. 0 p.b. 1 p.b. 2 p.b. minutes minutes per cent per cent per cent per cent per cent per cent 25 5 0 0 0 0 0 100 25 10 2 3 0 0 ? 89 25 15 20 6 0.2 2 3 69 25 20 36 2 0 34 12 16 25 25 71 2 0 18 9 0 25 30 94 1 0 3 1 1 25 35 96 2 0 1 1 0 25 40 88 0 0.1 7 3 2 40 5 to 40 0 0 0 0 0 100 55 5 to 40 0 0 0 0 0 100 showed 100 per cent activation. Large samples of eggs were then transferred from such dishes at various times after the beginning of the first treatment to 50 or 55 per cent sea water. The eggs were usually exposed to the second treatment for 5 to 40 minutes. They were examined in the dilute sea water and later in normal sea water to determine whether or not the polar bodies were suppressed by the treat- ment. The usual precautions in regard to the amount of water trans- ferred with the eggs, etc. were taken. The results of one such series of experiments are given in Table I. The eggs were first treated with 30 per cent sea water and an exposure of two minutes was found to give 100 per cent activation. All of the 220 ALBERT TYLER eggs in this two-minute dish later showed two polar bodies. The first polar body'appeared at 30 minutes and the second at 40 minutes after the first treatment. A sample of eggs was transferred to 55 per cent at 25 minutes after the first treatment ; that is, 5 minutes before the first polar body was due to appear, and treated for various lengths of time. About an hour later the percentage of cleavage was determined and also the presence or absence of polar bodies. As shown in the table, when the eggs are exposed for 5 minutes to the second treatment with dilute sea water the polar bodies appear and no cleavage is obtained as in the controls. However, upon longer exposures fewer of the eggs show two polar bodies, and after exposure of 25 minutes or more practically none of the eggs show two polar bodies. At the same time the percentage of cleavage increases from zero to 98 per cent. The great majority of the cleaved eggs have no polar bodies. A small per- centage of the divided eggs have one polar body (column 4 in the table) and a very few of the divided eggs show two polar bodies. When the second treatment is applied at 40 or at 55 minutes after the first treatment (i.e., after extrusion of the second polar body) the results are the same as for the control eggs — none of the eggs divide. Fourteen series of experiments of the type illustrated by Table I were run and all gave similar results. Cleavage was obtained when the second treatment was applied before the time of extrusion of the first polar body and was continued until after the time of extrusion of the second polar body. When the eggs were given equivalent treatments at any time after the second polar body had appeared no cleavage was obtained. Cleavage was sometimes obtained when the treatment was applied after the extrusion of the first polar body. Eggs were also isolated after the second treatment according to the number of polar bodies they showed, and of 200 eggs examined cleavage was obtained in 90 per cent of the eggs that showed no polar bodies, 15 per cent of the eggs with one polar body, and none of the eggs with two polar bodies. At the first cleavage the doubly treated eggs divided into two or three cells. Of 400 eggs on which counts were made 65 per cent divided into two cells and 35 per cent into three cells. Cleavage often stopped in the four-cell stage. Large numbers of abnormal top and bottom- swimmers but no normal embryos were obtained from the re-treated eggs. It is evident then that when the polar bodies are suppressed by means of a second treatment the eggs are then capable of division. Suppression of one polar body appears to be less effective in this regard than suppression of both. CLEAVAGE BY SUPPRESSION OF POLAR BODIES 221 Cytological preparations were made of the doubly treated eggs ac- cording to the method previously described (Tyler, 1932) in order to determine the behaviour of the chromosomes and centrosomes. Eggs were removed for preservation directly from the dilute sea water and also after their return to normal sea water. In the eggs preserved within twenty minutes after the application of the second treatment the achromatic figure was generally not visible, and the chromosomes ap- peared as condensed bodies, similar to their metaphase condition. They formed a single group at the pole of the egg. In eggs removed at later times from the dilute sea water the chromosomes were often found in two groups of about 12 to 18 each although generally they appeared in one group of about 18 scattered about in the polar region. "When the eggs were returned to normal sea water at 30 minutes after the second treatment and later preserved, they first showed two chromosome groups which were generally associated with two separate asters. At later stages the eggs showed a single group of chromosomes, presum- ably due to the fusion of the two separate groups. The asters are usually not visible at this time. At the time of cleavage an amphiaster develops, and the chromosomes are seen distributed irregularly about the spindle. The first cleavage divides the chromosomes irregularly and counts of anaphase groups ranged from 8 to 40, the two groups some- times containing equal numbers and at other times radically different numbers of chromosomes. Later stages were not followed. The examination of the cytological preparations shows that when the polar bodies are suppressed, the chromosomes first separate into two groups which later come together and distribute themselves more or less irregularly about the first cleavage spindle. DISCUSSION An important question involved in the cleavage of artificially acti- vated eggs concerns the origin of the amphiaster. The parthenogenesis experiments of Herlant (1918), Fry (1925), and others on the echino- derm egg are generally taken to mean that central bodies and asters may arise dc novo and either combine or divide to form an amphiaster. Although this argues against Boveri's view of the genetic continuity of the central bodies, more evidence has recently been presented in its favor from other sources (Sturdivant, 1931; Wilson and Huettner, 1931; Pollister, 1930; and Johnson, 1931). The parthenogenesis ex- periments on Cumingia (Morris, 1917; Heilbrunn, 1925) and on Urechis (Tyler, 1931a) show that cleavage is obtained when the eggs fail to extrude polar bodies. The question arises as to whether in such cases ALBERT TYLER the first cleavage spindle develops dc iwvo or whether it is directly continuous with the first polar spindle. Evidence of the similarity of the first cleavage spindle of such eggs and the normal first polar spindle has been previously presented (Tyler, 1932). In the results presented here it was shown that suppression of the polar divisions enables eggs to divide which would not ordinarily do so. The cytological work is insufficient to determine whether when the polar divisions are suppressed the first maturation spindle is converted into the first cleavage spindle. The two separate asters observed when the polar bodies are suppressed may have been derived from the poles of the first maturation spindle or may have arisen dc novo. The fact that similar treatments applied after polar body extrusion do not produce such effects favors the former view, but in the absence of more detailed cytological evidence the ques- tion as to the origin of amphiaster in the doubly treated eggs still remains open. The double treatment used here obviously differs from Loeb's clas- sical double treatment for sea-urchin eggs. In these experiments the agent used for the second treatment was of the same type as that used for the first; and its effect was to enable eggs to divide by suppressing the polar bodies. Moreover, Just (1922) has clearly shown that in the sea-urchin egg only a single treatment is necessary, whereas for several different agents used for single treatments on Urechis, the optimally activated eggs do not divide. Thus hypertonic sea water alone gives similar results to hypotonic sea water. Hypertonic sea water was also tried on optimally activated eggs after the extrusion of the polar bodies, but no cleavage occurred. SUMMARY 1. The polar bodies can be suppressed in artificially activated eggs of Urechis by means of a second treatment with dilute sea water. 2. The treatment must be applied before the time of extrusion of the first polar body and continued until after the time of extrusion of the second. 3. The eggs in which the polar bodies are thus suppressed undergo cleavage whereas ordinarily they would not do so. 4. Similar second treatments applied after the time of the extrusion of the second polar body do not induce cleavage. BIBLIOGRAPHY FRY, H. J., 1925. Asters in Artificial Parthenogenesis. I. Origin of the amphi- aster in eggs of Echinarachnius parma. Jour. E.rpcr. ZooL, 43: 11. HEILBRUNN, L. V.. 1925. Studies in Artificial Parthenogenesis. IV. Heat par- thenogenesis. Jour. E.rpcr. ZooL, 41: 243. CLEAVAGE BY SUPPRESSION OF POLAR BODIES HERLANT, M., 1918. Comment agit la solution hypertonique dans la parthenogenese experimentale (Methode de Loeb). I. Origine et signification des asters accessoires. Arch, dc zool. ex per. ct gen., 57: 511. JOHNSON, H. H., 1931. Centrioles and Other Cytoplasmic Components of the Male Germ Cells of the Gryllidae. Zcitschr. f. iviss. Zool., 140: 115. JUST, E. E., 1922. Initiation of Development in the Egg of Arbacia. I. Effect of hypertonic sea-water in producing membrane separation, cleavage, and top-swimming plutei. Biol. Bull., 43: 384. LOEB, J., 1913. Artificial Parthenogenesis and Fertilization. University of Chi- cago Press. MORRIS, M., 1917. A Cytological Study of Artificial Parthenogenesis in Cumingia. Jour. E.vt>cr. Zool., 22: 1. POLLISTER, A. W., 1930. Cytoplasmic Phenomena in the Spermatogenesis of Gerris. Jour. Morph., 49: 455. STURDIVANT, H. P., 1931. Central Bodies in the Sperm-Forming Divisions of Ascaris. Science, 73: 417. TYLER, A.. 1931a. The Production of Normal Embryos by Artificial Partheno- genesis in the Echiuroid, Urechis. Biol. Bull., 60: 187. TYLER, A., 1931b. The Relation between Cleavage and Total Activation in Ar- tificially Activated Eggs of Urechis. Biol. Bull.. 61: 45. TYLER, A., 1932. Chromosomes of Artificially Activated Eggs of Urechis. Biol. Bull., 63: 212. WILSON, E. B., and A. F. HUETTNER, 1931. The Central Bodies Again. Science, 73: 447. OSMOTIC PROPERTIES OF THE ERYTHROCYTE V. THE RATE OF HEMOLYSIS IN HYPO-TONIC SOLUTIONS OF ELECTROLYTES M. H. JACOBS AND ARTHUR K. PARPART (From the Department of Physiology, University of Pennsylvania, and the Marine Biological Laboratory, Woods Hole, Massachusetts) In an earlier paper in this series (Jacobs, 1932) it has been shown that, on the assumption that the rate of entrance of water into the erythrocyte in accordance with simple osmotic laws is the factor of chief importance in determining the rate of osmotic hemolysis, the theoretical relation between the time at which some given degree of hemolysis is attained and the osmolar concentration of the surrounding solution ought to be given by the equation : 0. C0-C R-- or, if the external medium be water alone, by where c0 is the osmolar concentration of the solution in osmotic equilib- rium with the normal erythrocyte, R the ratio of this concentration to that which will just cause the given degree of hemolysis, C the osmolar concentration of the external medium, F0 the initial effective osmotic volume of the cell, A its area (assumed to be constant — a not unrea- sonable assumption in the case of the biconcave erythrocyte) and k the permeability constant of the erythrocyte for water; that is, a numerical measure of the amount of water that would with unit difference in osmotic pressure between the cell and its surroundings pass through unit surface in unit of time. Since F0 and A are frequently not ac- curately known separately, the expression kA/V0 may for many pur- poses be used as a secondary constant, k', whose calculated values over a range of concentrations give indication in the same way as do those of k of the applicability of the equations in question. 224 OSMOTIC HEMOLYSIS In the case of hypotonic solutions of non-electrolytes, it has already been shown (Jacobs, 1932) that the observed times of hemolysis over a wide range of concentrations are in fairly good agreement with those predicted by means of the equations, if allowance be made for a decided increase in the " osmotic resistance " of the cells produced by exposure to such solutions. Since there is some reason to believe that this in- crease in resistance may itself be osmotic in nature, there is no need at present to postulate non-osmotic factors to account fairly well for the observed results with non-electrolytes. In the case of electrolytes, how- ever, which will be discussed in the present paper, conditions are some- what different. In passing from water through a series of hypotonic solutions of, for example, NaCl of increasing concentration, the prop- erties of the erythrocyte undergo a change, expressed quantitatively by a change in the value of the calculated permeability constant, which seems to depend on other than osmotic factors. Above a certain con- centration— roughly 0.02M in the case of NaCl — the behavior of the erythrocyte is in excellent agreement with that predicted by means of the equations ; that is to say, a constant calculated value of k' is obtained. Below this point, however, there is a fairly rapid increase in the value of k' with decreasing concentration which ceases only at very great dilutions of the electrolyte. This inconstancy of k', which almost cer- tainly depends upon non-osmotic factors, and which is influenced to a striking extent by the valence of the cations present in the solution, has been very briefly mentioned in a previous preliminary paper (Jacobs, 1930) but has not hitherto been discussed at any length. We believe that it is of possible significance not only in connection with the problem of hemolysis but with certain larger ones having to do with the general question of cell permeability as well. In the experiments here described, as well as in others omitted for lack of space, the blood used was that of the ox, obtained from freshly slaughtered animals, defibrinated immediately, and kept until needed in a refrigerator. A smaller number of experiments on the blood of man and of several other mammals gave essentially similar results. All ob- servations were made at 20° C. ± 0.2° with the employment of exactly the same technique as that already described in the fourth paper of the present series (Jacobs, 1932), which may be consulted for further details. II In Fig. 1 are presented the results obtained on the same sample of blood with sucrose on the one hand and with NaCl on the other. In order that the results may be strictly comparable osmotically, the ob- served times of hemolysis are plotted as ordinates, not against the 226 M. H. JACOBS AND A. K. PARPART concentrations of the two solutions, but rather against their freezing point depressions. The latter were calculated by the empirical equation for NaCl: A=3.6C--1.3C2 and for sucrose A = = 1.S6C + 0.2C2, which for the concentration range actually employed give a fairly satis- factory agreement with published freezing point data. 10 9 a 7 c 0 5 y (9 «0 4 E 2 F 0.1' 0.2° A in decrees C. FIG. 1. Rate of hemolysis of ox erythrocytes at 20° C. in solutions of sucrose and of NaCl. One part of blood to approximately 500 parts of solution. Ordinates represent times of 75 per cent hemolysis in seconds and abscissae calculated freezing point depressions of solutions. Two things are immediately apparent from the figure. The first is that over most of the range covered by the experiments hemolysis occurs far more slowly in NaCl solutions than in those of sugar of the same osmotic pressure. This difference is especially striking in the most dilute solutions (e.g., of A = = 0.1° or less) where the osmotic effect of the solute as calculated by equation 1 is almost negligible, and where the effect actually observed with non-electrolytes is equally insignificant. OSMOTIC HEMOLYSIS 227 but where that found with electrolytes is very pronounced. As will be shown later, this electrolyte effect, which is exerted on the rate of hemolysis rather than on the position of final equilibrium of the system, is especially marked when the valence of the cations present is greater than one. The second difference between the two curves to which reference has been made briefly above and at greater length in the preceding paper (Jacobs, 1932) is the lower critical hemolytic concentration, i.e., the higher osmotic resistance of the cells, in the case of the non-electrolyte solution. In this particular case the value of A for which 75 per cent hemolysis just failed to occur was 0.324° for NaCl and 0.280° for sugar. This effect, which is of the " equilibrium " type, is obviously in the opposite direction from that of the first or " rate " effect, since electrolytes within the range where it is operative tend to favor rather than to oppose hemolysis. Because of the different natures of the two effects, the curves in Fig. 1 cross at a A value of about 0.26° for which the time of hemolysis is equal in the two solutions. Above this point there is a relatively narrow concentration range within which hemolysis actually occurs more rapidly in the presence than in the absence of the non-electrolyte. It is to be noted, therefore, that the observed rate of the hemolytic process may be affected by a mere shift in the position of final equilibrium of the system. Similar cases have been discussed else- where by the authors (Jacobs, 1928, 1931 ; Jacobs and Parpart, 1932). The curve for NaCl in Fig. 1 shows very clearly the general rela- tion between the concentration of a typical electrolyte solution and the time required for it to produce hemolysis ; but for a more exact analysis of the extent to which such results are in agreement with osmotic laws it is necessary to employ more strictly mathematical methods. In Table I there have, therefore, been calculated by means of equations 1 and 2 for experiments with NaCl involving three separate samples of blood, values of the constant //, whose meaning is explained above and whose constancy over a given range may be taken as an indication of the ap- plicability for this range of simple osmotic laws. The value of R em- ployed for the calculations in each case was taken as the ratio of the freezing point depression of ox plasma (approximately 0.58° C.) to the freezing point depression of the NaCl solution in which, for the blood in question, the final degree of hemolysis was 75 per cent ; this critical hemolytic concentration being determined for each sample of blood by a separate experiment. It was mentioned in the previous paper that a greater constancy of k' is obtained with non-electrolyte solutions if a somewhat smaller value of R than this be employed ; but the theoretical justification for this latter procedure is rather questionable, and in the 228 M. H. JACOBS AND A. K. PARPART calculation in that paper of the true permeability constant, k, the same R was used as that here adopted. It should be emphasized that in view of the complexity of the material and of the various simplifying assump- tions made in deriving the equations a perfect agreement between theory and observation is never to be expected. For the present, therefore, it seems advisable to use the value of R which is most simply defined and most easily determined, even though a slightly different value may fit the data rather better in some particular cases. TABLE I Effect of the concentration of XaCl solutions on the time required for 75 per cent hemolysis of ox blood at 20° C. One part of blood to approximately 500 parts of solution. Each time is the average of four determinations. Concentration A Experiment 1 R = 1.63 Experiment 2 R = 1.79 Experiment 3 R = 1.69 Time seconds k' Time seconds k' Time seconds k' 0.00 0.000 1.35 1.06 1.42 1.34 1.30 1.23 0.005 0.018 2.48 0.62 3.00 0.64 2.28 0.73 0.01 0.036 2.72 0.58 3.75 0.55 2.68 0.64 0.02 0.072 3.60 0.48 4.82 0.48 3.78 0.51 0.03 0.107 4.70 0.41 5.65 0.46 4.68 0.46 0.04 0.142 5.12 0.42 6.15 0.48 5.05 0.48 0.05 0.177 5.65 0.43 6.88 0.49 5.60 0.50 0.06 0.211 6.40 0.44 7.85 0.52 6.68 0.48 0.07 0.246 7.80 0.44 10.62 0.47 7.48 0.53 0.08 0.280 11.02 0.39 35.90 0.20 10.78 0.48 0.09 0.314 30.22 0.20 — — - 130. 0.06 It will be noted in Table I that the value of k' for water alone in all three experiments is relatively high, i.e., 1.06, 1.34, and 1.23, respec- tively. These values may be compared with those of 1.16 to 1.48 found in the previous paper, when R was similarly determined, for water and fora wide range of concentrations of sugar solutions. It will be further noted that whereas with the non-electrolyte discussed in the earlier paper no appreciable change in A'' occurred in passing from water to solutions of a concentration of, say, 0.04M, in the case of NaCl, an enormous change appears on passing to a concentration of only 0.005M; and a further, though much slighter, change by an additional increase in the concentration to 0.01 M. Even allowing for the fact that the osmotic pressure of an NaCl solution may be twice as great as that of a sugar solution of the same concentration, it is evident that the striking retarda- tion of hemolysis caused by very dilute solutions of NaCl can scarcely be osmotic in nature. OSMOTIC HEMOLYSIS 229 Passing over the narrow range of concentrations from zero to 0.01 M or 0.02M, within which k' undergoes a considerable change in magni- tude, we find that for all higher concentrations up to 0.07M or 0.08 M the value of k' is not only remarkably constant for a given experiment but that the values obtained with different samples of blood are in good quantitative agreement. It is difficult to believe that the constancy of k' over such a wide range of concentrations is due merely to chance. The most reasonable interpretation of the facts is that within this ex- tensive range the concentration of an NaCl solution is related to the time of hemolysis by simple osmotic laws, as has already been found to be the case (with certain limitations) with non-electrolyte solutions. It is to be noted, however, that the value of the constant for NaCl solu- tions is only between one-half and one-third as great as for water and for non-electrolyte solutions. The same relation holds for the true permeability constant, k, which, for a given type of blood, is always a definite multiple of k' '. As to the complete lack of agreement between the last value of k' in each series with the remainder, it may be said that determinations of rates of hemolysis in solutions lying so close to the critical hemolytic concentration are notorious!}" unreliable, as has been pointed out by one of the authors elsewhere (Jacobs, 1928). Successive determina- tions under such conditions, even when carefully made, show such rela- tively enormous differences as to render exact quantitative work in this region almost hopeless. It is not unlikely that the very low values of k' at the highest concentrations, where the time of hemolysis exceeds about 10 seconds, may be significant, possibly indicating an escape of salts from the cell with a consequent retardation of hemolysis (see in this connection Ponder and Saslow, 1931) ; but in view of the difficulty of obtaining accurate data under these conditions we prefer to leave this point unsettled for the present. The important fact remains, neverthe- less, that over a wide range the effect of the concentration of NaCl solu- tions on the rate of hemolysis is in good agreement with that demanded by simple osmotic laws. Ill Turning now to the region of the lowest concentrations (i.e.. all be- low about 0.02M), it is apparent that in this region small changes in the concentration of the electrolyte solution affect the rate of hemolysis in a manner that is not at all in agreement with equations 1 and 2. As a matter of fact, such effects extend to much more dilute solutions than any included in Table I and are, as will now be shown, intimately related to the valence of the cations present. 15 230 M. H. JACOBS AND A. K. PARPART In Fig. 2 are presented the results of a typical experiment with a single sample of blood in which the time of hemolysis was determined in various hypotonic solutions of NaCl, Na2SO4, CaCL, MgCL and MgSO4. Since over most of the range employed osmotic effects must obviously be very slight, actual concentrations rather than freezing point depressions are used in the figure as abscissae. Furthermore, in order that a wide range of concentrations, including those of a number of extremelv dilute solutions, mav be covered, the concentrations are .0 3 -o C o C 0 H O.OOO156 0.00031} O.OOOfe23 O.OOI25 O.OO25 O.OO5 O.O1 Concentrotlon in mol*. per liter. O.02 O.O4- FIG. 2. Rate of hemolysis of ox erythrocytes at 20° C. in solutions of various salts. One part of blood to approximately 500 parts of solution. Ordinates repre- sent times of 75 per cent hemolysis in seconds and abscissas concentrations of solu- tions in mols per liter. plotted logarithmically, I.e., equal distances along the axis of abscissae are taken to represent equal multiples of concentrations rather than equal arithmetical increments. The figure is therefore comparable with those of Loeb (1922) to which reference will be made below. Included in the figure for comparison is a curve, labeled 6", which indicates the calculated, and also approximately the observed, effects of sugar solu- tions having the osmotic pressures of the indicated concentrations of NaCl. It will be noted that the true osmotic effects, which alone are found in such solutions, are entirely negligible over most of the range OSMOTIC HEMOLYSIS 231 covered by the figure and that most of the effects of the electrolyte solutions must therefore he of a different nature. An inspection of 'Fig. 2 brings out several additional points of in- terest. The first is that the salts fall into two sharply-separated groups, both with respect to the concentration at which a visible retardation of hemolysis first appears and with respect to the magnitude of the re- tardation at any given concentration. Thus, with CaCL, MgCL and MgSO4 a retardation of the order of 0.5 second or 40 per cent is present at a concentration of 0.00015M. A similar retardation is not reached with NaCl and Na,SO4 below a concentration of approximately 0.003M, and no detectable effect of any sort is found with either of the latter salts, or with KC1, which was studied in other experiments, below a concentration of about 0.001 M. Throughout the entire range em- ployed the relatively greater effectiveness of the salts of Ca and Mg is most marked. With salts of this type the valence of the cation appears to be the factor of chief importance, since there is little difference be- tween MgCl, and MgSO4. In the case of NaCl and Na2SO4, both of which are rather widely separated in their properties from the salts just mentioned, it would appear that Na.,SO4 is considerably more effective at a given concentra- tion than is NaCl. This difference is probably to be attributed to the fact that the salt of the dibasic acid furnishes twice as many cations as that of the monobasic acid, the cation being, as already indicated, the ion of chief effectiveness in influencing the rate of hemolysis. If in plotting the two curves the concentrations of the Na ions had been used as abscissae rather than the molecular concentrations, the curve for Na2SO4 would have been shifted to the right by an amount equal to that between two successive indicated concentrations ; and in that case the two curves would have almost coincided. In several other experi- ments, not described here, the times of hemolysis for Na2SO4 in the region below 0.01 M where osmotic effects are negligible were found to be somewhat below those for NaCl at the same Na' ion concentration. In other words, with the same concentration of Na', SO4" at times seemed, if anything, to favor hemolysis as compared with Cl', though it is to be noted that the concentrations of the two anions under these conditions were no longer the same and the differences were at best slight. In the case of trivalent cations, a number of experiments have been made with Al'", but the results are too complex to be discussed here, since they involve H' ion effects, agglutination of the erythrocytes, and other complications that have little bearing on the present problem. It may be mentioned, however, that in its ability to retard hemolysis at 232 M. H. JACOBS AND A. K. PARPART very low concentrations, Al"', under proper conditions, may very con- siderably exceed the bivalent ions. With it a distinct retardation of hemolysis is at times obtained at concentrations as low as 0.00001 M. The rather complicated nature of the effects of Al salts upon the eryth- rocyte will be discussed in detail elsewhere. In addition to the experiments here described, a considerable num- ber of others of the same general type have been performed. Because of the great rapidity of the hemolytic process in water and very dilute solutions, the quantitative accuracy of such experiments is not always as great as might be desired, and there are some slight discrepancies from experiment to experiment; but, on the whole, the results are in very satisfactory agreement and bear out the conclusion here reached, namely, that in dilute solutions cations tend to retard osmotic hemolysis in some non-osmotic manner with an effectiveness that increases greatly with an increase of their valence from one to two, and that anions have comparatively little influence on the process, though in some cases they seem with increasing valence slightly to favor it. IV As to the cause of the retardation of hemolysis produced by adding to distilled water electrolytes in concentrations from 0.01 M to 0.0001 M or even lower, it may be said with a fair degree of certainty that the osmotic pressure of the external solution* in such cases is a factor of little or no significance. This is indicated not only by the negligible osmotic effects of such solutions as calculated by means of equation 1, and as actually observed in the case of non-electrolytes, but by the enormous differences in the effectiveness of, for example, NaCl and MgSO4 at the same concentration, or of NaCl and CaCL at the same freezing point. The possibility nevertheless suggests itself that while in such cases the external osmotic pressure is of no importance, there might con- ceivably be produced by the solutions some indirect osmotic effects on the cells themselves which would influence the rate of the hemolytic process. We have already pointed out (Jacobs and Parpart, 1931) that the erythrocyte is unique among cells in the readiness with which its internal osmotic pressure is affected by apparently insignificant external changes of different sorts. Unfortunately for this explanation, such effects as might conceivably be produced in this way are, in the present case, in the wrong direction. As shown by the difference in the critical hemolytic concentration for electrolytes and for non-electrolytes (see Fig. 1 ) , the " equilibrium " effect of electrolyte solutions is in the direc- tion of favoring rather than of opposing hemolysis. An osmotic ex- OSMOTIC HEMOLYSIS planation of the observed results, either direct or indirect, seems there- fore definitely to be ruled out. A more plausible explanation, because it suggests analogies in both living and in non-living systems, is that the rate of entrance of water into the erythrocyte is affected by low concentrations of ions in a man- ner similar to that observed by Lucke and McCutcheon (1929) in the case of the Arbacia egg and by Loeb (1922) in the case of collodion- gelatin membranes on the alkaline side of the isoelectric point of the gelatin. The former workers have reported that cations inhibit the passage of water into the Arbacia egg to an extent which increases with their valence, while anions behave in the opposite manner. In the case of collodion-gelatin membranes, where the factors concerned are obvi- ously of a very simple physico-chemical nature, the results obtained are much the same; the nature of these effects has been discussed at length by Loeb. The erythrocyte differs from both the Arbacia egg and the artificial membrane in the much less prominent, and indeed somewhat doubtful, effect upon it of anions as compared with cations; but the striking difference between the ions of the alkali metals, on the one hand, and those of the alkaline earths on the other is found in all three cases, and may conceivably be due to the same causes. An alternative explanation is that the effect of ions is on the rate of escape of hemoglobin from the cell rather than on the rate of entrance of water into it (see in this connection the discussion by Jacobs and Parpart, 1932, of the effect of narcotics on hemolysis). This explana- tion, however, while not completely ruled out by the existing evidence, seems to us to be less probable than the other one in view of the fact that the " equilibrium " effect of electrolytes on hemolysis, unlike that of narcotics, is in the opposite direction from the " rate " effect. What- ever the explanation of the effect of traces of electrolytes on the rate of hemolysis may ultimately prove to be, however, the observed facts are themselves entirely definite; and the non-osmotic factors shown to be concerned in the process would seem to be worthy of consideration in connection with theoretical discussions of the nature of cell perme- ability. SUMMARY 1. In NaCl solutions of concentrations from about 0.02M to 0.07M or 0.08M the rate of hemolysis of ox blood is related to the concentra- tion of the solution as if the process were governed by simple osmotic laws. 2. The permeability constant for water over this range is between one-half and one-third as great as that previously found for non-electro- 234 M. H. JACOBS AND A. K. PARPART lyte solutions. At concentrations below 0.02M the calculated " con- stant " changes with the concentration of the solution in a manner indicative of the presence of non-osmotic factors of some sort. 3. The retarding effect upon hemolysis of dilute solutions of electro- lytes increases rapidly with the valence of the cations present. The valence of the anions is much less important but, if anything, acts in the opposite sense. 4. The tentative suggestion is offered that under certain conditions ionic forces may modify to an appreciable extent the rate of the osmotic intake of water by the erythrocyte. BIBLIOGRAPHY JACOBS, M. H., 1928. Am. Nat., 62: 289. JACOBS, M. H., 1930. Am. Jour. Med. Sci., 179: 302. JACOBS, M. H., 1931. Ergebn. d. Blol., 7: 1. JACOBS, M. H., 1932. Blol. Bull., 62: 178. JACOBS, M. H., AND A. K. PARPART, 1931. Blol Bull., 60: 95. JACOBS, M. H., AND A. K. PARPART, 1932. Biol Bull., 62: 313. LOEB, J., 1922. Jour. Gen. Physlol, 4: 463. LUCRE, B., AND M. MCCUTCHEON, 1929. Jour. Gen. Physlol, 12: 571. PONDER, E., AND G. SASLOW, 1931. Jour. Physio!., 73: 267. HIBERNATION AND DIAPAUSE PHYSIOLOGICAL CHANGES DURING HIBERNATION AND DIAPAUSE IN THE MUD-DAUBER WASP, SCELIPHRON C/EMENTARIUM (HVMENOPTERA) JOSEPH HALL BODINE AND TITUS C. EVANS ZOOLOGICAL LABORATORY, STATE UNIVERSITY OF IOWA It has been known for some time that many organisms enter periods of inactivity during winter or upon exposures to low temperatures. In some cases it has been clearly demonstrated that periods of rest or dia- pause are quite independent of external temperatures for their occur- rence. The parts played by heredity and environment in these phe- nomena have also been much discussed. Quantitative physiological observations on single individuals during these periods of quiescence, however, have been carried out on but few forms and particularly is this true for lower forms, especially the insects (Uvarov, 1931 ; Dreyer, 1932; Ashbel, 1932, etc.). The present paper is concerned with results of a detailed study of certain physiological changes taking place during the developmental life cycle of the common yellow-legged mud-dauber wasp, Sccliphron cfciucntarluni. The mud-dauber wasp, S. ccenicntarhiin, is extremely favorable mate- rial for physiological investigations since it can be readily obtained in its developmental stages in large numbers and is easily kept in the lab- oratory with a minimum of care. Its life-cycle is relatively simple. After completion of the mud nest, spiders are captured, paralyzed, and put into the individual cells of the nest. A single egg is laid by the female wasp on the abdomen of the first spider introduced. Other paralyzed spiders are added to the cell as food for the developing larva and the cell is then sealed. The egg hatches in a very few days (de- pending on external temperature) and the larva eats voraciously of the enclosed spiders and quickly attains the stage at which it spins a cocoon about itself. Within this cocoon case the animal goes through the remainder of its larval, prepupal and pupal life and eventually emerges as an adult wasp. The length of the larval stage is of considerable interest, since, normally, animals hatching late in the season (August- September) hibernate in this stage. Larvae from eggs laid early in the summer (June-July), in most instances, do not go through hibernation in the larval stage but develop uniformly, emerging out-of-doors in 235 236 J. H. BODINE AND T. C. EVANS from 19 to 25 days (Ran, 1918). It is reasonable to assume that two varieties are normally produced — one that goes through development from egg to adult at a fairly uniform rate with no marked periods of cessation — another that normally hibernates out-of-doors in the larval stage and thus does not develop during such periods. Since it is almost impossible to collect all the eggs laid by a single female wasp during the entire laying period, one cannot say definitely that both types of eggs are laid by the same individual. Indirect evidence would seem to indicate, however, that eggs laid early in the season invariably go through development without a cessation while those laid later in the season are usually of the diapause type. A somewhat similar observation for the codling-moth has been reported by Glenn (1922), Shelf ord (1927), and others. Both types of individuals have been obtained in Iowa as well as from New Jersey, Pennsylvania, Maryland, and Texas. Experi- mental materials for this investigation were taken over a period of three successive years. The procedure followed in obtaining animals was as follows : The mud-nests were collected at intervals throughout the year and larvae in various stages of their developmental life cycles, as far as diapause was concerned, were thus obtained. The animals were completely removed from the cocoon cases and kept separately in shell vials or gelatin cap- sules at known constant temperatures throughout the experiments. Eggs laid early in the season (July) were obtained at the time of lay- ing and after hatching were fed paralyzed spiders taken from the nests. The larvae in these experiments were kept at constant temperatures and grew in quite a normal fashion. It was thus possible to obtain in this manner accurately timed organisms for comparison with those taken at random. Inasmuch as the last larval stage is a non- feeding one the organism is relatively easy to keep under laboratory conditions. Body weights and morphological and physiological histories were kept for individual larva?. Oxygen consumption was determined by the modified Krogh manometer (Bodine, 1929). But one larva was used at a time and this always in the same manometer throughout the period of the experiment. Some 225 to 250 larvae have been individually studied. Inasmuch as the results obtained are qualitatively similar it seems de- sirable to express them graphically. This method shows most clearly the general course of the physiological and other changes followed by the larvae throughout their development. EXPERIMENTAL RESULTS I. Ar on-diet pause Type of Organ is in Oxygen consumption and body weight changes during the entire HIBERNATION AND DIAPAUSE 237 developmental life cycle of non-diapause individuals have been studied. All results obtained have been qualitatively similar so that only typical cases will be presented. At 28° C. the length of time required for the entire development from laying of egg to emergence of wasp is approxi- mately 25 to 28 days. The oxygen consumption of the egg during de- velopment steadily increases up to the time of hatching. During the active feeding and growing periods of the larva, body movements are so marked that satisfactory measurement of the rates of oxygen consump- ieo 170 160 130 no so 10' 0 V. 10 SO 30 40 60 80 70 80 90 100 1 io TIME - DATS •330 310 § M»o g 270 ?! o 250 Jj 230 2 a 210 2 o 190 * 170 160 FIG. 1. Rates of oxygen consumption and body weights during the development of diapause and non-diapause types of animals at constant temperature of 28° C. Ordinates at left, millimeters change of manometer fluid per hour per gram or- ganism (same manometer used throughout experiment; to convert readings into actual amounts of oxygen, results are multiplied by factor for the manometer in question). Ordinates at right, body weight in milligrams. Abscissae, time in days indicated. Letters indicate different periods in life cycle of organisms. Large letters for diapause type, small ones for non-diapause type. A = spinning of cocoon. B = beginning of diapause. C = end of diapause. D = formation of pre- pupa. E = pupation. F —. pigmented pupa ready to emerge. W • = body weight. - — oxygen consumption, non-diapause type. — = oxygen con- sumption, diapause type. = body weight, non-diapause type. body weight, diapause type. For further description, see text. tion can not be obtained. After the spinning of the cocoon, the animal becomes less active and it is largely for this reason that the results pre- sented in this paper begin at this point in the organism's development. Prior to and during the spinning of the cocoon, the alimentary canal is emptied of waste materials and this too adds much to the desirability of beginning measurements of oxygen consumption rates and changes in body weights at this stage. In Fig. 1 the rates of oxygen consumption and body weights of a typical non-diapause type individual are graphically shown. From an inspection of this figure it will be noted that the rates of oxygen con- sumption during the spinning of the cocoon are at first high but after 238 J. H. BODINE AND T. C. EVANS spinning they quickly drop to a minimum. This minimum value is at the time the animal reaches the prepupal stage. After the prepupal stage an increase in the rate of oxygen consumption occurs, during which the animal prepares for pupation. During pupation a drop in rate of oxygen consumption again takes place and this is followed by a steady and marked increase up to emergence of the wasp. This drop in oxygen consumption rate during pupation, or the so-called U-shaped oxygen consumption curve, is quite characteristic for this phenomenon since it has heen reported for many other forms undergoing complete metamorphosis (Taylor and Steinhach, 1931, and others). The pre- pupal drop in the oxygen consumption curve is equally characteristic hut seems to have been reported for but few forms (Fink, 1925). It is thus seen that during the development of the wasp definite cycles or rhythms in rates of oxygen consumption occur which are closely corre- lated with the morphological stages through which the animal passes. Changes in body weight are also of considerable interest (Fig. 1, curve w). During the spinning of the cocoon a rather marked drop in weight occurs, due largely to the emptying of the alimentary canal. After this, loss in body weight is gradual but continuous up to the time of emer- gence. II. Diapause Type of Organism 4 As noted above, eggs laid late in the summer (August-September) usually produce animals showing a diapause. These organisms are quite similar to the non-diapause variety in their development, the most striking difference being the length of time necessary for development at constant high temperatures (20-35° C.). In Fig. 1 there is graphi- cally represented for a diapause individual rates of oxygen consumption and body weights during developmental stages comparable to those de- scribed above for the non-diapause organism. From an inspection of these curves it will be noted that a marked decrease in rate of oxygen consumption down to a minimum value occurs during the spinning of the cocoon and in preparation for diapause. This minimum value at constant high temperatures (20-35° C.) is always considerably lower than that reached by the non-diapause form even though the morphologi- cal stages are similar in both cases. During diapause at constant high temperatures (20-35° C.) minimum rates of oxygen consumption are practically constant and at 28° C., as shown in Fig. 1, last some 40 to 50 days. As noted further on, this minimum rate of oxygen consumption and its duration are modified to a considerable degree by different tem- peratures. Changes in body weights in diapause organisms are quali- tatively similar to those undergone by non-diapause individuals, the HIBERNATION AND DIAPAUSE 239 only difference being a marked period of almost constant body weight during diapause. In general, the physiological changes through which both types of organisms pass during development are, with the exception of diapause, quite similar. The rhythmic or cyclic changes correlated with larval, prepupal, and pupal changes are strikingly indicated in both. The dia- pause individuals, at 28° C. (as indicated in Fig. 1), require from five to six times longer for their development. Questions as to the total amounts of energy involved in the development of the two types of individuals are of considerable interest, but since it is almost impossible to secure all the eggs from the same female, many uncontrollable factors enter which tend to make such calculations hazardous. Differences in initial body weights, amounts of food stored in nests, and similar con- ditions make absolute comparisons, as far as the total oxygen consumed and number of days required for development, impractical for this form. Similar experiments carried out on the eggs of the silkworm by Ashbel (1932), however, show that the amounts of oxygen consumed and the number of days required for development are more or less con- stant for diapause and non-diapause eggs. III. Reactions of Diapause to Temperature Inasmuch as diapause seems independent of external temperatures for its occurrence it was thought desirable to determine what the effects of different temperatures would be on its duration and intensity. Rather extensive experiments have been carried out using diapause type animals taken from out-of-door environments and determining their rates of oxygen consumption throughout hibernation and growth periods. In addition, organisms in similar physiological states have been experi- mentally subjected to controlled temperatures and their responses stud- ied. 1. Diapause animals under out-of-door temperatures. — Diapause larvae in the same morphological stages were collected from nests dur- ing different periods of the year, from August to April, put at constant temperature (28° C.) and their rates of oxygen consumption and growth followed. Results, typical of such experiments, are graphically indi- cated in Fig. 2. An inspection of this figure shows that animals col- lected late in August have rather marked periods of low oxygen con- sumption rates (diapause) similar to those pointed out previously for diapause animals kept continuously at constant high temperature (Fig. 1). With the approach of low out-of-door temperatures in November the length of low oxygen consumption rates (diapause) at 28° C. gets progressively shorter. Animals put at 28° C. late in December, after 240 J. H. BODINE AND T. C. EVANS being exposed to rather long periods of low out-of-door temperatures, show no periods of low oxygen consumption rates but develop in quite a uniform and normal fashion. That this gradual shortening of the low oxygen consumption rate (diapause) is not due to the animals being of different developmental ages when collected at different periods of the year, can be easily demonstrated. Individuals taken under identical environmental conditions early in the fall and kept out-of-doors under observation for the remainder of the year always give progressively shorter and shorter periods of low oxygen consumption rates the later they are put at constant high temperature (28° C.). In other words, exposures to low out-of-door temperatures during winter in some way K 0 160 K a "• 130 pa D O 110 CB £ 90 Or 10 AUG 28 30 1 40 OCT1 50 60 -f 70 NOV 1 TIME - DATS 90 t 100 t 110 120 DEC 1 DEC 17 FIG. 2. Rates of oxygen consumption for diapause-type animals collected out- of-doors at different periods during -year and then placed at 28° C. until emergence. All animals in same morphological stage (hibernating larval stage) at time of collection. Ordinates, millimeters change of manometer fluid per hour per gram of organism (see note under Fig. 1). Abscissae, time of year and days indicated. Each curve represents oxygen consumption for a single larva. For further descrip- tion, see text. or other shorten the length of diapause or diapause progresses during exposures to low temperatures. No morphological or developmental changes are evident in the animals during the periods of low oxygen consumption rates (diapause). 2. Diapause animals under controlled temperatures. — Animals of known history, as far as diapause was concerned, were collected in large numbers and put at constant temperatures of 2° C. where they were left for varying periods. At different intervals of time some were taken from 2° C. and put at 28° C. and left there until emergence of the adult animal. Rates of oxygen consumption and development were carefully studied on animals thus treated. Figure 3 shows, graphically, results typical of such an experiment. From an examination of this figure it HIBERNATION AND DIAPAUSE 241 will be noted that exposures of diapause animals to 2° C. cause a marked shortening of the period of low oxygen consumption rates (diapause) when the organism is subsequently transferred to 28° C. The degree to which this period is shortened is conditioned more or less quanti- tatively by the length of exposure to 2° C. Short exposures (5-10 days) cause but little change in duration while long exposures (50 -f- days) cause complete disappearance of the period. No morphological changes can be noted in the organisms kept for some time at 2° C. If the exposure is too long (over 3 to 4 months) a rather high mortality results. This, however, is a much longer time than necessary for com- 180 S 2 180 O 5 60 § 8 40 60 70 TIME - DATS FIG. 3. Rates of oxygen consumption for diapause-type animals kept at 2° C. for varying periods and then put at 28° C. until emergence. All animals collected on same date at beginning of experiment and in same morphological stage (hiber- nating larval stage). Ordinates, millimeters change of manometer fluid per hour per gram of organism (see note under Fig. 1). Abscissa?, time in days indicated. All oxygen curves begin on day animal was transferred from 2° to 28° C. Period of exposure to 2° C. indicated by number of days from start of experiment to time oxygen curve begins. Each curve represents oxygen consumption for a single larva. For further description, see text. plete disappearance of the period of low oxygen consumption rates (dia- pause). It is thus evident from such results that diapause progresses or is influenced by exposure to low temperatures (2° C.) and that such action of low temperatures out-of-doors must be a factor in the normal reaction of the organism to its environment. Temperature seems to have little influence on the occurrence of diapause but it is unquestion- ably a factor for its duration. In another series of experiments, larv?e just entering diapause were collected from out-of-doors and put at constant high temperatures (35, 28. 25° C.) as well as at 2° C. For those kept at the higher tempera- tures, oxygen consumption rates were determined during- the entire developmental stages. In the case of those at 2° C., individuals were 242 J. H. BODINE AND T. C. EVANS taken at different time intervals and transferred to 25° C., at which temperature their oxygen consumption rates and development were fol- lowed. Particular attention was given to the time necessary for dia- pause to disappear (as judged by the lack of low oxygen consumption rates at 25° C.) at this low temperature (2° C.) so that a relative com- parison between the length of diapause at the different temperatures could be made. Figure 4 indicates graphically typical results of such a series of experiments. An examination of this figure shows that low 60 70 TIME - DAYS FIG. 4. Rates of oxygen consumption for diapause-type animals kept at con- stant temperatures as well as for those kept at 2° C. and subsequently transferred to 25° C. All animals collected at same time and in same morphological stage (hibernating larval stage). Orclinates, millimeters change of manometer fluid per hour per gram of organism (see note under Fig. 1). Abscissae, time in days indi- cated. Each curve represents oxygen consumption for a single larva. Curve A, animals kept at constant temperature of 35° C. from beginning of experiment until emergence. Curve B , animals kept at 28° C. from beginning of experiment until emergence. Curve C, animals kept at 2° C. from beginning of experiment and then transferred and kept at 25° C. until emergence. Curve shown indicates period of exposure to 25° C. Period of exposure to 2° C. indicated by number of days from start of experiment to time oxygen curve begins. Curve D, animals kept at 25° C. from beginning of experiment until emergence. For further description, see text. rates of oxygen consumption (diapause) occur at the different tempera- tures (25-28-35° C.). Of particular note is the relative length of dia- pause at these temperatures. At 35° it took approximately 33 days, at 28°, 35 days and at 25°, 85 days. A temperature coefficient of approxi- mately 2.5 has been found for the duration of diapause at these constant high temperatures. In the experiments at 2° C. it took approximately 75 to 80 days for diapause, which is about equal to the time required at 25° C. In all experiments it has been found that low temperatures (10-2° C.) are almost as efficient for the progress of diapause as are temperatures as high as 25° C. It is thus evident from these results that diapause occurs independently of the external temperature but that HIBERNATION AND DIAPAUSE 243 its duration is to a marked degree conditioned by temperature. The most unusual nature of the effect of temperature on it is the fact that low temperatures, much below the threshold values for development for this species, affect it in a marked degree. The relative rates of oxygen consumption at different temperatures are also of considerable interest. At 28 and 35° C. the average rates 145 a 125 3 a S us 105 M O 9b W g W 65 O 45 36 86 16 25 30 TEMPERATURE -C 35 FIG. 5. Effect of temperature on rates of oxygen consumption of diapause- type animals at different periods of life cycle. Orclinates, millimeters change of manometer fluid per hour per gram of organism (see note under Fig. 1). Abscissae, temperature, degrees centigrade. Curve A, animal at beginning of diapause. Curve B, animal in middle of diapause. Curve C , animal at end of diapause. Curve D, animal coming out of diapause entering prepupal stage. Curve E, animal in pre- pupal stage entering pupal stage. Curve F, animal in pupal stage. Curve G, ani- mal ready to emerge. For further description, see text. expressed in terms of millimeter changes in manometer fluid per hour per gram organism are 18 and 12 respectively, while at 25° C. they are 3. At 2° C. rates are so low that it is practically impossible to get satis- factory results over long periods of time. At temperatures of 25° C. and below, rates of oxygen consumption are always extremely low in comparison with those at temperatures above this value. Such marked differences are, perhaps, due to a regulatory mechanism which enables the animal under out-of-door conditions to better meet and endure ex- treme fluctuations of temperature. 244 J. H. BODINE AND T. C. EVANS During diapause, however, responses to short exposures to different constant temperatures are quite in contrast to those for organisms in which diapause has ceased. In Fig. 5 there is graphically represented the effects of different temperatures on the oxygen consumption rates of organisms in different stages of diapause. From an inspection of this figure it can be readily seen that marked responses are given only after the period of diapause is well advanced. When reactions to dif- ferent temperatures are first noted they are of small magnitudes (Fig. 5, C—D). As diapause progresses and begins to wane progressively greater responses are given. Curiously, temperature coefficients for the different periods are quite similar even though the actual amounts of oxygen are of greatly different magnitudes. During diapause the organism is apparently quite dependent on the particular physiological state at which it happens to be for its response to different temperatures. In early periods it is less responsive than later ones and only approaches the response normally given by non-diapause animals after the effects of diapause have disappeared. SUMMARY AND CONCLUSION 1. Rates of oxygen consumption and changes in body weight during the development of diapause and non-diapause types of the mud-dauber wasp, Sccliphron cccincntarluni, at constant temperatures are presented and compared. 2. Diapause seems, for its occurrence, independent of environmental temperatures, but its duration is conditioned to an appreciable degree by these factors. 3. Cyclic or rhythmic changes in oxygen consumption and body weight during the developmental life cycle of the wasp are pointed out. 4. Quantitative responses to temperature are modified to an extreme degree by the diapause phenomena. LITERATURE CITED 1 ASHBEL, R., 1932. Sul Ricambio Gassoso Delia Uova Di Bachi Da Seta (Bombyx Mori L.) II. Protoplasma, 15: 177. BODINE, J. H., 1929. Factors Influencing the Rate of Respiratory Metabolism of a Developing Egg (Orthoptera). Physiol. Zoo!., 2: 459. DREYER, W. A., 1932. The Effect of Hibernation and Seasonal Variation of Tem- perature on the Respiratory Exchange of Formica ulkei Emery. Phvsiol. ZooL, 5: 301. FINK, D. E., 1925. Metabolism During Embryonic and Metamorphic Development of Insects. Jour. Gen. Physiol, , 7: 527. 1 The references cited contain more or less complete bibliographies and hence detailed references are omitted. HIBERNATION AND DIAPAUSE 245 GLENN, P. A., 1922. Codling-Moth Investigations of the State Entomologist's Of- fice 1915, 1916, 1917. Bulletin, Natural History Survey, State of Illinois, vol. 14, pp. 217-289. RAU, P., AND N., 1918. Wasp Studies Afield. Princeton University Press. SHELFORD, V. E., 1927. An Experimental Investigation of the Relations of the Codling Moth to Weather and Climate. Bulletin, Natural History Sur- vey, State of Illinois, vol. 16, pp. 310-440. TAYLOR, I. R., AND H. B. STEINBACH, 1931. Respiratory Metabolism During Pu- pal Development of Galleria mellonella (Bee Moth). Physio! . Zool., 4: 604. UVAROV, B. P., 1931. Insects and Climate. Trans. Enloin. Soc. London, 79: 1. 16 THE RESPIRATORY GAS EXCHANGE IN TERMOPSIS NEVADENSIS S. F. COOK (From the Division of Physiology, University of California Medical School, Berkeley, California) The series of studies made in recent years by the members of the Department of Zoology at the University of California, as well as other investigations, have thrown much light on the relations between environ- mental factors and the habits of termites, particularly with respect to diet, moisture, and symbiotic microorganisms. Questions have arisen at times in connection with other ecological problems pertaining to the utilization of and the dependence on oxygen by the termite colony. In particular, it is conceivable that small differences in the oxygen tension of various habitats might influence the distribution of species and life cycle of individuals. In considering the matter from the ecological point of view, one is confronted by the difficulty that very little is known about the normal respiration of the termite itself. It was to obtain in- formation concerning this point that the present study was made. Several hundred specimens of Termopsis nevadensis, secured from the vicinity of Santa Cruz, California, were kept in open dishes, and fed a combination of moist filter paper and wood scraps. For experimental work they were sorted roughly into sizes, use being made of only the nymphs. Although such is probably not the case, there is a possibility that the respiration of the soldiers and reproductives might differ from that of the nymphs. Both oxygen consumption and carbon dioxide pro- duction were measured manometrically by the Warburg method. Usu- ally the gas exchange of about twelve termites at a time was measured in each manometer vessel. The results were sufficiently consistent to indicate that variations between individuals were thereby eliminated. In most cases the termites were weighed and the results expressed as cubic millimeters of oxygen absorbed or carbon dioxide evolved per gram per hour. Occasionally, however, this was not necessary, particu- larly when the results were purely comparative in nature. It is, of course, very difficult to keep the termites motionless and otherwise achieve basal conditions. This must be constantly borne in mind in determining respiratory quotients. Cleveland (1925) has re- ported a quotient of 1.0, but since he does not state the conditions under which the measurements were made, it is to be assumed that he was deal- 246 RESPIRATORY GAS EXCHANGE IN TERMITE 247 ing with normal, active animals, feeding on a preponderantly carbo- hydrate diet, animals which would he expected to show such a quotient. However, in investigating the respiration of invertebrates, both anaer- obic and aerobic, it is not necessary to achieve that particular basal state demanded in a mammal, provided the conditions are uniform throughout the entire series of experiments. With termites these conditions may be attained satisfactorily, as is shown by the following experiment. Experiment 1. — The respiration of ten groups of termites with vary- ing number of individuals in each was measured at 32° C. The oxygen consumption in cubic millimeters per gram per hour was : 930, 820, 850, 740, 740, 830, 760, 750, 770, 860. The extreme variation is about 20 per cent. The possibility should not be overlooked, in dealing with large num- bers of termites of somewhat different ages, that the respiratory rate may vary with age. The possibility is eliminated, however, as indicated thus : Experiment 2. — The oxygen consumption of two groups of termites was measured at 20° C., the first group consisting entirely of individuals of an average length of 10 millimeters, the second of 15 millimeters. The consumption of Group I was 433 cu. mm. CX/gram )( hour and that of Group II was 413 cu. mm. (X/gram )( hour, the difference lying within the experimental error. The general problem of the relation between oxygen tension and oxygen consumption has been the subject of numerous investigations. Without entering at this point into any detailed discussion of the results of this work, it may be stated that there has not yet been found any clear correlation between the two. Some animals are able to acquire and use their normal amount of oxygen at extremely low tensions ; oth- ers are very closely dependent upon the tension. The characteristic reaction of each species, must be determined experimentally in every case. (For reviews of the literature, reference may be made to papers by Helff, 1928, and Hyman, 1929.) Experiment 3. — To ascertain the general relation between oxygen tension and consumption a series of comparative measurements were made with several groups of animals at 20° C., using oxygen concentra- tions of 100, 21, 10, 5, 2, and 0.8 per cent. For the lowest tension a tank of commercial nitrogen served with oxygen present to the extent of 0.8 per cent as an impurity. The procedure in all cases was to estab- lish the normal rate in air and then to replace the air in the manometer vessel with the gas mixture to be investigated. Frequently two or more mixtures were introduced successively and when this was done the order was varied from group to group. Check readings were made at 248 S. F. COOK the end in air to insure that no injury has been done the respiratory mechanism through subjection to low or high oxygen tensions. No abnormalities were observed save in the case of 0.8 per cent. The rate in air after exposure to this tension was a little low but soon rose to the normal. The initial low rate was probably due to the immobility of the animals which is induced by very low tensions. To the immobility TABLE I Relative Oxygen Absorption of Termopsis at Varying Oxygen Tensions Group no. Oxygen concentration Oxygen absorption per cent per cent of normal I 100 111 II 100 93 III 100 86 IV 100 82 I 21 100 II 21 100 III 21 100 IV 21 100 V 21 100 VI 21 100 I 10 88 II 10 93 I 5 86 II 5 83 III 2 64 IV 2 73 V 0.8 26 VI 0.8 30 may also be partially ascribed the low respiratory rate during exposure to 0.8 per cent oxygen. Since all these rates are relative and are com- pared to the rate in air as normal, they may be expressed on a percentage basis, taking the rate in air as 100 per cent. The data are summarized in Table I. It will be noted that as the tension decreases, the relative oxygen consumption also decreases, but not proportionally.1 The respi- 1 The effect of pure oxygen appears to be slightly inhibitory. This phenom- enon has been observed in the case of other organisms but has never been satis- factorily explained. A suggestion which might be advanced is that the oxygen at high tension inhibits the respiration of the microorganisms of the gut. This fauna is killed by prolonged exposure to oxygen and may be sensitive to shorter ex- posures. RESPIRATORY GAS EXCHANGE IN TERMITE 249 ration remains very near the normal until tensions below 5 per cent are reached and even with less than 1 per cent oxygen the respiration is al- most one-third of its usual value. This situation indicates a very high degree of independence of the oxygen tension on the part of the termite. Furthermore, there appears to he a very well-developed capacity for utilizing extremely small quantities of oxygen. This capacity was made evident in the above experiment when the gas used was 0.8 per cent oxy- gen in nitrogen. For here the rate of respiration underwent a steady decrease, indicating that the oxygen tension was being materially re- duced by the termites themselves and that the rate decreased along with the tension. It seemed advisable to investigate more thoroughly this ability of the termites to utilize very small concentrations of oxygen. Therefore a series of studies was made, the data for which are consolidated in Table II. For the purpose of orientation a group of normal termites was first used with a mixture containing 0.7 per cent oxygen and the gas exchange followed for several hours (data not shown in Table II). The oxygen uptake was steady for a short time (about 30-40 minutes), then began to fall off. The decrease continued until the gas consump- tion ceased. But on continuing the readings it was observed that a positive pressure appeared, suggesting that now some other gas was being evolved. The rate of evolution became constant within an hour and remained so as long as measurements were continued — a matter of sev- eral hours. Since the vessel contained strong potassium hydroxide, this gas could not be carbon dioxide. The experiment was repeated with an inset of 10 per cent sulphuric acid as well as alkali, but the general course of the reaction was similar. The termites therefore evolve a gas which can be absorbed by neither acid nor alkali. No further at- tempt was made to determine the exact composition of this substance, but there is a strong possibility that it may be hydrogen or methane, or a mixture of both. If so, a reasonable assumption is that the micro- organisms in the gut are responsible. The principal constituent of the termite diet is cellulose and the breakdown of this material is usually ascribed to the protozoa and possibly bacteria which inhabit the digestive tract. In other animals which utilize cellulose in a similar manner, such as cattle, large amounts of hydrogen and methane are produced. There is therefore considerable likelihood that we are dealing with an analogous situation in the termite, although naturally such a statement cannot be made with certainty in the absence of a quantitative analysis of the gas produced. In order to investigate the role of the intestinal fauna in the produc- 250 S. F. COOK tion of this material — which we may call for lack of a more exact description the " undetermined " gas — a number of termites were de- TABLE II Oxygen Absorption by Termites at Very Low Tensions \ Group I II in IV v VI VII B. Previous treatment of termites .... Nor- Nor- Nor- Defaun- Defaun- Defaun- Defaun- mal mal mal ated ated ated ated C. Initial O2, per cent. . . . 0.7 0.7 0.7 0.7 0.7 2.08 2.08 D Weight in mg 513 1067 457 552 1126 386 699 E. Duration of experiment in hours 6 61 7| 6^ 6 10^ 6i F. Net gas exchange in cu. mm -0.5 + 52 -4.5 -83 -88 -333 — ?71 G. Rate of production of undetermined gas in cu. mm. per hour . . . 11.3 19 9 0 0 0 0 H. Total production of un- determined gas in cu mm. 79 130 77 5 0 o o o I. Rate of production of undetermined gas in cu. mm. per gram termite per hour. . . . 22.1 17.8 19.7 0 0 0 0 J. Total gas in manom- eter vessel in cu. mm. 13,150 12,650 13,200 13,110 12,450 15,250 12,850 K. Oxygen at beginning of experiment in vessel in cu. mm. J X C . . . 92 88 92 92 87 318 270 L. Total oxygen absorbed in cu. mm. F -- H . . 79.5 78 74.5 83 88 333 271 faunated. This was done by the method of Andrew (1930) which con- sists of the application of oxygen at several atmospheres pressure. When defaunated termites are placed in 0.7 per cent oxygen there is no indication whatever that any gas is evolved. The oxygen consumption RESPIRATORY GAS EXCHANGE IN TERMITE 251 sinks to zero and remains there indefinitely. It seems legitimate there- fore to ascribe the inert gas production to the intestinal fauna. The complicating factor of formation of the undetermined gas may thus be eliminated by defaunation, but the question is introduced whether defaunated and normal termites react in the same way to low oxygen tensions. To investigate this matter and simultaneously to secure data concerning the original problem of utilization of small amounts of oxygen the following experiment was performed. Experiment 4. — Seven groups of termites were placed in low con- centrations of oxygen. Groups I, II, and III were normal, the remain- der were defaunated. The first three groups and also Groups IV and V were placed in 0.7 per cent oxygen and Groups VI and VII in 2.08 per cent oxygen. The oxygen content of these mixtures was checked carefully by analysis with the Haldane gas apparatus. Then the gas exchange in each group was measured until, in the case of the first three groups, the rate of production of the undetermined gas was clearly estab- lished, and in the case of the others the oxygen consumption had ceased for at least two hours. At the end of this time the readings were dis- continued and the net f/as exchange of each group calculated. This, of course, is obtained from the initial and final manometer readings (see line F in Table II). With the defaunated groups the net exchange is equal to the total oxygen absorption, since there is no other gas con- cerned. With the normals the net exchange is equal to the oxygen ab- sorbed plus the undetermined gas evolved. To find the total oxygen absorption we must subtract from the net exchange the total undeter- mined gas evolution. This involves the fairly reasonable assumption that the production of undetermined gas has the same constant value while oxygen is being taken up that it is observed to have after oxygen uptake ceases. This assumption, though reasonable, still remains an assumption, for definite evidence cannot be secured until a method is devised for differentially absorbing the undetermined gas in the vessel as fast as it is formed. There appears to be no method at present for doing this. If we, then, subtract the total undetermined gas evolved from the net gas exchange, we get the probable total oxygen consump- tion of the normal groups (see line L in Table II). Finally, since the volume of the vessels is known and also the percentage composition of the gas initially introduced, the actual initial quantity of oxygen can be calculated (see line K in Table II). The total oxygen usage may then be compared with the total oxygen available. From the data presented in Table II the following conclusions may be drawn. In the defaunated groups (IV-VII inclusive) there is a fairly close correspondence between the oxygen absorbed and the 252 S. F. COOK amount initially present in the closed system. At least it may be stated that there is no significant quantity of oxygen remaining in the system when the oxygen consumption of the termites ceases. This appears to be true irrespective of the initial concentration (compare Groups IV and V with VI and VII). In the normal termites there seems to be a slight residue of oxygen, a concentration of the order of 0.1 per cent or a tension of less than one millimeter of mercury. But this difference between the normal and defaunated animals is too slight to be of sig- nificance, particularly since (1) the difference is of the order of the experimental error (as indicated by the deviations in Groups IV-VII) and (2) the assumption of a constant rate of inert gas evolution, ir- respective of oxygen tension, may not be precisely consistent with the facts. In general, however, it is possible to state that both normal and defaunated termites are able to utilize substantially all the oxygen in the immediate environment even though the latter reaches exceedingly low tensions. Under anaerobic conditions at least the production of undetermined gas is very constant (see line I in Table II) at a rate of about twenty cubic millimeters per gram termite per hour. A further study of this gas production would be of interest with respect to the composition of the gas and also its possible bearing on the problem of cellulose digestion. The results obtained with low oxygen tensions suggest further ques- tions : ( 1 ) What is the aerobic carbon dioxide production in normal and defaunated animals? This involves also the determination of the respiratory quotient of both types of animal. (2) Is carbon dioxide produced anaerobically, and if so, can the termites incur an oxygen debt? (Such has been found to be the case with the cockroach by Davis and Slater, 1926, 1928.) Experiment 6. — Eight groups of termites were investigated (see Table III). Groups I— IV were normal animals. Groups V— VII were defaunated one day previously and Group VIII was defaunated two weeks previously. In every case the gas exchange was determined in air in the presence of 5 per cent KOH (line C, Table III) and then in air without alkali (line D, Table III). When no gases other than oxy- gen and carbon dioxide are involved the result of the first determination represents the oxygen consumption, since the carbon dioxide is quanti- tatively absorbed by the KOH. The observed gas exchange in the second determination then represents the difference between the oxygen consumption and carbon dioxide production. Since the former is known, the latter may be calculated (line E, Table III). It was ascer- tained in the previous experiment (Experiment 5) that there is no pro- RESPIRATORY GAS EXCHANGE IN TERMITE 253 duction of the undetermined gas by defaunated termites and therefore in Groups V— VIII the observed exchange of carbon dioxide and oxygen may be taken as the corrected, or true, exchange (lines H and I, Table III). With normal termites it may be assumed, as previously, that TABLE III Carbon dioxide production and respiratory quotient of termites. Gas ex- change in all cases expressed as cubic millimeters per gram termites per hour. A. Group and condition I Nor- II Nor- III Nor- IV Nor- V De- VI De- VII De- VIII De- mal mal mal mal faun- ated 1 day faun- ated 1 day faun- ated 1 day faun- ated 2 wks. B. Weight in mg 731 633 681 866 852 648 977 694 C. Observed oxygen consump- tion in air (KOH) 186 218 173 239 190 239 154 200 D. Observed gas exchange in air (no KOH) 11 14 21.5 22.5 -3.5 2.0 4.0 -36 E. Observed carbon dioxide production (C— D) 192 232 194.5 261.5 186.5 241 158 164 F. Observed R.Q 1.06 1.065 1.12 1.095 0.98 1.005 1.025 0.82 G. Production of undetermined gas (average from Table II) 20 20 20 20 0 0 0 0 H. Corrected oxygen consump- tion in air (KOH) C-G . . . 206 238 193 259 190 239 154 200 I. Corrected carbon dioxide production in air (no KOH) C-D 197 232 194.5 261.5 186.5 241 158 164 J. Corrected R.Q. 0.955 0.975 1.005 1.01 0.98 1.005 1.025 0.82 K. Total gas production in ni- trogen 46.5 52 48.5 48 50.5 55.5 43 43.5 L. Carbon dioxide production in nitrogen (K— G) 26.5 32 28.5 28 50.5 55.5 43 43.5 under aerobic conditions there is a constant production of about twenty cubic millimeters of the undetermined gas per gram termite per hour; otherwise the respiratory quotient of the normal animals appears to be considerably over unity (line F, Table III). This factor may be al- 254 S. F. COOK lowed for and corrected values of the oxygen and carbon dioxide ex- change obtained (lines H and I in Table III). Applying this correction, the values of the respiratory quotient approach very closely to unity (line J, Table III). The freshly defaunated termites also show the same approximate value, and with them no assumption concerning the production of the undetermined gas is necessary. These results con- firm the statement of Cleveland that the usual R.Q. of termites is prac- tically 1.0. It is of interest to note, however, that termites which have been defaunated for some time and consequently probably are under- going starvation show a much lower R.Q. — in the one case here investi- gated 0.82. - This is to be expected from what we know in general of the effect of starvation on the R.Q. of other animals. To determine the anaerobic carbon dioxide production the same termites were subjected to as low an oxygen tension as could readily be obtained, 0.7 per cent. Since it was ascertained in Experiment 5 that the consumption of this oxygen ceases or is too small to have mate- rial effect in about three hours, the termites were allowed to remain for this period in the closed vessels and then the gas exchange was measured for one hour. The readings showed a low but distinct positive varia- tion, indicating the production of gas (line K, Table III). With nor- mal animals part of the output (20 cu. mm./gm. X hr.) was due to production of the undetermined gas and therefore this amount was sub- tracted from the observed gas production to give the carbon dioxide value (line L, Table III). This value is considerably below that ob- tained under aerobic conditions but is evidence that some metabolic changes are still proceeding in the animal's tissues. To summarize this experiment and answer question 1, it may be said that both normal and freshly defaunated termites produce carbon dioxide under aerobic and anaerobic conditions; that the R.Q. of both types of animal is very close to unity; and that the anaerobic carbon dioxide production while present is much smaller in both types than the aerobic production. Experiment 7 . — Three groups containing equal numbers of termites at 20° C. were placed in air and their respiration measured with and without the presence of KOH. Group I was then exposed to 0.7 per cent oxygen in nitrogen in the presence of KOH for an hour and a half, or until the oxygen was nearly exhausted, at which time air was reintro- duced. The oxygen consumption began again and after a brief interval at a lower rate proceeded indefinitely at the same rate as at the beginning. - All the termites were normally mobile and active, even those which had been starved. All were therefore at the same " basal " level. RESPIRATORY GAS EXCHANGE IN TERMITE 255 The absence of any increase in the rate above the normal precludes the possibility of oxygen debt. Groups II and III were similarly treated except that KOH was not present. Air was readmitted at the end of three hours of anaerobiosis. The respiration rates (both carbon dioxide and oxygen) returned very quickly to their normal values and remained there for several hours, at the end of which time the experiment was discontinued. Experiment 8. — A repetition of Experiment 7 in which the results of the latter were confirmed. It seems clear that the termite (Termopsis at any rate) possesses a mechanism for the continued production of carbon dioxide even in TABLE IV Tlic Effect of Prolonged Exposure to Nitrogen Termite group Normal oxygen consumption Hours exposure to nitrogen Oxygen consumption after exposure Condition after recovery per cent per cent of normal I 100 1 89 Immediate recovery from immo- bility II 100 3 99 Same III 100 6 86 Recovery in 15 minutes IV 100 9 95 Recovery in 20 to 30 minutes V 100 24 99 Recovery in 1 hour VI 100 48 96 Recovery in 12 hours VII 100 96 42 Never recovered VIII 100 144 16 Never recovered the complete absence of oxygen. That this mechanism is not identical with that which presumably obtains in mammalian muscle and else- where is demonstrated by the complete lack of any indication, from gas measurements at least, of oxygen debt. That the anaerobic system is rather inefficient is shown by the fact that the carbon dioxide production under such conditions is less than one-half the normal value and further- more by the fact that the termites pass into a state of complete immobil- ity even though they continue to respire. This state resembles acute anoxemia in mammals in that it appears very soon after a sudden ex- clusion of oxygen and disappears very quickly after readmission of oxygen, a matter of minutes in both cases. It is worth while to determine how long Termopsis can endure com- plete anoxemia and still retain its capacity for normal aerobic respira- tion and normal activity. 256 S. F. COOK Experiment 9. — Eight groups of termites, after their normal oxygen consumption was measured, were placed in a desiccator. The latter was filled with nitrogen from which the oxygen had been removed by treatment with strongly alkaline pyrogallol. At intervals the termites were removed and their respiration measured as well as their general behavior observed. The respiratory rate was measured after the ter- mites had recovered their mobility, except in Groups VII and VIII which never recovered. The lack of mobility may therefore in part account for the low rate of oxygen consumption noted in these two groups. The effect of prolonged exposure to pure nitrogen is sum- marized in Table IV. It will be observed that the first noteworthy reduction in respiration occurs in the group which had been without oxygen for four days and that this group never recovered their mobility. These termites there- fore are not truly anaerobic in the same sense as, for example, yeast, which is able to survive and grow indefinitely in the absence of oxygen. Nevertheless the survival without apparent harm for two days is in itself a striking and significant phenomenon. In addition to the ability to withstand oxygen lack, Tcrmopsis shows high tolerance to the presence of carbon dioxide. The respiration of four groups of termites was measured in 20 per cent oxygen plus 5, 10, 20, and 40 per cent carbon dioxide respectively. The net gas ex- change of those in the 5, 10, and 20 per cent carbon dioxide, with no alkali present in the vessels, was approximately the same as that of termites in air and the appearance of the termites otherwise was per- fectly normal. It is unlikely therefore that any considerable upset oc- curred in their respiration. The same considerations concerning respi- ration apply to the group in 40 per cent carbon dioxide, but these ter- mites soon became immobile. The general conclusion may be drawn from all these experiments that Tcrmopsis ncvadcnsis possesses powers for meeting adverse en- vironmental conditions far in excess of its probable needs. The natural habitat is relatively well aerated wood in which the gas tensions are probably not far from atmospheric. Nevertheless, these animals are capable of extracting practically the last traces of oxygen in a closed space and then of persisting several hours, if not days, in the absence of oxygen. Furthermore, they can endure quantities of carbon dioxide which would seldom if ever be present naturally in their environment. These conditions might occur in wet soil or in other habitats frequented by some of the other species and genera of termites. If this is so, it is possible that ability to withstand such conditions is a general char- acteristic of all the members of the group which persists in certain RESPIRATORY GAS EXCHANGE IN TERMITE 257 species even though the actual need is seldom encountered. A com- parative study of the different species of termites with respect to their respiratory power and their environment is very desirable. SUMMARY The oxygen consumption of Tcrnwpsis nevadensis does not decrease materially with falling oxygen tension until a concentration of approxi- mately two per cent is reached. Below this tension the affinity of the animals for oxygen is so marked that substantially all the available gas is consumed. In the absence of oxygen the organism is able to respire anaerobi- cally, although at a reduced rate, for as long as two days without injury. During this time the animals are in a state of immobility from which they recover soon after readmission of air. After exposure to anaer- obic conditions no indication of oxygen debt was found. These termites are able to exist and respire normally in carbon di- oxide as high as 20 per cent. Higher concentrations tend to induce a condition of anaesthesia, which, however, is reversible. Under anaerobic conditions, and possibly also in the presence of oxygen, the termites evolve an undetermined gas which may be hydrogen or methane. The production of this gas depends on the integrity of the intestinal fauna, since it is not evolved by defaunated termites. The respiratory quotient of both normal and freshly defaunated termites is approximately unity, but in starved defaunated termites it falls considerably. LITERATURE CITED ANDREW, B. J., 1930. Method and Rate of Protozoan Refaunation in the Termite Termopsis angusticollis Hagen. Univ. Calif. PubL ZooL, 33: 449. CLEVELAND, L. R., 1925. The Ability of Termites to Live Perhaps Indefinitely on a Diet of Pure Cellulose. Biol. Bull, 48: 289. DAVIS, J. G., AND W. K. SLATER, 1926. The Aerobic and Anaerobic Metabolism of the Common Cockroach (Periplaneta orientalis). Part I. Bioclicm. Jour., 20: 1167. DAVIS, J. G., AND W. K. SLATER, 1928. The Aerobic and Anaerobic Metabolism of the Common Cockroach (Periplaneta orientalis). Part III. Biochcm. Jour.. 22: 331. HELFF, O. M., 1929. The Respiratory Regulation of the Crayfish, Cambarus im- munis (Hasten). Ph\siol. ZooL, 1: 76. HVMAN, L. H., 1929. The Effect of Oxygen Tension on Oxygen Consumption in Planaria and Some Echinoderms. PhysioL ZooL, 2: 505. FURTHER STUDIES OF THE AGGREGATING BEHAVIOR OF AMEIURUS MELAS * EDITH S. BOWEN WHITMAN LABORATORY OF EXPERIMENTAL ZOOLOGY, THE UNIVERSITY OF CHICAGO The investigations here reported are continuations of work on the role of the sense organs in aggregations of young catfishes (Bowen, 1931), and were directed toward securing further evidence concerning (a) the analysis of the factor concerned in the touch reaction; (b) the retention of touch and visual aggregating reflexes during prolonged isolation; and (c) the influence of aggregation and of isolation upon the physiological processes indicated by the rate of oxygen consump- tion. Taken as a whole, these studies together with those previously reported contribute to our understanding of the mechanisms concerned in the schooling behavior of Amciurus melas in particular and to some extent to that of fishes in general. SENSE ORGANS INVOLVED IN THE TOUCH REACTION The reactions of aggregating young catfishes of the species Ame'mrus melas have already been studied to determine which sense organs are responsible for receiving the stimuli that result in the formation of the aggregations (Bowen, 1931). Vision was found to be essential for the reactions concerned since blinded fishes did not aggregate and normal fishes failed to aggregate in the dark. There were indications, however, that responses to touch are also of great importance, and that the resulting stimulus may perhaps be the fundamental cause of the aggre- gations and that vision serves only as a means by which the fishes find one another. Individuals of this species are strongly thigmotactic ; they always rest with as many points of the body as possible in contact, and in the aggregations they push against each other continually when disturbed. The components of this contact stimulus were not deter- mined, but the early work indicated that it might contain both tactile and gustatory elements. Morphologically this is possible since (Her- rick, 1902) the skin of catfishes over the whole body contains not only end organs which are sensitive to pressure but also terminal buds, 1 The present investigation was aided by a grant to the University of Chicago from the Rockefeller Foundation administered by Dr. W. C. Alice, to whom I wish to express sincere thanks for interest and helpful criticism. 258 AGGREGATING BEHAVIOR OF AME1URUS 259 corresponding to taste buds in the mouth. These are most abundant on the barblets and diminish in number toward the tail. They are J sensitive to chemical stimuli through contact as contrasted with the olfactory organs, which are distance receptors (Parker, 1910). Observations of the behavior of catfishes in the tanks was also suggestive of a gustatory response. One cattish would often approach another, drag its barblets over it and push against it several times. This occurred to some extent when the second catfish was entirely inactive or when an active individual came in contact with a recently pithed fish. The same positive reaction was also given to goldfishes and mudmin- nows which were in the same aquarium with the catfishes. The re- sponse was negative to a weighted model, made of paraffin mixed with India ink to simulate a small catfish in color and form. The approach was the same as to another catfish, but the barblets were drawn over the model only once and then the fish turned away and swam off. It was considered desirable to analyze further the contact stimulus which one catfish receives from another, and to determine the relative importance of pressure and of chemical stimulation. An experiment was designed to determine whether there is a differ- ence in the reactions of a catfish to a scoured stone which presumably gives no chemical stimulus and to another catfish rendered incapable of movement and reciprocal pressure. Fishes were pithed and the wound was covered with melted paraffin to prevent the diffusion of body fluids and a resulting gustatory stimulation. Each of these freshly killed fishes was placed at the side of a crystallizing dish 24 cm. in diameter, and held down by a small pebble since the lifeless bodies floated to the top if unattached. A small stone about the size of a catfish was placed at the opposite side of each dish after which a normal individual about 3 to 5 cm. long was added. Observations of the positions of the cat- fishes were made at 5-minute intervals, or in a few cases after a longer period ; after each observation the normal fish was disturbed so that the position in the next observation was determined anew. After five series of records had been made the pithed fishes and the stones were interchanged so that a positive reaction to one place due to light or some other factor would not mask the results. In most cases the active fishes had come to rest at the end of 5 minutes, although in a few instances one or more were still active. If a fish was moving within a small area the location was noted ; otherwise these cases were not included in the results. The results are given in Table I (a). It is readily seen that the number of positive reactions to the pithed fish and to the stone are about the same, 24 and 29 respectively, whereas the total number of indifferent 260 EDITH S. BOWEN reactions where the fish was resting apart from either one totals 94. This indicates that there is no difference between the reactions to a stone and to a catfish if the latter is rendered inactive. Immediately upon the conclusion of the first series a control was run by removing the pithed fishes from each dish and substituting a normal fish. These results are shown in Table !(/>)• I*1 62 per cent of the cases the fishes were resting together as contrasted with 16.3 per cent in Table I (a). Since in these cases a response was recorded for each fish, this is equiva- lent to 31 per cent positive contacts per fish. The increase over the incidence (16.3 per cent) of positive contacts for fishes isolated with one pithed fish is due to stimulation of a visual nature from the moving partner as well as to reciprocal pushing after the two have met. Even with two stones present in the dish with two fishes (Table I, c} , the TABLE I Reactions of catfishcs to stones and to other catfishes as shoivn by resting positions Condition of experiment No. of fishes Positive reactions Indif- ferent Total Fish Stone a. Normal fish with stone and pithed fish 5 5 5 10 4 10 10 6 13 27 40 27 47 50 50 Total 15 24 29 94 147 b. Two normal fishes with stone 10 62 18 20 100 c. Two normal fishes with two stones 20 98 48 54 200 fishes aggregated about as much as in the preceding arrangement where the single stone allowed less possibility of purely indiscriminate contact. In about half of these observations the two fishes were found resting together, an equivalent of 24.5 per cent positive contacts per fish, and a decided increase over the 16.3 per cent in (a) where an inert partner was present. Since it had been found that catfishes responded negatively to a black paraffin model, it was possible that the small amount of paraffin upon the pithed fish might be preventing a positive reaction to the pithed fish itself. Accordingly the responses to a paraffined and to a non- paraffined stone were compared. An equal number of positive reac- tions, 24, was given to each object with 88 indifferent responses, so AGGREGATING BEHAVIOR OF AMEIURUS 261 that it may be concluded that there is no definite reaction to the paraffin itself. From the above results there is no evidence for a gustatory element in the stimulus which one catfish receives from another. When move- ment and pressure are eliminated the positive response is given in as many cases to a stone as to a catfish. The gustatory stimulus enters only to produce a negative reaction when some unfavorable chemical is present. Observations upon normal fishes in a dish with untreated paraffin models and similar models mixed with India ink to add color showed that a decidedly negative reaction was given only to the India ink models. To untreated paraffin models the reactions were the same as to an inactive fish. It was impossible to eliminate the sight reflex by using blinded fishes in these experiments, because a blinded fish remains active for long periods without coming to rest near any object; such fishes rest only after long activity and in the absence of recent stimula- tion. EFFECT OF ISOLATION UPON AGGREGATING BEHAVIOR The question as to whether the reactions concerned in catfish aggre- gations are entirely instinctive and automatic or are in part due to con- ditioning or at least susceptible to modification by conditioning is of interest in this simple type of social behavior found low in the vertebrate scale. Parr (1927) explained the schooling of pelagic fishes by an assumption of a simple automatic eye reflex which acts in the case of milling to produce a type of behavior with no apparent purpose. Learn- ing among fishes has, however, repeatedly been demonstrated. Thus individuals have been trained to associate a stimulus such as light, color (Hineline, 1927), or direction (Churchill, 1916), with food, or in Gold- smith's work (1914) with the nest; and Triplett, 1901, by separating a perch from minnows by a glass partition, conditioned the perch so that after the removal of the partition it made no attempt to reach the minnows by passing the line where the partition had been. Recently Bull (1929) has shown that Blcnnins pJwIis can be trained to use the senses of taste and smell in the capture of food, although normally they play no part in this behavior. He has demonstrated that purposive movement in fishes can be explained by a combination of unconditioned and conditioned responses in nature as well as in the laboratory. Two types of reactions of catfishes to one another had previously been observed (Bowen, 1931). Normally vision is the means by which individuals are enabled to come in contact with one another. In addition blinded fishes, previously grouped, were shown to respond to each other when the skin was intact by turning out of their course to- ward a passing fish in approximately 50 per cent of the cases where the 17 262 EDITH S. BOWEN two fishes came within two inches of each other. This reaction is apparently due to vibrations set up by the tail of the passing fish and received by the skin, and which act possibly more or less as a thigmo- tactic stimulus. This positive response may result only after fishes are blinded, but it serves to bring the fishes momentarily nearer one another, and may therefore be considered as a social response. If the satisfaction of a thigmotactic response is the fundamental basis of aggregating behavior in catfishes, then it seems possible that both the above types of response might be modified if, through isolation, thigmotactic responses from other fishes were prevented for a time. The first indication that such might be the case was found in work with gradient tanks (Bowen, 1931) in which blinded catfishes, isolated from the group for a week, showed a tendency to stay in the opposite end of the tank from the group of fishes when the latter were separated from the single individual by a wide-meshed wire partition. The observed difference in behavior was statistically significant ; at the same time blinded fishes which had been members of a group tended to stay in the end toward the group, although the difference from the controls was not significant with blinded fishes. To determine whether the response of blinded fishes could be con- ditioned by isolation, individuals were placed in two-quart jars contain- ing plants and fed regularly for a period of 20 days. They were then placed in groups of five in a porcelain tub and the reactions recorded for 15 minutes. At the conclusion of these observations they were kept grouped for one or more days and tested again as a control to determine whether they gave the normal number (50 per cent) of positive reactions after grouping. In almost all cases a positive reac- tion was given when two fishes actually touched. The few exceptions occurred when they met with some force and darted back in a negative response. In cases where the fishes passed within two inches of each other without touching, however, a difference between previously iso- lated and grouped fishes was evident. While the latter gave a total of 27 positive responses, 1 negative and 26 indifferent, showing 50 per cent positive reactions, isolated individuals gave 14 positive, no nega- tive, and 38 indifferent responses with only 27 per cent positive. The difference is, however, actually even greater between the two groups; during the first 9 minutes after grouping of previously isolated fishes there was only one positive response and 30 indifferent ones. During this period there were several positive reactions to contact, which ap- parently served to overcome the effects of isolation, so that the positive reaction was reestablished at the end of that time. In fact the positive responses for the next 6 minutes rose to 62 per cent of the total. AGGREGATING BEHAVIOR OF AMEIURUS 263 From this series it seems evident that the positive responses of blinded fishes to one another are eliminated by isolation, but are re- established again in a few minutes when the fishes are in the same container and contact may occur. The ease with which the response of blinded fishes to vibrations caused by another fish is broken down and the fact that several minutes are required for its reestablishment make it seem possible that this is a true conditioned response developed after the fishes are blinded, and in connection with the thigmotactic stimulus. It does not act with enough efficiency to enable the formation of aggregations, as one fish can determine the direction of another fish for only one or two turns and then loses it. In normal individuals vision directs the movements of the fishes toward one another at a greater distance than this stimulus can be felt so that it probably acts very slightly if at all in aggregations of normal fishes. The effect of isolation upon the aggregating behavior of normal fishes was also tested by a similar series of experiments. Several at- tempts were made to isolate very young catfishes to determine how readily the normal response might be altered before the characteristic social reactions had become established by long use. But it was im- possible to keep the very young fishes alive in the laboratory. They died in large numbers under the best conditions and none survived a spell of hot weather. Therefore most of these results were obtained upon older fishes. The method employed in recording the reactions of the fishes to one another was similar to that for the blinded fishes except that a distance of about a foot and a half was considered the maximum within which a positive reaction was possible, since normal fishes nearly always react to one another within this distance. In all cases the two or three fishes placed in the experimental tub at one time were aggregating closely in the course of a few minutes, so that reactions were recorded only until this occurred. These were of two types, touch reactions where two fishes came in contact without any observable turning from their course when within visual distance, and sight responses where one of the fishes turned definitely toward the other. When either resulted in typical aggregating behavior it was considered positive ; if not, indiffer- ent. Usually a touch stimulus resulted in a positive response but in a few cases, almost always with previously isolated fishes, the first touch response was indifferent, but was followed shortly by a positive response often quite clearly as a result of a sight stimulus. The results of all the experiments are given in Table II. The effect of isolation upon reactions responsible for aggregations seems to be shown most clearly by the number of indifferent reactions as compared 264 EDITH S. BOWEN with the positive sight reactions before aggregation is established. In each case where there were indications of a breakdown of the aggregat- ing behavior, the fishes were kept together for a clay and the experi- ment repeated as a control. Such procedure was necessary for com- parison since normal fishes which have been isolated for a long period of time require several days before they become as active or responsive as those from a group, a condition which might affect the proportion of positive reactions. The isolation of very young fishes for five days (Experiment 1, Table II) gave an indication of a breakdown of the aggregating re- TABLE II Summary of reactions of catfishcs to one another after isolation Indifferent reactions Positive Exp. Number of fish Number tested together Length Days isolated before positive response sight reactions before Sight Contact Total contact 1. 6 2 1 in. 5 6 1 7 1 K«).* 6 2 1 in. 0 3 0 3 0 2. 8 2 3-5 in. 38 0 1 1 3 3. 8 2 1.5 in. 52 0 0 0 6 4. 3 3 2 in. 161 4 0 4 0 4(«).* 3 3 2 in. 0 0 1 1 1 (from Exp. 3) 5. 12 3 2 in. 161 5 5 10 2 5(fl).* 12 3 2 in. 0 1 1 2 5 * In (a) of each experiment are the results when the fish were retested after one or more days together. spouses. In Experiments 2 and 3, where young fishes about six months of age were isolated for between one and two months, the behavior was not modified and was entirely typical of fishes from the group. In Experiments 4 and 5, however, where the individuals were kept isolated for over five months, there was decided evidence for an isolation effect. Under these conditions the fishes isolated for five months gave in all 9 indifferent sight responses and 5 indifferent touch responses as compared with one of each after they had been together for 24 hours. Only 2 positive sight responses were noted before contact occurred as contrasted with 6 in the control. As the sight stimulus still acted in two AGGREGATING BEHAVIOR OF AMEIURUS 265 cases without a touch stimulus, the former was not completely elimi- nated, but there seemed to be a decided tendency in that direction. Whether this sight reflex is instinctive or whether it is established soon after hatching as a conditioned response depending upon the posi- tive thigmotactic reaction is still a question. If instinctive, however, it is apparently subject to change by the conditioning processes involved in isolation if these act for a long enough period of time. Probably also this effect can be produced more readily the younger the fishes. Un- fortunately observations on the initial reactions of newly hatched fishes could not be made, hence the obvious and crucial test for the relative amounts of instinctive and conditioned elements in this behavior com- plex is lacking. However, the evidence indicates that conditioning plays a part in this aggregating behavior. From a survey of the results obtained to date in the investigation of the reactions responsible for the aggregations of young catfishes, we may conclude that the two senses concerned are sight and contact with taste playing no part in spite of the acuteness of this sense in these ani- mals. Sight is the sense by means of which the fishes normally aggre- gate and the visual response of catfishes may be an instinctive reaction to one another. Touch, however, probably has the more basic role, since the sight response can be diminished somewhat in effectiveness by isolation and is reestablished by contact. The mutual pressure of the fishes due to pushing seems to give the stimulus which maintains these aggregations. RESPIRATORY BEHAVIOR OF GROUPED AND ISOLATED CATFISHES It has been shown that grouping has an effect upon the respiratory rate of animals. Thus Alice (1926 and 1927) has found that in land isopods and the brittle starfish, OpJiiodcnna brcvispina, at least when out of the breeding season, the rate of respiration, as measured by the amount of oxygen consumed, is lower per animal for grouped individuals than for isolated ones for the first few hours. The rate of respiration decreases more rapidly for the isolated animals, however, so that after several hours these are respiring more slowly than the grouped ones. There is evidence that, in the case of the starfishes at least, the oppor- tunity for physical contact afforded by the other individuals present is responsible for the group effects upon respiration and for the longer survival of the individuals composing it. Schuett (1931) has investi- gated this phase of respiratory behavior among goldfishes, guppies, mudminnows and Fundulus hetcroditus, and has obtained results similar 266 EDITH S. BOWEN to those of Allee for land isopods and brittle starfishes. In his shorter experiments, which covered usually one to five hours, the group con- sumed less oxygen per fish than did the isolated individuals. Working independently, I have heen able to confirm Schuett's results with gold- fishes, the only one of his species tested. It seemed desirable to determine whether any difference exists be- tween the respiratory rate of isolated and grouped individuals of aggre- gating young catfishes, especially in view of the part that contact plays in such reactions. Accordingly experiments were run with both normal catfishes and with catfishes which had been blinded several days previ- ously in order to eliminate the important effect of vision upon aggre- gating behavior. The method followed the technique of Schuett. Each experiment was performed in parallel with normal and blinded fishes. In each case, four single fishes were placed in individual Erlenmeyer flasks, holding one liter, and four were placed together in a similar flask. The flasks were then arranged parallel to two windows to equalize the normal effect of light upon the behavior of catfishes. The flasks had previously been filled with well water which in most cases had been allowed to stand overnight to come approximately to room temperature and to saturation with air, although in the first experiments air was bubbled through the water. In all of the experiments the oxygen tension of the initial sam- ples varied only between 4.5 and 6.5 cc. per liter of water; in this range the rate of oxygen consumption is independent of the oxygen tension (Schuett and citations). After the introduction of the fish, a layer of heavy mineral oil was poured into the neck of the flasks to prevent gase- ous exchange at the surface. By means of glass siphoning tubes kept in the flasks, samples of about 15 cc., known to within 0.05 cc., were withdrawn for analysis for oxygen by the method of Winkler. One sample was taken immediately after the introduction of the animal and a second one one hour later. A control flask without fish was sampled similarly. The external end of the siphon terminated in a piece of rub- ber tubing closed by a clamp. In the first half of the experiments the siphons reached to the middle of the flasks. In the last ten experiments longer siphons were employed and the water was stirred with the siphon one minute before sampling. No difference was noted in the results with the later modification of the technique. The fishes to be used in these experiments were kept on the experimental table in glass-walled aquaria to avoid excessive stimulation while being transferred to the flasks. Table III shows the results of the experiments. The total oxygen AGGREGATING BEHAVIOR OF AMEIURUS 267 consumed by four isolated fishes in each experiment is compared with that consumed by the group. The difference which Schuett and I have found between grouped and isolated non-schooling fishes is not ob- tained here. In fact, with the normal catfishes the members of groups consumed on the whole more oxygen than did isolated individuals but TABLE III Total oxygen consumption for one hour of four grouped and four isolated catfishes, blinded and normal (expressed in cc. oxygen per liter) Normal Blinded Isolated Group Isolated Group 1 .56 .20 2 .79 .81 .40 .57 3 .41 .34 4 1.16 1.03 .80 .63 5 .47 .67 .95 .49 6 .64 .64 .42 .49 7 .62 .66 .67 .51 8 .65 .68 .68 .68 9 .26 .42 .73 1.03 10 1.02 .98 1.12 1.27 11 .80 .74 1.00 1.01 12 .72 1.11 1.29 1.02 13 .40 .70 .56 .72 14 .88 .48 .75 .50 15 .31 .54 .48 .42 16 .88 .55 .93 .51 17 .34 .61 .23 .61 18 .77 .55 .59 .39 19 .39 .50 .54 .36 20 .87 .88 .72 .88 21 .63 .63 22 .46 .57 .59 .53 23 .48 .45 .69 .86 24 .20 .45 .55 .53 25 .18 .41 .46 .42 Mean difference 0.05087 cc. Mean difference 0.05 cc. more oxygen for groups. more oxygen for isolated Probability 0.2158. fishes. Probability 0.2898. the difference is not statistically significant. In the case of the blinded fishes the isolated individuals consumed very little more oxygen, but here again the difference is not important. The results of three typical experiments are given in Table IV. The cause of the discrepancy from Schuett's and my results with goldfishes is not difficult to find, nor do these indifferent results offer 268 EDITH S. BOWEN any contradiction to his. With normal catfishes the isolated individuals remain quiet during practically the whole experiment. In the groups, on the other hand, the fishes are in motion most of the time, being stim- ulated by the presence of the other individuals. Thus the rate of respi- ration of the grouped fishes is increased and the total is higher than for the isolated fishes. The fact that the difference is no greater probably indicates that if the factor of movement could be eliminated with cat- TABLE IV Results of three experiments, shoeing oxygen consumption of isolated and grouped fishes, blinded and normal Normal Blinded Experiment Isolated Group Isolated Group 18 .23 .19 .09 .16 .18 .07 .27 .17 .77 .55 .59 .39 .04 19 .10 .07 .12 .13 .14 .15 .03 .19 .39 .50 .54 .36 -.04 20 .14 .28 .13 .03 .34 .18 .26 .23 .87 .88 .72 .88 -.06 fishes a lower rate of respiration for the groups would be found than for the singles, in agreement with Schuett's work on goldfishes, where problems connected with relative motion were apparently not involved. With the blinded fishes both isolated and grouped individuals are in mo- tion all the time, as are goldfishes, but more actively so. Here the dif- ference is in the same direction as in the case of goldfishes, but is less and not as consistently positive. This is not surprising with the amount of motion occurring with these blinded fishes. It was hoped that contact responses which gave as much satisfaction as they apparently do in catfishes might produce a difference in the respir- AGGREGATING BEHAVIOR OF AMEIURUS 269 atory rate, and that this might serve as a method of determining more accurately the role of the different sense organs in the social behavior of these fishes. So far this has not been possible, but the results do indi- cate an interesting difference between the respiratory rate of schooling and non-schooling fishes ; at least in these schooling fishes, such as Ainci- unis inclas, individuals stimulate one another to activity by sight or touch and thus offset or counteract the lower respiratory rate found among groups of non-schooling fishes, so that there is no significant difference in oxygen consumption between grouped and isolated individuals. With blinded fishes it may be that contact stimulation is effective in raising slightly the respiratory rate of the group so that here a significantly lower rate was not obtained. SUMMARY Catfishes showed no discrimination by contact between a scoured stone and an inactive catfish, nor between a paraffined and non-paraffined stone, but gave a negative reaction to models of paraffin mixed with India ink. There is no evidence for a gustatory element in the stimulus which one catfish receives from another. A gustatory stimulus acts only to produce a negative reaction when some unfavorable chemical is present. The positive responses which blinded catfishes give to one another in passing are eliminated by isolation for 20 days, but reestablished in a few minutes when the fishes are placed together. Touch responses by blinded catfishes are positive immediately after isolation when contact is gentle enough to prevent shock. The reactions of blinded fishes to one another due to a response to water vibrations, may be a conditioned response developed after the loss of eyesight, and is probably not effective among normal fishes. The sight response of normal fishes to one another was not com- pletely eliminated in all individuals by 161 days of isolation, although it was much less marked. It was reestablished in the course of a few minutes, usually soon after contact occurred. This sight response may be instinctive but is probably subject to modification by conditioning to some extent at least. A satisfaction evidently accrues to the catfishes from the mutual contact and pressure of the aggregations, and the importance of the thigmotactic response in these reactions is emphasized by these observa- tions. A comparison of the respiratory rate of catfishes, both normal and blinded, for a period of one hour gave no significant difference between the grouped and isolated individuals. 270 EDITH S. BOWEN BIBLIOGRAPHY ALLEE, W. C., 1926. Studies in Animal Aggregations : Causes and Effects of Bunching in Land Isopods. Jour. E.vpcr. Zoo!.. 45: 255. ALLEE, W. C., 1927. Studies in Animal Aggregations : Some Physiological Effects of Aggregation in the Brittle Starfish, Ophioderma brevispina. Jour. E.v- pcr. Zool., 48: 475. BOWEN, E. S., 1931. The Role of the Sense Organs in Aggregations of Ameiurus melas. Ecol. Mono., 1:1. BULL, H. O., 1929. On the Nature of Purposive Movements in Fishes. Kept. Doi'c Marine Laboratory, CnUercoats, Northumberland, 1929: 39-46. CHURCHILL, E. P., JR., 1916. The Learning of a Maze by a Goldfish. Jour. An. Behavior, 6: 247. GOLDSMITH, M., 1914. Les reactions physiologiques et psychiques des poissons. Bull, dc I'lnstitut general psychologique, Paris, 14: 97. HERRICK, C. J., 1903. The Organs and Sense of Taste in Fishes. Bull. U. S. Fish Comm., (1902) 22: 237. HINELINE (WHITE), G. M., 1927. Color Vision in the Mudminnow. Jour. Ex- pcr.Zool.,47: 85. PARKER, G. H., 1910. Olfactory Reactions in Fishes. Jour. E.vpcr. Zool., 8: 535. PARR, A. E., 1927. A Contribution to the Theoretical Analysis of the Schooling Behavior of Fishes. Occas. Pap. Bingham Occaiu/r. Coll., 1927, 32 pp. SCHUETT, J. F., 1931. Studies in Mass Physiology: the Effect of Numbers upon Oxygen Consumption of Fishes. Doctor's thesis, University of Chicago Library. TRIPLETT, N., 1901. Educability of the Perch. Am. Jour. Psych., 12: 354. MECHANISM OF MOVEMENT OF EPIDERMIS, ESPE- CIALLY ITS MELANOPHORES. IN WOUND HEAL- ING, AND BEHAVIOR OF SKIN GRAFTS IN FROG TADPOLES EARL H. HERRICK ZOOLOGICAL LABORATORY, HARVARD UNIVERSITY Probably no other animal tissue is more active than an epidermis in repairing lost or damaged parts. The process is not primarily one of growth. A deficiency, either large or small, in the epidermis is rapidly covered by centripetal movement of the surrounding epithelium, and later, by reorganization and growth, the original condition is re- stored. Rand (1915) in considering the wound reactions of actinians says (p. 207) : " In general, an epithelium will not tolerate a free edge. When such an edge arises, accidentally or otherwise, the epithelium ex- tends until, if possible, the free edge meets and unites with some other portion of the same layer or with another epithelium. The essential function of an epithelium is to cover a surface continuously." The investigations of Fraisse (1885), Barfurth (1891), Born (1896), Morgan (1901), Loeb and Strong (1904), Rand (1905), Eycleshymer (1907), Matsumoto (1918), Loeb (1920), Arey (1925), and Collins and Adolph (1926) point unanimously to the conclusion that wounds are at first covered by movement of the surrounding epithelial cells and that proliferation occurs later to restore the original thickness of the layer. The cause and mechanism of the cell movement are not so well agreed upon. Born (1896) believed that the cells flatten to cover a larger surface than previously. Rand (1905), however, referring to wound healing in earthworms, says (p. 46) : " There certainly is little evidence in favor of supposing that the concentric advance of the epidermis is due to the tendency of the individual cell to spread itself over the greatest possible surface." Barfurth (1891) proposed that, while the movement of the epidermis might be in part a passive " Verschiebung " due to relief from lateral pressure in the layer, it was in the main an active movement of cells which had become "embryonal beweglich (amoboid) ' (p. 417). Oppel (1912) described epithelial movement as an active movement— often a " Massenbewegung '; —resulting from change of form of the 271 272 EARL H. HERRICK epithelial cells which, however, are not ameboid. On the contrary, Holmes (1914), observing amphibian epidermis in tissue culture, de- cided that epithelial movement is not a " Massenbewegung " but is the result of essentially ameboid movements of individual cells. And, again, Collins and Adolph (1926) concluded that wound closure is accomplished by a mass movement or " pushing in " process of the epithelial layer and " not by the independent migration of cells " (p. 491). Loeb and Strong (1904, p. 282) considered it probable that " tension, either previously existing or called into play by the wound, is the cause "of the closure of the epidermal wounds. Morgan (1901, p. 70), describing wounds in very young tadpoles, stated that " the wound is covered not by individual cells wandering over the exposed surface, but by a steady advance of the smooth edge of the ectoderm toward a central point. ... As there are no muscle fibres present . . . , the result cannot be due to muscular contrac- tion. . . ." The artificial cultivation of epithelial tissue has given striking evi- dence that the cells are motile and their movement ameboid. This was suggested at an early date by Peters (1889). who studied wound closure in the cornea of frog eyes. Tissue culture studies have been carried out by Loeb (1902 and 1912), Harrison (1910 and 1914), Holmes (1913 and 1914), Oppel (1913), Uhlenhuth (1914 and 1915), Hooker (1914), Matsumoto (1918), Maximow (1925), and others. For the most part these investigators agree that movement of epithelial cells over a wound surface is ameboid and usually stereotropic. The cause initiating the cell movement is not definitely known. Rand (1905) suggested a wound stimulus : " The most important factor in the earlier part of the process of reparation is cytotaxis ; the individual columnar cells of the existing epidermis are affected by some directive stimulus and respond by an active migration, which results in the cover- ing of the surface of injury by a protective epithelium — the first de- finitive step toward regeneration" (p. 52). Taube (1923) concluded that materials of injured cells flow over wound surfaces and serve as a means of chemical stimulation for the uninjured cells at the edge of the wound, causing them to migrate and later to divide. Maximow (1925) suggested that regenerative proliferation may be the result of direct stimulation by the injury or of the action of specific chemical sub- stances. The trend of recent conclusions is in the direction of the activation of epithelial cells by a wound hormone or wound stimulus of some sort. In the following pages are reported studies upon the skin of tad- poles of Rana clamitans. The studies were carried out with a view to MOVEMENT OF EPIDERMIS IN WOUND HEALING further understanding the movement of epidermal cells over a denuded area, the behavior of host tissue in relation to a skin graft, and especially the mechanics of radial arrangement of epidermal pigment cells around a wound. The animals used in these experiments were collected from ponds or streams near Cambridge, Massachusetts. They were usually 70 to 80 mm. in length. THE MECHANICS OF WOUND CLOSURE It has been frequently noted that epidermal wounds close very quickly. In tadpoles, skin wounds 5 to 10 mm. across may be entirely Transition Migrating Cells r ?— ^ar . , -_.JJ. T.^^™^ "X*ri».J « ] :. HHK£._ H""1* """^ — **- "™^ ~~-& ""• -- i_i.'JLt^53 Nor ma) Celts i r , • v-^ ; -'•a^fe.ffc*.^ %*SFTSi? ^ --„"• ^^ *>*-.•• >; .•.^^•v,> ^"T-^P*:'-^ > V;V:>:v: S%-;* n Derm/s FIG. 1. Section perpendicular to the surface of skin near a wound, showing migrating cells and the transition to those not migrating. The wound was a short distance to the left of the cells illustrated. closed in from 6 to 24 hours. Upon the occurrence of a wound, a quantity of blood always flows into the area and there coagulates. The coagulum usually nearly fills the wound temporarily, being about equal in thickness to the skin. Continuous microscopical examination, im- mediately following the operation, revealed that the edge of the epi- dermis starts to move toward the center of the wound within a few minutes after the wound is made. The epidermal layer advances over the coagulum, moving equally from all sides toward the center of the wound. If the edge of the wound is irregular, the advancing layer is correspondingly irregular in outline. Over large wounds the advancing epidermis thins out until it may be only one or two cells in thickness. The thinning may extend several millimeters distant from the wound, depending on the size of the wound. Figure 1 shows the limit of the thinned region of the layer and the transition to the normal condition which in this case is about 5 mm. from the edge of the rather large wound. The advancing cells are quite undifferentiated. Those of the 274 EARL H. HERRICK cuticular layer are hardly distinguishable from those of other layers of the epidermis. Sections parallel to the surface of the layer were cut from normal epidermis (Fig. 2) and from areas adjacent to a closing wound (Fig. 3). In such sections the normal cells exhibited fairly regular polygonal FIG. 2. Section parallel to the surface of normal epidermis. The solid black areas represent unexpanded, epidermal melanophores. FIG. 3. Section parallel to the surface of epidermis near a wound, showing cells elongated in the direction of movement. The cells are moving in the direc- tion of the arrow. The elongate black areas represent expanded, epidermal melanophores. outlines while in an area near a closing wound the cells were from two to four times as long as wide, with the elongation in the direction of movement. This elongation of cells occurs at all points between the edge of the advancing layer and the region of transition to the un- disturbed epidermis. Several writers have suggested that epithelial cells elongate perpendicularly to the edge of a wound, but an extensive reading of the literature discovers no mention of appropriate sections having been made to establish the fact. Sections perpendicular to the surface merely suggest the possibility of elongation. When advancing edges of epidermis meet at or near the center of MOVEMENT OF EPIDERMIS IN WOUND HEALING 275 the wound, there often takes place a piling up of cells. This piling up suggests that the causal factor for migration continues to operate for a time after the advancing edges have come together. After the wound has been covered by a layer of epithelium, the coagulum disappears. AYithin two or three days mitosis begins and the layer over the wound continues to increase until it is somewhat thicker than the epidermis in other parts of the body. Also the deficient epidermis surrounding the wound is built up, by cell proliferation, to the normal thickness of the layer. In wounds made entirely through the skin the dermis is slow in closing, but the epidermis closes as usual, subsequently thickening until it is much thicker than the normal epidermis. Sometimes the thickness of the layer over a healed wound may be equal to the combined thickness of the normal epidermis and the dermis. When transplantations are made, the small gaps that necessarily exist between host skin and graft are closed by epidermis which grows in to fill the deficiency as if to maintain a smooth surface to the body. As the dermis is built up, the epidermis returns to its usual thickness. The function of epithelium may be not only " to cover a surface continuously," as declared by Rand, but also to restore a smooth external surface after small injuries. SKIN GRAFTS AND THE COMPATIBILITY OF TISSUES The diversely pigmented integument of amphibians, possessing ex- traordinary capacity for repair and regeneration and readily amenable to grafting operations, offers many obvious advantages for the study of problems in tissue compatibility. A patch of ventral, unpigmented skin from a tadpole was trans- planted to the back of the same animal. Twenty-six autotransplants of this kind were made. In each instance the transplanted piece was 5 to 7 mm. square. Several operations were performed in which a dorsal patch of skin was merely cut loose and then reimplanted where it had previously been. Other patches were rotated through 90° or 180° and reimplanted. The epidermis of an autotransplant united with that of the host immediately after the graft was made, but after that union had been effected there was no further movement on the part of the host tissue. In contrast to what happened in homoiotransplants, there was no in- vasion of the graft by host tissue. There was no sign of incompatibility between the ventral and dorsal skin tissue. Once the epidermal layers had united, the regenerative activity ceased. Dorsal pigmented patches which were rotated or reimplanted in their original positions were not disorganized and in a few weeks the limits of the grafts could not be 276 EARL H. HERRICK recognized. The half-white, half-pigmented, lateral patches which were rotated through 180° remained unchanged, the light area of the patch remaining unpigmented and the pigmented end retaining its pigment with none spreading into the surrounding white area (Fig. 4). Animals bearing this type of graft were observed for as long as 80 clays. Collins and Adolph (1926), however, observed disorganization of rotated skin grafts in experiments on Trititrus. It seems clear that no incompatibility exists between dorsal and ventral skin of an individual tadpole, although, in respect of pigment, they are locally specific. FIG. 4. Frog tadpole showing lateral, rotated graft. Homoiotransplants were made by grafting ventral unpigmented skin from one animal onto the back of another. The grafts healed in place very rapidly so that in a few hours a microscopic inspection of the tissue in the region of the union would scarcely reveal just which were host and which were graft cells. Not all of the movement, however, in covering such gaps is accomplished by the host epidermis, for that of the patch moves almost as rapidly until the two unite. The dermal layers of host and patch are much slower in uniting than the correspond- ing epidermis, requiring many days or even weeks. Sixty homoiotransplants were observed. \Yithout exception they be- came occupied by host epidermal cells ; and, if the animals lived long enough, dermal cells entered the unpigmented area. The grafts began to become pigmented in from 1 to 24 days after transplantation, the average being 9. The host epidermis, carrying the epidermal pigment cells, moved centripetally into the graft quite equally from all sides. Once having started, an average of 6 days elapsed before the pigment- bearing host epidermis reached the center of the patch. From a cursory examination of the surface, it was believed that the host epidermis grew over the patch, covering the epidermis as a whole. A study of micro- scopic sections, however, gave no indication of overgrowth or under- growth of the epidermis, in accord with the account given by Cole (1922). The epidermis of the patch evidently undergoes destruction and is simultaneously replaced by that of the host. What appeared to be the remains of disintegrated graft cells were found in the cells at the edge of the advancing host epidermis, suggesting phagocytosis. That epidermal cells may be phagocytic at times has been stated by Loeb (1902. 1912). MOVEMENT OF EPIDERMIS IN WOUND HEALING 277 The invasion of a graft by host epidermis is essentially a process of delayed wound healing. The delay is necessitated by the presence of the graft epidermis. Except for complications occasioned by the re- moval of the patch epidermis, the behavior of the host skin in relation to the homoiotransplant is nearly like that in ordinary wound closure. The host cells near the patch become elongated in the line of movement just as in wound healing (Fig. 3). Fourteen transplantations were made in which patches of dorsal skin were transferred to wounds on the ventral surface of the animal. Whether the graft was an auto- or a homoiotransplant, healing took place as usual, but in the course of a few days the patch showed signs of degeneration and subsequently sloughed off until all or the greater FIG. 5. Photograph of corner of skin graft showing the network of enlarged blood vessels. part was lost. It is believed that this condition was not necessarily the result of incompatibility between the tissues but that the ventral side of the body, with its very thin body wall, is an unfavorable place for a graft to receive nourishment. BEHAVIOR OF BLOOD VESSELS IN SKIN GRAFTS Nearly all skin transplants early become red because of the en- largement of their blood vessels. The degree of redness varies from slight traces to a deep, solid red covering the entire patch. The red- dening begins as blood vessels here and there over the patch become enlarged for very short distances. The enlarged regions then extend until a close network of such vessels may be seen over any portion or all of the patch (Fig. 5). The redness becomes apparent usually the 18 278 EARL H. HERR1CK second day after the graft is made, but it has been observed as early as 18 hours after an operation, with definite circulation in the enlarged vessels at 24 hours. The redness may persist for one to two weeks. In the greater number of cases the blood vessels enlarged until, in a few or many places in each graft, they broke to release blood into the loose dermal tissues, causing the appearance of a continuous layer of blood. This stage persisted usually less than a week. As it cleared, blood vessels of normal size could be seen, suggesting repair of the vessels. Since the vessels that occasioned the temporary redness be- came filled with blood within the short time of a few hours after the operation and because their arrangement is so similar to that of the vessels normally present in the ventral skin, it is probable that they are not newly developed after transplantation, but are the original vessels of the grafted skin. Cole (1922) has described the reddening of homoiotransplants of tadpole skin but states that autotransplants do not show this reaction. He says (page 391) : " The occurrence of such a process is evidence of a rather violent chemical reaction going on between the protoplasms of the graft and host, and a merely specific difference between the two could set up the reaction." Cole's conclusions, however, are not in accord with the experiments described in the present paper, for in these experiments both auto- and homoiotransplants became reddened in nearly all cases. Since white autotransplants never became pigmented and since there was no other indication of any reaction between tissues, it seems evident that increased vascularity is not an indication of in- compatibility between tissues. Furthermore, there is considerable evi- dence that enlargement of the blood vessels is due to mechanical stretch- ing of their walls. As circulation becomes established in the graft, the blood exerts pressure on the walls of the vessels. In normal tissues, tonus of the vessels balances this internal pressure. There is slight variation in the caliber of normal vessels to meet the varying needs of the animal, the tonus being controlled by the nervous system. In transplanted patches the nervous connections have been cut during the operation, leaving the blood vessels of the patch without normal tonus, only a certain elasticity intrinsic to the vascular tissue itself persisting. It is highly probable, therefore, that enlargement of blood vessels fol- lowing transplantation is due to stretching of the vessels from internal pressure in the absence of nervous control and not to a specific reaction between tissues as concluded by Cole. May (1924), after transplanting lizard (Anolis} skin, concluded (p. 553) : " Transplants of pigmented skin lose their colour-changing MOVEMENT OF EPIDERMIS IX WOUND HEALING 279 power immediately on being completely disconnected from neighbouring tissues, and regain it slowly as nerves regenerate." It has not been ascertained just when nerves regenerate to innervate the grafted tadpole skin, but the apparent resumption of control of the blood vessels- agrees in time with that given by May for the regeneration of nerves in the skin of Anolis. May's conclusions apply very well to the apparent loss and regaining of nerve control of the blood vessels of the tadpole grafts. RADIAL ARRANGEMENT OF EPIDERMAL PIGMENT CELLS AROUND A WOUND OR SKIN TRANSPLANTATION When expanded epidermal pigment cells lie upon or near a skin wound which is healing or a transplanted patch of skin which is be- FIG. 6. Photograph showing the nearly parallel arrangement of epidermal pig- ment cells near a wound. coming pigmented, they show a particular arrangement in relation to the wound or graft. In general, the axes of elongated melanophores are perpendicular to the edge of the wound or, in other words, they are parallel to the line of movement of the epidermal cells toward the wound or graft. In any one localized area the greater number of epidermal melanophores are parallel (Fig. 6). Some cells may be at an angle to others, or occasionally a melanophore may be perpendicular to the axes of the greater number, but such cells are too few to impair the con- spicuous radial arrangement of the majority. This arrangement may extend for several millimeters external to the edge of the wound or graft. The area over which radial arrangement may take place was found to coincide with the extent of the region within which migration 280 EARL H. HERRICK of the epidermal cells occurs. In Fig. 1 the limit of radial arrange- ment occurs at the point labeled " transition." This observation is very significant in relation to the mechanics of the radial arrangement. It is obvious that radial position of a pigment cell will not be evident unless the cell is expanded. Radial arrangement depends upon ex- pansion of the melanophores, but the causes of radial arrangement and expansion were found to be distinct. Since investigation of the causes of radial arrangement could be carried on only when the epidermal melanophores were expanded, it was desirable to find an artificial means of producing expansion. The following technique was used. An injection fluid was prepared by adding desiccated bovine pituitary gland to Ringer's solution and filter- ing the mixture through filter paper. The powdered gland material dis- solved and filtered better if the liquid were slightly warm. About one gram of desiccated gland was used in 10 cc. of Ringer's solution, but wide variations of this concentration proved to be effective. Intra- peritoneal injections were made, using a small hypodermic syringe with a dosage of about 0.2 cc. After an injection of this kind the melano- phores began to expand in about twenty minutes and continued expand- ing for one or two hours until maximal or nearly maximal expansion was reached. Allen (1917), Swingle (1921), and Collins and Adolph (1926) have studied in considerable detail the effect of pituitary gland and its extracts on pigment cells, but in the experimental work now being described the pituitary extract was used merely to expand melano- phores in connection with the study of the problem of their radial arrangement. Expanded epidermal melanophores near a wound which was being covered, or around grafts toward which epidermis was moving, ex- hibited more or less marked radial arrangement. Therefore radial ar- rangement of melanophores may be expected wherever there is a migrating epidermis whose melanophores are expanded. After a wound or graft has been covered by the surrounding epidermis and the layer has nearly returned to its normal thickness, the melanophores begin to lose their special arrangement. In homoiotransplants radial arrange- ment persisted from one to two weeks or even longer. In one large homoiotransplant, after it had become pigmented and the radial arrange- ment lost, two holes, about 3 mm. apart and each about 1 mm. across, were cut through the skin of the graft. The pigment cells were partially expanded. Twelve to eighteen hours afterward there was marked radial arrangement around each of the holes. In considering the mechanism of radial arrangement of melano- phores, the conditions of both normal and migrating cells must be kept MOVEMENT OF EPIDERMIS IN WOUND HEALING 281 in mind (Figs. 2 and 3). During the migration of the epidermal cells, their space relationships are continually changing. Further, the epi- dermal cells closest to the edge of the wound or graft move fa-ster than those successively farther from the ivound or graft. In the initial advance of cells in covering the wound, the cells closest to the edge of the wound travel a greater distance than cells successively farther from the edge. It is believed that the mechanism of radial arrangement may be completely described from the facts just stated. Unless a pigment cell is lying with its long axis exactly perpendicu- lar to the line of movement of the epidermal cells, the end of the melanophore nearer the edge of the wound will be carried faster than the remainder of the melanophore. The carrying of one end of the cell faster than the other in a given direction will tend to rotate the comparatively long pigment cell to lie in a new position, more nearly in the line of movement of the epidermal cells. Since a single epi- dermal melanophore may be as long as the diameters of twenty epider- mal cells, this action could easily occur. Continuous microscopic observations made on living epidermis at the edge of a wound support in every detail the suggestion just made. Such observations were made continuously on a single wound for as long as twelve hours. Suggestions for apparatus used in making the observations just mentioned were obtained from Clark (1912). A small glass box was constructed, part of one side of which was made of " cover slip " glass to permit the use of high-power microscope objectives. The box was filled with paraffin to a depth of about 5 mm. The tube of an ordinary compound microscope was turned to the horizontal position and the glass box was secured to the mechanical stage. With the box in this position, the open side was uppermost and the side made of the thin glass was next to the objective lens of the microscope. When a wound was to be studied, the tadpole was first anesthetized by placing it in water to which had been added a few drops of ether. As soon as anesthesia was produced, the animal was placed in the glass box, which contained a 0.05 per cent chloretone solution. Chloretone is a very slow-acting anesthetic and causes melanophores to expand. Ether was used as described merely to hasten immobilization of the animal before the melanophores could expand. The animal was held against the thin side of the box by means of pins thrust into the paraffin. Usually transmitted light from the mirror or directly from a lamp was used, but reflected light with or without transmitted light was some- times used. Sufficient light could be transmitted through the tail of a tadpole to permit the use of an oil-immersion lens, but the best results 282 EARL H. HERRICK were obtained by means of a 4 mm. lens and a 10 X to 18 X ocular. Magnifications up to nearly 1000 diameters were frequently obtained. The animals remained alive in the glass box for over 24 hours, so that very satisfactory records were obtained up to that length of time. A vertical drawing board was used to permit the employment of a camera lucida. Individual pigment cells were located and followed for several hours. Hence by making drawings at frequent intervals, the exact courses of the cells could be followed. The movement was too slow to be observed except bv means of a series of drawings. In following the movements of melanophores, the neighboring- epidermal cells were observed at the same time. It was found that the 11=00 3-00 12.00 4:00 Z:00 5:00 FIG. 7. First section of series of camera lucida drawings showing the paths taken by epidermal melanophores as they are carried toward a wound, coming to lie with long axes nearly perpendicular to the edge of the wound. The arrow indicates the direction of movement. melanophores did not orient themselves irrespective of epidermal cells but moved with them and at approximately the same rate. Several in- vestigators have concluded, from tissue culture studies, that melano- phores possess the power of self-movement. The above described ob- servations afford no ground for doubting that melanophores do have the power of self -movement, but there is every reason to believe that, under the conditions of the experiments in question, they are no more active than the other epidermal cells and that their radial arrangement depends upon the movement of the epidermal cells and not upon the activity of the melanophores themselves. Figures 7 and 8 are series of camera lucida drawings of groups of pigment cells which were near MOVEMENT OF EPIDERMIS IN WOUND HEALING 283 healing wounds. The drawings show the successive positions of the melanophores as they became more nearly perpendicular to the edges of the wounds. The time intervals between successive drawings are indicated. With regard to radial arrangement of pigment cells, there was a striking difference between auto- and homoiotransplants. There was never extensive radial arrangement around autotransplants. This is not strange, however, since in autotransplants there is little or no migra- tion of epidermis after the epidermis of the host meets that of the graft. In fact, this is further evidence in support of the proposed explanation for radial arrangement. 900 930 1100 FIG. 8. Second section of a series of camera lucida drawings showing the paths taken by epidermal melanophores as they are carried toward a wound, coming to lie with long axes nearly perpendicular to the edge of the wound. The arrow indicates the direction of movement. The writer wishes to express his sincere appreciation to Dr. II. W. Rand for his advice and aid throughout the period of this work. SUMMARY 1. Immediately following the occurrence of wounds in the skin of tadpoles of Rana clamitans, the surrounding epidermis moves to cover the area. Wounds 5 mm. across may be covered in from 6 to 24 hours. The advancing epidermis is reduced to one or two cells in thickness, thus covering a larger area than previously. The epidermal cells be- come considerably elongated in the direction of movement. After the 284 EARL H. HERRICK advancing edges have met, the cells resume their previous shape. Sub- sequent mitosis restores the original thickness of the layer. There is considerable evidence for stereotropism in epidermal migration. 2. In wounds which cause a deficiency in the dermis, the epidermis thickens over the wound area until the smooth contour of the body- surface is restored. 3. In the region surrounding a healing wound there is temporarily a deficiency in the number of epidermal pigment cells due to their movement into the wound. 4. Autotransplants of white ventral skin transferred to the pig- mented backs of tadpoles retained their specificity. They remained unpigmented for more than 100 days. Lateral and dorsal autotrans- plants in which the patches of skin were rotated through 90 or 180° also retained their specificity. There was no change in the pigmenta- tion or the cellular structure of such grafts. 5. In all cases where homoiotransplants of white ventral skin were transferred to the pigmentecl backs of tadpoles the region of the graft became pigmented. Epidermal pigment appeared at the edge of the graft in an average of 9 days after the graft was made, and in the course of about 6 more days had extended to the center of the patch. Dermal pigment entered the graft area many days after the epidermal pigment had completely covered the patch. 6. The initial pigmentation of grafts results from a replacement of the graft epidermis by that of the host. The pigment cells are carried along by the host epidermis as it covers the patch. There is considera- ble evidence that the graft epidermis is destroyed by phagocytic action of the host epidermis. However, the epidermal layers of graft and host always unite immediately after the graft is made, indicating some initial affinity between the two layers. 7. Patches of skin which were grafted onto the ventral side of the body always degenerated. The ventral side of the body is perhaps an unfavorable place for grafts to receive nourishment. 8. The replacement of the epidermis of a graft by host epidermis is, essentially, a process of delayed wound healing. 9. In nearly all cases grafts, soon after transplantation, became reddened as the result of an enlargement of the dermal blood vessels of the graft. There is considerable evidence that the enlargement of the blood vessels is due to blood pressure which mechanically stretches the vessels in the absence of the normal tonus, the nerve connections having been cut at the time of transplantation. Later, and probably in con- sequence of restoration of nerve connections, the dermal vessels became reduced to normal proportions. MOVEMENT OF EPIDERMIS IN WOUND HEALING 285 10. Radial arrangement of the epidermal melanophores occurs around homoiotransplants and skin wounds — that is, wherever there is translator}' movement of the epidermis. In any migrating epidermis in which the melanophores are expanded, their long axes tend to become parallel to the direction of migration. The position of the individual nielanophore, in radial or parallel alignment, is a consequence of the movement of the epidermal cells near a wound and is not due to inde- pendent orientation of the melanophores. The epidermal cells nearest the wound move first and most rapidly, therefore tending to rotate the elongated melanophores into the line of movement of the epidermis. Since the movement of the epidermis is centripetal, the melanophores become arranged radially around the point toward which the epidermis moves. BIBLIOGRAPHY ALLEN, B. M., 1917. Effects of the Extirpation of the Anterior Lobe of the Hy- pophysis of Rana pipiens. Biol. Bull., 32: 117. AREY, L. B., 1925. The Method of Repair in Small Wounds. Anat. Rcc., 29: 345. BARFURTH, D., 1891. Zur Regeneration der Gewebe. Arch, inikr. Anat., 37: 406. BORN, G., 1896. Uber Verwachsungsversuche mit Amphibienlarven. Arch. Entiv.- Mech., 4: 349. CLARK, E. R., 1912. Further Observations on Living Growing Lymphatics : Their relation to the mesenchyme cells. Am. Jour. Anat., 13: 351. COLE, W. H., 1922. The Transplantation of Skin in Frog Tadpoles, with special reference to the Adjustment of Grafts over Eyes, and to the Local Speci- ficity of Integument. Jour. Ex per. Zoo!., 35: 353. COLLINS, H. H., AND E. F. ADOLF 11, 1926. The Regulation of Skin-pattern in an Amphibian, Diemyctylus. Jour. Morfih. and Physio!., 42: 473. EYCLESHYMER, A. C., 1907. The Closing of Wounds in the Larval Necturus. Ant. Jour. Anat., 7: 317. FRAISSE, P., 1885. Die Regeneration von Geweben und Organen bei den Wirbel- thieren, besonders Amphibien und Reptilien. Cassel und Berlin. Verlag von T. Fischer. HARRISON, R. G., 1910. The Outgrowth of the Nerve Fiber as a Mode of Proto- plasmic Movement. Jour. Ex per. Zoo/., 9: 787. HARRISON, R. G., 1914. The Reaction of Embryonic Cells to Solid Structures. Join: Ex per. Zoo/., 17: 521. HOLMES, S. J., 1913. Behavior of Ectodermic Epithelium of Tadpoles when Cultivated in Plasma. Univ. Cal. Pnbl., Zoo/., 11: 155. HOLMES, S. J., 1914. The Behavior of Epidermis of Amphibians when Cultivated Outside the Body. Jour. Expcr. Zoo/., 17: 281. HOOKER, D., 1914. The Development of Stellate Pigment Cells in Plasma Cul- tures of Frog Epidermis. Anat. Rcc., 8: 103. LOEB, L., 1902. On the Growth of Epithelium in Agar and Blood-serum in the Living Body. Jour. Med. Res., N. S., 3: 109. LOEB, L., 1912. Growth of Tissues in Culture Media and its Significance for the Analysis of Growth Phenomena. Anat. Rcc., 6: 109. LOEB, L., 1920. A Comparative Study of the Mechanism of Wound Healing. Jour. Mcd. Res., 41: 247. LOEB, L., AND R. M. STRONG, 1904. On Regeneration in the Pigmented Skin of the Frog, and on the Character of the Chromatophores. Am. Jour. Anat., 3: 275. 286 EARL H. HERRICK MATSUMOTO, S., 1918. Contribution to the Study of Epithelial Movement. The corneal epithelium of the frog in tissue culture. Jour. Exper. Zool., 26: 545. MAXIMOW, A., 1925. Tissue-cultures of Young Mammalian Embryos. Carnegie Inst. of Washington. Contributions to Embryology, vol. 16, no. 80, pp. 47-113. MAY, R. M., 1924. Skin Grafts in the Lizard, Anolis carolinensis, Cuv. Brit. Jour. Exper. Biol., 1: 539. MORGAN, T. H., 1901. Regeneration. Macmillan, New York. OPPEL, A., 1912. Causal-morphologische Zellenstudien. V. Mitteilung; Die ak- tive Epithelbewegung, ein Faktor beim Gestaltungs- und Erhaltungs- geschehen. Arch. Entiv.-Mech., 35: 371. OPPEL, A., 1913. Demonstration der Epithelbewegung im Explantat von Frosch- larven. Anat. Anz., 45: 173. PETERS, A., 1889. Ueber die Regeneration des Endothels der Cornea. Arch. mikr. Anat., 33: 153. RAND, H. W., 1905. The Behavior of the Epidermis of the Earthworm in Regen- eration. Arch. Entw.-Mech., 19: 16. RAND, H. W., 1915. Wound Closure in Actinian Tentacles with Reference to the Problem of Organization. Arch. Entiv.-Mech., 41: 159. SWINGLE, W. W., 1921. The Relation of the Pars Intermedia of the Hypophysis to Pigmentation Changes in Anuran Larvae. Jour. Exper. Zool., 34: 119. TAUBE, E., 1923. Uber die histologischen Vorgange bie der Regeneration von Tritonen mit Beteiligung ortsfremder Haut. Arch. mikr. Anat. und Entw.-Mech., 98: 98. UHLENHUTH, E., 1914. Cultivation of the Skin Epithelium of the Adult Frog, Rana pipiens. Jour. Exper. Med., 20: 614. UHLENHUTH, E., 1915. The Form of Epithelial Cells in Cultures of Frog Skin, and its Relation to the Consistency of the Medium. Jour. Exper. Med., 22 : 76. OXYGEN DEFICIENCY AND SEWAGE PROTOZOA: WITH DESCRIPTIONS OF SOME NEW SPECIES * JAMES B. LACKEY (From the Seicagc Investigations Laboratory, X'cu.' Brunswick, X. J., and South- western College, Memphis, Tennessee) It is a common observation that the protozoa of hay infusions and other liquid media generally seek the top of the culture where oxygen is to be found. Investigation of a rich culture either in a jar or on a slide shows that only a few inhabit the oxygen-poor regions : Mctopus among the ciliates ; some of the very small flagellates and small amoebas, mostly of the limax type. Anaerobic protozoa are well known. Juday (1919) has described a freshwater anaerobic ciliate ; Lauterborn (1908) has discussed a number of species, many of them bizarre forms, from the oxygen-poor waters of the Rhine. Lackey ( 1925, 1926) has listed twenty-nine species common in sewage containing little or no dissolved oxygen, but often abundant H2S or COo ; while Cole (1921) has dis- cussed the oxygen supply of animals living under such conditions. Most of the protozoa inhabiting sludge tanks are small and do not occur in great numbers — except those which live at the very top and are not considered in this paper, for they have an abundance of oxygen and may be transient forms in the tanks as well — so that they are not often noted by students unless a careful search is made. All of those ciliates characterized by Noland (1925) as living in water whose oxygen saturation is below 45 per cent have been encountered at one time or another, but usually in small numbers, with the exception of Mctopus sigmoidcs. As a result of examining the protozoan fauna of the diges- tion tanks of five New Jersey towns, the Passaic River in some of its most polluted stretches, and septic tanks from three locations in Ten- nessee, it has been found that the same small group of protozoa is to be found in each of these situations. Of the forms listed previously (Lackey, 1925) two ciliates and one flagellate were described as new species. The present paper includes four new flagellates and two new rhizopocls. The wide distribution of this group argues that the factors limiting their occurrence are very largely those of anaerobism, for such conditions as pH, temperature, dissolved gases, and food substances must have varied widely in the 1 Journal Series Paper N. J. Agricultural Experiment Station, New Bruns- wick, N. J. Dept. Sewage Research. 287 JAMES B. LACKEY several locations examined. Sufficient check has been made on the first two of these conditions in sludge digestion to make this assumption plausible, and it is known that they certainly vary from time to time, and that the amounts of CO2 and H2S fluctuate considerably also in any given tank, apparently without a corresponding fluctuation of the protozoan fauna as a whole, in the same tank. It is also certain that while all sewage contains decomposition products which can possibly serve as food for the organisms, the variety of these substances is too great to be a probable limiting factor. To ascertain the effects of dissolved oxygen on two protozoa char- acteristic of sewage disposal plants, tall cylinders were set up, so that compressed air could be forced through porous plates in their bases. This allowed aeration of the sewage (eventually producing activated sludge) or kept the sewage under completely anaerobic conditions. Opercularia sp., a large peritrichous holozoic ciliate and an active bac- terial feeder, was selected as the obligatory aerobe, for it is abundant in the trickling filters, but never occurs in its active state in the tanks. Trepomonas was selected as the obligatory anaerobe, for it is charac- teristic of the depths of the tanks. Table I shows the behavior of these two forms in the cylinders under conditions varying from no aeration to constant aeration. Preliminary examination of the sewage at the treatment plant showed 3300 Trepomonas and 500 Opercularia per cc., and after standing in a refrigerator 24 hours there were 2900 Trepo- monas per cc., and no active Opercularia. It will be noted that the sludge when examined at the disposal plant contained both active Opercularia and Trepomonas. There was only a trace of oxygen pres- ent. The sample was taken to the laboratory and put in the refrigerator and 24 hours later contained no dissolved oxygen. Its protozoan popu- lation had changed in this time ; all colonies of Opercularia contained only closed-up individuals. After 6 hours aeration 65 per cent of the Opercularia were active, whereas with no aeration none of the Opercularia, were active. No count was made, but the normal population of Trepomonas seemed to be present. Under aerobic conditions, however, the 6 hours had re- duced the numbers of Trepomonas almost 96 per cent. Many dead or apparently dead ones were seen. Their protoplasm was greatly vacuo- lated and the cells sometimes much above normal size. Twenty-four hours of aeration was sufficient to cause them to dis- appear and they never reappeared in this sample. All the Opercularia, on the contrary, became active and showed an increase in numbers for 24 hours, then a gradual decrease. When the experiment was dis- OXYGEN DEFICIENCY AND SEWAGE PROTOZOA 289 continued after 196 hours, there were few protozoa of any sort in either aerated or unaerated sewage, but most of the Opcrcularia were active in the aerated, although they presented a rather starved appearance, due probably to disappearance of their food. Although the results practically speak for themselves, supporting evidence is easily obtained. During an examination of settled sludge from the Chatham, N. J., activated sludge plant, on August 2, the sampling showed no dissolved oxygen and there were present 9200 TABLE I Numbers of Organisms Present With and Without Aeration (Numbers refer to active organisms throughout) TIME AERATED ..UN- AERATED No. per cc. No. per cc. After 6 hours Trepomonas 140 No count Operculana 3200 0 After 24 hours Trepomonas 0 3300 Opercularia 12600 0 After 48 hours Trepomonas 0 4500 Opercularia 9300 0 After 120 hours Trepomonas 0 5400 Opercularia 8300 0 After 148 hours Trepomonas 0 4000 Opercularia 4500 0 After 172 hours Trepomonas 0 2000 Opercularia 4500 0 Trepomonas per cc., and only about 260 Opcrcularia per cc. The Opercularia were all closed up, inactive, while the Trepomonas were swimming about vigorously. This sample was then aerated for 3 hours, when an examination showed that all the Opcrcularia were actively feeding while about half the Trepomonas were killed and half active. At the end of 6 hours aeration there were no signs of living Trepo- monas in this cylinder, although the Opcrcularia were thriving and numerous small flagellates such as Bodo, Ccrcobodo, Pleuroinonas, Monas, and Tetramitus were active. After 49 hours aeration, aj larg£ 290 JAMES B. LACKEY protozoan fauna was present, the small flagellates of the group listed above being 15,300 per cc., and Trepomonas being absent. In a control of unaerated sludge there were, per cc., 5900 Trepomonas, 900 Tetra- mitus, active, and perhaps 175 inactive Opercularia. It is not intended to list the protozoa occurring in activated sludge or trickling filters or, in general, in sewage which contains much dis- solved oxygen. But from the above experiments and similar ones, it is apparent that many protozoa encyst when swept into the oxygen-free water of a digestion tank or the bottom of a river heavily polluted with sewage ; that like Opercularia, they are able to reduce their metabolic activities to a minimum and so survive for a time, at least, in this adverse environment. If conditions become favorable again, they excyst and become active. It is also seen that there are some forms which live under either condition and some which thrive only in the absence of oxygen. This latter factor is so sharply limiting that we find a small and cosmopolitan group of protozoa which may be regarded as common in Imhoff tanks, septic tanks, and the deeper waters of heavily polluted streams. Table II does not include many forms which have been seen, but not sufficiently studied for proper identification. Some of these will undoubtedly prove to be new species when better known. It is there- fore seen that careful investigation of this particular type of habitat will be productive of an acquaintance with a decidedly unusual and interesting group of protozoa. DESCRIPTION OF NEW SPECIES The following descriptions are concerned with two new rhizopods and four new flagellates, which have been found in studying the waters of sewage disposal plants and the water of a creek near Camden, N. J., which was heavily polluted with sewage. Only one of the flagellates, Chroomonas cyaneus, was found in the creek, the others being from sewage disposal plants. The new species are illustrated in Plate I, enlarged about 1000 times. VAHLKAMPFIA FRAGILIS, Sp. Nov. Fig. 4. Organism small, free-living, 5 to 15 microns long, naked, relatively abundant in some sewage, especially in cultures. Pseudopodia lobose, two to three in number, clear, steadily formed rather rapidly and always at end of cell opposite contractile vacuole and nucleus. One contractile vacuole, always posterior to the nucleus, constantly present, emptying seldom. Nucleus always oval, flattened, with two to OXYGEN DEFICIENCY AND SEWAGE PROTOZOA 291 three large endosomal granules (chromatin masses?) in center. Nu- clear membrane delicate. Several hyaline spheres of varying size, oil or albumen, located in posterior region. Endoplasm finely granular, never extending into pseudopodia. Xo flagellated stage found. TABLE II Group I Present Only in Small Xinnbers or Infreqitency in Digestion Tanks RHIZOPODA Dimastigamoeba gruberi Hartmanella hyalina Vahlkampfia limax Vahlkampfia albida Vahlkampfia guttula Chlamydophrys stercorea FLAGELLATA Mastigella simplex Dinomonas vorax Cercobodo longicauda Tetramitus- descissus Cercobodo crassicauda Tetramitus pyriformis Cercobodo ovatus Hexamitus inflatus Monas amoebina Clautriavia parva Monas minima Euglena gracilis Monas vulgaris Menoidium incurvutim Anthophysa vegetans Peranema trichophormn Helkesimastix faecicola Distigma proteus Bodo caudatus Petalomonas mediocanellata Bodo lens Petalomonas carinata Bodo mutabilis Heteronema acus Pleuromonas jaculans Entosiphon sulcatum Oicomonas termo Notosolenus orbicularis Cyathomonas truncata Chilomonas paramecium CILIATA Hexotrichia caudatum Colpoda cucullus Colpoda inflata Cyclidium glaucoma Glaucoma scintillans Paramecium putrinum Plagiopyla nasuta Group II . //ii'i/v.f Present in Absence of Oxygen in Tanks RHIZOPODA Chlamydophrys minor, sp. nov. Vahlkampfia fragilis, sp. nov. FLAGELLATA Mastigamoeba viridis Mastigamoeba radiosa, sp. nov. Mastigamoeba longifilum Trepomonas agilis Mastigamoeba reptans Bodo glissans, sp. nov. CILIATA Holophrya sp. Saprodinium putrinum Metopus sigmoides Trimyema compressa 292 JAMES B. LACKEY Nutrition apparently saprozoic. Division binary in active stage. Forms very small cysts, thick- walled, with a few slight protuberances. Excysts in fresh raw sewage (by putting cover-slips in Petri dishes of old cultures, cysts are collected which can be watched on being transferred to fresh material). Rather common in five samples obtained. MASTIGELLA RADIOSA, Sp. Nov. Fig. 2. Animal with spherical body, about 20 microns in diameter, free- swimming or floating. Pseudopodia much like those of Amoeba radiosa, occasionally branching, clear, up to 80 microns long. Single flagellum up to 100 microns long, used with a lashing or whiplike motion. One contractile vactiole. Nucleus central, spherical with endosome, mem- brane very thin, chromatin granules very small. Cytoplasm finely granular, no zones being present. Several types of granules such as oil spheres, or crystalline bodies, present. Some apparent binary fission stages seen. Nutrition saprozoic. Rather rare ; found in two of the tanks examined. M AST I G AMOEBA VOLUTANS, Sp. Nov. Fig. 1. Body elongate, flattened, 25 to 30 microns long, with a flagellum, vibratile in anterior third, slightly tapering, one and a half to two times body length. Cortical layer of protoplasm clear, hyaline, endoplasmic portion finely granular, reticulate, with numerous small, square crystals and numerous spheres of varying size. Due to the fact that they stain with methylene blue and that in some of these animals a clear vacuole was found which contained from one to three of these spheres, they are assumed to be volutin. Pseudopodia very short, numerous, rounded. Contractile vacuole single, in posterior part of cell. Nucleus near anterior tip of cell, round in outline, with a central endosome surrounded by minute chromatin granules within a clear zone of nuclear sap. Membrane thin, with a rhizoplast visible in the living animal, from the flagellar insertion to its surface. Nutrition apparently saprozoic. Reproduction not observed. Somewhat common in samples from six localities. CHLAMYDOPHRYS MINOR, Sp. Nov. Fig. 3. A very small form, showing constant differences from C. stercorca. Diameter of the transparent shell is 20 microns. Cytoplasm finely OXYGEN DEFICIENCY AND SEWAGE PROTOZOA granular, extending out as a few finely-branching pseudopodia, which only occasionally anastomose and never become thread-like. Contractile vacuole single, median. Nucleus centrally located, spher- ical, with a single endosome. A central zone of black, refringent, crescent-shaped granules is present. Nutrition apparently saprozoic. Reproduction not observed. Move- ment slow. Rather common in most sewage. BODO G LI S SANS, Sp. Nov. Fig. 6. A somewhat flattened elongate organism, rarely over 20 microns long, cell of definite shape, but quite plastic. Two flagella, emerging on the ventral surface, anteriorly, from a lip-like fold. Swrimming flagel- lum about body length, very tenuous and very active. Trailing flagel- lum slightly longer, thicker, used as an axis on which the animal glides. Forward movement rapid, path straight. "When the animal changes its path, its amoeboid nature becomes evident. No nucleus visible either in living specimens or those treated with neutral red or iodine. Some visible granules may be chromidia. One contractile vacuole, median. Protoplasm in anterior end mostly clear, but a few small spheres, probably oil, may be seen posteriorly. No visi- ble kinetic apparatus. Reproduction by longitudinal division, while ac- tive. Nutrition saprophytic. No cysts observed. Rather common in sewage from several locations. The animal is placed in the genus Bodo, following Calkins' (1926) key, but it is not holozoic, and the two flagella are not terminal but subterminal. None of the species described in Pascher's manual (1914) fit it. The position of the contractile vacuole and the inability to find an organized nucleus are also unique features. CHROOMONAS CYANEUS, Sp. Nov. Fig. 5. Organism pyriform, obliquely truncated anteriorly, tapering to a somewhat extended point posteriorly, never exceeding 10 microns in length. Two equal or subequal flagella emerge from a slight depression anteriorly. They are about body length, directed forward in swimming, and the path of the animal is a spiral. Two or more small contractile vacuoles are located at their base. There is no eyespot, nor can other structures be distinguished in this region. The nucleus is central and apparently there is no endosome, but only granular masses of chromatin. Chromatophores two in number, band-shaped, curved, of a bright blue color which speedily diffuses into the surrounding water when the animal 19 294 JAMES B. LACKEY disintegrates. Pellicle thin, rigid. One case of binary fission while active was noted, but no cysts were found. These organisms were not found in sewage, but were abundant in the waters of a creek, heavily polluted with sewage, near Collingswood, N. J. SUMMARY There is an unusual group of protozoa to be found in the oxygen- poor and oxygen-deficient waters of sewage treatment plants. Most of these are apparently facultative anaerobes. A few appear to be obligatory anaerobes. It is shown that the presence of dissolved oxygen in sewage allows Opercnlaria. to thrive, while the absence of oxygen is fatal to this pro- tozoon, if the condition endures for several days. The reverse condi- tions are found to obtain for Trcpoiuonas, except that active aeration of the water is quickly fatal. Six new species of protozoa from sewage or polluted waters are described. REFERENCES CALKINS, G. N., 1926. Biology of the Protozoa. Lea and Febiger, Philadelphia. COLE, A. E., 1921. Oxygen Supply of Certain Animals Living in Water Contain- ing No Dissolved Oxygen. Jour. Exper. Zo'oL, 33: 294. JUDAY, C., 1919. A Freshwater Anaerobic Ciliate. Biol. Bull, 36: 92. LACKEY, J. B., 1925. The Fauna of Imhoff Tanks. Bull. 417, N. J. State Agri- cultural Experiment Stations. LACKEY, J. B., 1926. Report of Zoologist in Fourth Annual Report of the Sewage Substation, N. J. State Agricultural Experiment Stations. LAUTERBORN, R., 1908. Zur Kinntnis einiger Rhozopoden und Infusorien aus dem Gebiete des Oberrheins. Zcitsch. f. IViss. Zool., 90: 645. XOLANB, LOWELL E., 1925. Factors Influencing the Distribution of Freshwater Ciliates. Ecology, 6: 437. PASCHER, R.. 1914. Die Susswasserflora Deutschlands, Osterreichs, und der Schweiz. Gustav Fischer, Jena. OXYGEN DEFICIENCY AND SEWAGE PROTOZOA 295 jAti^iO^^I^^ 7^j^L-J^-—--::-^f.J PLATE I MODIFICATION OF TRAITS IN MOSAICS FROM BINUCLEATE EGGS OF HABROBRACON * P. W. WHITING (From the Zoological Laboratory, University of Pittsburgh, and the Marine Biological Laboratory, Woods Hole, Massachusetts} Egg binuclearity has been suggested to account for the origin of a few rare mosaics in Drosophila and of hereditary mosaicism in Ly- inantria. Most of the mosaics which have been found in the parasitic wasp, Habrobracon juylaudis (Ashmead), are likewise best explained by this hypothesis. The theory advanced by the writer (Whiting, P. W., 1922) is that after extrusion of the first polar body, the second oocyte nucleus gives rise to two (reduced) nuclei which either take part in parthenogenetic cleavage (male mosaics) or segment after one has been fertilized (gynandromorphs). Post-reduction of binucleate eggs from heterozygous females would then result in mosaic males or, in case of fertilization of one nucleus, in gynandromorphs in which maternal contribution to male and female parts was different. Mosaics in Habrobracon are very infrequent. The present paper is based on 132 mosaic males and 92 gynandromorphs which have oc- curred scattered through the cultures of various investigators. When- ever a mosaic has been reported it has been given a serial number in the ' freak book " with a statement of its origin, a description, and any other pertinent data. The specimen, placed in a gelatine capsule, is then preserved in alcohol for future reference. MOSAIC MALES FROM HETEROZYGOUS MOTHERS Virgin females heterozygous for various genes have been bred in tests for linkage. Most of the mosaic males have been found in con- nection with these studies. They therefore arise from unfertilized eggs and are mosaic for one or more of those traits for which their mothers were heterozygous. It will be convenient first to give a record of these male mosaics as regards the various genes involved. Females heterozygous for orange, o (eyes) (Chromosome I), have produced thirteen sons mosaic for this trait. When tested by mating 1 The investigations here reported have been aided in part by grants from the Committee on Effects of Radiation on Living Organisms (National Research Council). The drawings for Plate I have been made by Kathryn A. Gilmore. 296 MODIFICATION OF TRAITS IN MOSAICS 297 with orange females two of these mosaics had type (black) daughters, one had orange daughters, and one had both black and orange daughters. The line of division between black and orange ommatidia is difficult to determine, showing more or less gradation. In general the orange re- gions of eyes of mosaics appear darker than in normal orange and ocelli contain more or less brown pigment so that it is sometimes diffi- cult to classify them as different from " wild type." Ocelli of normal " orange " contain more or less red pigment but not brown, while those of wild type contain much dark brown pigment. Females heterozygous for ivory, 0% have produced sons with mosaic eyes which may be classified as follows : nineteen had compound eyes described as mosaic for orange and black; nine for pale orange and black ; five had one eye ivory, the other ivory grading through orange to black ( Fig. 1 ) ; one had one eye black, the other ivory grading through orange to black. Observations were made on ocelli of thirty of these. They were classified as wild type, five ; more or less brown pigment, three ; very little brown pigment grading to colorless, five ; wild type grading to much red pigment, three ; more or less red pigment, ten ; completely devoid of pigment, four. Breeding tests made of ten of these showed five breeding as black, three as ivory, and two as black and ivory. Twenty showed mosaicisrn for other traits for which the mothers were also heterozygous. Females heterozygous for ivory have produced sons showing no very obvious mosaicism in eye color but which were mosaic in other traits of maternal origin affecting wings, legs, or body color. One with eyes and ocelli black bred as ivory ; one with ivory eyes and ocelli bred as black and ivory. One with orange eyes and a trace of brown pig- ment in ocelli bred as ivory. One with orange eyes and ocelli bred as black. Six had pale orange eyes among which one had a trace of brown pigment in ocelli and bred as black ; one had much red and brown pigment in ocelli ; one had much red pigment in ocelli ; while of the other three with colorless ocelli, one bred as black and one as ivory. Females heterozygous for ivory produced aberrant sons showing no obvious mosaicism in any trait. Eleven of these had eyes classified as orange among which two had colorless ocelli while nine had much red pigment in ocelli. One of the former and six of the latter showed by breeding test that they were actually mutants to orange and one of the latter bred as black and was therefore a mosaic. Seven had eyes classi- fied as pale orange among which one had no pigment in ocelli, two had much red pigment, while four had a small amount of red pigment. One of the last bred as black and was therefore mosaic and one bred as black stumpy and as ivory non-stumpy and therefore had mosaic gonads. P. W. WHITING The sixty-two exceptional males produced by females heterozygous for ivory all showed orange modification of eye color to greater or less extent except that one which bred as ivory had black eyes, and one which bred as black and ivory had ivory eyes. The entire group includes forty- four from untreated stock and eighteen from mothers X-rayed as larvae in twelve cases and as adults in six cases. Among the seven proved by breeding test to be mutants to orange only one was produced by an X-rayed (larva) mother. The seven mutants appeared entirely normal, showing mosaicism in no respect, not even as regards body pigmentation. There is no evidence that X-radiation has caused either modification or mutation to orange. An orange-ivory compound virgin female, o o' , produced a male (Fig. 2) with left eye orange, right orange dorsally, ivory ventrally with well-marked line of division between the two regions. Ocelli had red pigment granules characteristic of orange. The specimen bred as ivory. Females heterozygous for miniature, in (body, antennae, wings, legs) (Chromosome I), have produced two mosaic males showing clear-cut difference in antennae, wings, legs, and general body size. Females heterozygous for cantaloup, c (eye color) (Chromosome II), have produced fourteen mosaic males. Black and cantaloup re- gions are distinctly marked off in these mosaics in compound eyes and frequently even in ocelli in contrast to gradation observed between black and orange or black and ivory (Figs. 3—9). Females heterozygous for long, / (antennae, wings, legs) (Chromo- some II), have produced eight mosaic males, one breeding as long. Difference was clear-cut, being evident in antennae and wings, and could even be noted in legs. Females heterozygous for narrow, n (wings) (Chromosome II) have produced three mosaic sons with clear-cut difference showing in wings. Females heterozygous for Minnesota yellow, My (antennal seg- ments) (Chromosome II), have produced five mosaic sons. The char- acter is variable, modified by temperature, but the contrast is striking in antennae of these mosaics. Females heterozygous for reduced, r (wings) (Chromosome IV), have produced twelve mosaic sons. Difference is clear-cut and any one wing is either type or reduced. Tests showed four breeding as type and one as reduced. There were likewise produced from heterozygous mothers a reduced breeding as type and a type breeding as reduced, which were therefore also mosaic. Females heterozygous for fused, / (tarsi, antennae, wings) (Chromo- MODIFICATION OF TRAITS IN MOSAICS some IV), have produced two mosaic sons. Another male mosaic for fused, found in an inbred culture (No. 3), was in every other respect similar to Stock 3, hut proved sterile in observed matings with three females. The method of origin is uncertain, no other fused were found in the culture and it is therefore possible that this may have been a somatic mutant. The combination of traits is unmistakable and this specimen showed typical fused in tarsi, antennae and wings. Besides this case the locus has been known to mutate four times. Modification of fused in mosaics will be discussed below. A male mosaic for glass, g (eyes and antennae) (Chromosome IV), occurred in a mixed culture so that parentage is uncertain. The left antenna was very thin as in typical glass, the right type. The eyes were each mosaic with clear-cut regional distinction between glass and type (Figs. 20, 21), but the outline of ommatidia near the margin departs from the normal hexagonal form probably due to absence of pressure from other ommatidia in development. The glass regions are geneti- cally orange, the non-glass black. There is gradation of dark pigment into the orange regions as expected and outlines of bases of abortive ommatidia are rendered visible by the presence of this pigment. A male mosaic for wavy, wa (wings), and a male mosaic for broad, br (thorax) (Fig. 16), have each been produced from a heterozygous mother (Chromosome V). Two males mosaic for white, ivh (eyes) and one for attenuated, at (antennae) (Chromosome VI), have been produced by heterozygous females. White regions of compound eyes and of ocelli are, as in the case of cantaloup, sharply distinct from black, thus differing from genes in the orange series. (Figs. 10, 11, 12.) Two mosaics, each with left eye strikingly banded, were produced by mothers heterozygous for ivory and for cantaloup. The normal brothers had eyes of the three expected colors, — black (OC), cantaloup (Oc), and white (ivory, o'C, or ivory cantaloup, oV). Each mosaic had right eye orange or cantaloup in color which might be genetically cantaloup, Oc; ivory, o'C, modified by the presence of 0 in the mosaic; or ivory cantaloup, o'c. The last possibility is very unlikely since orange cantaloup, oc, is almost white and cantaloup with modified ivory, olc, should be no darker. The banded left eyes are shown in Figs. 13 (mosaic No. 550) and 14 (mosaic No. 540). Number 550 shows white dorsally bounded by a sharply marked-off narrow horizontal black band which is split anteriorly by an orange region into which it grades im- perceptibly. A broad horizontal white band then follows, sharply marked off from the black lying dorsally, and likewise from a narrow horizontal black line below, which grades ventrally through orange to 300 P. W. WHITING ivory. Number 540 shows a somewhat similar pattern (Fig. 14), but the median light band is distinctly pink, " cantaloup." It seems most reasonable to suppose that the sharply marked-off light regions of these eyes are genetically cantaloup, Oc, whether they appear white or pink, since eyes of cantaloup stock vary from white to red, becoming pro- gressively darker with age. The grading ivory or orange together with the black bounding bands are then ivory, o'C, modified by the dominant allelomorph, 0, in the cantaloup, Oc, regions. It is interesting to note that black color develops despite the absence of wild type, OC, tissue. Ocelli of No. 550 were colorless, which may have been either cantaloup or ivory, probably the former. In the case of No. 540, ocelli (Fig. 15) are mixed, the right ocellus as well as the right halves of the median and the left being colorless, " cantaloup." Brown pigment in the left halves of the median and left ocelli indicates modified ivory as in the left com- pound eye. Asymmetry in ocellar size and in pigmentation of ocellar region is to be ascribed to genetic difference in the tissues involved (Whiting, P. W., 1932). Females heterozygous for stumpy, st (legs) (Chromosome VI), produced fourteen males mosaic for stumpy and three that were modi- fied stumpy and thus suspected of being mosaics. Modifications of stumpy will be discussed below. Males mosaic for various other traits the genes for which have not yet been shown to be linked have been produced by heterozygous moth- ers. These include two for semilong, si, showing in antennae, wings, legs ; three for tapering, ta, showing in antennae ; one for yellow, Y (antennas) ; twenty-one for shot vein, sz1 (wing veins) ; three for club, cl (tarsi) ; three for cut, ct (wings) ; one for indented, in (wings) ; one for attenuated, at (antennae), one for twisted, tw, showing in antennae. Certain males from heterozygous mothers have been suspected of being mosaic for other genes but traits are too fluctuating to assert this with certainty. ORIGIN OF TRAITS IN GYNANDROMORPHS Most of the gynandromorphs found in Habrobracon are from pure stock or among the progeny of recessive females crossed with dominant males. The reason for this is that they come only from fertilized eggs and that very few offspring are bred from mated heterozygous females or from dominant females by recessive males. By far the greatest number of mated females that have been set are recessive and are used in connection with studies of ratios of biparental males. Thirty gynandromorphs have been found among progeny of orange- eyed defective, d (r4 wing vein) females crossed with type males. While sex of antennae can be readily determined, compound eyes vary MODIFICATION OF TRAITS IN MOSAICS 301 considerably in size so that there is no consistent sex difference. Color of eyes is, however, correlated with sex of antennae. Among the four- teen with both antennae male, eleven had orange eyes while three had eyes with some mixture of black and orange. Among the fifteen with one antenna male, the other female, eleven had eyes asymmetrical in color with the eye on the female side black, in which case the eye on the male side was orange or mixed ; or with the eye on the female side mixed, in which case the eye on the male side was orange. Of the other four cases with asymmetrical antennae, two had both eyes mixed while two had both eyes black. One wasp with both antennae female had both eyes black. Male regions of eyes then are orange, matro- clinous ; female regions show dominant black, patroclinous trait, and are presumably biparental. Ocelli are larger in male than in female and frequently in gynandro- morphs there is asymmetry, the female side showing the smaller ocellus surrounded by integument with characteristically less pigment. Among twenty-seven of these gynandromorphs sixteen had large male ocelli, orange in color, while five had small female ocelli. Six had ocelli asymmetrical in size and color, larger (male) and lighter on one side. In three cases the male ocelli were typically orange while in the other three they had more or less brown pigment. In five cases the female ocelli were described as " black " while in the other one the color was '' brown." Figures 17, 18, and 19 illustrate distribution of pigment in mixed eyes of gynandromorphs from orange females crossed with black-eyed males. Number 373 (Fig. 17) had small (female) black ocelli, and female antennae. It may be noted that the right eye is black (female) anteriorly. The left eye is orange (male). Number 375 (Fig. 19) is somewhat the reverse with male antennae, large (male) orange ocelli, while the left eye is orange (male) anteriorly. The right eye is orange (male). Number 397 (Fig. 18) shows a banded condition in the left eye with orange male ocelli, male antennae, and orange region anteriorly in the right eye. The grading margin may be noted between black and orange in these gynandromorphs in contrast to the clear-cut boundary between black and cantaloup or black and white. Wings of males are smaller than wings of females. The gene for defective rt vein, d, permits fluctuation in the character. Grade 4 de- notes the complete absence of the vein from a wing. Heterozygous females frequently show breaks classified as grades 1 or 2. Among the fourteen gynandromorphs with asymmetrical wings, the larger wings (female) had r4 classified as follows: type 10, dl — 3, d2— 1, while the smaller wings (male) had rt classified: d2 — 2, d3 — 5, d4 — 7. In each 302 P. W. WHITING individual the male wing showed the greater defect indicating maternal origin. We have already seen that mosaic males produced by mothers heterozygous for ivory usually show modification of this ivory color to orange. One gynandromorph (No. 322) from ivory female by black male had female abdomen and male head. Compound eyes were ivory showing no modification although female parts of body presumably were Oo'. Ocelli were male with a trace of brown pigment. Another (No. 481) from a similar cross with head male, thorax and abdomen mixed, likewise had ivory eyes but in this case the ocelli (male) con- tained red pigment. Another (No. 605) from a female heterozygous for ivory crossed with a black male had male head, female abdomen. Eyes were pale orange, but ocelli (male) were colorless. Another (No. 304) from ivory female by black male had head mixed, abdomen female. Left antenna was female, right male. Eyes were pronounced orange dorsally, black ventrally. Ocelli were male and of light color but show- ing some brown pigment. Another (No. 296) from a female hetero- zygous for ivory by an orange male, had one antenna male, the other female. Eyes were ivory, ocelli male and colorless. Other gynandro- morphs from crosses of females recessive for various traits by dominant males have shown male structures recessive (matroclinous) in the fol- lowing cases ; two for orange eyes, o ; three for cantaloup eyes, c ; three for long antennae, / ; one for type recessive to Minnesota yellow antennae, My. Gynandromorphs from similar crosses have shown female struc- tures, ocelli, dominant (patroclinous, presumably biparental), black in five cases of orange and in two cases of cantaloup. Only four cases have thus far been reported indicating that maternal contribution to male and to female parts is different. These are cases of crosses of heterozygous females by recessive males. One involved reduced, r (wings), showing one reduced and one type primary. Three involving orange had black and orange regions in the eyes. Ocelli were orange and female in one (No. 513) so we may suppose the black parts of the eyes were male, the orange female. In the second (No. 602) the larger ocellus (male) was orange, the smaller two (female) were black. In the third (No. 526) the smaller (female) lateral ocellus had no pigment. It was in a darker (male) region. The median and left ocelli contained orange pigment and although large (male) were in a somewhat lighter (female) region. MODIFICATION BY MOSAICISM OF FUSED AND STUMPY LEGS The mosaic "mutant" to fused (No. 507) mentioned above had antenme typical for fused. The left was slightly longer than the right. MODIFICATION OF TRAITS IX MOSAICS 303 Left wings were both type; right were both fused, the primary showing characteristic indentation at tip of radius vein, the secondary shorter than its mate on the left. The three left legs were type ; the right showed tarsi with segments typically fused. The propleuron was darker on the right side which was presumably composed of tissue bearing /. Since this specimen came from pure inbred stock it is likely that this pigmental difference is due to the factor / itself, in other words / is one of a number of genes causing darkening of integumental pigment. This mosaic proved sterile in observed mat- ings with three females. One of the mosaic males (No. 510) from a female heterozygous for fused had left antenna presumably type, but terminal segments were somewhat fused. Right antenna was fused and rather shorter than the average for this trait (Figs. 22 and 23). Primary wings were type but right secondary was short, probably fused. Left legs were type as was also the third right leg. First and second right legs were fused, the former showing much swelling in femur and tibia as is often the case in fused. Tarsus of this leg had joints completely fused (Fig. 24). Fusion of joints on the second right tarsus was incomplete (Fig. 25) so that it is possible to distinguish the five segments. The mosaic was highly fertile. Tests showed that it bred as wild type. The other mosaic for fused (No. 490) was produced by a female which was also heterozygous for semilong, si (antennae, wings, and legs). The wings of this specimen indicate that the two types of tissue present were wild type and semilong fused, for the left wings were both normal while the right showed the combined influence of semilong and fused. Eyes and ocelli were all cantaloup, but this gene does not presumably affect legs or antennae. Abdomen was normal male but external genitalia were missing (deficiency) except for a small clasper on the right. Left antenna was type but showing fusion of joints terminally. Right antenna was fused but rather long, perhaps due to the presence of the gene si (Figs. 28 and 29). The prothoracic right tarsus was type, while all the others were fused but showing more or less segmentation (Figs. 30—34). The meta- thoracic right was essentially similar to the left. A inetathoracic tarsus from stock fused is shown for comparison (Fig. 35). It is evident that these three mosaics illustrate considerable modifica- tion in the fused tarsi. No such difference occurs in fused stock or in semilong fused. Presumably the change in the legs at least is due to mosaicism. Whether the modification of the type ( ?) antennse is caused by mosaicism cannot be established from the two specimens. 304 P. W. WHITING Deficiency in genitalia of No. 490 may suggest correlated deficiency in antennae (Whiting, P. W., 1926). Distribution of type legs as associated with degree of modification of fused may be noted. In No. 507 there is no modification ; all the type are on one side, all the fused on the other. In No. 510 all are type on the left. On the right the third leg is type, the middle tarsus much modified fused, while the first is unmodified fused. In No. 490 the five fused tarsi show modification but the middle right which is next to the type first leg is extremely modified in contrast to its mate on the left. The left mate of the type leg shows definite but relatively little modification. These relationships suggest that modification of fused may be greater according to proximity to type on the same side but that there is little if any influence across the median plane. It may be noted that in No. 490 in which there are no type legs on the left, the left an- tenna is type and the degree of modification is less extreme on this side. Little of significance may be based on these three specimens, but in view of the relationships in modification of stumpy, it is regarded as worth while to call attention to these facts. In stumpy wasps the tarsal claws are very close to the tibia, the tarsal segments being reduced to minute chitinous vestiges (as in Fig. 39). Examination of over seventy-five specimens showed only six with a small tarsal segment on one or more legs. Legs showing this segment were prothoracic three, mesothoracic ten, and metathoracic one. There is therefore greater tendency for mesothoracic to possess a segment than for the other pairs. Moreover, an individual having a tarsal segment on one leg is likely to have it on others, for four of the six had segments on two legs while two had segments on three legs. Females heterozygous for stumpy have been bred in X-radiation experiments by Raymond J. Greb. Among the progeny there have been produced fourteen males that were mosaic for stumpy and three which, although not possessing any type legs, showed modification of stumpy and were mosaic for other traits. Modified stumpy legs of mosaics are shown in Figs. 36, 37, and 38. Of the entire group of seventeen, ten were produced by females X-rayed as adults, while seven were from non-rayed material. The modification of stumpy found in mosaics appears to have no relation to X-radiation of the mother. Among the fourteen mosaics for stumpy there were thirty-six type legs and forty-eight stumpy. There were only two instances of right and left pairs of type legs but nine of the fourteen mosaics had all type legs on one side, all stumpy on the other. This distribution of type A1ODIFICATION OF TRAITS IN MOSAICS 305 and stumpy tissue is considered due to the distribution of genetically diverse nuclei in cleavage. In order to have a quantitative measure of modification of stumpy, the legs of the seventeen mosaics (fourteen mosaic for stumpy) were roughly classified in grades from 0, denoting typical stumpy, to 6 with tarsi considerably over half normal length. The legs of the three mosaics that were not obviously mosaic for stumpy (with no type legs) were grade 0 — 10, grade 1 — 3, grade 2 — 4, grade 4 — 1. Stumpy legs that were opposite to type legs were grade 0 — 23, grade 1 — 3, grade 2 — 3, grade 4 — 3. Stumpy legs that were on the same side as type legs were grade 0 — 2, grade 2 — 1, grade 4 — 7, grade 6 — 2. It therefore appears that modification of stumpy is due to association with type on the same side and that there is very little if any influence across the median plane. If the grades of the various stumpy legs are totalled, the twenty prothoracic amount to 19 ; the twenty-four mesothoracic amount to 45 ; and the twenty-two metathoracic amount to 12. It therefore appears that mesothoracic are most subject to modification, prothoracic less, and metathoracic least. This order agrees with the six cases found in non- mosaics, but both ratio of legs modified and degree of modification is much higher in mosaics than in non-mosaics. A HAPLO-DIPLOID MALE MOSAIC A certain type of male mosaic not previously considered is a logical expectation from the fact that biparental (diploid) males appear in certain crosses and from the theory that a gynandromorph may be produced when one nucleus of a binucleate egg is fertilized. Shortly after the discovery of the mutation orange, there were found five patroclinous males from orange females crossed with black males, which, although showing no obvious mosaicism, bred as orange and were highly fertile (Whiting, P. W., 1921). The hypothesis was at that time advanced that the " black " parts were haploid and strictly androgenetic. In view of the abundant evidence now accumulated for egg binuclearity and for male diploidism as well as lack of evidence for androgenesis, it seems preferable to regard these patroclinous male mosaics as haplo-diplonts, developing from binucleate eggs in which one nucleus was fertilized and gave rise, not to female parts as in a gynandromorph, but to diploid male parts. Such a male, actually showing mosaicism, has recently been found and tested by Milton Franklin Stancati. An orange female (Stock 3) was crossed with a black-eyed male. This male was the original mutant 306 P. W. WHITING to indented, /// (wings). In addition to the regular black females — eighteen, and the orange males — nine, there were produced seven black- eyed biparental males, two of which occurred in the same vial (d) with the mosaic (No. 544). The mosaic was entirely male, with twenty- four segments in each antenna. Eyes were orange, matroclinous, except for black posterior region of the left. Ocelli were symmetrical in size and large (male). There was much brown pigment in left and median. The right had a very small amount of brown pigment (modified orange?). Body pig- ment was symmetrical. The mosaic was highly fertile, producing seventy daughters when mated with three females which produced forty-three sons (female ratio 62 per cent). By the two orange mates there were produced fifty- four orange daughters, while by a black-eyed heterozygous sister there were produced ten black and six orange daughters including no in- dented although three indented males appeared and one non-indented, proving this sister heterozygous as expected. There were also no in- dented among numerous descendants of the mosaic by the Stock 3 fe- males. The mosaic therefore bred like its maternal stock (No. 3). Since eight of the twenty-six fertilized eggs developing in the fra- ternity of No. 544 produced males, a binucleate egg would here have a very good chance of producing a haplo-diploid male mosaic. In view of the fact that eye color alone has generally been used as the criterion of male biparentalism. it is highly probable that haplo-diploid male mosaics have been missed on account of failure of obvious mosaicism, being classified either as normal haplonts or as biparental males. On the theory advanced, their frequency may bear the same relation to gynandromorphs as the frequency of biparental males bears to females, —haplo-diploid male mosaics /gynandromorphs = = biparental males/fe- males. A CASE OF FOUR- STRAND CROSSING-OVER Mosaic males from binucleate eggs would be expected to possess in their two types of tissue two of the various possible combinations of those genes for which their mothers were heterozygous. Only in those cases, however, in which the two or more pairs of genes affect the same structure can we be sure of what combinations are present. Thus, if eyes are mosaic for type (black) and cantaloup and wings are mosaic for type and long, there is no way of telling which wing is genetically cantaloup or which eye is genetically long. Virgin females heterozygous for long and for reduced have produced mosaic males having wings long on one side, reduced on the other, or having wings wild type on one side. MODIFICATION OF TRAITS IN MOSAICS 307 long reduced on the other. Gynandromorphs have likewise come from reciprocal crosses of orange and cantaloup having eyes type (black) if female, or showing the maternal color if male. One instance of recombination of genes affecting the wings may be noted as it involves four-strand crossing-over in ovogenesis. In con- nection with her experiment on the production of impaternate females, Kathryn A. Gilinore bred progeny from a virgin female heterozygous for four factors in Chromosome II, Mv/c.l.n. Minnesota-yellow, My, an irregular dominant, affects basal segments of antennae. It lies well to the left, 20 units ±, of the factor for cantaloup eyes, c. Long, /, lying 12 units ± to the right of cantaloup, shortens wings distally but lengthens antennal segments. Narrow, ;/, about three units to right of long, makes wings narrow. Among the progeny (male) segregating these four factors as expected there occurred a male (Xo. 552) with Minnesota-yellow long antennae and cantaloup eyes and ocelli. Left primary wing was narrow (Fig. 26) ; right, long narrow (Fig. 27) with defective, r4, venation. Character of antennae indicates a crossover somewhere between My and /. Origin of defective is uncertain and need not concern us. Left primary wing indicates that a crossover occurred between / and //. The right primary, however, shows a non- crossover combination. The facts indicate that we are concerned with a two-crossover ocicyte, two strands crossing over between Mv and c, the other two between / and n. There is no evidence for double crossing-over. Crossing-over between two strands in one region does not then prevent crossing-over between the other two in another region. For My and n we have pre- reduction ; for c and / post-reduction. The two ootid nuclei involved in parthenogenetic cleavage were My.c.l.ii and My.C.L.n, while iny.my.- Cc.Ll.Nn went out in the first polar body. SUMMARY Mosaic males from heterozygous mothers have shown clear-cut mosa- icism for the recessive eye colors cantaloup and white. Orange, how- ever, shows intergradation with black of wild type and ivory shows complicated types of modification and intergradation. A similar condi- tion obtains in the case of gynandromorphs. Males have shown clear- cut mosaicism for sixteen other traits. As regards either fused or stumpy legs, however, there is much modification, with evidence that influence is from wild type tissue on the same side of the body, but not on the opposite side. Further evidence is presented indicating that gynandromorphs show maternal traits in male parts of body while female parts are biparental. Maternal contribution to male and female 308 P. W. WHITING parts may be different. A male mosaic has been found which is best explained as a haplo-diplont, being in part biparental. A mosaic male from a mother heterozygous for four linked genes indicates such a combination of traits that two crossovers must have taken place in the tetrad, one between two strands, one between the other two, in the egg from which this male developed. BIBLIOGRAPHY WHITING, P. W., 1921. Studies on the Parasitic Wasp, Habrobracon brevicornis (Wesmael). Genetics of an orange-eyed mutation and the production of mosaic males from fertilized eggs. Biol. Bull., 41: 42. WHITING, P. W., 1922. Genetic Mosaics and Ontogenetic Abnormalities in the Parasitic Wasp, Habrobracon. Anat. Rcc., 23: 94. WHITING, P. W., 1926. Influence of Age of Mother on Appearance of an Heredi- tary Variation in Habrobracon. Biol. Bull., 51: 371. WHITING, P. W., 1932. Asymmetry of Wild-Type Traits in Habrobracon. Jour. E.i- per. ZooL, 62: 259. EXPLANATION OF PLATE I Figs. 1, 2, 3, 5, 7, 9, 10, 11, 12, 13, 14, 17, 18, 19. < 27. In compound eyes solid black represents black, stippling indicates reddish or orange color, white un- stippled regions are white. Fig. 16. < 17. Figs. 4, 6, 8, 15, 20, 21. X 64. Figs. 22, 23, 24, 25, 28-39. < 27. Figs. 26, 27. < 11. Figs. 1-15 show eyes and ocelli of males mosaic for various traits as follows : Fig. 1. Type and ivory (No. 383). Fig. 2. Orange and ivory (No. 341). Figs. 3 and 4. Type and cantaloup (No. 378). Figs. 5 and 6. Type and cantaloup (No. 465). Figs. 7 and 8. Type and cantaloup (No. 351). Fig. 9. Type and cantaloup (No. 593). Figs. 10 and 11. Type and white (No. 601). Fig. 12. Type and white (No. 600). Fig. 13. Ivory and cantaloup (No. 550). Figs. 14 and 15. Ivory and cantaloup (No. 540). Fig. 16. Mesonotum of male mosaic for broad, br (No. 465). Figs. 17 (No. 373), 18 (No. 397), 19 (No. 375). Heads of gynandromorphs with male regions of eyes orange, female regions black. Figs. 20, 21. Left and right eyes, respectively, of male mosaic for glass and for orange (No. 549). Stippling indicates brownish pigment. The glass regions are in general orange. FIGS. 22 (left antenna), 23 (right antenna), 24 (right prothoracic leg), 25 (right mesothoracic tarsus) from male mosaic for fused (No. 510). FIGS. 26 (left, narrow) and 27 (right, long narrow) wings of mosaic male (No. 552). FIGS. 28 (left antenna), 29 (right antenna), 30 (left prothoracic tarsus), 31 (right prothoracic tarsus), 32 (left mesothoracic tarsus), 33 (right mesothoracic tarsus), 34 (left metathoracic tarsus) from male mosaic for fused (No. 490). FIG. 35. Metathoracic tarsus of typical fused male. FIG. 36. Left mesothoracic tarsus of male mosaic for stumpy (No. 495). FIGS. 37, 38, 39. Two views of left mesothoracic tarsus and one of right meta- thoracic tarsus of male mosaic for stumpy (No. 497). MODIFICATION OF TRAITS IN MOSAICS 309 39 PLATE I 20 TEMPERATURE AND LIGHT AS FACTORS INFLUENCING THE RATE OF SWIMMING OF LARV.^ OF THE MUSSEL CRAB, PINNOTHERES MACULATUS SAY JOHN H. WELSH (From the Zoological Laboratory, Harvard University, and the Woods Hole Oceanographic Institution) Significant and numerous studies have been made in the past on the phototropism of plants and animals, and the quantitative aspects of the effect of light on photosensory systems have been extensively studied by Hecht and others ; however, little is known concerning photokinesis and the effect of temperature on free-moving, light-sensitive organisms. Some investigators have even denied an effect of light on the behavior of such forms other than on orientation. Davenport and Cannon (1897) found an apparent difference in rate of swimming of Daphnia in full light as compared with swimming under one-fourth this intensity, but they concluded that this was probably due not to a change in velocity but to more rapid and accurate orientation at the higher in- tensity. Yerkes (1900) substantiated the findings of Davenport and Cannon on Daphnia and found in addition a slight effect of intensity on rate of swimming of Cypris, but he came to similar conclusions, namely, that the apparent change in rate was due primarily to changes in accuracy and rapidity of orientation. In neither of these investigations was a very wide range of intensities used. Their conclusions agree with those of the majority of observers before and since, with few exceptions. Moore and Cole (1921) found that the rate of locomotion of Popillia japonica during upward geotropic progression was influenced by light and that the rate of movement was a function of the light in- tensity. Cole (1922a) obtained similar results regarding the upward creeping of Drosophila, and (1922&) found a distinct effect of light on the rate of creeping of Limulus. Mast (1923) expresses doubt re- garding the validity of Moore and Cole's (1921) results on Popillia for he states that they did not exclude the time required for the insects to orient and get under way. Mast also doubts the value of Loeb's (1890) observations on rate of movement of aphids as a function of light intensity, for he claims that temperature was not eliminated as a factor. Mast and Cover (1922) studied the effect of intensity of light and rate of locomotion of Phacits and Englena, and although with 310 EFFECT OF LIGHT ON SWIMMING RATE 311 Euglcna they found what may he a significant increase in rate at high intensities, they nevertheless concluded that light intensities sufficient to cause rapid and accurate orientation need not have any additional appreciahle effect on rate. The lack of agreement regarding the photokinetic effect of light, and the fact that the problem has considerable bearing on our proper understanding of movements of plankton organisms, led to an attempt to gain additional information. Crustacean larvae form one of the representative groups of the animal plankton of the sea and were found to be most satisfactory for the work as planned. The work was carried out during the summer of 1931 at the \Yoods Hole Oceanographic Institution. It was made possible largely through the kindness of Dr. H. B. Bigelow, Director of the Institution, and the author wishes to acknowledge his appreciation for the excellent facilities and equipment placed at his disposal. The author is also indebted to Professor W. J. Crozier for suggestions in the preparation of the paper. MATERIAL AND METHODS For the study of photokinesis a suitable animal should possess cer- tain characteristics which have been lacking in part in many of the animals previously studied. They should be positive or negative to light ; they should orient accurately and rapidly ; they should move in a straight line ; they should preferably be aquatic organisms, in order to make it possible easily to control the temperature. In addition to these requirements any satisfactory experimental animal must be obtainable in large numbers over a considerable period of time, or must be easily reared or kept in the laboratory. For satisfying these requirements it is difficult to conceive of a more suitable form than the young larvae of Pinnotheres inaculatus Say, the mussel crab. The adults are found living as parasites in the mantle cavity of Mytiliis cdulis, the edible mussel. Sixty-five per cent of the mussels collected from a bed near Grassy Island, Woods Hole, were infested with these crabs so the adults were easily obtained. A con- siderable number of females carrying eggs were found at all times during July and August. These are easily kept in the laboratory in bowls of sea water, and one can have one or more batches of larvae hatching daily, each batch containing several hundred individuals. The larvae are distinctly positive to light. They orient with head away from the light and by means of forward strokes of the swimming appendages move backward toward the light. This is similar to the orientation of the young larva? of the lobster as observed by Hadley 312 JOHN H. WELSH (1908), and of the young larvae of Palaemonetes as described by Lyon (1906). At all light intensities used there was extremely rapid orienta- tion consuming only a fraction of a second, and the larvae took a course toward the light which was a straight line in most cases. At a given light intensity and temperature the velocity of movement was quite constant for individuals of a given age, but after two or three days the rate of swimming decreased and at the age of four to five days the larvae had a tendency to be temporarily negative ; this necessitated using larvae of known age. As a careful control of either temperature or light intensity is neces- sary when studying the effect of the other factor, all experiments were performed in a dark room ; by means of a water-bath, the temperature could be accurately controlled, and with proper methods light of the desired intensity obtained. The water-bath consisted of a 30-gallon insulated tank with 4x6 inch plate glass windows set in opposite sides and near one end. Tem- peratures below room temperature were obtained by means of a cooling unit similar to one described by Stier (1931), with a mercury thermo- regulator operating a heating unit of 100-watt capacity. Rate as a function of light intensity was studied for the most part at temperatures slightly above room temperature which obviated the necessity of using the cooling device. The water in the tank was stirred by means of a motor-driven agitator, and was changed at frequent intervals in order to avoid loss of light by suspended particles which tended to accumulate in the tank. The light-source was a 6-volt, 18-ampere, ribbon filament lamp, shielded by means of a double housing in order to prevent leakage of light. The light passed through lenses which kept the rays practically parallel. The intensity was controlled by means of W ratten neutral tint filters, which, used singly or in combination, transmitted the follow- ing percentages of the original light: 50, 25, 10, 5, 2.5, 1, 0.5, 0.1, 0.05 per cent. The beam of light passed through a series of screens with apertures of the proper size, and through ground glass for diffusion. A second light for attracting the larvae to the opposite end of the trough consisted of a Spencer lamp with a 150- watt bulb and ground glass. The larvae were placed in filtered sea water in a trough of plate glass with inside dimensions 29 x 4 x 4 cm. This trough was covered with a glass plate and submerged to within a half centimeter of the top in the water of the bath. Here it was supported on hangers so that it was always at a given distance from the light source and so that the beam of light just covered the inside section of the trough and, the rays being parallel, reflection of light from the glass sides was negligible. EFFECT OF LIGHT ON SWIMMING RATE 313 Measurements of the light intensity within the trough were made by means of a Macbeth Illuminometer, by putting a small test plate in the water of the trough. The intensity of light at the end of the trough nearest the ribbon filament lamp (the variable source) when no filters were used was found to be 93 meter candles. The 150- watt lamp at the distance used gave an intensity of 68 meter candles at the end near- est this lamp. By means of the neutral filters it was possible, without changing the distance of the lamp from the tank, to obtain the following intensities of light at the end of the trough nearest the ribbon filament lamp : 46.5, 23.3, 9.3, 4.7, 2.3, 0.93, 0.47, 0.093, and 0.047 meter candles. The last intensity was the lowest that it was practical to use, for below this it was impossible to see the larva? distinctly and even at this intensity it was necessary for the observer to be adapted to complete darkness before making each observation. There was, of course, considerable absorp- tion of light by the sea water of the trough, and the intensity at the far end in each instance was considerably lower than the intensity given for the near end, and this varied throughout the trough; however, the total light reaching the larvae as they swam from the far end to the near end of the trough was proportional to the intensity at the near end, and varied as this was varied. In preliminary experiments observations were made on the rate of swimming of individual larva? as compared with the rates of the first or middle member of a small swarm, and the variations were no greater in the second instance than in the first. As it was more often possible to complete a series of data if several individuals were used instead of one, a small. swarm of 10-25 animals was used in most of the experi- ments. The writer realizes the possibility of greater variations in the results when swarms are studied but in this particular instance the be- havior of several larva? selected from a given batch showed as great uniformity as did single individuals, at least when the fastest and slow- est members had been discarded. In a particular experiment a group of larvae were selected of the proper age and placed in the trough and adapted to the temperature of the water-bath for at least a half hour. They were then attracted by the 150- watt light to the end of the trough away from the variable light source, and their swimming movements would keep them in close contact with the glass in their endeavors to continue their course toward this light. This light would then be cut off and at the same time the light ?t the opposite end turned on and a stop watch started. The time necessary for the fastest individual or group of individuals to traverse the 29 cm. was then taken. At a given light intensity or temperature 314 JOHN H. WELSH either five or ten trials were made and the results averaged. In this way it was possible to determine the rate of swimming of the larvae both as a function of temperature and light intensity. As stated above the • extremely rapid orientation of the larvae and the apparent absence of a latent period obviated the necessity of considering these factors and made the measurements easier and more accurate than would have been true with many free-swimming forms. TEMPERATURE AND VELOCITY OF LARV/E AT A CONSTANT LIGHT INTENSITY Preliminary observations indicated a decided effect of temperature on the rate of swimming of the larvae, and although only one series of measurements on individuals of the same age was made over a wide TABLE I Effect of temperature on rate of swimming of larvcc at a constant light intensity Temperature Time for swimming 29 cm. (Averages of 10 readings) P.E. of time Velocity 1 time for 29 cm. °C. seconds 27.0 12.38 ±0.100 0.0806 25.9 13.09 ± 0.090 0.0763 24.8 14.53 ± 0.046 0.0690 23.5 15.50 ± 0.057 0.0645 22.0 17.70 ± 0.099 0.0565 20.0 20.70 ±0.127 0.0483 18.5 24.79 ±0.175 0.0403 17.6 25.64 ±0.117 0.0391 16.5 27.72 ±0.139 0.0361 15.2 32.30 ±0.152 0.0310 14.2 36.11 ± 0.149 0.0277 13.4 39.05 ± 0.302 0.0256 range of temperatures, the results from this were quite significant. The range of temperatures used was from 27.0° C. to 13.4° C. The light toward which the larvae swam was kept at a constant intensity of 93 meter candles. The results are shown in Table I. The times given for swimming 29 cm. are, at each temperature, averages for ten read- ings. Between different temperatures one hour was allowed for adapta- tion to the new temperature. From these results one can determine the time for swimming a meter at different temperatures. At 27.0° this was 43.4 seconds, under the conditions of the experiment ; at 18.5° the same larvae required 1 minute 27 seconds for travelling a meter; and at 13.4°, 2 minutes 16.5 seconds. EFFECT OF LIGHT ON SWIMMING RATE 315 Temperature changes in the sea are slow and comparatively small yet this factor must be of some importance in determining the rate of move- ment of plankton organisms, particularly when near the surface. The probable errors of the times are shown in Table I and it may be seen that these are not large. There is a rather definite relationship be- tween the average time for swimming 29 cm. and its probable error, and it is interesting to note that they are both affected by temperature in much the same way ; as the mean time increases, the P.E. increases in proportion. The significance of this has been pointed out by Crozier (1929) and Navez (1930). A plot of velocity against temperature centigrade (Fig. 1) indicates that the increase in velocity with increasing temperature is not on a simple smooth curve ; it is evident that one curve does not fit the results. There is a break near 18.5° and near this point occurs the one velocity measurement which does not conform fairly well with the rest. It will ooe own 006 003 004 003 16 tO 12 Temperature *C 24 26 FIG. 1. Data of Table I plotted as velocity (reciprocal of time for 29 cm.) against temperature centigrade. Two curves are shown which intersect between 18° and 19°. This break is more evident when the same data is plotted as in Fig. 2. be seen in Table I that the mean time from which this velocity was obtained has a high P.E. compared with those above and below. When the logarithm of the velocity is plotted against the reciprocal of the absolute temperature (Fig. 2) it is more apparent that between 13.4° and 27.0° two lines must be drawn to fit the data ; these lines are straight and intersect at about 18.5°, thus indicating that the Arrlienius equation [v- M/_l l\ ^ = €2Vr T,) where Kl is the velocity at 7\° Abs., and K.. the KI velocity at T2° Abs.; £==2.718; ^ is a constant over a certain tempera- 316 JOHN H. WELSH ture range and designated by Crozier as a " temperature characteristic "] holds for the rate of swimming of these larvae as a function of tempera- ture. The values of /A as calculated are 16,900 below 18.5° and 12,800 above 18.5°. It is not necessary to go into the significance of these values and of the break, or critical temperature, for this has been done by Crozier (1924) and others, for many cases which are fundamentally similar ; it is sufficient to note that corresponding values for //, have been frequently encountered in studies of the influence of temperature on rate of many biological processes. 10 2.8 ez.v 00 26 25 - 17,800 ^•16.900 330 335 340 Ot«io») 345 350 FIG. 2. Same data as in Fig. 1, plotted as the logarithm of velocity against the reciprocal of absolute temperature X 105. That two lines must be drawn to fit the data is evident. Few studies of the effect of temperature on the rate of locomotion of organisms have been made, the only other on a free-swimming form with which the writer is familiar is that of Glaser (1924) for Paramecium. He found that the Arrhenius equation could be applied, and secured a value for /x of 16,000 below 16° and of 8,000 above 16°. It is realized that changes in density and viscosity of the sea water occur with changes in temperature and these factors enter in, and intro- duce slight errors in the measurements which are difficult to eliminate. To what extent they are significant remains to be seen. The effect of temperature on the velocity of swimming at different light intensities will be discussed in a later section. LIGHT INTENSITY AND THE RATE OF SWIMMING OF LARV/E AT A CONSTANT TEMPERATURE The importance of temperature as a factor influencing the rate of swimming of Pinnotheres larva? has been indicated in the previous sec- EFFECT OF LIGHT ON SWIMMING RATE 317 tion. During the work on the photokinetic effect of light the tempera- ture was carefully controlled and kept constant, for a given series, within ±0.1°. In addition to temperature, age was found to be an important factor •influencing the rate of locomotion, and although this was not carefully investigated the age was taken into consideration in the later determina- tions of the effect of intensity of light. Because either age or tempera- lure varied between separate series of experiments in some of the earlier work, it was impossible to average or compare much of the data, al- though the results were fundamentally similar. As an example of the effect of age on rate of movement it is necessary to cite only one instance. On August 23 at 2 :00 P.M. a certain larva swam the 29 cm. in 14.2 sec- onds, in a light intensity of 93 meter candles; on August 24 at 10:00 A.M. the same individual required 15.2 seconds to swim the same dis- tance. In each case the times as given are averages for five readings. It was also found that over a range of intensities, larvae 30 hours old were less sensitive to the light than larvae 15 hours old. It is obvious that age would not be of such importance if one were dealing with adult animals, but in the case of crustacean larvae, as is well known, a few days makes a great difference in the responses of the animals to light (Hadley, 1908, and others). Early in the course of the work on light intensity it was found that changes in rate of swimming were obtained over a comparatively small range of intensities. The maximum velocity, at temperatures near 23°, was reached at approximately 25 meter candles. This suggests that in much of the previous work on the photokinetic effect of light, intensities above the minimum necessary for eliciting a maximum response may have been used, and the conclusions that light intensity has no effect on rate of locomotion are perhaps unjustified in many instances. It is logical to assume that aquatic organisms are sensitive to a lower range of intensities of light than are land organisms, for they live constantly at reduced intensities. It should also be pointed out that the intensity of light to which the animals are previously adapted affects to a certain extent the rate of movement in subsequent intensities. Several trials were made to deter- mine the effect of dark adaptation on rate of movement in the light, with the expectation that for a few trials the animals would swim more rap- idly, due perhaps to an accumulation of a photosensitive material in the light receptors. In every instance the first few trips, after dark adapta- tion of several hours, occupied more time than subsequent trips. This phenomenon is perhaps similar to that noted by Davenport and Cannon (1897), who found that Daphnia responded more quickly and accurately 318 JOHN H. WELSH to the light after having made several trips in it. Hecht (1925) also found in dona that after dark adaptation of several hours the first re- action time to a given intensity of illumination was definitely longer than those which followed, and which remained constant for long pe- riods. These authors offered no explanation and as yet there seems to be no satisfactory reason why this should be true. The following experiment seems somewhat contradictory in view of these results on dark adaptation. Larva? were adapted to a series of light intensities ranging from 0.093 to 93.0 meter candles and the time for swimming toward a light of 68 meter candles subsequently obtained. The results are shown in Table II. After adaptation to light of 0.093 meter candles 13.4 seconds were consumed in swimming 29 cm., in an illumination of 68 meter candles. This time increased as the intensity of the adapting TABLE II Effect of intensity of adapting light on time for sivimming 29 cm. toward light of constant intensity. Age of larvae 30 hours. Temperature 25.4° C. Intensity of adapting light Intensity of attracting light Time for swimming 29 cm. toward light of 68 meter candles meter candles meter candles seconds 0.093 68 13.4 0.93 68 13.3 4.7 68 13.9 9.3 68 14.2 23.3 68 14.2 46.5 68 14.3 93.0 68 14.7 light increased, until after adaptation to 93.0 meter candles 14.7 seconds were required for swimming the same distance. This indicates a dis- tinct effect of the adapting illumination, and in the experiments to follow the larvae were, in every case, adapted to light of a constant intensity before each trial. From several series of experiments to determine the velocity of swimming in intensities ranging from 0.093 or 0.47 meter candles to 93.0 meter candles very uniform results were obtained. The maximum rate of swimming was reached in light of about 25 meter candles, and although this rate varied somewhat the variation was due to differences in age of the larvae, temperature of the water, or in some instances per- haps to slight differences between given lots of larvae or individual larvae. EFFECT OF LIGHT ON SWIMMING RATE 319 Table III gives one such -series of data on larvae 20 hours old, at a temperature of 24.5°. In this series readings were begun at the higher intensities and the intensities of the variable light source decreased by definite amounts between sets of readings. Progression from low to high intensities yields essentially the same results. In this particular series the faster individuals in the swarm travelled the 29 cm. in 16.6 seconds, at an intensity of 93.0 meter candles. At 46.5 meter candles there was no significant change. At 23.3 meter candles the time for swimming 29 cm. had increased to 17.5 seconds and from this intensity down to 0.47 meter candles, the lowest intensity tried, there was a grad- ual increase in time ; at the lowest intensity the time being almost exactly twice as great as at the higher intensities. A plot of this data as reciprocal of the time against the light intensity TABLE III Effect of light intensity on rate of sti'tinming of larva: at a constant tempera- ture (24.5° ± 0.1). Series 8.6. Age of larva? 20 hours. Intensity Time for 29 cm. (Averages of S readings) P.E. of time Velocity 1 time for 29 cm. meter candles seconds 0.47 33.9 ± 0.91 0.0295 0.93 31.8 ±0.80 0.0315 2.3 27.2 ± 0.70 0.0368 4.7 23.5 ±0.57 0.0426 9.3 21.5 ±0.62 0.0465 23.3 17.5 ±0.31 0.0571 46.5 16.7 ±0.41 0.0599 93.0 16.6 ±0.22 0.0602 yields a smooth curve as seen in Fig. 3. One point does not fall well on the curve but, as may be seen in Table III, the probable error of the time at this intensity (9.3 meter candles) is high in comparison with those above and below. The velocity at 93.0 meter candles is not shown on the graph as it is practically the same as that at 46.5 meter candles. It should be noted that in no instance were rates of swimming obtained in very low intensities or in total absence of light, for obvious reasons. It should also be pointed out that swimming movements of crustacean larvae do not stop even in absence of light ; instead the larvae remain at or near the surface, kept there by constant but random movements. One might expect, if the effect of light on rate of locomotion could be determined in the same way as the effect on orientation of the larvae, 320 JOHN H. WELSH to find them obeying the Bunsen-Roscoe Law as did Loeb and Northrop (1917) in their investigation of the orientation of Balanus larvae to a two-point source of light. If this law held for velocity of locomotion, the relationship between velocity and light intensity would be a linear one. That this is not true is quite evident, and the reason is perhaps obvious. Loeb and Northrop were concerned with the degree of turn- ing of the path toward the stronger of two lights and this in no way de- pended on previous velocity and to only a very slight extent on water resistance. The results obtained (Fig. 3) more nearly resemble those obtained 0.06 0.05 0.04 0.03 10 20 30 Intensity — meter candles 40 50 FIG. 3. Data of Table III, Series 8.6, plotted as velocity against intensity (meter candles). See text. by Moore and Cole (1921) for rate of movement of Poplllia japonica as related to light intensity, and later by Cole (1922) for Drosophlla. In both of these instances they found apparent conformity with the Weber-Fechner Rule, as Henri (1912) had claimed for the reactions of Cyclops to ultra-violet light, and Patten (1915) when using a graded series of absolute intensities of opposed lights in studying orientation of blowfly larvae. Although such conclusions are to a certain extent un- warranted by the fact that the Weber-Fechner Rule does not hold for intensity discrimination in certain forms, as pointed out by Hecht (1924, 1928), yet the approximate linear relation of response plotted against the logarithm of intensity is sometimes useful in analyzing such data. Crozier (1928) in discussing the case of Liina.r, where the EFFECT OF LIGHT ON SWIMMING RATE 321 amount of turning- per unit length of path is directly proportional to the logarithm of the light intensity, emphasizes the fallacy of consider- ing this an obeyance to Weber's Rule and yet suggests that the same empirical treatment is useful, where other perhaps more significant treatments are impractical. If the data given in Table III are plotted as velocity against the logarithm of the intensity, the relationship is found to be far from linear. This indicates that the velocity is not related directly to the logarithm of the intensity of illumination. In the work of Moore and Cole on the Japanese beetle they were dealing with an organism which, under a ruby light or in the dark, seldom showed any movement, but which was aroused to activity by illumination from any direction. As has been pointed out above, Pinnotheres larvae are constantly moving even in total absence of light, and this initial velocity must bear some relation to later velocities produced by illumination. 2.8 >>2.7 -M 1 > - _o 2.5 2.4 ---o o — J_ 1.0 0.0 1.0 log I — meter candles 2.0 FIG. 4. Complete data of Series 8.6 plotted as logarithm of velocity against logarithm of intensity. The graph is essentially rectilinear until the maximum velocity of swimming is approached. If we assume that the velocity of movement (V) is so related to the light intensity (/) that any increase in velocity (AF), produced by a small increase in intensity (A/), is a function of the velocity and also inversely proportional to the intensity, we obtain the following expres- sion: AF V A/ 7 Upon integration this yields the following equation : where k is the constant for the slope of the line, at any given temperature, and C is an integration constant. This expression indicates the way in 322 JOHN H. WELSH which the slope of the curve of velocity plotted against intensity de- pends, at any point, on both the previous velocity and light intensity. It may be checked by altering another variable such as temperature as is shown in the following section. If we apply this formulation to the data in Table III and plot the logarithm of velocity against the logarithm of intensity, we obtain a straight line over most of the range of intensi- ties as shown in Fig. 4. Two velocities are shown at intensities of 46.5 and 93.0 meter candles which were obtained at or near the maximum velocity possible at this temperature, and which, of course, do not fall on the curve as drawn, but between 0.47 and 23.3 meter candles the straight line fit is good. The theoretical significance of such an empirical treatment need not concern us. It is sufficient that we have a convenient method for com- TABLE IV Effect of light intensity on rate of szt'imining of larva at different tempera- tures. Age of larvoe 15 hours. INTENSITY SERIES 8.21 A TEMP. 13.4° ± 0.1 ° SERIES 8.21 B TEMP. 18.0° ± 0.1 ° SERIES 8.21 C TEMP. 27.1 ° ± 0.1 - Time for 29 cm. P.E. of time Time for 29 cm. P.E. of time Time for 29 cm. P.E. of time meter candles seconds seconds seconds 0.093 22.8 ±0.28 0.47 39.2 ±0.62 18.3 ±0.28 0.93 86.7 ± 2.16 35.9 ± 1.04 17.1 ±0.06 2.3 64.7 ± 1.06 31.0 ± 0.39 15.2 ±0.21 4.7 56.8 ±0.89 27.0 ±0.35 13.7 ±0.09 9.3 47.7 ±0.53 23.9 ±0.32 13.5 ±0.11 23.3 39.7 ±0.30 22.4 ±0.33 13.3 ±0.17 paring data obtained by varying the several factors such as temperature, light intensity, and age, which so evidently influence the rate of swim- ming. THE EFFECT OF LIGHT INTENSITY ON RATE OF SWIMMING OF LARVAE AT DIFFERENT TEMPERATURES In the preceding sections we have shown the effect of temperature on the velocity of Pinnotheres larvae at a constant light intensity, and the effect of light intensity at a constant temperature. Now it might be of value to compare light intensity curves from larvae of the same age obtained at different temperatures. Table IV gives three series of data on larvae 15 hours old, at tern- EFFECT OF LIGHT ON SWIMMING RATE 323 peratures of 13.4°, 18.0° and 27.1° C. The complete range of light intensities available was not used at each temperature, for the rate of swimming was extremely slow and rather irregular below an intensity of 0.93 meter candles, at the lowest of the three temperatures ; however, sufficient determinations were made for adequate comparisons. The curves for velocity plotted against light intensity are shown in Fig. 5. It may be seen that with increasing temperature, there is dis- tinct displacement of the curves upward, and an earlier arrival at the maximum rate of swimming as the temperature increases. Also the slope of the curves changes with the temperature. At the lowest tem- perature the curve is much flatter than at the higher temperatures. While it was impossible to determine the effect of temperature upon 0.07 0.06 0.05 o 2 0.03 0.02 0.01 0.00 • 154- 10 15 Intensity — meter candles 20 25 FIG. 5. Plot of data of Table IV as velocity against intensity. Open circles are Series 8.21 A ; half-closed circles, Series 8.21 B ; closed circles, Series 8.21 C. Temperatures as indicated. the velocity of swimming in absence of light, it will be seen that there is a distinct effect on rate of locomotion in the dark, and if the curves were begun at the zero point on the abscissa, they would intercept the ordinate at varying levels above zero, depending upon the temperature. This is due to changes in rate of general activity, and to changes in the viscosity of the sea water. While no attempt has been made to correct for viscosity changes due to changes in temperature, these are of con- siderable importance. It was shown by Ostwald (1903, a and b} that in comparison with water, which might be considered to have a viscosity 324 JOHN H. WELSH of 100 at 0° C., sea water of 30 per cent salinity has a viscosity of 102, and this is reduced to 52 at 25° C. Thus sea water at 25° C. is approxi- mately half as viscous as that at 0° C., and the same body would sink twice as fast at 25° C. as at 0° C. Viscosity changes with changes in temperature would therefore account for an appreciable part of the change in speed of swimming. If we plot the data of Table IV as the logarithm of the velocity against the logarithm of the light intensity, we obtain as before essen- tially rectilinear graphs (Fig. 6), which vary in slope and in position relative to the abscissa. As in Series 8.6 (Fig. 4), the points repre- senting velocities at or near the maximum do not fall on the lines, but over a definite range of intensities the straight line fit is good. The relative displacement of the graphs, and therefore the value of C in the 29 2.8 2.7 £2.6 IP M 2.4 ^o ~ 2.3 2.2 2.1 2.0 1.0 0.0 1.0 log I — meter candles 2.0 FIG. 6. Data of Table IV plotted as logarithm of velocity against logarithm of intensity. Symbols representing Series as in Fig. 5. See text. expression log V - k log I - -C, is seen to change considerably with a change in temperature. In addition, the slope of the lines is also seen to change. At 13.4°, k has a relative value of 0.225; at 18.0° k = 0.175; at 27.1° £==0.125. At a given temperature the effect of in- creasing the intensity of illumination, within certain limits, is to increase the velocity. The changes in slope of the lines in Fig. 6 indicate the effect of temperature on the relation between velocity and intensity. At 27.1° the slope is less than at the lower temperatures, and the maximum rate of swimming is attained at a lower light intensity than at 18.0°. The expression ±V V A/"/ implies that if V be increased by operation of a variable independent of EFFECT OF LIGHT ON SWIMMING RATE 325 /. the effect of increasing 7 must be correspondingly less — which is the fact. We have seen that temperature and light play an important part in determining the rate of locomotion of a crustacean larva. Other fac- tors such as age and changes in viscosity of the surrounding medium must also he taken into consideration. SUMMARY 1. Larva? of Pinnotheres maculatus Say are shown to be satisfactory animals lor the study of photokinesis. The velocity of swimming is found to be greatly influenced by temperature and light intensity. Age, although not carefully investigated at present, is also an important con- tributing factor in determining the rate of locomotion. 2. A series of measurements of the effect of temperature on the velocity of swimming, at a constant light intensity, showed the applica- bility of the Arrhenius equation, and yielded values of ^ of 12,800 above 18.5^° and 16,900 below 18.5° C. 3. The larvae are found to be sensitive to only a small range of light intensities. At temperatures between 20-25° C., the maximum possible velocity of swimming is attained at intensities between 10 to 25 meter candles. When velocity is plotted against light intensity a smooth curve is obtained. The same data when treated empirically according to the equation log V = k log I - - C, yields essentially rectilinear graphs which are more satisfactory for a comparison of such data. 4. When series of measurements are made to determine the effect of light at different constant temperatures it is found that, besides a marked effect on general activity, there is a change in the relationship of velocity to intensity; the slopes of the curves change, and the maxi- mum possible velocity of swimming for each temperature is reached earlier at the higher temperatures. LITERATURE CITED COLE, W. H., 1922a. Note on the Relation between the Photic Stimulus and the Rate of ^Locomotion in Drosophila. Science, 55: 678. COLE, W. H., \922b. Circus Movements of Limulus and the Tropism Theory. Jour. Gen. Physiol., 5: 417. CROZIER, W. J., 1924. On Biological Oxidations as Function of Temperature. Join: Gen. Physiol., 7: 189. CROZIER, W. J., 1928. Tropisms. Join: Gen. Psych., 1 : 213. CROZIER, W. J., 1929. The Study of Living Organisms. In Foundations of Ex- perimental Psychology. Worcester, Mass. : Clark University Press, pp. 45-127. DAVENPORT, C. B., AND W. B. CANNON, 1897. On the Determination of the Direc- tion and Rate of Movement of Organisms by Light. Join: Physiol., 21: 22. GLASER, OTTO, 1924. Temperature and Forward Movement of Paramecium. Jour. Gen. Physiol., 7: 177. 21 326 JOHN H. WELSH HADLEY, P. B., 1908. The Behaviour of the Larval and Adolescent Stages of the American Lobster, Homarus americanus. Jour. Comp. Neur. and Psych., 18: 199. HECHT, SELIG, 1924. The Visual Discrimination of Intensity and the Weber- Fechner Law. Jour. Gen. PhysioL, 7: 235. HECHT, SELIG, 1925. The Effect of Exposure Period and Temperature on the Photosensory Process in Ciona. Jour. Gen. PhysioL, 8: 291. HECHT, SELIG, 1928. The Relation between Visual Acuity and Illumination. Jour. Gen. PhysioL, 11: 255. HENRI, VICTOR, AND J. L. DES BANCELS, 1912. Sur 1'interpretation de la loi de Weber-Fechner. Compt. rend. Soc. Biol., Paris, 72: 1075. LOEB, J., 1890. Der Heliotropismus der Tiere und seine Uebereinstimmung mit dem Heliotropismus der Pflanzen. Wiirzburg. LOEB, J., AND J. H. NORTHROP, 1917. Heliotropic Animals as Photometers on the Basis of the Validity of the Bunsen-Roscoe Law for Heliotropic Reactions. Proc. Nat. Acad. Sci., 3: 539. LYON, E. P., 1906. Note on the Heliotropism of Palsemonetes Larvae. Biol. Bull., 12: 23. MAST, S. O., 1923. Photic Orientation in Insects with Special Reference to the Drone-fly, Eristalis tenax, and the Robber-fly, Erax rufibarbis. Jour. Exper. ZooL, 38: 109. MAST, S. O., AND M. COVER, 1922. Relation between Intensity of Light and Rate of Locomotion in Phacus pleuronectes and Euglena gracilis and its Bear- ing on Orientation. Biol. Bull., 43: 203. MOORE, A. R., AND W. H. COLE, 1921. The Response of Popillia japonica to Light and the Weber-Fechner Law. Jour. Gen. PhysioL, 3: 331. NAVEZ, A. E., 1930. On Temperature and the Breathing Rhythm of Canis mustelus and Squalus acanthias. Biol. Bull., 59: 104. OSTWALD, W., 1903a. Theoretische Planktonsutdien. ZooL Jahr., Abt. f. Syst, 18: 1. OSTWALD, W., 1903b. Zur Lehre vom Plankton. Natunviss. Wochensch., Neue Folge, 2: 481. PATTEN, B. M., 1915. An Analysis of Certain Photic Reactions, with Reference to the Weber-Fechner Law. Am. Jour. PhysioL, 38: 313. STIER, T. J. B., 1931. A Cooling Unit for Low-temperature Thermostats. Sci- ence, 73: 288. YERKES, R. M., 1900. Reaction of Entomostraca to Stimulation by Light. Am. Jour. PhysioL, 4: 405. BRANCHIAL RESPONSES TO ADRENALINE AND TO PITRESSIN IN THE EEL ANCEL KEYS* AND J. B. BATEMAN 2 (From the Physiological Laboratory, Cambridge, England) INTRODUCTION The gill per fusion method (Keys, 193 la) affords a convenient means of investigating the physiology of the branchial blood vessels in the fishes. In particular, it would seem to be desirable to study the effects of various hormones on the effective calibre of the branchial vessels and on the performance of the branchial chloride-secreting mechanism (Keys, 1931ft). We have investigated the effects of adrenaline and of pitressin on the gills of the eel, Anguilla vulgaris, by means of the ventral aorta-gill preparation as described in a recent paper (Bateman and Keys, 1932). Throughout each experiment the per fusion pressure (mean of systole and diastole of the pump) was maintained at a constant level by adjust- ment of the pump stroke and the reservoir level. Rates of perfusion were measured from the rates of inflow in some of the experiments and from the rates of outflow from the dorsal aorta in the other experiments. In the latter cases the dorsal aorta was cannulated at the level of the anterior portion of the liver. Practically identical results were obtained from these two methods. Venous escape from the cardinal and coronary systems was prevented in all cases except where specifically mentioned. The perfusion fluid used was the same as that given by Keys (1931o, p. 359), the concentrations being A = about 0.72° for eels from sea water and A = = about 0.60° for eels from fresh water. The external medium was either Plymouth sea water or Cambridge tap water, de- pending upon the medium in which the eels had been kept prior to the experiment. The net chloride exchange between internal and external media in the gills was determined by analyses of ingoing and outgoing fluids, using Keys' (1931c) method. Analyses were done in duplicate or triplicate. The effects of adrenaline were determined by the addition of adrena- 1 Fellow of the National Research Council. - Working on behalf of the Medical Research Council. 327 328 A. KEYS AND J. B. BATEMAN line chloride (Parke Davis) to the per fusion fluid immediately before use so as to give adrenaline concentrations between 1/300,000 and 1/1,000,000. Pitressin (Parke Davis), the pressor principle from the posterior lobe of the pituitary gland, was used in concentrations ranging 300 I i- 250 u P< 200 I O 150 100 15 30 45 60 75 90 105 120 135 150 Time in minutes FIG. 1. The influence of adrenaline (1/1,000,000) on the rate of flow through the gills under constant pressure. The encircled points are from measurements during adrenaline perfusion, and the arrows indicate the introduction and removal of adrenaline. from 25 to 50 international pressor units per liter by addition to the perfusion fluid just prior to use. The normal perfusion fluid was used as control in all cases. § 300 200 O O 2/o 8 *3£ i 3/0 3/ 2/0 I/O I/O I/O I/O I/O ^ FIGS. 5-8. Successive stages of diplotene. 1 Robertson and others have stated that the chromosomes are double before they pair but do not claim to have any direct evidence to show that this is so. The indirect evidence they bring forward consists in the well-known supposition of " anaphase duality " which is now seen to depend on a misinterpretation of spiral structure (Darlington, 1932). On the other hand, all direct observers of zygotene in plants and animals agree that the chromosomes are single at this stage (Gelei, Wenrich, Belling, Newton and Darlington). Speculations as to the pos- sible earlier division of the chromosomes at meiosis and mitosis conflict with the precocity theory of meiosis (Darlington, 1931) but are not valid evidence against it. 360 C. D. DARLINGTON pends on size differences. The loops first meet, revealing the number and position of the chiasmata, in the shorter pairs. As the loops open the doubleness of the chromosomes becomes detectable (Fig. 2). The chromatids can then be followed separately through the chiasmata and sometimes throughout their length. Later they swell considerably and cease to be separately identifiable, except at the chiasmata (Fig. 9). This swelling is seen in many animals (perhaps Pristiurus may be re- garded as the extreme type) and also in a gymnosperm, Taxus baccata 2/0 4/( 3/° '/° '/° y° '/• '/' 3/0 3/i 3/0 2/ 2/ l/o //. I/, 8 ft £*f in* 3/, 3/, 4/, 2/, 2/. I/. I/, I/, X 4/2 3/i 2A l/o l/o I/, l/o I/, X FIGS. 9 and 10. Diakinesis. FIG. 11. Pro-metaphase. FIG. 12. Metaphase, X chromosome lying to one side of the spindle. Note: the three long pairs of chromosomes have median spindle attachments, the rest terminal. (Dark, unpublished). It is characteristically different from the be- haviour in the angiosperms where the more swollen condition prevails throughout the post-pachytene stages and seems to prevent the chroma- tids being so clearly distinguished at diplotene as in some Orthoptera. The differences between angiosperm and orthopteran may, of course, be artefacts but it is impossible to say which is normal and which the artefact. Many would infer that the clearer observation is more true- to-life but such a conclusion does not necessarily follow. The increase in size has a curious effect. Owing to the differential but somewhat ORIGIN AND BEHAVIOUR OF CHIASMATA 361 variable rate of condensation of short and long chromosomes, their relative sizes during the diplotene phase change rather suddenly and interfere with the constant distinction of types (Figs. 3-6). t \ FIG. 13. Anaphase of the first division following failure of pairing of the shortest medium pair ; one of the univalents is about to divide at the first division. The chromatids that have been associated are connected by a dotted line. X ca. 3000. FIGS. 14 and 15. Second division. < ca. 3000. FIG. 14. Polar view of the eight autosomes in metaphase. FIG. 15. The two divisions of one spermatocyte. Above, anaphase; nine chro- mosomes including the X chromosome. Below, metaphase (chromosomes drawn separately) ; the distal ends of some of the chromosomes are coming together. Metaphase to Anaphase. — Between pachytene and metaphase the chromosomes contract to about one-sixth their length. The eight bi- valents lie on the equatorial plate and the X chromosomes to one side 362 C. D. DARLINGTON of it (Fig. 12). Repelling one another from their spindle attachments the pairs of chromatids associated at these points pass to opposite poles. Those pairs distal to the first chiasma therefore have to separate and the strain often draws them into a fine thread. Exceptionally a connection is seen between the separating chromatids (Fig. 13), as already noticed in Hyacinthus and Stenobothrns. This connection is as yet unexplained. Exceptionally also two of the shorter chromosomes are found to be un- paired and may lag and divide at the first division. This failure corre- sponds with observations during prophase and, as will be seen, is related to conditions of chiasma formation. SHOKT FIG. 16. Graph showing variation in chiasma frequency in the long, medium, and short types of chromosome. Second Division. — During the interphase and early metaphase of the second division the pairs of chromatids are held together only at their spindle attachments (Fig. 14). They suddenly come together, immedi- ately before anaphase, it must be presumed, for they are rarely seen together during metaphase and then only touching at the ends (Fig. 15). Syndiploidy. — Groups of from two to six spermatocyte nuclei are often found closely appressed, between pachytene and diakinesis ; and at metaphase, in a corresponding proportion of cells, fusion of the adjoining plates of chromosomes is seen. Where only two nuclei lie together the result is a regular fused spindle but, with a larger number, the chromosomes no longer orientate themselves regularly. These re- ORIGIN AND BEHAVIOUR OF CHIASMATA 363 suits are closely paralleled by Eisentraut's observations (1926) on Gomphoccnis maculatns and by numerous observations of syndiploidy in plants, although the exact time of fusion, the onset of metaphase, is not elsewhere so clear. It is evident that in all these cases fusion takes place after pachytene and, although four homologous chromosomes of each type are present (in both plants and animals), quadrivalents are TABLE I Summary of Observations of 100 Nuclei (800 Bivalcnts) Numbers of Terminal Chiasmata in each Bivalent Numbers of Chiasmata per Bivalent long type (three) medium type (four) short type (one) 1 2 3 4 5 6 0 i 2 3 1 '2 Mid [0 Diplotene I 1 (24 nuclei) 1 2 1 11 3 1 20 9 2 9 7 1 4 2 1 1 — • 67 3 11 15 , 8 14 2 Total 1 15 31 17 7 1 — 70 26 — 22 2 Late TO Diplotene I 1 (16 nuclei) \ 2 . 8 3 10 13 1 4 7 2 41 10 7 6 3 13 16 1 1 Total — 11 24 11 2 — 51 13 — • Diakinesis I . (40 nuclei) j * 2 14 12 1 17 30 12 2 15 10 1 2 2 1 90 43 25 1 4 35 Total 2 27 59 27 5 — 1 133 25 1 39 Metaphase I . (20 nuclei) j * 3 9 1 3 12 8 1 9 14 , — 31 30 18 1 4 15 1 Total — • 13 23 24 — — — • 61 18 1 19 96 1 4 4 Grand Total 3 66 137 79 14 1 1 315 82 2 Percentages 1 22 45.7 26.3 4.7 0.3 0.3 78.7 20.5 0.5 96 never formed. Evidently, therefore, the pairing of chromosomes at metaphase in these organisms can only be derived from pachytene pair- ing and does not arise from a direct affinity. This conclusion follows from the chiasma theory of pairing and contradicts the assumption that the chromosomes are paired at metaphase on account of any attraction operating between them at this stage (Darlington, 1931). 364 C. D. DARLINGTON Chiasma Frequency. — The numbers of chiasmata present in the bi- valents of each type remain without significant change from diplotene to metaphase. They show (Fig. 16) the interference curves of fre- TABLE II Mean Chiasma Frequencies per Bivalent (derived from Table I) Mid Diplo- tene Stage (24 nuclei) Late Diplo- tene Stage (16 nuclei) Dia- kinesis (40 nuclei) Meta- phase (20 nuclei) Total (100 nuclei) Observa- tions on Steno- bothrus parallelus Length No. Bivalents. . . Total No. Chi- asmata 72 233 48 148 120 366 60 191 300 938 Long Type No. Terminal Chiasmata. . . . Chiasma Fre- quency per Bi- valent 34 3.24 29 3.08 109 3.05 76 3.18 3.13 3.31 11.3/x Terminal Chias- mata per Bi- valent .47 .60 .91 1.27 No. Bivalents. . . Total No. Chi- asmata 96 122 64 77 160 186 80 100 400 485 Medium Type No. Terminal Chiasmata. . . . Chiasma Fre- quency per Bi- valent 18 1.27 16 1.20 69 1.16 49 1.25 1.21 1.45 4.1 „ Terminal Chias- mata per Bi- valent .18 .25 .43 .61 No. Bivalents. . . Total No. Chi- asmata 24 26 16 16 40 41 20 21 100 104 Short Type No. Terminal Chiasmata. . . . Chiasma Fre- quency per Bi- valent .... 16 1.08 13 1.00 36 1.02 16 1.05 1.04 1.04 1.6/i Terminal Chias- mata per Bi- valent .61 .81 .90 .80 quency variation constantly observed in all organisms so far studied (Haldane, 1931; cf. Darlington, 1932). They also show (Table II) the indirect relationship of length of chromosome to frequency already found in Stcnobothrus and probably very general in organisms with ORIGIN AND BEHAVIOUR OF CHIASMATA 365 wide range of size amongst the chromosomes (Darlington and Dark, 1932). This effect is not so pronounced, however, in the medium chromosomes as in Stenobothrus with the predictable result (on the chiasma theory of pairing) that the shortest member of this type occa- sionally fails to pair (Fig. 13, anaphase, and Table I, diakinesis). Terminalization. — The total number of chiasmata remains the same, but the proportion of these that are terminal increases during prophase and it increases in a characteristically different way in the different types of bivalent and in those with different total numbers of chiasmata (Table III and Fig. 17). This agrees with the assumption that the terminal chiasmata arise by movement of earlier interstitial ones to the ends. Thus, amongst those with 2, 3, and 4 chiasmata, the proportion that are terminal (the terminalisation coefficient) is the same at each stage. This indicates that originally all the chiasmata were interstitial as re- quired by the hypothesis. Further, as in Tidipa (Darlington and Janaki TABLE III Numbers of terminal chiasmata per bivalent in bivalents zvith different total numbers of chiasmata (derived from Table /) Stage Long Type Medium Type Short Type Mid Diplotene 2 .33 .27 .52 .84 3 .42 .62 .91 1.22 4 .53 .64 1.30 1.54 1 .04 .20 .48 .49 2 .58 .46 1.00 1.00 1 .64 .81 .90 .79 Late Diplotene . . Diakinesis . . . . Metaphase . . . . Animal, 1932), those with one chiasma show less movement, especially in the early stages, than those with two or more. Finally, the occurrence of more terminal chiasmata at the same stages in the corresponding classes of the shorter chromosomes is to be expected on the assumption of movement since the chiasmata are more concentrated in these chromo- somes and have not so far to go to reach the ends. The only change that takes place in terminalization in the long chro- mosomes in this organism is the expansion of closed loops at the expense of the distal arms, a change which according to the electrostatic hypothe- sis (Darlington and Dark, 1932) is due to a repulsion between dis- tributed surface charges on the paired chromatids. In the medium and short chromosomes there is also a slight movement of single chiasmata owing to the special repulsion of the spindle attachments, but this move- ment is ineffective in moving a proximal chiasma against the repulsion of a loop when two are present owing to the terminal spindle attach- ments always lying in open arms in which repulsion is less effective. 366 C. D. DARLINGTON The terminalization of the chiasmata in the short chromosomes is therefore very like that in the fragments of Fritillaria imperialis where all chiasmata are terminalized early although only the distal chiasmata move to the end in the major chromosomes with numerous chiasmata (Darlington, 1930). X5 O ja. a 1.0 o c 'I 0.5 _Q s ORIGIN MID- (LATE DIA- META- OF CHIASMATA DIPLOTENE DIPLOTENE) KINESIS PHASE FIG. 17. Graph showing that the increase in the numbers of terminal chias- mata is in proportion to the total number at all stages where more than two are formed but not in the earlier stages where only one is formed. The short chromo- some type and the late diplotene stage are omitted because the numbers of observa- tions are smaller than those given. SUMMARY 1. A study of meiosis in male Chorthippus clegans, Acrididse (2n = 16(-|-X) shows the chromosome behaviour to be similar to that al- ready described in Stenobothrus parallelus. Thus the chiasma fre- quency is an indirect function of length and has an interference curve of variation. Failure of pairing occurs in one chromosome type with a frequency in keeping with the curve. Terminalization depends entirely on the generalized repulsion in bivalents with closed loops and the lo- calized spindle attachment repulsions are of the minimum degree found in Fritillaria. ORIGIN AND BEHAVIOUR OF CHIASMATA 367 2. A more extensive quantitative study makes it possible to show in the long chromosome type, that the number of terminal chiasmata at each stage between diplotene and metaphase is proportional to the num- ber of interstitial chiasmata and increases from one stage to the next pari passu with the decrease in the number of the interstitial chiasmata. This confirms the earlier arguments that all interstitial chiasmata be- come terminal by movement without breakage while, on the other hand, terminal chiasmata always arise from earlier interstitial ones and in no other way. It is now clear that chiasmata always change their posi- tion after their formation at diplotene so that the configurations observed later are merely positions of changing equilibrium. 3. The opening of the diplotene loops has been followed in detail and shows that the chromosomes are single and undivided until this stage. The opening is therefore derived solely from the so-called " re- ductional " split, not from the " equational " one which only begins to appear at this time. 4. Syndiploidy occurs frequently just before metaphase. For bibliography, see p. 370 of the following paper by the same author. THE ORIGIN AND BEHAVIOUR OF CHIASMATA VI. HYACINTHUS AMETHYSTINUS C. D. DARLINGTON JOHN INNES HORTICULTURAL INSTITUTION, MERTON, LONDON The original frequency and distribution and the later behaviour of chiasmata have been made clear in a number of organisms by earlier studies in this series and the comparison of observations on Campanula, Tulipa, Fritillaria, and Stenobothrus has shown us how to analyse the changes undergone by chiasmata in terminalization and has enabled us to define the forces at work in producing these changes (Darlington and Dark, 1932). Hence it is now possible to recognise from typical meta- phase conditions of other organisms what the prophase conditions pre- ceding and determining them must have been. This method can be satisfactorily applied to many species in which the prophase is not amenable to direct study. A number of species of plants and animals are known with an ex- treme range in the sizes of their chromosomes (cf. Darlington, 1932). In species with slight size range the number of chiasmata formed is as a rule roughly proportionate to the lengths of the chromosomes at pachytene (e.g., in Hyacinthus orientalis, Fritillaria imperialis, and Vicia Faba). This would only be possible with a great range of size if the longer chromosomes had a very high number, for the shortest must always form one chiasma, according to the chiasma theory of pair- ing, to ensure that they pair regularly (Darlington, 1930). The indirect size- frequency relation expected on this theory has been found in Stenobothrus and Chorthippus where the extreme lengths are as 8 to 1. It may be inferred on the analogy of the observations in Brachystola, Yucca flaccida (O'Mara, 1931), and in the South Ameri- can Acrididse (Saez, 1930). In the last the abnormal frequency rela- tion is evidently due to localization of chiasmata near the spindle attach- ment so that the same length of chromosome is concerned in forming chiasmata in all the chromosomes. In the other examples the distribu- tion is even and the mechanism controlling the abnormal length-fre- quency relation cannot be so directly inferred. High size ranges occur in many species of monocotyledons. In Eucomis bicolor, 2n = 32 (Fig. 1) (cf. Miiller, 1912), and in Hyacin- 368 ORIGIN AND BEHAVIOUR OF CHIASMATA 369 thus ctiiicthystiniis. 2n--28l (Fig. 2), the range is about 20 to 1, i.e., the same as in DrosopJiila inclanogastcr. It is to be noted that the shortest chromosomes are too small to attain the characteristic chromatid breadth of the species. In HyacintJins the complement consists of 10 chromosomes about 5 /* long together with 18 less than 1 ^ long and an average about one-tenth the length of the longer chromosomes. At the first metaphase of meiosis 14 bivalents are regularly found in Hyaciiithns aiucthvstiiuts. Polar views show the five long pairs with two or three chiasmata (Fig. 3). The detailed structure of the nine short pairs cannot be determined from this aspect. As always, it is necessary to examine bivalents of this size in side view. (This is not FIG. 1 FIG. 2 FIGS. 1 and 2. Mitotic metaphases from the root tip. < 3200. FIG. 1. Eitcoinis bicolor, 2n = 32. FIG. 2. Hyacinthus amethystinus, 2n^28. yet generally appreciated.) It is then found (Fig. 4) that most have a single terminal chiasma, a few have a single interstitial chiasma and occasionally there is one with two chiasmata. The mean chiasma fre- quencies of long and short types in this division are 2.4 and 1.1 respectively. These observations show the closest analogy with Stcnobothrus and Chortliippus except that the departure from normal in length- frequency relation is even more pronounced. The degree of terminalization is the same : the terminalization coefficient for the longer chromosomes is .45 in Stcnobothrus and .42 in HyacintJins, for the shorter chromosomes .67 in StcnobotJinis and .60 in HyaciutJtus. It is therefore evident that the conditions of terminalization are similar in the two instances and that since chiasma frequency is not reduced during prophase in Stenobothnts 1 The somatic chromosome number is given by Heitz (1926) as 24. 24 370 C. D. DARLINGTON it is similarly unaltered in Hyacinthus, the only change being a move- ment of interstitial chiasmata to the ends of the chromosomes. FIGS. 3 and 4. First metaphase of meiosis in Hyacinthus amcthystinus. < 3200. The total numbers of chiasmata and numbers terminal are given. Sec- tions cut at 24 f--. FIG. 3. Polar view. The structure is only identifiable in the long bivalents. \/> l/i 2./i l/o I/ L2/> l/o L 3/2. L 2/o L 3/a I/ l/o L 2/i I/. FIG. 4. Side view, bivalents drawn separately. SUMMARY The longer chromosomes of Hyacinthus auiclhystinus are on the average ten times the bulk of the shorter ones but have only twice as many chiasmata per bivalent. Thus, although the longer chromosomes form only two or three chiasmata, the shortest chromosomes regularly form one chiasma which ensures their regular pairing. This abnor- mality is characteristic of particular species and, like other variations in chiasma frequency and distribution, it must be genetically con- trolled. It is therefore to be regarded as an adaptive property. BIBLIOGRAPHY BELLING, J., 1931. Chromomeres of Liliaceous Plants. U. Cal. Publ. in Bot., 16: 153-170. DARLINGTON, C. D., 1930. Chromosome Studies in Fritlllaria, III. Cytologia, 2: 37-55. DARLINGTON, C. D., 1931. Meiosis. Biol. Revs., 6: 1-43. DARLINGTON, C. D., 1932. Recent Advances in Cytology. Blakiston, Philadelphia. ORIGIN AND BEHAVIOUR OF CHI ASM AT A 371 - AND DARK, S. O. S., 1932. The Origin and Behavior of Chiasmata, II. Stcnobot lints parallclns. Cytologia, 3: 169-185. - AND JANAKI AMMAL, E. K., 1932. The Origin and Behavior of Chiasmata, I. Diploid and Tetraploid Tulipa. Bot. Gas., 93: 296-312. EISENTRAUT, M., 1926. Die spermatogonialen Teilungen bei Acridiern mit be- sonderer Beriicksichtigung der Uberkreuzungsfiguren. Zeits. itnss. Zool., 127: 141-183. HALDANE, J. B. S., 1931. The Cytological Basis of Genetic Interference. Cyto- logia, 3: 54-65. HEITZ, E., 1926. Der Nachweis der Chromosomen : Vergleichende Studien iiber ihre Zahl, Grosse und Form im Pflanzenreich, I. Zcits. f. Bot., 18: 625- 681. MULLER, H. A. C., 1912. Kcrnstudien an Pflanzen, I und II. Arch. f. Zellf., 8: 1-51. O'MARA, J., 1931. Chromosome Pairing in Yucca flaccida. Cytologia, 3: 66-76. SAEZ, F. A., 1930. Investigaciones sobre los cromosomas de algunos Ortopteros de la America del Sur. Rev. Mus. La Plata, 32: 317-361. DOMINANCE OF TWO KIDNEY ALLELOMORPHS IN HABROBRACON JUGLANDIS (ASH.) B. R. SPEICHER (From the Zoological Laboratory, University of Pittsburgh, and the Marine Biological Laboratory, Woods Hole, Massachusetts) INTRODUCTION An allelomorphic series of at least four factors affecting the size of the compound eye has been located on the first chromosome of Habrobracon. The present paper deals with the effects of two of these. One, extreme small (ke), was reported (Dunning, 1931) by W. F. Dunning and was kindly sent by her to P. W. Whiting. The eye size of mutant-type wasps showing this character is extremely variable, ranging from a total lack of eyes to those which, though approaching, never reach the normal size. The individual facets, when present, are of normal shape and size, the variation occurring as a decrease in the total number. The sizes of right and left eyes vary somewhat independently of each other. The ocelli are likewise affected and are also extremely variable, ranging from none at all to those apparently normal in size. Although no actual measurements have been made, it has been noted that the ocelli of any one individual tend to be of the same size. Aside from the modification of compound and simple eyes no other external effect is manifest. The mutant-type exhibits excellent viability when reared under standard conditions. Kidney (k), an X-ray mutation found (Whiting, 1932) by Anna R. Whiting, presents superficially all of the characteristics of extreme small, but shows a high percentage of lethality when reared at standard temperature (30° C.). At lower temperatures these wasps show a viability equal to that of extreme small. The striking similarity of the two mutants, coupled with their allelomorphism, suggested need for further study. Are they actually two separate factors having, except for the thermo-lethality of one, the same methods of expression, or is kidney a re-mutation to extreme small but carrying with it a closely linked factor causing lethality at higher temperatures? If the former hypothesis is correct, what are the conditions in regard to dominance? It was considered probable that a statistical study of the two types would bring out facts which, because of variability, would be obscure at casual inspection. 372 TWO MUTANT ALLELOMORPHS IN HABROBRACON 373 The writer is indebted to the Committee on Effects of Radiation on Living Organisms of the National Research Council for technical assistance through a grant to Dr. P. W. Whiting. Acknowledgment should also be made to Dr. Anna R. Whiting and to Dr. W. F. Dunning for making available the mutations here used, and to Dr. P. W. W'hiting, at whose suggestion this work wras done and wrhose help was greatly appreciated. The writer is likewise grateful to Miss Kathryn A. Gilmore for suggestions regarding technical difficulties. MATERIALS AND METHODS Crosses Made Two groups of females, one heterozygous for extreme small, the other heterozygous for kidney, were obtained by mating wild type (Stock No. 11) females to extreme small and to kidney males respec- tively. These heterozygous females were mated as follows: (1) 9 Kke by c (2) 9 Kke by c (3) 9 Kk by c (4) 9 Kk by c ? ke. ? k. ? ke. ? k. The offspring from one half of the females of each of these four crosses were reared at 23° C. and those of the other half wrere reared at 30° C. All progeny were graded according to (1) the diameters of the compound eyes and (2) the size of the ocelli. Considering the number of mutant eyes counted, any measurement by facet count was impracticable. Consequently the lesser diameter of the eye was used as a basis of comparison. Grading of Compound Eyes and Ocelli Six grades of eye size wrere arbitrarily marked off (Fig. 1), of which four wrere set off in the following manner. From a stock of mutant- type w^asps the individual possessing the largest eye was selected. With this eye as the upper limit of variability and the bare indication of the scleral ring surrounding the eye as the lower limit, four grades were established covering equal ranges in regard to eye diameter. These groups constituted Grades 1, 2, 3, and 4, in increasing order of size. During the process of grading the experimental material, a few eyes were found which exceeded the upper limits of Grade 4; these were placed in a separate group, Grade 5. Although, theoret- ically, Grade 5 extends from Grade 4 to wild-type, in no case did the sizes of any eyes belonging to this grade so closely approach that of type as to cause the two to be indistinguishable. It was considered advantageous to record separately the eyes which had failed to develop; 374 B. R. SPEICHER this group was designated Grade 0. Thus Grades 0 and 5 mark the extremes in variation of eye size with Grades 1, 2, 3, and 4 representing the intermediate stages. Eyes were selected (to be used as standards of comparison with the experimental material) which met the theoretical specifications of the grades. As an aid to more accurate grading, each standard was made to represent the upper limit of its grade. Eyes under question as to their proper position in the series could then be com- pared with the standard most nearly resembling them and if of the same size or slightly smaller would be placed in the same grade as Grade 2 Grade 0 FIG. 1. Eyes of wild and mutant-type wasps. Illustrations of mutant-type wasps represent the upper limits of their designated grades. X 33. represented by the standard, but if they were larger they would be classified as belonging to the next higher grade. Each wasp was com- pared directly with these standards, the sizes of right and left eyes being recorded separately. The ocelli were divided into three groups, absent, small, and normal. Although, as mentioned above, the ocelli of any one individual tended to be of the same size, in case of disparity the wasp was classified in the small ocelli group. Recording of Lethal Wasps In Habrobracon lethal individuals may be recognized by the pres- ence of shriveled larvae or blackened pupae still confined within the TWO MUTANT ALLELOMORPHS IN HABROBRACON 375 cocoons. While lethality occurring during the development of the egg may escape detection, some estimation of the number of lethal wasps may be made by scrutinizing the culture vials. Lethal counts given in this paper were made by using the above method. PRESENTATION OF DATA Comparative Lethality of Mutant Types Among the four classes of wasps which were reared at 30°, two had an excessive number of lethal individuals. As kidney is known to be lethal when reared at this temperature, the fraternities contain- ing kidney wasps may be expected to be the ones most affected. As seen in Table I, offspring of the crosses Kk X ke and Kk X k contain TABLE I Numbers of type, mutant, and lethal wasps resulting from different crosses reared at different temperatures. Origin Wasps reared at 30° C. Wasps reared at 23° C. Kke Xke KkeXk Kk xke KkXk Kke Xke Kke Xk Kk Xke KkXk Type Mutant 301 300 432 383 344 172 198 23 222 220 275 232 239 240 182 178 Lethal 71 44 119 134 47 54 33 45 by far the greater number of lethal wasps. The cross Kk X ke pro- duces kidney males and the cross Kk X k produces both kidney males and kidney females, thus accounting for the greater number of lethal individuals among the progeny of the former cross and the still greater number among those of the latter cross. The number of males produced in proportion to the females is an important factor affecting the comparative lethal ratios. An excess of males in a class coming from the cross Kk X ke would raise the entire lethal ratio and, since the males of the lethal group were not separated from the females, the cause for the increase would not readily be apparent. In this case the sex ratio of the mutant-type must be con- sidered, and may be assumed to be approximately the same as that of the type wasps of the same group. In the cross Kk X ke, reared at 30°, there was an excess of males, but not of statistical significance. In the other groups reared at 30° and 23° the sex ratios were likewise not significantly different. Offspring from the cross Kke X k had a low lethal ratio. The mutant-type females of this group, being extreme small-kidney com- pounds, should demonstrate the dominance of one or the other factor 376 B. R. SPEICHER in regard to thermo-lethality. The low lethal ratio and the abundance of viable compound females (Table I) indicate that the thermo-lethal trait associated with kidney is recessive. The lethal ratios of the four classes reared at 23° were not significantly different, since neither mutant factor causes an excess of lethals at this temperature. A peculiarity brought out in Table I which is worthy of passing notice is the excessive number of mutant-type wasps in classes reared at both temperatures. Since the number of lethals counted represents the minimum actually produced, and the major part of the lethal group consists of mutant-type wasps, this excess is obvious, although the cause is as yet unexplained. TABLE II Grades of eyes and of ocelli tabulated according to genetic constitution and temperature. As regards eye size, the numbers in the body of the table indicate numbers of eyes graded. As regards grades of ocelli, the numbers indicate individuals. Wasps reared at 30° C. Wasps reared at 23° C. ke keke kek •kk k ke keke kek kk k Eye size Grade 0 86 83 300 115 14 1.81 1 289 9 97 51 119 23 2 1.25 4 141 1 305 120 308 55 1.14 19 374 1 7 4 11 0.68 10 1 33 11 8 4 0.64 13 15 6 19 109 84 56 2.60 136 1 39 38 113 71 11 1.92 2 134 2 41 624 3 3.94 320 15 1 6 193 2 3.97 10 91 6 308 6 4.00 5 156 1 2 3 4 5 Mean Ocelli Lacking Small N^ormal Comparative Eye Sizes of Mutant Types The effect of a 30° temperature on the eye sizes of the various genetic classes is seen in Table II. Most of the offspring from each of the four types of matings possess eyes which fall into the lower groups, Grades 2 and 0 being particularly well represented. The fact is well demonstrated in the classes unaffected by thermo-lethality and is suggested in the kidney classes, although in these latter groups the numbers are not sufficient to be statistically significant. The dispersal of wasps according to the sizes of both right and left eyes is seen in Table III. Among the wasps reared at 30° the groups composed of extreme small females, compound extreme small-kidney females, and to a lesser extent the extreme small males show a pre- ponderance of individuals in four places, involving Grades 0 and 2. TWO MUTANT ALLELOMORPHS IN HABROBRACON 377 TABLE III Correlation tables showing the. relation of size of right and left eyes in wasps of different genetic constitution reared at different temperatures. Ordinates represent grades of left eyes and abscissae represent grades of right eyes. The numbers in the body of the table indicate the number of individuals reared, the eyes of which fall in the designated grades. 30° 23° 012345 012345 d,ke 5 4 3 2 1 0 7 1 1 5 10 17 5 7 24 19 1 12 17 3 12 15 84 35 1 3 5 29 22 3 1 1 9 22 2 2 3 2 4 1 16 5 16 4 1 9,keke 5 4 3 2 1 0 ! 2 1 1 1 4 3 4 2 8 16 13 2 22 13 27 3 1 5 4 2,8 13 2 13 5 6 3 2 2 12 2 19. 3 16 1 8 5 4 4 1 9,kek 5 4 3 2 1 0 2 1 10 296 1 5 4 15 6 7 17 50 18 78 5 1 1 8 14 19 5 78 28 45 6 9,kk 5 4 3 2 1 0 1 1 95 2 1 3 1 3 1 1 1 1 1 d\k 5 4 3 2 1 0 1 1 151 3 2 1 3 3 0 1 4 2 1 3 1 378 B. R. SPEICHER Thus in the compound extreme small-kidney group in Table III, the areas 0, 0 (78 individuals), 0, 2 (50 individuals), 2, 0 (45 individuals), and 2, 2 (78 individuals) possess an abundance of wasps as compared with the surrounding squares. This indicates a tendency for the wasps to have the eyes either entirely lacking or developed up to a stage represented by Grade 2, which is apparently somewhat of a limit for eye size of wasps reared at this temperature. The bimodality of each of these classes, as represented in Tables II and III, may be due in part to the method of measurement. The anatomical variation of the eye, although used as a basis of com- parison, is a secondary result of a more fundamental but unmeasured physiological change within the organism. In Grade 0 this variation reaches a lower limit of expression. The physiological variation, how- ever, does not stop here but grades down to the lethal point. Below the threshold of viability the effects are again perceptible externally, this time as lethals, the majority of which have the eyes entirely lacking and are noticeably micro-cephalic. There is, then, a range of physiological variation which can be seen externally only by Grade 0 and the lethal group, and as these two groups must contain those individuals falling into the intermediate physiological grades their numbers are thus increased. From this it appears that Grade 0 is not a true mode of eye size, the strong representation of individuals in this grade being a result of the methods used in calibrating the variation. The true mode is confined to the upper limit of variation, which for those classes reared at 30° centers around Grade 2. This explanation, although difficult to prove, is supported by the fact that in the arrays of eye sizes of wasps reared at 23° (in which cases the ranges of physiological variation have been raised) there is but one mode present, centering around the upper limits of variation. A difference in the expression of the mutant factors appears in the classes reared at 23°. As seen in Tables II and III, extreme small wasps possess eyes which cover the same grades as do those of the corresponding group reared at 30°, although their means (Table II) are slightly higher. In contrast to this is the array of kidney wasps which at 23° have the eyes concentrated around a single high grade. The compound extreme small-kidney females follow the kidney trait in the concentration of eye size around Grade 4, indicating a complete dominance of kidney over extreme small in regard to this character. A comparison of means of eye size is given in Table II. The difference between the means of extreme small and kidney wasps, both reared at 30°, suggests a difference caused by the two factors at this temperature, although the numbers are too small to be statistically significant. TWO MUTANT ALLELOMORPHS IN HABROBRACON 379 In judging the variation of the ocelli of wasps reared at 30° the modes of the classes containing the factor extreme small are in the "small" group, as represented in Table II. In the kidney wasps the mode appears to have shifted down towards the "lacking" group, although here again the small number of viable wasps is not sufficient to be of any statistical significance. In the 23° groups, however, the numbers are large enough to make comparisons. At this temperature the extreme small groups retain their small ocelli, but in the kidney groups the majority of ocelli appear normal in size. The compound extreme small-kidney females resemble the extreme small group at this temperature in regard to the size of the ocelli. Correlation in Size of Right and Left Eyes At 23° kidney and kidney-extreme small compounds showed little variation in regard to eye size, and at 30° kidney wasps were too few to give significant results. Correlation coefficients were calculated as follows : Extreme small males, reared at 30°, r = = 0.2939 ± 0.0352. Extreme small females, reared at 30°, r = = 0.1129 ± 0.0568. Extreme small-kidney females reared at 30°, r = 0.2375 ± 0.0353. Extreme small males reared at 23°, r = = 0.4520 ± 0.0458. Extreme small females reared at 23°, r = 0.2351 ± 0.0546. It is evident that correlation in eye size exists although it appears to be not very high under the conditions of measurement. SUMMARY AND CONCLUSIONS 1. Two mutant allelomorphs in Habrobracon, extreme small, ke, and kidney, k, superficially resembling each other in their effects on eye size, were studied to determine their interrelationships. 2. Kidney wasps reared at 30° are for the most lethal, appearing in general as eyeless small-headed pupae. Extreme small are viable at 30° and no lethal effect is evident at 23° in either type. Extreme small is dominant as regards thermo-lethality since extreme small- kidney compound females are of normal viability when reared at 30°. 3. In extreme small wasps size of compound eye is relatively little affected by temperature. In kidney, however, a lower temperature causes a much larger size. Kidney is dominant in this respect in the compound females. 4. In extreme small wasps size of ocelli is likewise relatively little affected by temperature, being "small" both at 30° and at 23°. In kidney, however, the majority reared at 23° have normal ocelli, while 380 B. R. SPEICHER at 30° ocelli are smaller or altogether lacking. Extreme small is dominant over kidney as to ocellar size both at 30° and at 23°. 5. There is a decided tendency in both mutant types for the com- pound eye either to be entirely absent or to develop to a certain modal size determined by temperature and genetic constitution. 6. There is a low but significant correlation as to size between right and left eyes. BIBLIOGRAPHY DUNNING, W. F., 1931. A Study of the Effect of X-ray Radiation on Occurrence of Abnormal Individuals, Mutation Rate, Viability and Fertility of the Parasitic Wasp, Habrobracon juglandis (Ashmead). Genetics, 16: 505. WHITING, P. W., 1932. Mutants in Habrobracon. Genetics, 17: 1. THE MOSAIC DEVELOPMENT OF THE ASCIDIAN EGG N. J. BERRILL DEPARTMENT OF ZOOLOGY, McGiLL UNIVERSITY, MONTREAL The development of ascidian eggs has long been considered, as the result of the classical investigations by Conklin, to be the perfect example of the mosaic rather than the regulative type. Different blastomeres of dividing eggs were killed or sufficiently injured to inhibit further development and the development of the remaining blastomere or blastomeres was followed. He found that each blastomere developed only into the tissues and organs that it would have produced had it remained a part of intact egg or embryo. In no case did a whole but dwarf larva result from one of the first two or four cells. The left blastomere of the 2-cell stage gave rise to the organs of the left side only. The anterior two blastomeres of the 4-cell stage gave rise to epidermal, neural, and chordal tissues, but failed either to gastrulate or to form a tail. In general the organization of amphibian eggs is similar to that of ascidian eggs, formative substances appearing rather later, although Hall has recently shown that they are delimited as early as the 8-cell stage. Isolated blastomeres of the 2-cell stage, however, may give rise either to a whole larva of half the normal size or to a right or left half-larva of normal size, according to whether or not there has been reorganization of the cell contents. In consequence of this it has seemed possible that the absence of regulation or reorganization in Conklin's experiments may have been due to the fact that the injured blastomeres were left in situ within the egg membrane. The presence of their inert mass in contact with the surviving part may have had an inhibitory effect. This possi- bility has been emphasized by Reverberi's conclusions, based upon experiments with Ciona eggs, that the eggs of ascidians should be placed properly within the regulative class. The great obstacle to experimental work with ascidian eggs has been the difficulty of removing the egg membranes and follicle cells without injuring the ovum. It is partly the object of this paper to record a method by which this may be done. Several ascidian eggs are shown in Fig. 1 , and it is seen that there is considerable variation in the character of these membranes. 381 382 N. J. BERRILL It has already been shown (Berrill, 1929) that the membranes are normally digested from within by a proteolytic enzyme secreted by the developing tadpole, that the enzyme is fairly stable and is active within the limits pH 7.0-10.0. Trypsin digests the membranes, but only very slowly, so that the protease concerned is probably a tryptic ferment of a less specialized nature than trypsin itself. There are, accordingly, two means by which membranes of ascidian eggs may be removed. The enzyme produced by developing ascidian tadpoles may be collected and concentrated, or a suitable and more readily obtainable substitute may be found. It was discovered that dona, Phallusia, Ascidia, and Ascidiella eggs remained about 100 per cent viable at 18° C. for a little over twenty-four hours, so that any enzyme mixture to be of use must be effective within that period. FIG. 1. Ascidian eggs. A, dona intestinalis; B, Molgula or Polycarpa; C, Ascidiella aspersa. ifc, inner follicle cell; im, inner membrane; ch, chorion or egg- membrane; ofc, outer follicle cell. The natural ascidian enzyme in sufficient concentration is prefer- able to any substitute, but it can be obtained only from species of the family Ascidiidae (i.e., the genera Ascidia, Ascidiella, Phallusia} and Ciona. To obtain it in effective concentration it is necessary to make mass cultures of some thousands of embryos. Normal develop- ment of large numbers of eggs necessitates the use of a relatively large volume of water, the larger the better, for it is only when development has proceeded to within two or three hours of hatching that concen- tration becomes effective. Then the embryos are allowed to settle on the bottom of the vessels and as much supernatant water as possible is poured off. The remainder, together with the embryos, is poured into a test tube or cylinder, the embryos again allowed to settle, and the water reduced to about 5 cc. This with the embryos is then spread over the bottom of a Petri dish to ensure a supply of oxygen to the embryos, left there until hatching is complete, and the water then MOSAIC DEVELOPMENT OF ASCIDIAN EGG collected. It should contain enough enzyme to digest off the mem- branes from unfertilized Ciona or Ascidia eggs in from five to ten hours. Such eggs may subsequently be fertilized. If it is desired to remove the membranes from merely a small number of eggs, the eggs may be placed among the developing tad- poles about three hours before they are due to hatch, when their membranes will dissolve about an hour after hatching (the developing tadpoles should be close enough together to be almost touching and to form a sheet one layer deep). It may be desirable to remove the layer of outer follicle cells by washing the eggs in a bag of fine bolting silk, when the enzyme can more readily attack the underlying mem- brane. Large Ascidia or Ciona are not, however, always obtainable, and it is often more convenient to use a substitute for the ascidian enzyme. Such substitutes may be obtained from any large carnivorous or omnivorous invertebrate, the most active mixtures being the stomach juices themselves rather than gland extracts or preparations. Those actually used were the clear juices found in the stomach of decapod Crustacea such as Munida, Maia, and Homarus. No doubt other forms would be equally effective. Undiluted, such juices are too potent and toxic, but a dilution of one part juice with fifty or one hundred parts of sea-water is both safe and efficient. The time taken for the membranes to be removed depends not only upon the activity of the enzyme but also on the thickness of the membrane and on the temperature. At 18° C. the membranes of Ciona, Molgula, or Ascidia are removed in from two to four hours, but the thicker membranes of Polycarpa or Styela require more than twenty-four hours. Unfertilized eggs of Ciona, Molgula, and Ascidia that have had their membranes removed in this way may subsequently be fertilized and undergo normal development. If fertilized eggs are subjected to such a mixture or if eggs are fertilized while yet within it, a somewhat unexpected feature appears. Nuclear division proceeds normally, at least to the 32-cell stage, but cytoplasmic division is inhibited. In the case of later embryos, the surface cells round off and tend to fall away. These phenomena are apparently due to surface action of the lipochrome contained in the original digestion mixture, and in consequence this method of removing the membranes is satisfactory only in the case of unfertilized eggs. In this last case the capacity for fertilization and normal development is unaffected. Once membranes have been removed, experiments of two kinds become possible. The naked eggs may be cut with glass needles, as 384 N. J. BERRILL used by Horstadius (1928), while blastomeres may be separated from one another. Attempts to cut the eggs were made only in the cases of Ascidia and Ciona. Those of Ascidia (Ascidiella aspersa) possessed a tough surface and a very fluid endoplasm, so that all attempts at cutting resulted in bursting them. The eggs of Ciona, on the other hand, are readily cut but possess such a viscous endoplasm that the injured surface does not re-form. In neither case, therefore, is experi- mental work of this nature feasible. It is just possible, however, that other eggs, such as those of Molgula or of other species of Ascidia, may be more suitable and of a type intermediate between those of Ciona and Ascidiella. Separation of blastomeres, on the other hand, is very easy to effect. They may be separated with a glass needle, as used by Horstadius, or more simply merely by slight shaking. Pouring water in which there may be some hundreds of naked eggs in the 2-cell or in the 4-cell stage from one vessel to another several times results usually in the complete separation from one another of all blastomeres. There is a tendency among the blastomeres of Ciona eggs to fall apart in any case, though eggs and early embryos of Ascidiella tend to fuse on contact. In the subsequent development of blastomeres isolated in the above manner no indication was found of any reorganization. Blasto- meres isolated in the 2-cell stage invariably give rise to lateral half- embryos that gastrulate imperfectly and have but twenty instead of forty notochordal cells which, moreover, fail to interdigitate. Anterior blastomeres isolated at the 4-cell stage formed small spheres of ecto- derm with about ten notochord cells loosely connected with the outer surface. Posterior blastomeres derived from the same stage formed small spheres of endoderm and mesoderm enclosed within ectodermal cells. Illustrations of such partial development are shown in Fig. 2. Each isolated blastomere apparently divides to form just those cells and tissues it would have formed had it remained a part of a whole embryo. Never was there any sign of development to form a whole larva of half or quarter normal size. It is concluded, therefore, that Conklin's results are in no way invalidated by his failure to remove the dead or injured blastomeres or the egg membrane, and his conclusions concerning the mosaic nature of the ascidian egg are confirmed. Reverberi, on the other hand, has reached conclusions of a very different nature. By an ingenious method of puncturing the chorion and compressing the ovum he succeeded in dividing Ciona eggs into two unequal parts, both parts subsequently being fertilized. His main conclusion, that the ascidian egg should definitely be assigned MOSAIC DEVELOPMENT OF ASCIDIAN EGG 385 to the regulative class, is based upon two discoveries. The segmenta- tion of the small extruded part of the ovum not only commences in a plane at various angles to the first cleavage plane of the ovum proper, but proceeds in a fairly typical manner; while eggs part of which have been extruded are often able to develop to form an apparently perfect tadpole larva. These statements are not questioned, but they are by no means incompatible with the mosaic conception. The development of the small extruded part is normal only in so far as the first three cleavages are in the three planes of space as in the normal egg. Gastrulation and coordinated development do not occur. Since, with the exception of the Nemathelminthes, the seg- FIG. 2. Development of Ascidiella aspersa (24 hours at 17° C.). -4, normal tadpole; B, development of isolated blastomere of 2-cell stage; C, of anterior blasto- mere of 4-cell stage; D, of posterior blastomere of 4-cell stage. mentation of almost all animal eggs is uniform, the first cleavage being vertical, the second vertical at right angles to the first, and the third horizontal, the ability of part of an ascidian egg to follow this same course is alone hardly evidence of strict regulation and any other sequence would be startling. The outstanding exception to such a course, namely the segmentation of the egg of Ascaris, results in a very unstable configuration. The spindle tends to lie along the longest cytoplasmic axis, and the fact that a fertilized part of an ascidian egg obeys this rule of Hertwig in the first stages merely emphasizes its validity. It has no real bearing upon the mosaic or regulative nature of the egg. The other experiments supposedly showing a regulative capacity on the part of the egg concerned the development of diploid eggs, 25 386 N. J. BERRILL part of which had been extruded after fertilization. These developed to form apparently normal tadpoles of reduced size. In no case was it found possible to orientate the eggs with regard to their ooplasmic contents. The only illustrations, however, show two larvae with tails about four-fifths full length and with a trunk region from about four- fifths normal to full size. The one with the smaller trunk has the longer tail, and if the part extruded consisted of part of the endodermal region of the ovum, the peculiarity of this tadpole is accounted for. Similarly, the large-bodied, short-tailed larva is explained if the part extruded contained more of the chordal crescent and less of the endo- dermal ooplasmic region. No growth in size, other than the swelling of notochord cells, occurs in dona until after the tadpole stage has been passed, so that the size of the tadpole is a fairly accurate indication of the size of the egg, and it is evident in these two cases that only a very small part can have been extruded. In any case the peculiarities of the two examples illustrated can be as readily explained on the basis of the mosaic conception as on the assumption that the egg is more regulative than is generally believed. SUMMARY A new method of removing the membranes from ascidian eggs in the mass is described. Essentially the method is the use of proteolytic enzymes. The development of isolated blastomeres is described, confirming Conklin's results that the development of each is strictly partial and that the whole development is a mosaic. It is also shown that the presence of the egg membrane and of injured or dead blastomeres in his experiments in no way invalidated those results. REFERENCES BERRILL, N. J., 1929. Studies in Tunicate Development. Part 1. General physi- ology of development of simple ascidians. Phil. Trans. Roy. Soc. London, B, Vol. 218. BERRILL, N. J., 1931. Studies in Tunicate Development. Part 2. Abbreviation of development in the Molgulidae. Phil. Trans. Roy. Soc. London, B, Vol. 219. CONKLIN, E. G., 1905. Mosaic Development in Ascidian eggs. Jour. Exper. Zool., 2: 145. CONKLIN, E. G., 1906. Does Half of an Ascidian Egg Give Rise to a Whole Larva? Arch.f. Entw.-mech., 21: 727. HALL, E. K., 1931. Puncturing Experiments on the Macromeres of the 8-cell Stage of Rana fusca. Arch. Biol., 42: 279. HORSTADIUS, S., 1928. tiber die Determination des Keimes bei Echinodermen. Acta Zoologica, 9:1. REVERBERI, G., 1931. Studi sperimentali sull'uovo di Ascidie. Publ. Staz. Zool. Napoli, Vol. 11. THE EARLY EMBRYOLOGY OF THE ECHIUROID, URECHIS \Y. \V. NEXYBY DEPARTMENT OF ZOOLOGY, UNIVERSITY OF UTAH INTRODUCTION The eggs of Urechis are being used for many investigations in experimental embryology and a need for a study of the normal develop- ment has'been felt. The object of the present investigation has been : first, to develop a technique whereby either normal or experimentally treated eggs may be stained and preserved for permanent record; second, to describe and figure the early stages of the development; and third, to compare the development of Urechis with that of Thalas- sema mellita (Conn), as described by Torrey, 1903. This paper is a preliminary report, preparations for the further study of the development of the worm having been made. The work was begun in the summer of 1931, at the Hopkins Marine Station at Pacific Grove, California. During the following winter it was con- tinued in the Zoological Laboratory of the University of Utah. It is with pleasure that the writer expresses his appreciation for the aid given him by Professor G. E. MacGinitie, of Hopkins Marine Station, in the collecting of the worms and the rearing of the larvae, and his indebtedness to Dr. Harold Heath, under whose direction the work was carried on. MATERIAL AND METHODS The worm Urechis caupo was described by Fisher and MacGinitie in 1928. It is an inhabitant of the mud flats of the bays and sloughs along the California coast and dwells in U-shaped, tubular burrows. The worms used in this work were obtained from Elkhorn Slough at Monterey Bay and from Morro Bay. They were brought to the laboratory in glass jars booled by ice packs and in the laboratory were kept in well-aerated aquaria. The sexes are separate in Urechis, the eggs and sperm being stored in the six segmental organs, from which they may be easily obtained by gently probing into the ducts of the organs with a thin, smooth, glass probe. The eggs were placed in finger bowls of fresh, filtered sea water with the water about an inch deep. Enough eggs were placed in each bowl to cover the bottom with a single layer. The sperm placed in a watch-glass made a rather thick fluid. This was diluted with about ten volumes of sea water. Insemination was 387 388 W. W. NEWBY effected by the addition of three drops of the sperm dilution to each bowl of eggs. This produced almost 100 per cent fertilization with only an occasional case of polyspermy. The developing eggs were kept in a poorly lighted basement room where the temperature was nearly constant at about 20° C. At this temperature the rate of development was very constant for all the eggs, being remarkably uniform after the 16-cell stage. The formation of the fertilization membrane, the polar bodies, and the first cleavage could be readily observed and samples of the eggs were fixed at each of these stages. After the 2-cell stage samples of the eggs were fixed every fifteen minutes. The eggs were removed with about 2 cc. of the sea water and placed in a vial. They were fixed by the addition, drop by drop, of an equal amount of corrosive sublimate and acetic acid fixative (Guyer No. 18). After ten minutes the fluid was removed with a pipette and the eggs were washed with distilled water for two minutes. This was removed and washing was continued in 50 per cent and then in 70 per cent alcohol for two days. This treatment resulted in a shrinkage of less than 12 per cent of the diameter of the egg. The eggs were stained for one hour and forty minutes in Mallory's phosphotungstic acid hematoxylin (Guyer No. 62). The stain was diluted to about 20 per cent and \vhen used at this concentration, spindles of dividing cells, chromosomes, and cell boundaries were dis- tinct. Little destaining was needed with this stain but when necessary it was done in a weak alkaline alcohol, the treatment being continued until the eggs were a light pink when observed with the low power of the microscope and with reflected light. The eggs were then de- hydrated, cleared, and mounted in balsam. In all processes after the washing the eggs were not handled, but were placed in a small dish and all fluids were added or withdrawn with a pipette. The balsam was added after most of the xylene had been removed and small drops of this balsam with the eggs were easily mounted. Contrary to a common opinion, the technique of staining and mounting is not laborious and the writer suggests that often much might be learned if experimentally treated eggs were thus preserved for closer examination. FERTILIZATION AND THE FIRST CLEAVAGE The process of fertilization has been described by Tyler (1931). The process may be outlined as follows: first, the attachment of the sperm; second, the appearance of a cone on the surface of the egg immediately below the point of attachment; third, the fertilization EARLY EMBRYOLOGY OF URECHIS 389 membrane appears on the surface of the egg and begins to lift; fourth, the sperm entrance cone of the egg begins to enlarge and the indenture of the egg begins to round out; fifth, the sperm head enters the egg. This process takes six and a half or seven minutes at 20° C. At a time about twelve minutes after the sperm attachment the nucleolus of the egg disappears and at fifteen minutes the nucleus or germinal vesicle disappears. At about thirty to thirty-five minutes the first polar body is extruded and about ten minutes later the second is extruded (Plate I, Fig. B). In many cases the first polar body divides so that three appear. About one hour and ten minutes after sperm attachment the egg elongates and the furrow of the first cleavage ap- pears and by one hour and twenty minutes the egg is in the two-cell stage. The second, third and fourth cleavages take place at half-hour intervals. This outline was made from the work of Tyler, confirmed by the writer's observations on several groups of eggs at 20° C. The unfertilized egg appears to have a marked polarity as the bottom of the indenture seems to mark one pole (Plate I, Fig. A). Tyler considers this as the animal pole, for in 85 per cent of the eggs observed by him the polar bodies appear at a point less than 10° from the egg surface, which before fertilization was at the bottom of the indenture. Taylor (1931) however, states: "The immature egg apparently has no mark of polarity. Neither the eccentricity of the germinal vesicle nor the position of the one or more depressions could be correlated with the subsequent place of extrusion of the polar bodies, as will be stated in some detail in later paragraphs." The position of the axis of the spindle of the first cleavage and therefore the plane of the cleavage is determined by the axis of the egg and some other point. Tyler has found that the plane of the first cleavage passes within ten degrees of the point of entrance of the sperm in 71 per cent of 178 eggs observed. The high correlation between point of sperm entrance and the position of the first cleavage furrow is confirmed by the work of Taylor. / DEVELOPMENT TO THE SIXTY-FOUR-CELL STAGE From the two-cell to the sixteen-cell stage. — The cleavage furrow first appears at the animal pole of the egg and spreads to the vegetal pole, cutting the egg into two equal-sized cells. These are at first almost spherical, being only slightly flattened on their adjacent sides. During the resting stage they further flatten and elongate and at this time the spindles of the second cleavage may be seen (Plate I, Fig. C). The axes of the spindles are parallel, and in stained preparations the chromosomes may be indistinctly seen and around the ends of the spindles red granules are visible. 390 W. W. NEWBY The second cleavage furrow also starts at the animal pole of each cell and then divides it into two. The egg is then in the 4-cell stage, the cells being equal in size and flattened on their adjacent sides (Plate I, Fig. D). The long axes of the cells are parallel with the axis of the egg and a distinct, equal-sided segmentation cavity lies between them. In Thalassema, Torrey describes the two spindles of the second cleavage as being not quite parallel so that the A and C cells are in contact at the upper pole and the B and D cells are in contact at the lower pole, the furrow at the upper pole being a segment of the first cleavage furrow. This method of cleavage was not observed either in the living or preserved eggs of Urechis. Preceding the third cleavage the four cells elongate vertically and become somewhat more rounded horizontally so that the area of contact between the cells decreases. As the spindles form with their axes coinciding with the long axes of the cells, the upper ends of the cells rotate to the right, clockwise; and the lower ends rotate to the left, counter-clockwise (Plate I, Fig. E). The resulting cleavage is thus spiral and dexiotropic. The upper moiety is the first quartet of micromeres and the cells are slightly smaller than the lower macromeres (Plate I, Fig. F). There is a rather brief resting period during which the cells become flattened on their adjacent sides and the whole egg becomes more spherical. Both the macromeres and micromeres then elongate, and their upper ends rotate to the left, counter-clockwise (Plate II, Fig. G). The resulting cleavage is therefore spiral and leiotropic and is some- what unequal. The first quartet now consists of eight cells, the lower cells, la2, etc., being somewhat smaller than the parent cells and the macromeres having divided to produce the second quartet, 2a, etc., the cells of which are the same size as the daughter cells of the first PLATE I All drawings were made with a camera lucida and have the same magnification. Figures A and B were drawn from living eggs. All other figures were drawn from fixed material and are therefore somewhat smaller. FIG. A. The unfertilized egg showing the indenture or depression. FIG. B. The fertilized egg with the polar bodies extruded, the fertilization membrane not being shown. FIG. C. A semi-polar view of the 2-cell stage showing the parallel spindles of the second cleavage. FIG. D. A polar view of the 4-cell stage showing the equal-sized cells and the equal-sided segmentation cavity. FIG. E. A lateral view of the 4-cell stage as the cells divide to form the 8-cell stage. FIG. F. A lateral view of the 8-cell stage. EARLY EMBRYOLOGY OF URECHIS 391 PLATE I 392 W. W. NEWBY quartet (Plate II, Figs. //, /, and J). The daughter cells of the first quartet are the prototroch cells. The egg is radially symmetrical and during a rather protracted resting stage the cells become flattened on their adjacent sides. In the 8-cell stage the first quartet cells are to the right of the macromeres. In the 16-cell stage, however, because of the pressure of the newly-formed cells, the stem cells of the first quartet are pushed to the left of the macromeres. The Iwenty-four-cell stage. — The next division is unequal and takes place simultaneously in the stem cells of the first quartet and in the macromeres (Plate II, Figs. K and L). It is dexiotropic and results in the formation of eight new cells, all markedly smaller than the parent cells and slightly smaller than the daughter cells of the pre- ceding division. The egg is now in the 24-cell stage and consists of the first quartet of twelve cells, four newly formed (la1-2, 161-2, etc.), the second quartet of four cells, the newly-formed third quartet of four cells (3a, 3b, etc.), and the four macromeres. During the rather brief resting stage the cells again flatten on their adjacent sides and by their pressure again push the stem cells of the first quartet to the right of the corresponding macromeres. The thirty-two-cell stage. — This stage is produced by a division of the prototroch cells, la2, 162, etc., and the second quartet cells, 2a, 2b, etc. Both divisions are dexiotropic and about equal, but the division of the prototroch cells precedes that of the second quartet cells, the first being in a late stage of division when the latter is in an early stage (Plate III, Fig. M}. In the resting period following this division the egg undergoes a change in shape. The cells of the first and second quartets become somewhat flattened and surround an enlarged seg- mentation cavity. The four macromeres remain in contact and have their adjacent sides almost vertical, the whole egg taking on a "mush- room shape" with the four macromeres appearing as four conspicuous cells at the base (Plate III, Figs. TV and 0). This change in shape marks the second conspicuous deviation in the development of Urechis as compared with that of Thalassema as PLATE II FIG. G. A lateral view of the 8-cell stage as the cells divide to form the 16-cell stage. FIG. H. A lateral view of the early 16-cell stage. FIG. /. A polar view of the early 16-cell stage. FIG. J. A lateral view of the 16-cell stage during the resting period. The extent of the segmentation cavity is indicated by the dotted line. FIG. K. A lateral view of the 16-cell stage as the cells divide to form the 24-cell stage. FIG. L. A lateral view of the 24-cell stage. EARLY EMBRYOLOGY OF URECHIS 393 PLATE II 394 W. W. NEWBY described by Torrey. The order of the cleavages has been the same and the general arrangement of the cells has been the same, but he figures the macromeres as being proportionally larger and he neither mentions nor figures the "mushroom shape" which characterizes the Urechis egg from this stage on to the rotating larval stage. The thirty-six-cell stage. — This stage is formed by a leiotropic division of the macromeres to form the fourth quartet (Plate III, Figs. N and 0). This division is unequal, the fourth quartet cells being even smaller than the cells of the second and third quartets. They are also pushed inward and flattened so that they form long cells extending radially into the egg and presenting a small surface to the outside. This division is similar to that of Thalassema in that there is nothing in either the time of cleavage, the shape, size, or positions of the cells to distinguish the 4d or M cell from the other fourth quartet cells. The forty-cell stage. — This stage is formed by a leiotropic division of the stem cells of the first quartet, la1-1, 161-1, etc. (Plate III, Figs. P and Q}. The division is very unequal and the smaller polar cells, la1-1-1, 161-1-1, etc. form the apical rosette; and the larger daughter cells, la1-1-2, 161-1-2, etc. form the stem cells of the cross (Plate III, Figs. P and <2). The rosette cells lie in a depression between the cross cells. The cross cells also become rounded upward and there is a shifting of the la1-2, 161-2, etc. cells to a position between the outer ends of the cross cells. These cells, the la1-2, etc. may then be called the inter- girdle cells. The terms "apical rosette," "cross cells," and "inter- girdle cells" are used because they have the same origin and relative position as similar cells in Thalassema, their ultimate fate not being as yet known in Urechis. The relatively smaller size and the position of the rosette cells in a depression is in contrast with the larger size and more superficial position of the cells in Thalassema. In this respect the apical rosette of Urechis resembles that of Podarke. PLATE III FIG. M. A lateral view of the 24-cell stage as the cells divide to form the 32-cell stage. FIG. N. A lateral view of the 32-cell stage as the macromeres divide to form the 36-cell stage. FIG. 0. A lateral view of the 36-cell stage, showing the characteristic "mush- room" shape of the egg. FIG. P. A polar view of the 36-cell stage showing the division of the micro- meres to produce the 40-cell stage. FIG. Q. A polar view of the 40-cell stage showing the apical rosette. FIG. R. A lateral view of the 40-cell stage as the cells divide to form the 48-cell stage. EARLY EMBRYOLOGY OF URECHIS 395 PLATE III 396 W. W. NEWBY The forty-eight-cell stage. — This stage is formed by an almost simultaneous, leiotropic division of the third quartet cells and the intergirdle cells, the division of the intergirdle cells generally preceding that of the third quartet (Plate III, Fig. K). The division of the intergirdle cells is somewhat unequal, the upper cells, lying between the cross cells, being a little larger. The division of the third quartet is also unequal, the upper cells in this case being slightly smaller and rather deeply imbedded between the surrounding prototrochal and second quartet cells. The fifty-six-cell stage. — This stage is formed by simultaneous, leiotropic, and equal divisions of the prototrochal cells, la2-1, la2-2, etc. (Plate IV, Fig. S}. The order of these last two cleavages is different in Thalassema and Urechis. In Thalassema there is a division of the third quartet followed by an almost simultaneous division of the intergirdle and prototrochal cells. In Urechis the first division takes place simultaneously in the intergirdle and third quartet cells and the second is in the eight prototrochal cells only. The sixty-four-cell stage. — This stage is formed by a division of the four cross cells, la1-1-2, etc., and the lower moiety of the second quartet cells, 2a2, etc. (Plate IV, Fig. T}. These divisions are important as they mark the end of the purely radial cleavages and initiate the bilateral and morphogenic cleavages. The divisions follow exactly those in Thalassema and the symmetry is probably the same. In two adjacent cells of the cross, If1-1-2 and Id1-1-2, a meridional, equal division occurs which is shortly followed by an unequal, leiotropic division of the other two cross cells, la1-1-2 and 1&1-1-2. This later division is almost meridional, the axes of the spindles being inclined only slightly to the left. The smaller cells are budded off toward the center and lie against the rosette cells (Plate IV, Fig. £7). Bilaterality is now clearly established. However, the relation between the plane of the first cleavage and this plane of symmetry; and the relation be- tween this plane of symmetry and that of the larva has not been determined at this time. PLATE IV FIG. S. A lateral view of the 48-cell stage as the cells divide to form the 56-cell stage. FIG. T. A lateral view of the 56-cell stage as the cells divide to form the 64-cell stage. FIG. U. A polar view of the 64-cell stage showing the bilateral symmetry. FIG. V. The right side of the 64-cell stage showing the 2c2-2 cell as being larger than the 2c--1 cell. This is also the case in the division of the 2a2 and 2b2 cells. FIG. W. A posterior view of the 64-cell stage showing the 2d--- cell as being smaller than the 2d--1 cell. This marks the posterior side of the larva. EARLY EMBRYOLOGY OF URECHIS 397 n.z V w 398 W. W. NEWBY of Cell-Lineage of Urechis caupo & 16 3Z 48 S6 Aprca/ ffoseffe t.ie.i Left Anterior /./Z.2 Arm of CroSS I nt er g i rdle. Pro to troch Second Quartet Third Quartet Fourth Quartet PLATE V A table which shows the order of the first eleven divisions of the egg of Urechis. As the cleavages in all four quadrants are the same at this stage, only the anterior, A and B, quadrants are shown. EARLY EMBRYOLOGY OF URECHIS It has been assumed that the first, equal, meridional divisions of the cross cells take place in the posterior, C and D, quadrants as they do in Thalassema. This is further confirmed by the mode of division of the second quartet cells. In the 2a2, 2b2, and 2c~ cells the unequal, leiotropic divisions have the smaller cells budding off upwardly to become deeply imbedded between the surrounding cells. The 2d~ cell also divides unequally but with the smaller cell downward, over- lying the 4d or M cell. This cell has been designated by Tread well for Podarke, and Torrey for Thalassema as the .r1-2 cell. "Treadwell ('97 and '01) in his discussion of cell-homologies calls attention to the fact that this cell has the same origin in all the forms of annelids and molluscs that have been studied." (Quoted from Torrey.) The identity of the four quadrants and the plane of sym- metry is thus established before the division of the M cell. The sixteen prototrochal cells have a rather superficial position on the surface of the egg and do not extend inward to the segmentation cavity. The fertilization membrane persists around the entire egg but at the animal pole has a depression that brings it in contact with the apical rosette cells. Development to the 64-cell stage takes five and one-half to six hours at 18° to 20° C. Cilia develop on the prototroch rather rapidly between six and a quarter and six and one-half hours and almost immediately begin to beat, causing the egg to rotate. The longer •apical cilia develop later. CONCLUSION Except for the shape of the egg and differences in the relative sizes of some of the cells the egg of Urediis at the 64-cell stage is identical with that of Thalassema and similar to the eggs of other annelids with equal cleavages. BIBLIOGRAPHY FISHER, \V. K., AND MAcGiNixiE, G. E., 1928a. A New Echiuroid Worm from California. Ann. and Mag. Nat. Hist., Ser. 10, 1: 199. FISHER, W. K., AND MACGINITIE, G. E., 19286. The Natural History of an Echiu- roid \Yorm. Ann. and Mag. Nat. Hist., Ser. 10, 1: 204. GUYER, M. F., 1930. Animal Micrology. Third Ed. Chicago. TAYLOR, C. V., 1931. Polarity in Normal and Centrifuged Eggs of Urechis caupo Fisher and MacGinitie. Physiol. Zool., 4: 423. TORREY, J. C., 1903. The Early Embryology of Thalassema mellita (Conn). Ann. N. Y. Acad. Set., 14: 165. TREADWELL, A. L., 1897. The Cell Lineage of Podarke obscura. Zool. Bull., 1: 195. TREADWELL, A. L., 1901. The Cytogeny of Podarke obscura Verrill. Jour. Morph., 17: 399. TYLER, ALBERT, 1931. The Production of Normal Embryos by Artificial Partheno- genesis in the Echiuroid, Urechis. Biol. Bull., 60: 187. POLYVITELLINE EGGS AND DOUBLE MONSTERS IN THE POND SNAIL LYMN^A COLUMELLA SAY CHARLES P. WINSOR AND AGNES A. WINSOR DEPARTMENT OF BIOLOGY, SCHOOL OF HYGIENE AND PUBLIC HEALTH, JOHNS HOPKINS UNIVERSITY The occasional occurrence of gastropod eggs containing more than one embryo is more or less familiar to workers with these forms. Such embryos are normally separate; but occasionally conjoined twins are found, and more rarely triplets, and instances of as many as five conjoined embryos have been reported. The literature on this subject has been well reviewed by Pelseneer (1920) and Crabb (1931), so that no extended historical discussion seems necessary. Our purpose is merely to present certain material which seems to us useful in supple- menting and perhaps in modifying the conclusions of other workers. The material here presented consists of the total egg production throughout life of a population of (initially) 400 snails of the species Lymnxa columella Say. The wild parents of these snails were collected in two ponds in the vicinity of Baltimore, designated here as the Falls Road pond and the Boyce Avenue pond. In addition to these wild ancestors of known origin, two snails isolated from laboratory aquaria furnished eggs for this experiment; nothing is known concerning their origin. These animals were isolated in the laboratory in finger bowls with about 150 ml. of spring water, fed with leaf lettuce, and their eggs collected daily. The eggs so obtained were allowed to develop for about a week, at which time healthy-appearing clutches were selected for the experimental population. These eggs were removed from the capsule and placed in finger bowls with spring water. Each clutch of eggs provided twenty eggs, which were distributed among six finger bowls, one at density ten and five at density two per bowl. Leaf lettuce was used for food. Records of egg production were kept for each dish throughout the lives of the animals. Early in the experiment appreciable numbers of polyvitelline eggs appeared and were regularly entered in the records. Occasional double monsters were found; such eggs were saved and their development followed until the death of the embryos. Table I summarizes the results obtained. The facts regarding the production of polyvitelline eggs are sum- marized in Fig. 1. This chart shows the percentage of polyvitelline 400 POLYVITELLINE EGGS IN LYMN^EA 401 eggs laid in each finger bowl in the experiment. The parentage of the snails in each finger bowl is also shown, together with the geographical origin of the parents. It is clear from an examination of the chart that there are very marked differences in the proportions of polyvitelline eggs occurring in different finger bowls; and that there is clear indication that the offspring of wild parents from the Falls Road pond, especially those TABLE I Production of Polyvitelline Eggs by Offspring of Different Wild Parents Per- Eggs having given number of vitelli Total cent- Wild ancestor No. of dishes No. of snails Total eggs poly- vitel- line age poly- vitel- 9 and eggs line 2 3 4 5 6 / 8 over eggs 1 5 8 3,570 58 1.62 47 9 2 — — — — — . 2 17 42 34,199 1,453 4.25 1,079 272 63 15 4 8 3 9 4 16 35 37,761 1,631 4.32 1,166 324 89 20 13 6 5 8 6 2 9 3,031 146 4.82 103 36 4 1 1 1 — - — • 8 12 22 9,857 144 1.46 110 23 4 3 2 1 — 1 Totals 1-8 52 109 88,418 3,432 3.88 2,505 664 162 39 20 16 8 18 Falls Road 12 12 23 14,508 96 0.66 85 8 2 — 1 — — — 15 6 14 9,647 81 0.84 65 12 — 1 — — 3 — 16 12 29 17,672 244 1.38 220 17 1 3 1 1 — 1 17 5 10 7,359 106- 1.44 83 14 4 1 3 — — 1 Totals 12-17 35 76 49,186 527 1.07 453 51 7 5 5 1 3 2 Boyce Ave. A2 7 10 13,411 227 1.69 209 8 3 3 1 — — 3 , A3 6 11 13,369 73 0.55 63 3 7 — — — — 1 Notes: Column 1 shows the wild ancestor. Numbers 1 to 8 inclusive were taken from the Falls Road pond; numbers 12 to 17 from the Boyce Avenue pond. Numbers A2 and A3 were isolated from a laboratory aquarium culture; their wild origin is unknown. from Nos. 2 and 4, laid a considerably higher proportion of polyvitelline eggs than those from the Boyce Avenue pond. There seems to be reason for supposing that these differences are the expression of some inherited character. This conclusion differs from that of Crabb, who holds that there is no indication that the production of polyvitelline eggs is influenced by hereditary factors. The data given by the Crabbs (1927) seem to us, however, consistent with the notion that inheritance does play a 26 402 C. P. WINSOR AND A. A. WINSOR part in the phenomenon. Neither their data nor ours, however, are adequate to suggest any Mendelian mechanism. /Jncestor A 3 • x> T •< L A 2 Boyce /4t 17 16 15 12 •'••<• • e ••;:•- 6 4 • •»•! ••• • • • • 2 ..../...•... • L ' i ; i i_Lj i i i i i 8 FIG. 1. Chart showing proportion of polyvitelline eggs in different finger bowls, segregated by ancestry. A paper by Tur (1910) (not cited by Crabb) is of interest in this connection. He found, in Philine aperta, that there were distinct local races, some of which produced polyvitelline eggs with some frequency and regularity, while animals from other localities rarely or never laid such eggs. EXPLANATION OF PLATE FIGS. 1-3. Lateral fusion with separate shells. Three views. Parent No. 27, grandparent No. 2. FIGS. 4-5. Dorsal fusion with separate shells. Two views. Parent No. 7, grandparent No. 4. FIG. 6. Visceral masses fused, single shell. Parent No. 12, grandparent No. 4. FIGS. 7-8. Triple monster, shell fused. Parent No. 26, grandparent No. 2. FIG. 9. Dorsal fusion, single shell. Parent No. 11, grandparent No. 4. POLYVITELLINE EGGS IN LYMN/EA 403 404 C. P. WINSOR AND A. A. WINSOR DOUBLE MONSTERS Double monsters in considerable numbers were observed in our material. Drawings of typical specimens (made for us by the labora- tory artist, Mr. Arthur Johansen) are shown in Plate I. As regards the anatomical details of these monsters, we can add nothing of importance to Pelseneer's account. So far as our observations went, they agreed with his. It will be noted that the wild ancestors giving rise to these double monsters came almost entirely from the Falls Road pond, as would be expected in view of their apparently certain origin through fusion of embryos. Actually, of 32 double and triple monsters, 28 (of which one was triple) had a Falls Road origin ; one was derived from Boyce Avenue stock; and three (one triple) from laboratory stock of unknown origin. SUMMARY Data are presented regarding the production of polyvitelline eggs and double monsters through the entire life of a group of Lymnxa columella. It is held that the data show that hereditary factors are concerned. LITERATURE CITED CRABB, E. D., AND R. M. CRABB, 1927. Polyvitelliny in Pond Snails. Biol. Bull, 53: 318. CRABB, E. D., 1931. The Origin of Independent and of Conjoined Twins in Fresh Water Snails. Roux' Arch. Enlw. mech., 124: 332. PELSENEER, P., 1920. Les variations et leur heredite chez les mollusques. Mem. Ac. Roy. Belg., Cl. Sci., Coll. in— 8°, 5: 1 (pp. 308-339 deal with double monsters) . TUR, JAN, 1910. Sur les pontes anomales chez Philine aperta L. Roux' Arch. Entw. mech., 30: 357. Detailed bibliographies will be found in Crabb (1931) and Pelseneer (1920). STUDIES ON AMPHIBIAN METAMORPHOSIS X. HYDROGEN-ION CONCENTRATION OF THF BLOOD OF ANURAN LARWE DURING INVOLUTION1 O. M. HELFF DEPARTMENT OF BIOLOGY, UNIVERSITY COLLEGE, NEW YORK UNIVERSITY INTRODUCTION The metamorphosis of anuran larvse, aside from the striking embryological and growth phenomena exhibited, presents perhaps an even more interesting series of degenerative changes, involving cellular autolysis and various histolytic effects, which in some cases results in the total atrophy of an organ. The range of tissues so affected and the degree to which they degenerate is subject to considerable variation, however. Thus we find the tail, the gills, and the areas of opercular integument through which the fore-limbs emerge undergoing complete atrophy resulting in the total resorption of the tissues con- cerned, with the exception of the gills which are represented in the adult by small spherical bodies, the "Kiemenreste," whose function, if any, is unknown. On the other hand, organs such as the intestines, stomach, and pancreas which undergo a reorganization in the develop- ment of the adult condition, are subject to a lesser degree of atrophy, although the amount of cellular histolysis is very great. Contrasted to the above two categories of degenerating tissues, we find those in which relatively little degeneration occurs as in the tissues of the jaws during the shedding of larval teeth, the early degenerative changes of integument in the formation of the tympanic membranes, and the general atrophy of the musculature of the back, the latter being subject to approximately a 38 per cent reduction in volume during metamorphosis. Although relatively little, if any, information is available regarding the histological changes operating in the case of the jaw tissues, pancreas, and gills, more definite data can be found in relation to the other organs and tissues as cited above. For such information the reader is referred to the papers of Barfurth (1887), Bataillon (1891), Bradley (1922), Clausen (1930), Helff (1926) (1928) (1931a), Helff and Clausen (1929), Kuntz (1922) (1924), Lindeman (1929a), Mercier (1906), Morse (1918), and Van Der Jagt (1929). 1 The expenses entailed in this investigation have been partially met by a grant of money from the Bache Fund of the National Academy of Sciences. 405 406 O. M. HELFF Although the endocrine glands, notably the thyroid and anterior pituitary, are generally conceded to furnish the initial stimuli for amphibian metamorphosis, relatively little information is available in respect to their mode of action. It seems quite certain, however, that certain developments and degenerative processes are not brought about by the endocrine hormones acting directly on the tissue con- cerned through the medium of the blood stream. Thus the dedifferen- tiation of integument to form the tympanic membranes is due to con- tact of the integument with the developing annular tympanic cartilage and columella (Helff, 1928 and 1931), while the total histolysis of opercular integument as it occurs in the formation of the fore-limb perforations is the result of integumentary contact with the atrophying tissue of the gills (Helff, 1926). On the other hand, the development and degeneration of tissues following transplantation to foreign regions strongly suggests the operation of metamorphic influences brought into play through the blood supply to the transplant. The development of the tongue (Helff, 1929), the columella (Helff, 1931a), the dermal plica? (Helff, 19316), the fore-limbs (Helff, 1926), the hind-limbs (Schubert, 1926), the adult pigment pattern (Lindeman, 1929a) and the nictitating membrane (Lindeman, 19296) are all cases in point in which the developmental stimulus is through the blood stream. The degenerative changes which have been shown to occur normally upon transplantation to foreign regions include the histolysis of tail integu- ment (Lindeman, 1929a, and Clausen, 1930), and the histolysis of tail muscle (Helff and Clausen, 1929). Indications also point to the probability that future work will add the atrophy of the gills, intestines, and pancreas to this latter group. While it is possible that all of the developmental and atrophic changes which occur during amphibian metamorphosis may be due to hormonic influences in the blood stream, it seems rather improbable that such will ultimately prove to be the case in relation to the partial or total histolysis of larval organs. Due to the considerable mass of work that has been done on various tissues in widely divergent animal groups in relation to the biochemical factors associated with and favorable to the autolysis of living tissues "in vivo" and "in vitro," one cannot escape the implication that the histolysis of larval organs and tissues in the amphibian is due to a more immediate and plausible causative factor which is the result, primarily, of course, of the meta- morphic hormones. An obvious factor and one that has been shown to be closely correlated with the autolysis and atrophy of a wide variety of tissues in other classes of animals, is that of the hydrogen-ion concentration of the medium in which the autolysis occurs. Briefly, AMPHIBIAN METAMORPHOSIS 407 it may be stated that in general the increasing of the hydrogen-ion concentration of the medium above a certain level either tends to activate the proteolytic enzymes concerned in autolysis or else brings about changes in the tissue proteins themselves which render them digestible by the omnipresent enzymes. The possibility of a lowering in the pH of the blood during am- phibian metamorphosis has been suggested by many writers as a likely explanation for many of the histolytic processes which occur. Thus Barfurth (1887), Bataillon (1891), Mercier (1906), Morse (1918), and Bradley (1922) all suggest the probability that the tissues of the larval tail undergo autolysis in response to a localized increase in hydrogen-ion concentration, the result of a partial or total occlusion of the dorsal aorta resulting from pressure exerted by the rapidly growing urostyle. Although their assumption of a lowered pH as the immediate causative factor may have been correct, the importance of the developing urostyle in this connection was found to be negligible as shown by subsequent work. In this connection it may be stated that the writer (1930) was able to show that the process of tail resorp- tion was not inhibited in any way in the complete absence of the urostyle, the latter having been extirpated while still in the anlage stage prior to the onset of larval involution. It was also shown by Lindeman (1929a) that tail integument previously transplanted to the back underwent normal histolysis during metamorphosis at the same time the integument on the tail autolyzed. The same result was obtained by Helff and Clausen (1929) working with tail muscle. A consideration of the foregoing brief survey regarding the problem of autolysis during larval involution clearly suggests that the question of hydrogen-ion concentration of the blood before, during, and after metamorphosis becomes a problem of considerable interest and im- portance. As far as the writer is aware, there is no previously pub- lished work relating to the hydrogen-ion concentration of larval blood either before or during metamorphosis. Numerous tests have been made on the blood of the adult amphibian, however, in relation to various environmental and experimental factors. In this relation the reader is referred to the papers of Rohde (1920), Hertwig (1927), Kamm (1930), et al. The present paper, therefore, relates the results of hydrogen-ion tests as made on the blood of non-metamorphosing larvae, larvae in various stages of involution, and young frogs shortly after the completion of metamorphosis. MATERIALS AND METHODS The initial stock used for all hydrogen-ion tests were large second- year Rana clamitans larvae obtained at regular intervals from a common 408 O. M. HELFF source near Cincinnati, Ohio, during the winter and spring months of 1930 and 1931. The selection of Rana clamitans larvae was based on two points both favorable to the work at hand ; first, in that the large size of the tadpoles insured a sufficient quantity of blood for testing purposes and second, that due to the neotenous nature of the species and the fact that they were second-year larvae, normal metamorphosis could be induced when placed in water at room temperatures during the spring months especially. Shipments of larvae, when received, were first placed in a large tank in order to acclimate them to the new water medium and also to allow the less hardy to die out. Under laboratory conditions larvae also frequently tend to develop edema and red-leg, cases of which were always eliminated during this preliminary quarantine period. In testing the normal or Stage 1 larvae, the tad- poles were first isolated in individual bowls of water for a period of at least three days prior to the hydrogen-ion testing. In obtaining the various states of metamorphosis tested, Stage 1 larvae were likewise isolated in individual bowls of water and allowed to metamorphose normally. In both cases no food was given to the animals. The time required at room temperatures for complete metamorphosis to occur following the onset of involution varied from four to seven weeks. The method employed in obtaining samples of blood for testing purposes was as follows: The larvae or newly metamorphosed frogs were first fastened, dorsal side down, to a moistened pad of cloth by means of rubber bands. It may be said in this connection that the elimination of the use of an anaesthetic in the securing of the blood samples was thought to be advisable in that the pH of the blood might be altered by such a procedure. Two incisions were now made through the integument of the anterior ventral surface; one running from side to side just anterior to the level of the heart, and the other along the mid-ventral line from the middle of the first incision anterior to the rim of the lower jaw. The two flaps of integument were now pulled aside, laterally, so as to expose the general region anterior to the heart. In the case of larvae this exposes the pericardial sac, while in the young frog it is first necessary to resect certain underlying muscular layers before the pericardium is brought to view. Following the exposure of the pericardial sac, the latter is cut at a point to expose the truncus arteriosus. In so doing, care must be exercised not to cut too far posterior or else premature bleeding may occur due to injury of the anterior abdominal vein. The cutting of the pericardial sac liberates, especially in normal larvae, a considerable amount of lymph which quickly flows out, however. The lymph is now carefully re- moved from the cavity so prepared immediately anterior to the heart. AMPHIBIAN METAMORPHOSIS 409 The truncus arteriosus is now cut through using a small Noyes' iris scissors and blood quickly flows out filling the prepared cavity anterior. The blood is quickly drawn up into a pipette which has been previously rinsed with a 2 per cent solution of sodium oxalate. The small amount of oxalate solution adhering to the walls of the pipette is sufficient to prevent the clotting of the blood. The blood is now immediately transferred to a small vial and the hydrogen-ion concentration test quickly made. The determination of hydrogen-ion concentration was made by means of the quinhydrone method. In this regard it seems advisable to mention some of the more important technical points found advan- tageous in securing greater accuracy with this method. Care must be exercised in the cleaning of the blood vials. The routine followed was to first thoroughly wash the vials and then rinse several times with hydrochloric acid followed by alcohol and distilled water. The vials were then allowed to dry thoroughly before using. Due to the small quantity of blood tested, very little quinhydrone had to be added. Solution of the latter in the blood was facilitated by gently tapping the vial for a few seconds. The platinum leaf electrode was next inserted to the bottom of the vial. By bending the platinum leaf at right angles to the supporting capillary glass tube, the former could be made to lie flat on the bottom of the vial and thus insure contact with the blood on both surfaces. Care should also be taken to care- fully cleanse the test electrode between any two consecutive tests by rinsing thoroughly in hydrochloric acid, alcohol, and distilled water. The electrode should be dried carefully before using again. The agar bridges employed were of the usual type except that quite small tubing was necessary due to the small diameter of the blood vials. Care should be taken to cleanse the end of the agar bridge which has been immersed in the blood of the test vial, between individual tests. The quicker the e.m.f. is determined the better, while use of a room with little variation in temperature also makes for greater accuracy, espe- cially where comparative results are desired. Finally, it is a good practice to standardize the set at frequent intervals by checking against a standard solution of 1/20 molar potassium acid phthalate which should give a pH of 3.98 at 25° C. In the present work such a standardization test was made between every two consecutive blood tests. RESULTS Blood pH of Normal, Non-metamorphosing Larvae The larvae used for the Stage 1 tests were normal tadpoles in all respects which would not have begun metamorphosis for at least a 410 O. M. HELFF -ai w - ' 8 a PQ "S- =: a ^ 01 5 •c V V c O SB MO o Z 2f 'C K o. ffi a si M 'S Leng ffi B. O GO — f O so "5 CN '— i •— iT-icNT-irsi»-i»— T-H -H ^ 00 O *-i r^ ^^ 0s *•"' '"' CO r^l '—i £""• ON ON CO ^D ^^ 0s tN»— ir^r-)T— ICNCN— H— H ^— c i— i cv) r>i T-I -t 10 -f irj to »* •«*< "~> ""> -t -f H^HI — i ON ON ^— < < — ION'— 'ON^-^^^ON^H^— i^^^— »ON*— I |xxxxxxxxxxxxxxxxxxxxxxxxx ON , oc oc r^ oc cc oc t— O vC oc 20 i~- r— bo C - O -4-J X bo C y X AMPHIBIAN METAMORPHOSIS 411 month. Their hind-limb lengths varied from 8 to 17 mm., while in total length the variation was between 92 and 105 mm. The weight of such individuals varied from 6.7 to 8.6 grams. Of the various stages tested, Stage 1 larvae were most favorable in respect to the quantity of blood obtainable. Although some variation was experienced in this regard, a sufficient amount was always present. On the average about one-tenth of a cubic centimeter could be secured per animal. Columns 1, 2, and 3 of Table I present the essential data for this group of tests. It will be seen that the range of pH was not great, being from a minimum of 7.43 to a maximum of 7.54 and repre- senting a variation, therefore, of but 0.11 of a pH unit. The average pH of the twenty-five larvse tested proved to be 7.50 ± .00459, while the standard deviation amounted to .034 ± .0032 (see Table II). We can thus consider a pH of 7.50 as typical for the blood of normal non-metamorphosing larvae of the particular species being tested. TABLE II Means, Standard Deviations, and Probable Errors Stage Mean Standard deviation 1 7.50 ± .00459 .034 ± .0032 2 7.39 ± .00650 .048 ± .0046 3 7.27 ± .00634 .047 ± .0045 4 7.20 ± .00607 .045 ± .0043 5 7.18 ± .00648 .048 ± .0046 Blood pH During Metamorphosis The various stages of metamorphosis tested were obtained by allowing natural metamorphosis to take place. The isolation of Stage 1 larvae in individual bowrls of water at room temperature acted as a stimulus to transformation with the result that the onset of meta- morphosis usually began in about three weeks following their transfer from the stock tanks of cold running water. All cases of edema and red-leg which occurred during metamorphosis were discarded. During the early stages of involution the mortality was relatively low. The acquiring of completely metamorphosed larvae, however, involved a mortality of over 80 per cent, necessitating the isolation of over one hundred and twenty-five Stage 1 larvae in the ultimate securing of twenty-five Stage 4 animals. To obtain twenty-five Stage 5 animals, it was necessary to isolate a still larger number of Stage 1 larvae, since the mortality was even greater in that about one-fourth of the Stage 4 animals died within three to four weeks following the attainment of 412 O. M. HELFF that stage. In all, some four hundred and thirty Stage 1 larvae were isolated in securing twenty-five cases each of Stages 1, 2, 3, 4, and 5. Stage 2. — Stage 2 larvae were tadpoles which had attained early metamorphic changes characterized by a definite acceleration in fore- and hind-limb growth, the loss of the full-bellied appearance typical of Stage 1 larvae, a slight anterior-posterior atrophy of the tail with pronounced histolysis of the dorsal and ventral finny portions, and a general darkening of integumentary pigmentation. The hind-limbs varied between 17 and 22 mm. in length, and total length between 88 and 100 mm. The weight varied between 4.6 and 6.4 grams. The fourth, fifth, and sixth columns of Table I present the essential data for Stage 2 tests. It will be noted that the range in pH was somewhat greater than in the case of the Stage 1 tests, being from 7.28 to 7.47 or a variation of 0.19 of a pH unit. The average pH of the twenty-five larvae tested amounted to 7.39 ± .00650, the standard deviation being .048 ± .0046. The results are clear, therefore, in that a definite change in pH from 7.50 to 7.39, representing a drop of 0.11 of a pH unit, occurs during this early stage of metamorphosis. Stage 3. — Stage 3 larvae represented a more pronounced stage of larval involution. At this stage the left fore-limb had just emerged, while the hind-limbs were considerably increased in length. The tail had undergone still further atrophy and the intestines were rapidly being transformed into the adult type. The hind-limb lengths varied between 23 and 37 mm., and total length between 82 and 97 mm. The drop in weight was especially pronounced, the larvae varying between 2.7 and 5.1 grams. The ninth column of Table I presents the hydrogen-ion concentra- tions determined for Stage 3 larvae. The range in pH is from 7.18 to 7.35 or a variation of 0.17 of a pH unit. The average pH of the twenty-five tests amounts to 7.27 ± .00634 with a standard deviation of .047 ± .0045. A still further drop of 0.12 of a pH unit on the aver- age has apparently occurred, therefore, due to the more advanced stage of metamorphosis attained. Stage 4. — Stage 4 individuals were characterized by practically complete metamorphosis. The animals were fully metamorphosed except for the persistence of a small 5-10 mm. tail stump and the lack of tympanic membrane development. Pigmentation, jaw develop- ment, and sitting posture were typically frog-like. Body and hind- limb measurements were not made, but the weight was found to vary between 1.8 and 3.0 grams. The individual pH determinations of Stage 4 animals are listed in column 11 of Table I. The range is from 7.10 to 7.29, or a variation AMPHIBIAN METAMORPHOSIS 413 of 0.19 of a pH unit. The average pH amounted to 7.20 ± .00607 and the standard deviation was .045 ± .0043 (Table II). The average pH therefore indicates a still further drop in concentration amounting, on the average, to 0.07 of a pH unit. Blood pH Following Metamorphosis Newly-metamorphosed young frogs were selected for the final group to be tested in order to determine whether or not the drop in pH, as it occurs during metamorphosis, is a transient phenomenon or is an actual adjustment of the blood to a hydrogen-ion concentration typical of the adult of the species concerned. The weights of Stage 5 animals varied between 1.4 and 2.5 grams. The average weight was somewhat less than that of Stage 4 animals, due no doubt to the fact that they had been without food for from 3 to 4 weeks. Column 13 of Table I lists the individual pH determinations made. The range is from 7.11 to 7.28 or a variation of 0.17 of a pH unit. The average pH of the entire group amounts to 7.18 ± .00648 with a standard deviation of .048 ± .0046 (see Table II). The difference between pH 7.20 (the average for Stage 4) and 7.18 (the average for Stage 5) is insignificant and well within the probable and experimental errors. Hence it seems probable that this pH represents, approxi- mately, the hydrogen-ion concentration of the adult frog's blood for the species in question. This point, however, is a debatable one and will be further taken up in the next section of this paper. DISCUSSION The variation in blood pH between Stage 1 larvae, although not excessive, calls for some explanation. The greatest difference between any two of the twenty-five larvae tested amounted to 0.11 of a pH unit. This variation can be partially accounted for by the limitations of the quinhydrone method which admits of an accuracy to within only 0.04 of a pH unit. It seems quite probable, however, that a definite amount of variation actually exists between various individuals of the same developmental stage in that even greater differences have been recorded in the adult frog. Thus Kamm (1930), working with fully- grown Rana esculenta and Rana temporaries, records blood readings ranging from pH 7.36 to pH 7.61 in cases of starved animals. On the same species Hertwig (1927) obtained variations ranging from pH 7.36 to pH 7.59, while Rohde (1920) recorded a minimum pH of 6.32 and a maximum pH of 7.13. Kamm has pointed out, however, that the extreme variations as recorded by Rohde were probably due to certain technical errors. The wide variations as recorded by Kamm and 414 O. M. HELFF Hertwig, however, give evidence of the differences in hydrogen-ion concentration of the blood that exist between individual adult frogs of the same size and developmental condition. Apparently the blood of the amphibian is very poorly buffered as compared with the con- dition found in the mammals. This point is strongly emphasized by Kamm's (1930) work in which the pH of frog's blood was quickly lowered by feeding with boric acid and as quickly raised by sodium carbonate feeding. It seems quite likely, therefore, that the blood pH differences as recorded in the present paper for normal larvae represent actual differences between individuals, the result of a relatively poor controlling mechanism for maintaining a constant hydrogen-ion con- centration of the blood stream. In this connection it should be stated, however, that the comparatively greater extremes recorded between individuals in Stages 2, 3, 4, and 5 were no doubt partially due to the difficulty of selecting twenty-five animals of exactly the same develop- mental stage. TABLE III Comparison of Mean Differences and their Probable Errors Difference Average compared between means of the difference Probable error of the difference decrease between means 1 and 2 .... 11 007957 13 8 per cent 1 5 2 and 3 .... 12 009080 13 2 1 6 3 and 4 .... 07 008777 8 0 1 0 4 and 5 .... 02 008879 2 3 3 The results of the present paper clearly indicate that an actual lowering of blood pH occurs during metamorphosis. The drop in pH would appear to be a gradual one, although somewhat more pro- nounced in the earlier stages of involution. That the various differ- ences recorded between the means of Stages 1 , 2, 3, and 4 are significant ones is shown by an analysis of such differences in relation to their respective probable errors (see Table III). Thus the difference be- tween the means of Stages 1 and 2 or 0.11 of a pH unit is 13.8 times its probable error while the differences between Stages 2 and 3, and 3 and 4, or 0.12 and 0.07 of a pH unit, respectively, are 13.2 and 8.0 times their probable errors. Hence the differences as recorded are significant and represent an actual increase in hydrogen-ion concen- tration. The difference between the means of Stages 4 and 5 (see Table III), or 0.02 of a pH unit, being only 2.3 times its probable error indicates AMPHIBIAN METAMORPHOSIS 415 that no significant change in hydrogen-ion concentration of the blood had occurred within 3 to 4 weeks following the close of the metamorphic period. Whether this represents the condition typical of young frogs newly metamorphosed in a natural environment is a questionable point. The experimental animals in the laboratory were not fed and hence were in a starved condition when tested. Young frogs in their natural outdoor environment would be feeding at this stage and in- creasing their weight instead of losing weight as was typical of the experimental animals (see Stage 4, Table I). In this relation it may be stated that feeding in the fully-grown frog apparently affects blood pH; according to Kamm (1930) who found the pH higher in the case of laboratory-starved animals as compared with freshly caught, well- nourished frogs. It seems quite probable, however, that the hydrogen- ion concentration of the blood may decrease somewhat in Rana damitans as the latter attain their full size during successive years of growth, especially since Kamm (1930) and Hertwig (1927), have recorded concentrations well above pH 7.3 for fully-grown adults of Rana esculenta and Rana temporaria. However, it is possible that the pH of adult Rana damitans blood is characteristically lower than that of the species tested by Kamm and Hertwig. The relationship of the increased hydrogen-ion concentration during metamorphosis to the various degenerative changes which occur may be of several types. In the first place, it is evident that the pro- nounced autolytic processes operating no doubt account for an accumu- lation of acid metabolites in the blood stream resulting in an increase in hydrogen-ion concentration of the latter. However, it is possible that the first lowering of the pH actually precedes the onset of, and may be causative for, the early autolytic degenerations which occur in that the increased hydrogen-ion concentration either activates the autolytic enzymes or increases their rate of action in the digestion of tissue proteins. Again the tissue proteins may simply be rendered more easily digested. Regarding the above points, it should be noted that the results of the present paper (due to the necessity of killing the larvae in order to obtain sufficient blood for the quinhydrone test) do not serve to indicate clearly whether or not a drop in pH actually precedes the first degenerative changes. It would be interesting in this connection to study the pH changes of the blood in the same individual larva immediately before and after the first degenerative tissue changes occur. This could be done with a suitable micro- electrode hydrogen-ion apparatus and without killing the larva due to the small amount of blood necessary for the test. Even if such tests should show that pH changes are purely secondary in time 416 O. M. HELFF sequence to early degenerative changes, the probability still would remain that the heightened hydrogen-ion concentration of the blood which has been shown to occur by the results of the present paper at a relatively early stage of metamorphosis, are either causative or at least favorable to the later autolytic changes which occur during larval involution. The fact that the various organs and tissues which undergo histolysis do not degenerate all at the same stage of meta- morphosis but have a definite time sequence coincident with various degrees of involution, strongly suggests a correlation between a definite blood pH and the onset of autolysis of each organ. The writer wishes to point out the possibility that carbon dioxide escape may have affected the actual pH readings recorded for the various tests. Such escape no doubt occurred to a certain extent during the collecting and transferring of the blood samples to the electrode vessels, although the general agreement of the results ob- tained for any one stage of metamorphosis would seem to indicate that if such losses did occur, they were not great. Consequently, although finer micro-electrode methods coupled with suitable carbon dioxide correction factors may ultimately serve to alter the average hydrogen-ion concentrations recorded for the various metamorphic stages, it seems quite unlikely that the approximate difference of 0.3 of a pH unit between normal larval and completely metamorphosed larval blood will be seriously affected. The important point to be derived, therefore, from the present paper is not that the results strive to represent a final determination of the exact average pH typical of each stage of metamorphosis, but rather that a definite drop in pH does occur during larval involution which may be and probably is correlated with various metamorphic structural and physiological changes. SUMMARY AND CONCLUSIONS 1. The blood pH of twenty-five normal larvae of Rana damitans was determined by means of the quinhydrone method. An average pH of 7.50 was obtained. 2. Two intermediate stages of metamorphosis were tested (twenty- five cases of each) showing an average progressive decrease in pH to 7.39 and 7.27, respectively. 3. The average blood pH of twenty-five animals, immediately following the completion of involution, amounted to 7.20. It is con- cluded that a definite increase in the hydrogen-ion concentration of anuran blood occurs during larval transformation amounting to an approximate drop of 0.3 of a pH unit. AMPHIBIAN METAMORPHOSIS 417 4. The average blood pH of newly-metamorphosed frogs, three to four weeks after the completion of metamorphosis, was determined as approximately 7.18. It is concluded that little, if any, change in blood pH occurs within three to four weeks following the completion of larval involution, although changes may occur during future adult growth. 5. The results are discussed in general and suggestions offered regarding the relationship between the increase in hydrogen-ion con- centration of the blood and the various degenerative changes which occur during larval involution. LITERATURE CITED BARFURTH, D., 1887. Die Riickbildung des Froschlarvenschwanzes und die sogenannten Sarcoplasten. Arch. f. mikr. Anal., 29: 35. BATAILLON, E., 1891. Recherches anatomiques et experimentales sur la meta- morphose des Amphibiens anoures. A nn. de V Universite de Lyon, 2 : Fasc. 1. BRADLEY, H. C., 1922. Autolysis and Atrophy. Physiol. Rev., 2: 415. CLAUSEN, H. J., 1930. Rate of Histolysis of Anuran Tail Skin and Muscle during Metamorphosis. Biol. Bull., 59: 199. HELFF, O. M., 1926. Studies on Amphibian Metamorphosis. I. Formation of the opercular leg perforation in anuran larvae during metamorphosis. Jour. Exper. Zool., 45: 1. HELFF, O. M., 1928. Studies on Amphibian Metamorphosis. III. The influence of the annular tympanic cartilage on the formation of the tympanic mem- brane. Physiol. Zool., 1: 463. HELFF, O. M., 1929. Studies on Amphibian Metamorphosis. IV. Growth and differentiation of anuran tongue during metamorphosis. Physiol. Zool., 2: 334. HELFF, O. M., 1930. Studies on Amphibian Metamorphosis. VIII. The role of the urostyle in the atrophy of the tail. Anal. Rec., 47: 177. HELFF, O. M., 1931a. Studies on Amphibian Metamorphosis. VII. The influence of the columella on the formation of the lamina propria of the tympanic membrane. Jour. Exper. Zool., 59: 179. HELFF, O. M., 19316. Studies on Amphibian Metamorphosis. IX. Integumentary specificity and dermal plica; formation in the anuran, Rana pipiens. Biol. Bull, 60: 11. HELFF, O. M., AND CLAUSEN, H. J., 1929. Studies on Amphibian Metamorphosis. V. The atrophy of anuran tail muscle during metamorphosis. Phvsiol. Zool., 2: 575. HERTWIG-HONDRU, LIDIA, 1927. Uber das Verhalten der Blutreaktion beim Frosch. Pfliiger's Arch. f. d. ges. Physiol., 216: 796. KAMM, B., 1930. Zur Frage der Konstanz des Froschblut-pH. P finger's Arch. f. d. ges. Physiol., 223: 214. KUXTZ, A., 1922. Metamorphic Changes in the Digestive System in Rana pipiens and Amblystoma tigrinum. Univ. of Iowa Studies, 10: 37. KUXTZ, A., 1924. Anatomical and Physiological Changes in the Digestive System during Metamorphosis in Rana pipiens and Amblystoma tigrinum. Jour. Morph., 38: 581. LINDEMAN, V. F., 1929a. Integumentary Pigmentation in the Frog, Rana pipiens, during Metamorphosis, with Especial Reference to Tail-skin Histolysis. Physiol. Zool, 2: 255. 27 418 o. M. HELFF LINDEMAN, V. F., 1929&. Development of the Nictitating Membrane of the Frog (Rana pipiens). Anat. Rec., 44: 217. MERCIER, L., 1906. Les processus phagocytaires pendent la metamorphose des batraciens anoures et des insectes. Arch. d. Zool. Exper. et Gen. ,4:1. MORSE, W., 1918. Factors Involved in the Atrophy of the Organs of the Larval Frog. Biol. Bull, 34: 149. ROHDE, K., 1920. Zur physiologie der Aufnahme und Ausscheidung saurer und basischer Farbsalze durch die Nieren. Pfliiger's Arch. f. d. ges. Physiol., 182: 114. SCHUBERT, M., 1926. Untersuchungen iiber die Wechselbeziehungen zwischen wachsenden und reduktiven Geweben. Zeitschr. f. mikr. Anat. Forschung., 6: 162. VAN DER JAGT, E. R., 1929. Histolytic Influence of Atrophying Gills of Anurans during Metamorphosis, with Special Reference to Resistance of Fore-limb Integument. Jour. Exper. Zool., 54: 225. SEXUAL PHASES IN THE AMERICAN OYSTER (OSTREA VIRGIN 1C A) W. R. COE OSBORN ZOOLOGICAL LABORATORY, YALE UNIVERSITY During the past three years there has been an accumulation of evidence which indicates that in at least three of the so-called dioecious species of the genus Ostrea a change of sex frequently occurs from season to season or between early life and full maturity. It may be recalled that at least ten of the more than sixty described species of the genus are regularly hermaphroditic and larviparous. Some of these exhibit a rhythmical sequence of alternating male and female phases, as Sparck (1925) and Orton (1926-27) have so fully described for 0. edulis and as Coe (1931, 1932) has more recently shown for O. lurida. Moreover, sex determination in other bivalves, as well as in some gasteropod mollusks, has long been known to be in such a labile condition that environmental changes may profoundly alter its expres- sion. It may not be surprising, therefore, to find that changes of sex, especially protandry, as well as various aspects of intersexuality, have been found to occur in dioecious and oviparous species of oysters. For example, Roughley (1928) concluded from his observations on 0. cucnllata that that species, formerly considered dioecious, is regularly protandric, for nearly all the very small individuals were found to be males. An experiment by Amemiya (1929) has been thought to indicate that in 0. gigas the sexual phase of each individual is deter- mined each winter without influence from its previous sexual condition. Many years ago Stafford (1913) found indications that the Amer- ican oyster (0. virginica) is protandric on the Canadian coasts and presented evidence that the young animal becomes a sexually mature male when it has reached a length of about 25 mm. At 32 mm. the gonads may be distended with spermatozoa. More recently Burkenroad (1931) has also shown that protandry occurs in this species on the coast of Louisiana. He concluded that although the sexes of the adults are morphologically separate, each individual is essentially a protandrous hermaphrodite. His evidence indicates that close association with large oysters causes some indi- viduals to assume or retain the male phase, although other oysters of the same size, but growing singly, are predominantly females. He 419 420 W. R. COE found small individuals to be almost always males regardless of their associations and interpreted his evidence as indicating that the likeli- hood of large oysters being males decreases rapidly as the distance from other individuals increases. Further evidence of protandry in this species has been furnished by Needier (1932) from oysters collected at various localities on the North Atlantic coasts of the United States and' Canada. She con- E \ A '.\§gf '••••• •<•!? •Vv ••••*. FIG. 1. Diagrams showing the primitive intersexual gonad and its trans- formation into the definitive spermary and ovary. A, early gonad with undiffer- entiated cells; B, intersexual phase, with differentiating ovocytes (oc); C, later intersexual phase with preliminary abortive spermatogenesis; D, spermary with only a few small ovocytes; E, spermary with many ovocytes, some of which (oc') are in process of degeneration; F, young ovary, with many spermatocytes (s£c), some of which (s£c') are pycnotic; G, nearly ripe ovary with residual cells and degenerating spermatocytes. eludes that the majority of individuals first mature as males and that many of them later change to females. The change may be hastened by favorable nutritive conditions and may possibly be retarded by close association with older females. She observed one instance where a three-year-old male changed into a functional female in the interval preceding the next breeding season. In order to determine more precisely the sequence of these changes in sexuality and the histological activities which accompany them, SEXUAL PHASES IN AMERICAN OYSTER 421 the gonads of a large number of oysters have been examined at frequent intervals during the first two years of their lives. Some of this material was collected from rocks along the shore of Long Island Sound in the vicinity of New Haven; some was taken from various natural oyster beds near Woods Hole, Massachusetts; many samples both of cultivated and untransplanted oysters from Quinnipiac River, New Haven Harbor, and Long Island Sound have been supplied by Mr. oc FIG. 2. Primary gonad in animal about four months of age, showing large ovocytes (oc), spermatogonia (spg), and primary spermatocytes (spc I) in spireme phase, with secondary spermatocytes (spc II) and a single spermatid (spt) bordering the ciliated genital canal (gc). Howard W. Beach, Chairman of the Research Committee of the Oyster Growers and Dealers Association of North America, while a most instructive series of first-year stages was furnished by Mr. J. B. Glancy from the floats of the same Association at West Sayville, Great South Bay, Long Island. From these collections the early development of the gonads and their transformations in the course of successive phases of sexuality have been studied both in life and by means of serial sections. 422 W. R. COE DEVELOPMENT OF THE PRIMARY BISEXUAL GONADS The profusely branching tubules of the primitive gonads can be found in young oysters at the age of six to eight weeks after setting or when the shell has reached a length of 6-10 mm. They ramify within the thin layer of connective tissue immediately beneath the body walls as Stafford (1913) has previously described. In these primary gonads the germinal epithelium consists of a thin layer of morphologically undifferentiated cells which lie upon the inner side of the tubular gonad (Fig. 1), while the cells which line the outer oc FIG. 3. Primary intersexual gonad, showing primary spermatocytes (spc I) and ovocytes in synapsis, secondary spermatocytes (spc II) and several large ovocytes (oc). Age about four months. wall, adjacent to the epidermis, become differentiated into the ciliated epithelium of the genital canals (Figs. 2, 4, 5, 6) as Hoek (1883-84) so fully described many years ago for 0. edulis and as Coe (1932) verified in 0. lurida. A few weeks later the germinal epithelium shows a differentiation into larger and smaller cells. The former are soon recognizable as ovocytes by the presence of fibrillar mitochondrial bodies, while many of the latter show by their rapid proliferation and later specialization that they belong to the male germ line. SEXUAL PHASES IN AMERICAN OYSTER 423 By the middle of October, at the age of about three months, the shells of some individuals have become 20-25 mm. in length, and in these the gonads have already become distinctly bisexual (Figs. 2-6). A typical section through the genital canal at this age (Fig. 2) shows more or less numerous indifferent cells remaining along the inner border, interspersed with large ovocytes in which the coarse mitochondrial filaments of the so-called yolk nuclei are always con- FIG. 4. Primary intersexual gonad in an animal about four months of age, showing indifferent cells (I) and several large ovocytes (oc), with spermatocytes (spc I) and a single spermatid (spt) bordering the ciliated genital canal (gc). spicuous. Morphologically less well differentiated are the ovogonia and spermatogonia which resemble each other so closely that they cannot ordinarily be distinguished. Both types of cells then pass through similar synaptic phases accompanied by spiremes of coarse, densely-staining chromosomes (Figs. 2, 3, 9B, 9C) as they transform into ovocytes and primary spermatocytes respectively. The spermatogonia proliferate rapidly, leading to the formation of the numerous spermatocytes wrhich soon give the intersexual gonad 424 W. R. COE a predominantly male appearance. In some individuals the spermato- cytes complete their meiotic divisions to form spermatids, thereby accentuating the resemblance of the gonad to a spermary (Figs. 4-6). In other animals of the same age the gonads may retain a closer similarity to an ovary (Figs. 2, 3), and it is improbable that any spermatids are formed in every individual. All degrees of inter- FIG. 5. Primary gonad, with ovocytes (DC) and primary spermatocytes (s£c I) in synapsis; a few spermatids (s/>/) are already present on the border of the genital canal (gc) ; age about four months. sexuality are found, although the vast majority of individuals are predominantly male. Not infrequently some parts of the system may assume a distinctly male appearance while adjacent follicles are characteristically female, as shown in Fig. 6. The process of spermatogenesis does not continue through the winter, however, in the localities investigated and no functional SEXUAL PHASES IN AMERICAN OYSTER 425 spermatozoa are formed. The gonads may continue to increase some- what in size during December, with an increase in the numbers of secondary spermatocytes and spermatids. Their activities are then interrupted until there is a rise in the temperature of the water the following spring. HIBERNATION During the long period of hibernation, when the valves of the oyster remain closed, relatively minor changes occur in the primary gonads and these are mainly regressive in character. There is little, if any, increase in the size of the gonads during the oyster's first oc oc FIG. 6. Portions of two follicles from the same individual at the age of about four months: one of these (A) is predominantly of the female type while the other (B) consists mainly of spermatogenic cells, although both show some indications of their intersexual character; oc', young ovocyte in spireme phase; other letters as in Fig. 2. winter or in the numbers of their constituent cells. On the contrary, many of the previously differentiated cells become obviously abnormal and evidences of cytolysis are frequently seen. Remains of dis- integrated cells sometimes occur in the lumens of the follicles and ducts. TRANSFORMATION TO FUNCTIONAL GONADS At West Sayville the animal retains its primary or immature gonad throughout the winter as the figures in Table I will show. As the water becomes warmer, however, spermatogenesis is resumed in some of the larger animals and in March about one-fourth of the young 426 W. R. COE oysters examined had transformed into males. No females were recognized until a month later. In an occasional individual, due apparently to poor nutrition, the primary bisexual type of gonad is retained until the second year, at which time a few functional spermatozoa may be formed. Males The transformation of the bisexual gonad into a spermary con- sists in the proliferation of spermatogonia and spermatocytes, usually accompanied by the disintegration of some of the previously formed ovocytes (Fig. 1). There is a rapid extension of the branching system of follicles, the new branches being, as a rule, exclusively male in appearance, for the reason that the ovocytes usually found in the older FIG. 7. Transition stages in the development of the mature ovary (C) from the primitive intersexual gonad (.4). B, developing ovary in an individual about ten months of age, showing numerous small pycnotic cells (p] in the lumen and between the ovocytes (oc) ; s, cells in spireme phase. parts of the system are not carried into the new follicles (Fig. 8). But this rule is not without exceptions, for many grades of inter- sexuality occur and the gonad is classed as a spermary in this report if it has a distinct preponderance of spermatogenic cells, even though many ovocytes are situated along the walls of the follicles (Fig. 12, A}. The term hermaphrodite is reserved for those cases in which there are extensive areas or large masses of the cells representing each of the sexes (Fig. 12, B). Females As the season advances an increasing proportion of the young oysters attain sexual maturity, as shown in Table I. During April SEXUAL PHASES IN AMERICAN OYSTER 427 1 • "S-x. / \ / ' \\ - - . . • \ .ft : ' VO Ix^X ^-^X t'/^'"-1 J • ? ; ** - 2f FIG. 10. Spermary early in second year, showing its continued intersexual character; gc, genital canal; oc, ovocytes. SELF-FERTILIZATION Among 55 sexually mature yearlings taken from the culture floats at West Sayville on June 23 were two functional hermaphrodites. Many of the ova in each of these proved capable of self-fertilization and apparently normal development in spite of the vast excess of sperm present. Needier (1932) has reported a similar observation. SPERMATOGENESIS The successive stages in the process of spermatogenesis are similar to those described by Coe (1932) for 0. lurida except that the spermato- 430 W. R. COE OC FIG. 11. Spermary at end of second year, showing its continued intersexuality by the presence of large ovocytes (oc) on wall of follicle; letters as in Fig. 2. SEXUAL PHASES IN AMERICAN OYSTER 431 cytes and spermatids are free to separate immediately after their formation (Fig. 8), instead of adhering in masses as they do in the latter species where the spermatozoa are united into sperm-balls. GONADS AT THE END OF THE BREEDING SEASON AND DURING THE SECOND YEAR Such individuals as become sexually mature during their first year produce but a relatively small number of gametes, retaining as residual cells a large proportion of the germinal cells composing the gonads. The residual germinal epithelium of the ovary is similar to that shown in Fig. 9, A (except for the absence of the single ripe ovum), while Fig. 10 shows a section of a spermary shortly after activity has been resumed near the beginning of the second year. f p FIG. 12. A, partial hermaphroditism; portion of spermary of hermaphroditic male, showing one of the scattered ovocytes. B, complete hermaphroditism; gonad with about equal areas of ovarian and spermatogenic tissue. Age about ten months. The gonads of both sexes after spawning usually retain indications of their continued intersexual character; in the males by the presence of ovocytes along the walls (Figs. 10, 11), and in the females by groups of small cells similar to those shown in Fig. 7. In the latter case, however, there is as yet no proof that these small cells actually belong to the male line, for active spermatogenesis has not been found in an individual classed as a female. A spermary may have few or many ovocytes which sometimes produce yolk in a normal manner, but the reverse conditions seem not to hold except in the relatively small proportion of individuals which exhibit functional hermaphroditism. SEX REVERSAL The observations of both Burkenroad (1931) and Needier (1932) indicate a strong tendency toward protandry in this species. The former concluded that the change to the female phase takes place 432 W. R. COE when the shell of the young animal has reached a length of about 40 mm., while Needier has positively proved that a change of sex may occur as late as the third or fourth year. It has hitherto been an open question whether all individuals are protandric; that is, whether the relatively few oysters which spawn as females at the end of their first year may have passed through a preliminary male phase the preceding autumn. It can now be answered that such is not the case in the localities under consideration. The ovary of yearling females develops directly out of the primary intersexual gonad by the growth of the primordial ovocytes and the elimination of spermatocytes and spermatids before spermatogenesis has been completed. The preliminary male phase is thus abortive in these localities although such may not be the case in the warmer areas farther south along the coast. CORRELATION OF AGE, SIZE, AND SEX DURING THE FIRST YEAR It has been stated above that at West Sayville, Long Island, a few of the more rapidly growing young oysters become sexually mature as males during March of their first year. Females were first recog- nized in April and these were mainly among the largest individuals of the group. Tables I and III show the numbers of individuals of each sex and of each size found at three different periods preceding the breeding season and once after spawning had commenced. The average size of the young females always exceeds that of the sexually mature males of the same age if considerable numbers of each are considered. This rule is shown by Needier (1932) to hold also for oysters at the age of two and of three years, but in still older animals the males are said to average as large as those of the other sex. The greater size of the females of the younger ages may be corre- lated (a) with a more efficient metabolism associated with the female sex mechanism, or (b) greater activity of the female in obtaining food, or (c) the actual differentiation of the individual into a female as the result either of its inherent metabolic potentialities or its favorable environmental conditions, or both. And, conversely, the responsible agency for the determination of maleness and slower growth may be a less favorable metabolism, either genetic or environmental, or, conceivably, the retarding influence of older, associated individuals of either sex. Inspection of Table I shows clearly that the collection made on March 23 represents a large proportion of oysters that would have become sexually mature later in the season and these would presumably have included both sexes. Omitting this collection and combining SEXUAL PHASES IN AMERICAN OYSTER 433 the two groups taken April 29 and May 21 shows (Table II) that shortly before the beginning of the breeding season in June, 41.6 per cent of 149 specimens sent for examination were still immature, while 41 per cent were males, 1.3 per cent hermaphroditic, and 16 per cent were females. By a curious coincidence these figures agree surprisingly closely with those obtained by Needier (1932) from the same locality TABLE I Correlation of Size and Sex during First Year, W. Sayville March 23 April 29 May 2 1 June 23 T (-1 L/cii §t n Im. M. F. Im. M. F. Im. M. H. F. Im. M. H. F. m m . 10-19 . . 6 — — . 16 — — 4 — — — 2 — — • — 20-29 . . 6 — — 19 1 — 8 — — — • 5 3 — — 30-39 . . 7 1 — - 4 12 2 9 14 — 1 2 8 — - — • 40-49 . . 7 3 — 1 12 1 1 7 — . 1 — 11 1 1 50-59 . . — 2 — — 5 10 — • 7 2 4 — • 7 — 7 60-69 . . — 1 — — — 3 — 3 — 2 — • 12 1 1 70-79 . . 3 Total . 26 7 0 40 30 16 22 31 2 8 9 41 2 12 TABLE II Correlation of Size and Sex; W. Sayville. Groups from April 29, May 21, and June 23 combined. Total Im. p.c. M. p.c. H. p.c. F. p.c. Less than 40 mm More than 40 mm 110 103 69 2 62.7 2.0 38 64 34.5 62.1 0 4 0 4.0 3 33 2.7 32.0 Less than 50 mm 146 71 48 6 68 46.6 1 0.9 6 4.1 More than 50 mm 67 0 0 34 50 7 3 4 5 30 448 Total 213 71 33 3 102 47 9 4 2 0 36 17.0 Im., immature; sex not determinable; M., male; H., hermaphrodite; F., female. at the end of June, 1931. She found 45.3 per cent immature, 38.7 per cent males, and 16 per cent females among 119 individuals which she examined. On June 23, 1932, however, the writer found that only 14 per cent of the 64 samples studied were still immature, while 64 per cent were males, 3 per cent functional hermaphrodites, and 19 per cent females 28 434 W. R. COE (Table I). The smaller proportion of sexually mature yearlings in 1931 as compared with 1932 may be accounted for by the somewhat higher temperature of the water during the growing season of the latter year. Combining the three groups examined April 29, May 21, and June 23, 1932, numbering 213 individuals, shows 71, or 3.33 per cent, immature; 102, or 47.9 per cent, males; 4, or 2 per cent, functional hermaphrodites; and 36, or 17 per cent, females (Table II). Considering only the 149 individuals of the present collections that were sexually differentiated we find that they comprise 109 males; 4 hermaphrodites, and 36 females, a percentage ratio of 73.2, 2.7, and 24.2 respectively. TABLE III Correlation of Size Groups and Sex First Year; W. Sayville March 23 April 29 May 21 June 23 I in. M. F. Im. M. F. Im. M. H. F. Im. M. H. F. mm. Less than 40 19 1 0 39 13 2 21 14 0 1 9 11 0 0 More than 40 7 6 0 1 17 14 1 17 2 7 0 30 2 12 Less than 50 26 4 0 40 25 3 22 21 0 2 9 22 1 1 More than 50 0 3 0 0 5 13 0 10 2 6 0 19 1 11 Total 33 86 63 64 Im., immature; sex not determinable; M. male; H. hermaphrodite; F. female. At the end of June, 1931, Needier found 46 males and 19 females among the 65 sexually mature individuals which she examined from the same locality. Combining these figures with those of Table II gives a total of 155 males, 4 hermaphrodites, and 55 females, a ratio of nearly 3 males to 1 female. Of the sexually mature individuals less than 40 mm. in length (Table III) there were 39 males to 3 females, but those of larger size showed 70 males and 33 females. But in the class measuring more than 50 mm. in length (Table III) there are nearly as many females as males, indicating a rather definite correlation of size and sex, as Burken- road (1931) has shown for the general population and Needier (1932) for the age groups. Probably the most reliable correlation of sex and SEXUAL PHASES IN AMERICAN OYSTER 435 size is shown by the collection made during the early part of the breed- ing season (June 23), representing as nearly as could be estimated an average sample of the entire population remaining on the floats. None of the 12 females was then less than 40 mm. in length and only one of them less than 50 mm. Approximately 27 per cent of the 41 males, on the other hand, measured less than 40 mm. and 54 per cent less than 50 mm. Those in the largest-sized group (70-79 mm.) were all females (Table I). Near the end of the spawning season (July 29) a collection of 70 individuals showed a sex ratio of 100 males to 23 fe- males. Of those measuring less than 50 mm. in length there were 37 males but no females, while those of larger size comprised 15 males and 12 females. Only a single individual was still immature. The correlation of sex and size is actually much closer than the tables indicate, for the figures given are based upon the length of the shell alone and not upon the actual size of the animal. Selection based on volume would undoubtedly give a still larger percentage of females at the end of the first year in the locality under consideration. In other areas, however, the sex ratios at the end of the first year are quite different, as Needier (1932) has shown by comparing one- year-old oysters from Prince Edward Island, Long Island Sound, and Great South Bay, Long Island. She concluded that in the colder areas, with a shorter season of activity, relatively few individuals became sexually mature as males before their second year and none- as females. FIRST-YEAR OYSTERS FROM QUINNIPIAC RIVER AND NEW HAVEN HARBOR Collections were made monthly from September, 1931, to July, 1932, and a very large number of young oysters were examined. Nearly a hundred of these were cut in serial sections. Conditions were unfavorable for a large set in these areas during the summer of 1931 and there was a high mortality of the young oysters in exposed situations during the ensuing winter. In November, when about four months of age, the shells of the largest individuals were from 20 to 27 mm. in length, but the majority measured only 5 to 15 mm. The gonads of the smaller individuals were beginning to branch out beneath the body walls, while those of the larger animals had already reached the primary bisexual phase, with differentiated ovocytes and spermatocytes (Figs. 2-6). A few spermatids were also present in some individuals. The rate of growth and the accompanying sexual changes through the ensuing months and until the breeding season in July are sum- marized in Table IV. The table shows that the first individuals to 436 W. R. COE become recognizably differentiated sexually were all males, and none of these could be identified with certainty until April or May. During June there was a relatively rapid growth in size, accompanied by an increasing number of sexually differentiated males among the larger individuals. But the correlation of size with sexual differentiation is not without exceptions, as Tables I and IV will show, although the sexual conditions may be definitely controlled by nutrition. Sper- matogenesis or ovogenesis may be inaugurated when nutritive condi- tions are favorable but growth may then be checked by the encroach- ment of other individuals, while gametogenesis continues. The result may be a dwarfed but sexually functional animal. TABLE IV Correlation of Age, Size and Sex, Quinnipiac River and New Haven Harbor Nov. Dec. Mar. May Julj Im. Im. Im. Im. M. F. Im. M. H. F. Total mm. 5-9 Many Many Many Many- 0 0 0 0 0 0 0 10-14 Many Many Many Many 0 0 7 0 0 0 2 15-19 Some Many Many Many 0 0 6 11 0 0 17 20-24 Few Some Some Some Few 0 9 41 0 0 50 25-29 Few Few Few Few Few 0 0 104 1 7 107 30-34. . 0 0 0 Few Few 0 0 116 1 3 120 35-39 0 0 0 0 0 0 0 8? 7 2 86 40-44 0 0 0 0 0 0 0 ?6 0 3 29 45-49. . . 0 0 0 0 0 0 0 / 0 1 8 50-54 0 o 0 0 0 0 0 ? 0 7 4 Total Few 0 17 389 4 13 423 An examination of 423 yearlings taken at random at the height of the breeding season in July indicated that about 96 per cent of the survivors from the set of the previous year had become sexually mature. Most of these were males filled with spermatocytes and ripe spermato- zoa (Table IV). The other 4 per cent still retained the primary inter- sexual gonads and indicated that sexual maturity had been postponed until their second year. All of these immature individuals, as would be expected, showed evidence of unfavorable nutritive conditions as indicated by their dwarfed size. The predominantly protandric nature of the group is shown by the fact that 389, or nearly 96 per cent of the 406 sexually mature indi- SEXUAL PHASES IN AMERICAN OYSTER 437 viduals examined were more or less fully ripe males, while only 13, or 3.2 per cent, were females. Four, or 1 per cent, were functional hermaphrodites with large masses of both ova and spermatozoa. The ratio of males to females is thus 100 : 3.3. While the number of females is too small to be statistically de- pendable, it will be observed that their mean length is somewhat greater than that of the males, but less conspicuously so than in the much larger group from West Sayville (Table I). If the volumes had been measured, instead of the lengths, the females would have shown a still greater comparative size. Comparison of Tables I and IV will show that the proportion of individuals which mature as females during their first year at West Sayville and in New Haven Harbor is about ten times as great in the former locality as in the latter. The size differences of the two populations of yearlings show that not only is there a correlation between rate of growth and sex as concerns individuals but that the proportion of females is several times higher in the locality where the conditions for growth are the more favorable or the growing season longer. There may be some question as to whether the evidence is sufficient to justify the conclusion that metabolic conditions at the time of sexual differentiation actually determine the direction taken by the primary intersexual gonad in its transformation into the func- tional organ of the first sexual phase. But until further experimental evidence is available it seems to be the most reasonable hypothesis suggested. FIRST YEAR OYSTERS FROM NEAR WOODS HOLE, MASSACHUSETTS Young oysters from small natural beds in the vicinity of Great Harbor, from the shores of the neighboring islands, and from Onset were examined in the summers of 1931 and 1932. In all of these areas growth is slow during the first year and the shell seldom reaches a length exceeding 30 to 35 mm. The more usual length is 6 to 20 mm. An examination of 389 yearling oysters from these localities showed that 373, or nearly 96 per cent, were still immature or the sex un- determinable, 9 were males, 3 were hermaphrodites and 4 were females. At the age of two years, when most of the young oysters in that region become sexually mature, the ratio of males to females is about one hundred to fifty-five, with approximately two per cent true hermaphrodites. At this age the average size of the females considerably exceeds that of the males, as was also the case in the other areas during the first breeding season. 438 W. R. COE COMPARISON WITH OTHER SPECIES The primary bisexuality of 0. virginica and the development of the definitive gonads, including later sex reversals, are in many respects similar to the series of sexual phases which characterize such strictly hermaphroditic and larviparous species as O. edulis (Orton, 1926-27) and O. lurida (Coe, 1931, 1932). In both the hermaphroditic and seasonally dioecious types the more rapid proliferation of sperma- togonia as compared with the ovocytes soon gives the ear.ly gonad its predominantly male characteristics. In the former type, however, all individuals are thought to become functional males before assuming the female phase, while in O. virginica from 3 to 30 per cent of the sexually mature young individuals at different localities show only an abortive male phase, the primary gonad developing into an ovary without completed spermatogenesis. In each type residual cells of both sex lines remain after the first spawning, providing a cellular mechanism which leads in some of the hermaphroditic forms to a series of alternating sexual phases, while in 0. virginica the sexual changes appear to be more or less faculta- tive, for it is known that in at least some individuals the same sexual phase may be retained for several years. This is presumably true of the great majority of adult oysters under a stable environment. Evidences of protandry and sex change have been reported in other species of oviparous oysters. Of a large number of very young 0. cucullata examined by Roughley (1928) all except about 5 per cent were males and he suspected that these exceptions might previously have spawned as males. He found nine functional hermaphrodites. These were thought to represent stages in the transformation of male to female, but it seems more probable that the sexual conditions in that species are not very different from those here described for 0. virginica, and that the change of sexuality in both species takes place in the interval between two breeding seasons. Among 3000 large adult oysters from thirty different localities he found a sex ratio of 270 females to 100 males. In 0. angidata also hermaphroditism occurs occasionally. Amemiya (1925) found two such individuals among 14 males and 59 females. But it is not known whether that species, which is closely allied to 0. virginica, experiences similar sex changes. A most interesting type of sex change has been recently reported by Amemiya (1929) for the Japanese oyster (O. gigas), previously considered dioecious. In one summer a small hole was made in the shell of each of several hundred oysters and the sex thereby deter- mined. The sexes were then placed in separate cages and returned to SEXUAL PHASES IN AMERICAN OYSTER 439 the sea. A year later it was found that oysters of both sexes were present in each cage. It was concluded that about 25 per cent of the females and 60 per cent of the males had changed their sex during the winter. If these conclusions prove to be well founded, Amemiya's hypothesis, that the sex of any individual of that species is determined each winter independently of its previous sexual conditions, will add a new phase to the many variants of sexuality in animals. Since there is such a wide diversity in the abundance and size of the ovocytes in the sexually mature young males of 0. virginica, it is pertinent to inquire whether there may not be two genetically distinct types of these males. The samples studied indicate, as shown in Table II, that at West Sayville about 48 per cent of the entire one- year group or more than 70 per cent of the sexually mature yearlings function as males, while in New Haven Harbor there are fully thirty males to one female at their first breeding season (Table IV). It is conceivable that those males with but few and very small ovocytes are genetically "true males" while those with more numerous and larger ovocytes may represent the protandric males. Perhaps it is only the latter that later undergo sex reversal, as Orton (1928) has suggested for Patella. It would probably be unwise to speculate further in this connection until more complete evidence is available. The appearance of the primary gonads, however, and the changes which they subsequently undergo, suggest that sexuality in this species may rest upon a basis somewhat comparable with that which Witschi (1932) has found in certain races of frogs. SUMMARY 1. Examination of the developing gonads of young oysters from various localities at frequent intervals during the first two years of life shows that a primary bisexual gonad is formed in each individual within a few months after setting. 2. The activities of the gonad depend upon the temperature of the water and apparently other conditions of nutrition, a much larger proportion of the animals becoming sexually mature during the first year in warmer than in cooler localities. 3. The primary gonad contains the antecedent cells of both sexes, with ovocytes upon the walls of the follicles and spermatocytes inter- mingled and bordering the lumens. 4. The protandric nature of the primary gonad frequently becomes manifest by the rapid proliferation of the spermatogonia and the formation of primary spermatocytes; the latter soon pass through the synaptic phases and lead to the production of secondary spermatocytes 440 W. R. COE and spermatids at the age of a few months. But no functional spermatozoa have been observed until the following spring in the areas investigated. The species is not strictly protandric, however, for 3 to 30 per cent of the sexually mature yearlings are females, the ovaries of which have developed directly from the primary gonads without the completion of a preliminary functional male phase. 5. The definitive sexual gland is a transformation of the primary gonad by the proliferation of spermatogonia and the disintegration of many of the ovocytes to form the spermary or, less frequently, the growth of ovocytes, accompanied by the disintegration of spermato- cytes and such spermatids as may be present, to form an ovary. But the intersexual character is usually retained to at least some extent in both types of gonads. 6. The proportion of male and female cells in the mature gonad is highly variable, a few large ovocytes being frequently found in some parts of otherwise typical spermaries, while in the ovary some follicles may retain characteristic male cells. True hermaphroditism was found in 1 to 4 per cent of the sexually mature oysters at the end of their first year. Apparently normal development follows self- fertilization. 7. In the warmer of the two principal localities investigated about 70 to 80 per cent of the oysters which became sexually mature during their first year were males. In the cooler locality the proportion of males exceeded 95 per cent. 8. Most of the relatively small number of females are among the largest of their age group, indicating a close correlation between sex and size. This may imply that the female is metabolically the more active sex or that she requires better nutritive conditions in order to mature, or that sex in this species is so labile that the nutritive condi- tions of the individual at the critical period of sex differentiation determine which of the alternative types of cells in the primary bi- sexual gonad shall predominate. Alternative genetical explanations are discussed. 9. After the animal has spawned as male or female the gonad may still retain its bisexual character; the sexual phase during the follow- ing year may again depend largely upon nutritive conditions. 10. The primary bisexuality of this species and the cellular mech- anism for sex reversal here reported are interpreted with reference to related species of the genus in which hermaphroditism and alternating sexual phases have been retained. SEXUAL PHASES IN AMERICAN OYSTER 441 LITERATURE CITED AMEMIYA, I., 1925. Hermaphroditism in the Portuguese Oyster. Nature, 116: 608. AMEMIYA, I., 1929. On the Sex-change in the Japanese Common Oyster, Ostrea gigas Thunberg. Proc. Imper. Acad. Tokyo, 5: 284. BURKENROAD, M. D., 1931. Sex in the Louisiana Oyster, Ostrea virginica. Science, 74: 71. COE, W. R., 1931. Sexual Rhythm in the California Oyster (Ostrea lurida). Science, 74: 247. COE, W. R., 1932. Development of the Gonads and the Sequence of the Sexual Phases in the California Oyster (Ostrea lurida). Bull. Scripps lust. Oceanog., Tech. Ser., 3: 119. GALTSOFF, PAUL S., 1932. Spawning Reactions of Three Species of Oysters. Jour. Wash. Acad. Sci., 22: 65. HOEK, P. P. C., 1883-84. De Voortplantingsorganen van de Oester: Les organes de la generation de 1'huitre. Tijdschr. Nied. Dierkund. Vereeniging, Supple- ment 1: 113-253. NEEDLER, ALFREDA BERKELEY, 1932. Sex Reversal in Ostrea virginica. Cont. Can. Biol. and Fish., 7: 285. ORTON, J. H., 1926-27. Observations and Experiments on Sex-change in the European Oyster (O. edulis). Jour. Mar. Biol. Assoc., 14: 967. ORTON, J. H., 1928. Observations on Patella vulgata. Part I. Sex Phenomena, Breeding and Shell Growth. Jour. Mar. Biol. Assoc., 15: 851. PRYTHERCH, HERBERT F., 1924. Experiments in the Artificial Propagation of Oys- ters. Bur. Fish. Doc. No. 961: 1-14. (App. xi, Kept. U. S. Comm. Fish. for 1923.} ROUGHLEY, T. C., 1928. The Dominant Species of Ostrea. Nature, 122: 476. SPARCK, R., 1925. Studies on the Biology of the Oyster (Ostrea edulis) in the Limfjord, with Special Reference to the Influence of Temperature on the Sex Change. Kept. Dan. Biol. Sta., 30: 1. STAFFORD, Jos., 1913. The Canadian Oyster, its Development, Environment and Culture. Comm. of Conservation, Canada: 1-159. YYiTSCHi, EMIL, 1932. Physiology of Embryonic Sex Differentiation. Am. Nat., 66: 108. A COMPARISON OF THE LIFE HISTORIES OF MICTIC AND AMICTIC FEMALES IN THE ROTIFER, HYDATINA SENTA ' JOSEPHINE CAROLYN FERRIS DEPARTMENT OF ZOOLOGY, UNIVERSITY OF NEBRASKA INTRODUCTION The life history of the rotifer, Ilydatina senta, involves the repro- duction of two kinds of females, amictic females, those which reproduce wholly by parthenogenesis, and mictic females which reproduce par- thenogenetically or bisexually. A comparison of these two types of females in regard to the periods of their life histories was made in an endeavor to find if there is a difference in the metabolic rate of the two and to discover its possible relation to the factors which regulate the production of these two kinds of females. Miller found that the three types of individuals, amictic females, mictic females, and males of Lecane inermis differed in the length of the total life period. The unfertilized mictic females and amictic females not only differ in length of life but also differ in rate of pro- duction, duration of fecund and post-fecund periods, and in the degree of correlation between fecundity and length of life. Miller states that the difference in length of life of these two kinds of females is due to the fact that egg-production is a less strenuous process in mictic females than in the amictic ones and consequently the mictic females survive the fecund period and pass entirely through the post-fecund, dying a natural death in old age. The mictic females produce fewer and smaller eggs than the amictic females at a slower rate and they cease egg-production at an earlier age. The differences in length of life would be due probably to the differences in the meta- bolic rate in the fecund period. Jennings and Lynch point out that in the case of Proales sordida there is no correlation between length of life and the number of eggs produced in the amictic females but that diversities in fecundity were due rather to the size of the eggs from which they have hatched. Smaller eggs which are supposed to have been produced early in the family produce less fecund individual daughters than the larger eggs produced later. 1 Studies from the Zoological Laboratory, University of Nebraska, No. 172. 442 MICTIC AND AMICTIC FEMALES IN HYDATINA 443 Like Proales sordida and Lecane inermis, there are four distinct periods which can be distinguished in the history of the individual in Hydatina senta: (1) the hatching or embryonic period; (2) the pre- fecund or adolescent period of rapid growth; (3) the fecund or egg- laying period; (4) the post-fecund or old age period. Hydatina senta is one of the larger rotifers commonly found in stagnant and foul ponds. It has been worked with a great deal in laboratories because it is easily cultivated, hardy, multiplies rapidly and has sexual and parthenogenetic generations. The females used in the present work were taken from a general culture which had been collected from a goldfish pool at Seward, Nebraska, in May 1931. The experiments and observations were made at the suggestion and under the supervision of Professor D. D. Whitney to whom the author wishes to express her indebtedness for advice and assistance given. MATERIALS AND CULTURE METHODS Hydatina senta is easily cultured in a variety of solutions. A very favorable one is made by using old hay tea as a basis. This is pre- pared by boiling 1 gram of ground timothy hay in 4000 cc. of tap water for 10 minutes. It is then strained and allowed to age for 4-6 weeks before using. The tap water from which it was made was placed in direct sunlight for several hours to remove an objectionable amount of chlorine. To 100 cc. of this aged hay tea there was added 1 cc. of 1 per cent urea solution, 1 cc. of 1 per cent ox-gall solution and 1 cc. blood solution. This combination of ingredients made a very favorable culture medium for these rotifers. The urea stock solution consisted of 1 gram of urea crystals dis- solved in 100 cc. of tap water and brought to a boil ; the ox-gall solution was prepared by using 1 gram of dried ox-gall plus 100 cc. of tap water and brought to a boil. The blood solution was prepared by using 1 gram of dried blood plus 100 cc. of tap water, brought to a boil and filtered. Throughout the experiment a pure culture of the flagellate, Poly- toma, was used as the food. This was prepared by using 1200 cc. of tap water that had been boiled, cooled, and put into a small battery jar. Into this was placed a muslin bag containing 200 grams of bone meal which previously had been brought to a boil and allowed to cool. A fresh hay tea solution also was added which was prepared by boiling for 10 minutes 1 gram of ground timothy hay in 100 cc. of sunned water. This culture was inoculated with Polytorna and placed at room temperature in a north light exposure. To maintain a good culture of Poly to ma, the bag of bone meal was changed every 48 hours and fresh 444 JOSEPHINE C. FERRIS hay tea solution added. After a few days, however, the culture water would become too foul and develop a red coating of bacterial growth on the walls of the jar. Whenever this occurred a new culture was made by pouring the top of the old culture into another sterilized jar and adding enough sterilized water to make 1200 cc. Then fresh bone meal and fresh hay tea solution were added as stated above. In this manner a vigorous culture of Polytoma was maintained for many months. Food for the rotifers from this culture was prepared daily by re- moving the film from the surface with a sterilized spoon and thus obtaining Polytoma in countless numbers. These were then washed twice with old hay tea solution by means of the centrifuge. One cubic centimeter of this concentrated Polytoma was diluted with 15 cc. of old hay tea solution. One drop of this was then placed daily in each individual watch-glass containing one female in 7 cc. of culture solution. The life histories of 184 amictic, 113 mictic females, and 88 mictic females whose eggs had been fertilized were studied and compared. The females were all cultivated under similar conditions and kept at a constant temperature of 16° to 17° C. This temperature was maintained by use of a double-walled temperature bath through which there was a continual in- and out-flow of tap water. During the winter months the temperature of this bath was kept quite constant and the entire observations were made during this period. The record for the life history periods began at the time of isolation of the eggs. This was done by isolating in a container a group of mothers that were about ready to lay eggs. They were given a great deal of food and then observed every hour. The first lot of eggs was not recorded due to the fact that some of them may have been laid previously and have been isolated with the mothers. However, be- ginning at the end of the first hour after they had begun to lay, the eggs were isolated every hour, placed in a container, and labelled. In this way it was known that the eggs isolated at any particular time had been produced during the preceding hour. These eggs were all placed in the temperature bath and carefully observed on the following day for the hatching of the young females. Upon hatching, the young females were immediately isolated and each placed in a separate Syracuse watch-glass with fresh culture solution. Observations were made twice a day on these young females and an effort was made to obtain within a few hours the hatching-time of the first offspring of each female. Thereafter throughout the experiment each individual was looked at twice daily and observations made. MICTIC AND AMICTIC FEMALES IN HYDATINA 445 During the fecund period the offspring of each female were re- moved twice daily. This was done so that if at the first counting any were overlooked they would be found and removed at the second. In the mode tables, the individuals were arranged in groups of 5's or 10's for convenience and to save space. The means used, however, in the calculation of the standard deviation and coefficient of varia- bility were obtained from all the individuals which had been carried out to the second decimal place. GENERAL LIFE HISTORY The non-sexual or amictic females multiply exclusively by par- thenogenesis. Their eggs carry the diploid number of chromosomes, are not capable of fertilization, and produce females, thus multiplica- tion by diploid parthenogenesis is carried on for many generations. However, at times from these eggs another kind of female hatches which produces small eggs that develop into males. These male- producing eggs carry the haploid number of chromosomes. The females which produce them are called mictic and are identical with the amictic females in outward appearance. Their eggs, however, are capable of being fertilized. If the eggs of the mictic females are fertilized, they produce instead of the haploid egg a larger, dark, thick-shelled egg which has the diploid number of chromosomes. This winter egg, as it is called, always develops into a parthenogenetic amictic female. The entire life of a female lasts usually about seven to eight days depending upon the temperature and other variable external condi- tions. The first swimming offspring appear from 32 to 57 hours after hatching of the mother. To all outward appearance, the mictic and amictic females of Hydatina senta are indistinguishable, but they differ markedly in certain features of their life histories, particularly in the periods of fecundity and the total length of life. EMBRYONIC PERIOD The hatching or the embryonic period lasts from the deposition of the egg until hatching. During this period embryonic development is taking place. The length of this period varies from 18 to 26 hours with the mean for 184 amictic females at 22.34 hours, for 113 unfer- tilized mictic females at 22.53 hours, and for 88 fertilized females at 21.40 hours. This shows that there is in all probability no difference in the hatching time. (See Table I.) 446 JOSEPHINE C. FERRIS PRE-FECUND PERIOD The pre-fecund or adolescent period is one of rapid growth. This period extends from the hatching of the egg to the beginning of the fecundity period as shown by the production of the first egg. The length of this period at a temperature of 16° to 17° C. varies from 32 to 57 hours with the mean for 184 amictic females at 41.66 hours, for 110 mictic females at 42.72 hours, for 80 mictic females whose eggs have been fertilized at 46.91 hours. (See Table II.) There was no mortality during this period of immaturity of the 385 individuals studied. FECUND PERIOD During the fecund or egg-laying period the eggs are laid one by one. The first were laid from 32 to 57 hours after hatching. After TABLE I Comparison of the embryonic or hatching period of amictic, unfertilized mictic, and fertilized mictic females. Number of hours 17 18 19 70 71 77 7S 74 7S 76 Total Amictic females 7 q 1 110 81 S7 76 P7 10 70 493 Unfertilized mictic females n 5 1 18 18 7 74 7S s 14 113 Fertilized mictic females 0 4 n 18 SS 13 11 5 3 1 88 Mean Standard deviation Coefficient of variability Mode Amictic females 22 34 ± 5105 3 1668± 0381 7 96± 1706 24 Unfertilized mictic females 22 53± 1355 213 ± 0957 9 49± 4266 23 Fertilized mictic females 21 40± 1126 1.57 ± 0795 7.33± 3615 21 the first egg, additional eggs were produced at varying intervals and this continued for a number of days. The parthenogenetic mictic female deposited 16 to 56 of the small male-producing eggs. The amictic female deposited 16 to 66 of the larger, female-producing eggs. The mictic female whose eggs had been fertilized deposited 3 to 26 of the large, thick-shelled fertilized eggs. The modal fecundity of 184 amictic females is 51 daughters and the mean 45.39 ± .3261. The modal fecundity for 110 mictic females is 46 male young and the mean 42.52 ± .5288. The modal fecundity for 88 mictic females w^hose eggs have been fertilized is 13 fertilized eggs and the mean 9.98 ± .3015. Thus the mictic and amictic females produce nearly the same mean number of offspring; 42.52 and 45.39 respectively. (See Table IV.) These results are much more nearly in agreement MICTIC AND AMICTIC FEMALES IN HYDATINA 447 W a m -S" < .-s GO O LT) CN O so LO T-H LO -f o ^H O o o o o o o o -^ 10 ON O -t o 10 oo cs oo ON IO — o O) 01 Ol '^ IO ' — i O 01 O O O O OO -t- r^ OJ O O 01 O sO 01 O T-l O O in o o 75 £ " c ~a -t O o o OO ^— i 00 H T3 O ^5 * — 01 r^ oo **> -t 10 ~t C — SO OO ^ .22 ^ 4J ij i-g -H „ OS QJ cti a> ^f 00 O oo' 01 oo' •d c *"}* O] o) so LO SO 0 ON "- O T— 1 CO rt — 13 -U -U II Tl •H rt > ro 1~^« *J 1) C/3-O rr; — f- ro O ^ •<*' 00 ro -t so ^ LO sO (s^ 0) If- SO cd OJ jj ij Tl Tl 4] ^ so 01 ^_i so r^ ON 35 sd jj 1 JJ JH C .- w-i -t-i en o o JV •£ '+J a c (— < S CO O CM S T-" OO 4-» (D rvi r^ co ^7 CO CM 00 CM OO ^ T-| OJ T-I CM "S s» t~~^ f^ f-^ f^*l CO O yi tn en ft OO O co T-H T-H CM •-1 -C -C JH s LO T-I -H>l CM O fM LO OO SO CM co O r^* LO LO r^ ,_^ •K. co so OO O sO •s, LO CO T— » T-H IO SO T-H CM T-I ON LO so O ^1 o -+1 s 13 LO T-H CM CO CM T-H II II II _-H CO CM CO CM 5^ CM O CM O -H -jJ CM OO LO LO — ' "C *H sO ON CO w | PQ »J* CO O OO T-H CO T-I CO t^* (~*> r^* ON -H t^. oo O r— l/~} __ ^ MD O 0 00 O NO -S « ^^ * ^H ^— 1 *^~* •g «£, o2 NO rs 00 O TH •ta s a 10 o 0 -H NO NO O OO O o vj" *^* lO IO OO TZ ^i i^J ON NC \O t~- NO OO O NO W**s» §i ^ X lO O »o If) 3 «• 10 r^ NO O -f I-1 00 NO • - T5 t-1 -^ s a lO CO O .y U *^ — . rt *™ -— • Jl^ u — o .^H *^ ^N ^ *•£ !_. MICTIC AND AMICTIC FEMALES IN HYDATIXA 451 as the amictic female. This means that the mictic female deposits its small male-producing eggs, on the average, in more rapid succession than the amictic females which produce the larger female -producing eggs. The unfertilized mictic female produces on the average one offspring every 1.2 hours; the amictic female produces on the average one offspring every 1.8 hours; the fertilized egg every 5.2 hours. These figures represent the mean fecundity divided by the mean duration of the fecund period. POST-FECUND PERIOD The post-fecund period extends from the deposition of the last egg until death of the individual. During the old age period the activities of the female gradually cease, structural degeneration sets in, and death follows usually about the seventh or eighth day. (See Table V.) In Table V is given the duration of the post-fecund period for all individuals of the amictic and unfertilized mictic females. Seventy- nine amictic females of 184 or 43 per cent died within 36 hours after deposition of the last egg; 37 per cent died within 24 hours, all died within 160 hours. Of the total 163 individuals having a post-fecund period 38 lived 25 hours, which was the mode and the commonest period of death; from 45-55 hours was the next commonest period. The maximum length of the period of old age for the amictics is 160 hours. For the unfertilized mictic it is 192 hours. The mortality rate reaches one maximal point at about the begin- ning of the period of old age, perhaps as the result of the exhausting effort of the production of the last eggs. Finally, towards the end of life, it rises to 100 per cent. A large proportion dies immediately after the period of egg-production. But those individuals which pass safely through this period live for some time; in such populations the old females are thick, heavy and sluggish in their movements. The mictic females cease to deposit eggs earlier, on the average, than the amictic females. Therefore the post-fecund period of the mictic females is extended, being on the average 59.29 hours, while in the amictic it is 53.94 hours. GENERAL DISCUSSION The amictic and unfertilized mictic females of Hydatina senta under controlled conditions differ considerably in length of life, rate of egg production, duration of the fecund and post-fecund period. In Hydatina senta, the difference in the length of life of the amictic and mictic female probably results largely from the differences in the metabolic rates. The amictic females live longer than the mictic females. (See Table VI.) 452 JOSEPHINE C. FERRIS t~~ 00 O ON CN 'Q VO t^ OO t— CN OO ~H ^ I"** ON) (~-^, f^ I^ CN O 8 t- -t t- -H •£, i-c CN g LO -H 00 — CN | t^- O t^- ""> •"3 t°*» f^*> t~**> ^O •£ 1-1 CN 8 CN lO i-H CN "i S^S* i t-* O *"•"- Os ^^ <"O 1-1 rN ^ £""•" f^i t~~-» ^^ ON fN I-H CN ^> s t~"~ ^—t r~* vo 00 *-i r~i TABLE ) of amictic 'a s Stf g 5 to QJ CU ~ rt rt £ CJ CJ If) +~> If) -t-> 1- CJ I- CJ o E o £ _C od -C rt O 0 0 O U !_ !-, 1- <1> 0) O) CU 3333 r 5 ^ 4 C 5 O H \ vC OC ^ ) ^ 4 vC O fN > CN \ ) I vC r^ > t^ 4 • v: i- | vC ( t- C* c 1 > NC ir X ) 1 ^ VC CN c 1 > NC > If H CN C If CN ) 1- > 1 1 vC f ) CN ) 4 vC t> > 1- 1 1 | vC CN V— > O 1 1 v vc fr CN > Tt > 1 ! vC T- ^ } 1— 1 1 \c CN CN > o 1 1 vC ^ ) 1- ) 1 VC ^ (N ) *^ 1 1 vC 0 ) 1- 1 C CN 1 CN ) I VC OC ) 1- ) C O ) i- ' (f C. •_ R rt C 4— " s ~ 4, S z c t C t ^ (4.) c "o "c '- CL .c 1— i ^: tn a 1-1 — 1 X | | o 4- c ioc 5 f ^ ) t^ t-. - f 4 > t- ex CN 5 "~ i ^ l- vC ^ - -t 5 i- 4 < t' * X** CN - c i > t^ if i— - C 3 H ) t- CN . rr ) 1 > r- -) - r< i r- i ) t~ if CN - C 3 I > r» (T • C*v ) 1 1 *- CN • i~ 4 ] i t- 1 t- - tr ) o 1 t^ ^ T— • ^ 4 i t~ CN CN • C 1 1 > r^ C ^ • CN ) 1 t~- ^ O • C 4 i— 1 > ! r^ O • CN I t^. c CN C ) 1 i t^ OC - t- 3 . pr v i^ i r^ i-~ ; '- $~* oc ^ . i/- ) I tr a. if - ^ - S ~~ o Z <. •4- •s. — •N 4 C 4_ a 5 n •_ ^-t-H O "3 <-t-H O *4H O c a a 1 X X X CJ o 2 21 S I Coefficient of variability -too OC >O CN -{^ 4^ vo o o 10 ON ^H oo CN i— i T3 C i-i ^. || 00 -t 10 0] ,-. to jH^ ?r ~* *J. ^ VO ON VO -f IO OO ON O ON •*+ <^5 -t* O CN i-! JJ -U JJ M TI ~n ~A ON IO -t 0) O CN 0} 2 OJ •1-1 S •4-J 900- 300- 15 30 TIME IN DAYS 45 FIG. 1. The total number of eggs produced during the lifetime of a fly. These oigmoid integral curves are obtained also when the data from other individuals are similarly plotted. Figs. 1 to 3 inclusive are curves drawn from the records (contained in Table I) for the same female. How often such spasms occur under uniformly stimulating conditions is unknown, but there is certainly no daily or weekly periodicity among different individuals," the data were not plotted simply as recorded. Guyenot (1913) remarks, ". . .la ponte se produit sous forme de decharges, causees par la surabondance des oeufs formes." The following means, consequently, were adopted for obtaining the curve shown in Fig. 2, a procedure which may be justified by a simple 462 HERBERT SHAPIRO hypothetical consideration, inadequate as it may be. The tendency to lay eggs in bunches has been noted. Thus, if at the height of its egg-laying period, a fly laid 120 eggs per day, it would not lay one egg regularly every twelve minutes as might be computed, but rather several eggs during this interval, and then none for perhaps the next half hour. For this reason one must have recourse to an integral curve, for considered from the standpoint of minutes, the egg-laying is irregular, but from the standpoint of days or weeks, it becomes quite regular. I OL Ld O. 120 100 80 a 60 40 20 10 15 30 35 40 45 20 25 TIME IN DAYS FIG. 2. The daily egg production. This curve was secured by graphical inter- polation from an enlarged plot of Fig. 1. A large integral curve similar to that shown in Fig. 1 was plotted on paper of dimensions 3| x 4 feet, and then, by use of the mirror tangentimeter described by Latshaw (1925), the number of eggs for each 24-hour interval could be pieced out of the curve with fair precision. The differential curve so secured rises to a maximum of about 130 eggs per day and then falls off gradually to zero at about the end of the fly's life. The area under the curve would represent the total number of eggs deposited. In the attempt to arrive at some general conclusions from the data, the figures were set up in a manner of which Table II is an example. No clean-cut generalizations sug- gested themselves from a study of such averages. With the apparent general similarity of the type of curve shown in Fig. 1 as drawn for RATE OF OVIPOSITION IN DROSOPHILA 463 different individuals in mind, it was decided to study individual cases, to determine whether some general relationship might be found into which they all would fit. It is found that if //In T is plotted against /, where / represents the time in days at which any given total (T) is attained, the points arrange themselves linearly, as may be seen by inspection of Fig. 3. It will be noted that the last four points are off the curve. This is due to the cessation, on the forty-fifth day, of egg-laying by the fly, which lived four days longer; the point for each of these post-laying days was calculated by using the same total, and these points arrange them- selves along another line. The curve shown in Fig. 3 has been drawn also for each of the other 92 females studied and gives an equally good fit for these too, except three others where the points scatter rather more widely on each side of the line. All the sets of data might be TABLE II Cross c : + d" X + 9 (with male) Total number of eggs laid Life span Day of cessation of egg- laying days 755 27 27 617 20 18 1176 41 28 1537 41 40 830 36 33 1097 35 32 Average 1002±109(A.D.) 35±2.5(A.D.) 29.7±1.2(A.D.) plotted in somewhat fan-shaped arrangement in the space of Fig. 3, those representing a smaller rate lying above the curve there drawn, those with a greater rate falling below. If instead of this the slope of each curve so obtained be plotted against the grand total of eggs produced, the points fall as found in Fig. 4. From Fig. 3 it follows that In T = at where / and T remain with the same meaning indicated previously, a and b representing respectively the slope of the curve and its inter- cept on the y axis. This yields an equation for the total number of eggs (7") already produced at any stated time (f): T = eat+b. 464 HERBERT SHAPIRO To get the change in total with time, the first derivative is taken dT _ dt ~~ (at + b)2 ' and to arrive at the time at which the rate of egg-laying is at its maximum, the second derivative is taken, set equal to zero, and solved - 2a(at (at + d*T dt2 For the fly whose data are plotted as shown in the figures, this maximum is thus calculated to be 5.7 days, when the values used for the constants a and b are 0.121 and 0.221, whereas the value found by inspection of Fig. 2 is about 7.5 days. Z j 10 15 20 25 30 TIME IN DAYS 35 40 45 FIG. 3. A typical curve, demonstrating the linear relationship between the variables indicated. The oviposition data are tabulated in the literature for the cases of some other invertebrates. These have been examined and found to fit the formulation just shown. They are as follows: Hyde (1921) reported what were considered three unusual cases of fecundity for mated D. melanogaster females, namely, totals of 1,613, 1,807, and 2,184 eggs. Similar values have been found quite regularly RATE OF OVIPOSITION IN DROSOPHILA 465 in the animals studied here, and with respect to the hybrids it is rather the rule for them to lay 2,000 to 3,400 eggs. Faure-Fremiet and Garrault (1928) give data for the egg production of Margaropus australis, an acarid. While studying the growth of the snail Lymnxa columella, Daily (1931) recorded also the egg-laying. The empirical equation applies also to the data presented by these authors, though not quite so well for some of the individuals of the last-mentioned instance as for Drosophila due to the more pronouncedly intermittent character of egg production in the snails; however, there is a definite fit. A curve plotted for one of the sets of data presented by each of 3000 2500 2000 o I- 1500 1000 500 • •-..i ••• . A .115 .125 .135 .145 .155 .165 .175 .185 SLOPE FIG. 4. Each point represents the slope (of a curve plotted as in Fig. 3) in relation to the sum total of eggs produced, for each of the flies studied. these authors is presented in Fig. 5, where it may be compared, for example, with the curve for the data derived from a female of cross a (D. obscura). It is to be pointed out here that it follows from what has been described that unless one is certain that females are producing eggs at the same rate, it is not a sound procedure to select flies at random from a stock and use some for experimentation dealing with egg production, and others for controls, and then to compare the averages of the results. This comparison might possibly lead to erroneous conclusions. From the nature of Fig. 2, it becomes evident that it 30 466 HERBERT SHAPIRO should be determined that the shapes of the egg-laying curves for both controls and experimental material, both as regards height of the maximum and length of the curve, are reasonably similar at the outset. In connection with this, an analogous situation may be cited. Daven- port (1931) inveighs against the procedure of drawing conclusions concerning growth processes from accumulated data. From the mass statistics of 100,000 children, one might decide that the velocity of growth is greatest at two periods, one in intra-uterine life, and one at about 14.5 years (in the case of the male). When, however, the study 10 10 20 30 40 TIME IN DAYS 50 60 70 FIG. 5. Comparison of the curves of various invertebrates. A, of a snail Lymncea columella (data of Baily); B, of D. melanogaster (data of Hyde); C, of D, obscura (Race A); D, of a tick, Margaropus australis (data of Faure-Fremiet and Garrault). In A the egg output amounted to 725; in D, 4346 eggs were produced. of individual children is made, the resulting curve of growth is found to be very different from that of the mass curves, and varies with different children. Instead of the maximum being reached rather gradually at the age of 14.5 years, a rapid growth of the individual at adolescence, of almost explosive rapidity, is found; the age at which this occurs, and its magnitude, varying with the individual. When one is dealing with a constantly changing quantity, and where this rate of change will vary for different flies at the same age, the averaging of data will give only a very approximate idea of what RATE OF OV1POSITION IN DROSOPHILA 467 is occurring in general in individual cases. If the proper flies be used, one may see that by averaging different kinds of curves such as are presented in Fig. 2, provided their maxima and duration be different, any kind of average curve, within limits, may be produced. This may perhaps account for the differences observed by Hanson and Ferris (1929), when in one experiment the averaged laying curve for mated white flies rose to a maximum of about 24 eggs per day, while in another similar experiment a maximum of about 52 eggs per day was reached. The maxima of averaged curves may be shifted depending on the nature of the maxima of the individual curves being averaged. Hanson and Ferris decide from their own data presented as aver- ages that mating results in heightened productivity by the female. There is lack of agreement as to whether or not, in general, the male serves to stimulate an increased production of eggs. Guyenot (1913) states that there is a considerable delay, which varies from case to case, in the initiation of egg-laying by certain Drosophila virgins, but that, once commenced, their production gradually approaches in inten- sity that of the mated individual. In the cases observed by the writer, the vestigial virgins started to lay eggs the second day after hatching, and produced eggs regularly thereafter, whereas type virgins delayed laying from two to eight days after hatching. It is of interest to note that examination of the data shows that certain of the virgins produced more eggs during their life than did mated flies, and this in females from a stock which had been well inbred; although the averages were greater for mated flies than for virgins. These figures, with the deviation of the mean (A.D.), are listed here; the numbers in paren- theses represent the number of individuals whose data are averaged. + virgins 677 ± 116 (8). + d" X + 9 (without male) 718 db 103 (12). + d" X + 9 (with male) 1091 db 115 (7). + d" X + 9 (with two males) 1402 ± 98 (4). Inasmuch as certain virgins, for example, will lay many more eggs than certain mated flies, one would hesitate to reach the generalization from these averages alone that mating stimulates the female to in- creased egg production, and cannot help suspecting that the increase may not be an intrinsic one. Eight type virgins laid the following numbers of eggs in the time in days indicated in parentheses: 165 (47), 378 (28), 400 (55), 454 (70), 698 (40), 758 (42), 1088 (70), 1474 (31). In view of the great individual variations, the averaging of the records of much greater numbers of individuals would be required to give a definite statistical answer. It might seem that a means may be 468 HERBERT SHAPIRO afforded for obtaining an indication of the course of events, in Drosoph- ila at least, by use of the equation described above. If mating results in increasing constantly and continuously the rate of egg production, then plotting as in Fig. 3 the data of an experiment where a virgin lays eggs for a given period and is then mated, there should be a break in the line toward the abscissa after the time of mating. Due, however, to the logarithmic nature of the plot, small deviations or changes in the rate do not become readily manifest; consequently this bend is not accentuated sufficiently to permit a definite and un- equivocal separation of the parts of the curve that would result. In an investigation of egg-laying in the domestic fowl, Faure- Fremiet and Kaufman (1928) advance the interesting hypothesis supported by histological evidence of what is termed a constant probability of transformation of oocytes, an interpretation entirely independent of the idea of senescence as offered by Brody, Henderson, and Kempster (1923). In seeking a factual basis for this idea, the first-mentioned authors studied the formation of oocytes in the hen, and the initial number of oocytes, and proposed an interpretation of the curve of laying according to such data. After a review of the experimental work done on this phase of the subject, it appears that the activity of the germinal epithelium can be restored in certain pathological or experimental conditions (Pearl, 1921), but they con- clude that in the hen, under normal conditions, the number of oocytes is quite limited after birth. Subsequent to birth, oocytes of the chick grow slowly and progressively. During the period of egg-laying (which may be of eight years' duration) some of the oocytes undergo a very rapid growth and increase their vitelline mass a hundred-fold in five to eight days, following which they may be laid. They propose then the following equation, similar to that of Brody and collaborators, to describe the decrease in rate of egg-laying: Nt = Ntie~Kt, where Nt - -- the number of oocytes still available at time t (reckoned in years), N0 = the initial number of available oocytes, e = the base of natural logarithms, and K = a constant. The average curve of decrease in a given race in an individual is said then to depend on two values: N0, the initial number, representing the stock of available oocytes; the other they term the probability of transformation of oocytes, a meaning which is attached to the value K, and l/K or 6, would represent the average life of the oocyte, or the average period during which the oocytes can remain at the initial state before under- going the very rapid growth of yolk accumulation. These constants RATE OF OVIPOSITION IN DROSOPHILA 469 are supposedly independent of environment, and are probably, accord- ing to the authors, hereditary. It is apparent that this equation is formally tantamount to that for a first order reaction, viz., TS 1 , a K = - In a — x where the rate of change of the concentration of substance A at any instant is proportional to its concentration at that instant; / represents time, a the original molar concentration, and (a •- x} the concentration of A after / minutes. When stated in terms of egg-laying this would mean simply, aside from any implications of a chemically analogous factual basis, that at any instant the rate of decrease of the number of eggs laid is proportional to the number of eggs remaining unlaid. This equation proposed by Brody, Henderson and Kempster, or the equivalent one of Faure-Fremiet and Kaufman, for averaged data of egg production by the hen, cannot be carried over to the situation presenting itself in Drosophila, where the rise in egg-laying to a maxi- mum is a regular, intrinsic part of the process. The equation of Faure-Fremiet and Kaufman can be applied only to the descending portion of the curve of laying of the tick Margaropus australis, whereas the equation for Drosophila can apply also to the entire curve for Margaropus. In the case of the arthropod, then, the hypothesis of a probability of transformation of oocytes appears to be inadequate. To bridge the discrepancy between the curve for laying of Margaropus and of the hen, Faure-Fremiet and Garrault (1928) introduce the con- ception of the progressive development of a "physiological factor," that is to say, of a complex of somatic conditions allowing yolk accumu- lation, as being probably responsible for the ascending portion of the curve. Inasmuch as a general equation is found which applies to the whole process, it is perhaps more desirable to conceive it as continuous, and operating throughout as part of the same mechanism, rather than to introduce the idea of a dichotomy, the operation of the second process remaining in abeyance until the completion of the first. However, the existence of such a factor as the first, not yet yielding to exact treatment, is by no means excluded. Grateful acknowledgment is due Professor D. E. Lancefield and Professor A. H. Sturtevant for reading and criticizing the manuscript. SUMMARY The rates of egg-laying of certain mutants of Drosophila melano- gaster and of two races of Drosophila obscura were studied and compared with certain other cases for which data are presented by the authors. 470 HERBERT SHAPIRO The fecundity of hybrids of Drosophila melanogaster was also studied. An empirical equation describes the egg-laying curves of all the flies studied (about 93 in number) and is T = eat+b where T represents the total number of eggs already laid at the time /, and e is the base of natural logarithms. The constants a are shown to be correlated with the total number of eggs deposited. LITERATURE CITED ADOLPH, E. F., 1920. Egg-laying Reactions in the Pomace Fly, Drosophila. Jour. Exper. Zool., 31: 327. BAILY, J. L., 1931. Some Data on Growth, Longevity, and Fecundity in Lymnsea columella Say. Biologia Generalis, 7: 407. BRIDGES, C. B., 1932. Apparatus and Methods for Drosophila Culture. Am. Nat., 66: 250. BRODY, S., E. W. HENDERSON, AND H. L. KEMPSTER, 1923. The Rate of Senescence of the Domestic Fowl as Measured by the Decline in Egg Production with Age. Jour. Gen. Physiol., 6: 41. CASTLE, W. E., F. W. CARPENTER, A. H. CLARK, S. O. MAST, AND W. M. BARROWS, 1906. The Effects of Inbreeding, Cross-breeding, and Selection upon the Fertility and Variability of Drosophila. Proc. Am. Acad. Arts and Sci., 41: 731. DAVENPORT, C. B., 1931. Individual vs. Mass Studies in Child Growth. Proc. Am. Philos. Soc., 70: 381. FAURE-FREMIET, E. ET H. GARRAULT, 1928. La courbe de decroissance de ponte chez Margaropus australis. Ann. de Physiol., 4: 218. FAURE-FREMIET, E., ET L. KAUFMAN, 1928. La loi de decroissance progressive du taux de la ponte chez la Poule. Ann. de Physiol., 4: 64. < .1 YENOT, EMILE, 1913. Etudes biologiques sur une mouche, Drosophila ampelo- phila Low. VII. Le determinisme de la ponte. Compt. Rend. Soc. Biol. Paris, 74: 443. HANSON, F. B., AND F. R. FERRIS, 1929. A Quantitative Study of Fecundity in Drosophila melanogaster. Jour. Exper. Zool., 54: 485. HUETTNER, A. F., 1924. Maturation and Fertilization in Drosophila melanogaster. Jour. Morph., 39: 249. HYDE, R. R., 1921. A High Fecundity Record for Drosophila melanogaster. Proc. hid. Acad. Sci., 31: 259. LANCEFIELD, D. E., 1929. A Genetic Study of Crosses of Two Races or Physiological Species of Drosophila obscura. Zeitschr. f. ind. Abstammungs- u. Verer- bungslehre, 52: 287. LATSHAW, M., 1925. A Simple Tangentimeter. Jour. Am. Chem. Soc., 47: 793. LAURINAT, K., 1931. Uber den Einfluss des Keimzellalters auf das Geschlechts- verhaltnis bei Drosophila melanogaster. Zeitschr. f. ind. Abstammungs- u. Vererbungslehre, 57: 139. MORGAN, T. H., C. B. BRIDGES, AND A. H. STURTEVANT, 1925. The Genetics of Drosophila. Bibliographic, Genetica, 2:1. NONIDEZ, J. F., 1920. The Internal Phenomena of Reproduction in Drosophila. Biol. Bull., 39: 207. PATTERSON, J. T., 1929. The Production of Mutations in Somatic Cells of Drosoph- ila melanogaster by Means of X-rays. Jour. Exper. Zool., 53: 327. PEARL, R. and W. F. SCHOPPE, 1921. Studies on the Physiology of Reproduction in the Domestic Fowl. XVIII. Further Observations on the Anatomical Basis of Fecundity. Jour. Exper. Zool., 34: 101. RATE OF OVIPOSITION IN DROSOPHILA 471 PEARL, R., 1926. A Synthetic Food Medium for the Cultivation of Drosophila. Jour. Gen. Physiol., 9: 513. RAU, P., 1910. Observations on the Duration of Life, on Copulation and on Ovi- position in Samia cecropia Linn. Trans. Acad. Sci. St. Louis, 19: 21. RICHARDSON, C. H., 1925. The Oviposition Response of Insects. U. S. Dept. of Agriculture. Department Bulletin No. 1324. 17 pp. STURTEVANT, A. H., 1915. Experiments on Sex Recognition and the Problem of Sexual Selection in Drosophila. Jour. An. Behav., 5: 351. STURTEVANT, A. H., 1921. The North American Species of Drosophila. Carnegie Inst. Wash., publ. 301. 150 pp., 3 pi. WARREN, D. C., 1924. Inheritance of Egg Size in Drosophila melanogaster. Genetics, 9: 41. A NOTE ON THE THYROID GLAND OF THE SWORDFISH (XIPHIAS GLADIUS, L.) WILLIAM H. F. ADDISON AND MAURICE N. RICHTER (From the Marine Biological Laboratory, Woods Hole, the Department of Anatomy, University of Pennsylvania, and the Department of Pathology, Columbia University) In teleosts, the thyroid gland is not so well defined as in elasmo- branchs and mammals. As a rule the follicles, in smaller or larger groups, are distributed within an abundance of soft connective tissue, and the boundaries of the thyroid tissue are formed by adjacent struc- tures, not by a distinct capsule. McKenzie (1884) in the siluroid, Ameiurus, describes the frame- work as consisting of loose connective tissue which does not form a limiting membrane, but merely passes over into the tissue surrounding the adjacent parts. The thyroid vesicles are scattered through this tissue, showing a tendency to arrange themselves in short rows. Gudernatsch (1911) made an extensive study of the distribution of the thyroid tissue in twenty-nine species of teleosts, belonging to twenty families, and found great variation in the compactness of the tissues, ranging from a complete dispersion of the follicles to a rather compact union of them. Compared with the fishes studied by Gudernatsch, the swordfish, Xiphias gladius, is noteworthy among the teleosts because of the concentrated character of the gland and its large size. In dissecting the heads of swordfish at the Marine Biological Laboratory at Woods Hole, we noticed that the main mass of the thyroid gland formed a large fairly well-circumscribed mass of tissue. It was situated in close relation to the cephalic end of the ventral aorta, and partially encircled it. The color was dark red, due to its great vascularity. Its consistency was moderately soft, but with care it could be separated from the surrounding structures. For con- venience in dissecting, it was usually left attached to the ventral aorta. The general appearance in cross-section and the relation of the thyroid to the ventral aorta are shown in Fig. 1. This specimen was from a medium-sized swordfish weighing 330 pounds after the head, tail, and viscera had been removed. After preservation in 10 per cent formalin the thyroid was a firm mass of a dark gray color. It measured 40 mm. in the sagittal direction and 35 mm. in its widest transverse part. The greatest thickness of the thyroid tissue was 17 mm. The organ was composed, for the most part, of four fairly well-defined 472 THYROID GLAND OF SWORDFISH 47.} PLATE I FlG. 1. Cross-section of the thyroid gland of the s\vordnsh partly surrounding the cephalic end of the ventral aorta. X 4. FIG. 2. A small portion of the thyroid gland of the swordfish, showing the compact arrangement of the follicles. X 200. 474 W. H. F. ADDISON AND M. N. RICHTER masses, which were separated from each other by thin connective tissue septa, continuous with the peripheral connective tissue. The two anterior masses, situated on either side of the median line, between the first and second branchial vessels, were larger than the two posterior masses. The latter extended between the second branchial arteries and the single stern for the third and fourth branchial vessels. Each of the four masses consisted of smaller closely-adherent masses or lobules, which were composed of compact thyroid tissue. Microscopically, the main gland mass shows epithelial-lined follicles containing colloid material (Fig. 2). Around the follicles is a very vascular but scanty supporting tissue in which are a few fat cells. Thus, in certain features, the gland resembles closely elasmobranch and mammalian thyroids. In transverse sections across the entire mass, separate small groups and rows of follicles were also seen. These were scattered in the loose connective tissue between the main part of the organ and the wall of the aorta and the branchial vessel, and showed the typical teleostean arrangement. The lumina of the follicles vary greatly in size. Of those containing colloid, a small one may measure 30 x 25 IJL, a large one 300 x 200^. The largest ones may measure several millimeters in length. One was found to be 3.5 x 1 mm., and another 2.5 x 0.75 mm. A great number, however, are of intermediate size, measuring about 120 x 80 ju. The general shape is round or oval, the large ones being usually elon- gated. The walls of the follicles are formed of simple columnar epithelium, which apparently does not rest on a basement membrane. The epithelial cells are columnar in form, and average 15x4.5/x after fixation in formol-Zenker. The cells are taller and narrower than the "chief" cells of Mustelus. According to Ferguson (1911) the latter measure 6-10 fj. in height; after fixation in formol-Zenker, how- ever, we find the Mustelus cells to be somewhat higher, 13-14 p.. They present many slight variations in shape, some being fusiform, others slightly curved. At places there is a pseudo-stratified appear- ance, because the nuclei of adjoining cells are alternately higher and lower in position in the cells. The cytoplasm, after staining with Dominici's stain or eosin-azur, is basophilic in the basal portion, and acidophilic towards the lumen. Sometimes the acidophilic zone is seen at the basal margin of the cells. This suggests a reversal of polarity in these cells. In a few cells, the portion of the cytoplasm adjoining the lumen contains acidophile substance in the form of globular masses of varying sizes. These globular masses resemble colloid in appearance. In Gudernatsch's study (1911), the "colloid" cells of Hurthle or of Langendorff were seldom seen. Typical ones THYROID GLAND OF SWORDFISH 475 are not seen here. The nuclei are usually situated in the center or in the basal third of the cell-body, and contain relatively little chromo- philic substance. In the follicles one usually finds homogeneous retracted colloid material. Within the colloid material in some, there are lighter staining spherical areas or vacuoles. Frequently, groups of ill-defined epithelial cells are also seen within the colloid, as well as hemorrhagic masses of blood cells. Such follicles often lie deeply in the lobule where they would be well protected from mechanical injury. Throughout the greater part of the gland, the epithelium is sepa- rated from the blood by only the endothelium of the thin-walled vessels. The blood-vessels, however, are numerous, and are of rela- tively large size. Lymph-vessels containing colloid are also seen. The amount of fibrous tissue between the follicles is very small. In addition to the interfollicular supporting tissue there are wider strands of connective tissue between the lobules. In these interlobular strands are sometimes rows of fat cells and single follicles, or small groups of them. Around the peripheral lobules, the connective tissue fibers are arranged in a parallel manner, somewhat closer together than in the adjoining loose connective tissue. After orcein staining, elastic fibers of small size are demonstrable. These are much thinner than those in the walls of the blood-vessels, and are arranged in a loose network. This peripheral connective tissue could scarcely be regarded as a true capsule, but it is somewhat modified from the ordinary loose connective tissue. The consistency of the organ is thus due more to the close arrangement of the follicles than to the presence of a definite peripheral covering membrane. Pigment cells are distributed in the loose connective tissue around the organ, and are often perivascular in position, in some places lying against the walls of the blood-vessels. Studies in the thyroid of fishes have revealed several interesting conditions. In the thyroid of the dogfish, A. T. Cameron (1913) finds by chemical analysis that the iodine content is higher than in that of any mammalian thyroid yet examined. Marine and Lenhart (1911) find that hyperplasia of the thyroid tissue develops in brook trout (Salvelinus) kept under certain conditions, and that iodine stops the hyperplasia and causes the thyroid to return to the colloid or resting state. For chemical or physiological studies of the teleostean thyroid it would appear that the swordfish thyroid, on account of its large size and the homogeneous structure of the main part of the organ, would afford satisfactory material. 476 W. H. F. ADDISON AND M. N. RICHTER LITERATURE CITED CAMERON, A. T., 1913. Note on the Iodine Content of Fish-thyroids. Biochem. Jour., 7: 466. FERGUSON, J. S., 1911. The Anatomy of the Thyroid Gland of Elasmobranchs. Am. Jour. Anat., 11: 151. GUDERNATSCH, J. F., 1911. The Thyreoid Gland of the Teleosts. Jour. Morph., 21: 709. MARINE, D. AND C. H. LENHART, 1911. Further Observations and Experiments on the So-called Thyroid Carcinoma of the Brook Trout (Salvelinus fontinalis) and its Relation to Endemic Goitre. Jour. Exper. Med., 13: 455. AIcKENZiE, T., 1884. The Blood-vascular System, Ductless Glands, and Uro- genital System of Amiurus catus. Proc. Canad. Institute, Toronto, 3d Ser., 2: 418 (thyroid, pp. 434-435). INFLUENCE OF HYPOPHYSECTOMV ON THE PANCREATIC DIABETES OF DOGFISH OSCAR ORIAS1 (From the Marine Biological Laboratory, Woods Hole, and the Laboratories of Physiology in the Harvard Medical School) Using toads and dogs, Houssay and Biasotti (1930a, 19306) have shown the marked influence of hypophysectomy on the course of the diabetes produced by pancreatectomy. In their pancreatectomized animals the diabetes was definitely milder when the hypophysis had been removed. It was the purpose of this investigation to determine whether or not the general conclusions reached by these investigators were applicable to a lower form. The smooth dogfish (Mustelus canis) was chosen on account of its availability and its position in the vertebrate scale. Furthermore, its cartilaginous skull renders hypo- physectomy a comparatively simple procedure. PANCREAS, HYPOPHYSIS, AND GLYCEMIA ix ELASMOBRANCH FISHES Elasmobranchs have a large pancreas, its tissue being made up of glandular acini and of insular cells functionally equivalent to the islets of Langerhans in the pancreas of mammals (Jackson, 1922). The isolation and removal of the entire pancreas are procedures easily performed without any serious bleeding. Herring (1911) has shown that the hypophysis of elasmobranchs is developed almost entirely from Rathke's pouch. This same investigator (Herring, 1913) states that "the elasmobranch pituitary differs from all other pituitaries in not possessing a posterior lobe. The brain wall of the embryo merely evaginates to form a paired saccus vasculosus, but no pars nervosa is formed." And although de Beer (1926) showed that there is an exten- sion of neuroglia fibers from the border of the infundibular cavity which penetrates the posterior lobe (pars intermedia), it is certainly true that the selachian lacks a true pars nervosa. This means that one has to deal in this form with a hypophysis composed of (a) an anterior lobe of eosinophile and basophile cells, (b) a posterior lobe (pars intermedia) of basophile cells, and (r) a ventral lobe composed also of basophile cells, but containing, in addition, certain curious large cells staining with eosin and of undetermined significance (de Beer, 1926). In the hypophysectomies of the present investigation 1 Fellow of the Rockefeller Foundation from Argentina. 477 478 OSCAR ORIAS no attempt was made to discriminate between these various parts. All removals of the gland have been total, and were performed accord- ing to the technique described by Lundstrom and Bard (1932). The normal glycemia in fish has not yet been satisfactorily worked out. All the recent studies dealing with the blood sugar of fishes have shown great discrepancies between the values found not only among specimens of different species, but also among different individuals belonging to the same species. Although no one has been able to determine precisely all the factors responsible for these discrepancies, undoubtedly the variable degrees of asphyxiation involved in the process of obtaining the blood samples for analysis stand as a cause of paramount importance in explaining the widespread variability of results. There is no doubt that asphyxia induces a condition of hyper- glycemia which may last during a period of several days even when the fish is replaced in sufficiently oxygenated water. This has been shown by McCormick and Macleod (1925), Simpson (1926), and Kisch (1929). The asphyxial blood-sugar rise is due to a mobilization of glycogen from the liver (Simpson, 1928; Kisch, 1929). Apart from the influence of asphyxia, the blood-sugar level seems to vary with the different species of fish because of the differences in their habits of life. According to the investigations of Gray and Hall (1930), who studied fifteen species of teleosts, fast-swimming fishes depending on the speed of their movements to catch their prey have a higher blood-sugar level than those less active bottom-feeders that live on crustaceans and other slow-moving creatures. Before proceeding to the main problem, an attempt was made to determine the normal blood-sugar level of Mustdus canis. Blood samples were directly withdrawn from the heart by means of a syringe with the needle inserted through the ventral median line at the anterior edge of the pectoral girdle. As a rule 0.5 cc. of blood was used for each determination. The reducing substances of blood after precipi- tation of proteins were determined by the Shaffer and Hartmann method (1921) and computed as glucose. Blood-sugar determinations were performed on each fish used (total of 39) in this work immediately after its arrival in the laboratory from its place of capture in traps situated out in the ocean. The values obtained ranged between 72 and 250 milligrams of glucose per 100 cc. of blood. Sometimes very low values were recorded (from 0 to 50 milligrams per cent), but animals with such low blood-sugar figures were either moribund, or, if apparently normal, they invariably died within a few hours. In this respect our observations entirely agree with those of Scott (1921). After this initial blood-sugar deter- HYPOPHYSECTOMY AND PANCREATIC DIABETES 479 mination, in order to let the fish recover from the effects of asphyxia and rough handling incident to its capture, it was placed in an im- mersed floating cage, several meters off shore, exposed to tides and marine currents. After 48 hours, a new sample of blood was with- drawn as quickly as possible and the amount of sugar determined. The figures thus obtained in 10 specimens of Mustelus canis ranged between 65 and 137 milligrams of glucose per 100 cc. of blood, the average being 105 ± 5 milligrams, and the standard deviation ± 22 milligrams. No food was given to the animals. These results, as well as the data found in the literature regarding the blood-sugar level of elasmobranchs, seem to show a clear difference between the amount of glucose in the blood of the species of the genus Mustelus (smooth dogfish) as compared with that in the blood of species of the genus Sqnahts (spiny or horned dogfish). In effect, the "weighed average" of all the analyses on fish of the genus Mustelus reported by Fandard and Ranc (1914), Scott (1921), Menten (1927) and Fremont-Smith and Dailey (1932), turns out to be 99 milligrams of glucose per 100 cc. of blood, for a total of 34 animals examined. Denis (1922), omitting to say how many animals she investigated, reports for the blood of Mustelus canis amounts of glucose ranging from 80 to 181 milligrams per 100 cc., the majority of her results falling between 90 and 110 milligrams. In so far as the genus Squalus is concerned, the "weighed average" of the analyses performed by Claude Bernard (1877), Lang and Macleod (1920), and White (1928), is of the order of 36 milligrams of glucose per 100 cc. of blood for a total of 15 animals examined. It is impossible to account for this difference on the basis of the view that the difference in the blood-sugar levels of these two genera is related to their different habits of life in the way pointed out by Gray and Hall for teleosts. While according to Bigelow and Welsh (1924), the spiny dogfish (Squalus) is a strong, fast-swimming animal, the smooth dogfish (Mustelus} is a bottom fish, feeding principally on crustaceans. Elasmobranchs of the genera Torpedo and Scylliiiw, which according to the analyses of Diamare (1905 and 1906), Diamare and Montuori (1907), and Kisch (1929) have also lower amounts of blood sugar than Mustelus, are slow-moving animals of the sea bottom (Couch, 1868). PANCREATIC DIABETES IN FISH Capparelli (1894) was apparently the first to study the effect of pancreatectomy on fish. He removed the pancreas from eels, and was able to find marked glycosuria as a consequence. Diamare (1905, 1906, and 1911), working on elasmobranchs, found considerable 480 OSCAR ORIAS amounts of sugar in the blood of Scyllium and Torpedo after removal of the pancreas, whereas he was unable to detect any sugar in the blood of these animals before the operation. Probably Diamare and subsequently Diamare and Montuori (1907) failed to find sugar in the normal blood of Scyllium and Torpedo because of the inadequate methods available at that time, as they themselves suggested. More recently McCormick and Macleod (1925) found in Myoxocephalus (sculpin) marked hyperglycemia as a consequence of the ablation of the principal islets, easily removed in this animal, leaving intact the remaining pancreatic tissue. Simpson (1926) working on Myoxo- cephalus and Ameiurus confirmed the observations of McCormick and Macleod. EXPERIMENTS AND RESULTS Limiting the present investigation to animals of the same species and following as uniform a procedure as possible in handling the ani- mals before, during, and after the operations, it is possible to a certain extent to make "constants" out of the several factors, known and unknown, which modify the blood-sugar level, aside from the experi- mental conditions created for purpose of the study (pancreatectomy, hypophysectomy, etc.). Finally, a statistical treatment of the data obtained will enable us to get a more complete idea of the significance and validity of the differences between the average values obtained under the different experimental conditions (Dunn, 1929). Medium-sized animals were chosen (from 70 to 90 centimeters long), regardless of sex, but pregnant females were rejected. Blood- sugar determinations were always performed by Shaffer and Hart- mann's method (1921). The first step of our procedure was always to withdraw a sample of blood from the heart as previously described, for a blood-sugar deter- mination. Then the necessary operations were performed, with care to avoid asphyxia as much as possible. A constant flow of sea water was maintained through the mouth and gills and under these circum- stances the respiratory movements proceeded in normal fashion. The pancreas was removed through an abdominal incision about three centimeters long and the abdominal wall was subsequently closed in layers by silk sutures. The hypophysis was extirpated through a buccal approach. No anesthetic was used. The longest operation lasted about twenty minutes. After they had been operated the animals were treated as previously described. Forty-eight hours after operation a new sample of blood was secured and another blood-sugar determination performed. The forty-eighth hour after operating proved to be a critical juncture. HYPOPHYSECTOMY AND PANCREATIC DIABETES 481 ~ -a SJ **-> ~ ~ « t w £ •*• ffl fel < s £ *£. j_, i-i OJ 3 " O - ^ 0 ^ -. "i\ IO 0 , -i OO O "O "* s T) cs o r-H m OO >, >, S g ir> 00 i~— C ^^ 10 O to r^ Tj J^ >o 1 I <^ ^^S'^°S$^^^ 00 OO 41 ^ -H ^•«2 CM 41 J3 c3 2 es "5, >- o *-* CM a 1 0)-- Ig IO LO *O LO 'O I"— ^ 1O t^» ^O j-^. ^H t^- r*-j ^O ^ ^ — ' ^O ^ 41 JJ w a c. s§ Tl Tl o •^ CO ^ t fc 1 -HOOCO^OOOOO 41 CN IT) cs o *O ^O *-O 5 O] 41 5 (N W | ^ Tl cn „ C-4 1 1 0 0 s O -w O -H 00 Os 41 vO 41 ' OO O *— ( S a r- 0 41 0 C, cu.o ° s ^toooo^^ito +! r3 IO ^ "^ § P >0 t" - — . *D "~ ^H 41 41 0 '^ Tl Tl en v-H IH S o 0 10 LO UO UO '0 "0 MD ^ ^ 00 ^ 4ija rt\ 41 41 p1 o CO ^ 41 es o " 0 Pi si[ns3^{ junpi.vipuj W JJ U5 GH ** o OJ o s "O LH y faO 4J s OJ C O sc < -M £ 5 482 OSCAR ORIAS Before that time the effects of asphyxia and rough handling inherent in the operations were still too marked, and beyond the 48 hours the mortality began to be rather high. The animals fall into five groups. The results show respectively the separate effects on blood sugar of (a) simple laparotomy, (&) hypophysectomy, (r) pancreatectomy, (d) pancreatectomy and hypo- physectomy, and (e) pancreatectomy and injury of the hypothalamic region of the brain. The operations were performed in such a way as to permit having animals of different groups simultaneously exposed to the same environmental conditions. As Table I shows, neither the laparotomy alone, nor the simple hypophysectomy, exerted any significant influence on the glycemia. The values, of course, were above those considered normal, but of the same order of magnitude as the values found before the operation. The removal of the pancreas, as was to be expected, caused a marked increase of blood sugar values: from 159 ± 6 milligrams of glucose per 100 cc. of blood as the average for 9 unoperated animals, to 402 ± 7 milligrams per 100 cc. of blood as an average for the same animals 48 hours after the operation. When both pancreas and hypophysis had been removed in the course of the same operation, a condition of hyperglycemia also en- sued, but the average for this whole group as well as the individual figures were lower than those encountered when the pancreas alone was taken out. That this difference (-• 114 ± 19 milligrams) which is certainly significant, is due to the absence of the pituitary body and not to the influence of some direct nervous factor brought into play by the operative traumatism, is demonstrated by the fact that in animals of the fifth group in which the pancreas was removed, leaving intact the hypophysis but injuring the adjacent nervous tissue (hypo- thalamus), the blood-sugar values were even higher than those found when the pancreas alone was extirpated. Shortage of time and animals prevented the study of the action of pituitary grafts and the action of pituitary extracts, but the data here reported support the conclusion that in the dogfish (elasmobranch fish) just as in the toad (batracian) or in the dog (mammal) the hypophysis exerts an aggravating influence on pancreatic diabetes, the mechanism of which is still obscure. SUMMARY Laparotomy or hypophysectomy does not change the blood-sugar level in the dogfish. Pancreatectomy produces a marked hyper- glycemia. The hyperglycemia is, however, less marked if pancrea- HYPOPHYSECTOMY AND PANCREATIC DIABETES 483 tectomy is accompanied by hypophysectomy, but it is slightly more marked if in addition to pancreatectomy the hypothalamus is injured. In conclusion I wish to express my indebtedness to Dr. Philip Bard for valuable help and suggestions. BIBLIOGRAPHY BERNARD, CLAUDE, 1877. Legons sur le Diabete et la Glycogenese animale. J. B. Bailliere et fils, Paris, p. 204. BIGELOW, H. B., AND \V. \V. WELSH, 1924. Bull. U. S. Bureau of Fisheries, 40: 1. CAPPARELLI, A., 1894. Arch. Hal. de Biol., 21: 398. COUCH, J., 1868. A History of the Fishes of the British Islands, London. DE BEER, G. R., 1926. The Comparative Anatomy, Histology, and Development of the Pituitary Body. Edinburgh. DEMS, W., 1922. Jour. Biol. Chem., 54: 693. DIAMARE, V., 1905. Zentral.f. Physiol., 19: 545. DIAMARE, V., 1906. Zentral. f. Physiol., 20: 617. DIAMARE, V., 1911. Arch. Hal. de Biol,, 55: 97. DIAMARE, V., AND A. MONTUORI, 1907. Rend. d. R. Accad, d, Sci. Fis. e Afutcin. di Napoli, 13: 348. DUNN, H. L., 1929. Physiol, Rev., 9: 275. FANDARD, L., AND A. RANC, 1914. Compt. Rend. Soc. Biol., 71: 68. FREMONT-SMITH, F., AND M. E. DAILEY, 1932. Biol. Bull., 62: 37. GRAY, I. E., AND F. G. HALL, 1930. Biol. Bull, 58: 217. HERRING, P. T., 1911. Quart. Jour. Exper. Physiol., 4: 183. HERRING, P. T., 1913. Quart. Jour. Exper. Physiol., 6: 73. HOVSSAY, B. A., AND A. BIASOTTI, 1930a. Comp. Rend. Soc. Biol., 104: 407. HOUSSAY, B. A., AND A. BIASOTTI, 19306. Comp. Rend. Soc. Biol., 105: 121. JACKSON, S., 1922. Jour. Metabol. Research, 2: 141. KISCH, B., 1929. Biochem. Zeitschr., 211: 276. LANG, R. S., AND J. J. R. MACLEOD, 1920. Quart. Jour. Exper. Physiol., 12: 331. LUNDSTROM, H. M., AND P. BARD, 1932. Biol. Bull., 62: 1. McCoRMiCK, N. A., AND J. J. R. MACLEOD, 1925. Proc. Roy. Soc., Series B, 98: 1. MENTEN, M. L., 1927. Jour. Biol. Chem,, 72: 249. SCOTT, E. L., 1921. Am. Jour. Physiol., 55: 349. SHAFFER, P. A., AND A. F. HARTMANN, 1921. Jour. Biol. Chem., 45: 365. SIMPSON, W. W., 1926. Am. Jour. Physiol., 77: 409. SIMPSON, W. W., 1928. Quart. Jour. Exper. Physiol,, 19: 197. WHITE, F. D., 1928. Jour. Biol. Chem., 77: 655. 31 MELANOPHORES INDUCED BY X-RAY COMPARED WITH THOSE EXISTING IN PATTERNS AS SEEN IN CARASSIUS AURATUS1 GEORGE MILTON SMITH ANATOMICAL LABORATORY, SCHOOL OF MEDICINE, YALE UNIVERSITY In a recent publication (Smith, 1932) it was shown that if gold- fishes (Carassius auratus] were exposed to X-rays, an eruption of corial melanophores occurred varying greatly in intensity in different fishes. In some fishes a general cutaneous melanosis resulted, leading even to death of the fish. Generally speaking, the X-ray eruption of melanophores is a transient affair, appearing about the fifth day after exposure to X-ray or somewhat later. Newly formed melanophores increase rapidly in numbers to form pigmented areas often visible to the eye. By a process of degeneration, these melanophores disappear, and there results a restoration to normal coloring. At water temperatures of 70° F., depigmentation consumes roughly from two to four weeks, after which the cutaneous regions once more assume a normal color. X-ray-produced melanophores in the goldfish behave, therefore, in much the same manner as melanophores produced in the same type of fish by trauma or in the healing of wounds or fractures as noted in earlier experiments (Smith, 1931). It became of interest to learn what relationship, if any, melano- phores newly produced from X-raying held toward groups of melano- phores already existing in the goldfish in the form of cells massed to form a definite pattern of the body, head, or fins; and further, to note any evidence of degeneration in existing pattern cells following radia- tion. Forty-five goldfishes, possessing various black patterns for the most part resembling those seen in Plate I, Fig. 1, were studied after exposure to 7 human erythema units of X-ray (sufficient usually to induce melanophores in the goldfish), and the behavior of melano- phores in patterns and melanophores formed by X-raying were com- pared. The technique of X-raying, found satisfactory in earlier experi- ments, was the following. The goldfish, anaesthetised in a solution 1 Aided by Grant from Blossom Fund. 484 X-RAY INDUCED MELANOPHORES IN GOLDFISH 485 of chloretone 1 to 2000 of water, was removed from this solution and placed on a folded towel directly under the X-ray tube with the entire left side facing directly upward. One unit of human erythema dose consisted of 100 k.p.v., 5 milliampere, 8-inch target-skin distance, no filter, 72 seconds exposure. Seven erythema units involved, therefore, an exposure of 504 seconds. After exposure, the fishes were kept under conditions of ordinary laboratory light in tanks of still water (70°-78° F.) supplied with a current of air. The X-raying was done through the courtesy of Dr. William LaField, Mr. E. E. Furbush of the New Haven Hospital, and Dr. Samuel Atkins of St. Mary's Hospital, Waterbury. The following are two illustrative experiments, in which X-ray eruptions were intense enough to permit photographing. Experiment 1. Goldfish, 5 cm. in length from snout to base of tail (Plate I, Fig. 1, Fish A, photographed before exposure to X-ray) with markings of massed melanophores on head and fins. This fish received 7 human erythema units of X-ray, exposing entire left side of fish. Eleven days later an active development of melanophores occurred in the exposed surfaces of the fish. Fourteen days after exposure the X-ray eruption appeared to have reached its height and the fish was photographed (Plate I, Fig. 2). Dense masses of X-ray-induced melanophores occupied chiefly the left side of head, body, and fins, and closely encroached upon the periphery of black pigmented patterns. Depigmentation of X-ray-induced melanophores began approximately 3 weeks after exposure. Plate I, Fig. 3, shows fish 27 days after exposure with depigmentation greatly advanced. The head region has cleared, leaving the pattern undisturbed. Complete disappearance of X-ray melanophores was noted on the fifty-fifth day after exposure. Plate I, Fig. 4, was taken on the seventy-ninth day after exposure and shows head pattern practically unchanged. The pigmented markings of the fins showed microscopically no apparent difference from the original arrangement as seen before X-raying. Although melanophores composing pre-existing patterns remain usually undisturbed by exposure to X-rays in doses of 7 human erythema units, as in the above experiment, in three instances there seemed to be definite evidence of an induced degeneration of melano- phores composing a pattern, and the following illustrative experiment describes such a degeneration of pattern following a typical X-ray eruption and depigmentation. Experiment 2. — Goldfish measuring 6 cm. from tip of snout to base of tail received 7 human erythema units of X-ray, the entire left side of the fish being exposed to the X-ray tube. Plate II, Fig. 1, 486 GEORGE MILTON SA1ITH is a photograph of this fish 19 days after exposure, showing a marked massing of melanophores, XR., on the left side ot the head and operculum, approaching in distribution close to the small head pattern P. At the periphery of the head pattern, P., a close intermingling occurred of both pattern cells, P., and X-ray melanophores, XR., seen in higher magnification in Plate II, Fig. 2. Degeneration of melanophore masses induced by X-ray began in the third week, and advanced by the twenty-seventh day to a complete clearing of the head region, the upper part of the operculum alone showing still massed X-ray-induced pigment cells, XR. (Plate II, Fig. 3). The head pattern, P., at this time retained the details of its original form except at the caudal tip of the pattern where there was noted some degeneration of melanophores adjacent to an opaque zone (Plate II, Fig. 3, 0-0'), a point where xanthophores had also disappeared. Plate II, Fig. 4, taken 42 days after exposure, shows the head pattern partly degenerated. Plate II, Fig. 5, 55 days after exposure, shows the head pattern no longer existing; and it was noted that the black pigmented markings on the various fins had disappeared to a very large extent. This particular fish, relatively sensitive to X-ray, showed beside melanophore degeneration, within the first two weeks a degeneration of xanthophores in several areas indicated in the photographs by the letter 0. Zones of xanthophore degeneration appeared in life as streaky grayish opaque areas, confined chiefly to the left side of the body, which had been directly exposed to X-ray. Ten fishes failed to give any eruption whatever of melanophores after exposure to seven human erythema units. A month later me- chanical injury was produced by crushing the left operculum with an artery clamp. Five days later numerous melanophores developed near the wound in each fish, temperature of water being 76° F. It was believed that in these experimental fishes, X-ray injury was not severe enough to elicit a melanophore reaction. EXPLANATION OF PLATE I FIG. 1, A, B, C, are types of goldfishes employed in these experiments, with black pigment patterns. Fish A, photographed on day before X-raying. FIG. 2, Fish A., 14 days after X-raying showing pigmentation from X-ray- induced melanophores (XR.) chiefly on left side or exposed side, encroaching upon region of existing pattern of head, P. FIG. 3, Fish A., 27 days after raying. Depigmentation of X-ray-induced melanophores greatly advanced, eruption showing only at points XR. FIG. 4, Fish A., Photograph shows fish A on the seventy-ninth day with head pattern intact. Complete disappearance of X-ray eruption occurred on the fifty- fifth day. X-RAY 1XDUCKD MELAXOPHORES IX GOLDFISH 487 XR, xn. '9*- p. PLATE I 488 GEORGE MILTON SMITH 0. 0, PLATE II X-RAY INDUCED MELANOPHORES IN GOLDFISH 489 COMMENT Results of experiments indicated that melanophores of already existing black patterns were for the most part not influenced by single doses of X-ray as high as seven human erythema units. At the periphery of such patterns an active development of new melanophores from X-ray exposure might occur here and there, so that pattern melanophores and what may be designated X-ray melanophores grew in close apposition (Plate II, Fig. 2) with much intermingling at the time of the height of the X-ray eruption. Disappearance of the X-ray melanophore eruption left the melanophores of the pattern usually in an intact condition without alterations in the morphology of the cells. Thus, melanophores of two kinds were found to exist in the same fish, reacting differently to the effect of X-ray. On the one hand, melanophores composing existing patterns usually remained stabile, and rarelv degenerated. On the other hand, masses of new melano- phores evoked by X-ray, pursued a comparatively short life cycle of active growth and early degeneration with complete subsequent de- pigmentation. Degeneration of pattern cells was noted definitely only three times among the 45 fishes studied, as illustrated in Experiment 2 cited above. As seen in this experiment, the head pattern slowly degenerated after disappearance of the X-ray-induced melanophore eruption. An extensive though incomplete degeneration of the fin patterns occurred simultaneously with that of the head pattern. In this particular fish, areas of degeneration of xanthophores were noted as well. It is riot unlikely that further investigation will show that a dosage somewhat higher than 7 human erythema units will be neces- % EXPLANATION OF PLATE II FIG. 1. Shows goldfish referred to in Experiment 2 with X-ray eruption (XR.) 19 days after raying. X-ray eruption (XR.) and head pattern (P.) are in close apposition. Letter 0 points to an area of xanthophore degeneration. FIG. 2. A higher magnification of a part ot the field in Fig. 1. XR points to eruption of massed melanophores induced by X-ray. P. represents the massed melanophores forming existing pattern of head. Melanophores of eruption and pattern intermingle at the periphery of pattern P. Magnification X 6. FIG. 3. Same fish as in Fig. 1 with partial depigmentation of X-ray eruption of melanophores 27 days after X-ray exposure. The head region has cleared and shows no eruption of melanophores except at the upper part of the operculum (XR.) Pattern P. is intact except for slight degeneration posteriorly at 00' where xantho- phores have also disappeared. FIG. 4. Same fish 42 days after raying. Head is cleared of X-ray eruption; pattern P. is degenerating. Xanthophores have degenerated at point marked 0. FIG. 5. Same fish 55 days after raying. Head pattern has degenerated com- pletely. Xanthophore degeneration at point marked 0. 490 GEORGE MILTON SMITH sary to produce uniformly degeneration of pattern melanophores of this fish, influenced by weight and size of fish. The melanophores following X-ray exposure may bear a close morphological resemblance to the melanophores of an existing pattern, so that the two types are distinguished often with difficulty. Usually, however, X-ray melanophores look more delicate and smaller than pattern melanophores; their processes are more irregular and reach out into different planes in the tissue spaces. The pattern melano- phores appear more flattened as they lie at rest spread out immediately beneath the transparent epithelium. Their borders with processes parallel to the surface appear more sharply circumscribed and deeper pigmented especially when in a somewhat contracted state. As Fukui (1927) and Goodrich and Hanson (1931) have shown, the young goldfish is normally dark colored as the result of the presence of melanophores. Depigmentation begins irregularly after a few weeks of life and the fish gradually assumes a yellowish, golden color. The extent and completeness of depigmentation determines the pattern of adult conditions, subject probably to still further slow changes in black pigmentation later in life. Melanophores of patterns are prob- ably fully differentiated cells, and closely affiliated with the nervous system as shown by Ballowitz (1893), von Frisch (1911) and other investigators; whereas melanophores evoked by X-ray, or by mechan- ical injury function perhaps in behalf of body defense and repair, when certain chemical conditions are produced in the corium of gold- fishes possessing potential pigment-forming cells. In the present experiments, areas composed of pattern cells did not seem to develop new X-ray-induced melanophores to any extent except temporarily at the periphery of the pattern. This fact suggests that conditions did not exist in the central parts of the patterns for the development of new pigmented cells, under the conditions of dosage employed, or possibly that the massed flattened pre-existing pattern cells offered enough protection against the effects of X-rays to inhibit the formation of new pigmented cells. SUMMARY In the goldfish exposed to X-ray (7 human erythema units) existing patterns remained for the most part intact in the presence of an induced temporary eruption of corial melanophores caused by X-raying. In several fishes, however, a degeneration and disappearance of the patterns, partial or complete, was noted, and this followed after depigmentation of an eruption of X-ray-induced melanophores. X-ray thus produced two effects relative to melanophores, (a) an eruption X-RAY INDUCED MELANOPHORES IN GOLDFISH of new melanophores with a short life cycle, (b) occasional degeneration of melanophores in existing patterns. LITERATURE CITED BALLOWITZ, E., 1893. Die Nervendigungen der Pigmentzellen, ein Beitrag zur Kenntnis des Zusammenhanges der Endverzweigungen der Nerven mit dem Protoplasma der Zellen. Zeitschr. f. Wissenschaft. Zool., 56: 673. VON FRISCH, K., 1911. Beitrage zur Physiologie der Pigmentzellen in der Fischhaut. Arch.f. ges. Physiol., 138: 319. FUKUI, K., 1927. On the Color Pattern produced by Various Agents in the Goldfish. Folia. Anat. Jap., 5: 257. GOODRICH, H. B. AND I. B. HANSEN, 1931. The Postembryonic Development of Mendelian Characters in the Goldfish, Carassius auratus. Jour. Exper. Zool., 59: 337. SMITH, G. M., 1931. The Occurrence of Melanophores in Certain Experimental Wounds of the Goldfish (Carassius auratus). Biol. Bull., 61: 73. SMITH, G. M., 1932. Eruption of Corial Melanophores and General Cutaneous Melanosis in the Goldfish (Carassius auratus) Following Exposure to X-ray. Am. Jour. Cancer, 16: 863. INTRACFLLULAR CRYSTALLIZATION OF HEMOGLOBIN IN THE ERYTHROCYTES OF THE NORTHERN PIPEFISH, SYNGNATHUS FUSCUS ALDEN B. DAWSON (From the Zoological Laboratories, Harvard University, and the Marine Biological Laboratory, Woods Hole, Mass.) It has long been recognized that the hemoglobins of different animals vary widely in solubility and ease of crystallization and that hemoglobin rarely crystallizes within the red blood corpuscles. In a previous paper (Dawson, 1930) intracellular crystallization of hemo- globin was described for the erythrocytes of the urodele, Nectunts maculosiis. In this case crystallization was apparently favored by previous poisoning with lead acetate, although it had been occasionally encountered in normal animals. A similar phenomenon has been observed in the erythrocytes of the northern pipefish. The animals were obtained in the Eel Pond, at the Marine Biological Laboratory, Woods Hole, and appeared to be normal in all respects. The observation was made incidentally while studying supravitally the blood cells of the common marine fishes of that locality. The crystallization of hemoglobin in the erythrocytes of the pipefish is readily induced by slowly drying, in the air, rather thick smears of blood, and is most uniformly obtained when the humidity is relatively high. Preceding the appearance of definite crystals the cells lose their typical oval form and show an increasing tendency towards angularity. The majority finally assume a triangular shape but some become rhomboidal. However, at this stage the hemoglobin gives no evidence of crystal formation. Soon well-defined clefts appear in the cell contents and definite crystals then appear. The number of crystals formed in individual cells is subject to some variation. Three crystals, forming the three sides of a triangle with the nucleus in the center, are most commonly encountered. Occa- sionally four crystals forming a rhomboidal figure, and more rarely two crystals arranged parallel with the long axis of the cell are present (Fig. 1). Frequently a variable number of very small, slender crystals may be associated with larger ones. They usually lie irregularly about the nucleus. The size of the larger crystals is also somewhat variable, but the shape is relatively constant. Practically all are 492 CRYSTALLIZATION OF HEMOGLOBIN 493 modified on the side next the cell membrane, being rounded rather than straight. In addition many are notched on the inner side especially if they are in contact with the surface of the nucleus. DISCUSSION Little is known of the factors involved in maintaining the hemo- globin within the red blood cell in solution. In the present instance the only obvious cause of the crystallization of the hemoglobin in the ,.; FIG. 1. Four selected areas from a preparation of pipefish blood showing the characteristic numbers, size, form, and position of hemoglobin crystals within the erythrocytes. The turbidity of the background is caused by the laking of many cells due to the injury produced in transferring the cells from the slide to a coverslip in order to obtain a preparation thin enough to photograph. Magnification X 1150. erythrocytes of the pipefish is the slow withdrawal of water with whatever attendant injuries that may occur when drying takes place. It is of interest to note that the cell is deformed by the changed orientation of the hemoglobin molecules before any change in the nature of the hemoglobin can be observed with ordinary transmitted light and that the form acquired by the cell, triangular or rhomboidal, foreshadows the appearance of three or four major crystals within it. Moreover, the clefts which mark the amount of hemoglobin to be apportioned to each crystal also become evident while the hemoglobin still appears unmodified. One of the striking features of all erythro- 494 ALDEN B. DAWSON cytes is their tendency to return to their specific form after deforma- tion, but in the case of incipient crystallization the shift in orientation of the hemoglobin molecules is sufficient to produce a permanent distortion. In the case of Necturus, previously described, and in the pipefish the crystals of hemoglobin are large and relatively few. In other instances that have come under my observation while studying supra- vitally the erythrocytes of many vertebrates, the crystallization of hemoglobin has been quite different, the crystals being numerous and very small, producing a granular effect. Such crystallization has been encountered on a few occasions in Necturus as well as in another urodele, Eurycea bislineata. It has also been noted in several fishes such as the common mackerel, menhaden, alewife, and sea bass. In all of these cases the cause of the crystallization was unknown and appeared irregularly in preparations of fresh blood. In a review of the literature one finds few references to intracellular crystallization of hemoglobin. Guerber (1927) observed it in the erythroblasts of embryos of the pig and cow. Kranz (1928) described crystals in mammalian erythrocytes after fixation with potassium bichromate and acetic acid, followed by paraffin imbedding. Celloidin imbedding gave negative results. He believed that the crystals were not pure hemoglobin but a product resulting from the reaction of hematin with the chromic and acetic acids. The work of Kranz was subsequently repeated by Tschachmachtschian (1932) who concluded that the crystals described by Kranz were entirely an artefact, the result of paraffin imbedding, and were not directly related to the hemoglobin content of the erythrocytes. Jokl (1925), while studying fresh preparations of skate's blood, observed certain erythrocytes in which the cell content was divided obliquely by two or three peculiar light stripes. These light stripes appear comparable to the clefts which appear in the hemoglobin of the red cells of the pipefish, preceding the appearance of the large crystals. Intracellular crystallization of hemoglobin was encountered in certain teleosts by Yoffey (1929), although he failed to recognize it as such. He states: " In the Gadus group the erythrocytes may assume a very curious shape. At first round, they then become oval, as in other fishes. They then show an increasing tendency towards angu- larity, and finally may become perfectly triangular in shape (Fig. 15). The relative proportion of triangular to oval red blood corpuscles varies from one animal to another. The illustration shown is from a blood film of Gadus minutus in which the majority of the erythrocytes CRYSTALLIZATION OF HEMOGLOBIN 495 are triangular. On the other hand there are many cases in which only a few of the corpuscles are triangular, and the majority are of the normal shape. The triangularity is not artificially produced by the fixative because it may be observed in specimens of perfectly fresh and unfixed blood, though the angles may not be sharp as in the fixed film." (p. 336.) From Yoffey's description it is obvious that in Gad us minntns he was dealing with intracellular crystallization and, in the photograph reproduced in his Fig. 15, three large crystals are clearly seen in almost every cell. Apparently in the Gadus group crystallization of hemoglobin occurs as readily as in the pipefish and was induced by Yoffey unconsciously by slight variations in his technique. SUMMARY Crystallization of hemoglobin within the erythrocytes of the pipe- fish is described. This phenomenon is readily produced by slow drying, especially in a humid atmosphere. Preceding the appearance of the definitive crystals the erythrocytes lose their characteristic oval form and become angular, triangular and rhomboidal forms predominating. Then definite clefts appear, fol- lowed soon after by the appearance of typical crystals. The number of crystals within individual cells varies. Two, three, and four large crystals are most commonly encountered, but a more variable number of minute needle-like forms may also be present in the erythrocyte. It is of interest to note that the erythrocytes exhibit deformation due to the changing orientation of the hemoglobin molecules before any evidence of crystal formation can be detected with ordinary transmitted light. LITERATURE CITED DAWSON, A. B., 1930. Changes in the Erythrocytes of Necturus Associated with the Intracellular Crystallization of Hemoglobin. Anat. Rec., 46: 161. GUERBER, A., 1927. Endocytare Haemoglobinkristalle. Sitzungsber. Ges. Be/order, ges. Naturwiss. Marburg, 62: 294. JOKL, A., 1925. Uber vitalfarbbare Erythrozytengranulationen ("Substantia meta- chromatico-granularis") beim Rochen, nebst weiteren Bemerkungen iiber das Blut dieser Tiere. Zeitschr. f. mikr. anat. Forsch., 2: 461. KRANZ, H. \V., 1928. Kristallbildungen im Innern cler roten Blutkorperchen. Zeitschr. f. Biol., 87: 258. TSCHACHMACHTSCHIAN, H., 1932. Uber die von H. W. Kranz beschriebenen Kristallbildungen im Inneren der roten Blutkorperchen. Zeitschr. f. Zellforsch. u. mikr. Anat., 15: 114. YOFFEY, J. M., 1929. A Contribution to the Study of the Comparative Histology and Physiology of the Spleen, with Reference Chiefly to its Cellular Con- stituents. I. In Fishes. Jour. Anat., 63: 314. SALT REQUIREMENTS AND SPACE ORIENTATION OF THE LITTORAL ISOPOD LIGIA IN BERMUDA1 T. CUNLIFFE BARNES OSBORN ZOOLOGICAL LABORATORY, YALE UNIVERSITY INTRODUCTION The important problem of the invasion of terrestrial or fresh-water habitats by marine organisms has received considerable attention in recent years (cf. Schlieper, 1929; Pearse, 1929; Pantin, 1931). The Isopoda extend from great depths of the ocean to terrestrial and fresh- water environments and should afford especially interesting material for these investigations. Tait (1916) in particular has studied the longevity of the littoral form, Ligia oceanica, in dilutions of sea water. The structure of Ligia has been described in detail in the monograph of Hewitt (1907). The present paper deals chiefly with the effect of changes in the salt content of sea water on Ligia baudiniana Milne-Edwards, the common isopod in Bermuda. HABITS Ligia baudiniana may be found in great numbers on the rocky shore (Verrill, 1903), especially in the intertidal zone at low tide. L. oceanica (Tait, 1916) sometimes remains covered with the tide but L. baudiniana retreats as the water advances. When isolated on stones in tidal pools, the isopods will run from one side of the rock to the other as if seeking a dry pathway to the shore. Occasionally I have observed them leaping from one stone to another to avoid the water. However, they are dependent on the sea water to keep the surface of the gills moist and we have never found specimens more than seventy feet from the sea; they appear in great numbers on rocks and walls several feet from the sea on rainy days. If placed in a terrarium containing a small pool of sea water, they will crawl to the edge of the water, turn around, and dip the ends of theuropodal spines in the water. By bringing the spines close together and altering the distance between the spines in a slow rhythm, the isopod moistens the gills with the water which rises between the spines by capillarity (Fig. 1). In this way a large drop of water may form on the gills. 1 Contribution from the Bermuda Biological Station for Research. 496 SALT REQUIREMENTS OF LIGIA 497 These spines, which are usually long, are also used as swimming fins and feelers as in other species of Ligia. It is probable that Ligia enters the sea to release the young from the brood pouch for this occurred only in submerged specimens. An examination of the gut contents revealed vegetable debris and unicellular algae. According to Hewitt (1907) L. oceanica feeds largely on decaying vegetable matter. L. exotica is described as omnivorous by Pearse (1931). Although Pearse (1929) is undoubtedly correct in pointing out that food supply alone is probably not the most important factor in determining the habitat of littoral forms, yet it is of interest to note that Ligia baudiniana subsists largely on the green coating of rocks in the intertidal zone. FIG. 1. Ligia baudiniana wetting the gills by the capillary action of the uropodal spines which are dipping into the sea. CONCENTRATION EFFECTS As a basis for comparison for subsequent experiments, the longevity of the isopods was first determined in sea water, air, and in fresh water. Individual specimens, carefully collected to avoid injury, were placed in finger bowls containing 100 cc. of water or solution. The average duration of life was only four hours in distilled water, seven and one-half hours in fresh water and thirty-four hours in sea water (Table I). The maximum longevity is also given in the tables. Sea water, changed every twenty-four hours, gave an average duration of life of fifty-eight hours, and in running aerated sea water the average longevity was one hundred and ninety-two hours, but the maximum was recorded for unchanged sea water. The large surface exposed to the air in the finger bowls permitted considerable diffusion of O2 and CC>2 as is indicated by the maximum of twelve and one-half days in unchanged sea water. In dry glass dishes the average duration (and the maximum) was eleven hours, but in bowls containing damp sand the isopods lived for very long periods (fifteen days). It is clear that moist air is a far more favorable medium than sea water. Ligia dies in about one hour in dry air in the sun at 30-37° C. In sea 498 T. CUNLIFFE BARNES water the isopods survive overnight at a temperature of 5° C., but the gills cease beating at 15° C. The life of Ligia in sea water is markedly curtailed by dilution below 50 per cent or by doubling the concentration of salts (Table II). In dilute sea water there is a slight increase in the frequency of gill TABLE I Longevity of Ligia in Air and in Water Medium Average Duration of Life Maximum No. of Specimens Tested Distilled water hours 4 5 21 Fresh water t\ 8 6 Sea. water 34 297 93 Sea water (changed daily) 58 198 20 Running sea water 83 192 26 Air 11 12 8 Air over damp sand 360 625 31 TABLE II Concentration Effects on Ligia Medium Average Longevity Maximum No. of Specimens 25% sea water hours 6 10 6 50% sea water 20 70 7 75^%) sea water 51 172 8 200% sea water 3 7 5 250% sea water 1 1 4 400% sea water 1 2 3 M/2 glycerine 7 10 5 AI glycerine 5 12 13 M/4 glycerine in sea water 11 24 8 M/' glycerine in sea water 9 19 6 A I glycerine in sea water 4 5 6 2 M glycerine in sea wrater 2 2 3 movements. The brief existence of the isopod in distilled water is not due to the decreased osmotic pressure, for the addition of glycerine has little beneficial effect (Table II). On the other hand, the death of Ligia in concentrated sea water appears to be due largely to osmotic factors as will be seen in the records of rapid death for sea water in which the osmotic pressure was increased by glycerine. SALT REQUIREMENTS OF LIGIA 499 SPECIFIC ION EFFECTS Of solutions containing a single salt isosmotic with Bermuda sea water (5/8 M), the isopods lived longest in NaCl, eight hours, and CaCl2, seven hours; while the average duration of life in MgCU and KC1 was only four and one and one-half hours respectively (Table III). KC1 exerted an immediate paralyzing effect on the gills which normally began to vibrate as soon as the animal was immersed in any of the solutions mentioned in this paper except KC1. The fre- quency of gill movements was taken at intervals in all solutions, but no vibrations were ever observed in KC1 although the animals ap- peared quite normal for the first half hour. The average time for ten beats was three and one-fifth seconds in sea water (27°); and approaching death was indicated when this increased to five seconds. Immature specimens (12-15 mm.) had a faster rate, two and one-tenth seconds, and were not used; in addition it was found that immature specimens showed greater resistance to all solutions tested. TABLE III Specific Ion Effects on Ligia Solution Average Length of Life Maximum No. of Specimens 5/8 M NaCl hours 8 14 12 5/8 M CaCl* 7 13 19 5/8 M MgClo 3 5 7 2 5/8 M MgSO4 4 4 10 1 M MgSO4 3 4 8 5/8 M KC1 H 3 10 IONIC ANTAGONISM Combinations of two ions were tried in various proportions, but no satisfactory solution was found. No specimens lived for more than a very few hours in binary mixtures but some antagonism seemed evi- dent between Na and Ca. In solutions containing Mg or K the longevity seemed to be controlled by the amount of the most toxic ion present. In artificial sea water (i.e., 100 NaCl, 11.6 MgSO4, 2.2 KC1 and 2.5 CaCl2 in 5/8 M cone.) the duration, forty hours, com- pared very favorably with natural sea water. In artificial sea water containing no magnesium, the same average duration of life was exhibited (Table IV). If the KC1 was omitted, the isopods lived for twenty hours; in the absence of Ca, fourteen hours; and they died within two hours in artificial sea water containing no Na. 500 T. CUNLIFFE BARNES The next step was to determine how long Ligia would survive if the concentration of individual ions were increased in sea water. In the case of Na, Ca and Mg concentrations not exceeding M/8 made up in sea water (i.e., one-eighth of the molecular weight added to a liter of sea water) did not exert an appreciable toxic effect but KC1 showed a limiting concentration of M/10 in sea water (Table V). TABLE IV Longevity of Ligia in Antagonistic Solutions Solution Average Length of Life Maximum No. of Specimens Artificial sea water hours 40 123 8 Same without Na 2 2 5 Same without Ca 14 23 10 Same without K 20 28 9 Same without Mg 41 120 22 TABLE V Effect of Increasing Concentration of Ions in Sea Water Solution Average Longevity Maximum Number of Specimens 5/8 M NaCl made in sea water hours 9? 36 10 1/4 M NaCl made in sea water 11 27 5 1/6 M NaCl made in sea water 6 11 8 1/8 M NaCl made in sea water 57 244 5 2.5/8 M Cad-) made in sea water H 2| 9 1/8 M CaCl<> made in sea water 88 248 8 2.5/8 M MgCl2 made in sea water 3 6 5 1/8 M MgSO4 made in sea water 91 258 5 5/8 M KC1 made in sea water 2 3 13 1/4 M KC1 made in sea water . . . 3| 5 9 1/6 M KC1 made in sea water . . 5 5 9 1/8 M KC1 made in sea water . ... 6| 16 9 1/10 M KC1 made in sea water 31 201 14 The gills were completely inhibited in 5/8 M KC1 in sea water but exhibited the usual rhythm in M/4 KC1 in sea water. In all these solutions in which salts were added to sea water it is probable that the increased osmotic pressure was significant judging from the short life of isopods in sea water containing glycerine (Table II). SPACE ORIENTATION Ligia bandiniana inhabits a very restricted zone along the shore line, and a number of experiments were performed to discover what SALT REQUIREMENTS OF LIGIA 501 tropisms or other reactions restricted the distribution of the isopod. One may mention first the inability to survive in sea water or in dry air and the presence of food (unicellular algae) on the intertidal rocks. These facts, however, do not explain the curious ability of the animal to orient towards the sea when released a short distance from the shore. It was noted that the isopods appeared to be reacting to the inclination of the land sloping gradually to the sea and it was found that under controlled conditions (in a photographic dark room under dim red illumination) pronounced geotropic orientation was exhibited. TABLE VI Orientation of Ligia on Slopes near the Sea Date Temp. Direction of Slope Inclina- tion Num- ber Re- leased Direction of Creeping 0 C. degrees July 10 27 Towards sea 30 6 5 down 11 27 Towards sea 40 2 down 11 27 Away from sea 40 2 down Aug. 1 26 Towards sea 40 9 7 down 1 up 1 went up but turned 12 27 Away from sea 30 6 5 down 14 25 Towards sea 40 5 down 14 25 Right angles to sea 60 4 3 down 14 25 Away from sea 20 4 3 down 1 up 14 25 Away from sea 45 4 2 straight down 3 down obliquely 24 28.6 Away from sea 60 4 3 down 25 27 Towards sea 50 3 down 25 27 Away from sea 20 5 4 down 25 27 Away from sea 30 10 4 down 6 up To test this hypothesis, specimens were released at various distances from the sea on ground (sand, grass, or rocks) sloping towards and also away from the sea (Table VI). Of sixty-four isopods tested, fifty showed positive geotropic orientation and crawled downward regardless of the direction of the sea. They also showed a less pro- nounced tendency to crawl in the direction of open patches of sky and exhibited positive phototropism under controlled conditions. When isopods are released in the sea at a distance not greater than eight feet from shore, they swim energetically to shore or crawl over the bottom directly to shore. In general, the animal crawls over the bottom and seldom exhibits the typical swimming movements unless 32 502 T. CUNLIFFE BARNES in deep water. The cause of this orientation to shore in the sea is unknown; it is independent of currents or the direction of the sun. The animal is negatively rheotropic and will swim against currents in an aquarium even after the removal of antennae and uropodal spines, but this has nothing to do with the shoreward orientation. Luther (1930) has shown recently that the antennules of crabs are receptors for rheotropism, but in Ligia the antennules are extremely small (cf. Hewitt, 1907), and it seems probable that currents in the water stimulate the legs. DISCUSSION Ligia affords a striking example of a marine organism which is invading the land through the intertidal zone — an approach to terres- trial life which has not received sufficient attention in theoretical considerations of the evolution of land animals. In tropical islands, lacking fresh water, and where there is no pronounced temperature difference between sea and air, the intertidal zone becomes an impor- tant route for the invasion of the land (Pearse, 1929). Ligia bau- diniana contrasts sharply with L. oceanica which, according to Tait (1916), may live over eighty days in sea water. However, like the beach crab Ocypode, L. baudiniana is dependent on sea water to keep the gills moist for aerial respiration. It also resembles the beach crab in its inability to withstand fresh water or diluted sea water, and is thus quite different from the marine Gammariis, which lives long periods in sea water diluted to .5 per cent (Adolph, 1925). The death of Ligia in distilled water and in glycerine solutions indicates that, like Gammarus (Loeb, 1903), loss of essential salts is more detrimental than osmotic disturbances in the medium. In spite of its terrestrial life, Ligia baudiniana is clearly a poikilosmotic form, although in diluted sea water the respiratory rate increases, which on Schlieper's (1929) theory might be due to osmotic work performed in partially resisting disturbance of the water and salt equilibrium. The order of toxicity of single ions, K > Mg > Ca > Na, appears to be about the same for several Crustacea, i.e., Gammarus (Adolph, 1925), Artemia (Martin and Wilbur, 1921), Daphnia (Berger, 1929), Cambarns (Helff, 1929), and is the reverse of the toxicity series for the egg of the sea urchin (Page, 1929). The rapid death in KC1 is prob- ably due in part to the lack of ventilation of the gills, which are unable to move in this solution. Zoond (1931) has shown that ventilation of the gill surface is of vital importance in Crustacea due to the extremely slow rate of diffusion of C>2 in water. The recent experiments of Bialaszewicz (1932) have demonstrated that the high toxicity of KC1 for Crustacea is associated with its rapid disappearance from the blood SALT REQUIREMENTS OF LIGIA 503 nto the tissues. According to Loeb (1903) Na, K, and Ca are neces- sary for the gill movements of Gammarus, but it resembles Ligia in certain other salt requirements, i.e., there is no satisfactory binary mixture and Mg appears to be a dispensable ion. The orientation of Ligia to the sea resembles that of young logger- head turtles described by Parker (1922). The orientation of the iso- pod to the shore, when in the sea, appears to be an instance of definite orientation which is not governed by a simple type of tropistic response and is not unlike the orientation of Onchidium (Arey and Crozier, 1921), and of ants (Barnes, 1929). The fact that the orientation of Ligia in the sea is not exhibited at distances greater than ten feet from the shore precludes the possibility that the isopod reacts to the blue color of deeper water as in the loggerhead turtle (Hooker, 1910). SUMMARY 1. Ligia baudiniana moistens its gills by the capillary action of the uropodal spines. 2. The isopod survives best in damp air and is unable to live for long periods in sea water. 3. Changes in the concentration of sea water are detrimental, i.e., Ligia is poikilosmotic. 4. The cations of sea water are toxic in the following order: K > Mg > Ca > Na. K exerts a specific paralyzing effect on the gill movements. 5. On land Ligia orients toward the sea. Positive geotropism appears to be the most important factor. I am greatly indebted to my co-worker, Mr. Frank Gilchrist, who performed most of the early experiments. It is a pleasure to express my gratitude to Dr. J. F. G. Wheeler who placed the facilities of the Bermuda Station at my disposal. CITATIONS ADOLPH, E. F., 1925. Some Physiological Distinctions between Fresh Water and Marine Organisms. Biol. Bull., 48: 327. AREY, L. B., AND W. J. CROZIER, 1921. On the Natural History of Onchidium. Jour. Ex per. Zool., 32: 443. BARNES, T. C., 1929. The Positive Geotropic Orientation of an Ant (Crematogaster lineolata). Jour. Gen. Psychol., 2: 517. BERGER, E., 1929. Unterschiedliche Wirkungen gleicher lonen und lonengemische auf verschiedene Tierarten. Pfliigers Arch., 223: 1. BIALASZEWICZ, K., 1932. Sur la regulation de la composition minerale de 1'hemo- lymphe chez le Crabe. Archiv. Internal, d. Physiol., 35: 98. HELFF, O. M., 1929. Toxic and Antagonistic Properties of Na, K, Mg and Ca Ions on Duration of Life of Cambarus clarkii. Proc. Soc. Exper. Biol. Med., 26: 797. 504 T. CUNLIFFE BARNES HEWITT, C. G., 1907. Ligia. L. M. B. C. Memoirs. HOOKER, D., 1910. Certain Reactions to Color in the Young Loggerhead Turtle. Carnegie Inst. Pub., No. 132: 69. LOEB, J., 1903. On the Relative Toxicity of Distilled Water, Sugar Solutions and Solutions of the Various Constituents of the Sea Water for Marine Animals. Univ. Calif. Pub. Physiol., 1: 55. LUTHER, W., 1930. Versuche iiber die Chemorezeption der Brachyuren. Zeitschr. f. vergl. Physiol., 12: 177. MARTIN, E. G., AND B. C. WILBUR, 1921. Salt Antagonism in Artemia. Am. Jour. Physiol., 55: 290. PAGE, I. H., 1929. The Toxicity of Monovalent and Divalent Cations for Sea Urchin Eggs. Biol. Bull., 57: 449. PANTIN, C. F. A., 1931. The Adaptation of Gunda ulvae to Salinity. Brit. Jour. Exper. Biol., 8: 63. PARKER, G. H., 1922. The Crawling of Young Loggerhead Turtles toward the Sea. Jour. Exper. Zool., 36: 323. PEARSE, A. S., 1929. Observations on Certain Littoral and Terrestrial Animals at Tortugas, Florida, with Special Reference to Migrations from Marine to Terrestrial Habitats. Papers Tortugas Lab., Carneg. Inst., 26: 207. PEARSE, A. S., 1931. The Ecology of Certain Crustaceans on the Beaches at Misaki, Japan, with Special Reference to Migrations from Sea to Land. Jour. Elisha Mitchell Sci. Soc., 46: 161. SCHLIEPER, C., 1929. Uber die Einwirkung niederer Salzkonzentrationen auf marine Organismen. Zeitschr. f. vergl. Physiol., 9: 478. TAIT, J., 1916-17. Experiments and Observations on Crustacea: Part I. Immer- sion Experiments on Ligia. Proc. Roy. Soc. Edin., 37: 50. VERRILL, A. E., 1903. Zoology of the Bermudas. New Haven. ZOOND, A., AND E. CHARLES, 1931. Studies on the Localisation of Respiratory Exchange in Invertebrates. I. The respiratory mechanism of the fresh- water crab Potamonautes. Jour. Exper. Biol., 8: 250. THE RELATION BETWEEN ABSORPTION AND ELIMINA- TION OF WATER BY TERMOPSIS ANGUSTICOLLIS S. F. COOK AND K. G. SCOTT (From the Division of Physiology, University of California Medical School, Berkeley, California) The purpose of this investigation was to study the water relations of the termite from the point of view of the animal itself. The question has been of interest in its ecological aspects with reference to the cli- matic conditions favorable for termite growth and to moisture and dryness as factors determining the distribution of species. As far as we are aware no studies have been undertaken to determine under controlled conditions from what source the termite derives its water, how it undergoes water loss, and how much drying it can suffer without ill effect. These questions are considered in the data here presented. The organism used was Tennopsis angusticollis, the common large wood termite of the Pacific Coast. The termites were collected in the vicinity of Berkeley and kept in jars of moist rotten wood until needed. The first step was to determine the normal water content of Termopsis and establish the relation between the partial pressure of the water vapor in the atmosphere surrounding the termites and their water content. Experiment 1. Three groups of 50 termites each were placed in desiccators. In the desiccator with Group I was placed a vial con- taining water. At equilibrium, therefore, the relative humidity was 100 per cent and the partial pressure of water vapor from 15 to 17 mm. Hg, since the temperature varied from 17° to 20° C. It was not possible to maintain an absolutely constant temperature, but the variation of approximately 3° was too slight to affect the validity of the result. In the second group the vapor pressure was lowered to 7 to 9 mm. Hg, the variation again being dependent on the temperature. This was accomplished by placing in the desiccator with the termites a saturated solution of calcium chloride over solid calcium chloride. The vapor pressure should theoretically be of the order of magnitude mentioned. This was checked by evacuating a vessel containing a similar mixture of solid and dissolved calcium chloride and allowing the system to come to equilibrium, the actual pressure of the water vapor being read on a mercury manometer. The use of this mixture 505 506 S. F. COOK AND K. G. SCOTT is also of advantage since the absorption of water from the termites does not alter the concentration of the dissolved calcium chloride when the solid phase is present in excess. Since the termites were transferred directly from conditions of water saturation to this tension, any effects due to the reduced tension should become immediately apparent. The third group of termites was placed in a desiccator with an- hydrous calcium chloride where the vapor tension was less than one millimeter. All three groups of termites were kept under the conditions ob- served above for two days. Food was provided in the form of punky wood which had been allowed in each case to come into equilibrium with the appropriate vapor tension before the experiment began. Starvation, therefore, was eliminated as a factor in reducing the weight TABLE I Water loss by termites at different vapor tensions. In each case the initial number was fifty. A. Group I Group II Group III B Vapor tension in mm Hg 15-17 7-9 0-1 C Average fresh weight in mg 42.4 29.25 34.5 D Average weight after experiment in nig 39.8 24.8 26.3 E Average dry weight after experiment in mg 11.1 5.9 8.0 C — D 6 2 15 3 23 1 G. Water content of termites at beginning in per cent: C — F X 100 73.8 79.9 76.8 C H. Water content of termites at end of experiment: D ~ E x 100 72.2 76.1 69.6 C of the termites. The termites were weighed before and after the exposure to the different degrees of humidity and the results expressed as average weight per termite. This was permissible because there was a sufficiently large number of animals in each case (50) to justify an average. Finally the dry weight was determined. The data are summarized in Table I. From this table it is evident that the percentage of water lost is inversely proportional to the water vapor tension. Furthermore, the water content and the average weight appear to stand in inverse ratio. This probably can be explained by the fact that the older insects have developed wing pads and a thicker integument, thus lowering the relative water content. Nevertheless the extreme variation in relative water content which may be ascribed to this factor is of the order of 6 WATER RELATIONS OF TERMITE 507 per cent (Line G, Table I), whereas the extreme observed variation is approximately 17 per cent (Line F, Table I). There is little doubt, therefore, that the water loss is substantially proportional to the decrease in vapor tension. It is highly probable that most of this water is lost through evaporation from the tracheae and the body surface. Experiment 2. In order to check the above results, an experiment was performed in which the water actually given off by the termites was compared with the loss in weight of the animals. Fifty termites were placed in a test tube through which dry air was passed for 212 hours. Before reaching the animals the air was dried thoroughly by passing it through ten feet of one-inch glass tubing filled with anhy- drous calcium chloride and then through a flask of pure, concentrated sulphuric acid. The water and carbon dioxide given off by the termites was absorbed by anhydrous calcium chloride of the finest mesh obtain- able, packed tightly in a medium-sized U-tube. Since the total quantity of carbon dioxide and water produced was small, and since the rate of flow was very slow, there can be little doubt that the absorption by the calcium chloride was practically complete. The carbon dioxide and water were not determined separately because the average carbon dioxide production of Termopsis has already been determined by one of us (Cook, 1932). According to the data pre- sented in the paper mentioned, the average production is approxi- mately 8.9 milligrams per gram termite per hour. The calculated value of the carbon dioxide produced during any time interval could then be deducted from the gain in weight of the calcium chloride with the reasonable assurance that the balance represented water. The initial weight and that at the end of the experiment were obtained, also the weight of food consumed (filter paper) and the weight of the feces (see Table II). Lender the conditions described, viz., a very slow current of air, the water loss proceeded at a fairly regular but diminish- ing rate until the effects of the drying were apparent in the behavior of the termites. The effect of rapid drying is described elsewhere. The total gain in weight of the calcium chloride in 212 hours was 628 milligrams, of which 150 milligrams (0.369 milligram per hour x 212 hours x 1,910 milligrams) may be ascribed to carbon dioxide, leaving 478 milligrams as the weight of the water. Meanwhile the loss in weight of the termites was 369 milligrams. The excess water found may be accounted for in two ways. In the first place, since the diet was almost exclusively cellulose, and had been for weeks previous, and since the R.Q. under such conditions is very nearly unity (Cook, 1932), the predominant oxidation must have been that of carbohydrate. This should, of course, yield one molecule of 508 S. F. COOK AND K. G. SCOTT water to each molecule of carbon dioxide, and if the total weight of carbon dioxide was 150 milligrams, as suggested above, the corre- sponding weight of the water produced would be 61 milligrams. In the second place, the three termites which died and were eaten must have contained, on the basis of the data presented with Experiment 1, about 75 per cent of water, or 90 milligrams. This water must have appeared in the calcium chloride tube directly by evaporation from the three termites while living or dead, or have been consumed by the others. In the latter case it would have been lost again from the surviving animals before the completion of the experiment. These two sources of water combined would furnish, therefore, a maximum of 150 milligrams, which when subtracted from the total of 478 milligrams found equals 328 milligrams. This is comparable with the net loss in weight of 369 milligrams. TABLE II Water Loss by Fifty Termites in a Slow Current of Dry Air Initial total weight Final weight Filter paper consumed Feces Water lost in indicated intervals Total water lost at end of indicated periods mg. mg. mg. mg. mg. mg. 2,030 1,541 105 92 83 in 19 hrs. 83 in 19 hrs. 83 in 23 hrs. 165 in 44 hrs. 43 in 14 hrs. 208 in 58 hrs. 29 in 10 hrs. 236 in 68 hrs. 47 in 24 hrs. 283 in 92 hrs. 78 in 48 hrs. 360 in 140 hrs. 77 in 48 hrs. 438 in 188 hrs. 40 in 24 hrs. 478 in 212 hrs. The water loss obtained in Group III of Experiment 1 was 23.1 per cent. In this experiment the loss as based on the water found in calcium chloride is 19.3 per cent, and as based on the loss in weight is 17.2 per cent. The results obtained by the two methods of estimation in Experiment 2 show a reasonably close correspondence. The differ- ence between the degree of water loss in Experiment 1 (23.1 per cent) and in Experiment 2 (19.3 and 17.2 per cent) is not sufficient to invali- date the conclusion that under prolonged conditions of dryness the termites lose water to the extent of approximately 20 per cent of their weight. Experiment 3. The water loss having been shown to be primarily through evaporation, the mode of intake was next investigated. There are but two possibilities. The water must be absorbed from the atmosphere or taken in with the food. If the termite can use WATER RELATIONS OF TERMITE 509 atmospheric water, it must be able to absorb the vapor from a fully saturated atmosphere. If it can supply its needs adequately in this way, it should be able to utilize very dry food. If it cannot, then at least some water must be presented in the liquid form with the diet. To differentiate between these two possible sources of water, three groups of 25 termites each were kept under different conditions of atmospheric and food moisture. All three groups were fed wood which had been dried in a desiccator over calcium chloride for several days prior to the experiment. Owing to the great power of cellulose to absorb water even from the most powerful drying agents known, the water content could not have been zero, but it was lowered sufficiently to demonstrate the inability of the termite to live on relatively dry food. It is obviously impossible to feed termites dry wood in a saturated atmosphere or damp wood in a completely dry atmosphere. Therefore the separation of food and atmospheric water had to be made in time. This was done as follows: Group I was kept 22 hours per day in satu- rated air without food, and was placed in a desiccator with dry air and dry food two hours. Group II was exposed to moist air 16 hours and to dry air and food 8 hours. Group III was exposed to each set of conditions 12 hours per day. In no case could the termites get much water from the food, but if it is possible for them to utilize dry food and atmospheric water, then at least Group I should be able to absorb as much water in 22 hours at a high humidity as it would lose during two hours at low humidity, and thus the normal water content should be maintained. But if they are wholly dependent on food moisture for their water intake, then all the groups might be expected to lose water irrecoverably during the period of exposure to dryness. The water loss was determined by weighing after four days, but the termites were kept under the same conditions until they died in order to secure information concerning their viability. The data are presented in Table III. It will be observed that the water loss in all groups is substantially the same and is of the order of magnitude observed in the two previous experiments (20 to 23 per cent).1 This seems to be the case even though the periods of feeding varied so that in Group I the termites had only two hours to lose water and twenty- two in which to regain it. But they lost as much water in four feedings, or a total of eight hours, as Group III, which was placed in dry air six times as long. 1 The effect of starvation in four days may be neglected, particularly in view of the fact that in Experiment 2 it was shown to play no significant role even in eight days. 510 S. F. COOK AND K. G. SCOTT The substantially equal loss of water with such widely varying total time of exposure to dryness as 8 and 48 hours presents an ap- parent anomaly. But this anomaly in itself shows that the water loss in dry air and dry food is cumulative and that intervening periods in damp air of as long as 22 hours per day do not tend to decrease the entire water loss incurred during the combined time of the dry feeding periods, even though this total time is as short as 8 hours. In other words, after losing water during a short exposure to dryness the termites cannot make up the loss by even a long sojourn in a highly saturated atmosphere. This view is further substantiated by the data presented in Experiment 4, below, from which it is evident that rapid drying extracts the major portion of the water in about eight hours. In Group III the desiccation of the animals probably reached TABLE III The Water Loss and Viability of Termites under Varying Degrees of Atmospheric and Food Moisture Group I Group II Group III Moist air Moist air Moist air 22 hours 16 hours 12 hours per day. per day. per day. Dry air Dry air Dry air and food and food and food 2 hours 8 hours 12 hours \verage initial weight in ing . . . 33.4 37.3 33.7 \verage weight after 4 days in mg. . . . 26.0 29.6 25.9 \Vater loss in per cent . . . . 22.4 20.8 23.1 Per cent mortality after days specified! 4 8 4 0 6 48 4 4 8 80 44 24 . 12 96 96 76 15 .... 100 100 92 at the end of the first 8 to 10 hours the same degree as that attained by Group I at the end of four days. But with Group 1 1 1 the remaining forty odd hours of dryness did not serve to make any material addi- tions to the water loss. Experiment 4 indicates that at this stage of drying water becomes very difficult to remove and it also must be remembered that the animals of Group III were consuming consider- able amounts of cellulose which, although dry, nevertheless must have contained a slight quantity of hygroscopic water. From these findings the conclusion seems to be justified that the termite cannot utilize the water in the atmosphere to replace that lost by evaporation. The viability data indicate that during prolonged exposures death occurs primarily as a result of drying, and secondarily, from the starvation which accompanies it. The course of mortality WATER RELATIONS OF TERMITE 511 was similar in all three groups but life was most prolonged in the twelve-hour group, and least prolonged in the two-hour group. Since the former had a longer feeding period, it seems possible that they received slightly more nourishment despite the general unavailability of the dried wood. Nevertheless, the fact that practically none sur- vived more than fifteen days is additional evidence that dry food, even when accompanied by long periods of atmospheric saturation, will not support these animals, and that the moisture in the atmosphere can function only to prevent evaporation and thus to diminish the water loss. If now we exclude the water vapor surrounding the termite as a source of water we are forced to conclude that normally the water is taken in with the food.2 Experiment 4. The uniformity in the water loss of the termites in Experiment 3, even with a 600 per cent variation in time of exposure to very dry air, raises the question of just how rapidly the loss of water occurs, for apparently the first 20 per cent of the water is lost within 8 TABLE IV Rate of Water Loss in Termites under Dry Conditions Hours in desiccator Average weight per termite Water loss Rate of water loss mg. per cent mg, per hour per termite 0 51.0 0 4 45.3 11.2 1.4 8 39.9 21.6 1.3 18 36.7 28.4 0.35 (Eight dead) hours. To make certain of this point 25 termites were placed in a desiccator over anhydrous calcium chloride and were weighed at frequent intervals. No food was given. From Table IV it may be seen that the total water loss rose as high as 28 per cent after 18 hours exposure, but that the termites were rapidly dying. Since there was no food, it was impossible for the termite to get water, even in traces from wood, but the sharp fall in water content to a 20 per cent loss in 8 hours fits in with the result obtained in the previous experiment. In fact, here the death rate was even more rapid than in Experiments - Concerning the form in which the water may be ingested it may be pointed out that in addition to "free" water in the intercellular spaces of wood the hygroscopic water may be of importance. In short leaf pine at a relative humidity of 100 per cent (but with no condensation) the amount of water held in this way is approxi- mately 30 per cent (Schorger, 1926). All forms of cellulose may absorb hygroscopic water up to a concentration of 17 per cent. The ability of the termite to exist in rela- tively dry places (but not completely dry) may depend upon the availability of this water after digestion of the cellulose in the gut of the termite. 512 S. F. COOK AND K. G. SCOTT 2 and 3 (18 to 20 hours compared with 10 to 15 days), a difference which again may be correlated with the presence or absence of small amounts of water contained in wood. Our observations indicate that the termite survives a water shortage quite well until the loss reaches about 20 per cent of the normal fresh weight. At this point the animals begin to become moribund. Death is certain when the loss reaches about 28 per cent. In conclusion, the chief results of this study may be briefly sum- marized. Termopsis loses water rapidly when placed in dry air, and unless the loss is compensated it may be fatal. The water for replace- ment is provided as liquid in the food. The animal is unable to take up water vapor actively from even a moisture-saturated atmosphere. LITERATURE CITED COOK, S. F., 1932. The Respiratory Gas Exchange in Termopsis nevadensis. Biol. Bull., 63: 246. SCHORGER, A. W., 1926. The Chemistry of Cellulose and Wood. McGraw-Hill, New York. (Cf. particularly p. 12.) INDEX ABSORPTION and elimination of water, relation between, in Ter- mopsis angusticollis, 505. ADDISON, WILLIAM H. F., and MAURICE N. RICHTER. A note on the thyroid gland of the swordfish (Xiphias gladius, L.), 472. Adrenaline, branchial responses to, in the eel, 327. Aggregating behavior of Ameiurus me- las, 258. ANDERSON, BERTIL GOTTFRID. The num- ber of pre-adult instars, growth, relative growth, and variation in Daphnia magna, 81. Annual report of Marine Biological Laboratory, 1. Anuran larvae, hydrogen-ion concentra- tion of blood of, during involution, 405. Ascidian egg, mosaic development of, 381. J£ ARNES, T. CUNLIFFE. Salt require- ments and space orientation of the littoral isopod Ligia in Bermuda, 496. BATEMAN, J. B. Sec Keys and Bate- man, 327. BERRILL, N. J. The mosaic develop- ment of the ascidian egg, 381. BODINE, JOSEPH HALL, and TITUS C. EVANS. Hibernation and diapause. Physiological changes during hi- bernation and diapause in the mud- dauber wasp, Sceliphron casmenta- rium (Hymenoptera), 235. BOWEN, EDITH S. Further studies of the aggregating behavior of Amei- urus melas, 258. ENTRAL body structure in Chse- topterus at metaphase, first cleav- age, after picro-acetic fixation, 149. Chiasmata, origin and behavior, in Chorthippus elegans and Hyacin- thus amethystinus, 357, 368. Chromosomes of artificially activated eggs of Urechis, 212. CLAUSEN, H. J. Rate of regeneration of partly histolyzed anuran tail skin, 129. Cleavage, production of, by suppression of polar bodies in artificially acti- vated eggs of Urechis, 218. COE, W. R. Sexual phases in the American oyster (Ostrea virgin- ica), 419. Colpidiutn campylum, experimental pro- duction of chains in, 187. COOK, S. F. The respiratory gas ex- change in Termopsis nevadensis, 246. COOK, S. F., and K. G. SCOTT. The re- lation between absorption and elim- ination of water by Termopsis an- gusticollis, 505. Crustacean eye hormone as vertebrate melanophore activator, 108. Crystallization, intracellular, of hemo- globin, in erythrocytes of northern pipefish, 492. Cyclotrichium meunieri, sp. nov., 74. PVAPHNIA magna, number of pre- adult instars, growth, relative growth, and variation, 81. DARLINGTON, C. D. The origin and be- havior of chiasmata. V. Chorthip- pus elegans, 357. DARLINGTON, C. D. The origin and be- havior of chiasmata. VI. Hyacin- thus amethystinus, 368. DAWSON, ALDEN B. Intracellular crys- tallization of hemoglobin in the erythrocytes of the northern pipe- fish, Syngnathus fuscus, 492. DAWSON, ALDEN B. The reaction of the erythrocytes of vertebrates, espe- cially fishes, to vital dyes, 48. Diabetes, pancreatic, influence of hypo- physectomy on, in dogfish, 477. Diapause in the mud-dauber wasp, 235. Diapause, physiological changes during, in Sceliphron caementarium, 235. 513 514 INDEX Dogfish, influence of hypophysectomy on pancreatic diabetes, 477. Dominance of two kidney allelomorphs in Habrobracon juglandis (Ash.), 372. Double monsters in pond snail, 400. DRAPER, JOHN W., and DAYTON J. ED- WARDS. Some effects of high pres- sure on developing marine forms, 99. Drosophila, rate of oviposition in, 456. Dyes, vital, reactions of erythrocytes of fishes to, 48. gDWARDS, DAYTON J. See Draper and Edwards, 99. Eel, branchial responses to pitressin and adrenaline in, 327. Elimination and absorption of water by Termopsis angusticollis, 505. Embryology of Urechis, 387. Erythrocytes, of fishes, reactions to vi- tal dyes, 48. EVANS, TITUS C. Sec Bodine and Evans, 235. Eye hormone, crustacean, as vertebrate melanophore activator, 108. pERRIS, J. C. A comparison of the life histories of mictic and amictic females in the rotifer, Hydatina senta, 442. FRY, HENRY J. Studies of the mitotic figure. I. Chaetopterus : central body structure at metaphase, first cleavage, after picro-acetic fixation, 149. Fundulus eggs, effects of high pressure on, 99. consequences of experi- mental production of chains in Col- pidium campylum, 187. Genetics and development, 337. GOLDSCHMIDT, RICHARD. The Fourth Reynold A. Spaeth Memorial Lec- ture. Genetics and development, 337. Growth and variation in Daphnia magna, 81. ABROBRACOX juglandis (Ash.), dominance of two kidney allelo- morphs, 372. Habrobracon, modification of traits in mosaics from binucleate eggs, 296. HELFF, O. M. Studies on amphibian metamorphosis. X. Hydrogen-ion concentration of the blood of anu- ran larvae during involution, 405. Hemoglobin, intracellular crystalliza- tion of, in erythrocytes of northern pipefish, 492. Hemolysis, rate of, of erythrocytes in hypotonic solutions of electrolytes, 224. HERRICK, EARL H. Mechanism of movement of epidermis, especially its melanophores, in wound healing, and behavior of skin grafts in frog tadpoles, 271. Hibernation and diapause in the mud- dauber wasp, 235. Hyacinthus amethystinus, origin and behavior of chiasmata in, 368. Hydatina senta, comparison of life his- tories of mictic and amictic females, 442. Hydrogen-ion concentration of blood of anuran larvae during involution, 405. Hypophysectomy, influence on pancre- atic diabetes of dogfish, 477. TN STARS, pre-adult, number of, in Daphnia magna, 81. Involution, hydrogen-ion concentration of blood of anuran larvae during, 405. J ACOBS, M. H., and ARTHUR K PARPART. Osmotic properties of the erythrocyte. V. The rate of hemolysis in hypotonic solutions of electrolytes, 224. J^EYS, ANCEL B., and J. B. BATE- MAN. Branchial responses to adre- naline and to pitressin in the eel, 327. Kidney allelomorphs, dominance of, in Habrobracon juglandis (Ash.), 372. KROPP, BENJAMIN. See Perkins and Kropp, 108. LACKEY, JAMES P. Oxygen defi- ciency and sewage Protozoa : with descriptions of some new species, 287. Light, effect on rate of swimming of larvae of mussel crab, 310. Ligia. salt requirements and space ori- entation in, 496. INDEX 515 Lymnsea columella Say, polyvitelline eggs and double monsters in, 400. ARINE Biological Laboratory, thirty-fourth report, 1. Melanophore activation in vertebrates by crustacean eye hormone, 108. Melanophores, induced by X-ray, com- pared with those existing in pat- terns as seen in Carassius auratus, 484. Melanophores, mechanism of movement, in wound healing, 271. Mictic and amictic females, comparison of life histories in Hydatina senta, 442. " Mitogenetic rays," critique of yeast- detector method, 113. Mitosis, studies on, central body struc- ture in Chsetopterus, 149. Monsters, double, in pond snail, 400. Mosaic development of ascidian egg, 381. Mosaics, modification of traits in, from binucleate eggs of Habrobracon, 296. Mussel crab, temperature and light ef- fects on rate of swimming, 310. EWBY. W. W. The early embry- ology of the echiuroid, Urechis, 387. (~)RIAS, OSCAR. Influence of hypo- physectomy on the pancreatic dia- betes of dogfish, 477. Orientation in Ligia in Bermuda, 496. Osmotic properties of erythrocytes, 224. Ostrea virginica, sexual phases in, 419. Oviposition, rate of, in the fruit fly, Drosophila, 456. Oxygen deficiency and sewage Protozoa, 287. pARPART, ARTHUR K. Sec Jacobs and Parpart, 224. PERKINS, EARLE B., and BENJAMIN KROPP. The crustacean eye hor- mone as a vertebrate melanophore activator, 108. Pinnotheres maculatus Say, influence of temperature and light on rate of swimming, 310. Pitressin, branchial responses to, in the eel, 327. Polarity of egg of Urechis caupo, 145. Polyvitelline eggs and double monsters in Lymnsea columella Say, 400. POWERS, PHILIP B. A. Cyclotrichium meunieri sp. nov. (Protozoa, Cili- ata) ; cause of red water in the Gulf of Maine, 74. Pressure, effects on developing marine forms, 99. Protozoa, sewage, new species, 287. REGENERATION, rate of, in partly histolyzed anuran tail skin, 129. Respiratory gas exchange in Termopsis nevadensis, 246. RICHARDS, OSCAR W., and G. WELL- FORD TAYLOR. " Mitogenetic rays " —a critique of the yeast-detector method, 113. RICHTER, MAURICE N. See Addison and Richter, 472. QALT requirements of Ligia, 496. Sceliphron csementarium (Hymen- optera), physiological changes dur- ing hibernation and diapause, 235. SCOTT, K. G. See Cook and Scott, 505. Sewage Protozoa, new species, 287. Sexual phases in American oyster, 419. SHAPIRO, HERBERT. The rate of ovi- position in the fruit fly, Drosophila, 456. SMITH, GEORGE MILTON. Melanophores induced by X-ray compared with those existing in patterns as seen in Carassius auratus, 484. SONNEBORN, T. M. Experimental pro- duction of chains and its genetic consequences in the ciliate pro- tozoan, C o 1 p i d i u m campylum (Stokes), 187. Space orientation of Ligia in Bermuda, 496. Spaeth Memorial Lecture, genetics and development, 337. SPEICHER, B. R. Dominance of two kidney allelomorphs in Habrobra- con juglandis (Ash.), 372. Swimming, rate of, in larvae of mussel crab, as affected by light and tem- perature, 310. Swordfish, thyroid gland of, 472. Syngnathus fuscus, intracellular crys- tallization of hemoglobin in erythro- cytes of, 492. 516 INDEX BAYLOR, G. WELLFORD. See Rich- ards and Taylor, 113. Temperature, effect of rate of swim- ming of larvse of mussel crab, 310. Termopsis angusticollis, relation be- tween absorption and elimination of water by, 505. Termopsis nevadensis, respiratory gas exchange in, 246. Thyroid gland of the swordfish, 472. Traits, modification of, in mosaics from binucleate eggs of Habrobracon, 296. TYLER, ALBERT. Chromosomes of arti- ficially activated eggs of Urechis, 212. TYLER, ALBERT. Production of cleav- age by suppression of the polar bodies in artificially activated eggs of Urechis, 218. TYLER, ALBERT. The polarity of the egg of Urechis caupo, 145. ^JRECHIS caupo, polarity of egg, 145. Urechis, chromosomes of artificially ac- tivated eggs of, 212. Urechis, early embryology of, 387. Urechis, production of cleavage by sup- pression of polar bodies in artifi- cially activated eggs, 218. , relation between absorp- tion and elimination of, by Termop- sis angusticollis, 505. WHITING, P. W. Modification of traits in mosaics from binucleate eggs of Habrobracon, 296. WINSOR, CHARLES P., and AGNES A. WINSOR. Polyvitelline eggs and double monsters in the pond snail Lymnsea columella Say, 400. Wound healing, mechanism of epidermal movement in frog tadpoles, 271. ^-RAY-INDUCED melanophores, compared with those existing in patterns as seen in Carassius aura- tus, 484. YEAST-DETECTOR method in mi- togenetic ray study, critique of, 113. Volume LXIII Number 1 THE BIOLOGICAL BULLETIN PUBLISHED BY THE MARINE BIOLOGICAL LABORATORY Editorial Board GARY N. CALKINS, Columbia University E. G. CONKLIN, Princeton University FRANK R. LlLLEE, University of Chicago E. N. HARVEY, Princeton University CARL R. MOORE, University of Chicago SELIG HECHT, Columbia University GEORGE T. MOORE, Missouri Botanical Garden LEIGH HOADLEY, Harvard University T. H. MORGAN, California Institute of Technology M. H. JACOBS, University of Pennsylvania G. H. PARKER, Harvard University H. S. JENNINGS, Johns Hopkins University W. M. WHEELER, Harvard University E. E. JUST, Howard University EDMUND B. WILSON, Columbia University ALFRED C. REDFIELD, Harvard University Managing Editor AUGUST, 1932 Printed and Issued by LANCASTER PRESS, Inc. PRINCE & LEMON STS. LANCASTER, PA. THE BIOLOGICAL BULLETIN THE BIOLOGICAL BULLETIN is issued six times a year. Single numbers, $1.75. Subscription per volume (3 numbers), $4.50. Subscriptions and other matter should be addressed to the Biological Bulletin, Prince and Lemon Streets, Lancaster, Pa. Agent for Great Britain: Wheldon & Wesley, Limited, 2, 3 and 4 Arthur Street, New Oxford Street, London, W.C. 2. Communications relative to manuscripts should be sent to the Managing Editor, Marine Biological Laboratory, Woods Hole, Mass., between May 1 and October 1 and to the Institute of Biology, Divinity Avenue, Cambridge, Mass., during the remainder of the year. INSTRUCTIONS TO AUTHORS Preparation of Manuscript. In addition to the text matter, manuscripts should include a running page head of not more than thirty-five letters. Footnotes, tables, and legends for figures should be typed on separate sheets. Preparation of Figures. The dimensions of the printed page (4% x 7 inches) should be borne in mind in preparing figures for publication. Draw- ings and photographs, as well as any lettering upon them, should be large enough to remain clear and legible upon reduction to page size. Illustrations should be planned for sufficient reduction to permit legends to be set below them. In so far as possible, explanatory matter should be included in the legends, not lettered on the figures. Statements of magnification should take into account the amount of reduction necessary. Figures will be reproduced as line cuts or halftones. Figures intended for reproduction as line cuts should be drawn in India ink on white paper or blue-lined coordinate paper. Blue ink will not show in reproduction, so that all guide lines, letters, etc. must be in India ink. Figures intended for reproduction as halftone plates should be grouped with as little waste space as possible. Methods of repro- duction not regularly employed by the Biological Bulletin will be used only at the author's expense. The originals of illustrations will not be returned except by special request. Directions for Mailing. Manuscripts and illustrations should be packed flat between stiff cardboards. Large charts and graphs may be rolled and sent in a mailing tube. Reprints. Authors will be furnished, free of charge, one hundred re- prints without covers. Additional copies may be obtained at cost. Proof. Page proof will be furnished only upon special request. When cross-references are made in the text, the material referred to should be marked clearly on the galley proof in order that the proper page numbers may be supplied. Entered October 10, 1902, at Lancaster, Pa., as second-class matter under Act of Congress of July 16, 1894. LANCASTER PRESS, INC. LANCASTER, PA. CONTENTS Page THIRTY-FOURTH REPORT OF THE MARINE BIOLOGICAL LABORATORY 1 DAWSON, ALDEN B. The Reaction of the Erythrocytes of Vertebrates, Especially Fishes, to Vital Dyes 48 POWERS, PHILIP B. A. Cyclotrichium meunieri Sp. Nov. (Protozoa, Ciliata); Cause of Red Water in the Gulf of Maine 74 ANDERSON, BERTIL GOTTFRID The Number of Pre-adult Instars, Growth, Relative Growth, and Variation in Daphnia magna 81 DRAPER, JOHN W., AND DAYTON J. EDWARDS Some Effects of High Pressure on Developing Marine Forms 99 PERKINS, EARLE B., AND BENJAMIN KROPP The Crustacean Eye Hormone as a Vertebrate Melanophore Activator 108 RICHARDS, OSCAR W., AND G. WELLFORD TAYLOR "Mitogenetic Rays" — A Critique of the Yeast- Detector Method 113 CLAUSEN, H. J. Rate of Regeneration of Partly Histolyzed Anuran Tail Skin 129 TYLER, ALBERT The Polarity of the Egg of Urechis caupo 145 Volume LXIII Number 2 THE BIOLOGICAL BULLETIN PUBLISHED BY THE MARINE BIOLOGICAL LABORATORY Editorial Board GARY N. CALKINS, Columbia University E. G. CONKLIN, Princeton University FRANK R. LlLLIE, University of Chicago E. N. HARVEY, Princeton University CARL R. MOORE, University of Chicago SELIG HECHT, Columbia University GEORGE T. MOORE, Missouri Botanical Garden LEIGH HOADLEY, Harvard University T. H. MORGAN, California Institute of Technology M. H. JACOBS, University of Pennsylvania G. H. PARKER, Harvard University H. S. JENNINGS, Johns Hopkins University W. M. WHEELER, Harvard University E. E. JUST, Howard University EDMUND B. WILSON, Columbia University ALFRED C. REDFLELD, Harvard University Managing Editor OCTOBER, 1932 Printed and Issued by LANCASTER PRESS, Inc. PRINCE & LEMON STS. LANCASTER, PA. THE BIOLOGICAL BULLETIN THE BIOLOGICAL BULLETIN is issued six times a year. Single numbers, $1.75. Subscription per volume (3 numbers), $4.50. Subscriptions and other matter should be addressed to the Biological Bulletin, Prince and Lemon Streets, Lancaster, Pa. Agent for Great Britain: Wheldon & Wesley, Limited, 2, 3 and 4 Arthur Street, New Oxford Street, London, W.C. 2. Communications relative to manuscripts should be sent to the Managing Editor, Marine Biological Laboratory, Woods Hole, Mass., between May 1 and October 1 and to the Institute of Biology, Divinity Avenue, Cambridge, Mass., during the remainder of the year. INSTRUCTIONS TO AUTHORS Preparation of Manuscript. In addition to the text matter, manuscripts should include a running page head of not more than thirty-five letters. Footnotes, tables, and legends for figures should be typed on separate sheets. Preparation of Figures. The dimensions of the printed page (4%x7 inches) should be borne in mind in preparing figures for publication. Draw- ings and photographs, as well as any lettering upon them, should be large enough to remain clear and legible upon reduction to page size. Illustrations should be planned for sufficient reduction to permit legends to be set below them. In so far as possible, explanatory matter should be included in the legends, not lettered on the figures. Statements of magnification should take into account the amount of reduction necessary. Figures will be reproduced as line cuts or halftones. Figures intended for reproduction as line cuts should be drawn in India ink on white paper or blue-lined coordinate paper. Blue ink will not show in reproduction, so that all guide lines, letters, etc. must be in India ink. Figures intended for reproduction as halftone plates should be grouped with as little waste space as possible. Methods of repro- duction not regularly employed by the Biological Bulletin will be used only at the author's expense. The originals of illustrations will not be returned except by special request. Directions for Mailing. Manuscripts and illustrations should be packed flat between stiff cardboards. Large charts and graphs may be rolled and sent in a mailing tube. Reprints. Authors will be furnished, free of charge, one hundred re- prints without covers. Additional copies may be obtained at cost. Proof. Page proof will be furnished only upon special request. When cross-references are made in the text, the material referred to should be marked clearly on the galley proof in order that the proper page numbers may be supplied. Entered October 10, 1902, at Lancaster, Pa., as second-class matter under Act of Congress of July 16, 1894. LANCASTER PRESS, INC. LANCASTER, PA. CONTENTS Page FRY, HENRY J. Studies of the Mitotic Figure. I. Chaetopterus : central body structure at metaphase, first cleavage, after picro-acetic fixation 149 SONNEBORN, T. M. Experimental Production of Chains and its Genetic Conse- quences in the Ciliate Protozoan, Colpidium campylum (Stokes) 187 TYLER, ALBERT Chromosomes of Artificially Activated Eggs of Urechis 212 TYLER, ALBERT Production of Cleavage by Suppression of the Polar Bodies in Artificially Activated Eggs of Urechis 218 JACOBS, M. H., AND PARPART, ARTHUR K. Osmotic Properties of the Erythrocyte. V. The rate of hemolysis in hypotonic solutions of electrolytes 224 BODINE, JOSEPH HALL, AND EVANS, TITUS C. Hibernation and Diapause. Physiological changes during hibernation and diapause in the Mud-dauber Wasp, Sceli- phron caementarium (Hymenoptera) 235 COOK, S. F. The Respiratory Gas Exchange in Termopsis nevadensis. . . . 246 BOWEN, EDITH S. Further Studies of the Aggregating Behavior of Ameiurus melas 258 HERRICK, EARL H. Mechanism of Movement of Epidermis, especially its Melano- phores, in Wound Healing, and Behavior of Skin Grafts in Frog Tadpoles 271 LACKEY, JAMES P. Oxygen Deficiency and Sewage Protozoa: with Descriptions of Some New Species 287 / WHITING, "]*. W. Modification of Traits in Mosaics from Binucleate Eggs of Habrobracon 296 WELSH, JOHN H. Temperature and Light as Factors Influencing the Rate of Swimming of Larvae of the Mussel Crab, Pinnotheres macu- latus Say 310 KEYS, ANCEL, AND BATEMAN, J. B. Branchial Responses to Adrenaline and to Pitressin in the Eel 327 Volume LXIII Number 3 THE BIOLOGICAL BULLETIN PUBLISHED BY THE MARINE BIOLOGICAL LABORATORY Editorial Board GARY N. CALKINS, Columbia University E. G. CONKLIN, Princeton University FRANK R. LlLLEE, University of Chicago E. N. HARVEY, Princeton University CARL R. MOORE, University of Chicago SELIG HECHT, Columbia University GEORGE T. MOORE, Missouri Botanical Garden LEIGH HOADLEY, Harvard University T. H. MORGAN, California Institute of Technology M. H. JACOBS, University of Pennsylvania G. H. PARKER, Harvard University H. S. JENNINGS, Johns Hopkins University W. M. WHEELER, Harvard University E. E. JUST, Howard University EDMUND B. WILSON, Columbia University ALFRED C. REDFIELD, Harvard University Managing Editor DECEMBER, 1932 Printed and Issued by LANCASTER PRESS, Inc. PRINCE & LEMON STS. LANCASTER, PA. THE BIOLOGICAL BULLETIN THE BIOLOGICAL BULLETIN is issued six times a year. Single numbers, $1.75. Subscription per volume (3 numbers), $4.50. Subscriptions and other matter should be addressed to the Biological Bulletin, Prince and Lemon Streets, Lancaster, Pa. Agent for Great Britain: Wheldon & Wesley, Limited, 2, 3 and 4 Arthur Street, New Oxford Street, London, W.C. 2. Communications relative to manuscripts should be sent to the Managing Editor, Marine Biological Laboratory, Woods Hole, Mass., between May 1 and October 1 and to the Institute of Biology, Divinity Avenue, Cambridge, Mass., during the remainder of the year. INSTRUCTIONS TO AUTHORS Preparation of Manuscript. In addition to the text matter, manuscripts should include a running page head of not more than thirty-five letters. Footnotes, tables, and legends for figures should be typed on separate sheets. Preparation of Figures. The dimensions of the printed page (4% x 7 inches) should be borne in mind in preparing figures for publication. Draw- ings and photographs, as well as any lettering upon them, should be large enough to remain clear and legible upon reduction to page size. Illustrations should be planned for sufficient reduction to permit legends to be set below them. In so far as possible, explanatory matter should be included in the legends, not lettered on the figures. Statements of magnification should take into account the amount of reduction necessary. Figures will be reproduced as line cuts or halftones. Figures intended for reproduction as line cuts should be drawn in India ink on white paper or blue-lined coordinate paper. Blue ink will not show in reproduction, so that all guide lines, letters, etc. must be in India ink. Figures intended for reproduction as halftone plates should be grouped with as little waste space as possible. Methods of repro- duction not regularly employed by the Biological Bulletin will be used only at the author's expense. The originals of illustrations will not be returned except by special request. Directions for Mailing. Manuscripts and illustrations should be packed flat between stiff cardboards. Large charts and graphs may be rolled and sent in a mailing tube. Reprints. Authors will be furnished, free of charge, one hundred re- prints without covers. Additional copies may be obtained at cost. Proof. Page proof will be furnished only upon special request. When cross-references are made in the text, the material referred to should be marked clearly on the galley proof in order that the proper page numbers may be supplied. Entered October 10, 1902, at Lancaster, Pa., as second-class matter under Act of Congress of July 16, 1894. CONTENTS Page GOLD SCHMIDT, RICHARD The Fourth Reynold A. Spaeth Memorial Lecture. Genetics and Development 337 DARLINGTON, C. D. The Origin and Behaviour of Chiasmata. V. Chorthippus elegans 357 VI. Hyacinthus amethystinus 368 SPEICHER, B. R. Dominance of Two Kidney Allelomorphs in Habrobracon juglandis (Ash.) 372 BERRILL, N. J. The Mosaic Development of the Ascidian Egg 381 NEWBY, W. W. The Early Embryology of the Echiuroid, Urechis 387 WINSOR, CHARLES P. AND AGNES A. Polyvitelline Eggs and Double Monsters in the Pond Snail Lymnaea columella Say 400 HELFF, O. M. Studies on Amphibian Metamorphosis. X. Hydrogen-ion concentration of the blood of anuran larvae during involution 405 COE, W. R. Sexual Phases in the American Oyster (Ostrea virginica) . . . 419 FERRIS, JOSEPHINE CAROLYN A Comparison of the Life Histories of Mictic and Amictic Females in the Rotifer, Hydatina senta 442 SHAPIRO, HERBERT The Rate of Oviposition in the Fruit Fly, Drosophila 456 ADDISON, WILLIAM H. F., AND MAURICE N. RICHTER A Note on the Thyroid Gland of the Swordfish (Xiphias gladius, L.) ... 472 ORIAS, OSCAR Influence of Hypophysectomy on the Pancreatic Diabetes of Dogfish 477 SMITH, GEORGE MILTON Melanophores Induced by X-Ray Compared with those Ex- isting in Patterns as Seen in Carassius auratus 484 DAWSON, ALDEN B. Intracellular Crystallization of Hemoglobin in the Erythro- cytes of the Northern Pipefish, Syngnathus fuscus 492 BARNES, T. CUNLIFFE Salt Requirements and Space Orientation of the Littoral Isopod Ligia in Bermuda . • 496 COOK, S. F., AND K. G. SCOTT The Relation between Absorption and Elimination of Water by Termopsis angusticollis 505 MBL WHOI LIBRARY blH 17IF M